Early Japanese Trade, Administration and Interactions with the West 9781912961078

Citation preview

EARLY JAPANESE TRADE, ADMINISTRATION AND INTERACTIONS WITH THE WEST

Louis Cullen, Blackrock, Co. Dublin, Autumn 2019

Renaissance Books Distinguished Asian Studies Scholars – Collected Writings. Vol. 3.

Early Japanese Trade, Administration and Interactions with the West ™

By

Louis Cullen Trinity College, Dublin

Renaissance Books Distinguished Asian Studies Scholars. Vol. 3. ISSN 2515-0626 EARLY JAPANESE TRADE, ADMINISTRATION AND INTERACTIONS WITH THE WEST

First published 2020 by RENAISSANCE BOOKS PO Box 219 Folkestone Kent CT20 2WP Renaissance Books is an imprint of Global Books Ltd © Louis Cullen, 2020 ISBN 978-1-912961-06-1 [Hardback] ISBN 978-1-912961-07-8 [e-Book] All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any electronic, mechanical or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without prior permission in writing from the publishers.

British Library Cataloguing in Publication Data A CIP catalogue entry for this book is available from the British Library

Set in Garamond 11 on 12 point by Dataworks Printed and bound in England by CPI Antony Rowe, Chippenham, Wilts

To the memory of Matsuo Tarō, Eileen Katō and Tsukahara Sueko

v

Contents ™

Foreword by Shunsuke Katsuta Acknowledgements & Transliteration of Japanese Introduction: The Route to Japanese Trade

vii xi xiii

Part I: Interactions –Ancient and Modern 1. 2. 3. 4. 5.

Sakoku, Tokugawa Policy, and the interpretation of Japanese history Knowledge and Use of Japanese by the Dutch on Dejima Island, Nagasaki Review of James W. White’s Ikki: Social Conflict and Political Protest in Early Modern Japan Gulliver in Japan: Swift’s Gulliver’s Travelsࠉ Japan in a Changing Asia: Achievements and Opportunities Missed

3 18 44  56

Part II: Statistical Resources of and Interactions with Tokugawa Japan Population: Tokugawa Population: The Archival Issues Coastal Trade: Statistics of Tokugawa Coastal Trade and Bakumatsu and Early Meiji Foreign Trade. Part 1: Coastal Trade in Tokugawa Times 8. Post-1859 Foreign Trade: Statistics of Tokugawa Coastal Trade and Bakumatsu and Early Meiji Foreign Trade. Part 2: Trade in Bakumatsu and Early Meiji Times. 6. 7.

vii

75

134

183

viii

EARLY JAPANESE TRADE

Archives: Japanese Archives: Sources for the Study of Tokugawa Administration and JapaneseHistory 10. The Foreign Trade of Nagasaki: The Nagasaki Trade of the Tokugawa Era: Archives, Statistics and Management 9.

Notes Glossary Index

234 273 315 355 359

Foreword by Shunsuke Katsuta University of Tokyo

™

It is easy enough to talk about Tokugawa Japan’s uniqueness, with its domestic political stability, steady economic growth, controlled and peaceful foreign relations, and development of various social values and attitudes, all of which were sustained for two and a half centuries. However, strictly speaking, its uniqueness can only be fully appreciated when it is compared with the history of other countries. Such comparisons are rare and it is hard to think of any historian who has attempted to do so in any depth – with the solitary exception of Louis Cullen. Over the course of his fifty years’ career as a historian, he has conducted extensive research on the early modern history of Ireland, Britain and France, later extending his interest to Japan. He has written eight monographs plus five shorter studies and edited six volumes of studies in history or conference proceedings. Cullen’s ‘research’ is by no means confined to reading manuscripts at particular archives; importantly, he never uses documentary sources before he understands how they were created, preserved, copied or destroyed and, at times, went missing. Consequently, spending, as he has, a considerable amount of time and labour in Japanese as well as European archives and libraries has enabled him to compare the personalised and non-systematic way of record-keeping in Tokugawa Japan and the more formalised way of the European counterparts. In Japan, shogunal records were regrettably thin, while the papers of bakufu officials at Edo and elsewhere in the country were simply regarded as the personal property of individual office-holders. However, there was a positive side to the Japanese style of record-keeping. Because of the number of copies that had been made, often by private hands, compilaix

x

EARLY JAPANESE TRADE

tion in later years was possible – as Cullen’s careful tracking of Katsu Kaishū’s editorial activities shows – which has enabled historians to catch a glimpse into the past even though the original documents no longer exist. More specifically, Cullen’s insight into sources is illustrated by his important study of the population figures of Tokugawa Japan, which, though no figures exist for three census years, survive in detail for eleven of the remaining nineteen. The data were collected by the domains and then converted by shogunal officials to a kuni basis before a grand total was arrived at. He also shows that many of these existing censuses are copies transcribed from official compilations, or even secondary copies of earlier copies, which inevitably contained errors. In other words, as Hayami Akira, one of the leading historians of Japanese demography, admits (‘Population structure and change in early modern Japan’, Bulletin of Japan Academy, 62–3 (2008), p. 292), Cullen’s article brings into question the common understanding that Japan’s population remained more or less static in the later Edo period. Another of Cullen’s research subjects in Japanese history is trade, both domestic and foreign. In a mountainous country like Japan, coastal trade played a more significant role in overall domestic trade than in Europe. As a historian familiar with early modern European economy, Cullen concludes that Tokugawa Japan’s coastal trade was probably the largest in the world. In explaining its size one may refer to Edo and Osaka, respectively the biggest and one of the biggest cities in the world, but Cullen points out that Tokugawa Japan was exceptional in another way: its creation of records of the coastal trade. While only part of these records has survived, European countries of the same period had virtually no such records. Interest in archival issues leads Cullen to reconstruct the story of the survival of the record for Osaka trade for the year 1714 in three sequences of copying. The return (or more probably what was already a copy) had been copied by a private individual. In time it was recopied for a university library. The survival of its information in turn depended solely on being transcribed on three different occasions between 1904 and 1913 (on one of them by a distinguished historian) before the library’s copy (itself the only known surviving item of this sequence) was consumed by fire in the Great Kanto Earthquake of 1923. After exploring the possibility of recordmaking in other years and having carefully studied the story of the survival over time of the 1714 figures, Cullen proceeds to examine the existing coastal trade records for other years.

FOREWORD

xi

The lively and expanding domestic trade of Tokugawa Japan stands in sharp contrast with its foreign trade. Cullen is cautious against a prevalent tendency in Japanese historiography to see the four portals of foreign trade, namely Nagasaki, the Ryukyus, Tsushima and Ezo, as equal in importance. Nagasaki, the only official port directly under the control of the bakufu, was by far the most important. Although its trade saw a long-term decline beginning in the early eighteenth century, Cullen sees the process as a complex one: following an increase immediately after the introduction of sakoku, the Dutch trade began to decline steadily to the point of Japan’s fearing a Dutch abandonment of the man-made ‘trading island’ of Dejima, while Chinese trade somehow recovered in the late eighteenth century. Unlike documents at Edo, those at Nagasaki for foreign trade were very well kept until they were lost in 1868. However, using the now scattered surviving documents Cullen reconstructs the figures of Nagasaki trade though the role of Edo in the finalising of figures remains obscure. For coastal trade figures the central figure may have been the Osaka bugyō already enjoying shogunal backing during the reign of Yoshimune. Things become even more elusive regarding the question of smuggling. However, here again Cullen’s careful consideration of the information, admittedly indirect, as well as his knowledge of European trade, which unlike Tokugawa Japan was conducted with high custom duties and hence prompted smuggling, helps him to argue that the significance of Japanese smuggling, centred around Satsuma, has been overestimated by historians. In 1858, Japan’s foreign trade entered a new era with the opening of several ports to the West leading to Europeans demanding trade statistics compiled in Western style. Added to this were the creation of custom houses (and with it the concern for aggregate statistics) for the first time in Japanese history and the currency question (gold and silver / dollar and yen). The Europeans, particularly the British, were at first frustrated and critical of the Japanese ‘lack of system’, but according to Cullen the Japanese officials coped well with the new task (even in the difficult years of the late bakumatsu), as the British consuls were in a few years able to rely on aggregated statistics prepared by the Japanese custom officers. The consular reports were sent home and published in the British Parliamentary Papers (BPP), now the best source available for the Japanese foreign trade in the late Edo and the early Meiji periods. However, even in BPP variations and contradictions in statistical figures exist, which has been a problem for generations of historians. Here

xii

EARLY JAPANESE TRADE

again Cullen’s thoroughness in using sources and expertise in trade helps him to explain two of the major and apparent contradictions. Cullen’s fascination with Nagasaki is not solely because of his interest in foreign trade. His command of the Dutch language leads him to research another subject: how well versed in the Japanese language were the Dutch in Nagasaki? The conventional view is that the Dutch were prohibited from learning Japanese. After citing cases of the Dutch use of Japanese, Cullen makes a distinction between prohibition and on formal occasions constraints on the use of either language by the Dutch. This reflected Japanese etiquette which required that the formal conversation of august persons with Japanese and Dutch alike was conducted exclusively through intermediaries. Informal settings in which the Dutch were addressed directly did in fact exist. The best known instances are calls on the Dutch by daimyo in disguise, a not uncommon event. Disguise was simply a convention to indicate that the daimyo was dispensing with the formality of rank The above question may urge one to raise a broader issue, which has fascinated many historians and which shows that Cullen has been not only an archival expert but also an insightful reader of historiography, namely, Tokugawa Japan’s relationships with and its perception in the West. While he is critical of the stereotyping of Japan by contemporary Westerners, he admits that the Japanese coped remarkably well with the Russians and the Americans in the nineteenth century. Cullen is also against the Marxist interpretation of Tokugawa Japan by many of the Japanese historians of preceding generations, as well as some of the Americans who in the atmosphere of the Cold War tried to see elements of Western-style dissent in Tokugawa Japan. Conventional assumptions and sweeping assertions are Cullen’s chief enemies. He thinks, ‘History is a set of hypotheses to explain complex social realities; these hypotheses are never perfect and will constantly shift.’ (L. M. Cullen and Kevin Whelan, ‘Watching the detective’, History Ireland, 2–2, (1994), p. 12). This collection of essays is thus the fruit of long, extensive, but careful and considered research, and together with his earlier book, A History of Japan, 1582–1941: Internal and External Worlds (Cambridge U.P., 2003), this volume puts Cullen on the list of those Westerners who through their writings have made great contributions to the understanding of Japan. Shunsuke Katsuka Professor, Graduate School of Humanities and Sociology, University of Tokyo April 2020

Acknowledgements ™

For permission to reproduce material, the publisher and the author are indebted to The Transactions of the Asiatic Society of Japan Routledge (Taylor & Francis Group) The Japan Foundation Newsletter Journal of the Irish Society of Japan Japan Review, Journal of the International Research Center for Japanese Studies Fuller bibliographical details appear at the head of each paper TRANSLITERATION OF JAPANESE

The rendering of words from Japanese into other languages presents a potentially large and on occasion unsightly clustering of macrons, capitals and italics In this volume macrons are not added to long vowels in words and place names likely to be relatively familiar to readers. For less familiar words the macron is retained. Though capitalization does not exist in Japanese, capitals appear in this volume in the romaji (romanization – transliteration – of the Japanese written language) of personal names and titles of major government offices. These instances apart, Japanese words appear in romaji and italics (with capitalization appearing solely in the opening letter of the first word of book titles). Generic terms, referring to administrators – bugyō – and to shogunal offices in many locations – bugyōsho – are italicized without an initial capital. Weights and measures in romaji are italicized, as are Japanese phrases where they occur and unfamiliar individual words. An exception is ‘koku’ and ‘tael’ which appear widely in English-language writing. In Japanese writing there is no spacing between words, leaving open, in Transliteration, the option of a space or its absence. e.g. the alternatives of xiii

xiv

EARLY JAPANESE TRADE

machi bugyō or machibugyō. The former is the norm chosen in this volume. The term han or ‘han’ in roman script or italics is left unchanged in earlier papers – as opposed to the more recent usage of ‘domain’ in later ones. Likewise ‘Tempō’ and ‘Deshima’ in early papers are replaced by ‘Tenpō’ and ‘Dejima’ repectively in later papers. Names of Japanese follow Japanese practice with surname preceding given name.

Introduction

The Route to Japanese Trade Louis Cullen ™

LANGUAGES AND JAPAN

I had never envisaged Japan as a part of my future academic life. In a teaching career spanning forty years (1963–2003), Japan featured actively only in the eighteen years from 1985. It had then to be combined with teaching in Irish history and longstanding research in French history, inevitably resulting in a drastic curtailment of plans for an exhaustive study of the Irish merchant communities in France. My first contacts with Japan date from 1972, when Matsuo Tarō, a lecturer at Hosei, took two years’ leave of absence in Ireland to benefit from supervision and to attend undergraduate lectures. Turned down by University College Dublin (which he regarded as the preeminent national university and hence his first choice) on the grounds that his English was inadequate, I met him by appointment. While he had difficulties in spoken English, he read it with ease and had a very coherent plan for his research in rural history. He was subsequently admitted to TCD (Trinity College Dublin), and thus began a lifelong friendship. Between 1979 and 1992 he visited Ireland on four further occasions, for sixteen months in all, and was due to return for a full year in 1997 when he was struck down by a fatal illness. His long first stay fitted into the emerging pattern of Japanese academic visits to Europe. From the early 1980s Japanese scholars sought the status of ‘visiting academic’ at Trinity, and from the mid-1980s research students became commonplace. xv

xvi

EARLY JAPANESE TRADE

Invited to Japan by Professor Matsuo in 1985 for a month-long stay to give a series of four lectures in Irish history at Japanese universities, I started to study Japanese in advance. This urge was prompted less by the invitation itself than by an active interest in languages, and by this unexpected opportunity to study an oriental language. Years before, in wartime isolation in a neutral country, old textbooks in French and German had served as an avenue to the outside world and fired my imagination. I even wrote to the largest bookseller in Dublin for grammars in both languages. With books of any sort from Britain in short supply, the only book on offer was Hugo’s Dutch in Three Months Without a Master. At eleven or twelve years of age I duly mastered it. The irony was that the choice of a Dutch-language textbook was to prove invaluable in later years, both in allowing me to translate French brandy sources (which as often as not were in Dutch) and, latterly, in understanding the writings of the Orandatsūji, or Dutch-Japanese interpreters. Requiring an extra subject for my final school examinations to qualify for a state university scholarship, I chose a language, Spanish. As there was no teaching in the college where I was a student, I posted coursework to a priest who had spent years in Spain and was a member of the order that ran the school. After completing my studies at University College Galway and the London School of Economics, from 1959 I spent four years in diplomatic service, mostly in Rome and Paris, before returning to academia. FIRST ENCOUNTERS WITH JAPANESE HISTORY

Though, as we have seen, I was in contact with Japanese academics from 1972, my direct involvement with Japan began some years later, when I undertook a short visit to the country in 1985. In 1989, I introduced a course in Japanese history at TCD. I had it in mind eventually to turn the lectures into a textbook, although I had no plans to do so with any haste. I already knew Richard Fisher, Cambridge University Press’s history editor at the time, as the result of long-drawn-out negotiations with the Press over my proposal to bring together the French brandy trade and the Irish merchant houses of France within the confines of a single large volume. It was he, in the course of a conversation at a history conference in 1991, who encouraged me to prepare a book on Japan for the Press, which in turn prompted me to consider writing a more ambitious work than I had previously considered, and sooner

INTRODUCTION

xvii

rather than later. Extended visits to Japan in 1993 and, again, in 1995 as a Japan Foundation fellow, widened my study of the language, and from 1995 I was able to read Japanese, albeit still with some difficulty. The book, which would come to bear the title A History of Japan, 1582– 1941, gradually took shape. By April 2003, at the end of a year as a Visiting Scholar at Kyoto’s International Research Centre for Japanese Studies, or Nichibunken, the final revision was completed, much aided by access to the Centre’s splendid library. With no academic background in the subject, such progress in Japanese language and history could not have been achieved without help from many sources. The most important were Tarō Matsuo and Kawakatsu Heita, whom I first met at All Souls College, Oxford, in 1987. Their support and encouragement were vital in myriad ways. Eileen Katō (née Lynn) and Tsukahara Sueko ሯཎᮍᏊ, two non-academic helpers, also provided a great deal of assistance, often by sending books to me. Eileen and I shared classes in Spanish in 1950–1951, and when I next met her in Tokyo in 1992 she was the widow of a distinguished Japanese diplomat, and a goyōgakari, or special adviser, at the imperial court. For her part, Tsukahara Sueko, who years before had been Matsuo’s kanbun teacher, introduced me to many of Nagasaki’s libraries and archives, and accompanied me on my first visits to them. Dr Honma Sadao, head of the local history section of the Nagasaki KenritsuToshokan (Nagasaki Prefectural Library), lent me much assistance. In Tokyo, Professor Masato Miyachi, at the time director of the Shiryō Hensanjo (Historiographical Institute), helped me to understand the role of the Hayashi; and, later, Dr HōyaTōru supplied me with copies of the famous 1804 scrolls when they were otherwise inaccessible due to the earthquake proofing of the Shiryō Hensanjo. Of the many others who provided support of one kind or another, I must draw especial attention to Professor Saitō Osamu (Hitotsubashi University), and to a former student, Professor Katsuta Shunsuke, now of the University of Tokyo, for the generosity with which they offered their time and help. In recent years, I have been particularly obliged to Professor Katsuta, whose assistance I have called upon a great deal. I am also greatly indebted to the late Professor Yamamura Kōzō (University of Washington), who commented on the text of my 2003 book. I owe much, too, to Jim Baxter and John Breen, as editors, and to the anonymous reviewers who critiqued the papers I submitted to the Japan Review. Whilst there is not space here to thank all of those who have assisted me, Professors Kasaya Kazuhiko (Nichibunken), Mitani Hiroshi

xviii

EARLY JAPANESE TRADE

(University of Tokyo), Gotō Hiroko (Hosei University), Kuwajima Hideki (Hiroshima University) and Takagi Masao (Ritsumeikan University) all deserve a mention. WRITINGS ON JAPANESE HISTORY

In Japan’s huge historical literature, there are some very rewarding books in English, notably Ishikawa Shigeru’s Economic Development in Asian Perspective (1967), and Visions of Virtue in Tokugawa Japan: The Kaitokudo Merchant Academy of Osaka by Tetsuo Najita (1987). Works also exist on the early eighteenth century, specifically Told Round a Brushwood Fire: The Autobiography of Arai Hakuseki by Joyce Akroyd (1979), Kate Wildman Nakai’s Shogunal Politics: Arai Hakuseki and the Premises of Tokugawa Rule (1988), and, more recently, Beatrice M. Bodart-Bailey’s perceptive The Dog Shogun: The Personality and Policies of Tokugawa Tsunayoshi  (2006). On trade and sakoku (that is, the isolationist foreign policy of the Japanese Tokugawa Shogunate), research in English is slight, though of late invaluable books have been published by Shimada Ryūto and Suzuki Yasuko. In both languages, the material available on the role of ‘senior councillor’, effectively ‘prime minister’ at the head of a cabinet of four or five rōjū, or ministers, is uneven. The negative portrayal of Matsudaira Sadanobu in J.W. Hall’s uncritical 1955 appraisal of his predecessor, Tanuma Okitsugu 1719–1788: Forerunner of Modern Japan, contrasts with studies on Tanuma by Ōishi Shinzaburō and on Matsudaira by Herbert Ooms (in English ) and Fujita Satoru. Fujita’s magisterial work of 1987on Mizuno Tadakuni is incomparably the best study of a senior councillor but is sadly lacking an English translation.1 Of his other books one might single out Kinsei kōki seijishi to taigai kankei (2005) for his determined effort to track down documents of the mid-1820s on the uchiharai policy (which sanctioned the firing upon all foreign ships approaching Japan), an approach all too rare in Japanese writing.2 The negative view of Japanese history espoused by Japanese Marxist historians, which became if anything more prominent post-1945, was gradually replaced by the more positive appraisals of later generations. Whilst academics have continued, to a greater or lesser degree, to focus on the bleaker aspects of the country’s past – such as arbitrary government, harsh criminal justice (including torture, conviction based on confession, crucifixion of criminals, and numerous deaths in prison), famine, a defective economic system, a Shogunate said to be unsure of itself,

INTRODUCTION

xix

and what can be read as expressions of alienation in private writings – this has been tempered by other considerations. Herman Ooms, in a justly celebrated work, has shown evidence of a fair system of civil law, through which private individuals were able to pursue cases, and could even gain access to an appeals process.3 The potential tension between emperors and shogun, moreover, has often been overstressed. It is worth recalling that Emperor Kōmei (1846–67), although a vigorous critic of open trade (largely due to the perceived future menace posed by the presence of foreigners in Osaka, close to the Imperial Palace in Kyoto), remained a strong backer of the Shogunate itself.4 With such a wealth of history writing from which to choose, it is not easy to select individual authors for comment, but the task does become easier if one limits oneself to scholars who have taken a wideranging approach to the subject. Of these, one of the most interesting is Kōda Shigetomo ᖾ⏣ᡂ཭(1873–1954), a pupil of Ludwig Riess (1861–1928), father of Western-style historiography in Japan. Kōda supervised the completion of the first five volumes of the Osaka-shi shi (‘A history of the city of Osaka’) published between 1911 and 1915, which has since served as a forerunner to all similar works. He maintained a keen interest in the sources and statistics of trade and urban population until, after spending a substantial amount of time in the Netherlands in the 1920s, he turned his attention to relations between Japan and Europe during the Tokugawa period, a subject that was to remain his focus for the rest of his career. His memory lives on; a small group met every year at least into the 2010s in the Nichibunken to commemorate him. His versatility is not often matched. But of the more recent historians who have covered a remarkably broad spectrum of topics, Ōishi Shinzaburō ኱▼ៅ୕㑻(1923–2004) is notable for the scope of his books and articles, which deal with economic, political and statistical themes; whilst Fujita Satoru ⸨⏣ぬ (1946–), already mentioned in the course of this Introduction, can probably be judged the most versatile historian of his generation. In an entire category unto himself, meanwhile, is Hayami Akira ㏿Ỉ⼥ (1929–2019) who, inspired by family reconstitution exercises in France and England, used the much more intractable village population registers of Japan for similar ends, thereby creating the first substantial contributions to Japanese historical demography. This brief overview may seem unduly narrow to some, coloured as it is by a personal interest in trade, administration and archives. It certainly

xx

EARLY JAPANESE TRADE

ignores the work of historians who have written widely on other areas, and could also be termed unfair in its disregarding of two generations of historians whose research has focused on Nagasaki. JAPANESE ARCHIVES

By 1995, I was at last able to read in Japanese, and immediately threw myself in at the deep end by going straight to the archival sources, expecting to find systematic, chronologically ordered series of documents in the European fashion. If many years’ study of European history meant that I had only a late-acquired knowledge of Japanese language, the advantage was that I was well-trained in working with centuries’ worth of archival material relating to English and French history. As a graduate student researching in the Archives Nationales in Paris in 1954–1955, furthermore, I had had the good fortune of attending the opening semester of the Stage technique internationale d’ archives, which sharpened my awareness of archives as a fundamental component of the central framework of administration. In examining in 1995 the huge Tsūkō ichiran ㏻⯟୍ぴ (TKIR), or Survey of foreign relations (1853), as well as the challenge of understanding the texts within the TKIR (which vary from easy to very difficult), there was a need of establishing whether coherent archival series existed. I found myself reliant on the terse indications of sources provided by the TKIR at the end of each section for guidance on modus operandi and a source base at large. Japanese archives, as has been noted by Kikuchi Mitsuoki, onetime president of the National Archives, or Kokuritsu Kōbunshokan ᅜ❧බᩥ᭩㤋 (formerly Naikaku Bunko as established in 1885), fell far short when compared to the scale and continuity of European archives.5 It is not that archives were not kept; they were, in fact, in both Edo (modern-day Tokyo) and the domains. What is lacking in its entirety, rather, at least as far as Edo is concerned, are the papers of decision-makers; the files, in other words, of the four or five men who made up a cabinet. The TKIR, invaluable though it is in drawing on a wide range of sources, simply confirms the elusive nature of the final stages of government-level decision-making. The fact that Nagasaki bugyō, that is, magistrates or administrators, took their papers with them at the end of their periods of office, is another complication. Not only this, but the Japanese pattern of employing several figures to carry out a given task has ensured that, where documents exist, they do so in

INTRODUCTION

xxi

a profusion of copies that may or may not have been retained by their creators, and whose continued existence is dependent on the whims of descendants. For the most sensitive matters, moreover, it was customary in Edo for members of three categories of rank, namely kanjō bugyō ຺ᐃዊ⾜, jisha bugyō ᑎ♫ዊ⾜, and machi bugyō ⏫ዊ⾜㸦bugyō exercising responsibility in finance, supervision of temples and shrines. and town administration respectively) to work together. The sources drawn on by Hayashi Akira’s team for the TKIR in the late 1840s and early 1850s were heterogeneous, but the compilers had no access to papers that rōjū had removed when their terms were completed, and which survived, if they did so at all, in domain archives. The knock-on effect is to render evidence of decision-making almost invisible. This is despite the fact that the nineteenth-century heads of the Hayashi family were themselves involved with high-level consultations. Hayashi Jussai ᯘ㏙ᩪ (1768–1841) had the same status as officials of kanjō bugyō rank in policy consultations of the 1820s whilst, thirty years later, Hayashi Akira ᯘ㡐(1801–1859) was titular head of a team tasked with negotiating with the Americans. As head negotiator he was a respected, active and well-briefed figure, yet, if we were to go by the surviving correspondence, he would seem to have played no part in policy or decision-making. His involvement in the talks may, however, explain why the TKIR Zokushū (the second instalment of the TKIR, completed in 1856) drew on some fifty sources from a wide, though very miscellaneous, range of private and official documents for 1854, the key year of the negotiations. In European states, by contrast, well-defined ministries existed, whose ministers, whilst regarding much of the correspondence addressed directly to them as a private archive, ensured that the integrity of the central core of records on policy and administration was respected. The papers generated by such ministries were managed systematically, and the contents of volumes were often numbered and sometimes bound, a practice that was especially common in France. Even in smaller centres, such as custom houses in Britain and naval stations in France, incoming and outgoing letters were laboriously transcribed into two separate runs of letter books. As far as archives in the domains are concerned, J.W. Hall’s seminal writings of the 1950s set out in detail the scale and nature of the records of Okayama Domain. More recently, in a book on domains, Kasaya Kazuhiko has described the extensive holdings of seventeen domains. Domain material was to prove the backbone of the Dai Nihon

xxii

EARLY JAPANESE TRADE

ishin shiryō ኱᪥ᮏ⥔᪂ྐᩱ (DNIS), literally ‘Historical Records of Restoration Japan’, of which only the initial nineteen volumes were destined to be published, and that only many years later in 1938– 1943. The venture’s origins can be traced back to Shimazu Hisamitsu, regent to the last daimyo, or lord, of Satsuma domain, and his desire to vindicate the domain’s past political stance. Beyond this, we can also look to the proposal for an editorial board of national history made in 1890 by Kaneko Kantarō 㔠Ꮚሀኴ㑻, who was long to remain an active campaigner in the field. This project bore some fruit, specifically the compiling of 4215 volumes of copied records, which was financed from 1911 by the Monbushō (Ministry of Education).6 This relative abundance of domain-based archival material, at such odds with the lacuna represented by central records, explains why much of the work produced by American graduate students (who account for the majority of native English speakers carrying out research on Japanese history at this level) has been concerned with aspects of administration, politics and economy in the domains. Examining these sources has produced students who are at ease with reading historical documents in the original Japanese, and who are well-placed to build on such early encounters, even whilst the absence of other records poses a challenge to early and seasoned researchers like. THE LOSS AND RECOVERY OF ARCHIVES

The story of Japanese archives is a unique one, combining both loss and partial recovery. At the outset of the Meiji ishin, or ‘Meiji Restoration’, documents were at risk both in Edo and elsewhere. There was no coherent archives policy until 1884–1885, which meant that any surviving documents in Tokyo ran the risk of being passed from institution to institution casually. Even after 1885, policy remained indecisive. The records of jisha bugyō, machi bugyō and Hyōjōsho (supreme judicial council) were transferred very late in the day to the Imperial University Library, where they were consumed by the fires caused by the earthquake of 1923. The fate of the records produced by the Kanjōsho (Finance Office), is unclear, despite the assumption that the archives became the property of the Ōkurashō (the new, post-Meiji Restoration Ministry of Finance) in 1871, and were then lost in the fires of 1923. Be that as it may, it is evident that there was little for the Ōkurashō to inherit, whether as the result of past accident or due to the total collapse of record retention – in Nagasaki style – in 1868. Ōkuma Shigenobu, minister of finance in the

INTRODUCTION

xxiii

1870s, aware in 1878 of the magnitude of the archival deficit, launched a programme to gather copies of Tokugawa-era documents. The wellknown efforts of Katsu Kaishū in his 1887 Suijinroku ྿ሻ㘓 (a history of the Tokugawa Shogunate’s finances) ran parallel to a larger enquiry headed by a team of about ten officials, which finally resulted in Nihon zaisei keizai shiryō (‘Historical Documents for Japanese Finances and Economics’), whose ten volumes were published from 1921–1925. Although this lengthy work paid tribute to the sheer scale of past copying (some of it in compendia of documents), it was something of a ragbag collection of texts that had already lacked a meaningful context. Takimoto Seiichi, who was an active figure in this project, was also the editor of a vast Nihon keizai taiten (‘Collected Works of Japanese Economics’). The Nagasaki bugyōsho (‘governor’s office’), almost on a par with the Kanjōsho in terms of size, has a happier story to tell about the recovery of papers abandoned in 1868; although not a completely happy one. The papers that were retrieved included those that had been in the bugyōsho in 1868, along with a huge flotsam and jetsam of copies made privately at an earlier date. Typically, in some instances the original records only survived because the descendants of office holders had retained them. A striking case in point is a diary compiled by the administrators of the Saga watch house in Nagasaki of the diplomatic mission of N.P. Rezanov in 1804 which sought to open the harbour to Russian trade and ships. The diary passed down in the family of the head official, and was finally deposited in the library of the small town of Isahaya near Nagasaki in modern times.7 The fate of the 1663–1715 nikki, or ‘diaries’, of the Nagasaki interpreters of Chinese, in the 1880s, already missing eighteen years of this time-span, represents another cautionary tale of the hazards facing records. Kanai Toshiyuki 㔠஭ಇ⾜, head of a Nagasaki city ward from 1886, had copies drawn up of ten volumes of nikki.8 No doubt the duplicates were made because he had failed to acquire the originals. One of the copies was later lost. Of the surviving nine, seven were held by the city after his death in 1897.9 The final two volumes, the so-called ‘Mukai volumes’ (the Mukai being the family who, by tradition, presided over the Chinese Confucian temple, or Seidō ⪷ᇽ), reveal some small differences in dimension and in transcribing style, from which it is evident that they were transcribed separately, in circumstances which are not clear. While the originals are unaccounted for, the twocopies were finally acquired in 1934 by the Nagasaki ken kyōiku-iinkai 㛗ᓮ┴ᩍ⫱ጤဨ఍(Prefectural Education Board). It is safe to assume

xxiv

EARLY JAPANESE TRADE

that the originals of the nikki copied for Kanai, and possibly also those of the ‘Mukai volumes,’ were by the 1880s in the hands of several interpreter families or their descendants.10 Whilst all nine copies were lodged in the City Museum by 1959–1960, the loss of the originals, whether at the hands of the Mukai or others, is an illustration of the ongoing attrition of papers in private hands. The absence of nikki for years after 1715, either lost or retained in private possession, contrasts with what appears the lodgment of other papers in the Seidō in the 1860s. With no Chinese vessels entering Nagasaki in the 1860s, the interpreters may have precipitously abandoned their now irrelevant files. One intriguing survival is a unique run of documents that lists the details of cargoes from 1709–1714, bound by an unknown hand into four volumes. The run is itself incomplete, consisting of only thirteen of the original twenty satsu (files), and even those that survive lack some individual cargo returns. An effort was made to fill the gaps for 1711, and we also have a manuscript written in several hands under the title 㛗ᓮᚚ⏝␃ Nagasaki goyōdome (a form of shahon, meaning ‘notebook’ or ‘workbook’), which contains transcriptions of fūsetsugaki 㢼 ㄝ᭩ and Tōjin domo mōshiguchi ၈ேඹ⏦ཱྀ (terms that denote collective statements by Chinese), interspersed among the cargo returns. This shahon itself was evidently created with a precise purpose in mind, almost certainly by the man who had the cargo returns bound into four volumes. It is worth noting that both these volumes and the Nagasaki goyōdome survive in Tokyo rather than Nagasaki, hinting at an obscurechain of past ownership. Nagasaki, as we can see, represents the outstanding instance of recovery of papers in Japan. Serious rescue attempts began in the 1880s at the behest of two main protagonists (one of them a head of a Nagasaki city ward), both of whom were determined to acquire, rather than simply copy, documents. By way of contrast, later recovery operations, undertaken on a grander scale in the first decade of the twentieth century for the Osaka-shi shi and the Osaka Chamber of Commerce respectively, borrowed documents for copying before returning them to their holders, whose names they apparently failed to retain. INTERPRETATIONS OF JAPANESE HISTORY

There is a sharp contrast between Engelbert Kaempfer’s irritation in his 1727 History of Japan at being policed by his hosts, and his much more mellow take on a closed Japan in an appendix to the same work, which

INTRODUCTION

xxv

appears in some but not all editions. The appendix is drawn from a work composed in 1712, written reflectively in Latin by Kaempfer, and far removed from the petty frustrations recorded in the History, which appear to rely on diaries of 1690–1692.11 This same tension runs through the story of the small Dutch community on Dejima, as recorded in the dagregister (diary maintained by the opperhoofd). Interpretations of Japanese history have been subject to a number of specific influences that have had an unusual degree of impact. One of the most noteworthy is the Marxist approach that was prevalent in the interwar years. Maruyama Masao (1914–1996), who avoided conceding the eclecticism of Japanese thought in his single-minded search for the origins of Japan’s militarism, became a close friend of the Canadian diplomat, E. H. Norman (1909–1957). Norman’s Japan’s Emergence as a Modern State, the first analytical study of Japanese social history in English, rested on his reading of texts by Japanese Marxist historians of the era. Together, Maruyama and Norman sought to give significance to the works of the highly eccentric eighteenth-century philosopher and medical doctor, Andō Shōeki.12 Marxist thought accounts, in part, for the weight given to ikki୍ ᥡ(rural unrest or grievances) in retellings of Japanese history. Their origins were far from straightforward, but they were seized upon by American historians, who were eager to find a tradition of Japanese dissent to bolster the prospects of the modern democratic system that had been imposed on the country.13 For example, Hugh Borton, one of two Columbia University historians who were pioneers of Japanese studies in the USA, wrote an extended study of ikki in 1938, his take on which is neatly summarised by a chapter title from his 1955 Japan’s Modern Century: ‘The Dictatorship Collapses 1857–1867.’14 Another influence on Western writing on Japan, and one which was closely aligned with Marxist thought, was a focus on foreign trade. The restricted trade possible under sakoku was a cause of immiseration, or so it was argued, and was thus a factor in rural unrest. Such views were reinforced by Iwao Seiichi and Tashiro Kazuo’s assertion that a Japanese readiness for open trade did, in fact, exist, and had done so for more than a century from the time a request by an English vessel, the Return, was turned down in 1673 (allegedly because the monarch’s consort was Portuguese). Their argument was that in the absence of a request from an English vessel in later years the freedom was simply not tested. One can look, too, to an attempt by Hall to turn the eighteenth-century

xxvi

EARLY JAPANESE TRADE

reformer, Tanuma Okitsugu, into an exponent of open trade, in a 1955 work entitled Tanuma Okitsugu,1719–1788: Forerunner of Modern Japan. Ronald Toby was also involved in these endeavours, taking up the Iwao–Tashiro argument to argue that Matsudaira Sadanobu’s rejection of a Russian request in 1793, on the grounds of exclusion by ancestral or immemorial law, represented the creation of a new policy.15 Negative perceptions of Japanese attitudes in Tokugawa times have nevertheless lingered on. Even in a fresh and insightful recent book, Ravina sees ‘an increasingly desperate policy of isolationism.’16 In other writings the old belief that foreigners were prohibited from learning Japanese is still repeated, no matter that no such prohibition ever existed (and was certainly not mentioned by Kaempfer). It is true that the Dutch traders of the 1640s, in a heavily-policed Dejima were refused assistance in learning Japanese and, although they had recruited their own interpreters in Hirado, were forced to use interpreters employed by the Shogunate for all formal business, but this is as far as it went. The assertion that the Dutch were prohibited from even speaking Japanese, which was still doing the rounds in the 1820s and 1830s, never made sense, or did so only in as much as Japanese etiquette insisted that high officers be addressed through intermediaries (who, in the case of the Dutch inhabitants of Dejima, were necessarily interpreters). ROBUSTNESS OF ADMINISTRATION

In tracing Japan’s eighteenth-century history, one can observe that there was originally a pattern of active Tokugawa shoguns with coherent policies. Three shoguns, in particular, were to maintain a crucial role in decision-making. The first was Tsunayoshi (shogun from 1680– 1709). The second was Yoshimune (1716–1745), who decisively settled a protracted conflict of interest by regulating trade on his assumption of office. He was also the most statistically-minded shogun, introducing a population census for every sixth year and, in the course of trying to raise the price of rice, supporting the Osaka machi bugyō in generating statistics relating to trade between Osaka and Edo. The third, Ienari (1787–1837), although he did not publicly put a personal stamp on policy, maintained in an elusive fashion the good order of the bakuhan taisei (that is, the political structure established by the Togukawa shoguns of the period). Under his rule – the longest in shogunal history – Japan retained a good relationship with Satsuma (a sensitive undertaking due to the Ryukyus), and elsewhere showed caution in handling

INTRODUCTION

xxvii

the Russian presence. Apart from crisis at the very outset of his rule, and again in his final four years, there were more than three decades of good harvests, although this was somewhat qualified in the latter years of his reign by the fall in rice prices persisting. Reaching their nadir around 1818, such low prices, though good for the consumer, entailed fiscal distress since rice was the basis of shogunal income, and triggered the debasement of the currency from 1818–1830. Japan possessed a solvent administration. Its often acute fiscal problems were caused, instead, by a lack of capacity to create new taxes, and hence were characterised by efforts to secure economies or to support rice prices, which underpinned shogunal income. Yoshimune’s great reform programme, implemented during the fiscal crisis of the Kyōhō period, along with those of the Kansei and Tenpō years, demonstrate that shogunal endeavours could balance the books. This remarkable achievement – which stands in stark contrast to the massive British and French debts of the era, which in the former case was run up despite a system that was able to raise taxes efficiently, and in the latter, with its greater levels of disorder, ended in the Revolution – was made possible by the Pax Tokugawa which was maintained by a mere token military force. When shoguns after mid-century retired from direct involvement in governance, the senior rōjū became the effective leader, facing on his own authority the challenges posed by the factionalism of Japanese political life. Such factionalism was rife both in Edo and beyond, and was a particular issue in Mito (the domain of a collateral branch of the Tokugawa family), Chōshū (which had a civil war in late 1864), and even in the home base of the powerful Shimazu family, daimyo of Satsuma (Kagoshima). With the retreat of the shoguns from direct management, the senior councillors, at the head of a cabinet of four or five men, were left politically vulnerable. This is especially evident in the years of Tanuma Okitsugu (1772–1786), Matsudaira Sadanobu (1787–1793), Mizuno Tadakuni (1839–43), and Ii Naosuke (1858– 60), all of whom had to face novel problems. The conflict in the 1860s was particularly acute, unprecedented in Tokugawa times. Yoshinobu, regent to the shogun, was under pressure in late 1864 to abandon a campaign against Chōshū, which was then relaunched in the ‘summer war’ of 1866 before being formally abandoned by Yoshinobu, now de jure shogun, in early 1867. The war, it should be noted, had lacked support even among some of his supporters. This urge to end conflict, which was more strongly felt outside the shogun’s immediate circle,

xxviii

EARLY JAPANESE TRADE

accords with the words of Satow, who had witnessed the events of the 1860s, and who commented in 1882 that the Koreans ‘[differed] from the Japanese, among whom the feeling of comradeship is in most cases so strong’.17 In terms of foreign policy, the rōjū were not paralysed, although they acted with extreme caution for several reasons. Matsudaira, for instance, had resisted calls for activism in Ezo, and caution continued to be the order of the day after his departure; in the response to Russian raids in Ezo in 1806 and 1807; in the delicate episode of a Russian naval commander, Golownin, being held prisoner; and in the face of the new challenges posed by the appearance of whalers both north and south of Edo in the 1820s. The arrival of the whalers was a particular cause for concern, provoking fears that their activities were, in fact, exploratory ventures for a future attack on Edo’s supply lines. This is understandable in the light of the Phaeton incident at Nagasaki in 1808, which had painfully revealed the naval vulnerability of Japan. Fujita has demonstrated, too, Mizuno’s concern regarding potential threats to Edo’s sea-borne food supplies in the wake of the Opium War in China. Post-1841, moreover, there was increasing alarm that violent encroachments would take the place of polite requests for trade, especially in the aftermath of English and French campaigns in China. The Japanese had at least acquired knowledge of the size and force of Western naval powers in betsudan fūsetsugaki (that is, Dutch special reports), and in 1856 made a very detailed review of incoming trade in Edo.18 Japan negotiated these obstacles very successfully, a vital achievement given that the treaties of 1858 were open to renewal after twelve years. I commented on this success in papers I delivered in Kyoto and Tokyo in 2002. In Tokyo, Professor Katō reminded me gently that he had made the same point two years earlier. An American author in 2004 made a similar argument.19 PAPERS IN JAPAN REVIEW

After the year 2002–3 spent in Kyoto, and in later visits to Kyoto and Tokyo, I explored in greater detail issues that I could broach only briefly in my 2003 book. In the five papers that I wrote for Japan Review between 2006 and 2017, two concerns are dominant. One is the issue of archives; the second that of statistics, which, population apart, have not enjoyed sustained attention. One paper dealt explicitly with archives. Of the remaining four papers, one dealt with

INTRODUCTION

xxix

population, and the remaining three with trade (coastal trade, post1859 trade, and Nagasaki as the sole centre of Tokugawa foreign trade). The papers provide the main corpus of writing in this volume and, within a single section, are numbered no. 1 to no. 5. Their themes are summarized very briefly below. 1: Population

The study of population statistics in Japan started in the late 1870s. The gold standard is work published in 1957 and 1958 by Sekiyama Naotarō, supplemented by work by Takahashi Bonsen (1955, 1962) and Minami Kazuo (1978). While domains had records of their populations, the process at shogunal level of converting domain figures into kuni totals (the kuni, as an older land unit that had lost direct administrative importance, had stable boundaries, unlike domains) and of compiling national totals is totally obscure. The argument of the paper is two-fold. First, it maintains that the suggestion that secrecy ensured that the figures remained unknown is unfounded. The copies hint, rather, at a process of private copying. Second, it demonstrates that variant or secondary counts for domain populations, contrary to what is sometimes said, did not exist. What did exist was a distinction between the commoner populations occupying the fiefs held by the daimyo (and thus the source of his wealth), and the commoner populations who lived in fiefs held by collateral branches of the ruling family, in fiefs held directly by samurai and on land held by shrines and temples (all conferring no income to the daimyo). The daimyo fiefs measured the wealth of the daimyo; the other fiefs served solely as a population count. This means that domain population counts could either take the form of a count of population on daimyo lands or, with the addition of the populations of other fiefs, of a total commoner population. These figures continued to exclude towns and samurai families. 2: Coastal Trade

In contrast to maritime Europe, the coastal trade of Japan was documented to an extent, at least in quantitative terms. In Osaka, inward shipments greatly exceeded outward ones (and loans to the domains were secured on domain rice cargoes). In Edo, rice from the Tōhoku fed the population; in Osaka, rice from the southwest fed the city itself

xxx

EARLY JAPANESE TRADE

and its densely populated hinterland. The surviving sources are isolated documents, invariably copies rather than originals. The most important contributors to the analysis of this meagre harvest have been Kōda Shigetomo, editor of the first Osaka-shi shi of 1911–1914; Yasuoka Shigeaki, a participant in the 1964–1966 publication by the Osaka Chamber of Commerce of documents copied sixty years earlier, and a perceptive contributor in 1990 to the Shinshū Osaka-shi shi (‘A New History of the City of Osaka’); Miyamoto Matao, an expert on the rice trade; and Ōishi Shinzaburō, for his careful analysis of the trade data for 1714 and 1724–1739. Counts of Osaka imports and exports in 1714, of imports in 1736 (and, presumably, of exports, though the records have since been lost), and of the gross aggregates of inward and outward trade for 1766 (although not of the detail of the commodity counts) point to uncommon occurrences. While the 1714, 1736 and 1766 returns lack figures for domain rice, estimated values can fill this lacuna. Evidence of commodity figures recorded on other occasions over the period points conclusively to the fact that the Osaka bugyōsho made some regular, though not necessarily comprehensive, counts of the intake of goods by sea. 3: Post-1859 Foreign Trade

As for foreign trade, post-1859 figures are relatively abundant. British returns compensate for gaps in Japanese figures in the 1860s. Early Japanese studies drew on the annual consular reports. There is some suggestion that there were two counts of trade; one accounted for by the consular figures, the other by the Chambers of Commerce. But this is a misunderstanding. The Chambers of Commerce simply sought to create an inventory of stocks. They took the figures for imports, and by private enquiry made an estimate of the scale of disposals, that is, sales within the same period. To some extent they also altered commodity prices 4: Archives

The wholesale copying of documents in the Tokugawa era was replicated on a vast scale by the transcriptions of the Meiji era. A common feature of both eras was that, once copied, originals were returned to their holders. The major project was the Dai Nihon ishin shiryō. It had grown out of the tozama domains’ pre-existing interest in vindicating their past actions but, under the management of the Monbushō from 1911,

INTRODUCTION

xxxi

it finally resulted in over 4000 manuscript volumes. The fact that the rōjū (daimyo of fudai rank) took their papers with them at the end of their terms of office make this source the key insight into the high politics of the period. A concerted effort to keep track of the ill-monitored record of relations between officials, foreign residents and representatives of their governments led to manuscript compilations of such documents, which were completed in two stages: the Tsūshin zenran ㏻ಙ඲ぴ (‘Full survey of correspondence’), which was finalised in 1867 and related to the years 1859–1863; and, in 1877, the Zokutsūshin zenran ⥆㏻ಙ඲ぴ (‘Full survey of correspondence (sequel)’ to cover the following years, which through delay was more an academic exercise than a response to contemporary needs. Published in 1983–1988, the two series are comprised of six and fifty-five volumes respectively. Two Osaka compilations from the first decade of the twentieth century and, less impressively, the ten volumes of copied documents compiled by the Ōkurashō from material from outside its own offices (published 1921–1925) are further examples of large-scale copying. 5: The Foreign Trade of Nagasaki

Nagasaki, in the sakoku years the sole centre of foreign trade, warrants a study in its own right. The peak year was 1661, with a recovery in the China trade in the 1690s and the following decade. Thereafter, trade stagnated or fell, though the China trade recovered modestly again just ahead of the turn of the nineteenth century, to the point that it was worth six times the value of the languishing Dutch trade (worth a mere 1000 kanme). In these altered conditions, Chinese traders were now welcome visitors, rather than victims of harsh and arbitrary treatment, as they had been in the past. Satsuma was permitted to trade in Chinese goods in Nagasaki via the Ryukyus, although only to a limited extent, and the bugyōsho in Nagasaki, lacking the wholehearted backing of Ienari on this sensitive issue, and powerless to prevent Satsuma trading directly with other domains, perceived the trade to be larger than it actually was. ADMINISTRATION, ARCHIVES, STATISTICS

The five papers from Japan Review all stress the vigour of Japanese governance. The surviving data provide a picture at a routine level of a

xxxii

EARLY JAPANESE TRADE

highly organised bureaucracy that was reliant on paperwork, the archival integrity of which at the time was compromised only by retention of policy papers by bugyō and rōjū. The evidence that statistical data continued to be compiled in the 1850s and 1860s supports the picture of a bureaucracy functioning effectively amid the tribulations of the two final decades of Tokugawa Japan. There was even the successful emergence of a new category of statistics from 1859, relating to foreign trade at three open ports. Despite reservations expressed by British consular staff, such documentation was compiled efficiently, but the loss of the original accounts shines an arresting light on the vulnerability of records of the 1860s. From as early as 1879, indeed, Japanese writers on foreign trade had to depend largely on British consular reporting of the data. The records held by Yokohama’s ‘governor and superintendent of customs’ (initially answerable to senior colleagues engaged in foreign relations, and finally in 1869 and 1870 to the Gaimushō) do not appear to have been aggregated at the time into grand totals for the open ports. That was to emerge as a by-product of the takeover by the newly created Ōkurashō in 1871. A copy made of tables of exports dating back to 1866 indicates that a general retrospective aggregation by the Ōkurashō, part of it now lost, did exist. As for population statistics, the few surviving returns for 1860s Edo and Osaka are enough to suggest that compilation remained active. Sekiyama (1958) argues that the political weakness of bakumatsu Japan led to the breakdown of data transfer between the domains and the Kanjōsho, and thus to the abandonment of the sextennial census.20 However, while plausible for the census year 1864, when the centre of political action had temporarily shifted from Edo to Kyoto, it is an unconvincing explanation for the census years of 1852 and 1858. For these years it is likely that a census did, in fact, take place, and while only a few incomplete or defective returns exist for the census years immediately preceding these, it is salutary to remember that the complete or better copies on which contemporaries drew were as unknown to Katsu Kaishū as they are to us. The Kanjōsho dealt primarily with collection, expenditure and audit of revenue on shogunal lands (tenryō ኳ㡿), and less routinely with fiscal crises and the rice market. Its role meant that it was well-equipped to amalgamate the census returns from the domains, and to deliver the final return for each population census to the rōjū with the signature of an ōmetsuke ኱┠௜ (senior inspector) and a kanjō bugyō ຺ᐃዊ⾜.

INTRODUCTION

xxxiii

Its most senior officers – a mere eight or so kanjō bugyō – spent much of their time engaged in policy issues beyond those that there were routine for their office, and from the 1780s were increasingly involved in protecting sakoku.21 As was customary, they would have retained their papers on leaving office. The Kanjōsho faced growing difficulties, political as well as economic, in the 1860s. The recruitment, unprecedented in its scale, of around fifty kanjō bugyō (many of them, moreover, serving for short periods) in the years 1860–1868 hints at the magnitude of its problems. While some of the new post-holders may have been appointed to handle the novel issues posed by dealing with foreign residents and their governments, others had to struggle with the deepening challenge of reconciling the falling income of the 1860s with rising defence expenditure, the effects of acute harvest failure in 1865, inflationary pressures, and currency matters arising from the complex (and now semi-open) Japanese monetary system. A host of expedients – currency debasement, suppression of posts, reduced emoluments, delays in paying creditors – made it possible for the administration to continue to function, though the gap between income and expenditure is far from clear.22 Heavy compulsory levies (goyōkin) on the merchants of Osaka made a reappearance. One shogunal measure, planned in advance of Osaka becoming an open port, was the intended takeover of the management of the port. As Osaka houses were central to the finances of southern and western domains through loans on the security of rice cargoes, this proposal ensured that the rising gulf between tozama and Shogunate was economic as well as political. Fire is another archival hazard that must be given close consideration, as there has been a tendency to accept underestimations of more prosaic archival losses in institutions, which had little or no relevance to a new Japan. The widespread and largely silent bakumatsu archival collapse warrants further reflection. Fires had been common in Edo, as can be read in graphic dagregister accounts of the annual hofreis visits to the city. Kyoto and Osaka too had their own great fires. For Osaka, in particular, the results of fires and other obscure circumstances can be seen in the many gaps, and in some instances incoherent or incomplete texts, in the two great compilations of the first decade of the twentieth century. The extent of losses in fire is, however, subject to qualification. As far as the Kanjōsho is concerned, losses in its office in the honmaru (the central Edo Castle precinct or shogunal court) in 1859 were more than

xxxiv

EARLY JAPANESE TRADE

offset by the fact that its major store, or sōko ಴ᗜ, on the perimeter moat of the Castle at Ōtemonbashi, entirely escaped fire damage. The Momijiyama archive, despite its location close to the heart of the Castle, was left astonishingly unscathed by an unprecedented run of fires in 1859, 1863 and 1867, and again in Meiji 6 (1873). Its archive losses of precious books and papers on the latter occasion were all of items that had been transferred to the imperial court.23 In Nagasaki, runs of some individual document types for a significant number of years hint at the likelihood of habitually high survival rates; the most celebrated instance being, of course, the Hankachō ≢⛉ᖒ, a register of decisions on criminal charges from 1666 to 1867 (even allowing for the fact that many of the support files are missing). The records of both the Kanjōsho and the Nagasaki bugyōsho lost their relevance in the wake of changing governance. The methods and personnel of the Kanjōsho (confined to taxing rice in the tenryō, or shogunal domain, which accounted for approximately a quarter of Japan’s rice output) were out of place in a post-1868 Japan, in which domains and tenryō alike had been abolished, and in which the entirety of the country’s rice production became the object of national taxation. As Totman noted in his study of the 1860s, no records of actual land tax income appear to have survived for the years after 1864.24 In Nagasaki, the elaborate and overstaffed apparatus of a huge bugyōsho and Kaisho (‘Trade Office’), not to mention two corps of interpreters, were already losing their relevance well before 1868. As a result, the survival of the Dutch archives has been of singular importance to the study of Nagasaki’s trade. Apart from returns for 1709–1714, almost no original official Japanese documents for either the Dutch or Chinese trades have survived, and even copies drawn at the time from now-lost sources are few. By contrast, the Dutch archives, beyond recording the Dutch trade itself, also noted much or even all of the Chinese trade (a few gaps notwithstanding), and therefore provided the source for Nagazumi’s magisterial reconstruction of the Chinese trade. The details of fires and of the losses that ensued are not always clear. The conflagration in Dejima in April 1798, which destroyed two thirds of the structures on the island, is well documented. In this case, the Dutch records were saved by the prompt intervention of the Dutch themselves. The island’s interpreter post, on the other hand, was destroyed, and if Dutch accounts of the lethargic Japanese response to the fire are accurate it seems likely that its papers were lost.25 The losses would have been mitigated, however, by the fact that a central core of

INTRODUCTION

xxxv

trade papers was held or duplicated at another location by the Kaisho, and by papers that had been retained by interpreter families. Even better documented is the great fire in Yokohama in 1866, which laid waste to four wards in the city.26 It destroyed all of the city’s shogunal records, including the customs records of the port of Yokohama, and the copies of returns from other open ports sent to the ‘Superintendent of Customs’ in his general supervisory role. Here too losses were mitigated by the retention of copies of figures in the ports themselves, and by the obstinacy, in the circumstances fortunate, of the British consulate in holding out against moving from Kanagawa to Yokohama. The destruction of archives in the Tokyo earthquake of 1923 represented far and away the single most calamitous event of its kind in Edo’s history, but even in this instance much survived thanks to the Shiryō Hensanjo and Gaimushō. Moreover, although the Ōkurashō was itself burned down, ten volumes (1921–1925) of stray documents copied over the preceding forty years survived, since they had already been published or were being prepared at other locations for publication. ARCHIVES: RESIDUAL PROBLEMS IN INTERPRETATION

The story of Japanese archives has two distinct elements. The first is concerned with the loss and later recovery of records or in idiosyncratic Japanese archival practices, at least of copies. The approach adopted in the studies in this volume, in tracing a pattern of loss and preservation, rests on a large though not exhaustive mass of heterogeneous material. Its arguments can be pursued further. The second lies in the fact that the archives provide substantial supporting evidence for the underlying institutional strength of Japanese society. Despite widespread emphasis on archival losses, much more material had survived than is generally assumed, for all of the reasons set out above. My personal experience of studying in Japanese archives, indeed, has been one of moving from a focus on loss and confusion to a recognition that varied patterns of strength, as well as weakness, exist. These issues are in themselves independent of the delay in creating a national archive until 1884–1885, at the outset of five years of decisive and radical political and constitutional change. The great mass of material held outside modern archives is made up of three categories: the actual copies held by officials and inherited by families; copies made privately, often with whimsical titles and lacking

xxxvi

EARLY JAPANESE TRADE

signatures and seals; and archives abandoned at the Ishin. While the abandonment of archives in 1868 was real, even though the details remain hazy, it was not so much the archives in their entireties that were at risk than holdings relating to bodies that, in the wake of the Ishin, were no longer relevant. There is still more to learn, for example, about the collapse of the Kanjōsho (‘Finance Office’) and the disappearance of its papers, the absence of which increases the likelihood that census records existed for both 1852 and 1858. For Nagasaki, the records are sufficiently abundant to allow for a fresh appraisal, even though it will remain of necessity incomplete. The Nagasaki study of 1997, which lists a scatter of bugyōsho records across the northern half of Kyushu, certainly deserves to be more widely known, being as it is the parameter for further reflection on the impact of the ending of sakoku on archival survival.27 Sweeping changes in the control of trade followed the treaties of 1858–1859, which opened up trade and residence to Westerners, a transformation highlighted by the fact that the Dutch were no longer symbolically prisoners of Dejima. The Chinese trade – understood as trade carried on Chinese vessels – was not covered by the treaties and effectively ceased: no Chinese vessels were counted in the last seven years of the Tokugawa era. What we do know is that from 1859 the English enclave in Shanghai, now able to trade with Nagasaki, essentially provided a cover for Chinese traders, who entered the city on English vessels either as agents for British houses or to trade on their own account.28 The value of Nagasaki trade, according to the new statistics produced by Nagasaki officials at the beginning of 1860, was circa 8000 kanme (converted from dollars). Yet recovery of trade, though real, was deceptive. Nagasaki was, in effect, a peripheral location, and from the outset Kanagawa (Yokohama), later joined by Osaka (Hyōgo), conducted a much larger volume of business. The ‘new Chinese’ settled freely in Nagasaki, much as their Western counterparts did. The Tōjin yashiki, once the compulsory, seasonal Chinese residence, decayed over the course of the 1860s, whilst interpreters of Chinese may have begun to lose their employment even before 1868. Interpreters of Dutch appear to have fared better. The work of the English consulate with Nagasaki officials initially continued to be conducted in Dutch, and the nomination of nenban tsūji (interpreters of rapporteur rank) as late as 1865 and 1867 shows that the interpreter ‘college’ at least formally still existed. The consular reports, however, refer only to ‘custom house officials,’ and it is unclear whether these were new officers or former interpreters cast in a new role.

INTRODUCTION

xxxvii

Relevancy was a major factor in the survival of archives. Some categories of the 1850s and 1860s remained in good order right to the end of the Tokugawa regime. The most famous instance is, of course, the Hankachō, which dates back to 1666 and was passed on in toto to the new police, who ironically found no use for them. As the closeknit Chinese community and its management collapsed, some of their records found a home in the Seidō ⪷ᇽ. A temple, however, was not a natural place for commercial papers to be stored, and some of them may, in fact, have passed into private hands. This ‘deposit’, which even at the time did not comprise all of the existing papers, was still an uncharacteristic one in bakumatsu Nagasaki, and so was the first source to attract the attention of the professional rescue operation that began in the 1880s. Despite the fact that Dutch interpreters had, on the whole, fared better in the twilight years of the Tokugawa regime, their documents were also widely abandoned and dispersed over the course of the two post-1868 decades, with only the emergence of collectors and a growing antiquarian market providing a passage for papers into safer hands. These circumstances account for the huge gaps in, and general incoherence of, the surviving mass of archives. In some cases, papers owe their integrity to their retention in families of former officials. Amongst such collections, the most remarkable are the papers of two interpreter families, the Motoki and Nakayama. The large Motoki archive experienced mixed fortunes over time, including some losses and separation of material.29 The Nakayama papers, by contrast, formed a compact and comprehensive collection, and were later donated by descendants of the family to the Siebold Museum in Narutaki, site of the celebrated Siebold school of the 1820s. JAPAN’S RELATIONS WITH EASTERN ASIA AND WESTERN COUNTRIES: GAINS AND LOSSES

Leaving the archives aside, there remains the challenge of explaining the conduct of a country that had opted out of relations with the wider world in the 1630s and, when forced reluctantly to open its doors in the late 1850s, gradually moved towards aggressive engagements with neighbours and Westerners alike. Was this due to an internal flaw within Japanese society, in which the seeds of what was later to become militarism were sown, as Maruyama has argued, or was the Japanese response dictated by new menaces in the East, especially British interference in China from 1839, and later Russian encroachments in Korea and Manchuria?

xxxviii

EARLY JAPANESE TRADE

The relative calm of the country’s internal affairs in the 1880s – which saw the formation of a cabinet government in 1885, the creation of national archives in 1884–1885, the approval of a constitution in 1889, and the introduction of a ‘modern’ (that is, Western-style) system of justice – contrasted with growing concern about external problems within the Asia-Pacific region. Local factions in Korea, for example, encouraged opportunistic Chinese and Russian interventions, which in turn led to Japanese retaliation, and ultimately to wars with China (1894– 1895) and Russia (1904–1905). The railway building that was taking place in Manchuria over these same years, with its south-bound thrust and many implications, also fuelled fears. In the light of these developments, it is not surprising that the moderate wing in Japanese politics at times lost ground to more reckless elements, such as those that were to make the notorious twenty-one ‘demands’ on China in 1915, which were only watered down – by the omission of seven demands (the socalled ‘Group V demands’) following public outcry. In the new postVersailles Treaty world, interventions were made in the late 1920s to protect Japan’s commercial interests in Shantung (China), which was followed in the early 1930s by the creation of a puppet state in Manchuria, condemnation of Japan by the League of Nations, and finally by Japan’s withdrawal from the League in 1933. Although the United States had no territorial stake in China, it did have numerous merchants and missionaries settled there, and so was an aggressive partisan of the Open Door Policy, which called for a system of trade in China that was open and equal to all countries. With very unclear aims, Japan then invaded China in 1937, and the countdown to war in 1941 followed. Did Japan really have a choice between moral leadership in East Asia and participation in the scramble for power in the region, as is argued in a paper in this volume? Japan was, at this stage of history, unique in being the only Eastern country that had totally preserved its independence, but had conceded extraterritorial rights in treaties of 1858–1859. The recovery of these was to remain the central aim of foreign policy for three decades. The frustration of such efforts in the 1880s profoundly coloured official thought, and stoked widespread xenophobia in the second half of the decade.30 Extraterritorial rights ended in the 1890s but, faced with the alternatives of either speaking for Asia or carving out enclaves on the Asian coast, like Western countries, it chose the latter. Its activities even had Western encouragement; hardly surprising when one considers how welcome was Japan’s defeat of Russia, the external power best-placed to interfere in the affairs of

INTRODUCTION

xxxix

the landmass to the west of Japan, and how convenient was Japan’s intrusion in Korea, which secured internal order and helped to forestall further Russian and Chinese intervention in Korean politics (no matter that the price was war). The idea that Japan might emerge as leader or defender of Asia, improbable though it may now sound, gained added weight in the face of Western jostling on the crowded coasts of the East China Sea. Japan’s defeat of a Western power had encouraged a not insubstantial number of Chinese students to study in Japan. Even Sun Yat-Sen, founding father of the Republic of China, spent time in Japan, and it was to Sun’s relationship with Japan that that distinguished scholar of Japanese history, Marius Jansen, devoted his first book. Satow speculated in June 1918 that if Japan cast itself in the role of defender of China, a great role lay ahead of it.31 Matsukata Masayoshi ᯇ᪉ṇ⩏ (1835–1924), a former finance minister and prime minister, had already made a similar comment in 1916.32 These were fleeting moments, however, since Japan was to negotiate for ratification of its gains in China and in the German islands of the Pacific in 1919. The resulting tensions between Japan and the United States led to a sharp rise in friction, causing Satow to write to one of his former colleagues that ‘I cannot help fearing that some day these two countries may have a war.’33 Forces of both moderation and military adventurism had been present in early Meiji and Taisho Japan, and Makino Nobuaki, who had effectively been the Japanese negotiator at Versailles, still optimistically believed in 1932 that moderation would prevail.34 But his hopefulness, and that of Joseph Grew, the American ambassador who recorded the conversation, proved out of place. Japanese animosity towards jealously protected Western enclaves in China, and in the face of the new Western hegemony in the Pacific, was the major long-term determining factor, deepened by simmering resentment of American opposition to the Racial Equality proposal of 1919. The more immediate one was Japan’s occupation of Manchuria in 1931, its rejection of Western condemnation, and its ensuing withdrawal from the League of Nations. Overconfident and reckless actions by a series of unstable cabinets finally paved the way to war. It is clear that Japan was no innocent victim when it entered the world stage, whatever the provocations of Western policy The ending of the Second World War had a happier outcome for the world than that created by the Versailles Peace Conference a quarter of a century earlier. For Japan, this result was in part due to General MacArthur’s remarkable achievements as proconsul, but it also reflected

xl

EARLY JAPANESE TRADE

the role played by Joseph Grew in heading the Japanese section of the State Department in 1944–1945.35 The fact that many in 1930s Japan had counted on moderate interests in the end prevailing over military factionalism ensured that outright demonisation of the Japanese was avoided. As with the Marshall Plan in Europe, this was a high point in the modern history of enlightened American self-interest. In the 1930s, scarcely a handful in the West spoke or read Japanese. Higher-level studies were almost non-existent, and serious books were few, work on literary texts aside. The effect of World War II and its aftermath was to generate a need for language expertise, which led to the development of crash courses, and in turn to the creation of a new breed of specialists in translating, interpreting and reporting. Many of them were already, or became, scholars and strong advocates of Japan. Donald Keene, Louis Allen, Ronald Dore are among the better-known names36 Edwin Reischauer, son of a missionary in Japan who had served briefly under Grew, created the Harvard Institute of Japanese History in 1973, in the wake of an ambassadorship in Japan. At the same time, a sizeable increase in visits to the West by Japanese students and professors fostered a flourishing tradition of Western studies in Japan. A more modest example of cultural exchange can be found in the experiences of Japanese who chose to serve in kitchens in Italy, bringing their newfound skills (and conversational knowledge of Italian) back to Japan. Japanese prestige peaked at the end of the 1980s, in a curious parallel to the allure that the country held in the 1880s and 1890s. The bursting of the economic bubble in 1991 led to a future of relative stagnation in Japan, eerily foreseeing the financial crash that would hit the West in 2007. If relations between Japan and the West have been close, with old rancour now largely forgotten, Japan’s relations with its neighbours remain unstable. A peace treaty with Russia is still unsigned, North Korea is a small but menacing nuclear power, and visits by government ministers to the Yasukuni Shrine to Japan’s war dead, along with their lacklustre apologies for the past, have time and time again caused outcry in Korea and China. Finally, China, weakling of the 1930s, has replaced Japan as the second largest economy in the world, and has alarming pretentions to control the China Sea and the islands in it.

PART I

Interactions – Ancient and Modern

Source: The Transactions of the Asiatic Society of Japan, fourth series, Vol. 18 (2004), pp.17–31

1

Sakoku, Tokugawa Policy, and the Interpretation of Japanese History

™

Japanese history poses problems of sources, because many sources have been destroyed, and in some instances were not even preserved in the form of the originals at the time. Copying was sometimes conducted haphazardly, but in some cases, careful and regular copying into journals was carried out. Some of these journals were official, but many were personal to the office-holder, and if there was an official record it was made by further copying of the record he had retained, and was then circulated. This mode of proceeding, however, meant that there was less a centralised general record than private or semi-private ones maintained by officials and in many cases retained by them or their descendants. In some cases, there appear to have been multiple copies, as on the occasion of the Rezanov expedition (the Russian embassy which sought trade with Japan) in 1804 which absorbed Japanese attention for months: that visit still awaits full study, and the effort spent on it should be repaid by the light it throws on Japanese administrative methods and record-keeping as much as on the political story itself. Much administrative work simply took the form of copying, and all foreigners from the seventeenth century onwards who made contact in some official capacity commented on the number of note-takers and the fact that everything was noted down. This method of proceeding also accounts for the constant going back over subjects already covered, which so exasperated foreigners. The pattern is amply confirmed in 3

4

EARLY JAPANESE TRADE

surviving Japanese records: one finds in them a minute recording of everything, and, apart from the record of negotiations, they reveal a curiosity about details, some of them apparently trivial, not only with regard to ships and armaments, but to dress, tools, weapons and uniforms. Sketches, often coloured, confirm that at the time the copyists sketched as well as wrote. The physical structure of offices needs to be taken into account. Then as now subordinate officials, as we can see from illustrations of Tokugawa-period work spaces, tended to work in a large room under the supervision of one man at its head. Moreover much of the work consisted of copying documents. In Edo Castle, the senior officials, the metsuke and kanjō bugyō, were accommodated in two separate spaces or rooms. A clear picture of how this structure worked in detail awaits the completion of further study. What is clear however is that there was not a structure of departments or ministries (themselves internally compartmentalised) such as existed in the West. Instead, there was a fluid system, in which officers were deputed to tasks as necessary, and subordinate workers were drawn from what was a pool. In other words not only did senior officials shift from task to task as occasion demanded, but they did not have the advantage of a team of regular subordinates whom they would command indefinitely and who through a lifetime would build up a familiarity with all the details of the work. It is not surprising that in such a loose structure heavy responsibility fell on the shoulders of the metsuke, who had to cope with wide-ranging and ever-changing challenges. Administration consisted of a handful of heavily overworked senior officials, and a large number of lesser officials, mostly underworked. In other words, there was both undermanning and overmanning, These observations, in some respects tentative, are prompted by my own current work, which has moved on from a study of the broad lines of political and economic life to a study of the bureaucracy and its functioning, especially that in the period when officials began to realise in a slowly unfolding external context from the 1790s onwards that Japan had to confront a foreign and ultimately major challenge.1 The system could work very well, as in the case of the survey work in Sakhalin and the southern Kuriles from the late 1790s made necessary by Russian encroachment: the task itself was well-defined, conducted in remote centres and hence free from interference on a day-to-day basis by vested political interests, and the team of some 200 men was led by a man of real ability, Kondō Jūzō (a middle-ranking officer

SAKOKU, TOKUGAWA POLICY, AND THE INTERPRETATION

5

promoted only later to high office). On the other hand, in 1839 in the smaller but more technical survey of Edo bay, which in effect took place around the capital for the purpose of selecting gun emplacements and deciding on the type of cannon, a flimsy structure of command led to a falling-out among opinionated individuals all of whom were very much out of their depth. The best known office in Japan, and in many respects the one with the most elaborate internal structure, was the huge Nagasaki bugyōsho. We know a good deal about it, and uniquely for Japan some of the internal evidence of its functioning survives. There exists in the records of the bugyōsho a very attractive series of sketches of its structural arrangement and also much detail about its staffing (the main site has been excavated, and a new museum recreating the central office now stands on the site).2 There was some centralised keeping of records, as illustrated by a goyōheya-hiki-tsugi mokuroku surviving from 1850, the sort of list that an official made when he was entrusting the appurtenances of office to a successor.3 Interesting though this document is, it has details of relatively few documents of earlier times, and some categories are entirely lacking. It has been suggested that such documents were held in other sections of the bugyōsho. If so, it illustrates the loose pattern of record-keeping, especially as much material seems to have been held independently by officials, and documents as sensitive as the fūsetsugaki (reports which from 1644, upon the arrival of their vessels, the Dutch were required to make on events in the outside world, and which were translated into Japanese by the interpreters for transmission to the shogunate in Edo) seem to have been held mainly by the interpreters, as was all the correspondence with the Dutch. The interpreters, Dutch and Chinese alike, acted in collegiate fashion and also maintained for each year a regular diary of business. Diaries of only two isolated years survive for the Dutch interpreters; Chinese diaries later than 1715 are unaccounted for. Of ten remaining diaries copied in the 1880s none of the originals have survived. A trade in the 1860s conducted by British and Chinese traders on board British vessels sailing from Shanghai, seriously disrupted the status quo, undermining the role of the interpreters and leaving archives at large at risk.4 We have however some idea of the Dutch intepreters’ archives through two collections of papers of interpreters, the Nakayama family records now in the Siebold Kinenkan in Nagasaki, and the huge Motoki collection, divided between Nagasaki and Kobe.5

6

EARLY JAPANESE TRADE

Rich though the surviving store of records is, given a society devoted to copying, the diffuse character of the material and the fact that much has disappeared, means that there is less of a centralised record than there is for Europe. In some respects, given the hazards of fire in wooden buildings, this means that more has survived than if there had been a centralised Western-style system. The exception to this form of insurance, if one can describe it as such, is the Nagasaki office, which in one way or another maintained a great number of papers to the end. It was favoured by the fact, uncommon in Japan, that its last big fire occurred in the 1660s. For all these reasons, any generalisation about the nature of this society and the purposes of its public policy has to be framed carefully. However, no less important than the problem of sources is the problem of stereotyping, made all the easier amid the gaps in the record by the readiness of many historians, for several reasons, to bridge them with supposition. Stereotyping of Japan began in the West in earnest when, as an accompaniment to the final exploration of the north Pacific, the wish to seriously challenge Japan’s sakoku began to emerge around 1820. Concrete knowledge of Japan, already slight, remained limited. Western encyclopaedias are an interesting measure of this, judging by how little space Japan took up in their pages and how knowledge continued to be based on little new information until the 1850s. However, if the intelligence itself was limited, the perception up to the 1820s of Japan, if vague, was often positive. From this point on, it began to become negative. This can be seen graphically in the English edition, published in 1819, of Golownin’s Recollections of Japan. The long editorial introduction reads almost as if the editor had not read Golownin’s sensitive text.6 Ironically, this new and commercial curiosity about Japan was more deadly to understanding than the relative indifference which had characterised the eighteenth century. The Jesuits a hundred years before Kaempfer had been better disposed, and their accounts, though lamenting in particular Japanese sexual morals, tended not only to be favourable but to treat of Japan from the perspective of an equality between Japanese society and their own. While in his famous book of 1727 Kaempfer’s comments, in what are in effect a transcript of diaries he maintained while in Japan, are critical of sakoku, his observations in an appendix, written years later upon reflection in Europe, were more friendly (and were translated into Japanese in 1801 as part of a famous defence of the policy). The overall picture he gave was favourable:

SAKOKU, TOKUGAWA POLICY, AND THE INTERPRETATION

7

later and more hostile commentators attributed similar conclusions reached by the few visitors not to observation of facts on the ground but to mere copying of what Kaempfer had written. If a negative tone now permeated writing on Japan, a further element had also entered into the equation: rivalry among Westerners in an empire-building race which was beginning in Asia. What is striking about Hawks’ account of the Perry expedition – the official report published in 1856 – is that it was if anything more hostile in its comments on Western countries than to Japan, which in a patronising way were often favourable.7 The purpose behind American interest was commercial, and the astonishing depth of hostility directed against other Westerners in the report reveals the real purpose behind American interest in east Asia. The argument that the United States had no interest in Japan other than fair treatment of mariners and the provision of coaling stations for vessels on the long route from San Francisco to China is not convincing. The report also had – and continues to have – the negative result that the comparatively successful modus vivendi established between Japan and the West during the 1850s and 1860s was simply seen as the consequence of Western forbearance (dictated by the fact that its major interest lay in China) and not something in which Japan’s initial understanding of the stakes and its own successful diplomacy played a large part.8 The negative dimension of a policy of isolation (in the past often viewed in a benign way) was increasingly denounced by mid-century: it was seen not only as a departure from what were declared to be civilised norms, but as a policy maintained solely by despotic power. In later days, this stereotyping was enhanced from the inside by the lack of sympathy for the Tokugawa era shown by the new Meiji regime, and it carried over into accounts by Westerners in Japan and is re-echoed in the pages of early issues of the Transactions of the Asiatic Society of Japan. It was then deepened further in the 1920s and 1930s by the progressive spread of a Marxist approach in the universities, and by the consequent urge, in a travesty of reality, to see past history solely in terms of poverty, oppression and ikki (unrest). Western Marxism then embroidered the story further, E. H. Norman’s facile and overrated generalisations setting the fashion. While his account was of course, whatever its qualities, the first analytical account in English of modern Japanese history, yet, resting as it did on the negative and Marxist picture found in histories written by Japanese in the 1930s, it was not an independent insight into Japanese historical circumstances.

8

EARLY JAPANESE TRADE

Maruyama Masao, a friend of Norman and a semi-Marxist himself, further added to the dismal picture in the post-1945 years. Marxist interpretations in Japanese universities were to remain dominant until the generation of young teachers who had come to the fore in the second half of the 1940s reached retirement: unevenly the conclusions reached in a huge corpus of earlier writing still linger on. These Marxist-style interpretations were reinforced, ironically from entirely opposite motives, by the early American historiography beginning with the Occupation, which was coloured by the Cold War and the confrontation between the US and the Soviet Union. Early post1945 study of Japan by Americans rested firstly on the urge to create a Western-style democracy in Japan using the shaky argument that Japan in its history had had a covert tradition of political dissent, and secondly resting on the assumption that Japan’s development after 1868 had sprung from a capitalistic modernisation process, hence from a model of economic development which could counter the Indian model drawn from the Soviet Five-Year plans. Further, a whole mythology, sustained by a highly negative image of Japan, persisted about obsessive secrecy and its ruthless enforcement within Japan. In particular maps, it was said, were not to be shown to foreigners; the metsuke were primarily spies whose role was to ensure compliance with the rules of a despotic shogunate; foreigners were prevented from learning Japanese. The reality was more complex. Maps were surprisingly readily shown to foreigners. Only in the case of maps of Ezo did this erupt into conflict in the late 1820s. This was a sensitive issue for security reasons, as Japanese officials were well aware of Western interest in these waters: the sudden appearance of Western whaling vessels from the 1810s was also an unsettling issue. Contact with the rough, sometimes violent and uncooperative crews was also bedevilled by problems of interpretation: were they distressed mariners as they claimed, or were they spies? There was a real division of opinion on this issue. When Mogami in 1826 showed a map of Ezo to Siebold, he laid such an emphasis on secrecy that Siebold entered the conversation in his diary in Latin. As for the metsuke, they were not spies. Along with the kanjō bugyō they were the linchpins of the administration. They were senior officials with heavy responsibilities, and often, moving from one responsibility to another, gave the system a flexibility that it would otherwise have lacked. It has also been argued that the Dutch were prevented from learning Japanese. However, the reality is more complex. The real

SAKOKU, TOKUGAWA POLICY, AND THE INTERPRETATION

9

problem was that the opperhoofd (head of the Dutch factory) was limited to a year’s service at a time in Japan (a requirement introduced by the Japanese in the early years of sakoku), and though he could return with a later fleet, he had in the case of a single visit to leave within a year (a problem which the Chinese at large also encountered and from a somewhat earlier date). However, some of the subordinate officials in the Dutch factory who, unlike the Chinese and the opperhoofd, were not affected by this requirement, spoke Japanese very well. For an illustration of the divergence of views in modern scholarship raised by the problem of stereotyping and interpretation, it is instructive, for instance, to compare Professor Bodart-Bailey’s views in her review in the Australian journal Japanese Studies with the arguments in R. H. Hesselink’s Prisoners from Nambu: reality and make-believe in seventeenth-century Japanese diplomacy. The few foreigners who had known Japan had been well-disposed. Thus Isaac Titsingh, opperhoofd in 1779–80 and for a stay in 1781-2 extended for a further year, though impatient with Japanese bureaucracy and no pushover in his dealings with officials, got on very well with them. He was an intellectually exhausting man, as we can sense from his writings. His long account of life after death, which he had prepared for translation for the head of the Kaisho or trading office, still survives in the original Dutch.9 It is a demanding piece of reading even for a Westerner reading in one of his own languages, and its translation would certainly have made heavy demands on a translator. However Titsingh’s readiness to treat his counterparts as intellectual equals also meant that he was genuinely liked in return. He argued that if there was little contact, the reason was failure on the part of the Dutch; much of his later life was devoted to compiling accounts of Japan from the great mass of material he had accumulated. As his treatise on life after death shows, he was a man not put off by complex issues, and the vast amount of information he acquired through the interpreters as a reward for his ability to communicate, is impressive. The Russian naval captain Golownin spent two and a half years in Ezo 1811–1813, and his account was very favourable to his guardians, especially to Murakami Teisuke, like Golownin himself a likeable man. Sir Ernest Satow (an admirer of Golownin’s writing), because of his own singular ease in mastering the language (and also drinking or carousing with officials in a way which the unbending Titsingh would hardly have approved of ), also got on well with them. He intensely disliked the habit of his boss, Sir Harry Parkes, of shouting at them.

10

EARLY JAPANESE TRADE

Some important firsthand accounts of Japan are, however, lacking in any underlying sympathy or deeper insight. Thus, Siebold (a man denounced incidentally in Hawks’ account and who made himself intensely disliked by Westerners after he reappeared in Japan in 1859), showed in his book published in 1835 a lack of any real warm feelings for Japan.10 His insensitivity to the delicacy of some of the issues and to the tragedy he was responsible for is remarkable. It is hard to avoid the conclusion that he remained silent by choice on an issue which did him no credit. He was, to put it crudely, a very ambitious man on the make. The other important account of these early decades is that given by Hendrik Doeff, a short and relatively factual account and one which is not as well known as it should be because until recently it lacked an English translation.11 It is by a man whose stay in Nagasaki was the longest uninterrupted one by a Dutch official, a fact occasioned by Dutch contact with Europe having been cut off by the war with England from 1795 to 1815. However, it may overemphasise his own initiative, and underestimate Japan’s independent reasons for its urgency, in what was a period of crisis, to improve its knowledge of the West. It is perhaps invidious to single out individual figures in Japanese history, as many played an important role, as did for instance the shogun Tsunayoshi, who is often seem as a dilettante. However it is hard to resist the temptation to single out two men: the remarkably curious and many-faceted Yoshimune (shogun 1716–1745) and Matsudaira Sadanobu, prime minister 1787–93, who had to face the famine crisis of 1786–7 and the Russian threat. Intellectually Yoshimune is in some respects unique. Though Japan was not a statistically-minded society (Japanese had begun however to appreciate Western leadership in mathematics), Yoshimune tried to create statistics, both demographic ones and also figures for the trade of Osaka, commercial nerve centre of Japan. This was prompted by administrative concerns for Japan’s welfare but also by personal curiosity. The Nagasaki deputy opperhoofd, for instance, noted in 1723 that he had been told by an interpreter that the bugyō “had received orders from court compelling him to conduct a census to find out how many people aged between 80 and 90 were living in Nagasaki.... Nobody knows why the [shogun] has ordered this census.”12 In 1734, Yoshimune also asked nine han to provide information about earlier population trends. While the census first launched in 1721, and relaunched in 1726 in a format which required it to be taken every sixth year, continued to be compiled, it became, as the chronic stereotyping in its presentation suggests, a mere soulless

SAKOKU, TOKUGAWA POLICY, AND THE INTERPRETATION

11

exercise by rote. The curiosity displayed under Yoshimune was never repeated, and no statistical innovation or even expression of statistical curiosity can be detected again before the end of the Tokugawa era. Yoshimune also gave an impetus to the study of knowledge of the West. It is the first case of personal shogunal interest in it, insatiable moreover, as the dagregister shows, when it came to the questions with which the Dutch were plied on their annual visit to Edo. Some of this interest was just narrowly curious, but its serious side laid the basis for systematic study. While the famous fūsetsugaki were state documents, modest steps in privately translating medical and technical texts began as early as the 1650s. Under Yoshimune this study finally got political backing. While from the Dutch accounts it can be seen that the curiosity had sprung above all from Yoshimune’s inexhaustible inquisitiveness, under later shoguns it was driven entirely by the doctors and astronomers. They acquired some conversancy with Western knowledge and, more importantly, supported its pursuit by others. The annual meetings with the shogun, however, lost the colour and at times the eccentricity which had formerly characterised them, and the Dutch reporting became routine (apart from occasional complaints of exhaustion from being plied in long sessions with questions by shogunal scholars). Matsudaira Sadanobu stands out for several reasons. Descriptions of him are often negative because of the false argument that his predecessor Tanuma was a proponent of the opening of Japan, from which Matsudaira allegedly retreated. Matsudaira was beset by problems, economic and political. Against shrill exhortation for bolder steps, he created an enduring pattern of avoiding confrontation in Ezo with Russia. He defined what the future foreign policy was to be for several decades, and his action seems to have been justified in view of the fact that in the 1810s Russian interest in the Kuriles receded for a whole generation. He boosted political study of the West: under him one can see the first evidence of a shift from an intellectual interest to a political and even military one. He aimed at the systematic training of officials: his famous decree on heterodoxy in 1790 was directed not against thought itself but against the teachers in the Hayashi academy pedantically luxuriating in the minutiae of academic hair-splitting. He proposed kakari to deal with external problems, a step reactivated with great success in the mid-1840s. Sakoku itself was a defensible policy as long as the retreat of Westerners from the waters around Japan in the 1630s made its pursuit

12

EARLY JAPANESE TRADE

realistic. For decades at a time not a single Western vessel (except for Nagasaki–bound vessels) crossed the horizons of the seas around Japan. This reassuring context began to change in the late eighteenth century, when Russian encroachment in the Kuriles came dangerously close to Japan. In 1793 the Russians sought trade; and the French under La Pérouse in 1787 and the British under Broughton in 1796–7 launched surveying missions. The learning of Russian became important in 1804. The curious semi-declaration of war delivered in a French text by some maverick Russians in the Ezo islands in 1807 seemed to make a knowledge of French necessary. The intrusion of the English warship Phaeton into Nagasaki in 1808 added English to the list of necessary foreign languages. In 1811, two major steps were initiated on the language front. The first was that in Nagasaki, from 1811, the interpreters started drawing up their famous Haruma or Dutch-Japanese dictionary. Secondly, some of the interpreters had already been sent to Edo in 1807–8, to translate urgently from Dutch works on both Russia and the West. This step anticipated the foundation in 1811 of the Bansho-wage-goyō (Office for translating foreign books). The staff of the Bansho-wage-goyō consisted of some 30 officials in all. The greatest Dutch expert in the office was Baba Sajūrō (1787–1822), a young interpreter transferred to Edo as the brightest star of the Nagasaki interpreter corps. As well as the Dutch required in his profession, he had acquired some modest command of English and French in Nagasaki, and in the autumn of 1811 found himself charged with deciphering the letters in Russian found on the Russian captives in Ezo. While the Russian threat receded from the 1810s, it was replaced by the pressing and novel one of English-speaking whalers. No less than ten instances had been noted between 1818 and 1822. To deal with the problem, the questions drawn up in Dutch for putting to foreign mariners were translated by the versatile Baba Sajūrō into a halting English, and he, we know, also interviewed whalers on several occasions. Couched in polite terms, the list of English phrases set out the prohibition on coming to Japan, offered fuel, water and food, and instructed the visitors to leave without further delay and not to return. The text included a small glossary, beside their Japanese equivalents of words in English and Dutch both for a wide range of foods and for the practical requirements of the anticipated contacts. Japanese policy was driven by caution. Despite keeping the uchi harai decree against Russia on the books (and this was of course the

SAKOKU, TOKUGAWA POLICY, AND THE INTERPRETATION

13

occasion of the confrontation with Golownin in 1811), the Japanese displayed a remarkable amount of pragmatism in dealing with the Russians. It was the advent of the whalers (all English-speaking) in the late 1810s and early 1820s which created a problem for which the Japanese had no precedent, and to deal with which they were poorly equipped. Because whalers lacked the easy identification provided by flags and the insignia of officers’ uniforms, with which the Japanese already had some acquaintance in their Western dealings, these vessels did or at least could appear sinister. Somewhat surprisingly, given the importance the problem had assumed in Edo policy-making in 1824–5, the whalers momentarily disappeared from Japanese waters (at any rate to the extent that Japanese preoccupations on the score ceased, as far as we can judge from the surviving documentation): the appearance in 1837 of the Morrison, an English-speaking but nonwhaling vessel, is the sole disturbing episode between 1825 and 1844. The need for a knowledge of the English language, which emerged in 1808 and was evident again in the early 1820s, declined; the modest competence among the intepreters evaporated. However, in the mid-1840s when foreign vessels, whaler and non-whaler alike, became numerous, acquisition of English became urgent once more. Moriyama, the main interpreter for the American visits in the 1850s, had picked up his English from Ranald MacDonald, this time a real infiltrator of the late 1840s (carried to Japan aboard a whaling vessel). One American author has made the sweeping statement that the uchi harai order of 1825 was “little more than a tired restatement of Tokugawa isolationism, which revealed the incapacity of the bakufu to see beyond the immediate implications of events”. The famous uchi harai policy (firing on and expelling foreign vessels), though formulated as a concept in 1793, became applicable by decree only in 1807 and was at that stage confined to Russian vessels. The term itself was an older one, applied to driving off Chinese vessels hovering off the coast for the purposes of smuggling. The policy was extended, because of the whaling problem, to all Western vessels from 1825, and when it was seen that it could prove dangerously provocative, it was amended in 1842 to admit of succour to the crews of vessels in distress, and a proposal in 1848 to restore uchi harai in its 1825 form was rejected. The years 1806, 1807, 1825, 1842 mark stages in the defining of policy, just as the 1848 decision confirmed adherence, in the face of opposition, to the modified policy. These instructions given in 1842 were perfectly clear in the case of a vessel calling and asking for fuel, water and food.

14

EARLY JAPANESE TRADE

But what was the position of people who were already ashore: in other words, those who did not present themselves while still aboard a vessel or in its wreckage, politely or desperately asking for help? Were mariners who were not identified at the moment of landing and who made their way inland or across territory infiltrators? Officials made very heavy weather of this as the interrogations show – their immense detail shows how much time this problem absorbed – and also varied in their responses. To some extent, the stand maintained in 1848 was itself a response to those who had wanted to deal with the dilemma by a restoration of unlimited uchi harai. When the real challenge came in 1853 and 1854 with Perry’s two visits, the Japanese were surprisingly able to deal with it, both in terms of interpretation, and the sophisticated handling of the issues. Four observations summarise the situation. (i) The Japanese had prepared for the confrontation from the 1790s, and were able to face the challenge. Knowledge of the West had grown, and, after the factionalism so evident in the late 1820s and more publicly in 1839, this learning, or rangaku as it was called, was finally depoliticised. As events in 1839 proved, much knowledge was held by opinionated or tactless men, who did as much to cause the problem as did other Japanese, conservative and equally tactless, with whom they worked or whom they foolishly confronted. (ii) They negotiated very well, and the whole period 1853–1868 has to be seen as a single sequence of negotiation. (iii) Divides were not over principle (no one wanted to open the country) but over the risk and likely outcome of war if concessions were refused. (iv) These divides, on points of real substance, were complicated (a) by factionalism over the 1858 shogunal succession which divided both daimyo and their senior officials into two camps, and (b) by what over several years emerged as a point of real political substance: would Japan face the challenge as a more unitary country or as a confederation of han. Control, whether by shogunate or han, of the wealth of Osaka, when it would, in accordance with the agreement reached in 1862, finally be opened for trade on 1 January 1868, also became a key point from the 1860s.

The Japanese public sphere is at first hard to comprehend because all business was private: in other words there was not a public sphere. However, such a lack of distinction was unworkable in a society as sophisticated and complex as Tokugawa Japan, and one where the

SAKOKU, TOKUGAWA POLICY, AND THE INTERPRETATION

15

circulation of copied documents in itself ensured that many individuals, often those remote from high office, were familiar with the issues. In the West by contrast, a defined boundary between the two spheres existed and a clear-cut policy operated. Sensitive documents were marked secret; some were even in code. Administrative documents at large were also kept in neat chronological and – frequently – numbered order within the offices of government ministers and high officials. They were not circulated, and copying was made only with authority and for express administrative purposes (the documents copied also tended to deal with routine matters, not, as in Japan, with topics of high sensitivity). In Japan, the system of maintaining a private sphere for state business was unworkable. There seems to have been no clear distinction between copying required by raison d’éat and for reasons of private curiosity or convenience. The pattern even within the administration of the ceaseless copying of documents was, in other words, unstoppable. Many documents, moreover, were kept by the “colleges” of Chinese and Japanese interpreters. The result was that these were often multiple copies even within official circles, and were loosely held either by individuals or at varied locations. The Tsūkō ichiran (a vast compilation of documents concerning foreign policy) was compiled from 1849 onwards, the first section up to 1825 being ready by 1853, the second section by 1856. It was compiled by a team of twelve men working under the direction of the formidable and polished Hayashi Akira, not from official documents in the Western sense, but from a large number of compendia of documents. The Tsūkō ichiran acknowledged its sources, and it is possible to reconstruct the range of material the team used. Many of the records have since disappeared, but the Tsūkō ichiran is an illustration of how a system of decentralised and multiple record-keeping worked. Golownin’s detailed Narrative of his Japanese experiences itself enters into the story at this juncture.13 When it was translated from Russian into Dutch, a copy was received in Nagasaki. The interpreters lost no time in translating it (the overworked Baba Sajūrō began the task but died before he could complete it). There are at least seven copies in existence,14 and that is without taking into account its reappearance in the pages of the Tsūkō ichiran, where it extends over many volumes and is interleaved with the story as set out from Japanese sources. As we come to study administration, we can see how numerous and decentralised the sources were. The one thing that was not circulated widely was stadstics. It has been assumed that this was because of

16

EARLY JAPANESE TRADE

secrecy, but that is not the reason. Copies of the population statistics (which in modem times have made it possible to rebuild the profile of the population) seem almost invariably to have been held by individuals in a private capacity, and in one case details of the complete census for two years were held by the Machida family, a family so obscure that the name does not occur in biographical dictionaries. A general lack of statistical curiosity (the curiosity that Yoshimune was the first and last man in public life to display) is the reason. The demographic figures were never put in tabular form, so that knowledge of them circulated in a strange flotsam-jetsam of isolated and sometimes incomplete copies. There is also an interesting history to relate about the dispersal and later partial reconstitution of Tokugawa records. Where records have come together again, as in the mass of paper at Nagasaki, the largest surviving collection of Tokugawa administrative documents, it is frequently impossible to distinguish between originals which had once lain in the Bugyōsho and copies made and held privately in late Tokugawa times. In 1868 when the new Meiji politicians took over the reins of government, documents for the 1850s and 1860s were to a large extent retained by the new administration. Earlier documents seem to have been abandoned or sold off. The Hankachō, records of criminal process under the bugyō, were given to the police: they then sold them to a merchant of second-hand wares. Later they returned to public ownership, and are to-day the most impressive single series of original official documents in Japan (uniquely for Tokugawa records, almost complete from the 1660s, the decade of the great fire in the bugyōsho).15 In the Meiji period, individuals acting on private initiative, including one man who was an official, salvaged a large corpus of material variously abandoned by officialdom or already in private possession. In other words the story of the records in Nagasaki is less a case of an archive surviving into modem times than of the salvage of a great mass of paper of varied provenance. That explains why those fūsetsugaki which are now in the Kenritsu Toshokan in Nagasaki are such a curious mixture of carefully crafted documents and very cursive copies. Fortunately, Japan, a society addicted to recycling paper, had a highly developed antiquarian sense. The antiquarian value of the paper was probably low, but high enough to ensure its survival until first in the 1880s a city ward head and then and later schoolmaster érudits, serious collectors of old documents, came along. To an administrative historian, they, as much as the bugyō and interpreters of Tokugawa times, are heroes.

SAKOKU, TOKUGAWA POLICY, AND THE INTERPRETATION

17

Louis Cullen is professor emeritus of modern Irish history at Trinity College, Dublin. He has been a visiting professor at Hosei University and was a visiting scholar at the International Research Centre for Japanese Studies from 2002–3. His research interests are Irish, French and more recently, Japanese history. His latest books include The Brandy trade under the Ancien Régime: regional specialisation in the Charente (Cambridge University Press: 1998) and A history of Japan 1582–1941: internal and external worlds (Cambridge University Press, 2003).

Source: Asia and the History of the International Economy: Essays in Memory of Peter Mathias, Eds. A.J.H. Latham and Heita Kawatsu, Routledge, pp.72–94. Reproduced with permission of the Licensor through PLSclear

2

Knowledge and Use of Japanese by the Dutch on Dejima Island, Nagasaki ™

INTRODUCTION

The view that the Dutch were prohibited from learning Japanese has long been repeated. However, there is no contemporary evidence of a prohibition. While a few contemporary Dutch references appear at first sight specific, their context may imply some doubt or qualification. The most comprehensive modern statement is that by the respected American historian Jansen: It was Japanese policy to discourage the Dutch from studying Japanese lest the outsiders get too close to those they would contact. On a number of occasions representatives of the VOC (Verenige Oost-Indische Compagnie or United East-Indies Company) were told to leave Japan because their knowledge of Japanese was becoming too good. The Japanese wanted to keep the contact on their own terms.1

Jansen’s ‘a number of occasions’ needs to be qualified. They were not numerous. Apart from cases in 1669 and 1670 and in the years 1739 and 1741 (three individuals in two instances), the sole other case was that of the medical doctor in 1718. The better-documented case of Domburg in 1784 was entirely in a different category. The argument – assertion of existence accompanied by elusiveness of any concrete evidence – is typical in the sense that it accorded with other undocumented prohibitions: for example, the argument that maps were not to 18

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

19

be shown to foreigners. In fact, they were, on many occasions. In the wake of the Laxman visit to Ezo, a geographer in a group of shogunal doctors, and geographers visiting the Dutch at their inn, showed maps of Japan he had been given by some of the officers on Laxman’s vessel.2 The (Von)Siebold affair revolved in part about sensitive maps of the Ezo region, and their export: the explorer Mogami had intimated to Siebold the sensitivity of the topic, so much so that Siebold recorded Mogami’s conversation in Latin in his diary.3 Arguments supporting the existence of various restrictions have seemed all the more plausible because the Dutch in Dejima complained often about restraints, variously serious and petty they suffered. Kaempfer, who in 1712 almost two decades after his Japanese sojourn as doctor in the Dutch factory had mellowed enough to include in his immense Latin text Amoenitatum exoticarum, a relatio which concluded with a remarkable section Regni Japonica status post occlusionen prosperissus) (the highly prosperous condition of the Japanese kingdom after its closure).4 This favourable account of sakoku contrasts with his critical comments drawn from what appears to be accounts or a diary kept at the time of residence in Nagasaki of the restraints both in Dejima and on the route to and from Edo. The first attempt to explore the universal belief of prohibition is a scholarly paper by Tanaka-Van Daalen in 2009 which drew on documents from 1642, 1675, and 1683, relating to issues of learning and to a Dutch wish that Dutch interpreters of Japanese should be involved in official dealings. Unable to find evidence of a prohibition from the Japanese end, “the Dutch”, she observed,“deemed it wiser to keep the practice[their use of the language] under wraps.”5 There are two questions, which this account raises. First, were the Dutch secretive in their speaking of the language, or was it simply that they used their knowledge of the language, such as it was, in secret contacts in search of commercial intelligence, to counter a prohibition from 1682 on providing them with details of the cargoes of their Chinese rivals in trade? Secondly, why was the belief that learning Japanese was prohibited so persistent, or, put more precisely what did statements more or less to this effect actually mean? Kaempfer did not mention prohibition; simply observing that the interpreters were “employed to assist us and to discourage foreigners from learning the language, so that they can be kept ignorant of and blind to internal conditions of trade.”6 Thunberg, writing of his 1775–6 year as doctor on Dejima stated that “the government permits no foreigners to learn their language.”7 The idea lingers vaguely even

20

EARLY JAPANESE TRADE

in Titsingh’s account in 1784 of the expulsion of Domburg. Doeff in his account of his long sojourn as opperhoofd (head of the Dutch factory) mentioned in 1833 the obstacle presented by “The existing law that no foreigner may ever learn the Japanese language” and Fisscher in his book of the same year referred to the prohibition, adding that Japanese could be studied only in the greatest secrecy.8 Yet by this date the language issue had been wholly dead for several decades. Meijlan (opperhoofd 1826–30) writing in 1830, the sole author to confront the language issue at some length, made no reference to prohibition, simply expressing annoyance that opperhoofden were obliged to rely on the interpreters and were not listened to if on formal occasions they addressed their remarks in Japanese directly to officers of bugyō (governor) rank.9 SHOGUNAL OPPOSITION TO USE BY FOREIGNERS OF JAPANESE IN TRADE AND DIPLOMACY

The concepts of prohibition and simply of discouragement are fairly close to one another. Thunberg may be correct in this sense, possibly drawing on recollection of earlier language-learning proposals that had come to nought. The absence of provision for the Dutch to learn Japanese (more precisely the refusal to help the Dutch to learn Japanese) contrasted sharply with arrangements which existed in the other direction for Japanese apprentice interpreters to receive instruction from the Dutch. As there were no learning aids, grammars or dictionaries, learning Japanese was difficult. Perforce it relied on oral instruction, and hence depended on the good will and assistance of Japanese. In reverse the Japanese would have made slow progress from Portuguese (along with Chinese, a lingua franca of the east), which the Dutch and Japanese alike knew, to Dutch, Portuguese lingered on in the hands of some Dutch and Japanese into the early eighteenth century. Books and grammars prepared by Jesuit missionaries did exist in Portuguese, but they were unknown in the wake of collapse of Portuguese society and more importantly, they were for Japanese user-unfriendly as they were written in western script. The story of the use of the Japanese language by the Dutch can be divided into three sequences. First, what appeared at least in retrospect to be the halcyon control-free years from 1609 to 1639 in Hirado. Second, the early decades in Dejima when they faced a new style of shogunal administration relying for many purposes on Japanese interpreters who were acquiring – or improving – a knowledge of Dutch. The Dutch

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

21

had by the 1680s conceded failure in attempts to have the principle of Dutch interpreters conversant with Japanese established. The attempt failed because the Japanese were adamantly negative on the subject. In the third sequence, the period from the 1680s to the 1730s, Dutch access to intelligence of the trade of their Chinese rivals was denied in 1682 and the prohibition was repeated in some form or other in subsequent years to as late as 1732. There was at first some real success in covert enquiry (for which knowledge of Japanese was indispensable). But success waned from 1708; and apart from intelligence received in 1718–27, the channel was closed off. The 1730s were finally to herald a thaw from mid-decade in official sanction, though as usual in Japan change was made in practice rather than promulgated as a principle. In the first sequence, the years in Hirado the Dutch were relatively little subject to regulation, and benefited from the usually benign attention of the daimyo of the domain of Hirado.10 Both Portuguese and Chinese, each a lingual franca of intercourse in Asia, served as the languages of communication. The Japanese language posed no problem; it was simply a supplementary linguistic aid, and the Dutch began to acquire knowledge of it. In 1639 Caron who had been in Hirado for eleven years (in his final two years as opperhoofd ) was at ease in Japanese, and another man H. Esserd, an under-steward was also said to be fluent. The second phase was that of the immediate aftermath of the transfer of the Dutch to the artificial island of Dejima in Nagasaki bay. The transfer was the final stage of implementing the sakoku policy of the 1630s intended to confine all foreign trade to Nagasaki under the close supervision of governors no longer drawn from local grandees but dispatched from Edo. Liaison between governor and the Dutch was henceforth maintained by a governor-appointed corps of interpreters with an office of their own on the island of Dejima intended to police both the trade and the ten to twenty Dutch inhabitants resident on the island. Business was itself to be conducted through the interpreters. For the Japanese there was a steep learning curve in mastering Dutch. While one would require more evidence, it is conceivable that, with Portuguese easing the pressure on language learning, at first the Dutch made more progress in learning Japanese than did the Japanese in acquiring Dutch. Not surprisingly the Dutch were uneasy: and even in the following centuries they felt that the language command of many of the interpreters was negligible. The Chinese also lost freedom they had formerly enjoyed. No longer free to establish themselves in

22

EARLY JAPANESE TRADE

locations around Kyushu, confined to Nagasaki they had to deal with a numerous shogunal corps of interpreters conversant with the Chinese language (necessary as kanji characters of themselves did not make Chinese comprehensible to the Japanese). The interpreters also became the enforcers of a strict regime. Eventually from 1689 they were confined to a yashiki or enclosure referred to by the Dutch as “the Chinese island”. Much more numerous (at times seasonally exceeding 2,000) the Chinese lived under a far harsher regime than anything experienced by the minute Dutch community on Dejima. The Japanese interpreters of Dutch were the eyes and ears of the bugyōsho (office of the bugyō or governor). A key problem was that the defective knowledge of the interpreters, at the outset learners, left the Dutch both uncertain as to whether their statements were fully understood and uneasy as to failings in translation from Dutch into Japanese. In 1642, the Dutch governor-general in Batavia complained in a letter to the Japanese government about the rigid controls they were now subject to. Faced with language problems, the Dutch sought to get around the restrictions during the tenure of opperhoofd Overtwater in 1642 in “secret talks about trade between H. Esserd, an understeward who was fluent in Japanese, and the interpreter Siroyemon.”11 Such efforts were destined to fail and the Dutch attributed the exclusion of Dutchmen in 1669 and 1670 to their fluency in Japanese.12 The price the Dutch paid for tolerance of their stay in Japan, the sole western presence to survive, was dependent status: this excluded them from equality in negotiation. Language was a central part of this process. Confinement of negotiations to the Dutch language ensured that the Japanese retained exclusive control of the formal record. i.e. final version or compte rendu in Japanese of what had been recorded by the interpreters in dealings with the Dutch. In other words diplomatic equality did not exist, and the Japanese decided the formal record unilaterally without any knowledge by the Dutch of what it said. The ease at an earlier date of the Dutch condition in Hirado could be exaggerated. Kaempfer saw the contribution of the Dutch to the campaign against Shimabara in 1637–8 as a humiliating price for their stay in Japan.13 In Dejima the interpreters had an office (and kept their records in it): the Dutch were never free from a Japanese presence on the island. They were policed, a situation which to be bearable required sensitively or tact on both sides. The Dutch and Japanese alike were not always tactful. The major resentment of the Dutch was the Japanese failure to provide the agreed amount of copper for export; but the

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

23

Japanese were powerless to meet the demand in quantity or in time, as the output of the copper mines was fitful. Irritation could affect even the judgment of Titsingh, far and away the most sympathetic resident in the history of Dejima. The issue of Dutchmen learning Japanese and acting as interpreters to ensure common agreement on translations was raised in 1675. The governor general of the VOC in Batavia proposed that several Dutchmen should be allowed to learn Japanese, “as was the custom in Japan in former times” and to have the status of interpreters. As Tanaka-van Daalen has noted, though the governor was supportive, a satisfactory outcome does not appear to have occurred.14 The issue was raised again among other things in a letter from the Governor-General in 1683. In response the Dutch were permitted to add an additional member to their group on the hofreis (or court visit annually to the shogunal court in Edo) in 1684, and chose Moses, a slave in the factory, who had a good knowledge of Japanese. He ran into difficulties (the interpreters professed not to understand). In an aftermath to the formal part of the audience, when the shogun asked whether there were persons among the Dutch who could speak Japanese, the answer was that the opperhoofd understood a little, and Moses spoke passages in Japanese in a loud voice.15 Back in Nagasaki in the wake of the hofreis the opperhoofd noted “speaking Japanese to people who do not want to listen is of no use” but also observed the “advantage of being able to speak Japanese.”16 The issue was carefully noted by Japanese officialdom. Commissioners from Edo who happened to be in Nagasaki to conduct an enquiry into the administration, were to ask the opperhoofd “why the Dutch want to learn Japanese.”17 In 1687 and 1688 the opperhoofd recorded doubts about any prospects of success on the interpreter issue. In 1688, in response to a letter from the Company in Holland broaching the idea of sending Dutchmen specifically to learn the language, the opperhoofd indicated that agreement on interpreters was not likely and that, if the proposal from Holland was acted on, its only outcome would be employing the knowledge of Japanese for covert use in Nagasaki or Edo.18 The Dutch believed that the possession of a fluent knowledge of Japanese was responsible for the exclusions. The belief was an assumption by the Dutch, not a response made by the Japanese. It also encounters the problem that some fluent speakers were expelled, but not other fluent speakers. Exclusion does not of itself imply that language had been the reason for the exclusion. Ease with the language pushed sensitivity or tact to the limit. The Domburg expulsion, occurring during

24

EARLY JAPANESE TRADE

the tenure of Titsingh with his remarkable personal relations with the higher officials in Nagasaki is the illustration of complexity. Titsingh attributed the expulsion not to a prohibition but to the enmity of the interpreters. In his words, Domburg who had served for eleven years “incurred their enmity and revenge, for they fear that someone who has not only exposed them, but also has such an adequate knowledge of their language that he is able to make himself understood, will rise to a higher office and will eventually outsmart them.”19 It was not language, which led to Domburg’s downfall. In contrast to past cases and perhaps a reflection of novel openness in the 1780s, the governor cited reasons, claiming that he had assaulted several coolies. Allegedly the coolies while unloading sugar were also stealing it. According to Titsingh at the end of October 1784, one coolie only was assaulted, and feigned injury: “From time to time I enquired of them and of Domburg about the outcome, but every time they said it had been settled until to-day the banishment struck like a bolt of lightening ...plotted for a long time by a few requests brought about by the stealing of the coolies”. More serious than the brawl with one or more coolies as a source of enmity, in Titsingh’s view, was that Domburg had stood up to “a senior interpreter Eizaemon, especially by turning him away from the opperhoofd’s house during the time of the court journey when Eizaemon insisted that he point out the camphor-wood chest.” The brawl with a coolie or coolies did not end Domburg’s list of offences. He was accused by the governor of other incidents, physically assaulting workers on the sampans ferrying goods between vessels and the shore. The Governor’s order referred also to how on an outing to Nagasaki he failed to return by the appointed hour and spent the night in the prostitutes’ quarter.20 Whatever Titsingh’s high opinion of Domburg, he may have been arrogant and highhanded, ironically displaying qualities that Titsingh deplored among the Dutchmen on the island. The Dutch factory included ten or more members. Doeff, writing in 1833, enumerated them precisely at ten; opperhoofd and doctor, warehouse man, scribe, four clerks and two sailors. Fisscher gave much the same figures. The number was smaller than in the past, when there were around twenty in Hirado and remained high in the earlier Dejima decades. These figures do not include slaves, brought to Dejima by individual Dutch members of the factory and perhaps destined to live out their days there. They spoke Japanese well. Their ability was a subject of comment at the shogunal audience in 1684, and one was excluded in 1739, allegedly because of fluency in Japanese.21

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

25

The Dutch at large have been painted in unflattering terms. But this comment, prompted by envy of the Dutch presence, may do less than justice to the Dutch as a whole. Some of them were educated, and the criticism of them by Titsingh was not of their competence but of their arrogance or crudity in dealings with the Japanese. In the celebrated case of Sidotti, the Italian Jesuit arriving in 1709 with no Dutch or Portuguese, his interrogation with help from the Dutch was possible only because one of the Dutch knew Latin. Less surprisingly incidental reference suggests that some had knowledge of French. The doctors were the most outstanding case. If they remained in the island for more than a year - they averaged several years, and some had long and in a few cases remarkably long stays - they acquired a working knowledge of the language. Local doctors sought contact with them, and this contact without the presence of interpreters was perforce in Japanese. Duration of stay in Japan was the decisive factor in mastering Japanese, allied to permission for women to visit the island and a freedom widened in time to visit the pleasure district, if accompanied. In time a presence of womenfolk and family units on the island was permitted. In other words there was ample opportunity to become fluent, though one must assume that it was in onna kotoba (women’s language)!. Thunberg, medical doctor in 1775–6, who with the help of some of the interpreters, learned some Japanese, found that though many members of the factory “seemed to be able to call for any thing they wanted in the Japanese language,” there was no systematic knowledge in the factory.22 PROHIBITION OF INTELLIGENCE OF THE CHINESE TRADE. 1682–1732: GROWTH IN DUTCH PRIVATE ENQUIRY IN JAPANESE

In the wake of the 1682 ban, the competence in Japanese of some of the Dutch made it possible to recruit informants at a low or less orthodox level to provide some of the intelligence they no longer received from the interpreters. The record itself would suggest that the services of Mottogi Tarroyemon (sic)ᮏᮌኴ㑻ྑ⾨㛛ࠊon whom for long the Dutch had counted, had declined before 1695, the year of his death, hardly surprisingly because of the prohibition on Chinese interpreters providing information to the interpreters of Dutch. Many of the Dutchmen however had a grasp of the language. In the 1690s, apart from Harmanus Mensingh mentioned later, there were several others. In November 1697 Wm. Brunt, was according to the opperhoofd

26

EARLY JAPANESE TRADE

summoned “to help me with his knowledge of Japanese against these villainous tolken [interpreters].”23 Two other Dutchmen in the 1690s, Joost Visser (in the 1690s a junior merchant in Japan first in 1679 and later for eight years into the 1690s) and Pieter de Vos, a senior merchant, had a reasonable command.24 The result was that the Dutch were able to acquire a good deal of information over a span of ten years centered on 1700. The effort to tap Japanese who visited Dejima became a determined one. In November 1696 leerling Tominanga Niffei (sic) was recruited by Lodewijs Wijs, second in command of the factory, to break his oath of secrecy.25 A year later, as he was being scrutinized by the “tolken”, he asked permission to keep a low profile for some time and to refrain from giving information about Chinese matters.26 He was later convicted of accepting a loan from the Dutch and of having stolen the opperhoofd’s watch. Sentenced to death in April 1700, the Dutch had to witness his crucifixion.27 A (wholesale) tailor Hansa (Fansa) provided information in November and December 1697. He seems to have provided information a decade earlier in May 1685: “the snijider Fansa brings some news of the Chinese junks and their trade”28 (and one may assume that there were other occasions.)29 On one occasion when he was about to give information on copper exported, the meeting was interrupted by the arrival of interpreters. In November 1699 details were received of seven outward cargoes from a private Japanese: “but he wants his name to be kept secret”.30 Information at times almost casually came from what are described as “our servants”. This was a vague term to cover Japanese other than the interpreters who serviced the needs of Dutchmen and interpreters alike either residing or working in the island. Jansen listed a total of sixtythree not including cooks, grooms and other lower servants.31 Information about copper in Chinese outward cargoes dried up temporarily in December 1703 as a result of an incident described (without detail) as having occurred on “Thieves’ day.” Thieves’ day was a traditional day when servants working on the island gathered there. On this occasion the opperhoofd recorded that as a result of the incident “the servants have not been anxious to visit our island.”32 The relevance of the quotation is simply that it hints at the existence of a sizeable regular flow of servants into and out of Dejima. Moreover, merchants had occasion to visit the island to inspect goods. Hansa, the tailor seemed a regular visitor (no doubt to inspect imported fabrics) and for that reason must have had access to officials of the Dutch and Chinese trades alike.

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

27

Through this flow, a good deal of information, even if much of it poor or garbled, was secured. Promises were sometimes made, but unsurprisingly were not always honoured. From 1699 the dagregister refers to entries in the junk booklets recording the cargoes of Chinese vessels though as detail is not repeated in the dagregister and the junk booklets have been lost, we do not have precise details of quantities. Language competence was not confined to the junior members of the factory who often spent years on the island. Opperhoofden equally had some or much conversancy. While opperhoofden were confined to a single year for a visit, they were free to serve at later dates. For fortyone years from 1641 to 1682, fifteen opperhoofden spent a sole year in Dejima, but eleven had spent two periods and one had three periods. In the fourteen years 1696–1710 four men, each spending two to three years accounted for eleven of fourteen tours of service. One of them, Harmanus Mensingh, had formerly been bookkeeper, on one occasion left in charge of the factory in the absence of the opperhoofd on the hofreis or annual visit to Edo. His knowledge of Japanese and his writing with a brush was noted in the 1690s.33 He was later to be opperhoofd in 1705–6, 1707–8 and 1709–10. Kaempfer records that in hia first hofreis in 1691 the opperhoofd was Henrich Von Butenheim “a gentleman well acquainted with Japanese customs and the Japanse langauge”, and in 1692 his successor was Cornelius Outhorn “an experienced and well-read gentleman familiar with the language”. Von Butenheim was serving for his third year (with a fourth to follow). Outhorn had already served in 1688–9 and was later to serve for a third time.34 Remarkably, the authorities seem to have been slow to wake up to leakages in information, responding only from 1706. In March 1707 the opperhoofd noted failure to get details of the cargo of a junk: “A Japanese told me that he could not get any information about the cargo of the junk: all who worked on the Chinese island were afraid, because some had been deported last year after having given information to others. I will ask the one who had informed me two years ago about it.”35 In December 1708 the opperhoofd did not know even the mere destinations of cargoes: “no one wants to tell us anything.”36 In November 1711 Hansa, a willing informant from earlier years, after promising a list of prices for Chinese goods, provided only general comments, as “the writers of the Chinese islands had not given him a list of prices.”37 In failing in June 1714 to get details of copper on Chinese junks, the opperhoofd observed, “They obtained more copper than last

28

EARLY JAPANESE TRADE

year according to our servants who cannot always be trusted.”38 Trust may imply less deceit than an inability to gather more than title-tattle. Dutch competence in the language did not wane. In May 1715 such was the competence of Six, a Dejima resident, that “his knowledge of the Japanese language permits him to communicate with the Japanese without the mediation of the interpreters.”39 It was not greater or lesser Dutch competence in Japanese, but more effective Japanese control of the flow of information from the Chinese island that stemmed the flow. The weakness of controls was that there was always some leakage from interpreters, either in exchanges in informal contact, or by some degree of bribery or easy credit. Effective control lay less in policing of the interpreters of Dutch with their daily contact with the Dutch than in tighter control of an outflow of information from the Chinese island. This fact is remarkably evident in an extraordinary change of fortune from 1718 to 1727, made possible by the fact that a Dutch interpreter was brother to the chief interpreter of the Chinese island. It also suggests that for the Dutch the problem was less unhelpfulness of interpreters of Dutch than the extent of success of an official clamp down in the Chinese island. In March 1718 Ichijirōzaemon, interpreter of the Chinese island, provided “a bill of loading” (sic) for 9 outward junks,40 and the information was provided by his brother on later occasions. Though information was sometimes conveyed in bills of lading, it seems to have been most frequently verbally given in round figures without a breakdown by vessel. The flow ended in July 1727 when the opperhoofd was told by an interpreter that “on pain of corporal punishment, the governor had forbidden the Japanese to supply us with information concerning Chinese imports and exports.”41 In 1730 an interpreter, asked for a price list, replied “the governors had forbidden the servants of the Chinese island to disclose the prices”.42 Faced with this situation the opperhoofd seems to have formed hopes of an outcome by way of covert enquiry in Edo. In September 1729 a member of the Dutch factory, Keijser (recte Johan Georg Keysering),43 who had first accompanied to Japan horses for the shogun and exceptionally spent periods in Edo as a result of shogun Yoshimune’s passion for horses and riding, was charged with the task. Together with two Japanese interpreters he was to be sent to Edo in relation to trade matters. The reason for his recruitment was that “While in Edo, he will meet influential people and – since his Japanese is reasonable – when he stays there he will become fluent in it: the more reason to give him a memorandum.” The opperhoofd saw these contacts, which were to be

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

29

kept secret from the interpreters, as freeing the Dutch from reliance on mediation of the interpreters in Nagasaki.44 The prospects of succes in Edo were slim, and the project simply reflected desperation. Despite an unpromising start, the 1730s were to prove a decade of change. The early years were poor in information as the crack down remained in force, but changes, first for incoming cargoes from 1735, and more slowly for copper, then occurred. There was no formal declaration, and it was evident simply in a resumption of the ready flow that had ceased in 1682. The change had official backing. It occurred in the late 1730s when either interpreters of rapporteur (nenban tsūji) rank or having some degree of seniority provided information. Suenaga Tokuzaemon, a junior rapporteur for the year to March 1739 was a frequent informant45. He was supplemented on infrequent occasions by Yoshio Chūjirō, oppertolk (senior Interpreter). As early as March 1738 apparently under pressure he had provided details of the cargoes of eleven outward junks.46 In 1740 under Tōsaburō, what was to prove the pattern of the future emerged .As holder for the year of the rotating office of rapporteur he reported in all from March to October on eighteen occasions. USE OF JAPANESE AT SHOGUNAL AUDIENCE AND IN THE EDO INN

Restriction on use on formal occasions of Japanese did not entail any prohibition on using Japanese outside the field of formal contact. Indeed as the Dutch were allowed to consort with prostitutes, and stable unions some with offspring existed, it makes nonsense of the assumption that there was a restriction on acquiring Japanese and that the Dutch concealed their competence. Moreover some use of Japanese as well as Dutch in private conversations at times extended into more formal learning of the language. This depended on the existence of a close rapport with individual interpreters. Knowledge of Japanese can be identified in the case of Kaempfer, and Titsingh, and for later times for Blomhoff and Meijlan we have more than supposition for their competence. Any doubt about the absence of prohibition of speaking Japanese, is removed by the fact that Japanese was spoken at shogunal audience. The topic came up in 1684. The deputation was asked by the shogun about the Japanese language in 1692, and in April 1693 its members were asked to count up to twenty in Japanese.47 The formal essence of audiences with the shogun was a ceremony, at which the shogun received the obeissance of the Dutch. It could by the

30

EARLY JAPANESE TRADE

time of Doeff take no more than one minute. In earlier times, questions were sometimes asked of the Dutch at this stage. But more commmonly the visit was protracted by the shogun retiring behind a screen to join wives and family for what was in effet an informal audience and a source of entertainment, lasting on occasion for as long as four hours. Kaempfer provides a very interesting account of the occasion for two years. In 1691 the shogun from behind the blind “as close as he could ...had us take off our kappa or ceremonial robes, and sit upright so that he could inspect us; had us stand up and walk, now pay compliments to each other, then again dance, jump, pretend to be drunk, speak Japanese....”48 A year later the audience was not in essence different, but two shogunal physicians were summoned the senior of whom “asked me at what time an abscess became dangerous, and when and how and in the treatment of which illness people were bled. He also indicated that he knew about our ointments and mentioned their names with awkward pronounciation, and I helped him with my awkward Japanese. And since the names I was saying were partly Latin and partly awkward Japanese, the shogun wanted to know in which language we were conversing, and what language the Dutchman was speaking. The answer was Japanese but badly.”49 The court journey in April 1701 was somewhat similar. After the initial formal audience, the shogun concealed himself behind blinds. “We are asked to take off our coats, to walk around, to sing and dance and to speak Japanese which we were not able to do, with the exception of (Dr) Wagemans, who could say “chia jaracenti” (bring tea).50 Whether this is a confession of deficiency on the part of the doctor or mere brevity of the account is not clear. But he was already into a second tour of duty, and was probably capable of better. He was to serve for another eleven occasions between 1706 and 1717.51 In 1714 the opperhoofd made a point of noting Dr Wagemans’ “spirited” conversation with Kurisaki “the famous” Japanese doctor at their inn in Edo in the course of the hofreis.52 In April 1725 (with the second part of the proceedings quite exceptionally taking place a day later), with the shogun observing through vents in the lacquered screen, “the court physician talked with Ketelaar (the Dutch physician) in Japanese about the patients.” 53 Ketelaar was serving his fourth and final year. A degree of informality was part of the contact with the Dutch in Edo in a way for which there was no parallel in Nagasaki. During the hofreis the Dutch were usually entertained at the home of the Nagasaki governors whose turn it was to spend a year in Edo. Shogunal physicians

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

31

were already regular visitors to the doctor in the mid-1690s.54 Visits by some daimyo to the Dutch at their inn later emerged, and in time the daimyo of Mito and Satsuma became regular visitors. Some sense of the informality which existed is conveyed in the account by Boockestijn in the dagregister of the 1732 hofreis. In the group of Japanese and Dutch waiting in the antichamber for the audience, Suō the Nagasaki governor accompanying the Dutchmen was asked by some of those present if the opperhoofd could speak Japanese. He put the question to the opperhoofd. He asked me: Capitan Conjits Edo Sammaoeijka(Is it cold in Edo?). I replied: Wataxie gatting hito bekerri nihon ima Edo hito bekerri sampko.’ (I understand some Japanese, and it is presently rather cold in Edo.).

The Dutch of the published text is a translation into English, though the Dutch phonetics are untouched.55 In kanji the exchange reads ࠕ࢝ࣆࢱࣥࠖࠊ௒᪥ࠉỤᡞᐮ࠸࠿ ࠕ⚾⮬㌟ࠖ㸹⚾ࡣྜⅬ୍ࡘࠉࡤ࠿ࡾ᪥ᮏ㸦ㄒ㸧ࠊ௒Ụᡞ୍ ࡘࠉࡤ࠿ࡾᐮ࠿

The opperhoofd’s phonetics (in Dutch) are poor, reflecting knowledge of conversational rather than written Japanese. The language, in Kyushu dialect (hitobakari୍ࠉࡤ࠿ࡾin place of hitotsubakari (࣮ࡘࠉࡤ࠿ ࡾ) is clear enough and may imply that he may have had in fact a reasonable command of colloquial Japanese. The occasion is instructive on several counts. Not least is the informality of the occasion, and the absence of the governor standing on his rank and his readily addressing himself directly in Japanese to the opperhoofd. OPPERHOOFD KNOWLEDGE OF JAPANESE

The doctor and the opperhoofd by definition held a special position in the factory (and hence at the audience). Of the doctors we can say little because accounts by them, Kaempfer, Thunberg and Siebold apart, are non-existent. Doctors remained for variable and sometimes long intervals. In addition to medical or scientific backgrounds, a few had an academic one. Doctors learned Japanese quickly because of the eagerness of Japanese doctors to make contact with them. Doeff in

32

EARLY JAPANESE TRADE

his recollections wrote that Kaempfer spoke no Japanese,56but in fact he did, however imperfectly. Kaempfer and Thunberg are well known because of the accounts they wrote, but most of the doctors are poorly documented. Siebold provided a large volume of information on Japan, but fewer observations on the more ordinary aspects of Dutch experience in Japan. Of opperhoofden we know much more. While in the first century few opperhoofden spent more than two tours of duty in Japan, the fact is often overlooked that many had spent several years on Dejima at a lower level before assuming the role of opperhoofd. A few opperhoofden could actually read Japanese. But read it or not, some had an intelligent awareness of the language. De Laver (opperhoofd 1733–4) who had earlier service at a junior level in Dejima observed in August 1734 that “Since the time I arrived here in1725 I have learned that a lot of Japanese, can read Chinese but do not understand its meaning. The Japanese language is very concise. One line of characters equals one written page in Dutch.”57 Opperhoofd Boockestijn quoted above in conversation with governor Suō was serving the final one of three tours spread over five years. It would be surprising if he did not know Japanese (and the phrases quoted are correct though the evidence is too slender to hazard an estimate on the extent of his conversancy. The opperhoofden were in the main intelligent men. The frequent negative comments by them on their meetings with doctors and astronomers reflect the inordinate amount of time the meetings took, the visitors asking technical questions in long and exhausting sessions (totally oblivious of western social norms of invitation and length of stay) in which until the late eighteenth century the conversations were primarily in Japanese. By the late eighteenth century when knowledge of Dutch was growing rapidly in a post-1770 Edo circle, both the number of visitors and facility in Dutch grew. The enthusiasm of this new breed of rangakusha (followers of Dutch studies) knew no bounds. Some of them affected western dress, and twenty-nine of them celebrated the Dutch new year on 1 January 1795.58 The scale of contact with the Dutch is suggested in the account by Fisscher who served on the 1822 hofreis. On the twelfth day of their stay sixteen doctors visited the Dutch in their inn and remained for five hours. On the following two days, the visitors were astronomers and some doctors.59 For opperhoofden a pattern of several stays, punctuated by intervening years in Batavia, became more common in the 1690s and first decade of the eighteenth century. Though the pace of change declined for two

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

33

decades, the pattern of single-stay opperhoofden became uncommon from the 1740s and was to change dramatically from 1770 onwards. For the years 1770–83, one year excepted, three men monopolized the role. Armenault (three years), Feith (six years) and Titsingh (three years) accounted for eleven of the twelve years. Wartime conditions had changed things. Faith served for two successive years 1777–9 and Titsingh for 1781–3. In a novel variant Titsingh’s stay in 1784 was a short one for the buying season (August- November ) only, and. Romberg, already in the last days of his 1783–4 stay, was left in command for 1784–5. Two more tours of service were later to follow for Romberg. Previous service in Dejima at a lower level was a major factor in securing knowledge of Japanese, Romberg spent twenty years (five as opperhoofd) in Nagasaki in all, Feit sixteen years (six as opperhoofd), Van Homoed, sixteen years (three as opperhoofd ), Duurkoop (thirteen years (one as opperhoofd).60 The non-appearance of one (or both) of the two annual vessels could affect succession making necessary an unexpected prolongation of a stay. The limitation of the trade to a single vessel annually from 1790 added to the problem. Wars in the early1780s and more seriously during the years of the French Revolution and Empire when Holland had become a French satellite were more serious causes. In wartime Titsingh who had already spent the trading year 1779–80 on Dejima, found his tour of duty in 1781–2 extended into a second year as no vessel appeared in 1782. The pattern repeated itself in the following years of peace. Van Rheede tot de Parkeler, already opperhoofd in 1785–6, found his next visit extended into two years (1787–9), and Romberg’s successor Chassé from1790 spent two years in a row. Prolonged wartime problems were to make longer stays become the norm. Hemmij who arrived in 1792 was given permission, as a Dutch ship did not arrive in 1794 to remain till 1798.61 Willem Wardenaar spent 1800–03 in Dejima preceding the long and eventful stay of Doeff in 1803–17. Doeff’s was the longest uninterrupted stay in the history of the office (though exceeded for total years of service at all levels in Dejima by Romberg), his long stay accounted for by prolonged war, absence of ships and challenges from the English to Dutch survival in Dejima itself. Doeff in 1803 was already in Dejima at a junior level for his fourth year. The two exceptions to a rise from the ranks after years of junior service were Titsingh and Van Rheede tot de Parkeler, of high professional and aristocratic background respectively. Neither man served in Dejima before assuming office.62 When Titsingh left Dejima he was to remain in correspondence with Rheede

34

EARLY JAPANESE TRADE

who wrote a now much-quoted letter in 1787 with comments on the interpreters.63 Titsingh, and Van Rheede are unique in the survival of personal correspondence throwing light on them. Longer stays meant that the head of the factory, as well as speaking Japanese had a greater involvement in Japanese official life. In these years, with long stays by opperhoofden, their ease in the language had greatly increased. Even Romberg, grumpy or difficult though his reputation was, conformed to the demands of the hofreis, entertaining daimyo and experts alike and trying to respond to the demands of the latter. Doeff was followed by Blomhoff (who had already served in Dejima 1806–13) as opperhoofd for 1817–23. From 1826 four years was to become the norm into the 1840s. Either personal qualities or the opportunism created by longer stays ensured that opperhoofden have a clearer profile than their predecessors. This was the case for Titsingh and Van Rheede, for whom the Titsingh correspondence illustrate the Dejima stays of both men and the ties between them (enriched also by the letters of his father to Van Rheede). In later years Blomhoff was on the evidence of a letter to his wife an intelligent man; he had close ties with some of the doctors, and made a collection of Japanese materials now in the Dutch National Collections.64 Meijlan’s’ character is suggested on the evidence of his book in 1830, practical and modest and touching on the key issue of language in a way no other account has done.65 Fisscher, next to the opperhoofd in rank, wrote the most detailed account by a member of the factory. It may hint at the qualities counted on for the office of opperhoofd (he had been seen as a possible future one). Meijlan in his book reveals impatience with the protocol of audiences and is critical of the prevention of direct engagement in conversation with the governor on formal occasions Er zijn overhoofden genoeg geweest, die op the Japansche taal zich togelegd, en zoo veel kennis van dezeslve verkregen hadden, om zich daarin zeer verstaanbar uit te drukken. Er zijng er ook geweest die, met voorbjigang der tolken, hoogen abtenaar, in het Japansche onmiddeelijk heben aangessprokke doch te vergeess; de man hield zich als verstond hij niets, en wees hen tot den tolk over, voor hetgeen zij te leggen had..66 (There have been opperhoofden enough who focussed on the Dutch language and had acquired such a knowledge of it that they could express themselves in a very understandable way. There have also been some who, dispensing with an interpreter, immediately addressed a high office-holder, but in vain. The man held himself as if he understood nothing, and pointed to the interpreter, to whom he had to explain).

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

35

Meilan had a good knowledge of Japanese. He observed that there was no language with as wide a difference between the written and spoken language as Japanese. His observation on addressing a high officeholder leaves one with the impression that he was writing from personal initiative at formal audiences with the governor. FORMAL AND INFORMAL OCCASIONS

Meijlan, impatient because he could not use his Japanese, made a point of the distinction between formal occasions and less formal ones. It is central to understanding the status of Dutch use of Japanese. Life was more formal in Nagasaki than Edo, something explained by the language routine established in the 1640s, also by life revolving around a single figure (the governor i.e. bugyo-), head of what was far and away the largest single administration in Japan, and the interpreters’ vested interest in maintaining Dutch as the language of business in what was in trading terms now a small and contracting world. On formal occasions, Japanese etiquette excluded direct communication of Japanese and foreigners alike with high officers of state (according to Meijlan from the rank of oppeeerbanjoost upwards).67 This was modelled on the etiquette revolving around the shogun himself which Kaemfper commented on.68 Meijlan (opperhoofd in 1826–30) observed that at formal audiences the governor addressed a secretary who in turn spoke to an interpreter who finally translated for the benefit of the opperhoofd. A reply followed the same sequence in reverse. A parallel existed, as Blomhoff observed, in opperhoofd contact with daimyo in two-way passage of conversation addressed through intermediaries with replies in turn relayed from junior to senior levels.69 When daimyo came in disguise, the intention was not to conceal their identity. Disguise was the symbol that they did not stand on rank, and were dispensing with the etiquette to which their rank entitled them. Kuchiki Masatsuna (1750–1802) daimyo of Fukuchiyama called in disguise on the Dutch at their inn on their way to Edo in 1776.70 When Mito-sama (i.e the daimyo of Mito) arrived incognito at the Dutch inn in Edo in 1818, accompanied by a retinue of male and female attendants and a metsuke, there could be no doubt as to who he was. However, the disguise signified that rank was being dispensed with. Conversation was man to man: without the presence of interpreters and a fifteen-minute conversation was conducted in Dutch.71 Daimyo were entertained to dinner and sometimes with shogunal doctors on

36

EARLY JAPANESE TRADE

the same occasion. The remarkable friendship of Titsingh and Shimazu Shigehide (1745–1833), daimyo of Satsuma, is an instance of close ties. The daimyo had a knowledge of Dutch, was a visitor when in Kyushu to Dejima and in Edo to the Dutch at their inn. In 1822 the Dutch descended from their norimono as they passsed the Satsuma mansion and greetings were exchanged. On the occasion of the following hofreis in 1826, the elderly Shimazu did the Dutch the high honour of ぢ㏦ miokuri (accompaniment to the point of departure) as they left Edo, also warmly recalling his friendship with Titsingh. The little world of Nagasaki began to change in the 1780s and 1790s. It was also the beginning of a period in which sustained stays by opperhoofden began to become the norm. Moreover as foreign vessels on the coasts began to become more numerous, interpretation acquired an importance beyond mere trade demands in Nagasaki. Fisscher was damning in his comments in 1833. The interpreters, he wrote, spoke Dutch little outside trade dealings; wrote it even less often, and required no great knowledge to perform their work.72 Kaempfer’s comment had been even more damning in 1691–2, ending with the words that “following the idiosyncrasies of their own language[,] one would require a second set of interpreters to understand the first.”73 As a general comment it is too sweeping, as new demands revealed the ability of some interpreters to meet them. A broad language change dates from the first decade of the new century with the Russian visit to Nagasaki in 1804, the letter in French in 1807 and the visit of the English warship Phaeton to Nagasaki in 1808, pointing to new language requirements. The movement of interpreters to Edo was a feature (the brightest of the junior interpreters Sajūrō was summoned to Edo in 1808), and in reverse from this period on to the 1850s, visits to Nagasaki by self-improving individuals, rarer in earlier days, became common. The Dutch language was no longer sufficient for dealing with the outside world, and the interpreters had to respond to the need. A novel instance had occurred a century earlier on the occasion of the Sidotti case, when two junior interpreters were sent to Dejima to learn Latin.74 That need was short lasting, and needs did not arise again until 1804. But between 1804 and 1808 knowledge of French, English and Russian seemed important. The letter in French by marauding Russians in Ezo prompted an urge to read French. Consultation with the Dutch answered some of the short-term issues, but a corpus of independent Japanese knowledge was necessary. Hence the foundation in 1811 of the ⻅᭩࿴ゎᚚ⏝Banshowagegoyō (centre for translating foreign books),

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

37

headed by the remarkable Takahashi Kageyasu; the movement of interpreters from Nagasaki to Edo (Fisscher from the 1822 visit named six75); and the launch in the same year 1811 of compilation of the Haruma dictionary in Nagasaki TITSINGH AND DOEFF

Two men, Titsingh and Doeff are outstanding in the exercise of the office of opperhoofd, both unique in different ways and having a major impact on the history of the factory and its place in Japanese society. Titsingh, the more exceptional in his Japanese contacts, had a scholarly relationship with seven interpreters76 but also close contact with the two most senior officials in Nagasaki, the governor Kuze Hirotamiஂୡ ᗈẸ㸦ࡃࡐࡦࢁࡓࡳreferred to as “de brave gouverneur”), and the head of the Kaisho (trade office) Gotō Sōzaemon (1725–80). Titsingh’s contacts with Kuze reached to meetings, correspondence and proposals for improving the standard of ship construction in Japan. For Gotō he wrote a treatise on the spirit, obviously implying philosophic conversations the two men had.77 It is a daunting task to translate. There is no evidence that it was translated: it lies in the papers of the Motoki interpreter family (in any event the death of Gōto in 1780 made translation superfluous). He also had ties with Kuchiki Masatsuna and Shimazu Shigehide, daimyo of Fukuchiyama and Satsuma respectively. With these men Titsingh spoke and wrote Dutch, and a correspondence lingered on after he left Japan in 1784. He spoke Japanese; he had an attachment to a yūyo 㐟ዪ (a lady from the pleasure district) with whom after his final departure from Japan he remained in contact, sending presents. In conventional terms this attachment, similar to that of other Dutchmen, helped to give him speaking knowledge. But from early in his Dejima career he could read Japanese, a rare achievement, though for a polyglot like Titsingh less surprising than for others. The evidence for his competence tends to be oblique rather than direct. His method is suggested however in one passage in English when he described to William Marsden from Amsterdam in 1809 in a letter that he sought to represent Japan in its sources rather than by imposing an interpretations on it: “to obtain this end, I applied during my stay in Japan to some friends, reputed as men of learning, and free from all national prejudices – they procured me such works on various topics, as enjoyed with them the highest regard.78 Titsingh, according to Meijlan compiled a small dictionary

38

EARLY JAPANESE TRADE

(woordenboekje).79 No doubt intended for his own guidance, it was still there to turn to, as noted by Meijlan, and it must have been the basis of knowlege passed on to others. An English captain Wilson made a compilation of words entitled “Japanese words obtained by Captain Wilson from the Dutch Resident at Chinsura, Mr. Titsingh.”80 The dictionary can hardly be the 147 pages of dialogues in the hand of Titsingh and of an interpreter respectively with Dutch transliteration of Japanese with introduction by Titsingh (it survives in the Bibliothèque Nationale, Paris).81 The dictionary does not appear to have survived and remaining, one assumes, in use in the factory in Dejima must have been lost in time. There are two things, which make Titsingh unrepresentative of foreigners in Japan in the entire Tokugawa period. First he was a polyglot. That is reflected in his interest in Chinese. With the help of his friends in 1790, absorbed in study of a copy of a Chinese - Latin dictionary of missionary origin, he feared it was the cause of a pause in his Japanese studies, leaving his collection of Japanese materials marking time.82 In 1789, writing to Masatsuna, he noted that he was learning Chinese to improve his grasp of Japanese, and hoped that in a few years he would write him a letter in Chinese.83 Titsingh’s approach is perfectly realistic but it is the attitude of a scholar, not that of a normal businessman. The second factor is that he continued his study of Japan and Japanese long beyond the years of his stay in Japan. The reason was that he had in mind the improvement of his grasp of the material he had collected and its translation into European languages. The rapidity of his learning Japanese was uncommon. One has to think forward to the years of Satow and Aston to find something comparable in the way of evidence of ease in mastering Japanese and, though more modestly, of scholarship. Unlike the ambitious Siebold, Titsingh’s scholarly commitment to Japan was disinterested, not seeking to enhance his own reputation and dedicated to eventually making the country better known. Contrasting western knowledge of China with poor awareness of Japan, he sought to redress the imbalance. His approach and unique personal rapport at the highest level contributed not only to improving the relationship between opperhoofd and high officials but to altering it. Remarkably, though governors, like the rōjū in Edo, never provided justification for actions, the reason for the expulsion of Domberg were set out in a detailed case at the end of October 1784. Though Kuze had already been transferred in March, his successor seems to have responded to

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

39

changed circumstances and in all probability as a courtesy to Titsingh. Angered by the Domburg expulsion, however, Titsingh in a rare loss of his composure, referred to the despotic rule of the new governor. But this may be less than fair. Titsingh’s stay in the autumn of 1784 was a short one and the governor was a new man. It may be observed too that, marked though his Japanese sympathies were, Titsingh was a formidable defender of Dutch business interests, and in the Domburg case, this clearly took precedence. Titsingh was in some ways an inflexible and uncompromising man. His experience in his short tour of service in 1784 led him to make uncommonly harsh comments. Despite the contretemps, circumstances had changed however. Even an uncommon run of short tenures of governors did not prevent the indignity of the opperhoofd being subject to a body search on arrival in Japan and on departure - long a subject of protest by the Dutch– though it was to end three years later in 1787. Longer and uninterrupted periods of service, though in part a consequence of war, may also have created a changed perception of the headship of the factory, and provided the basis for what at times, as in Doeff’s tenure as opperhoofd, was a closer working relationship with the shogunal authorities. The governor had always counted on the aid of the Dutch when linguistic problems arose. But the ambitious dictionary project of 1811 rested less on consultation than on co-operation, and the need for Dutch assistance or advice, now acquiring a strategic or political significance, was more frequent in the last decades of the shogunate. Doeff’s stay in Dejima is the fourth longest in the story of Dejima, 18 years from 1799 to 1817 (and uniquely for 1803–1817 the longest as opperhoofd ). As in many other cases he had female relationships and a Japanese family. Apart from his achievement in protecting the factory from encroachment by the English in 1808 and 1813, what was exceptional in his case was his close collaboration in the creation of a great dictionary. It was not the first Haruma (an Edo Haruma, modeled on a Dutch/French dictionary by Halma had already been completed in 1795). Work commenced in 1811 with the participation of 10 interpreters, and the first draft was completed in 1816 (revision continued up to 1833). Doeff claimed that he originated the idea. This is unlikely to be the case. The launch of the project in Nagasaki coincided with the creation in Edo of the Banshowagegoyō. Fisscher in an unpolemical comment in what was in some ways a book with a polemical purpose, had said that it began met voorkennis der Japansche regering (with the foreknowledge of the Japanese government).84

40

EARLY JAPANESE TRADE

Unlike the Haruma in Edo, the Dejima Haruma included entries of sentences illustrating the use of words. Doeff had a close rapport with interpreters, most famously perhaps with Baba Sajūrō 㤿ሙబ༑ 㑻̽ who had haunted his house every day. He was later from his Edo base to prove the most versatile of the interpreters, his career halted only by premature death in 1822. Fisscher, meeting him almost on the eve of his death, referred to him as de genoemdeondertolk (the renowned junior interpreter).85 Quite unlike Titsingh, Doeff does not appear to have developed close contacts with the higher echelon of Japanese society. On the hofreis his world appeared that of interpreters and scholars, and though daimyo visits must have continued during the hofreis in 1806, 1810 and 1814, he did not develop a close relationship with daimyo. His book in 1833, boastful though it is on many issues, is silent on daimyo.86 The dogmatic tone is evident in his comments in 1824 when he became aware of Titsingh’s correspondence with governor Kuze and with eenige hooger op in ‘t land woonende Jappaners ( some higher-rank people in Japan). He claimed op eigene ondervinding en beter weten gegrond (on the basis of my own experience and better knowledge), that the danger (‘t gevaar is te groot) made such correspondence impossible.87 Doeff’s repute back in Holland was later to be somewhat clouded by claims both by Siebold and in a book in 1833 by Fisscher for credit for dictionary work. In 1826 in a report to the governor general, Siebold made a claim that he had put together in Latin a treatise about the Japanese language.88 Fisscher in his book recognized Doeff’s early work and the loss at sea in 1819 of the copy of the dictionary taken by Doeff as it stood at the time when he left Japan in 1817. But he went on to claim that after Doeff’s departure the project marked time until he came across the material; and relaunched it in 1823, widening its scope to incised sentences to illustrate the use of individual words.He brought back a copy to Holland presenting it to the royal library. In different ways Doeff and Fisscher sought to enhance the account of their achievement. Both men referred to a prohibition on learning Japanese, though their account of their work makes nonsense of the existence of language prohibition. They also ignored the momentum of the project, which continued through the period and beyond it. It reached a complete draft in 1822 (which appears to contradict Fischer’s statement of an unfinished work in 1823) and work continued until 1833.89 Ease in Japanese was now commonplace. And, simply because it was taken for granted, occasions to refer to it became few. Only in excep-

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

41

tional circumstances such as those of of a Dutchman made prisoner by the crew of the English warship the Phaeon in 1808 and put ashore on the potentially dangerous mission of carrying the peremptory demands of the English captain, did it warrant mention. The Dutchman was saved by his lingusitic competence from a probably fatal challenge by the Japanese forces on the shore line. “Fortunately Gozeman had a good enough command of the Japanese language to make himself understood.”90 CONCLUSION

The story of contact with the west is in some respects that of a fluctuating pattern on the use of two languages, Japanese and Dutch. Beyond the Dejima base of the interpreters, Japanese knowledge of Dutch increased rapidly in Edo itself in the years from the 1770s in the hands of doctors and astronomers. Dutch knowledge of Japanese also increased in these years. Informal conditions in Edo, and a growing circle whom the Dutch met on the hofreis meant that both languages were readily used. Nagasaki changed less than Edo both because one man dominated the administration and was served by a large and underworked corps at a much lower social level. Change was ruled out by three things; the established pattern of business in Dutch settled in the 1640s, the exalted status of a governor heading one of the largest bureaucratic offices in Japan requiring address via intermediaries, and the interest of. the interpreters in faithfully preserving the status quo. There was some private use of Japanese with governors in the years of Titsingh’s stay in Nagasaki, but evidence for other years is elusive. A process of change was marked by novel movement of interpreters to Edo, and by Nagasaki becoming a potent center of attraction for serious visitors from elsewhere. Though the governor always relied on the Dutch for assistance or help, when occasion demanded, the dependence became more marked in the remaining decades of the Tokugawa regime. The Dutch too were emboldened in 1844 to advise the shogunate that it should open its borders, and a Dutch warship entered Nagasaki bay not as a threat but as a reminder of the formidable armaments of the outside world. The warning did enhance Japanese awareness of the external threat. Between 1850 and 1855 the Japanese received a series of ูẁ㢼ㄝ᭩betsudan fūsetsugaki (special reports) setting out details of the number of vessels and armaments of Western navies.91

42

EARLY JAPANESE TRADE

In the 1810s and 1820s serious study and use of foreign languages reached regularly beyond Dutch. Some interpreters had already been ordered to study English. Blomhoff too taught English to some of the interpreters.92 Baba Sajūrō, attached to the Banshowagegoyō, had drawn up a document in halting English to answer the needs of dealing with the crews of whaling vessels, advising them of uchiharai and enquiring what were their needs in food, water and firewood.93 What is at first sight puzzling is that the knowledge of English had receded by the time American whalers in the northern waters off Japan began to become numerous in the 1840s. There may have been some pause in whaling in the 1830s. The Siebold Incident in 1829 culminating in the death of Takahashi in prison and punishment of those interpreters who were suspect of acting as intermediaries between Edo and Siebold may also have reduced the Banshogowageyō to a mere translation bureau, ending active language learning and the movement of interpreters between Nagasaki and Edo.94 When the crews of whalers were brought to Nagasaki, the Japanese relied on the assistance of the opperhoofden in their interrogation.95 One of the interpreters, Moriyama Einosuke  ᳃ᒣᰤஅຓ) (1820 – 1872), did make progress in mastering English. The report on the negotiations with Perry in 1854 referred to him as a man “who spoke a little English.”96 This negative judgment may have been influenced by the strong Dutch accent of his English which made it difficult to understand him. But it is clear from the American record that he impressed the Americans he encountered, and as the only Dutch interpreter competent in English, his role was vital for Japanese acceptance of the English version of the treaty. Two years later Townsend Harris (1804–78), the American consul now installed at Shimoda in Shizuoka Prefecture on a first meeting noted him as “a superior interpreter”, whose comments he came to regularly record in his diary.97 The old order of Dutch as the language of external communication did not outlast the 1850s. The treaty in 1854 with agreed texts in Chinese, Japanese and Dutch (the Americans were the first foreign mission in Tokugawa times adequately equipped in interpretive skills) ended the one-sided negotiations of the past two centuries. Dutch itself remained unavoidably the language in negotiating the 1854 and 1858 treaties, simply because, Moriyama apart, none of the team of Dutch interpreters knew English. In future English replaced Dutch; it also quickly became the language of social intercourse. While in the Meiji period there were foreign residents who acquired ease in Japanese, notably

KNOWLEDGE AND USE OF JAPANESE BY THE DUTCH

43

officials like Satow and Aston, private individuals like Chamberlain, and many Protestant missionaries. International trade and diplomacy were henceforth to rest on use of English by both sides. The trade statistics of the 1870s, themselves for Japan a novel publication, had an English-language version as well as a Japanese one. Few of the new traders in the open ports, now with no need to do so, got to grips with Japanese in the way that the Dutch residents of Dejima had done over two and a half centuries.

Source: Review of James White’s book, Cornell University Press, The Japan Foundation Newsletter, vol. XXIV Nov. 1996, pp.21–2

3

Ikki: Social Conflict and Political Protest in Early Modern Japan James W. White ™

In any discussion about interpreting Japanese history anew, the question of ikki (riots) is often a stumbling block. In Japan, as in other countries, a large literature on unrest in various forms, often highly specialized, tends to lie outside the mainstream of broader historical investigation. This compounds a problem concerning coverage of Japanese history, whether by Japanese or non- Japanese, of giving only a self-contained history written in terms of an isolated debate, and further distorted when non-Japanese concepts are applied outside their original contexts. The most striking problem connected with ikki is that they are usually related to a negative appraisal of Tokugawa history (1603–1868). Immiseration caused unrest, and if there is any doubt of the misery, the records of the ikki clearly demonstrate severe social unrest. Ikki are represented as rising exponentially to a peak in the 1860s; however, this neat profile and dénouement smack somewhat of Ernest Labrousse’s increasingly discredited version of French history before the revolution. Professor White’s book, however, while working within the framework of ikki counts and typology, turns the debate on its head. Ikki, he argues, did not arise from misery but resulted from increasing affluence. Though growing in number, ikki, defined broadly as contention 44

IKKI: SOCIAL CONFLICT AND POLITICAL PROTEST

45

(ranging over fifty-two categories), represented not desperation but a rational calculation. They played a dynamic role in preventing the state, whether han or bakufu administrations, from increasing revenue. Halting fiscal demands and creating surpluses for rural occupiers, economic development itself was sustained. In turn a rational calculus by countrymen helps to explain both the uncommon height of unrest in the decade of greatest political upheaval, the 1860s, and the ability of the new Meiji government (from 1868) to end it. A stronger Japan, unified and effective in contrast to the old bakuhan taisei, could count on the rationality of its citizens as well as on its enhanced military and administrative singlemindness. This interpretation eliminates the problem of accounting for the contrast between the apparently better order of the Meiji era and that of the Tokugawa period. Professor White’s belief is consistent with new thinking on bakufu economic history, which, in contrast to describing economic decay and social strife, argues that the economy actually grew. Professor White also suggests a shift in unrest from political to social, which he accounts for by increased social differentiation. Caused by economic development, this in turn led to social friction between richer and poorer commoners. Thus, in his analysis of the ikki counts, Professor White contends that unrest was primarily in regions close to metropolitan economic centers or at least within their orbit. Elsewhere it occurred in regions where han government was weak. The absence of ikki is not due to benign government but to the ability of effectively adminisiered domains to control discontent. However, there are problems with this approach, which, though the conclusions are highly original, continues to rest on the existing ikki counts and on the historiography. A crude count of events, many small and on a single day. and drawn from a heterogeneous and uneven range of sources, poses serious problems of comparability. Though Professor White concedes this in principle, it does not prevent him from making an elaborate analysis including multiple regressions. Static and unequal evidence of this sort can be best used, as in one chapter, simply in analyzing the identity of victims and perpetrators. If extended beyond this capability, difficulties inevitably multiply. The limitations of his approach are particularly clear in Professor White’s unreal contrast between political and social events. For example, opposition to tax is treated as political, whereas riots or the like are seen as social. In reality, of course, both occurrencies are closely related to the same economic difficulties and are thus not really separate phenomena.

46

EARLY JAPANESE TRADE

This problem becomes most obvious when Professor White describes the period covering the Tenmei, Tenpō, and Keiō crises. This period was heralded by the unrest experienced on a national scale for the first time in the Tenmei years. For this reason alone the eighteenth century itself is described in the book as the most significant period in the history of ikki. This, however, potentially contradicts the previously stated picture of improving economic conditions. As a result, the argument is advanced that while major crises became more widespread and aggressive, the magnitude of the average event was diminishing (p. 138) and that undue weight should not be attached to the economic crises (p. 244). Whether this approach is realistic is debatable, but certainly it underestimates the gravity of economic and social problems in the 1780s and 1830s for the state and citizens alike. Professor White demonstrates a little uncertainty at times. If, for example, unrest was less widespread in Japan than abroad, and if many areas had no recorded ikki at all, and if the magnitude of individual events declined over time, it is hard to see how ikki could have been so decisive in quelling the avarice of the slate or in increasing an awareness against encroachments of the state. The book, though, does have genuine merit in that events are put in a fiscal framework. Even so, this potentially fruitful approach is not developed as far as it might have been. This is ultimately because the author’s view of the state is not only negative but sweepingly so. Such a view is starkly obvious when Professor White contends that “by the middle of the century the state had for all purposes capitulated to popular resistance” (p. 78), or that “in the mid-nineteenth century . . . respect for the state was at an all time low” (p. 302). His argument regarding the state retreating from effectively taxing the rural inhabitants is reinforced by a stress on the reduction of samurai stipends and the generation of non-agriculture- related taxes, though at one point the author concedes that few han were successful in this regard. However, this argument is questionable. The cut back in samurai stipends predates Genroku times, and the taxing of other sources applies more cogently to the Mizuno Tadakuni years than to those before or after. Prior to and following the Tenpō reform, state behavior was consistent with a basic moderation (for which antecedents exist even from the early Tokugawa period). At the end of the eighteenth century there was a fear of the simultaneous occurrence of three problems— unrest, famine, and foreign invasion. This worry became a recurrent preoccupation of policymakers. From this period the bakufu

IKKI: SOCIAL CONFLICT AND POLITICAL PROTEST

47

set a difficult course of not making, or at least of moderating, demands on both han and individuals and at the same time covering costs and providing for defense. The 1860s, with an increased foreign military presence, created an acute problem, however, with regard to how revenue and expenditure could be managed. The circumstances which, under an unchallenged policy of isolation for so long, had made a working and consensual bakuhan taisei realistic and, arguably successful, no longer held. In moving away from the negative view of the Japanese economy and by recognizing explicitly the fiscal implications of Japanese administrative history, Professor White has given us an original and important volume. However, ultimately, it operates only within the limits of existing ikki counts, and its sophisticated analysis only serves to underline further this fundamental weakness.

Source: Journal of the Irish Society of Japan, No.33, March, 2014, pp.74–83

4

Gulliver in Japan: Swift’s Gulliver’s Travels ™

I

Swift’s largest and most original corpus of writing, that effected in the 1720s, included besides the Wood’s Halfpence pamphlets and the Modest Proposal of 1729, the celebrated Gulliver’s Travels in 1726, a fictional work with underlying serious social and moral themes. It contains also a section on Japan. The section on Japan was the first published account to appear in English in the 1720s on Japan. It is one of only two works between 1706 and 1726 which had a Japanese theme. By coincidence or otherwise, it was published in 1726, a year ahead of Kaempfer’s History of Japan, which was to remain in several languages the main source of information on Japan until 1853.1 Swift’s inclusion on the eve of the appearance of Kaempfer’s great book of a brief chapter on Japan in Gulliver’s Travels raises the question whether it is a case of coincidence or of conscious anticipation, through some awareness, of the ambitious project of publishing Kaempfer’s manuscripts, which were firmly in the hands of Sir Hans Sloane from 1724 and which after a very belated and hurried process of translation from German into English, finally appeared in 1727. There was an interest in the translation, as the list of highly-placed subscribers collected in 1726–7 suggests. The fact that Japan should appear in a volume of fabled travel in 1726, just ahead of Kaempfer (1651–1716) has long attracted attention. The study by William A. Eddy in 1923 could find no model for Japan in the context of travel fables: he considered the possibility that a Japanese account published at a later date in 1774 might have 48

GULLIVER IN JAPAN: SWIFT’S GULLIVER’S TRAVELS

49

served as an inspiration. He speculated that awareness of fabled travels might have been brought back by the Dutch from Deshima and that Swift might have learned of the tales when acting as secretary to Sir William Temple, ambassador to Holland, in the 1680s. But this is a very fanciful chain of reasoning, Eddy himself observing that “all this, however, is but pernicious conjecture.”2 It was not unique as an imagined account of Japan. In 1706 a work entitled Man unmask’ed: being a wonderful discovery lately made in the island of Japan written in the Japanese language by the spirit of contradiction, and translated into English for the benefit of the publick, by Sir Tristina Nerebegood, both of them eye-witnesses to the whole affair had been published in London.3 It was devoid of any basis of reality and there is no evidence that it had any influence on Swift. Swift himself was to write in 1728 (though not published until 1765) An account of the Court and Empire of Japan, an imagined account, lacking any basis in fact, and in any event set in the year 331 A.D.4 These works do however draw our attention to the fact that whether in the context of an interest in exploration, or independently of it, there was some interest in Japan, as seven editions of Kaempfer’s History in three languages between 1727 and 1732 suggest.5 Swift’s account of Japan stands out in Gulliver’s Travels’ as Japan is the only real country as opposed to mythical ones among the nine territories, which his hero in varied circumstances visited (unless the island of Lugnagg (thus spelled on Swift’s map, but Luggnagg in the text), south-eastwards of Japan rather than to the south, may be, as suggested below, the Ryukyus (modem Okinawa)). Swift is known to have had an interest in travel literature, and in 1722, at a time when he had begun drafting Gulliver’s Travels, he noted that he was reading “many diverting books of History and Travel.”6 There is for our present purposes one important and revealing inaccuracy. Gulliver, in leaving Lilliput and Blefuscu, the first islands visited, in a small boat was rescued by an English merchant ship returning from Japan (the English however had been excluded from Japan from 1623).7 In contradiction of that inaccuracy, Swift notes in Part 3 (which is known to have been written at a later date than Part 1) correctly that the Dutch alone could trade with Japan.8 His account also adverts to numerous ties between Lugnagg and Japan.9 The pages on Japan itself come at the end of the somewhat ragbag third Part, recognised in literary study as the weakest of the four Parts of the work. This section describes stays in sequence in five territories,

50

EARLY JAPANESE TRADE

Laputa, Balnibarbi, Glubbdubdrib, Lugnagg, and finally Japan itself. Though it appears as the third of the four divisions of the work, it was written in 1726. Why it was written at all is not clear. It may indeed have followed on as an afterthought to his interest in the far northern reaches of the Pacific and in contemporary debate about lands in the vast northern expanse between Asia and America. Brobdingnag, the location for the adventures in Part 2, was, according to the map preceding the account attached to North America by a narrow isthmus. Beyond the isthmus, the king’s dominions ran northwards in a coastal fringe three to five miles in depth, cut off from the interior by a ridge of impassible mountains “thirty miles high”. This description reflected the fruits of Swift’s speculation about the conundrum of the North Pacific. By adding Part 3 at a later date Swift may have intended to balance Gulliver’s travels on the eastern side of the Pacific by travels in a region close to the elusive Tartary and its adjoining waters on the other side. It is unlikely that awareness of Japan, created by Sir Hans Sloane’s interest in the Kaempfer mss, was responsible for the section, though it may have prompted him in London in 1726 to consult Amoenitatum exoticarum, a work in Latin published by Kaempfer in 1712.10 Swift’s account of Japan is brief, depending on the edition consulted a mere three pages or so. Within Part 3, it follows after longer accounts of the four other islands and it is the briefest account of any of the locations chosen by Swift in the entire work for his hero’s adventures. It also lacks the sustained moral argument present in the account of other sojourns by his hero. In other words, Japan, a known - even if imperfectly known - archipelago on the fringe of Tartary, may have simply served as a vehicle for bringing his hero back to Europe (leaving Nagasaki on a Dutch ship bound for Europe). As for hard detail, all the account says is that Gulliver arrived at Xamoschi on the western shore of what appear to be the narrow mouth of Edo Bay, and later was escorted from Edo to Nagasaki. If his point of arrival is correct, it has to be Uraga. The word Xamoschi may be Swift’s creation and does not warrant further speculation. The benign behaviour of the emperor and the courteous reception of a stranger are in contrast to the accepted view held at the time of the hostile treatment of foreigners. As a favourable view of Japan was uncommon, it prompts speculation as to what influenced Swift. Swift’s reference to stamping on the crucifix or other Christian symbols was moreover not made in a critical vein. Introduced to mock the Dutch, it simply reflected Swift’s

GULLIVER IN JAPAN: SWIFT’S GULLIVER’S TRAVELS

51

dislike of the Dutch. When Gulliver who was passing himself off as a Dutchman sought to be exempted, the Emperor expressed surprise, saying that “he believed I was the first of my countrymen who every made any scruple in this point, and that he began to doubt whether I was a real Hollander or no.” Swift’s reference is any event incorrect. The requirement applied not to the Dutch or other foreigners, but solely to Japanese suspect of being Christians. His misunderstanding of the practice would suggest strongly that he had not read the text of Kaempfer’s History. II

The question about Japan’s place in Gulliver’s Travels is: why Japan at all and why in 1726? The part embracing Japan and neighboring islands was included, it seems, in revision in 1726. It reflected, especially in the detail of Swift’s map for Part 3 (which shows Ezo, the northern territory of Japan connected to Tartary), both the existing interest in Tartary and the active speculation about uncharted waters and lands to the east. A point of particular interest was posed by the mysterious southern shores of what later came to be recognised as the Kamchatka Peninsula, a question unresolved at the time. Were these shores part of a landmass attached to the northern coast of Tartary, or were they part of a larger mass which extended to North America, and if so were they broken at some point by a strait leading into polar waters? Awareness of these issues, long debated in Europe, was however quite independent of Sloane’s project of translating Kaempfer’s work. Swift’s knowledge of these issues did not need access to Kaempfer or to some knowledge of the content of the History. The map in Gulliver’s Travels includes Ezo and the “Straits of the Vries” (i.e. the strait of De Vries the Dutch explorer of 1644). This strait is depicted as dividing Ezo from the “Company’s land” (meaning Dutch East India Company’s land, an elongated strip to the east of Ezo and implying a territorial claim once made by the Dutch but not pursued later). All this information rests on earlier and accessible knowledge. De Vries’s account was available in English in Amoldus Montanus, Atlas japannensis, being remarkable addresses by way of embassies from the Eastindies Company of the United Provinces to the emperor of Japan, translated by J. Ogilby, London, 1670. Other possible sources for detail of Tartary are Nicholaas Witsen, Noord en oost Tartarye (1692 and 1705) and J.F. Bernard, Recueil de voyages au nord 1715. In fact speculation on these lands was widespread. Even if these works were not seen by

52

EARLY JAPANESE TRADE

Swift, they had reechoes in the general debate among European geographers in the early decades of the eighteenth century. Drawing on geographical knowledge in Europe, Kaempfer wrote in the manuscript of his History (completed by 1712) of the lands to the north of Ezo (later identified as the Kamchatka Peninsula). This is the country our geographers know exists but do not know whether it is connected to Tartary or America... Nor do they know whether it is connected to the two continents, and whether consequently no sound or passage exists. However hard I tried to find out something about these northern lands, I was unable to learn anything definite and credible... All of them seem to indicate that to the east an isthmus connects the great land of Tartary, to a solid landmass which they believe is America; they were therefore of the opinion that there is no passage at that point between the ocean and the polar sea.11

Knowledge by the Japanese themselves was equally slight. Kaempfer had recorded that “I have seen a variety of their maps at the Edo mansion of the Nagasaki governor. But like all maps they are slipshod, without any gradation; moreover they differ from each other.”12 The map in Gulliver’s Travels, reflecting Swift’s knowledge of the debate about this region, is reechoed in his words in the preceding Part (no. 2) on Brobdingnag that “For it was ever my opinion, that there must be a balance of earth to counterpoise, the great continent of Tartary, and therefore they ought to correct their maps and charts, by joining this vast tract of land to the north-west parts of America.”13 It was on the western rim of this uncharted world of land and water in the north Pacific that Swift located the islands of Part 3. Lugnagg was placed to the southeast of Japan in relatively well-known waters and at what his text states is a distance of a mere 100 leagues. The island of Balnibarbi however, described in the map in Gulliver’s Travels as “discovered A.D. 1704”, lies far out into the Pacific, and the island of Laputa is simply indicated by a name written on the map unembodied in a well-defined location. In other words, Japan’s relevance for Swift appears to arise from the existing curiosity among Europen geographers about what lay beyond Tartary in the vast expanse of the northern Pacific. Japan was simply a territory, well-known despite sakoku, on the fringe of what were variously unknown and little-known areas. These issues rather than Japan for its own sake may explain why Japan featured in his account.

GULLIVER IN JAPAN: SWIFT’S GULLIVER’S TRAVELS

53

III

Swift was at home in the highest social and intellectual circles of Dublin; he was equally so in London, though he had not been in London for many years before his 1726 visit. Some of his geographical knowledge could well have come through his contacts with Robert Molesworth (who died in 1725), a cosmopolitan figure well-known in both Dublin and London and greatly admired by Swift.14 The dates 1726 and 1727 are tempting. The fact that Part 3 in contrast to Part 1 correctly states the Dutch monopoly of Western access, suggests that Swift in 1726 had a more assured grasp of Japan than he had in 1722 when he wrote Part 1 and had referred to an English ship returning from Nagasaki. This glaring contradiction also has the interesting corollary that in adding or revising Part 3 in 1726 he failed to advert to the inconsistency which remained in an unrevised Part 1, which no doubt had relied on old and accessible English accounts of the pre1623 trade with Japan. His account of Lugnagg, very plausibly the Ryukyu Islands with their special and regular relationship, commercial and diplomatic, with Japan seems to imply Kaempfer in some form as the source. Hence the question remains as to whether he had access to Kaempfer’s manuscript of the History of Japan or to precise knowledge of its content. The ingenious, though inconclusive and at times somewhat forced arguments, in a book in 1977 that Swift “made specific use of Kaempfer” rest on speculation that “Swift might well have had access to Kaempfer’s work in manuscript form as it circulated among his friends in the Royal Society for more than ten years before the brilliant young translator, J.G. Scheuchzer, succeeded in arranging for its publication.”15 The speculation however loses credibility from the facts that Swift had long been absent from London until he paid a visit in 1726, that the translation came late in the day and that Scheuchzer, an employee of Sloane’s, made it under pressure within a remarkably short period of two years. A more recent author Robert Markely in a very well-documented and scholarly study of English awareness of the East takes the theme up again, speculating that Swift may have been made aware of Kaempfer’s work as early as 1724 through his friendship with Lord Carteret (lord lieutenant in Dublin from 1724 to 1730), who was a subscriber to the book: “Swift’s only opportunity to have seen Kaempfer’s work would have been two years later when he was in London and making final revisions to Gulliver.”16 There are several

54

EARLY JAPANESE TRADE

questions in this. The subscribers were sought in 1726–7: we can not count on Carteret being aware of Kaempfer’s manuscript in 1724. Carteret moreover was not the only subscriber with Irish connections: there were also the archbishop of Armagh, the bishop of Kilmore, and Lord Orrery. They suggest a readiness in intellectual circles to support publication of Kaempfer’s work. Swift, given his associations with this circle and with his interest in travel literature, would certainly have been aware of the project at least in general terms. However, he did not need to have access to the manuscript or even to be aware directly of its content. The interest in the project could have served at the time to draw attention to an earlier work of Kaempfer, written and already published in 1712, his huge Amoenitatum. exoticarum.17 This work written in Latin in five fascicles, amounting in all to 900 pages, was a compendium of Kaempfer’s learning on Asiatic themes including Japan (Kaempfer mentioned in it also that he had completed a manuscript History of Japan). Apart from Japanese flora dealt with elsewhere in the volume, relatio XIV on pp. 478–502 of the second fascicle provides a broad general account or Japan under the title of Regnum Japoniae, optima ratione, abgressu civium, et exterarum gentium ingresu et communione, clausum (The kingdom of Japan on good grounds closed to the exit of its citizens and entry of or communion with foreign peoples). The relatio concluded with a short section entitled Regni japonica status post occlusionem prosperissimus (The highly prosperous condition of the Japanese kingdom after its closure). This remarkable tribute also appears as an appendix in English in the published translation of his History of Japan in 1727. It was written in residence in Germany as much as almost two decades after his stay in Japan. In some respects its tone contrasts quite sharply with the main text of his History of Japan, which has detailed accounts of Japanese life and almost diary-like reporting on the two official journeys of the Dutch from Nagasaki to Edo. They record frequently his irritation at the restrictions he and his Dutch colleagues experienced. No doubt, with Japan remote in time and space, and dwelling on its many remarkable features, his view had become much more benign. It may have been this brief section which helped to shape the warm account of Japan in the few pages which describe the Japanese visit in Gulliver’s Travels. The brevity of Swift’s passages on Japan, and the fact that unlike the case of the other islands there is no moral argument, show that Japan’s only use to him was as a vehicle to provide in very few words a plausible location for Gulliver, after his many adventures, to find shipping for Europe.

GULLIVER IN JAPAN: SWIFT’S GULLIVER’S TRAVELS

55

The entire third part of Gulliver’s Travels is lacking in the coherence and clear focus of the other three parts. One is tempted to think that having described in Part 2 Gulliver’s adventures in the uncharted waters of the North American side of the Pacific, Part 3 was added rather maladroitly as a balancing theme on the lands close to Tartary, on the far side of the Pacific. While we can not be positive that he consulted Amoenitatum exoticarum, it seems probable: it would account for exact knowledge which went beyond the fleeting mention of Japan in Part 1. The direct value of the information for him was slight: it did however make his sparse statements factually accurate. Wider issues or moral themes based on Japan’s imposed and, according to Kaempfer, happy isolation from the outside world are avoided. In other words Swift seems to have had no interest in Japan itself, and his fanciful An account of the Court and Empire of Japan, though written in the late 1720s, only confirms the impression. His pages on Japan have attracted interest primarily because Japan as a theme was mostly absent from English writing of the period.

Source: First published in this volume

5

Japan in a Changing Asia: Achievements and Opportunities Missed1 ™

I will start with observations from two men, whose knowledge of Japan was considerable, and whose names, one to two centuries later, still ring with authority. The first is Vassily Golownin, captain of the Russian war ship Diana and a prisoner in Ezo (Hokkaido) in 1811– 13. The second is Sir Ernest Satow, acquaintance of Saigo Takamori, friend of Ito, Kido, Inoue and others, and thirty years later successively British minister in Japan and China. The observations of these two men praised the capacity for success of the Japanese. Golownin’s words were glowing, and I quote them: What must we expect of this numerous, ingenious and industrious people, who are capable of everything, and much inclined to imitate all that is foreign, should they ever have a sovereign like our Peter the Great: with the resources and treasures which Japan possesses, he would make it become in a few years the sovereign of the eastern ocean.2

If one can not pick out such a striking passage in Satow, it is only because insightful comments were scattered in many place. Even in later years, when he had become critical of Japan, he speculated on what a role Japan could play in forthcoming events if she were cast in the role of defender of China. The purpose of referring to these two men is because their views reached far beyond economic matters. You might expect me to talk about economic matters, because most western commentators have looked at economic aspects and speculated on reasons for Japanese success in Meiji 56

JAPAN IN A CHANGING ASIA: ACHIEVEMENTS AND OPPORTUNITIES

57

times or again in later years. Concern with economic development has dominated western writing on Japan. Many, perhaps most undergraduate courses and many text books still concentrate on the century after 1868 and primarily on reasons for Japan’s successful industrialisation. Moreover, most books and virtually all undergraduate courses in Japanese studies, were launched post-1945. Their approach rested on the fashionable assumption that Japan’s development after 1868 could be explained by a modernisation process, and drew on new theories of development in the 1950s intended to make impossible a recurrence of the depression of the 1930s and to quicken diffusion of the benefits of growth to less developed countries. Walt Rostow’s Stages of economic growth in 1960 picked Japan out as the sole case of an allegedly less developed country which had attained take-off. The interest in Japan’s success was in no small measure inspired by the cold war, and by the fact that India had modelled its development plans on the Soviet and centrallyplanned model. Hence as a model on private enterprise principles or at least on politically more orthodox principles and of proven success, Japan was seen as an alternative to the new and ideologically suspect Indian model. However Japan’s economic success does not require special reasons. Tokugawa Japan was, if different from Europe, nevertheless highly developed. Japan’s large internal market was all the more effective because it was one of relatively equal consumers: that is there were relative few very rich, and outside the Tohoku also very few poor. In modern times the only periods in which Japan’s foreign markets accounted for a high share of income were the 1910s and 1930s. Both periods were years of growing inequality in the distribution of income. The year 1918 was in consequence a year of great unrest, and if there was little unrest in the late 1930s, social alienation was growing rapidly and would surface strongly in 1945. Before the nineteenth century Western thought had not been patronising about Japan. The Jesuits paint a favourable picture, as do the four later writers, Kaempfer, Thunberg, Titsingh and Golownin, who knew Japan and wrote about it. Europe’s material and military success in the nineteenth century did however lead to a patronising and arrogant outlook, which coloured policy and attitudes. Japan, it was argued, benefited aftei mid-century by exposure to the superior values of the West, and in the last analysis, Japan, lacking originality, was an imitative country. In relatively recent times, this was reflected in gross underestimation of Japanese military capacity and Japanese military matériel. Pearl Harbor among many other things was a strategic and

58

EARLY JAPANESE TRADE

technical triumph of its day. The only time when this outlook changed was in the late 1980s when it was thought that Japanese management was the secret of the country’s remarkable business success in the 1970s and 1980s. It was therefore argued that if it should be imitated, the West would have the same economic success. It was the first and only occasion when the advocacy of imitation was reversed. Alas the world of simple explanations is not a real one. The truth is that much Japanese and American management is discredited, and American and Japanese success alike, both in their way remarkable, can not be explained by a simple model of management. If there was imitation of the West, the urge was less a wish to adopt Western styles, than the need to be able to resist a challenge of political take-over, and to adopt the weapons and literally to match the fire power which the west displayed in its onslaught on Asia in the 1840s and 1850s. Fukoku kyōhei was a term from the Chinese classics, not one invented in the 1850s. And when it was popularised it did not mean the building of an army. Its translation in the context of the 1850s is a wider one of defence, not army as such The military obsession of the 1850s was canon, not manpower, and in the 1860s steam ships and rifles. The army itself was for decades to remain a small one, too small in point of fact for the needs of the country. The cultural urges towards imitation were limited in their chronological expression, in essence confined to two periods, rather far apart, the 1870s and the 1950s. These two periods were of course traumatic ones. The first is explained by the need to undo the unequal treaties, and by reducing the cultural distance between East and West, to persuade Westerners that they would be safe in surrendering their extraterritorially or the immunity from Japanese law which they enjoyed. This was responsible in particular for what one might describe as the creation of a Western infrastructure of justice and law courts. The second period reflected defeat and occupation in 1945. Life styles in the 1930s, upper middle class and the new salary men of Osaka and Tokyo apart, were remarkably similar to those of the 1870s. Wider change came only in and after the 1950s: some of the changes were the fruits of Japan’s new economic momentum, some primarily part of a phenomenal world economic boom spread over a quarter century. Think for instance of the simultaneous dramatic changes in rural Ireland or rural Spain. In some respects, in social as opposed to material ones, from the Yoshida Shigeru years until the 1990s, the promise of change was slower. Politically, change has not only been slow, but in recent times has fallen behind

JAPAN IN A CHANGING ASIA: ACHIEVEMENTS AND OPPORTUNITIES

59

the popular will. The central impasse in political decision- making is not much different in 2002 from that between 1890 and 1912, the first two decades of Japan’s new parliamentary institutions. If one looks at Japans’ encounter with the West, two things stand out. Japan was able to cope with the challenge and also displayed much confidence in doing so. Japan had long studied the West, something that no other Asian country did. Japanese also from the time that it was clear that they would have to live with the West immediately experienced the urge to go abroad. Yoshida Shōin famously had sought to board Perry’s vessels in 1854, and Tokugawa Nariaki, a staunch supporter of Japanese defence, proposed taking 400 rōnin to the West (the germ of the idea of the Iwakura mission fourteen years later). Both privately and on either shogunal or domain account people went abroad from 1860. There was a series of official missions to the West. There was no reason in 1860 to deposit the ratification instrument of the 1858 treaty in Washington. It could just as easily and much more cheaply have been done in Japan. However, the mission of 1860 sought to deposit it in America and the delegation made a point not only of doing so but by making the voyage on a Japanese steamer. The Western cultural fascination with Japan which was strong in the 1870s and 1880s was caused in no small measure by curiosity about a people who while they had not wanted outsiders to come to them, had then, uniquely among Asiatic nations, taken the initiative in visiting the West, and with great formality. The famous Iwakura mission in 1871–3 was set both in the longstanding pattern of study of the west, and in the more recent one of missions sent abroad. Kume Kunitake, secretary of the mission and who wrote its report, saw Western technological leadership as something that had emerged in the preceding half century. His account is an absorbing one of the West, and is instructive for Westerners to read.3 But of course it is even more interesting because of what it tells us about Japanese outlook. Kume, immensely impressed by the West though he was, was confident that Japan could catch up, and that the technological change in Europe in the preceding fitly years, could be repeated in a short time-span in Japan. In studying Japanese relations with the west several preliminary observations are necessary. First, Western concepts have been widely used, even by Japanese. Thus, Maruyama Masao started his great study of Japan thought in 1952 with the name of Hegel, the German philosopher, mentioned in the very first line. The book at large is written

60

EARLY JAPANESE TRADE

in a western theoretical framework, and is greatly overrated as an explanation of the evolution of Japanese thought. On a less intellectual level, sakoku has been seen as an oddity, as something which was not normal and represented the secretive power of the shogun and the xenophobia, of a nation Yet it was realistic for its time, something proved by the fact that it faced no challenge for almost two hundred years. The question was simply whether what was realistic in 1635, 1771 or 1793 was still realistic in 1840. The second of these preliminary observations is that Japan is seen as having posed a problem to other countries by having closed its doors for two and a half centuries. However, it would be just as easy to say that over centuries Europe was Japan’s problem or, more widely still, Asia’s problem: a proselytising Christianity, revived military aggression in the 1840s, and from the time that the north Pacific was exploited intensively in mid-century, a wish to turn it into a Western and increasingly American lake. The Western presence in Asia was inherently distorting. In the early seventeenth century, it had already threatened to destabilise East Asia, but the challenge receded as the West found India and North America as much as its naval capacity could cope with. All that changed in the 1830s. A series of voyages of discovery and surveying had progressed from the 1780s; whereas Kaempfer and Thunberg’s work had been published only with long delays, rapid publication and even more strikingly almost simultaneous translation into other major languages of new books began in the 1790s. The publication of Titsingh’s writing had marked time in his own life, but several years after his death, over fifteen years from 1819 it appeared in both French and English. The invasion of China in 1840 represented the beginning of aggression in its modern context. The Opium War must be among the least defensible wars in history; its immorality was repeated in 1856 in the Arrow War, Britain was now joined as villain by France. The destruction of the Summer Palace in Peking was an extraordinary piece of gratuitous vandalism, wantonly singling out a historical or cultural monument and heralding a sort of Western derision of other cultures. Intrusion into the affairs of Japan was justified on several grounds. Japan, it was claimed, was a despotic country, and even its people would welcome a Western intrusion which would destabilise its government. Its people wanted, or would want change, and if they had long been deprived of something so beneficial as foreign trade, they would want it when they saw what was on offer. Older books had been very favourable in their judgement of Japan in a number of respects: they

JAPAN IN A CHANGING ASIA: ACHIEVEMENTS AND OPPORTUNITIES

61

had also argued that trade prospects were not promising. In the midnineteenth century they were rubbished by the claim that their good opinion came not from facts observed but from mere copying one of another The emphasis by the Americans in 1853 and 1854 on succour to distressed mariners and on the law of nations was no more than an excuse to justify blatant intrusion into a foreign country which for two and a half centuries had troubled no neighbours. If a country did not harm its neighbours, it was not under an obvious obligation, under international law such as it was, to open its frontiers, and decidedly so when the 1825 uchi harai decree was modified in 1842.4 Given the moral dilemma of forcing access on another country, the Russians in 1792 and in 1804, and in 1837 the American vessel, the Morrison, dressed out their démarche with an effort to return Japanese castaways. It is only for the year 1825–1842 that uchi harai really existed. It also dealt with a specific problem which began in 1818: the appearance of whalers, English-speaking, and hence novel. As the incident of the English warship the Phaeton in Nagasaki in 1808 was still uppermost in Japanese thinking, the fear was that they were investigatory missions for a future attack. In any even, if there had been some ten whalers between 1818 and 1824, for one reason or other few foreign vessels were sighted for the next decade. I am not clear that any were fired on before the Morrison in 1837, which of course made much larger claims that the whalers, and was also the first instance in a century and a half of a demand for trade by English-speaking visitors. A decade or more later American opinion was angered by the long detention and questioning of a number of Americans shipwrecked in Matsumae in 1848. It is not clear that all the American claims about the treatment of this crew are not in fact exaggerated, though it is true than, alarmed by events in China from 1840, cases of shipwrecked mariners, even after the modification of uchi harai in 1842, ran into obsessive Japanese suspicion and take up many pages of the Tsūkō ichiran zoku shū. There was failure by the Japanese authorities to distinguish between foreigners making an orthodox (and open and polite) request for fuel, fire and water and those who first shipwrecked and moving around, aroused suspicions of infiltration. The number of actual post-1842 cases is few; shipwreck – not a feature of the relatively numerous cases in 1818– 1824 which had triggered the 1825 decree – was more common; their location in the underpopulated and the vulnerable north was novel; and doubts by officials were fuelled also by the evidence of a

62

EARLY JAPANESE TRADE

growing number of foreign war vessels in Japanese waters. In doubtful cases the mariners were required to stamp on Christian symbols, the procedure in resolving past doubtful cases. It was of course objectively inappropriate for cases of shipwreck and more positively still for a greatly changed international environment. In at least one case the behaviour of the crew arguably contributed to the unhelpful circumstances.5 The public professions of American motives in 1853 were not to be taken at face value. They were already the second largest purchasers of tea at Canton, and there was a consuming interest in commercial expansion in the Pacific. It has been argued that the west was not interested in Japan, that its real goal was China, that Japan was a secondary interest, and therefore as long as Japan partially opened it would be free from serious Western threat. It is true that China was the major interest, because of the belief that its vast population meant a potentially vast market. However, Western ambitions were limitless and competitive. China was ruthlessly and cynically plundered. The French started their conquest of Indochina in the 1850s. The little island kingdoms of the north Pacific, long almost ignored by Westerners (occasionally visited but rarely a subject of a claim to sovereignty), began finally to fall prey from the early 1850s to acquisition by French, Germans and British. From 1898 also America carved an empire in the ocean. The story of Japan and the West has been perhaps looked at too readily in terms of the assumption that what the West was interested in was simply aid for mariners and coaling stations for ships from California destined for China. It is worth reading the official account of the American expedition; the trade interest is explicit. For an official report, it was surprisingly hostile to other Western countries. Resorting to a bluntness rare in official documents, it emphatically blackened them all in their ties with Japan. What saved Japan in a sense was that the race to dominate the East, confined as late as the 1830s to Britain, now had other participants. In China, France became an ally of Britain in the warfare of the late 1850s. In the case of Japan, America, Britain and Russia almost simultaneously made their case, all with a persuasive display of naval power. There was an avalanche of treaties in 1854–5 and again in 1858. Aggressive action by any one was ruled out by the presence of others. The attack in 1864 on the Chōshū batteries in the Shimonoseki strait was carried out in concert by the warships of four nations. At first sight this might appear ominous, but in fact it was the protection

JAPAN IN A CHANGING ASIA: ACHIEVEMENTS AND OPPORTUNITIES

63

of Japan. The pattern reoccurred in the course of trade negotiations in November 1865 in the concerted anchorage of Western fleets in Osaka and again for the opening of Osaka on 1 January 1868: the big countries had large fleets, the smaller ones a token presence. Western occupation of lands had an unwritten law in the nineteenth century: prior claim by rivals (of course only Western ones) was honoured; so much jostling by outsiders in Japan ruled out the staking of either exclusive or territorial claims. If despite appearances to the contrary, this situation was benign, the progression of Western intrusion in other weaker or less wise countries was unsettling, especially in the surge of Western imperialism of the 1880s and 1890s, a counterpart to the reckless exploitation of Africa in the same years. The acquisition of Hawaii and the Philippines by the United States was the basis of a dominant strategic presence by the Americans in the north Pacific, and created a new significance for naval power. It is no accident that an American naval officer, Mahan, had already written in 1892 the classic book on sea power, as well as in 1900 a revealing though now less well-known book, The problem of Asia and its effect upon international relations.6 A race to stake commercial interests in Manchuria and China quickened. In one sense this was an effort, as all the Western countries were stymied by their moral code of prior claims, to exploit commercial opportunities, when territorial claims were not feasible. The Americans, fortified by their brand new territorial position in Pacific islands from 1898, losing no time, loudly called for an Open Doors policy in China. This morality was reflected in the famous triple intervention in 1895 (by Russia and two new contenders, France and Germany) to compel Japan to give up its newly acquired stake in the Liaotung Peninsula. In this instance very high standards were imposed by the members of a not very moral alliance on an Eastern country. But Western morality was itself of course one of opportunism. Tempted by China’s increasing weakness, two of the partners in the triple intervention, proceeded to stake claims, Russia in the Liaotung Peninsula which Japan had just been compelled to surrender, Germany in the Shantung Peninsula. Not to be outdone, Britain, though not a party to the triple intervention, acquired Weihaiwei in the Shantung Peninsula. The effects of this cynicism had a devastating effect on Japanese official thinking, showing that there was no morality in Western-style international relations. The Western presence, in some ways repeating the pattern of the 1860s, could however still benefit Japan. The Western powers with remarkable insistence sought Japan’s

64

EARLY JAPANESE TRADE

participation in the protection of the foreign compounds in Peking in the Boxer Rebellion, of course only as a junior partner, but nevertheless as an associate. The Japanese contingent was the largest, by far the most disciplined, but, given the pecking order of West first and last, was not given the overall command. More strikingly still, as all the Western countries were alarmed by the politically unstable nature of Korea, the Japanese interest there became increasingly welcome as the most effective defence, before and after 1904–5, against Russian encroachment, from its growing power base in Siberia. The third of my preliminary observations is that Japanese in conversation sometimes see Japan as small, or poor. What they mean, I think, is vulnerable. Yet the sea, or an insular position, is an advantage. After all, by the end of the First World War, it was landlocked Germany which was starving, not insular Britain. But there is, I think, something wider in such observations, a reflection of an abiding uncertainty about Japan’s status relative to other nations. This theme ran through the years from 1902 to 1919, and into the demands of Japanese negotiators in the later naval agreements. Today, in proportion to its wealth, the fact that it has not a seat on the UN security council seems to imply a lack of status. Yet, there are two questions. If Japan were a permanent member, would it have a clear sense of the role it should play? Would it be prepared to stand up to the United States, and does Japanese public opinion know clearly what its role should be, or perhaps more precisely would opinion be in agreement as to what role it should play? There is a more fundamental reason about permanent seats on the Security Council. The concept of permanent members in a nominally egalitarian world is outmoded. It was a consequence of the outcome of the war, as indeed in a preceding war were the the five “first-class” states at the treaty negotiations in Versailles in 1919. Essentially, it is a divisive concept. It needs to be rethought fundamentally. Both Britain and France have ceased to be world powers. In the short term, moreover, with the collapse of the Soviet Union, we face the terrifying situation of a single superpower. China, Russia, and France stand up to a degree to the US; Britain does not; does one really want another permanent member which does not? The issues are stark. Either permanent members, using their weight for what it is, must be prepared to stand up to pressures; or better still, if the reconstitution of permanent membership should come up for reconsideration, the whole concept should be scrapped. These observations make it possible to return to the politics of the Edo period, and the issues of opening up Japan. Two considerations

JAPAN IN A CHANGING ASIA: ACHIEVEMENTS AND OPPORTUNITIES

65

have to be stated. The fist is that sakoku was a successful policy. It did not damage Japan economically and it did not harm Japan politically. Among other things, it made possible the demilitarisation of Japan. That aspect of Japan was described in Noel Perrin’s undeservedly neglected book7 later translated into Japanese by Professor Kawakatsu. Perrin meant the absence of artillery and rifles. But demilitarisation also meant that swords and especially the armour of samurai aged. Samurai became civil servants, and a samurai in full armour became unknown, so much so that in 1866 when the second expedition against Chōshū was being prepared samurai cut an absurd figure by appearing in the armour of their grandfathers. Demilitarisation is itself a remarkable phenomenon, as firepower has historically had the attractions of alcohol: for most once tasted, it is impossible to give it up. There are costs in what happened in Japan. But let me be clear about what they are. The first was that the country was undertaxed (and the Japanese taxation system is still inferior to Western ones as a means of detaching people from their money), and the second that the country had no military traditions The military traditions of the 1890s were new ones, not old ones, though with the progress of militarism there was an effort, accelerating in the 1930s, to suggest that they represented profound historic traditions. The second consideration is that the Japanese negotiated very well in the 1850s and 1860s. The confrontation has always been looked at in terms of Perry calling for the opening of Japan, and Japan having to acquiesce, not simply only because of Western pressure, but because of internal support for change. When a divide emerged in the 1850s it was not over the merits or demerits of sakoku: it revolved around the consequences if Japan refused to make concessions. This was a rational debate. It had nothing do with the merits or demerits of a closed country; it had every thing to do with the consequences of war, and especially with Japan’s ability to resist the Western countries with their advanced vessels and firepower. Confidence in the 1850s did not come by accident, nor simply out of Japan’s history and the personality of its people, though these things helped. Japan had long studied the West. Isolated country though it was, it had also consciously developed a capacity to negotiate when the challenge came. Through the limited channels left open by sakoku, Japan had gradually deepened this competence. From the outset conversancy with Portuguese and then with Dutch as the successive lingua franca of Asia existed among a small corps of linguistically competent officials.

66

EARLY JAPANESE TRADE

When Hirado was closed in 1641 and the Dutch transferred to Dejima, the artificial island in Nagasaki bay, the interpreters not only moved but from the status of private employees of the Dutch became direct employees of the shogunate. From the 1780s, when fear of the Western threat for the first time since the 1640s recurred, Japan began a wide-ranging though limited political study of the West. This awareness of the outside world was accompanied by a gradual creation, starting when Matsudaira Sadanobu (1758–1829) was senior councillor or prime minister (1787–93), of an administrative competence to cope with foreign challenges. The result was that Japan had some elements of strengthened administration for foreign affairs by the 1850s, a highly competent knowledge of Dutch (and even some knowledge of other languages), and a practical if incomplete understanding of the West. This is too long a story to present, but in Nagasaki the Oranda tsūji, in Edo and Osaka a small but wellplaced circle of officials, medical doctors in many parts of Japan, the Banshowagegoyō, and finally the Hayashi-compiled Tsūkō ichiran all are part of it. It was moreover Hayashi Akira, the head of the team compiling the Tsūkō ichiran who headed the negotiations with the Americans. The Americans were impressed by him. When the real challenge came in 1853 and 1854 from the largest groups of warships ever seen off its coasts Japan was surprisingly capable of dealing with it.8 In 1853–4 concessions were kept to a minimum and from 1857–8 Japan not only in realistic mode made concessions but in tortuous negotiations succeeded in dragging out over a period of ten years their full application. While no one wanted outsiders, a degree of consensus was established by acceptance of the argument that the unwanted treaties would buy time and, when renegotiation became possible under treaty terms in 1872, it would take place from a position of strength: foreigners could then be confined to a few Nagasakistyle enclaves in isolated centres. The missions to the West from 1860 to 1871 gradually revealed that the West was too powerful to admit of the radical renegotiation that in 1858 or in the early 1860s had seemed attainable, and for some optimists or bold spirits sooner rather than later. Hence the concept of radical undoing of the treaties was replaced by a limited and realistic one of bringing to an end the humiliating concession of extraterritorial sovereignty wrung from a defenceless Japan in 1858. The prospect of achieving this lay in creating new institutions reassuringly like Western ones and under which Westerner residents would feel safe.

JAPAN IN A CHANGING ASIA: ACHIEVEMENTS AND OPPORTUNITIES

67

Thus, if one takes the overall view, Japan did very well. This achievement was at the time observed by Asian countries, and there was an interest both in Japan’s modernisation, and in its comparative success in holding foreigners at bay. Even the king of tiny Hawaii in 1881 proposed an alliance of Eastern states with Japan at its head. Japan itself seemed to go from strength to strength. It negotiated the end of the unequal treaties in or from 1894; it made a treaty on a basis of equality with Britain in 1902; it defeated a Western power in 1905. The obverse side to Japan’s success is that it did not appreciate the wider role which it could play. It despised the other Eastern countries, precisely because they were not able to stand up for themselves. In 1885, noting Chinese inability to stand on its own feet in defence of its interests, Fukuzawa abandoned the idea of common interest among Asian countries: “Japan should part with Asia”. Japanese who had served in Korea and Manchuria during the Russo-Japanese War of 1904–5 are said to have returned home with undisguised scorn for the barbarity, poverty, filth and servile nature of the inhabitants they had encountered. In 1917, Gubbins, formerly a British diplomat in Tokyo, in reading the Japanese press for the British Foreign Office reported that the Japanese were aware of Chinese ill will occasioned by the contempt with which they were treated by the Japanese in both China and Japan.9 The process of Japan and Asia parting began with Korea. Harry Parkes, the British representative in Tokyo, in 1876 wryly noted that Japan had imposed on Korea the sort of treaty that it complained of loudly in its own context. In Korea, far more sealed from the outside than Japan had been, divides opened up between conservatives who wanted to maintain under a Chinese umbrella its isolation and others (“liberals”) who wanted Korea to establish direct contact with the outside world. Koreans were already present unofficially in Tokyo in 1880, and at that time an official mission was shortly expected “to study recent improvements in Japan on the European model”. As late as the 1890s those who looked to Japan were disappointed that Japan’s interest had been less vigourous than they would have wished, and even in the first decade of the new century there were Koreans who favoured the ties with Japan. China too had looked to Japan as a model for reform and as an example of successful response to external encroachment. Sun Yat-Sen had lived in Japan, and 1300 Chinese students entered Japan in a single year 1904. In Waseda alone 651 students had graduated up to 1910.10 It should be stressed that any Japanese leadership of Asia could only have been moral or intellectual. Real divides within countries between

68

EARLY JAPANESE TRADE

conservatives and progressives ruled out any more formal ties, and in any event had they occurred, they would have exacerbated the growing tensions with the Western powers. Japan’s Korean policy at the outset was also driven by the fear that if it did not act with haste, the West would get there first, with consequences which were incalculable. Japan’s future was not created in 1928 and 1931, and by the militarists who ignored both the war ministry and the cabinet. The failure began in and after Versailles in 1919, and by moderate politicians. While Japan spoke in 1919 about equality, at the same time it wanted Shantung against the wishes of China, and America. The welcome in 1919 by Hara, the prime minister, of news of success and even more the tone of public and press reporting, though comprehensible in the sense that the Japanese negotiators would not return empty-handed, betrayed total insensitivity to the underlying implications. In other words Japan had pushed herself further into the club of European powers which against American wishes held extraterritorial interests in China. The Shantung issue had a significant role in the deteriorating relationship with the United States. The year 1918–19 was a remarkable year; the rights of nations became an issue worldwide and high expectations were entertained by oppressed countries as far apart as Ireland and Korea of the prospect of recognition at Versailles of their rights. Disappointment on that front led to popular manifestations in Korea and to brutal repression in 1919. Versailles, abandoning the rights of China, led to the 4th of May movement in China in 1919. Nationalism in Asia acquired a new force. For Japan the price of the Shantung Peninsula was a heavy one. The two issues, recognition of racial equality among nations, and the formalising of Japanese occupation of Shantung were intertwined in a complex fashion. Though they could be seen (once Chinese rights were disregarded) as independent of one another, they were interrelated in the sense that recognition of Japan’s economic presence in Shantung would not only give Japan a new status in China but by doing so at a full-blown international conference, put Japan on a footing of real equality with the Europeans. China’s delegation, though vociferous on the question of Shantung, did not count as one of the five major powers at Versailles (the United States, Britain and France, plus the “new” powers of Italy and Japan). Subtly it changed Japan’s position from defence against Western encroachment into an exploitative role identical to that of other countries. Japan had no colonial past. The tragedy is that it threw away the moral advantages of being the sole country to negotiate itself out of a

JAPAN IN A CHANGING ASIA: ACHIEVEMENTS AND OPPORTUNITIES

69

status of imposed inferiority, and in the case of Russia of defeating a Western power. In 1919, adding Shantung to its interests in the Liaotung Peninsula and its possession of the strategically sensitive Korea, it had become the major regional power, effectively outweighing the combined Western presence. The West was for the first time in a century in a situation of inferiority in East Asia. Moreover, Japan’s position as a regional superpower was secure. Though not wholly clear at the time, Britain was totally overshadowed by the United States in the Pacific, and was no longer either an independent or a decisive force. While Russia was arguably a menace, the Russian threat was one which could have been used to command support for a strong Japan from the other Western powers. There was no immediate justification for the expensive and sustained military and naval build-up from 1918 to 1922, which threw into doubt Japan’s good intentions. More seriously, it took place without a clear framework as to how it served Japan’s long-term interests. It represented army and naval politics: the army seeking a challenge in Siberia to justify its role, the navy one in the Pacific for its expansion. The problem is that success had come too quickly. For Japan Versailles was dominated by the question of status, recognition of racial equality (where it failed), and material gain (where it succeeded). The experience of Versailles was all the more bitter, however, because it showed that Japan’s inclusion among the big five quickly was even at the time a hollow prize. While this status prompted the proud refrain that Japan had become” a firstclass nation”, the actual negotiations were conducted in quite a different spirit. The big decisions were taken by Wilson, Lloyd George and Clemenceau at private meetings, and the status of Japan and Italy, the newcomers in this select club, was maintained vestigially in a series of much less important meetings of foreign ministers. On racial equality, the Japanese encountered dismissive attitudes: in particular the Englishspeaking countries failed to see the point of the Japanese concern. For the Americans, the issue was on the part of the Japanese a mere bargaining counter, which they intended to trade in, in return for an American climb down over Shantung. At the first plenary session of the conference, Makino Nobuaki, the effective head of the Japanese delegation, in a powerful and passionate address raised the principle of equality. A vote in the committee devising the charter of the League of Nations highlighted the humiliation. On their proposal to include a reference to racial equality in the preamble of the Charter, the Japanese had a majority on their side: the Americans, the British and Australians were

70

EARLY JAPANESE TRADE

outvoted by the French and other nations (the actual vote was 11 to 6). However, President Wilson, highhandedly declared that the issue could not be pursued further, as it implied interference in the internal affairs of nations and hence would require unanimity. Japan was isolated diplomatically not from 1933, the year of the Manchurian Crisis and walk-out from the League of Nations, but from the years 1919 to 1922, which should be seen as years of diplomatic failure, imposed incidentally as much by public opinion as by politicians. One can of course see how things rankled in the 1920s: American arrogance in 1919 over the recognition of equality, the opposition in naval conferences to any recognition of a real change in the balance of sea power in the western Pacific, and the immigration issues of 1920 and 1924. The immigration issues, while not in material terms important for Japan, were profoundly so in a racial context, reinforcing the recollection of the painful racial attitudes displayed at Versailles Under American pressure Japan had to withdraw its troops from Shantung in 1922. But retaining its economic interests, the issue remained volatile,11 The net effect of all these events was that Japanese isolation was more complete than at any time since the beginning of its modern relationship with the outside world in 1853. From 1858 it had enjoyed a very good relationship with the United States (which was the first country to display in concrete terms a readiness to concede its rights under the unequal treaties); it had an alliance with Britain from 1902 to 1922. In earlier years, Japan was looked up to in Asia, and after the war in 1894–5, many younger Chinese looked to Japan as a model of regeneration. However, Japan’s moral authority in Korea dissipated as early as 1909 (a year before the formal annexation of Korea) and the year 1919 ended any lingering prospect of a meaningful relationship with China. China posed difficult problems for Japan: should it be in essence cautious as in the Shidehara foreign policy of the 1920s or more forceful as in the 1910s and 1930s. The Western countries every bit as much as Japan should have displayed more insight into the existence of new forces in China. A key problem is that Japan and the West alike retained doubts about the other’s intentions. A problem too was the persistent political instability of China, which meant that while the West was loudly critical of any dramatic steps by Japan, even in the 1930s an active Japanese presence was seen as part of the reinforcement of the status quo. In other words, an external presence by all in China was seen as permanent, and a Japanese role on mainland Asia was welcome, even necessary, provided that it did not upset the existing

JAPAN IN A CHANGING ASIA: ACHIEVEMENTS AND OPPORTUNITIES

71

balance of external interests within the country. It was the welfare of external interests, not the welfare of China, which was uppermost in the minds of all. If Japan’s success had come too quickly for Japan, it came too quickly for other countries too. Thus the weakened Britain and the ambitious US presented a united front at the naval conferences of 1922 and 1930. With hindsight one might say that all should have behaved differently. Japan should have played a longer and more patient game. The others should have seen the changing balance of power, and not like King Canute have sought to halt the tide. As Britain’s mantle was spread over a large area, both Indian Ocean and Pacific, effectively the conferences sanctioned American naval power in the broad reaches of the north Pacific itself, a phenomenon itself as new as the rise of Japanese naval power. While they covered sea power in the Atlantic as well as the Pacific, the essential concern of the conferences, as the very precise detail on regulating naval bases showed, was the Pacific. With Germany eliminated by defeat and treaty from the Atlantic, post-1919 American navy power was essentially in the Pacific. The conferences were in effect an effort to halt the naval militarisation of the north Pacific, though more by preserving Western supremacy there than by concession to a changed balance of power in East Asia. Account must be taken too of public opinion, or more precisely public opinions, racially-tinged in the West, especially in America, and in Japan both strong and volatile (and hence further diminishing the room of manoeuvre of already politically impotent politicians). America and Japan were on a collision course, and tensions, dramatically exposed at Versailles, over China reinforced the drift. American intervention had of course brought together in 1905 the two warring countries of Japan and Russia at Portsmouth. The intervention, however, was prompted primarily by America’s new role as a Pacific power and by a fear that Russia might in the end prevail on land in East Asia. The tradition of goodwill to Japan, a long-standing one, was beginning to change rapidly through America’s transformation almost overnight in 1898 into a Pacific colonial power. Its commercial stake also expanded in both Manchuria and in China. In Japan, political weakness at the centre, the inability of cabinets to control ministries, and the fact that army and navy were rivals, meant that Japan was setting out, more or less without an overall policy, to seek the impossible. It was setting out on a course of rivalling the two largest armies in the world (China and Russia), and the two largest navies (the United States and Britain).

72

EARLY JAPANESE TRADE

No one could have anticipated the scale or rapidity of military success in late 1941 and early 1942. It could of course not last in the long term both for logistical reasons and because of Japan’s finite material resources, manpower and industrial capacity. Yet if it was a failure and inevitably so, the collapse of all the long-standing empires, British, Dutch and French, meant that like Humpty Dumpty they could not be put together again. The West variously retreated or was driven from Asia after 1945. Japan’s earlier role model for Asia as the first country to match the West, and the dramatic though short-lived success of its armed advance in 1941–2, effectively ended empire in Asia. The irony is that Japan itself, though author of all this change, forfeited the moral authority that would have given it a secure and privileged base in Asia. In some respects, Japan’s international position is extraordinary and unique: a political dependence on the United States unprecedented for a major economic power, a comparatively good relationship with European countries, and a relationship with its Asian neighbours whose future course is far from clear and the rancours of which are cast in an abiding historical framework.

PART II

Statistical Resources of and Interactions with Tokugawa Japan

Source: Japan Review, No. 18 (2006), pp. 129–180

6 Population

Tokugawa Population: The Archival Issues1 ™

While the surviving Tokugawa demographic data have been brought together in a number of studies, notably in Sekiyama’s work published in 1958, the sources themselves as an archival residuum – as opposed to the actual figures contained in them – have never been analysed systematically. Scholars have ascribed primacy, as a source, to Suijinroku, published in 1890 by Katsu Kaishū, who had served both as a bakufu and Meiji government official. In addition, historical population research has proceeded on the assumption that Tokugawa secrecy policies were effective and resulted in a lack of circulation of information. For the han, some knowledge exists of how census returns to the shogunate were compiled. Processing by the shogunate is totally obscure. Apart from converting han data into kuni equivalents, no adjustments seem to have made by shogunal officials to the han figures, which thus retain both the uncertainties associated with the raw returns and the original variations in coverage. Of the surviving summary tables for individual census years over the signature of an ōmetsuke and a kanjō bugyō, a mere four can be regarded as compete copies. Moreover, surviving returns are not official documents (or copies by officials acting in an office capacity), but copies made privately, often from existing private copies. Copies usually relate to a single census, and a mere five or six examples incorporate data from more than one census. Time series of census data were unknown, and the long table of Osaka population constructed by a machi bugyō named Isshiki, who drew on registration data, is unique. 75

76

EARLY JAPANESE TRADE

TOKUGAWA DEMOGRAPHIC DATA

This paper originated in a study of Tokugawa statistics as part of a wider study of administration in the key 1790–1853 period when Japan showed that to a degree it could strengthen its institutions, and when 2 the prompting came from the fear of a foreign threat. While they do not relate directly to a foreign challenge, statistical sources are important in terms of assessing the nature of administration, and they also throw light on an important issue, i.e., the assumed hermetically sealed and secretive nature of government. In fact, administration was open in the sense that information was disseminated quite widely though privately. While the surviving Tokugawa demographic data have been brought together in a number of studies, the sources themselves, as opposed to the actual figures, have never been analysed systematically. That is the object of this paper. A later paper will deal with trade figures. Apart from data relating to precious metals, trade statistics sparingly feature in Tokugawa sources. They emerged primarily from the Nagasaki bugyōsho ዊ⾜ᡤ, and related to regulation of trade with the object of ensuring that the outflow of precious metals and copper remained within predetermined limits. Some of the figures appear in the Tsūkō ichiran ㏻⯟୍ぴ, a vast compendium of documents on external relations compiled by the Hayashi family, in the early 1850s. Tsūkō ichiran illustrates how figures originally compiled in Nagasaki were available in copies made or kept in Edo, and still could be consulted one or two centuries later. For other branches of trade (coastal trade), statistics existed also at least in theory for both Edo and Osaka, as a result of decisions made at much the same time as shogunal population compilations for Japan began to emerge. Yoshimune, the eighth shogun, is central to both endeavors which corresponded to an attempt, with public welfare in view, to invigorate central administration. In Meiji times, an interest in demographic statistics, or at least awareness of them, grew with Katsu Kaishū ຾ᾏ⯚or more accurately as a result of the publication of his compilation of documents from various sources. Before Katsu’s population figures appeared in his Suijinroku ྿ ሻ 㘓 in 1890, some Tokugawa data were published in works by Yokoyama Yoshikiyo ᶓᒣ⏤Ύ, Honchō korai toguchi kō ᮏ ᮅྂ᮶ᡞཱྀ⪃ (1879), Hosokawa Hiroyo ⣽ᕝᗈୡ, Nihon teikoku keisei sōran ᪥ᮏᖇᅜᙧໃ⥲ぴ(1883), and Komiyama Yasusuke ᑠᐑ 3 ᒣ⥀௓, Kinsei jinkō no hanshoku ㏆ୡேཱྀࡢⶽṪ(1889). The section in Suijinroku dealing with demographic data and for a much larger

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

77

4

number of years than the preceding authors, had an immediate impact. The best evidence of this is the article on Tokugawa population by Garrett Droppers in 1894 in Transactions of the Asiatic Society of Japan, which brought together all the known gross totals of Japanese population (and which is widely quoted even in modern studies): his paper spe5 cifically acknowledged its debt to Katsu, and discussed his figures. In 1904, Inoue Mizue ஭ୖ⍞ᯞ published a series of papers in which, while presenting no data for years for which figures were wanting (and unaware of a few data for other years which had already appeared), the author provided fuller data for 1828, for which only summary data were 6 at the time known. Honjō Eijirō ᮏᗉᰤ἞㑻, a pioneering economic historian, from 1916 onwards presented the first really close analysis of 7 population. In 1958, Sekiyama Naotarō 㛵ᒣ┤ኴ㑻provided a mag8 isterial survey of the figures; his work remains the basic reference source of data for demographic study, and it appeared to make redundant any further look at the census sources themselves. Minami Kazuo ༡࿴⏨ made an important contribution notably with identifying and dating figures for 1840, a census year for which, before his paper appeared, no 9 data appeared to survive. Takahashi Bonsen 㧗ᶫᲙ௝also dealt with population in several works. His first contribution was a general account 10 of population. His major contribution, however, has been to publish figures from sources in the han, notably Nanbu, Mito, Sendai, Tosa, for 11 which long runs of figures can be built up. Because Katsu was an official, and because he provided the most clearcut (though far from complete) run of data, it has been assumed that he obtained his figures from official data to which he enjoyed privileged access. Sekiyama saw Katsu’s data as originating in information provided 12 by former rōjū or by shogunal officials. In fact, while his own role as an official in the 1860s gave him an almost unique knowledge of and access to surviving documents of the 1860s (and to a small number of papers from Tenpō times) which he put to good effect when commissioned by the government to make a wide-ranging compilation of documents, Katsu was not particularly advantaged in terms of getting statistics. Modern work has been content to rely on the figures as assembled by Sekiyama rather than to rework the sources. Such reworking might not add to our statistical knowledge, but it should offer a partial answer as to how and in whose hands the rather meagre information had survived. The lack of ongoing review of the sources has reinforced belief in both Katsu’s primacy as an authoritative source and the prevalence of Tokugawa secrecy, with its resultant lack of circulation of information.

78

EARLY JAPANESE TRADE

Japanese population figures are superficially impressive, all the more so as they are based on returns from a large number of component administrative units. On a superficial basis they seem comparable to Scandinavian ones. National counts begin earlier, from 1721 in Japan as against ca. 1740 for Sweden and Norway. Doubts have long been expressed about their merits by Honjō Eijirō and more recently by Hayami Akira. But they are still widely used in surveys of population, are frequently quoted to reinforce a picture of relative demographic stability, and are sometimes incorporated in complex or sophisticated 13 analyses such as James White’s on ikki. The stability they suggest may in particular be spurious. While, as a statistical phenomenon, in theory stability could be an accurate mirror of reality, the probability evaporates when the trends at regional level are viewed, or when the apparent rate of growth between the last surviving count in 1846 and the first modern near-census count in 1872 hints at variable omissions from actual figures of the past. HAN POPULATION COUNTS

The instructions for compiling the Tokugawa censuses from 1721 are known in outline from surviving later copies of the original instructions. The first stage in the compilation of national returns lay in action within the han. How this was done is not perfectly clear in detail, and the instructions left much latitude to the han as to how the counting was to be conducted and what was to be counted. Where shūmon aratame chō ᐀㛛ᨵᖒ(the original purpose of which was the winkling out of Christians) existed, the modus operandi is – or seems – perfectly clear. At village level, again at least in theory, public ceremonies – held once annually – recorded population in terms of affiliation to the three entities respectively of goningumi (five-household groups), village and temple. The latter dimension of the operation provides the basis for the title of the documents, shūmon aratame chō (or SACs as for brevity in recent demographic literature in Japan they are referred to). However, not all han compiled shūmon aratame chō. In some han the exercise, it has been claimed, was not conducted annually, in others it appeared only later or else never existed in the literal sense. What existed in many han were Ninbetsu aratame chō ேูᨵᖒ (or for brevity NACs), some examples of which preceded the SACs and the purpose of which (even if the term shūmon sometimes recurred in their styling) was not religious.

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

79

One copy of the outcome of the exercise, whether shūmon-type or otherwise, was retained by local officials at the village level, one forwarded to the han authorities. Of the second stage – the process of aggregating data from the registers at han level by officials in the castle town – we know less than for the first (or village) stage. Although the local registers – of which copies were sent to the han – were detailed and somewhat complex in structure, in theory the work in compiling a han-wide total for population should not, assuming that satisfactory returns were received for all villages, be a time-consuming exercise. Each register as returned to the han authorities ended with gross totals of population for the village. Nevertheless, doubts remain about how the han conducted its count. While surviving village registers on the internal evidence of successive documents in runs of records for individual villages can be seen as conscientious compilations, it has still to be proved that surviving series (emanating from village headmen and often surviving in the hands of their descendants) are representative, even within a han of all the returns from the villages. In other words, did – or could – deficiencies exist in the returns? And if deficiencies existed in the primary returns from villages, how were the gaps made up by han officials? Did they simply repeat a preceding year’s figures? Did they use other sources, and even more importantly, as shūmon aratame chō or SACs were not universal, how were numbers counted, as we know they were, in han where SACs did not exist? For those han which supposedly held a SAC exercise but not every year, how were the counts which in some cases appear to have existed for intervening years compiled? SACs appeared first in the tenryō ኳ㡿and then spread to other han. The Shimabara revolt in 1637–38 has been seen also as the factor 14 behind an impetus to their spread. They were therefore already fairly numerous in the central reaches of Japan by 1665 when the shogunate required them to be compiled nationwide (from 1671 annual compilation would be ordered). However, at this stage the requirement entailed no reporting to the shogunate, nor were any enforcement procedures prescribed; indeed for the fudai and tozama requirements for stringent application of specific rules were arguably not constitutionally possible. Han enjoyed wide latitude as to how they responded to the directive. Even in studies which postulate a widespread existence of shūmon aratame chō, an exception is conceded for the Tōhoku 15 and for Kumamoto. Population counts did appear to become universal. (Nor were they a Tokugawa innovation. There were of course preSAC models for them.) That such counts were widespread is illustrated

80

EARLY JAPANESE TRADE

in the response to Yoshimune’s request in 1734 to nine han for earlier population data: two han provided data as far back as 1665, one for 1669, and the remaining seven with one exception had a first count 16 before 1700. With population counts becoming common or frequent (even outside the cases where SACs either existed or were maintained), a wide range of response became possible. Modern study is bedevilled by confusion in the terminology, which is not elastic enough to cover a varied range of response, doubly so as the word shūmon was at the time often incorporated into the styling of exercises which did not have a religious objective for their operation. It is important to stress this, as it has been said that SACs were in principle universal: “Still, with the exception of clans like the Nihonmatsu clan, the population registers (NACs) mentioned earlier were compiled very sporadically and by very few domains, but the SACs were, in principle compiled annually 17 nation-wide.” Repeating these views, Yamamura and Hanley noted that the exercise “was carried out on a nationwide basis, theoretically 18 every year.” In Takahashi Bonsen’s formulation, the first returns were SACs but the exercise widened later in the century to serve two purposes. On this argument they were near-universal, but simply deepened in their coverage of information with the passage of time. Hayami likewise suggested that they had a dual purpose, being compiled not simply to stifle Christianity but to provide information about popu19 lation and households. In a much-quoted article with Cornell, he stated that “after 1670 the information in population registers became 20 more detailed.” A formulation later in the same article that “a number 21 of lords also took what amounts to a census (ninbetsu)” could imply, probably unintentionally, that two types of register had existed simultaneously. The assumption made by Sekiyama and others that two types of register blended is a rationalization of a situation, in which radically different types of register existed in Japan, if not necessarily at one time 22 in a given place. It also can provide a basis for an assumption that two registers were kept simultaneously within a single han. There is no evidence of this, and there is the added challenge that such a transition would have to be dated. The relatively small number of surviving village registers and the very large gaps in their sequence make it possible to assume a complex pattern over the years, something added to by a varied terminology, shūmon aratame chō, ninbetsu aratame chō 23 and shūmon ninbetsu aratame chō, styles which often do not quite correspond to the nature of the contents. A recent study by Takagi of the Sendai sources has argued that too much attention should not be

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

81

attached to the incidental use of language, and that for Sendai at least the underlining motivation was the han urge to know more about 24 its own population. Hence, on this argument, while their title may vary somewhat, the documents, were essentially NACs. This seems to be supported by the evidence of Mito, where the 18th century registers are clearly not shūmon aratame chō in the more radical sense: whatever about earlier SACs, the eighteenth-century counts there seem to have been compiled as population counts. The prospect of overlapping sets of SACs and NACs, or a conscious pursuit of two different objectives simultaneously, is somewhat unreal. While the original SACs had a distinctive purpose, they were so broad in their coverage that, where they existed, they made a supplementary exercise unnecessary. The shūmon aratame chō in extreme cases involved the ceremony of fumie and in all cases noted the population with the temple seal appended to the names of recorded inhabitants. The ninbetsu aratame chō even if in most cases their origins coincide with or follow closely on the orders of 1665 and 1671 (that is if we choose to ignore earlier precedents of population counting) were not an exercise in recording by formal ceremony inhabitants on a temple basis but a form of koseki or registration in some form or other of the population and which, as it could be used as a basis for collection of the rice levy, established itself as the basic form of record. This type of document was easier to compile and revise at local level than the complex and ambitious SAC exercise. Of course devoid of the public scrutiny (with or without fumie) which was the hallmark of the shūmon aratame chō, they were capable of becoming fossilized as statistical sources tradi25 tionally have become in all societies. That population registers existed universally is implied in the order of May 1870 to make a count of the number of individuals and households, “using existing population 26 registers.” In some cases there is little evidence that shūmon-type registers (i.e., registers enumerating the population on a religious basis) 27 were ever compiled. This is the case for Tosa, and Satsuma. However, they did compile registers of population. In another instance, Chōshū, 28 known examples of the registers exist only for the nineteenth century (which implies an earlier existence, even if now undocumented). Given the latitude enjoyed by han, and lack of central supervision of the process by the shogunate, registers of population, either existing ones or new ones lacking the features of a shūmon aratame chō exercise in the complex and literal sense of the early SACs were widespread. As levies on rice output were the basis of the income of han authorities and

82

EARLY JAPANESE TRADE

of their samurai, registers of inhabitants (at least of honbyakushō) were essential for the guidance of han officials. In a sense if some form of register did not already exist, registers would have had to be invented. They could easily be used for a population count, where shūmon aratame chō were not compiled or even to replace them (hence the importance of precision about what records existed to eliminate any suggestion that two types of register existed simultaneously). The regular compilation of shūmon aratame chō themselves is said to have been onerous enough for some han like Mito and Kii, significantly both collateral Tokugawa 29 houses, to compile them only every sixth year. Through detailed information in some form of registers at the disposal of the han daikan (the officials responsible for supervising collection of levy on rice output), counts of population could easily be made. It was all the easier because with fief holders, who effectively managed their own land either already few in the fudai han or becoming so, daikan increasingly took over direct superintendence of the entire rice output within han. In han such as Mito where officials calculated population in the intervals between formal han counts, while the title of documents remained unchanged, the exercise itself change its nature radically. As revenue was a central preoccupation, if not for the original exercises in counting or registering population, then for their continuation, a distinction often came to be drawn within them between the rural population (subject to levy) and the rest of the population (who were not liable for the levy). In Mito, for instance, almost all the surviving 30 available data are for the rural population alone. Such coverage shows by definition that shūmon aratame chō, if they had existed in the literal 31 sense there, had ceased to be the basis of counting. As revenue was a key objective in keeping registers up to date, a distinction between the commoner population existing on fiefs and that on directly retained land was also important. As a result, in a pattern which varied between han, aggregated figures of population in many han did not include the population resident on fief-holdings, or later came to exclude them, or for internal purposes han administrations even compiled two sets of figures, one comprehensive, the other confined to the population on the directly-held lands of the han, In Nanbu han (Morioka), as we will see below, confusion caused by the distinction in the records between figures including fief holdings and figures excluding them has become the basis for a modern assumption that two distinct types of counts of population existed simultaneously. In Sendai han, the basic regular count in han records is of gunkata (rural population). While this defini-

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

83

tion appears to exclude town population, much more importantly the returns omitted the very large population residing in the fiefs. In Mito the decline in rural population between 1730 and 1770, almost unparalleled in its scale as a decline spread over several decades of the mid-eighteenth century, is probably a consequence of some shift in the treatment of fiefs and hence in counting commoner population in fiefs. As the trend was away from fief holding, it is more likely to reflect a decision to exclude from han totals existing lands held by fiefholding samurai or collateral branches of the daimyo family than an increase in fief-holding occupiers. Given the acute pressure on the straitened income of daimyo administrations, it is difficult to visualize a voluntary decision which would have the effect of worsening the fiscal plight of han by increasing fief holds. At the same time the scale of fief holding helps to explain why in many large han and chronically, the “fiscal” status of daimyo administration was so precarious. As the basis of record keeping, whatever its original purpose, was to learn more about the population of the han, an urge to acquire further information about wealth was easily answered. Data about livestock and horses for instance were noted in many of the han aggregates which survive. Of the village of Kōriyama in Nihonmatsu han with a uniquely complete run of NACs from 1685 to 1871, in Hayami’s words, “it could be said that they are the most outstanding survey documents for 32 that time.” An offsetting drawback to this easy method of population counting, was that it was all too easy simply to take the registers literally as a measure of household population: in other words, families were assumed to exist – or to survive – because the names of householders were already on the record – and survived – on registers. The collection of the levy on rice output was an ongoing annual activity. The aggregation of data for population counts however required some further clerical work, and it was easy either to take a shortcut or simply to follow the path of inertia. The consequence was two-fold. The first was that estimation of household size – the number of members within each household – if attempted, was often not an actual count but an 33 arbitrarily or notionally derived one. Secondly, the counts themselves could become static, with the registers adjusted only marginally, ignoring in the process both long term trends and short term upheavals. The methodology would also explain why for major famines, the data are at times to a surprising degree insensitive. Their quality for these reasons could also vary between han. For Nanbu fossilization is clearcut. For Sendai on the other hand, the recording of population in the

84

EARLY JAPANESE TRADE

registers was more sensitive. In particular it fell sharply in the crisis decades, thereafter recovering slowly. It did not however fall dramatically in the 1750s which suggests, if the quality of the han records is superior as seems to be the case, that the scale of the Tohoku famine of 1755 may be exaggerated. Counting of population even if rather notional, as in the case of Nanbu could also provide the vehicle of calculating aggregate figures of deaths and births for each year and of movement into and out of 34 the han. In the short term, the Nanbu figures fluctuated from year to year, suggesting that they were certainly not the rule-of-thumb counts that the gross population totals, visibly independent of them, all too often because. They reveal from year to year a somewhat indeterminate pattern of adjustment. They were however insensitive to demographic disasters, as net changes within the year are small. While Takahashi saw the Nanbu figures as reflecting for instance the impact of the famine 35 of 1783, that crisis is not in any meaningful sense visible in them. For what they are worth, the totals of births and deaths in the 1780s do not point to any real crisis, even if unlike the population totals they show fluctuations – very modest ones – from year to year. For the very long term higher totals of births and deaths appeared in the few Nanbu counts surviving for the nineteenth century, which might suggest some improvement in the counts. How the figures were calculated is a matter of conjecture. Whether the exercise was good or bad, the real point of these observations is that some form of returns must have been made regularly from the villages of raw demographic data of births, deaths and movement, which were then aggregated in the han capital. Both the gross population data for Nanbu, because of its notional basis of little value as an indicator of short term change, and the totals of births and deaths (with the unanswered questions they raise) imply a poor sense, if any, of statistical realities. They suggest administrative functions executed mechanically, something entirely consistent with the remarkable failure of statistical counts to enter into Tokugawa discourse on social 36 and economic problems. Within the various han, the copies of returns from the villages sent to the han authorities do not survive at han level. At that level they were got rid of, sooner or later. The only exception is Suwa han in Shinano kuni or province, where the registers were kept till 1868, remark37 ably some of them then surviving subsequent vicissitudes. In lands under the control of the shogunate, things may have been better. At any rate for the huge and competent Nagasaki shogunal administra-

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

85

tion, some registers survive. With these exceptions, what data survived in central administrations did so only by contemporary abstraction of the data (from returns which sooner or later were got rid of ) and by their recording annually as isolated entries in various han journals. Where records survived in the families of headmen, even if the individual runs are sometimes impressive, the total number of registers is a mere fraction of what was compiled. In Sendai han for instance surviving NACs come from a mere three of the han’s twenty-one gun or districts or, illustrating more strikingly their narrow range, from ten of 38 the han’s 970 villages. Population figures were in effect simply a by-product of practical administrative purpose. As a result the coverage not only varied from han to han, but could vary over time within a han. As cultivating occupiers were growing in numbers in the seventeenth century, the relatively small number of counts for that century in individual han, in essence a mere listing of existing and new rice-levy paying units, recorded a rise in population. In the eighteenth century as registered household numbers ceased to increase, the registers themselves inevitably became more or less static. To take the example of Nanbu, once the seventeenth- and early eighteenth-century growth of population had ceased, Nanbu returns were static with only minor adjustments: from 1752 they entered a plateau, which became even more rigid from 1762. From that date, significant subtraction from or addition to the household total for the han scarcely arose. Of course even data derived on such a defective basis does have a value, reflecting the seventeenthcentury rise in population, and in the eighteenth century, anchored at least loosely to reality. The figures present other difficulties. The kuni as historical provinces had invariable boundaries, and this was in essence the reason for the purpose of creating the shogunal grand total, of aggregating han counts into figures for kuni. Han were either smaller than kuni, or consisted of territories within several kuni. Their boundaries were very rarely coterminous with those of the kuni. A further potential source of confusion, not to contemporaries but to modern students, is that powerful families sometimes ruled over several kuni, the case notably for the rulers of Satsuma, Chōshū and Kaga. Throughout the period returns by the Shimazu family for their historical or hereditary base – the kuni (and han) of Satsuma – are small, not taking account of other territories ruled by the family. Satsuma as returned in the 1721 census was a mere 149,039 persons, dramatically smaller than returns in 1734

86

EARLY JAPANESE TRADE

to the shogun for 1698 and 1732 (and which appear to include all the territory they administered). To take the opposite pattern, the territory administered by the daimyo of Mito han was only part of a much larger entity or kuni of Hitachi (twice the population of Mito). However, at least the profile for Mito han and Hitachi kuni is similar: a sharp fall a few decades either side of 1750, and an emerging stability, little ruffled 39 statistically at least by the 1780s. A lesser problem, though more for the seventeenth century than for the eighteenth, was that han boundaries could change: some han disappeared or were even created afresh; some han were or became mere creations of widely scattered territories or even of territories intermingled intimately with those of another han. These posed practical problems for the compilers, and could especially if processing was perfunctory have added to the arbitrary nature of the operation, whether in the primary return from the han or in later aggregation. We know nothing about the actual process whereby han figures were converted by the shogunate to kuni equivalents. It is unfortunately impossible as a general rule to compare data for han with the data for kuni. However within han records internal divides seem to have existed for the various districts within the han. A guess has to be that returns to the shogunate contained some form of breakdown that, as the need arose, guided shogunal officials in distributing population between different kuni. The most serious problem of all in the elusive attempt to compare han and kuni data is that all the han of the north-east were lumped together by the shogunate in a single grouping (taking an older and territorially larger definition of the kuni of Mutsu). That accounts also for the the variations in modern documents in the number of kuni. They have been given variously as 66 to 71, but should more strictly be counted as either 69 or 74 if two aggregations of two and five kuni respectively 40 are disaggregated into the component kuni. Equally difficult is the question as to whether returns for the han and their subsequent conversion into kuni data were on a consistent basis. There is no documentary evidence to establish how they were adjusted by shogunal officials or whether the shogunal officials had any concern on this score. In fact, the very absence of any circulars issued over the period 1726–1864 regarding revisions of approach suggests that officials acted in a passive manner by processing figures in the form in which they were submitted. We can establish with confidence that Satsuma data were not adjusted to take into account the population account excluded in the han data (see Section 4 below). In the case of

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

87

Bizen no kuni (Okayama) and Tosa, in both of which the han and kuni were almost identical in area, shogunal data (and hence the figures previously submitted by the han) seem to have omitted the population of 41 the towns, itself about 10 percent of the han gross total. In the case of Mito the fact that shogunal census data for Hitachi followed the same profile as data from han sources for Mito hints in a crude way that shogunal figures in compiling the figures for the kuni of Hitachi automatically followed adjustments in the han data. For shogunal officials the han population figures may not have even had much practical significance. The kokudaka also was returned in every census. Whatever the limitations of kokudaka returns (which understated real output), they more certainly embraced the entire han (and in contrast to the population data the conversions onto a kuni basis were more consistent). They, not the population figures, were already long before census figures began to emerge, the well-established basis for appraising the wealth or status of daimyo. HAN COUNTS: METHODOLOGY

Population totals could be compiled in several ways. First, they could be based on shūmon aratame chō returns (SACs), at least in theory universally made across the territory of many han, and carefully aggregated by han authorities. Secondly, they could derive from counts, which were or became in essence fiscal, of households and perhaps also (certainly the case for Sendai). of a count, more or less realistic, of the number of members within the household units, Thirdly they could consist of a mere count of the number of households (without an actual count of household members) with minor adjustments to the number of households from year to year or from census to census: the household count was converted into a population total by an arbitrary multiplier. The exercise could on occasion deteriorate even into identical returns from year to year. This procedure would also explain why in some instances, as has been noted by Hanley and Yamamura for Nanbu, the sex ratio 42 remained invariably within a range of 1.1 to 1.2 for 160 years. The really serious problem in these exercises is that certain categories were omitted. These problems have been well set out in numerous accounts from the 1920s onwards. In English there is a very good account from the 1930s by Honjō Eijirō, one of the pioneers of modern study of demographic sources: in more recent times Hanley and 43 Yamamura summarize the problems succinctly but clearly. Most

88

EARLY JAPANESE TRADE

famously the shogunal census omitted samurai, and in the case of commoner population also variably omitted the young at ages from below one or two up to below eight years (or even fifteen years). There were other exclusions in addition to these obvious omissions. As revenue was a prime purpose of the counts, the internal treatment of population was influenced by the consideration whether they provided revenue or not. In some cases the listing of rural population was confined to honbyakushō: craftsmen or the transient inhabitants of rural society were ignored. A more substantial omission was that of population resident in temple and shrine domains and in towns. The numbers in these categories were however not necessarily unknown: the categories were often counted or estimated. The real problem is that practice varied as to whether they were carried forward into a grand total for the han. Again this problem would be easier if there was consistency between han or within the individual han over time. In quantitative terms a still serious problem, and, one that is understressed, or not recognized sufficiently in the literature about census omissions, is that where samurai fiefs existed, commoner population on fiefs was in many cases omitted. In other words, in some cases, the counts were confined to the population resident on directly retained lands – kurairichi ⶶධᆅ – and whose revenues went into the coffers of the daimyo administration, and the population of fiefs – kyūchi ⤥ᆅ, or chigyōchi (also described as kyūsho ⤥ᡤ) – which would not provide an income for the han were omitted. This problem has been glossed over because it has been seen as a minor appendage to the exclusion of samurai which also comprehended their servants. However this is a misleading formulation, because far more was involved than servants. For stipendiary samurai (i.e., non-fief-holding samurai), especially those on lower incomes, resident in the castle town, it was certainly true that dependants would have been mere household servants and hence the number per household small. However, for fief holders, what is involved is not primarily household servants but the peasants who worked their own fields, and paid a rice levy to their samurai fief holder. While samurai were few in the many small fudai han, the issue had larger implications for the tozama han where samurai were numerous, and some of whom also had large fiefs. The implications were all the greater because collateral branches of the daimyo family were or could be counted among the fief population for the same reason that they provided no revenue to the han. The problem is made more awkward by the fact that internal bookkeeping practice in han sometimes

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

89

may have fluctuated in including collateral branches with the kura or han income of the han, or in excluding them as fiefs. In other words the overall problem may be not only the existence of a further category of exclusion, but a variety in population totals between han, or within a han variation over time. The exclusion of urban population is relatively simple, though in internal counts itself lacking in clarity and in consistency. In the case of Nanbu in its retrospective return to Yoshimune in 1734, the population figures were 245,635 for 1669, 306,142 for 1703, and 322,109 44 for 1732. While the first two figures are identical – or almost identical – with the data in the Nanbu han nisshi, the return for 1732, a shogunal census year, is lower than the figure of 345,825 in han records of that year, plausibly by not giving figures for the non-rural population. The fact that the very few surviving general counts for the early nineteenth century are below the figure for the han gross total, suggests that towns were still omitted. As we have no copies of returns actually made to the shogunate (and han figures are concealed in data for the huge kuni of Mutsu), this can not be verified on any regular basis. However the profile in the Mutsu figures suggests that, excluding towns or not, the returns were made on a consistent basis. The trend in the Tosa data (excluding town population, drawn from the now lost nisshi) is close 45 to the returns as made to the shogunate. This may be the case for Mito too, as almost all the surviving counts within Mito exclude town 46 population. For Sendai the figures for gunkata by definition appear to exclude towns. These comments bear on omissions of urban population. Urban population was however a small percentage of total population. On the other hand commoner population on fiefs could be large. Fief population was excluded in Satsuma, and that exclusion was reflected in turn in the data assembled by the shogunate. Within Sendai, also, the exclusion of fief populations must have remained in the final version returned to the shogunate. The sharp rise in the population of the five kuni denominated under the single heading of “Mutsu” in 1846, to 2,294,915 in 1872 from 1,607,881 in 1846, suggests that all or most of the han that comprised this territory must have excluded fiefresident commoners from their returns to the shogunate. The inclusion in 1872 of Sendai’s former large fief-resident population of 200,000 would account for a substantial part of the rise. Other han in Mutsu, equally with large areas in fiefhold, would also have contributed to the rise. Nanbu (whose kuni, Rikuchū, was comprehended in the Mutsu

90

EARLY JAPANESE TRADE

total) likewise must have excluded its fief-resident population from its returns. There are of course dilemmas in the internal records of Nanbu, which regulary contain estimates for total population. For the few surviving internal returns for the early nineteenth century, the figures seem full ones, net of an amount which seems to equal the population of the towns. In other words, totals existed which included fief-resident population. That did not mean, however, that returns of gross population were made to Edo. As will be seen in section 4 below, the existence of a variety of han totals in the internal records of Nanbu did not mean, as has been assumed, that returns to Edo involved a deliberate concealment of part of the population of the han. Nanbu’s data are unusual only in the degree of detail which has survived. Returns such as Nanbu’s for 1732, which was close to the han gross figure, or the ups and downs of Suruga and Kazusa totals – unparalleled in other kuni – in the surviving twelve full shogunal census returns, may hint at abnormalities or administrative uncertainties. But the evidence is that gross returns were the exception rather than the norm. Just across the frontier from Mutsu, in Mito, the large fall in numbers over the mid-eighteenth century suggests that action was taken to exclude a section of the population from han totals and hence automatically from kuni figures (in this instance probably through a shift of territories to collateral branches – in effect a process similar in its effects to enlarging the lands of fiefholders – of the ruling family). If the exclusion of fief-resident population in many cases (such as Satsuma, Mito, Sendai and Nanbu) is statistically significant, it was much less so for most fudai han and for the tenryō. In their case the number of samurai was usually modest, and they were also mainly stipendiary and castle-town resident (and hence the only dependants of samurai families were household servants). For han statisticians the difficulty was posed by the instruction to exclude samurai and dependants from the six-yearly returns to the shogunate. Fief-resident commoners, holding no direct link with officials acting on behalf of the han, were as much dependants as the house servants of samurai were. While no revenue accrued to the central han administration out of agricultural production on fiefs, han administrations were nonetheless interested in household counts, partly because of the need to know about all their population and on occasion partly to measure the benefits that might accrue if they dared to degrade the status of enfeoffed samurai to that of mere stipendiary samurai. In a sense the census instructions which enjoined han to exclude samurai and dependants from returns to the shogunate involved the necessity of making

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

91

a distinction between figures which served some internal purpose and figures for shogunal returns. NANBU AND MITO HAN

Nanbu and Mito han merit some special attention. In Nanbu’s case it is because its records are both extensive and contain many subdivisions within the data. In Mito’s case the question as to whether its population fall in the mid-eighteenth century was real or not is important not only for study of demographic sources but for comprehension of Mito’s own economic history. The population figures for Nanbu (Morioka) han, are available for 1653 and some decades later in relatively uninterrupted fashion from 1680 up to 1803; thereafter only for a mere four years up to 1840. It has even been suggested that two series of documents, han nisshi 47 and shūmon aratame chō were simultaneously maintained. The situation is confused by the fact that the term kirishitan shūmon aratame chō (Christian religious registers), while applied primarily to the earlier documents, is at times loosely applied in the abbreviated form simply of 48 shūmon aratame chō (religious registers) to later documents. These are, however, more properly described as ninbetsu aratame chō (registers of individuals); from ca. 1680, driven by a broader administrative curios49 ity, some han collected a wider range of information in these. Population totals for Nanbu han survive in han nisshi (han journals), also called zassho (records of miscellaneous subjects). The zassho in all comprise 192 50 satsu, or volumes, for the years between 1644 to 1841. The very loose use of language in modern writing can mislead us into assuming that there were simultaneously two sets of records. A table of population for 1669–1790 constructed by Mori Kahee ᳃჆ර⾨ in a 1934 article in Shakai-keizai shigaku, which appears to give lower-bound figures for population for the years 1756–1790, identifies his source simply as 51 shūmon aratame chō for the han. His figures for the years 1756–1790 contrast with higher-bound population figures in Takahashi’s later study of Nanbu population, said to be drawn from the so-called han nisshi 52 ⸬᪥ㄅ. Susan Hanley and Kozo Yamamura, in wrestling with this problem – the origins of which lay in the very terse identifications of the sources in the writings of Mori and Takahashi, and also in the Iwateken shi – based their arguments on the existence of two sets of figures, one drawing on the nisshi (or zassho 㞧᭩, which Takahashi cited as his source), and one on shūmon-aratame chō (which Mori’s 1934 article

92

EARLY JAPANESE TRADE

appeared to suggest as his authority). For the lower figures from the supposed shūmon aratame chō, a purpose was proposed: Most domains based their regular population report to the Bakufu on the shūmon aratame records and officials could well have doctored these records at the domain level in order to have the Bakufu believe that the domain was in straits and thus unable to bear the burden of additional Bakufu levies.53

Elsewhere it is repeated that “one set was for Bakufu eyes, others for the 54 domain’s own internal use.” The consequences of such an assumption are not small: Thus, according to the figures compiled from the religious investigation records, the total population of the domain dropped by 33 percent during the half century from 1740 to 1790, in contrast to a fall of only 2.8 55 percent in the set of figures in the Han nisshi.

In actuality, there are neither two locations for the records nor two categories of sources. What Mori in his 1934 article actually stated was that the figures were drawn from a work by Nitobe Sengaku ᪂Ώᡞ௝ ᓅ, Kyū Nanbu han shūmon aratame chō ᪧ༡㒊⸬᐀㛛ᨵᖒ, edited from sources for the history of Nanbu population and contained in a writing or writings in the archives of the former ruling family of 56 Nanbu han. The location of the surviving han records was identified more precisely by Takahashi as Kawai-mura in Shimohei-gun (presumably the residence of the daimyo’s descendants); he added that 57 the records were now held in the Morioka-shi Sangyō Bunka Kan. The Iwate-ken shi, dealing with population at some length, identifies its sources simply as the nisshi, plus the works of Mori, Takahashi, and 58 Nitobe. For the years 1756–1790 its main population table, drawing on the han records at large, presented two sets of data, one identical to the lower-bound population figures used by Mori, and one 59 identical to the higher-bound figures employed by Takahashi. While lower-bound figures for population are lacking in Takahashi’s tables, this is simply because his intention was to present a total population of the han. Published data on the number of households added unnecessarily to the obscurity of the situation: except for the year 1776, lower- and higher-bound figures for households appeared in the Iwate-ken shi only from 1781, when they make an appearance for a sequence of nine years.

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

93

From 1781 to 1789 in addiction to the ongoing two sets of figures for population, there are both lower and higher bound household counts (in 1781 63,228 and 56,185 households with the lower-bound figure falling further from 56,082 households in 1783 to 45,527 in 1784). For 1790 a higher-bound figure alone appears for households (i.e., a lower-bound figure ceased to exist after 1789). On the other hand, for the year 1790 – for the last time – both lower-and higher-bound figures were given for population as opposed to households. The precise distinction conveyed by higher – and lower-bound – figures of population or of households, even in the case of the large change in 1784 is not set out anywhere explicitly in writing. From 1756, one series of han calculations (measured in the existence of upperbound and lower-bound figures) excluded what would seem the commoner population of many samurai fiefs, and more briefly and more radically from 1784 of all fiefs. In all probability the dilemma must have been whether collateral branches were to be viewed as part of the ruling family, or, given their independent income which made no contribution to the receipts of the ruling house, simply as very large fiefholds. In a crude sense the population deductions excluded from the grand total of han population were some 54,000 commoners resident on fiefs in 1756 and an additional 61,000 in 1784. In terms of household counts, the exclusion in 1756 was approximately 7,000 households, a figure which rose to 18,000 households in 1784. A multiplier of five to seven members per household would give a crude total of 90,000– 126,000 commoners (a total not far removed from the figure created by combining the 54,000 and 61,000 fief-residents excluded in two stages in 1756 and 1784 respectively). The bookkeeping calculations in 1784–89 were not made for idle or arbitrary reasons. The acute concern about the inadequacy of han income led, as detailed han calculations for 1790 suggest (see next paragraph), to a transfer in the 1780s of income from fiefholds to the kura or han administration. Going back to the emergence of two runs of figures for population, in 1755 there had been a sole figure of 358,222; in 1756 alternative figures of 356,005 and 301,686 appeared. A fall in lower-bound population from 306,077 in 1783 to 245,963 in 1784 corresponded to a matching reduction in 1784–89 in the household count. In other words the lower -bound population figure did not at any stage purport to represent the total population of the han, and came from a distinction between total households within the han, and a total net of those resident on fiefholds which provided no income to the han authori-

94

EARLY JAPANESE TRADE

ties. The matter is made confusing by the fact that, while two sets of population figures exist, the corresponding lower- bound household figure is not given for early years, but is available in a regular fashion only for the years 1781–89. However, even if not made regularly, the distinction between kura – or daimyo – income and total income (which combined kura and kyūsho – or fief – income) existed before 1781, and even well before 1756. Detailed estimates of rural population for some years specifically distinguish fief and kura populations. Thus for 1683 in the case of a rural population of 247,053, the kura 60 figure was 154,878 and the kyūsho population 92,175. In 1712, in a total population of 350,596, the kura count was 178,138 and the kyūsho count was 114,876. In 1752, in a rural population of 286,877 (and total population of 353,725), the respective figures for kura and 61 kyūsho were 165,089 and 109,183. For 1790, another calculation was made, no doubt to measure, inter alia, the outcome of a large-scale operation at the end of the 1780s in which many samurai were demoted from fiefhold status to mere stipendiary rank. The commoner population on kura lands showed a dramatic rise from 1752, to 217,493; on fiefhold lands it had fallen to 76,843. In other words, a considerable 62 revenue-increasing operation had taken place. Of itself the operation did not generate revenue, as the income received from the peasants in the first instance merely funded the stipends that had to be paid out of han central funds to the former fief holders. Han income rose only to the extent that the number of samurai subject to deductions from stipends (a theoretically temporary practice that progressively became wider and quasi-permanent as the eighteenth century wore on) was now greatly increased. The apparent lack of further estimates suggests that no similar major operation occurred in later years. From 1790, a higher-bound household figure alone is given, and from 1791, a sole population figure appeared (although for the year 1790 itself, despite the absence of a surviving lower-bond household total, lower and higher-bound data for population both appeared). The practice of adhering to a higher-bound figure only seems to be departed from in a reduced population figure of 326,262 for 1807 and 1816 (identical in both years!), a figure which implies an omission of a section of the population. That the reduced figure was itself not intended to represent the whole population can be seen in the fact that the household count itself continued to be given in the 63,000 range. The omissions in 1816, it was speculated in the Iwate-ken shi, included 63 both town population and the temple/shrine population. However,

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

95

this is misleading; the shortfall very nearly matches the town population (ca. 30,000). Had it included other categories, the residual population should be still lower. In the only later years for which data have survived, the full population is 351,332 (in 1838), 357,207 (1839), 64 and 356,269 (1840). Higher-bound figures, whether of households or of population, where both higher-bound and lower-bound data exist in the nisshi, are invariably the intended basic figure for the total population of the han. The statistical weaknesses of both upper-bound and lowerbound figures, it hardly need be said, are identical, inevitably so as the mechanics of compiling the aggregates, for both lower-bound and higher-bound returns, were common to both sets of figures. In Nanbu the annual grand total for han population was to a remarkable degree stagnant. From 1762 it scarcely changed, adjustment from year to year being merely notional. Thus the figure of 358,857 in 1764 differs but slightly from 357,810 recorded in 1803. In some years even the small notional adjustments made from year to year, were not executed: there were identical figures for households in 1755 and 1756, and again in 1772 to 1777. For the years 1762–1840 the figures for households, the basis for calculating the total population, were (except in the few years in which both higher-bound and lower-bound counts exist), invariably 65 some 63,000 with the addition at most of some odd hundreds. The Iwate-ken shi noted in passing that the han returns were probably not realistic figures, though the reason advanced is that the population level was inconsistent with the high level of mortality for some 66 years reported by contemporary officials. Takahashi in his study simply noted that the population figures related primarily to the settled per67 manent population or honbyakushō. Neither of these explanations is convincing or sufficient. As Hanley and Yamamura noted in relation to the comments in the Iwate-ken shi, contemporary reports of famine were exaggerated and could be countered with other and more sober 68 recitals. As for Takahashi’s observation, even if the figures are in reality confined to the settled population, the stability from year to year is totally unrealistic. The han authorities routinely took the number of houses as more or less fixed, and adjustments for actual population – conducted on a basis of which we know nothing – produced little change in population from one year to the next. Crisis years characterized by significant mortality (even if the alarmist figures of some officials are to be discounted) do not show up. The change in the 1780s – though coincidentally occurring in 1784 in the wake of the 1783 famine – is

96

EARLY JAPANESE TRADE

simply one of definition; this explains why, whether in lower-bound or higher-bound totals, the total remained unchanged immediately after 69 the devastating 1786–87 famine. In the case of Mito something similar is implied by the fact that the slide in population returns from the 1730s to the 1780s had ended by 1783. As a consequence the difficult decade of the 1780s is poorly reflected in the Mito figures also. Mito han, like Nanbu han, is an apposite case for the attempt to determine whether coverage of han figures was comprehensive or not. Whereas the population count from 1762 for Nanbu han was remarkably static in trends, the count for Mito han at this time was moving downwards. The population fell from 309,711 70 in Kyōhō 17 (1732) to 250,807 in An’ei 3 (1774). This might arguably suggest a more conscientious recording of population, and hence could be construed as evidence both reflecting – and confirming – opinion about crisis in Mito. However, the matter is not that simple. Counts in Mito were not confined to the one year in six and to the shūmon aratame 71 chō basis suggested by Hayami. There are, apart from earlier counts in 1697, 1702, and 1703, four surviving counts between 1729 and 1794 72 in years which were not shogunal census years. This may suggest that numbers were counted in simple manipulations of data in many and perhaps in all years rather than in more ambitious – and traditional – collecting operations. There is some confirmation of this in the relative stability of numbers between 1703 and 1732 (305,649 in 1703 and, with some limited variation in the interval, 309,711 in 1732), and in the still more pronounced stability from 1783 to 1828. A downward movement between 1732 and 1774 (or 1783) might at first sight seem evidence of a real trend. However, the figures in population data for Mito are of rural and fishing population only. This strongly suggests that they did not originate in SAC exercises, which would have sought to be more inclusive. If the count is based on tax registers, the probability of changes of definition or of omission of kyūsho cannot be excluded. In theory the movement could reflect a contraction in the number of honbyakushō on han registers: in Motoki village, for example, the number of honbyakushō fell from 103 in 1696 to 85 in 1772 and to 76 73 in 1813. But there are also instances in Mito where the pattern is the reverse, and in any event changes either way of this sort would probably have been marginal to the total population. The existence of a substantial fief population would emphasis that a large segment of han output was beyond the control of the han administration, and would help to account for the penury which was a persistent complaint of the han

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

97

74

authorities and its direct retainers. Han population as noted above was increasingly stable from 1783 onwards (or equally from 1786, a shogunal census year). Significantly, despite famine in 1783 and 1786, the figures changed little either in the 1780s or at the outset of the 1790s. The parallel with Nanbu in the same years is arresting. In a crude sense, even after famine the counts reflect not mortality, but merely the fact that “taxable” households (either survivors or notional fiscal entries) remained on the books. Of course, by definition honbyakushō were better off, and hence less vulnerable to being wiped out, than marginal families of laborers and tradesmen. The suggestion of rigidity emerging in counts is very striking for the four surviving ones from 1810 to 1828; the total varied between a minimum of 227,170 and a maximum of 75 227,732. LONG TERM TRENDS IN COVERAGE IN HAN RETURNS

Nanbu, long characterized on statistical evidence by stable population, and Mito, with stability of population a feature from the 1780s, might be seen perhaps as exceptions. However, other han have figures showing stability or moderate increase. Tosa and Aizu are both cases in point. In the case of Tosa, figures survive for 1681–1798, 1834, 1841, 1842, 76 1854 and 1855. The Tosa figures are upwards with no real dip. In Aizu han the figures surviving for 120 years within the period 1648–1805 77 remain relatively stable. The figures for Sendai han offer an interesting contrast. Sendai was a well-administered han in which a range of figures survives in numerous han documents. Surviving ninbetsu aratame returns are fewer and more isolated: for instance for the post1721 period, when national census taking began, they amount to a mere 32. The regularly collected figures in the central han records are for gunkata 㒆 ᪉ (rural population), and they exclude figures for fiefs. More detailed calculations for a few isolated years reveal that the population resident on fiefs was about 26 percent (202,000) of the estimated population in 1695 of 770,000, and 22 to 23 percent 78 of a population of 820,000 in 1742. Though samurai numbers can not be readily quantified, the large number of retainers suggests that a high proportion resided in rather large fiefs (some of them certainly collateral branches of the ruling family rather than samurai, however exalted). Gunkata figures – the only systematically surviving figures and thus the basis of han population returns – while not figures of overall population, represent demographic trends meaningfully.

98

EARLY JAPANESE TRADE

The long-term trend emerging from the gunkata returns is one of stability or of a slight fall in population. However, in terms of shortterm fluctuations, two things emerge. First, in contrast to data for other han, population figures were not stable. They fell in difficult periods, and then revived in subsequent years. The 1755 crisis is only modestly reflected in the figures. Tenmei population however declined by 20 percent and Tenpō by 14 percent. Second, in periods of population decline, the drop in households was less that the fall in popula79 tion, a circumstance which suggests something of a real-life count. All these considerations suggest that the Sendai data are relatively sensitive, though the comparative stability for the years between crisis periods is striking (yet explicable if a regime of relatively low fertility is assumed). In summary, whatever the original reasons for registering population, the primary long-term concern was revenue. Han counts (and national ones) as already set out, in what are the best known omissions, did not list samurai, and excluded the young to a variable age limit. Hence, by definition the figure for han population was less than comprehensive. So far so good, and if this was the only complexity, the profile of population would be easy to comprehend. More seriously, han counts often excluded – for the same reason, i.e., that they were not sources of revenue – Buddhist and Shinto domains, towns, and the commoner population resident on fiefs of samurai themselves, and on the lands of collateral branches of the daimyo family. The significance of this is sometimes underestimated because accounts state simply that the servants of samurai were omitted. But the omissions were larger. Not only servants, but, in the case of fiefs, commoner landholders were left out. This problem is made the more treacherous by the fact that practice as to how the figures were presented could vary over time. The most interesting case is that of Nanbu where dramatic contrasts exist within han figures which is a subject of potential confusion in interpretation unless the reasons for the deductions are understood. This did not pose problems for han administrators at the time as they were working with current or recent data. But modern tabular presentation will give an impression of confusion to a degree which did not exist among contemporary officials who had no concern of maintaining a consistent time-series with long term comparison in mind. The modern creation of tables is artificial. While the officials were not statistically-minded ones in the modern sense, they knew what they were doing, and their methods answered their requirements.

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

99

These problems are not only mirrored in but to a degree can be measured in comparison of the census counts of 1846 and the first modern style count from the registers for 1872 (returns from which were finalized in 1874). The populations of the Tōhoku and of Kyūshū appeared to rise sharply between 1846 and 1872, while at the other extreme the Kinki and Kanto rose little. In the Kyūshū and Tohoku regions population rose by 42 and 38 percent respectively compared with preceding 80 stagnation in these regions. Kinki and Kantō, on the other hand, had a relatively static population from the 1780s to 1846 and by comparison only a relatively moderate rise occurred between 1846 and 1872. To some extent the contrast between on the one hand both Kantō and Kinki and on the other the tozama of more distant regions was an abiding consequence of the sharp reduction of samurai numbers in tenryō and subservient fudai during the seventeenth century. Henceforth not only was their number small, but in the smallest han, a mere vestigial administration existed. However, comparison between 1846 and 1872 is complicated less by the fact – real enough – of the omission of samurai than by the meaningful survival of fiefs in eastern and western regions of Japan. Satsuma had both numerous samurai and numerous fiefs. Sendai, on the other hand, while having modest samurai numbers, had a large population of commoner-residents on fiefs. Total population in the case of some han – or kuni – appeared to rise dramatically in 1846–1872. The territories occupied by the Shimazu family provide the outstanding instance. The large domains administered by the Shimazu from their castle town of Kagoshima consisted of two kuni. The population of one of these, Ōsumi, appeared to rise by 159 percent, and that of the other, Satsuma proper, by 127 percent. From the 1750s onwards the population of Ōsumi had fallen somewhat. The population of Satsuma no kuni rose until 1786, then stagnated. Hayami has referred to the speculation by some that the profile of the data suggested that there may have been at some point in the 81 past a decrease in the population of Satsuma. In the two cases, however, the data may not have been arbitrarily defective. The Shimazu family had the largest following of samurai in Japan, a majority living in the countryside. In essence they were fief-holders (hence recipients of rice from peasants below them and themselves free from paying a rice levy to a daikan). The census counts must have embraced only occupiers on the directly held domains of the lord, in other words in districts contributing rice to the daimyo’s kura. The relative long-term stability of the figures also gives the impression that han

100

EARLY JAPANESE TRADE

officials worked from notional counts rather than from more laborious regular compilations. Rough estimates will give one an idea of the situation. In the Tokugawa census of 1846 the combined population of Ōsumi and Satsuma was 341,009; in 1872 it was 808,256. Samurai, who of course were not enumerated, would have been about 30,000, or including immediate family members (and servants) a total on the 82 order of 150,000–210,000. Most though not all of this population would have been fief-resident. Deducting 150,000–210,000 for samurai and their families from the 465,247 rise in population from 1846 to 1872 a balance of 255,247 to 315,247 still remains to be accounted for (if the deduction should be smaller to allow for samurai with their servants resident in the castle town, the residue would be even larger). The great majority of this residue of 255,247– 315,247 would have been commoners resident on fiefs. In other words, the deficiency of 465,247 in comparing the census figure of 1846 with 1872 almost vanishes, if to the 1846 total are added the lower-bound estimates of 150,000 in samurai families (and servants) and of 255,247 fief-resident commoners, all of them unenumerated, in 1846. Under the watchful and harsh eye of so many small fief holders, the commoner population would have lived in debased conditions: in contrast to peasants elsewhere who cultivated with a degree of independence, they would have been little more than servants or laborers to rustic samurai. Evidence of harshly imposed targets in the output of commercial crops, for which Kagoshima is well known in the modern literature, would only have reinforced poverty. The low level of literacy in Kagoshima-ken in 1884 mirrors their degraded status. Kagoshima (i.e., Satsuma) was not of course alone in presenting problems in its census counts. Another case of a problem han is Suruga in central Honshū: in Hayami’s view, the exceptional ups and downs in 83 its census returns were a result of bad data. However, the data could as easily reflect variations in the reckoning of kura and kyūsho lands, and which were repeated in the returns to the shogunate. Manipulations in Nanbu figures illustrate some of the distinctions which arose at least internally. In the census figures for Satsuma no kuni a rise in population between the 1750s and the 1780s could conceivably mirror some reduction from an even higher earlier total of fief-resident samurai. In striking contrast to the rise in population in the Tōhoku and Kyūshū regions between 1846 and 1872, the change in Kinki and Kantō was modest. Population had fallen in Kinki and Kantō in the 1780s, and on census evidence that of the Kantō did not recover. In the Kantō,

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

101

there was a decline in 1786 (sharper even than in the Tōhoku), persistent stagnation in subsequent returns (even the Tōhoku managed modest growth), a further fall in 1834 and an upswing from 1834 to 1846 (a period when population fell in Kinki). The Kinki was much the more stable of the two regions. Population decreased little in the 1780s and did not rise sharply between 1846 and 1872. Thus comparing Kantō and Kinki, on the statistical evidence, trends diverged in the 1780s; in 1834– 46 they diverged again. Once more in 1846–72, they contrasted, when a modestly substantial rise of 16 percent in Kantō significantly outpaced a 6 percent increase in Kinki. While the graveyard theory has been suggested as an explanation of the comparatively low demographic profile of both regions (heavy migration to towns and high death rate), this does not explain the contrast between the two regions themselves. The fact that statistical stagnation in Kinki preceding 1846 was followed by a very modest rise between 1846 and 1872 suggests that the figures were close to reality, with only modest adjustment arising in 1872. A limited fall of 2 percent in 1834–46 was matched by a modest 6 percent rise in 1846– 72. Kinki was a developed region, much of it highly regulated tenryō land, and as samurai were few, issues of including or excluding kyūsho (fief land) scarcely existed. While a highly developed pattern of tenancy, in other words rather large numbers of cultivators of non-honbyakushō status, also existed, the limited change between 1846 and 1872 suggests that registration was close to covering the entire population. Kantō was both economically and geographically a more heterogeneous region. The population data of Mito, which are included in the population aggregates of the Kantō region as defined in modern study, point starkly to the problems that the bare data can pose. The population of Mito appeared to fall in the middle decades of the eighteenth century, 84 when comparable declines occurred nowhere else in Japan. It has long been commonplace to seize on the figures for Mito as support for the picture painted from other evidence of a han beset by an extraordinary degree of crisis. The penchant to seize on poverty as a major explanatory factor for trends in contemporary statistical data is potentially dangerous to interpretation. The data and pessimistic contemporary comment have frequently been seen as reinforcing one another. Superficially the widespread decline in population in Kantō might be presented as a consequence of the 1783 famine (and of problems which persisted subsequently). However, as the data for Hōreki 6 (1756) can be compared only with data for Tenmei 6 (1786) (with the exception of some data for intervening years surviving at least readily only for Mito), too much

102

EARLY JAPANESE TRADE

weight is placed on a single year, and by no means all the provinces in Kantō stood in an identical situation in relation to the volcanic fallout from the eruption of Asama in 1783. What is striking about the post-1798 data for Kantō is that the performance of kuni varied widely. While the population of the region as a whole either stagnated or rose marginally, there is a contrast between two quiet difference statistical trends. On the one hand, Musashi, Sagami, and Awa rose, and Shimōsa remained stable; all were kuni where tenryō predominated. On the other hand, Shimotsuke and Kōzuke (both kuni with some tenryō but also a substantial amount land in small fudai han) trended downward over the period. To these should be added Hitachi where a substantial drop in 1834 and a very sharp rise in 1846 broke with the stability 85 of immediately preceding decades. In the case of kuni where tenryō predominated, Kazusa alone departed from the common profile: in a markedly unstable pattern its population fell in 1822, rose in 1828, fell again in 1834, and rose once more by 1846. These four kuni (Shimotsuke, Kōzuke, Hitachi, and Kazusa) contrast with a more stable situation in the rest of Kantō. In 1834 they recorded a sharp decrease not recorded elsewhere; they later experienced, compared with the rest of Kantō a relatively sharp increase in 1846–1872. All these factors suggest differences or changes in the comprehensiveness of the coverage within 86 han, and automatically in the transference of data to the shogunate. The prevalence of tenryō appears to be a factor in stable counting (though, as the case of Kazusa suggests, it did not give immunity from erratic 87 processing). Where tenryō and minute fudai han existed, samurai were inevitably few. Musashi, Sagami, Awa, and Shimōsa had some of the features of the Kinki region where the population of kuni was relatively stable both before and after the 1780s. In other cases, though the proportion of tenryō was large, the pattern was unstable. This may suggest the possibility of variations in treatment of the commoner population on land held by collateral branches of the daimyo or on fiefs. However, in the case of Kazusa, as beyond Kantō for Suruga, instability rather than any clearly defined trend is the striking feature of what are the two extreme instances among all the kuni in Japan. NATIONAL FIGURES

These considerations remove any confidence about the shogunal figures as a useful guide to trends at han level. At best, over time, the quality of the figures varied widely between han and within han, their profile on

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

103

occasion infuenced by adjustments introduced in administrative circumstances of which we know little. How closely the intrinsic quality of the data supplied by the han were scrutinized in Edo at the stage of compiling national aggregates, is a matter of conjecture. However, we can be confident from the detail that no effort was made to adjust the han figures to make them fully comprehensive; urban populations and commoner populations on fiefs in many instances were omitted. The only figures for Japan as a whole which survive are figures in summary tables for all the kuni (depending on the actual count of kuni or some bunching in modern tables, the unit count would have been properly between 69 and 74), returned to the rōjū over the signature of an ōmetsuke ኱┠௜and a kanjō bugyō ຺ᐃዊ⾜. The absence in the surviving documents of any papers at an intermediate stage in the processing, or of contemporary tables containing sequences of census years, suggests that the treatment in Edo was casual. If there was sometimes enough curiosity for unofficial copies of individual census returns to be made, the total absence of tabular data from the records that are left for us is arresting. It suggests that tables were not circulated and may never have existed. Where comparisons were made, they are likely to have been with the preceding census alone. Matsudaira Sadanobu’s observation of the change in population between 1780 and 1786 seems unique amongst a dearth of general comment about statistical matters by high officers of state. Moreover, the issue of tabular presentation apart, by international standards a continuous compilation in an identical fashion of returns and of the introductory apparatus without modification is itself remarkable, and revealing of an operation conducted by rote. Officials at han level and in Edo castle went through the motions every six years. Wider scrutiny of aggregates was non-existent. It is hard to imagine elsewhere any pattern of statistics being compiled in rigidly unchanged format for as long as the period from 1721 to 1846. That also reinforces the impression that in the shogunate (as in the han for the returns from the villages), the primary returns were probably destroyed soon after the event. The view of Ogyū Sorai Ⲷ⏕ ᚂᚙ(1666–1728), so admired in the eighteenth century for his commentaries on public life, on the rule-of-thumb nature of administration or the low culture of record keeping is apposite. No Office should fail to keep records of business. At present it is the general practice to deal with business on the base of precedents and established procedures committed to memory. It is entirely due to the

104

EARLY JAPANESE TRADE

lack of records of business that the officials are vague and ignorant of the 88 duties of their offices.

Population figures hardly admitted of being committed to memory. But they were deployed only as a short-term fruit of administrative processes. They were never looked at except in the immediate context of the time of their original creation. The absence of contemporary tabular runs of data suggests that at best not only memory but paper never ran beyond the immediately preceding data. If copies were made, it was of individual censuses. Where returns survive of figures of several censuses, they do not survive in official documents (or copies of them), but in compilations made by private individuals at dates later than the final date of two or three censuses, probably from randomly circulating copies that have since been lost. Only in two cases did pre-1868 writers refer to data for three censuses, and only in one case to full aggregates (for two of the three years). This pattern suggests that figures, or at least some knowledge of them, did float around with some ease, if not with abandon. Nothing approaching a consolidated table existed, however. That raises the question of whether the figures were secret, or whether the real factor was simply a lack of statistical sophistication or even plain curiosity. The answer to this question lies in part in analysis of the surviving returns. Were they in fact secret documents, or at least documents that never circulated outside a confined circle? Or is it the case that they did circulate? The fact that they did not circulate more widely or did not merit more penetrating comment might support the conclusion that the real problem 89 is less secrecy than a lack of interest. From what one can judge from the small number of surviving national census returns (variously complete or incomplete), the form of the actual return was remarkably unchanging (sometime evident also in other forms of statistical compilation in the Tokugawa period). The one period in which an interest in demography was actively entertained by the shogunate was the 1720s and early 1730s, when it reflected Yoshimune’s vigor and all-embracing interest in problems. In addition to ordering the once-off census of 1721 and the new census of 1726 (as one to be held henceforth every sixth year), Yoshimune uniquely had an interest in putting population in a longer perspective: in 1734 for instance he instituted population enquiries in regard to preced90 ing trends in the population of nine large tozama han. Even as early as 1723 the Nagasaki deputy opperhoofd noted that he had been told

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

105

by an interpreter that the bugyō “had received orders from court compelling him to conduct a census to find out how many people aged between 80 and 90 were living in Nagasaki. Nobody knows why the 91 [shogun] has ordered this census.” These years were an unprecedented occasion of broad-ranging administrative vigor in demographic counting in Tokugawa history. Even if innovatory, the reporting of data to Edo, as the latitude allowed to han showed, was less an imposition of a new task on them than harnessing existing counting within han to a national framework: it formed part of the policy pursued by Yoshimune of enhancing the role of the shogunate. There was also in these years the institution of trade figures for traffic between Osaka and Edo. Again, this was not in itself an innovation: detailed figures had already been compiled for Osaka for 1714. Yoshimune’s contribution (or that of his officials), as part of his administrative drive, was to put their compilation on a regular basis. There is no evidence in later times of statistical curiosity comparable to that of Yoshimune. A guess would be that the figures are closest to reliability at their first launch in 1721 or 1726, when they were novel and driven by Yoshimune’s inquisitiveness, thereafter in all probability becoming fossilized (except when interrupted – at han, not Edo level – by random revisions). The poor survival of data – and at the level of han administration near total absence of evidence of how the surviving han totals were arrived at – raises questions about the administrative processes themselves. The population returns from the villages (whether described as shūmon aratame chō or not) despite the fact that large quantities of other categories of paper at large have survived for some han administrations, were not retained by the han authorities. Where the primary returns have survived, it is in some extensive sequences retained by the families of former village headmen (sometimes still in family possession, in other cases in library custody). At the center, nothing whatever survives for the Edo authorities. It is easy to assume the contemporary or later destruction of records by natural disaster as an explanation for the poor survival of primary data. However, Japanese administration was conducted on the basis of much copying of documents. What is surprising is that in a society which copied documents and equally often made multiple compilations from them (the prime means by which Tokugawa records have survived), so few statistical sources exist. The copies survive in isolated contexts, not in compilations, and in the few cases where figures exist for two or more years, they are very obviously by individuals from outside shogunal administration (in a mere three

106

EARLY JAPANESE TRADE

cases – and for individual years, not for sequences of years – does a hint of an official association exist, and even then only in the ambiguous or vague sense that they survive in collections which inherited paper which existed ca. 1868). The absence of sequences of figures for successive censuses – in other words tables – reflects how official interest in the census was itself circumscribed to the immediate objective of finalizing by rote individual censuses. In the case of the han where documents are more numerous, the data remained embedded in nisshi of various sorts, and were never collated. Japan still lived in a pre-statistical age, in effect. It was not secrecy that was the handicap, but lack of analytical interest. The nisshi are in essence administrative documents compiled routinely, year after year. URBAN POPULATION

Despite their many limitations, han figures are by no means arbitrary. If all too often unsophisticated in the approach which han sources suggest, the differences are caused by changes of definition rather than by erratic handling. The problem of consistency did not pose itself for administrators simply because they did not view the data in a chronological and hence tabular framework. Urban populations were recorded by han officials (though they were not necessarily entered into han totals and hence, the independently-counted shogunal towns apart, were not entered by the shogunate into kuni figures). It is now necessary to look at urban population. In contrast to rural society, the registering of urban population was not fundamentally driven by revenue purposes. A consequence was that data for categories other than rural population (who were subject to a rice levy) were not necessarily carried into han aggregates, and the omissions, where they existed within han totals, appear to have been repeated in the general returns for census purposes to Edo. There is of course an inconsistency in this. Han effectively were free for statistical purposes to treat their urban population as they wished; the shogunate on the other hand not only counted the population in the shogunal cities, but dutifully entered the figures in the totals for kuni. In the case of han, where han sources are complete or detailed, estimates of town population (both jōkamachi and informal townships), whether or not carried into the han aggregates, exist. In some cases ambiguity arises as to whether towns were included in the han returns and consequently in the kuni data compiled by the shogunate from the

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

107

submitted information. Though in the case of towns under shogunal control, population returns were made directly by daikan or by bugyō to the shogunate, the survival in the capital itself of material from shogunal cities (Edo itself apart) was even in early Meiji times almost nil. Overall for shogunal towns, the survival of data is much poorer than it is for han. Even for the city of Edo itself, the survival of material has been meagre: what we have is mostly data entered into later compilations, from earlier sources which no longer exist, and of which we know singularly little. In the case of Osaka, best documented of the shogunal cities in point of population, figures variously survived into post-1868 times in documents of the time or in copies made directly from pre-1868 sources. However, while continuous for most of two centuries, the surviving detail for Osaka is mostly summary. All these data for shogunal cities (and, as far as can be seen, for towns in the han) originated in registers of residents. These were maintained from an early date. For Osaka, data have been attributed in Osaka sources to several pre-1664 years, though figures can be attrib92 uted reliably to a sole year 1634. Katsu, who compiled population data for Edo, also reproduced the population of Kyoto and of Osaka in 1665 and 1681 as well as providing a somewhat wider range of detail 93 for Osaka in the latter year. Figures for Kyoto exist for 1634, 1665, 94 and 1669. A small number of figures exist for other cities, the earliest 95 for Nagasaki in 1616. While some taxes fell on houses, revenue was a minor consideration. The prime purpose of maintaining registers of population was administrative: security, food supply, and fire fighting all warranted a knowledge of town population. In the figures based on registers some additional data were compiled even if they have survived only fitfully if at all: the number of household units, whether the status of occupier was owner or tenant, and in some instances particulars of where residents were born. In particular some of the surviving Edo data 96 for later years distinguish between birth in Edo and elsewhere. These data have a relevance to the belief, in periods of crisis, that the towns were overpopulated, and to the advocacy of removing population (in Edo in the Tenpō crisis the authorities made some attempts to implement this, following a less wholehearted Kansai precedent). As in the case of rural population, samurai were not counted. This means that registers understate population in the castle towns or jōkamachi. While samurai are by definition a larger proportion of castle-town population than of the han population at large, samurai (including family members and servants) are a minority in the sense

108

EARLY JAPANESE TRADE

of not exceeding crudely a range of 20–40 percent of gross population. Edo itself, through the effect of the sankin kōtai, was the great exception: the enumerated commoner population of 500,000 should be doubled to take account of daimyo and samurai of all categories, 97 including dependants. While this samurai population would include servants as well as family members, except for hatamoto (the larger of whom maintained almost daimyo-like lifestyles), retainers were few. The omission of samurai is not a serious problem for Osaka, as few samurai resided there. Doubts can of course arise about the urban counts based on the registers (which in Edo itself were updated twice a year). Lack of close revision and substantial omissions are both possible. A distinction was made between those residing as of right at their address and those who only rented the property. Such a distinction exists, for instance, in the 98 figures for Fukuya-machi in Nagasaki from 1742 to 1863. However a distinction of this sort may not be sufficient to either catch or cover all temporary residents, more particularly in a large metropolis with much movement between the city and a large adjoining hinterland. The real problem was a large floating population of indeterminate status, occupation and even residence. With revenue a very minor consideration, there is no guarantee that the registering of population was conducted with special care once the basic details of household heads and recognizable tenant occupiers were confirmed – or modified annually or biannually. The scale of the problem is further highlighted by the high proportion of urban residents who were born elsewhere: in the late Edo period, between a quarter and a third of the population of Edo came from elsewhere, and more than 30 percent of the popula99 tion of the five provinces of the Kinai lived in Osaka, Kyoto, or Nara. As Osaka data on birthplaces are fragmentary and for the earlier years only, this high proportion implies a similar large-scale movement into the three big cities of the Kinki region. One may hazard some guess at problems: informal consensual occupation of space by transient figures, the omission of menial or casual servants whose occupation was uncertain, and in general a limitation of close enquiry once a household head was established or an existing one confirmed. The most striking case of underestimation is of women in Edo. There, in surviving counts for 1721–47, drawn from the registers, women made up less than 40 percent of the population. It has to be assumed that there is some systematic bias, either of status or of occupation. The discrepancy between males and females for the districts that were under the

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

109

direct control of the bugyō is repeated in almost identical proportion in the separate data – where they survive—for the very large number residing on temple and shrine lands in Edo. The combined omission, of the order of one hundred and forty thousand, is a very large one. Given a floating and transient population and some distinctions of which we are not clear, registers of their nature would probably have been more reliable counts of households than of individuals, though 100 the survival of data for the number of household units is very limited. These inbuilt limitations were not dissimilar from ones in han data. However, the fact that when figures for Edo resume after a total gap in data for the years 1748–1831, the discrepancy had been narrowed, on balance speaks well for the statistics, and suggests that the basis of counting had been revised radically at some intervening date. It was further narrowed in a progressive fashion over the remaining years. By 1867, male and female population in Edo counts were almost equal. Likewise a sustained downward trend in Osaka population from a peak in Meiwa 2 (1765) suggests that figures were sensitive to under101 lying changes in population. The omission of women on such a scale in earlier years poses a problem for the estimation of long-term trends in Edo population: if 140,000 were added to the total of population for the years up to and including 1747, the profile of Edo population in later years becomes less positive. This would also have in comparative terms the consequence of emphasizing the sheer dynamism of Osaka as its population expansion continued to a peak in 1765. However, these speculations simply underline the inherent problem we face when attempts are made to draw conclusions with any confidence from a pitifully small mass of data of whose compilation methods we know nothing. The survival of documents, not simply of originals or Tokugawa-period copies, but even of sources that were preserved long enough to be drawn on by early post-1868 compilers, is very poor. There are some pre-Meiji counts for Osaka, but for Edo there seems to be a single pre-1868 document – “Gojōka machi Toshima-gun, Ebara-gun, Katsushika-gun no uchi ninzū cho” ᚚᇛୗ ⏫㇏ᔱ㒆ⲥཎ㒆ⴱ㣭㒆அෆேᩝᖒ – which by definition has to be 102 post-1840, as it includes census data for that year. Where data for towns exist in han sources, they are usually contemporary, even if surviving in the form of entries in han compilations rather than in series of primarily demographic data. For shogunal cities the data are much more fragmentary and often in miscellaneous documents (or copies) that are not in essence official. Kyoto, Osaka, and

110

EARLY JAPANESE TRADE

Edo, in a class apart because of their size as well as status, have limited data and few documents that can be regarded as having been directly copied from official documents. For Kyoto, only a handful of figures 103 exist. For Osaka and Edo, while the sources cannot be described as rich, things are somewhat better. In the case of Edo, Kōda Shigetomo ᖾ⏣ᡂ཭in 1938 devoted a famous article to the subject of Edo popu104 lation and the sources for it. He asserted that the data were set out in 105 three modern authorities. One was “Edo shigai tōkei” Ụᡞᕷ⾤⤫ ィ, published in Edo kai zasshi Ụᡞ఍㞧ㄅ in 1889 by Yamashita Shigetami ᒣୗ㔜Ẹ, who drew on what he said were zakki 㞧グ, but offered no detail about his source (or sources). The second work was Katsu Kaishū’s Suijinroku, also from 1890 (though Kōda dated it to 1887); Katsu identified two sources for his data. The third source was an article by Yuzuki Jūzō ᰆᮌ㔜୕and Horie Yasuzō ᇼỤಖⶶ, “Honpō jinkō hyōchū Edo no jinkō” ᮏ㑥ேཱྀ⾲୰Ụᡞࡢேཱྀ, in Keizai shi kenkyū, 1930. The figures in this last study, drawn from wide but heterogeneous sources, are often irreconcilable with one another, 106 Kōda noted; in his view, they are not to be considered useful. However, accurate or not, they imply access to sources that contain figures differing, for whatever reason, from the numbers in the sources available to Yamashita and to Katsu. All these now-unknown sources of varied origin imply that copies of data survived at least until the 1920s; these copies appear to have been in private hands, in the main and perhaps in all cases. The provenance of the data is wider still if it is borne in mind that both Yamashita’s and Katsu’s compilations made a distinction between data drawn from the biannually revised registers and from the returns in the census held at intervals of six years (though census counts were of course themselves based in the last instance on the registers). Yamashita was not explicit about his sources beyond the statement that they all came from zakki. Katsu was more forthcoming, 107 as he presented his data in two sections in his Suijinroku. The first was entitled “Edo jinkō shoki” Ụᡞேཱྀึᮇ, the second, “Shōtoku yori Kōka made Edo chōsū jinkōsū” ṇᚨࡼࡾᘯ໬ࡲ࡛Ụᡞ⏫ᩘேᡞᩘ. The first section contained summary data (a single gross total for each census) from 1721 to 1834; these are clearly census data. Except for the first year, they are also identical, according to Kōda, to the data given by Yamashita. The second section, for which Katsu cited “kōzuka” ዲ஦ᐙ – meaning private collector – as the primary source, was by definition drawn ultimately from the registers. It also began earlier than the cen108 sus; the first count was from 1713. Katsu’s counts continued to 1845.

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

111

The summary census data from both Yamashita and Katsu are of importance because for the period of the gap in totals from the registers for the years 1748–1831, they are the only population counts we have for Edo. Neither Katsu nor Yamashita have data for the census year 1840. The full text of the return by the bugyō to the Kanjōsho for that year survives, however, in a document included in a compilation entitled Tenpō sen’yō 109 ruishū ኳಖ᧝せ㢮㞟. While it is a copy, not an original source, it is a contemporary one. Moreover as a report to the bugyōsho over the signature of the two machi bugyō, it may be a copy of the census report for that year. It is of course all the more interesting as, although a return for Edo alone, it is one of only two documents which contain detail at 110 any level (urban or national) of the 1840 census. In the first post-Restoration published source for Edo data, Yamashita in 1889 simply attributed his information, census and registration data alike, to zakki. The word zakki means collections or a collection of miscellaneous information, and hence implies strongly that the source (or sources) itself did not have an immediate official connotation. Katsu 111 is more interesting. While he gives no sources for his census data, for his section with data based on the registers he does. For two years he attributed his data to the Kanjōsho. It is doubtful if this is to be taken literally. It has to be assumed that they are mere copies; whether officially made or simply ones held by officials at or near the time Katsu got them, is not clear. As they contain data for only two isolated years 112 (1731 and 1737) and he has no data from the same sources for later years, direct provenance in late Tokugawa times in Kanjōsho records seems highly unlikely. They are in other words likely simply to be full texts of the reports. Katsu’s categorization of his source for the two early census years as official must be simply that he was drawing on a full text and hence on one which contrasted with his second and major source, demonstrably in its terse and incomplete details, of unofficial origin. By characterizing his source as kōzuka, a collector or collectors of things by curiosity or antiquarian motivation, he hints that the materials are miscellaneous and not dissimilar from the zakki that Yamashita used. It is likely that the data were drawn from the records of several compilers, as the detail varies from one year to another, and supplementary though terse comment for individual years also varies. If, however, he had been working from a single collector’s papers, their nature suggests that the collector himself compiled them not from access to official documents but from highly miscellaneous sources. While sources described as zakki or as kōzuka are uncommon (if not unknown) for demographic

112

EARLY JAPANESE TRADE

information, sources of this sort for other categories of information were numerous in Edo times. For instance some of the sources consulted even for a highly political purpose by a group as exalted as the team deployed as the Hayashi family to compile in 1849–56 the Tsūkō ichiran, were collections, mostly not now surviving, whose titles often implied quite literally miscellaneous or hearsay origins. Edo gross totals were arrived at by aggregating figures for population of the three shogunally-administered sections of the city and for the large population – some 60,000 – resident on lands under the dominion of temples and shrines. In the case of counts surviving from register origins (one or two for each year) for a total of twenty-two years between 1733 and 1844, the distinction between population under bugyō administration and temple-shrine administration respectively survives for most years. Where a single summary total existed, the coverage at times was a subject of confusion. For instance, the data on census population compiled by Katsu, while intended to cover the entire commoner population (directly administered and temple/shrine populations), are at times inconsistent. For some years, though he was unaware of the problem, they demonstrably cover the shogunally administered population only. Thus the jump in 1732 compared with 1726, of the order of 60,000, is a consequence of inclusion in the later year of the temple and shrine population. It jumped again in 1750, 113 conceivably for the same reason. This is important to note because the inconsistencies in the data might suggest greater confusion in the original statistics themselves than is the case. The problem for Katsu arose from the very miscellaneous information available to him and from his reliance on the styling of data for individual years by the compiler(s) on whom he relied. Thereafter consistency may be assumed: the data (together with Yamashita’s) may be taken to be inclusive of both categories, for the years 1756–1828, for which moreover they are the sole surviving data for Edo population, as figures from the population registers do not survive for these years. Osaka is the best documented town for population data. For the years to 1664, an isolated figure for 1634 is recorded in the Nendai chobunshū 114 ᖺ௦ⴭ⪺㞟. Data attributed in another source to an even earlier year, 1625, are identical to data for 1669 in an early Meiji document that compiled data under the title of “Beishō kyūnikki” ⡿ၟᪧ᪥グ (Old Dairies of a Rice Merchant); the data are obviously wrongly dated. From some five sources data survive from 1665 to 1756 inclusive, for 115 twenty-three individual years. By contrast, for later years, there was a

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

113

very large gap until a document compiled by the Osaka machi bugyō Isshiki ୍Ⰽ came into Kōda Shigetomo’s hands in 1911, two years after the end of his long stint (1901–1909) in managing research on 116 the first Ōsaka-shi shi. They covered all the years from 1757 to 1856. Some of these years and, especially beyond Isshiki’s terminal year of 1856, the years 1857–62 (except for 1860, which is omitted in all known 117 sources) are covered between two other sources, “Tekagami” ᡭ㚷 which already had some data for earlier years, and “Shōkisai hibi zakki” 118 㙂ወᩪ᪥ࠎ㞧グ. Thus in all, including Isshiki’s document, at least seven sources were used in the modern Ōsaka-shi shi. Nor do these documents exhaust totally the information. Data for 1689 (an otherwise undocumented year) exists in “Settsū shō” ᦤὠ㕒, and an existing 119 total for 1703 is duplicated in “Jikatayaku tekagami” ᆅ᪉ᙺᡭ㚷. There are no data for the final bakumatsu years, for which untypically a dearth of Osaka figures contrasts with an Edo count for 1867. The document by Isshiki is of particular interest not only because it provides a complete run of data from 1757 to 1856 (and is unrivalled as a Tokugawa document in that regard), but because it constitutes a unique survival of a working document of population compiled by a serving bugyō and even more strikingly in his own hand. In other words it is not simply a copy of data from other sources, but a very rare holograph document and as a statistical source a unique record. Isshiki’s collection of documents consisted of papers from his years of office together with copies of records for earlier years (some made long before the years of his office-holding). Although he had retained papers after the ending of his period of office in 1861, his documents are confined to his years as bugyō or to preceding years; he did not, in later years, engage in work on Osaka’s past. His population document with its authentic feel or appearance as a document of its times is 120 certainly contemporary with his years of office. Unique in Japanese demographic history though that makes the document, what is even more unusual is that it is a tabular document. It is the only known case of such a demographic document for the entire Tokugawa period. The immediate reason for the Osaka compilation is not clear, though compilation within the years 1858–61 may imply a reason, whether 121 some sense of the known economic crisis in Osaka’s affairs at the time or – and not unrelated – the prospect of the city being forced to open sooner or later under the terms of the 1858 treaties for foreign trade. Isshiki’s access to data was not complete. Though his document started from 1750, suggesting a wish to collect figures as far back as that year,

114

EARLY JAPANESE TRADE

actual figures appear only from 1757. While he attempted to give data for five districts (Sangō – i.e., the city itself – plus three outlying districts and the eta mura ✧ከᮧor outcast community), his data became 122 fully comprehensive of the five categories only from 1797. It seems clear that Isshiki had no data for the years 1750–56. Indeed the reason that he did not attempt to go even further back may simply have been that he did not have ready access to earlier data. There is an irony in all of this as in modern times outside Isshiki little at all survives for the later years. As surviving post-1756 documents, Isshiki apart, have very little information on Osaka population, it is clear that he had access to sources which no longer exist. While Isshiki’s document could conceivably be a mere transcript of another document, it seems more likely that it was drawn up directly from existing records. If it were a mere transcript it would seem more likely to have been compiled by a clerk. An overall survey of the information surviving for pre- and post1856 years yields the impression that there was an incredible attrition of paper over time. While available for many fewer years and in a smaller number of sources (some of which still survived in early Meiji years but are now unaccounted for), the Edo information is richer in content than Osaka information, as it gives a breakdown of men and women; some details of house numbers, births, and dekasegi figures also exist. Osaka information, though more continuous (complete, apart from sixty-seven 123 years, for the entire period from 1669 to 1862, all of the missing years except one year occurring before 1756), is for the most part very summary. While Isshiki gives information for a long uninterrupted sequence of years and it is possible that he was not interested in data other than summary totals, equally, fuller data may not have been available to him. Despite the long run of summary figures and a moderately large number of sources, all surviving data for Osaka in the last century of Tokugawa administration lack the division into men and women. However, the fact that surviving data on men and women, birth places, houses and territorial distinctions are fragmentary and 124 for early years, and the telling fact that Isshiki’s own breakdown of the five districts in and around Osaka data was incomplete for 1756–96, suggests that he too had limited access to data. In the case of both cities, information survives only in copies, and the pre-1860 documents are mostly figures in isolation in compilations which in turn drew on other – unknown – sources. The contemporary copy of the report on Edo population in 1840 cited by Kōda is cer-

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

115

tainly not, as he suggests, a Kanjōsho document in the sense of a 125 document physically originating in the Kanjōsho. The discussion in the appendix to that document as to the basis on which the information was compiled would have been superfluous to an official of the Kanjōsho itself. While the copyist clearly had access to a text for 1840, the fact that he referred to eight preceding censuses without citing the data (and conceivably moreover compiled his copy in 1841–43, years when policy towards excess population in Edo was a subject of debate), gives the impression either that he had no access to such information or, despite making his observations in a time-frame which took account of preceding censuses, chose not to create a table. The presentation in documents drawn from miscellaneous compilations, and in many cases for individual cities, of mere bald summaries, suggests a large undergrowth of figures that circulated and were copied in a somewhat random way. It would be possible of course to argue that lists similar to Isshiki’s existed either for Japan at large or for some of the cities but have not survived. However, the absence of tabular returns in the widespread copying of information by largely private copyists suggests that they did not exist in the first place. This would be consistent not only with the often summary and random character of the surviving data but with the striking absence in all economic and social commentary of a statistical context. The upper officials of the Kanjōsho, overworked and understaffed in an administrative framework that lacked specialized bureaux and tasked officials with highly heterogeneous responsibilities, almost certainly had no leisure to look at things other than in a short time scale. The problem was compounded by the lack of specialized staff below them. On the other hand a machi bugyō, able to rely on a large and stable staff working within comparatively narrow or specialized terms of reference (the finite and repetitive task of administering a large city) in the bugyōsho, might at least dream of viewing problems in a longer-term perspective. However, the unique Isshiki document suggests that even if a bugyō desired to do this, statistical information was not immediately to hand in a time series, and he would have to rely on personal endeavor to produce a summary table. ARCHIVAL ASPECTS

A career in bakumatsu and Meiji government service such as Katsu Kaishū’s was itself no guarantee of greater access to material than others enjoyed. This fact supports the argument that population figures served

116

EARLY JAPANESE TRADE

little administrative purpose and the exercise of collecting data had fossilized. To observe this is less to criticize the Tokugawa administration than to take note of the lack of statistical sense in Tokugawa society at large. If deficient in a statistical sense, the scale of the administrative exercise remained substantial, involving statistical work of varying quality within han, and every sixth year a response by han (tozama as well as fudai) to the shogunate. At local, han, and national levels, the pattern, once instituted, survived for 120 years. Survival however also reflected the weakness of Japanese administration: a tendency for procedures once set in place to be faithfully though narrowly conducted, without scrutiny of their applicability in later times. A further feature – consonant with the constitutional character of Tokugawa Japan – was the absence of any national co-ordination or supervision. Thus they reflected the characteristics of Japanese society, administratively competent in many ways, but responding to the need of innovation only when a pressing challenge presents itself. The widening diffusion of the shūmon aratame chō exercise from the 1620s to 1680s (a period in which a sense of crisis existed from Europeans up to the 1640s and from events in China through the entire period), the systematization ca. 1680 (in han which did not have shūmon aratame chō) of an earlier and apparently occasional model of counts drawn from registers of households, or the whole range of initiatives launched by Yoshimune – at grips with the economic and social problems of Japan in the 1720 and early 1730s – reflect an impressive ability to respond in a bureaucratic sense to crisis, whether serious or simply perceived as such. Thereafter while the national census exercise instituted in 1726 continued, the quality at both han level and at national level deteriorated. Summary figures, either a mere total or more complete copies of 126 the returns for Japan, exist in various sources in the period. Often there is no more than a single figure for Japan (sometimes without even a breakdown between male and female); in other words, a mere total was as a rule better known than any breakdown, large or small, into component parts. The key data, apart from a mere gross total, were the population breakdowns for individual kuni (and for both kuni and Japan at large, the distribution between male and female). The official return invariably recorded the kokudaka of the kuni. The official return over the signature of two officials, an ōmetsuke and kanjō bugyō, was presented in a stereotyped format from 1726 to 1846. Some data (variously a mere national total or fuller data) survive for nineteen of twenty-two censuses up to and including that of 1846 (i.e., ignoring

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

117

the censuses that may or may not have been compiled for 1852, 1858 127 and 1864). For eleven censuses a breakdown at the kuni level survives, and for twelve censuses, males and females are distinguished, either in national figures or broken down for all individual kuni. For four censuses, one or more copies of a report to the rōjū exist in a full text (date, formal introduction, signatures of the two officials, breakdown into population and kokudaka respectively for kuni), and for seven censuses, a semi-complete text giving the important population figures for all kuni but omitting one or more of the supporting details (i.e., full date, covering letter, signatures of the two officers, kokudaka, or in one case, though presenting data for the individual kuni, the grand total). The lack of a date in surviving copies has some times caused confusion, though if the officers’ names are given, it is possible to identify the date of the census from the term of office of the officials. For a further census (1732) for which we do not have a breakdown, the reference in Katsu Kaishū’s published work in 1890 – unless his information was totally wrong – could be taken to imply that such a source which he had not seen, existed. Thus the overall pattern is as follows: Date

Degree of Completeness

Comment on Contents

1721

Incomplete text; compiled in the 1830s

Contains data for all kuni, but lacking a distinction between males and females, also lacking grand total and the kokudaka

1732

Only a bare total is known

Katsu Kaishū’s reference seems to single out 1732 as a year for which a report fuller than his bare national total existed

1750

Incomplete text

Recorded in the 1770s, the sole data being population figures for the kuni; a second version, said by some to be of independent origin, may in fact not be of independent origin

1756

Incomplete text

Recorded in 1770s from the same source as that which recorded the 1750 data

1786

Incomplete text

From a 1930 publication drawing on a copy made (or extant) in early Meiji times; full version, apart from lacking the signatures of the two officials

1798

Full text

Formerly in the possession of the Machida family; has internal errors in transcription

1804

Incomplete text

By Katsu Kaishū, complete except for dating, where the month but not the year is indicated

118

EARLY JAPANESE TRADE

1804

A second text, also incomplete

Said to be from the Meiji period; full, except for missing date; transcriber dated the document to census of Kansei 10 (1798)

1822

Full text

Originally in the possession of the Machida family; has internal errors in transcription

1822

A further text may have existed

Containing kuni population, this item, now lost, appears to be the source from which two later summary documents drew their information

1828

Full text

In a ms chronicle of Tokugawa times in Wakayama University

1828

Incomplete text

Detail presented by Inoue (1904); may have originated in a copy transcribed in Kaei 6 (1853)

1828

A further text may have existed

Contained kuni population; two later summary documents appear to have drawn information from this now-missing text

1834

Full text

In National Diet Library; may be a copy held by an official in the 1860s; a further text, edited in 1917, may be a fuller version of the same document giving further detail, but with some errors in transcription of figures

1834

A further text (ca. 1834) The Sanka manroku, which records details from may have existed the 1721 census, refers in its title to a 1834 census which is missing from the surviving ms

1840

Incomplete text

University of Tokyo Faculty of Law Library; full, except for lacking date

1846

Incomplete text

Reproduced in Katsu Kaishū’s Suijinroku; lacks introduction and signatures, but contains kokudaka

A modest number of heterogeneous copies thus existed as the basis for our modern knowledge. Significantly, most of these were compiled later than the actual censuses on which they report. Several of these no longer survive. In the case of the 1732 document, its existence even in 1890, when Katsu Kaishū’s work was published, is conjectural. The Kanchū hisaku ᐁ୰⛎⟇ manuscript by Nishiyama Genbun すᒣඖᩥis no longer extant, but at least a portion of the original (fortunately retaining the demographic data) did survive until 1879, when it was published. Similarly, later citations of summary data for 1822 and 1828 imply that there were sources for those years that contained data on all the kuni, at least, but those sources no longer exist. In the case of the 1834 census, there may not in fact be, as assumed in modern times, two independent surviving sources, but a single document more fully transcribed in

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

119

one modern account than in the other. For this census of 1834, the existence at an early date of a copy that has since been lost is implied by the Sanka manroku ୕ᬤㅿ㘓, which purported to have detail of 1834 as well as of 1721. In the case of the population of towns, data for Osaka and Edo circulated in summary form in a miscellaneous range of privately complied sources. Edo sources that still existed as late as 1890 are now unaccounted for. INTERCENSAL AND TABULAR PRESENTATION

The full text of a document is important in an archival sense, as it can give a very clear idea of whether the document is itself a copy or an original. “Original” in this instance denotes a document made in the Kanjōsho or by officials who had served in it. In cases where we have multiple survivals of other documents, as in the case of fūsetsugaki 㢼ㄝ ᭩(reports submitted by the Dutch to the shogunate on arrival of their vessels in Nagasaki), it is easy to identify originals (or at least fair – or very carefully made – copies), other copies, and cursive or casual copies. If the document is itself a copy, an incomplete or semi-complete format implies the prior existence of other copies; it must be, unless one assumes a careless transcriber, a copy of an incomplete copy, and hence not a copy (a wholly faithful copy) of a lost original. As far as can be judged from what has been edited or published, no copy of census returns seems to have been a paper which originated physically in the Kanjōsho. The fact that Nishiyama Genbun, compiler of Kanchū hisaku in the 1770s, could give summary figures for Japan in 1744 and full figures for all kuni for the two years 1750 and 1756, is particularly interesting. The balance of probability is that a document such as his – compiled at a date relatively close to the actual census – had to be based on a copy or copies (probably not the originals, and hence a copy – or copies – drawn from one or more documents, at one or several removes from the originals). If it was drawn from a single source, of course, the documentary basis is by definition limited. If on the other hand the Kanchū hisaku is a modest compilation from several sources, it implies a somewhat wider circulation of the demographic data. At the very minimum, the fact that he presented only a round figure for an earlier year (1744), while he gave fuller data for two later years (1750 and 1756), suggests that the compiler was operating on the basis of access to at least two sources. The only other case of data being drawn from three censuses is that of Ōta Nanpo ኴ⏣༡␇. Writing as late as ca. 1800,

120

EARLY JAPANESE TRADE

Ōta produced a more modest amount of data: mere summary figures for three years (gross totals only for three years, 1721, 1726, and 1732, but including kuni population for ten kuni in 1721). All his data were drawn from a single 1735 source (whether original or copy is not clear). In treating some or much data for three census years, Nishiyama Genbun and Ōta were exceptional. There are four instances (three certain and one probable) which contain data from more than a single census, or did so at one time. The first of the three certain cases is a document the title of which mentions both 1721 and 1834; this suggests that the compiler had access to details of two censuses, and that the data for 1834 later became lost or detached. The surviving data for 1721 itself is incomplete in that it does not distinguish the male and female population. This now incomplete or mutilated document is clearly not simply a copy, but probably a copy of a copy. The second case is the Tenmei Kansei ninzū chō ኳ᫂ᐶᨻேᩘᖒ, supposedly for 1786 and 1798, in which the compiler had access to data for two census years in fairly full form. The data for the second census lacked a date, and upon careful inspection it is apparent that the copyist erred in attributing it to 1798 instead of 1804. This suggests that it was already not only a copy but an imperfect one. The third case is the the Shokoku ninzū chō ㅖᅜேᩘᖒ, a modest volume containing census data for 1798 and 1822. It is in fact the manuscript with the most complete survival for the data of two census years. Not only that, but it provides two of the meagre four fully complete texts in existence. In a technical sense it is in every way a complete document apart from errors in transcription by a copyist. It originated with the obscure Machida family (Saku-gun, Mimasemura in Shinano), a family name which crops up in no biographical dictionaries of Confucian scholars, monks or prominent people. The fourth case can only be inferred, because the documents or transcripts of them do not survive; it is suggested by the peculiarly con128 structed summary of the censuses of 1822 and 1828. These four instances of reporting the results of two census years cover seven distinct census returns. The copies, all made in the period 1804–1834 or somewhat later (taking the date of the second census in each case as suggesting an approximate date of the copying), testify to a pattern of survival of data from earlier censuses, as well as to a private knowledge of current or past censuses. In the first three cases none of the compilers held official responsibilities. The highly idiosyncratic quality of the fourth instance also points to a private initiative.

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

121

What is remarkable is the absence of data in a tabular form, and that where data for two or three censuses survive, the copies appear to be private. Despite the combination of widespread copying and random access to census data, tabular presentation of demographic data was 129 unknown. Isshiki stands out as the sole example of a tabular presentation. Katsu (open to Western ideas on arranging statistical data) is one of the first Meiji compilers to give a tabular presentation of the figures, and it has been assumed that his official status before and after 1868 favored him in access to the figures he presented. However he was able to reproduce full data for Japan for only two years, 1804 and 1846. Moreover, as his data for both years were not full texts, he was working from what were already incomplete copies. He was, it is clear from other evidence, a very conscientious transcriber, and the incompleteness of his return for 1804 implies that it is a copy he made from an incomplete copy, itself possibly a survivor at the end of a whole sequence of copying. His return for 1846 lacks both the covering letter and names of the ōmetsuke and kanjō bugyō, and hence is even less complete than his 1804 data. However, Katsu Kaishū is the sole authority for both the mere total and the detail of the last known census. Except for this chance survival, we would be in exactly the same position as we are in regard to the years 1852, 1858 and 1864; that is, we would have to conclude, in the absence of specific information, that a census may not have been made. Data, at least summary figures, were reproduced in a largish number of sources in Meiji times. For instance, while Katsu Kaishū reproduced data for many years, Komiyama Yasusuke in Kinsei jinkō no hanshoku (published a year earlier than Katsu’s Suijinroku) had data for 1756, 1828, and 1834 which Katsu lacked. Hosokawa, who wrote in 1883, had totals for 1721 and 1732, but he was not aware of Ōta’s work of ca. 1800 and had to draw from a different source. Much later Inoue Mizue, writing in 1904, while not adding new general totals, had access 130 to a fuller account of the census of 1828 than had preceding writers. Yokoyama Yoshikiyo (1879) had a variant total for 1744, and hence may have had access to a different document to those known. While Katsu was able to give full figures for only two years (1804 and 1846, the second year moreover in very incomplete form), in his summary table he reproduced data for ten other years (a single national figure for seven of those years, and a national figure plus a sex breakdown for three years). It does not at all follow that he obtained his data from former officials or by virtue of his earlier official position. His information probably came from various sources, reflecting

122

EARLY JAPANESE TRADE

the diffusion of documents rather than from any access his privileged place in a circle of officials with a foot in the pre-1868 world gave him. His citation of a work containing reference to the 1732 census shows that in the one case where he gives us a faint clue to his modus operandi, he seems to have been relying on an unofficial source. His total lack of post-1804 data, apart from the 1846 return, is remarkable, and seems to confirm that for his statistical work he enjoyed little real advantage. Indeed, the evidence suggests that he was relying on more “literary” sources. In other words, he was himself, on the demographic front if not in the case of political documents, a mere compiler of scattered information. This is even clearer for his data for urban population; when he identifies his sources, they are mostly miscellaneous. Apart from the data compiled from the population registers, the figures for Edo he gives for census years varied from one census-year to another, again suggesting that he drew on highly miscellaneous sources. He was no better than others in some respects, as Komiyama’s data for 1828 and 1834 prove. The singular feature of his data is a return for 1846, a year that figured in none of the contemporary surviving returns or in early post-1868 writing. As in the case of many other Tokugawa documents, what ensured the survival of source data (of which originals were either not retained or over time were lost or destroyed) was copying. Individual writers were aware of some documents, but the copying seems to have been almost random. As a result of this randomness, there can be no easy generalization about the process. It does however contradict the assumption of secrecy or inaccessiblity of the information in the first instance. Contemporary manuscripts (i.e., copies contemporaneous with the census years) are few or non-existent. The 1721 return survives only in a copy made ca. 1834; the 1750 and 1756 returns in a manuscript compiled in the 1770s that itself now survives only in a work published in 1879. Data for 1798 and 1822 survive in a source once held by the obscure Machida family in Shinano province, and which though undated may have been made at a date relatively close to the 1822 census. The 1828 census version surviving in a Wakayama manuscript nenpyō (chronicle) is by definition close to the date of the census. In the case of the data for 1840 and 1846, while they were probably copied in Tokugawa times, a date can not be determined. Losses by fire in Ansei 6 (1859) in the honmaru of Edo Castle must have included papers in the smaller of the Kanjōsho’s stores in the Castle. Plausibly, this store could have held the final results of the

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

123

1852 and 1858 censuses. The very existence of a census in these years has been held in doubt, but there is no concrete evidence either of censuses or of losses. There is no evidence of census data being preserved by officials in the 1850s and 1860s. The texts for 1804, 1834 and 1840 may have been copies in the hands of officials, but that is itself far from certain. There is, however, in Katsu Kaishū’s summary listing of census totals, tantalizing reference to possession of data for the kuni for 1732 by a kanjōsho official of late bakumatsu times named 131 Ishikawa Sōjirō ▼ᕝⲮḟ㑻. While Ishikawa may have had data for a census in a remote year, other former officials with whom Katsu was in regular contact had little to offer him, although they would have been aware that Katsu was working at the behest of the Ministry of Finance. In this scarcity of evidence of official survival lies strong confirmation that the copying and dissemination of data in Tokugawa times was largely a private process. The surviving documents may be the tip of a modest iceberg of copying. Census data were not unknown outside official circles, and more importantly they were no more secret than other documents. All Tokugawa documents were private (in the sense that the business of state was seen as private), but copies circulated. Thus for trade figures, the Tsūkō ichiran reproduces data available in Edo (often in compilations, and hence quite literally in copies of copies), and not in data freshly sent up from Nagasaki. The fūsetsugaki circulated, and more so in later times than earlier times. Despite Katsu Kaishū’s poor access to census data for the entire period after 1804, such existed none the less. The 1822, 1828, 1834 censuses – all three unknown to early Meiji writers – and the sole copy of the 1840 census have come to light only in relatively modern times. But for the discovery relatively recently by Minami Kazuo of the 1840 data and the survival of the 1846 data in a sole source – Katsu Kaishū’s compilation – we would conclude, in the absence of any other reference, that effective census taking halted not with the 1846 census (as Katsu argued) but with the 1834 census. The question remains open whether census taking continued beyond 1846. It is striking that in general fewer figures survive for the more recent censuses than for earlier ones. What this means is that documents after their creation quickly disappeared unless the slow process of private copying was allowed time to take effect. Survivals of census materials are better for earlier years than for the last fifty years of known census taking. At the national level, data remain unknown for 1812 and 1818. As official records suffered attrition at an alarming

124

EARLY JAPANESE TRADE

rate, survival of data very often depended on private copying. Thus, Katsu could offer data for only two years, 1804 and 1846, in the nineteenth century, although he had some details for earlier national censuses. Only with the progress of enquiry over the first sixty years of the twentieth century were some of the gaps covered – and then only from very heterogeneous sources. A rather similar story emerges for the shogunal cities. Though city registers were well maintained into the 1850s and probably beyond, none of the data, themselves all from highly miscellaneous sources, survive for Osaka beyond 1862; and in Edo they do so for a sole year. While private copies too, like official ones, were subject to loss, as the disappearance of material still in existence in 1890 shows, private copies alone have preserved for posterity some outline of the figures. This perusal of sources and their survival also suggests, firstly, that Katsu Kaishū’s access to demographic information was not particularly good, and secondly, that any assumption that information had become secret or more inaccessible is not convincing. It was the private copying, conscientious in some cases, cursory in other cases, operating moreover not necessarily from “originals” (copies made directly from documents in official custody) but quite literally from copies of copies, which more than any single factor ensured that any data survived into modern times. Within such a haphazard survival pattern, writers or copyists had access to data in a random fashion, often for years separated by large intervals. The real problem was less secrecy than the absence of tabular recording of data. Lesser data (han data) survive within han nisshi; before 1868, however, these were not abstracted into tables. While contemporaries worried about economic stagnation, they were working on impressions, rather than on officially recorded serial data. Such comparisons as exist, admittedly private rather than official, are random comparisons of censuses. In any event, as argued above the data had an inbuilt tendency to be static, because static registers of households, more than serious counts, determined the outcome in many instances. Ironically, in a statistical sense, the great crises of Kyōhō, Tenmei and Tenpō were worse that census figures suggest. Officials on the ground were of course well aware of the crises, and described these in alarmed reports which sometimes exaggerated the scale of catastrophe. There is no concrete evidence of the census data at either han or shogunal level playing a part in the gloomy conclusions of officials on longer trends. At best they must, like Matsudaira himself, have been aware of changes from six years before, or, like Isshiki, have had to compile their own personal

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

125

lists of the figures. Certainly population policies existed in some han, but they were not informed by statistical concerns. Ironically, modern writing on population, like that of Mori (who quoted han figures) and more directly Takahashi, has been influenced or even prompted by the 132 interest in Tokugawa population policies. The danger in this is that it makes it all too easy to take the statistical stability which seems to present itself in shogunal population data as a reality. In modern study of population, there has long been a search for evidence of infanticide, sometimes finding it and sometimes failing to find it. It may be well to end on a cautionary note from within Tokugawa Japan. In the prosperous Bunka period (1804–18), in counterpoint to evidence of Japanese authorities’ preoccupation with population policy, the Russian captive Golownin was told by an official that there was no official obsession 133 with inquiring into infanticide. APPENDIX Tokugawa Censuses: Sources and Coverage (Table includes sources documenting only gross totals as well as full-coverage censuses) Census Modern Source Date

Earliest Source Attribution or Date

Extent of BreakCoverage down by Sex

1721

Ōta, Chikkyō yohitsu besshū (ca. 1800)134

Kyōhō 20 (1735), fifth month

10 kuni

No

1721

Suda Akiyoshi, Tokugawa jidai jinkō (1954)135

“Kokuchū ninzū kokudaka no koto” in Sanka manroku (post-1834)136

All

No

1726

Katsu Kaishū, Suijinroku (1890), Katsu Kaishū zenshū, vol. 6 (1974 ed.), p. 294

Source not identified

No

1726

Ōta, Chikkyō yohitsu besshū (ca. 1800)137

1735 (summary reference to source)

No

1732

Ōta, Chikkyō yohitsu besshū (ca. 1800)

1735 (summary reference to source)

No

1732

Katsu Kaishū, Suijinroku (1890), Katsu Kaishū zenshū, vol. 6 (1974 ed.), p. 294

Source not identified, but existence of a full report implied138

Yes

126

EARLY JAPANESE TRADE

1732

Nihon zaisei keizai shiryō, vol. 9, p. 1246139

“Kinotomi zakki”

1738

n.a.

n.a.

1744

Partial survival of original publication in Gakugei sōdan, ed. Toki Takashi (1878)

Tsushima official Nishiyama Genbun, Kanchū hisaku (1770s)140

No

1744

Variant figure in Yokoyama 141

1750

Suijinroku (1890), Katsu Kaishū zenshū, vol. 6 (1974 ed.), p. 295

Source not identified

No

1750

Kurokawa, Gakugei sōdan (1878)

Kanchū hisaku (1770s)142

All

Yes

1750

Suzuki, Tōkyō keizai zasshi, no. 125 (1882), pp. 1090–1091

“Kan-en sannen okuni jinkō hyō”143

All

Yes

1756

Gakugei sōdan (1878)

Kanchū hisaku (1770s)

All144

Yes

1762

Suijinroku (1890), Katsu Kaishū zenshū, vol. 6 (1974 ed.), p. 295

Source not identified

No

1768

Same as above (Katsu, Suijinroku)

Source not identified

No

1774

Same as above (Katsu, Suijinroku)

Source not identified

No

1780

Same as above (Katsu, Suijinroku)

Source not identified

No

1780

Matsudaira Sadanobu, Uge no hitokoto145

Implied population total in comment on 1786 population

No

1786

Suijinroku (1890), Katsu Kaishū zenshū, vol. 6 (1974 ed.), p. 295

Source not identified

No

1786

Takimoto Seiichi, Nihon keizai taiten, vol. 48 (1930), pp. 169–188

“Tenmei kansei ninzū chō”146

1786

Matsudaira Sadanobu, Uge no hitokoto147

Comment on change since 1780

No n.a.

All

n.a.

Yes

No

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

127

1792

Suijinroku (1890), Katsu Kaishū zenshū, vol. 6 (1974 ed.), p. 295

Source not identified

No

1792

Matsuura Seizan (daimyo of Hirado), Kasshi yawa (ca. 1800)148

Source not identified149

No

1798

Suijinroku (1890), Katsu Kaishū zenshū, vol. 6 (1974 ed.), p. 295

Source not identified

No

1798

Matsuura Seizan (daimyo of Hirado), Kasshi yawa (ca. 1800)150

Source not identified

No

1798

Sekiyama, “Kansei 10-nen oyobi Bunsei 5-nen kunibetsu jinkō,” Keizai riron (1957)

“Shokuku ninzū chō” (Monbushō archives, ms formerly in private possession)

1804

All

Yes

Suijinroku (1890), Katsu Report with text of covKaishū zenshū, vol. 6 ering letter by ōmetsuke (1974 ed.), pp. 296–306152 and kanjō bugyō (“last year eleventh month”)

All

Yes

1804

Takimoto, Nihon keizai taiten, vol. 48 (1930), pp. 179–188

All

Yes

1810

n.a.

“Tenmei kansei ninzū chō” (has names of ōmetsuke and kanjō bugyō but lacks date; copy from Meiji period)153

1816

n.a.

1822

Nihon keizai taiten, vol. 54 (1930), p. 392

“Chōkaiki” in “Tokugawa rizai kaiyō”154

No

1822

Nihon zaisei keizai shiryō, vol. 4, pt. 2, p. 1193

“Ninbetsu aratame okanjōsho chōsho jō no utsushi”155

No

1822

Sekiyama, “Kansei 10-nen oyobi Bunsei 5-nen kunibetsu jinkō,” Keizai riron (1957)

“Shokuku ninzū chō” (Monbushō archives, ms formerly in private possession)156

All

Yes

128

EARLY JAPANESE TRADE

1828

Takimoto, Nihon keizai taiten, vol. 54 (1930), p. 392

“Chōkaiki” in “Tokugawa rizai kaiyō”157

No

1828

Nihon zaisei keizai shiryō, vol. 4, pt. 2, p. 1193

“Ninbetsu aratame okanjōsho chōsho jō no utsushi”158

No

1828

Inoue, “Dai Nihon korai “Kaei 6 ushi rōgetsu jinkō kō,” Tōkeigaku zasshi shirabe kōkoku sō (1904), pp. 7–9 ninbetsu yorichō”159

All

Yes

1828

Wakayama Daigaku Toshokan160

“Taihei nenpyō” (ms chronicle of Tokugawa times)161

All

Yes

1834

Shinozaki Ryō, in “Tenpō 5 umadoshi Tōkeigaku zasshi (1917), shokoku ninzū- chō”162 vol. 32, no. 369, pp. 14–15

All

Yes

1834

Takahashi, Nihon jinkōshi no kenkyū 2 (1955), pp. 332–343

“Tenpō 5-nen kōgō shokoku ninzū-chō “ (National Diet Library ōmetsuke and kanjō bugyō report163

All

Yes

1840

Minami Kazuo, Bakumatsu Edo, pp. 164–178

Shokoku ninzū chō in University of Tokyo Faculty of Law Library, Historical Materials Room (lacks date)164

All

Yes

1846

Suijinroku (1890), Katsu Kaishū zenshū, vol. 6 (1974 ed.), pp. 307–317

Report lacking covering letter and names of ōmetsuke and kanjō bugyō but giving date, Kōka 3-nen, twelfth month

All

Yes

REFERENCES

Cornell and Hayami 1986 L. L. Cornell and Akira Hayami. “The Shūmon Aratame Chō: Japan’s Population Registers.” Journal of Family History 11:4 (1986), pp. 311–28. Cullen 2002 L. M. Cullen. A History of Japan: Internal and External Worlds, 1582–1941. Cambridge University Press, 2003.

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

129

Deshima dagregisters 1990 The Deshima dagregisters, V, 1720–1730. Ed. Paul Van der Velde. Leiden: Centre for the History of European Expansion, 1990. Droppers 1894 Garrett Droppers. “The Population of Japan in the Tokugawa Period.” Transactions of the Asiatic Society of Japan, vol. 22 (1894), pp. 253–84. Golownin 1819 W. M. Golownin. Recollections of Japan. London, 1819. Hanley and Yamamura 1977 S. Hanley and K. Yamamura. Economic and Demographic Change in Preindustrial Japan 1600–1868. Princeton University Press, 1977. Hayami 2001 Hayami Akira. The Historical Demography of Pre-modern Japan. University of Tokyo Press, 2001. Hearn 1907 Lafcadio Hearn. Kokoro. Orig. pub. 1907. Reprinted in Writings from Japan, ed. Francis King. Harmondsworth: Penguin, 1984. Honjō 1972 Honjō Eijirō ᮏᗉᰤ἞㑻. Nihon shakai shi, Nihon jinkō shi ᪥ᮏ♫఍ ྐ᪥ᮏேཱྀྐ. In Honjō Eijirō chosakushū ᮏᗉᰤ἞㑻ⴭస㞟, vol. 5. Osaka: Seibundō, 1972. Honjō 1935 Honjō Eijirō. The Social and Economic History of Japan. Kyoto: Nihon Keizaishi Kenkyūsho, 1935. Hōreki gan hitsujidoshi yori Ansei san tatsudoshi made Sangō narabini etamura Hyōgo Nishinomiya Shiwaku-tō ninzūdaka-chō “Hōreki gan hitsujidoshi yori Ansei san tatsudoshi made Sangō narabini eta-mura Hyōgo Nishinomiya Shiwaku-to ninzūdaka-chō” ᐆᬺඖᮍᖺࡼ ࡾᏳᨻ୕㎮ᖺ㎾୕㒓ᖼ✧ከᮧරᗜすᐑሷ㣬ᓥேᩘ㧗ᖒ. In Isshiki monjo, Kōda Bunko, Hitotsubashi University, Qfq14, no. 2. Inoue 1904 Inoue Mizue ஭ୖ⍞ᯞ. “Dai Nihon korai jinkō kō” ኱᪥ᮏྂ᮶ேཱྀ ⪃, part 4. Tōkeigaku zasshi ⤫ィᏛ㞧ㄅ217 (May 1904), pp. 127–129. Iwate-ken shi 1963 Iwate-ken shi ᒾᡭ┴ྐ, vol. 5 (kinsei hen, pt. 2). Morioka: Toryū Insatsu, 1963.

130

EARLY JAPANESE TRADE

Katō and Toyama 1984 Katō Akira ຍ⸨❶and Toyama Mikio እᒣᖿኵ, eds. Waga machi no rekishi Nagasaki ࢃࡀ⏫ࡢṔྐ㛗ᓮ, no. 1. Bun’ichi Sōgō Shuppan, 1984. Kawaguchi 1996 Kawaguchi Hiroshi. “Population Increase Policy after the 1783 Great Famine in North-Eastern Tokugawa Japan.” Annales de demographie historique (1996), pp. 151–168. Reprinted in Eurasian Project in Population and Family History, no. 2 (May 1997), pp. 151–168 (original pagination retained). Kōda 1972 Kōda Shigetomo ᖾ⏣ᡂ཭. “Edo no chōnin no jinkō” Ụᡞࡢ⏫ேࡢ ேཱྀ. Orig. pub. in Shakai keizai shigaku ♫఍⤒῭ྐᏛ, no. 627, vol. 81 (1938). In Kōda Shigetomo chosaku shū ᖾ⏣ᡂ཭ⴭస㞟, vol. 2, pp. 244–265. Chūō Kōron Sha, 1972. McEwan 1962 J. R. McEwan. The Political Writings of Ogyū Sorai. Cambridge University Press, 1962. Matsuura 1978 Matsuura Seizanᯇᾆ㟼ᒣ. Kasshi yawa ⏥Ꮚኪヰ, vol. 6. Heibonsha, 1978. Minami 1978 Minami Kazuo ༡࿴⏨. Bakumatsu Edo shakai no kenkyū ᖥᮎỤᡞ♫఍ ࡢ◊✲. Yoshikawa Kōbunkan, 1978. Mito-shi shi 1968 Mito-shi shi Ỉᡞᕷྐ, middle series, part 1. Mito: Mito-shi Shi Hensan Iinkai, 1968. Mito-shi shi 1969 Mito-shi shi, middle series, part 2. Mito: Mito-shi Shi Hensan Iinkai, 1969. Mito-shi shi 1984 Mito-shi shi, middle series, part 3. Mito: Mito-shi Shi Hensan Iinkai, 1984. Mori 1934 Mori Kahee ᳃჆ර⾨. “Meiji shonen ni okeru Iwate-ken no ikushi seido” ᫂἞ึᖺ࡟࠾ࡅࡿᒾᡭ┴ࡢ⫱Ꮚไᗘ. Shakai keizai shigaku ♫఍⤒ ῭ྐᏛ4:1 (1934), pp. 73–84. Nihon zaisei keizai shiryō 1922a Nihon zaisei keizai shiryō ᪥ᮏ㈈ᨻ⤒῭ྐᩱ, vol. 1, pt. 1. Ōkurashō Zaisei Keizai Gakkai, 1922. Reprint Zaisei Keizai Shiryō Kenkyūkai, 1970.

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

131

Nihon zaisei keizai shiryō 1922b Nihon zaisei keizai shiryō, vol. 4, pt. 2. Ōkurashō Zaisei Keizai Gakkai, 1922; reprinted Zaisei Keizai Shiryō Kenkyūkai, 1971. Ōhashi 2000 Ōhashi Yukihiro ኱ᶫᖾὈ. “Kirishitan kinsei to shūmon aratame seido” ࢟ࣜࢩࢱࣥ⚗ไ࡜᐀㛛ᨵไᗘ. In Jūnana seiki no Nihon to higashi Ajia ༑୐ୡ⣖ࡢ᪥ᮏ࡜ᮾ࢔ࢪ࢔, ed. Fujita Satoru ⸨⏣ぬ. Yamakawa Shuppansha, 2000, pp. 69–99. Ōsaka hennen shi 1969 Ōsaka hennen shi ኱㜰⦅ᖺྐ, vol. 6. Ōsaka: Ōsaka Shiritsu Chūō Toshokan, 1969. Ōsaka-shi shi 1911–1914 Ōsaka-shi shi ኱㜰ᕷྐ. Ōsaka: Ōsaka-shi Sanjikai. Vol. 1 (1913), vol. 2 (1914), vol. 5 (1911). Ōsaka-shi shi 1990 Ōsaka-shi shi ኱㜰ᕷྐ, vol. 4. Ōsaka: Ōsaka-shi Shi Hensan Iinkai, 1990. Ōta 1976 Ōta Nanpo ኱⏣༡␇. Chikkyō yohitsu besshū ➉ᶫవ➹ู㞟, vol. 8. Kyūko Shoin, 1976. Saitō 2002 Saitō Osamu ᩪ⸨ಟ. Edo to Ōsaka: kindai Nihon no toshi kigen Ụᡞ࡜኱ 㜰㸸㏆௦᪥ᮏࡢ㒔ᕷ㉳※. NTT Shuppan, 2002. Saitō 2005 ———. “Pre-modern Economic Growth Revisited: Japan and the West.” Working Papers of the Global Economic History Network, Working Paper No. 16/05. London School of Economics, June 2005. Sekiyama 1957 Sekiyama Naotarō 㛵ᒣ┤ኴ㑻. “Kansei jū-nen oyobi Bunsei go-nen no kunibetsu jinkō: shin shiryō no hakken” ᐶᨻ༑ᖺཬࡧᩥᨻ஬ᖺࡢᅜู ேཱྀ㸸᪂ྐᩱࡢⓎぢ. Keizai riron ⤒῭⌮ㄽ37 (1957), pp. 61–74. Sekiyama 1958 ———. Kinsei Nihon no jinkō kōzō: Tokugawa jidai no jinkō chōsa to jinkō jōtai ni kansuru kenkyū ㏆ୡ᪥ᮏࡢேཱྀᵓ㐀㸸ᚨᕝ᫬௦ࡢேཱྀㄪᰝ ࡜ேཱྀ≧ែ࡟㛵ࡍࡿ◊✲. Yoshikawa Kōbunkan, 1958. Seya and Toyosaki 1973 Seya Yoshihiko ℩㇂⩏ᙪand Toyosaki Takashi ㇏ᓮ༟. Ibaraki ken no rekishi Ⲉᇛ┴ࡢṔྐ. Yamakawa, 1973.

132

EARLY JAPANESE TRADE

Shinozaki 1917 Shinozaki Ryō ⠛ᓮு. “Tenpō go-nen chōsa shokoku ninzūchō” ኳಖ஬ ᖺㄪᰝㅖᅜேᩘᖒ. Tōkeigaku zasshi ⤫ィᏛ㞧ㄅ(1917), pp. 14–15. Suda 1954 Suda Akiyoshi 㡲⏣᫛⩏. “Tokugawa jidai jinkō ni kansuru shin shiryō” ᚨᕝ᫬௦ேཱྀ࡟㛵ࡍࡿ᪂ྐᩱ. Jinruigaku zasshi ே㢮Ꮫ㞧ㄅ 63:4 (1954), pp. 26–28. Suijinroku 1890a Suijinroku ྿ሻ㘓, part 1, 1887. In Katsu Kaishū zenshū ຾ᾏ⯚඲㞟, vol. 6, Keisō Shobō, 1974. Suijinroku 1890b Suijinroku, part 5, 1887. In Katsu Kaishū zenshū, vol. 10. Keisō Shobō, 1978. Suzuki 1882 Suzuki Kentarō 㕥ᮌๆኴ㑻. “Kansei sannen okuni jinkō hyō” ᐶᨻ୕ᖺ ᚚᅜேཱྀ⾲. Tōkyō keizai zasshi ᮾி⤒῭㞧ㄅ125 (1882), pp. 1090– 1091. Takahashi 1955a Takahashi Bonsen 㧗ᶫᲙ௝. Nihon jinkō shi no kenkyū ᪥ᮏேཱྀྐஅ◊ ✲, vol. 1. Gakujutsu Shinkōkai, 1955; reprinted 1971. Takahashi 1955b ———. Nihon jinkō shi no kenkyū, vol. 2. Gakujutsu Shinkōkai, 1955. Takahashi 1962 ———. Nihon jinkō shi no kenkyū, vol. 3. Gakujutsu Shinkōkai, 1962. Takahashi 2003 Takahashi Nanako 㧗ᶫ⳯ዉᏊ. “Kōda Shigetomo no keizai-shi kenkyū to sono shiryō: Hitotsubashi daigaku fuzoku toshokan zō: Kōda bunko o chūshin ni” ᖾ⏣ᡂ཭ࡢ⤒῭ྐ◊✲࡜ࡑࡢྐᩱ㸸୍ᶫ኱Ꮫ㝃ᒓ ᅗ᭩㤋ⶶᖾ⏣ᩥᗜࢆ୰ᚰ࡟. Keizai shiryō kenkyū ⤒῭ྐᩱ◊✲ 33 (2003), pp. 29–43. Takagi 2004 Takagi Masao 㧗ᮌṇᮁ. “Zenkindai no jinkō chōsa: Sendai han ninzū aratame chō no seiritsu to tenkai” ๓㏆௦ࡢேཱྀㄪᰝ㸸௝ྎ⸬ேᩘ ᨵᖒࡢᡂ❧࡜ᒎ㛤. Ritsumeikan Daigaku Jinbunkagaku Kenkyū Sho, Kyoto, SDDMA working paper, 2004.

TOKUGAWA POPULATION: THE ARCHIVAL ISSUES

133

Takimoto 1930a Takimoto Seiichi ℧ᮏㄔ୍. “Tenmei Kansei ninzū chō” ኳ᫂ᐶᨻேᩘ ᖒ. Nihon keizai taiten ᪥ᮏ⤒῭኱඾, vol. 48 (1930). Takimoto 1930b ———, ed. Nihon keizai taiten, vol. 54 (1930). White 1992 James W. White. The Demography of Socio-Political Conflict in Japan, 1721– 1846. Berkeley: Institute of East Asian Studies, University of California, 1992. せ᪨ ᚨᕝᮇࡢேཱྀ̿ྂᩥ᭩ࡢၥ㢟Ⅼ̿ ࣝ࢖࣭࢝ࣞࣥ ᚨᕝᮇࡢேཱྀ⤫ィࢹ࣮ࢱࡣࠊ1958ᖺห⾜ࡢ㛵ᒣࡢ◊✲ࢆ➹㢌 ࡟ࠊከࡃࡢ◊✲࡛㞟✚ࡉࢀ࡚࠸ࡿࠋࡋ࠿ࡋྂᩥ᭩࡟グࡉࢀࡓ ᐇ㝿ࡢᩘᏐ࡛ࡣ࡞ࡃࠊྂᩥ᭩ࡢṧవ࡜ࡋ࡚ࡢྐᩱࡑࡢࡶࡢ࡟ ࡘ࠸࡚⣔⤫ⓗ࡟ศᯒࡉࢀࡓࡇ࡜ࡣ࡞࠿ࡗࡓࠋᖥᗓ࡜᫂἞ᨻᗓ ཮᪉࡟௙࠼ࡓ຾ᾏ⯚ࡀ1890ᖺ࡟ฟ∧ࡋࡓࠗ྿ሻ㘓࠘࡟◊✲⪅ 㐩ࡣྐᩱ࡜ࡋ࡚ࡢඃ఩ᛶࢆㄆࡵࠊࡲࡓᚨᕝࡢຠᯝⓗ࡞ᶵᐦᨻ ⟇ࡢ⤖ᯝࠊ᝟ሗࡢὶ㏻ࡣḞዴࡋ࡚࠸ࡓ࡜ࡢ௬ᐃ࡟❧ࡗ࡚Ṕྐ ⓗ࡞ேཱྀㄪᰝࡀ㐍ࡵࡽࢀ࡚ࡁࡓࠋ⸬࡟࠾࠸࡚ࡣᑗ㌷࡬⏦࿌ࡉ ࢀࡓேཱྀㄪᰝࡀ࠸࠿࡟⦅⧩ࡉࢀࡓ࠿ࡀ୍㒊▱ࡽࢀ࡚࠸ࡓࠋࡋ ࠿ࡋᖥᗓഃ࡟ࡼࡿㅖ⸬ࡀᥦฟࡋࡓㄪᰝሗ࿌ࡢྲྀࡾᢅ࠸࡟ࡘ࠸ ࡚ࡣ඲ࡃ▱ࡽࢀ࡚࠸࡞࠸ࠋ⸬ࡢࢹ࣮ࢱࡀᅜ༢఩࡬᥮⟬ࡉࢀࡓ ௚ࡣࠊᖥᗓࡢᙺே࡟ࡼࡿㄪᩚࡢ㊧ࡣぢࡽࢀ࡞࠸ࠋࡑࢀࡀࠊᮍ ⦅㞟ࡢ⏦࿌࡛࠶ࡿ⏕ࡢࢹ࣮ࢱ࡜ඖࡢㄪᰝ⠊ᅖࡢࣈࣞ࡟㛵ࢃࡿ ୙☜ᐇᛶࡀಖࡓࢀࡓཎᅉ࡟࡞ࡗ࡚࠸ࡿࠋ኱┠௜࡜຺ᐃዊ⾜ࡢ ⨫ྡ࡟ࡼࡿᩘᖺ࡟ཬࡪ⌧Ꮡࡍࡿಶࠎࡢㄪᰝࡢせ⣙⾲ࡢ࠺ࡕࠊ ಙ៰ᛶࡀ࠶ࡿ࡜⪃࠼ࡽࢀ࡚࠸ࡿࡢࡣഹ࠿㸲ࡘ࡛࠶ࡿࠋࡑࡢ ୖࠊ⌧Ꮡࡍࡿ⏦࿌᭩ࡣࠊබᩥ᭩࡛ࡣ࡞ࡃ㸦࠶ࡿ࠸ࡣᙺே࡟ࡼ ࡗ࡚సᡂࡉࢀࡓබᩥ᭩࡜ࡋ࡚ࡢᶵ⬟ࢆᯝࡓࡉࡎ㸧ࠊࡑࡢከࡃ ࡣಶேࡀᡤ᭷ࡋ࡚࠸ࡓ」෗࠿ࡽࡉࡽ࡟ಶேⓗ࡟」෗ࡋࡓࡶࡢ ࡛࠶ࡿࠋࡇࢀࡽࡢ⏦࿌ࡢ」෗ࡣ㏻ᖖᡈࡿ༢୍ࡢேཱྀㄪᰝ࡟ᇶ ࠸࡚࠾ࡾࠊ」ᩘࡢேཱྀㄪᰝࢆྵࢇ࡛࠸ࡿࡢࡣഹ࠿㸳ࠊ㸴౛ࡢ ࡳ࡛࠶ࡿࠋ᫬⣔ิࡢேཱྀㄪᰝࡣ▱ࡽࢀ࡚࠸࡞࠿ࡗࡓࠋ୍Ⰽ࡜ ࠸࠺ྡࡢ⏫ዊ⾜ࡀேཱྀⓏ㘓࠿ࡽᵓ⠏ࡋࡓ኱㜰ࡢேཱྀࡢ㛗኱࡞ ⾲ࡣ௚࡟㢮ࢆぢ࡞࠸ࡶࡢ࡛࠶ࡿࠋ

Source: Japan Review, No.21 (2009), pp. 183–223

7 Coastal Trade

Statistics of Tokugawa Coastal Trade and Bakumatsu and Early Meiji Foreign Trade PART 1: COASTAL TRADE IN TOKUGAWA JAPAN

™

Evaluation of archival sources on Japanese trade in the Tokugawa period and the first fifteen years of the Meiji period is the primary purpose of this two-part essay, of which Part 1 appears on the following pages and Part 2 will appear in Japan Review, Number 22. Surviving Japanese statistics of coastal trade (or “inland” trade, as opposed to foreign trade) are remarkable. In Europe, comparable figures were not compiled. For Osaka, some count appears to have been kept of a variable number of import items, and probably also of six sensitive “exports,” that is, items shipped from Osaka to Edo. Counting in 1714, 1736, and 1766 yielded aggregate figures of the total trade of the port – rare statistics today, for other totals have not been passed down to us. For Edo, a consumption center, figures for outward shipping were never compiled. The only known coastal trade statistics for Edo are for the years 1724–1730, during which eleven items were recorded in Osaka (outwards) and in Edo (inwards), 1726 and 1856, for which a massive return survives. As for foreign trade, after the opening of the ports in 1859, statistics were directly compiled by bakufu officials from traders’ invoices. Previously the officials had relied on data furnished by tonya and guilds for coastal trade. 134

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

135

PART 1. COASTAL TRADE IN TOKUGAWA TIMES

1. Trade Statistics

Japanese trade is unusual in terms of its statistical documentation. In a mineral-rich country, balance of trade considerations in the European mercantilist sense did not arise. Some record of foreign trade was maintained only because of the fear that without control foreign demand would carry away the precious metals required for the domestic currency. But this was more a process than a rigorous general quantitative counting, and the surviving details reveal that there was little if any creation of general aggregates. Those details often exist only in Dutch records.1 What survives of the work of the Nagasaki Kaisho 㛗 ᓮ఍ᡤ, a bakufu office in Nagasaki for trade matters that in 1715 was expanded and given the added function of buying and selling commodities, consists of business accounts rather than a quantitative count of trade itself.2 In Europe, if statistics were prompted in part by an urge to measure the balance of trade (in which a deficit would have to be settled by precious metals from abroad), what made possible the remarkably detailed record of trade was the fiscal system. In contrast to Japan, indirect taxation (customs and excise) was highly developed in Europe, and hence revenue collecting generated detailed information,3 which could easily be harnessed to provide statistics of trade. Statistics began to make an appearance from the outset of the seventeenth century; after 1700 they had become full-blown trade ones. If Japanese figures for the limited foreign trade that was possible under sakoku were inferior, the record of coastal trade, on the other hand, though only part of it has survived, was virtually unique. The figures of coastal trade are all the more striking as they had lacked fiscal impetus. They reflect three circumstances: the central trading importance of Osaka, the rising consumer market of Edo, and the scale of trade in sensitive commodities between these two dominant ports. The coastal trade was probably the largest in the world. In a shadowy way the bugyō ዊ⾜or magistrate of Osaka was a central figure, acting on prompting from Edo. The thin surviving documentation, though richer for Osaka than for Edo, however provides no conclusive evidence of the reasons for these statistical exercises or how some of the detail survived. The relative scale of coastal and foreign trade differed between Europe and Japan. In the West the foreign traffic greatly exceeded coastal traffic, more particularly in value. In Japan the coastal traffic (“inland” trade in the technical parlance of English) exceeded foreign trade both in

136

EARLY JAPANESE TRADE

volume and value, and with the contraction of foreign trade after 1685, became ever more the dominant force. By contrast in Europe there is little record of coastal trade despite the fact it was closely watched to prevent its becoming a vehicle for illegal transfer of taxed goods. At its peak around 1661 the foreign trade of Nagasaki was worth about 50,000 kanme of silver; it halved to about 22,000 kanme in 1685. It continued to decline thereafter, and was perhaps a mere 5,000 kanme in the early nineteenth century.4 By the late 1840s, as few as four or five Chinese vessels a year entered Nagasaki; only three came in 1859.5 There was after 1791 a single Dutch vessel per year from Batavia. The trade through the Ryukyus increased the volume of trade somewhat, but did not greatly alter the picture. Converted into sterling, an export trade of 10,000 kanme in the late seventeenth century was worth about £233,000 sterling. Compared with the exports of Ireland at the time, which at 800,000 Irish pounds were relatively high for a small island economy, at first sight this might suggest a large turnover for the business of a single port.6 However the number and tonnage of shipping was small. It was a high-value exchange of silver, gold, and copper for silk and a small number of others goods. In other words, the Nagasaki trade was inherently an unstable one, dependent on high mining output in Japan and on a uniquely large trade in expensive silk from China, triggered by Japan’s high prosperity and rapid expansion. In time Japanese mining output would fall, and the growth of a large market for silk would prompt expansion in indigenous production. Decline was inevitable, that is to say, and with it a fall in the population of Nagasaki which had grown in the expansive seventeenth century. Nagasaki became progressively a commercial backwater. Even in the open Japan of 1859, these drawbacks could not be overcome: recovery in the early 1860s did not give the port a new momentum. Within the decade the thrust of growth shifted decisively to Yokohama and Osaka, and Nagasaki’s relative place receded further. The monitoring of Osaka’s trade for 1714 and 1766 affords a measure of the vast scale of its coastal trade. Its combined exports and imports were estimated in greatly inflated prices at 382,361 kanme in 1714; at lower, more realistic prices, a total of 271,145 kanme for 1766 suggests that Osaka held or more than held its own. The sheer volume of the trade can be conveyed in comparative terms: it was about half the size of the export and re-export trade of England. It is hardly surprising in these circumstances that the fleet which carried the coastal trade of Japan was probably almost as large as the entire fleet engaged in both

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

137

the coastal and foreign trade of England and Wales.7 This was the case, moreover, despite the huge tonnage of coal carried from Newcastle to London. An intense use in multiple voyages of this hardy and efficient fleet meant a relatively small fleet in proportion to the huge and growing tonnage of coal. As a result Japanese shipping, much less economically employed, was not much short of the total English tonnage. As coal was low-priced, the comparison of the two coastal trades in value would have remained overwhelmingly to Japan’s advantage. Edo and Osaka alike were approached by shallow channels drawing as little as two feet at low water. Goods were enumerated in both Edo and Osaka at the wharves, but counts of vessels were more complex. For Edo some goods were transferred to smaller craft at Uraga; for Osaka some goods were transshipped at Sakai. We cannot be sure that it is possible to compare like with like. The number of vessels entering Edo was inflated by transfer of goods from ocean-going vessels to smaller craft at Uraga. There was rivalry between Edo and Uraga tonya in the dispatch of goods for Edo bay.8 In Osaka enumeration, in the anchorage outside the bar in the Yodogawa ᾷᕝ, vessels below 200 koku, which could sail at high tide to enter the Kizugawa ᮌὠᕝand Ajigawa Ᏻ ἞ᕝchannels without offloading their cargoes into lighters, were not counted. The number of such vessels was added to by small craft sailing from Sakai, which, with much better anchorage facilities, served as an outport to Osaka for large vessels, especially from the Hokuriku. In a crude comparison, the number of ships was much higher in Edo than in Osaka: it was 7424 in 1726, and 7741 in 1871. The figures usually given for Osaka are doubtful. Counts of 2,000–3,000 ships are far too low (rice alone would have required 1000 to 2000 vessels).9 Perhaps 4000 to 5000 ships passed through the offshore anchorage of Osaka.10 There is no doubt however that, as Edo’s population climbed to a million people in the course of the first half of the eighteenth century, the total of shipping at Edo, especially if rice is included, greatly exceeded Osaka’s total. Crude counts of shipping of course do not of themselves measure precisely the commercial vitality of ports. While in Edo, primarily a center of consumption, vessels arrived with cargoes, they mostly departed in ballast. Osaka on the other hand received goods, sent outwards the product of its hinterland, and especially redistributed towards the south goods received from Ezo, and dispatched northwards goods received from Nagasaki. If the rice trade of Edo reached the high figure of 3,000,000 koku suggested for 1856, it would have accounted for more than 3,000 of the 7,000-plus vessels that arrived each year.11

138

EARLY JAPANESE TRADE

A decline in the population of Osaka from a peak in the mideighteenth century may suggest that the volume of trade in Osaka became relatively static. The trend is not however clear. The city’s rice trade rose in the prosperous Bunka and Bunsei period. Special pleading in some sources also made an exaggerated case for decline. Yamagata Bantō ᒣ∦ⶽ᱈(1748–1821), an advisor to many of the business houses maintained in Osaka by han, made the case for their being losers (though his argument was not presented in statistical terms). Bugyō Abe Shōzō㜿㒊ṇⶶ, in pleading the case of the Osaka guilds ᰴ௰㛫 in 1842, did make a statistical case for decline. Seeking to defend the city’s guilds, which had long been accused by han interests of monopolizing tendencies and which were disbanded in the same year, he maintained that the guilds had themselves been victimized for decades by commodities being withheld by outside interests from Osaka market. However, whatever the trends, which for want of statistical evidence remain obscure, any problems in trade were more than compensated for by the enhancing of the financial business of Osaka. Even Tōhoku domains banked there, and their borrowings were collateralized by their growing rice trade to Edo. With its multiple ties and warehouses maintained by more than 100 han, Osaka has in effect served in modern studies as a surrogate for the study of the coast trade of Japan. Even though usually little rice went from Osaka to Edo (but with some change in Bunka/Bunsei times), the fact that Osaka drew rice both from the West and much further afield from the Hokuriku made it the recognized barometer of the national rice market. Osaka also either produced or supplied goods, notably sake, oil, finished cotton cloth or the intermediate product of ginned cotton, and redistributed silk, sugar and goods from Nagasaki. Osaka for itself and its hinterland was a heavy importer of oil and rice, and from the eighteenth century also of goods, from Ezo. Osaka or Sakai channeled copper, the vital support which alone kept the Dutch and Chinese trade in existence, to Nagasaki. Pharmaceuticals were another commodity of consequence. Though details of quantities imported in Nagasaki in 1735–1820 are lacking, information on the wide range of items, prices and the redistributive role of Osaka is abundant.12 The histories of Osaka, the first Ōsaka-shi shi ኱㜰ᕷྐ, the Ōsaka-fu shi ኱㜰ᗓྐ, and the Shinshū Ōsaka-shi shi ᪂ಟ኱㜰ᕷྐ, provide a good framework of information on its trading history. Monographs by Ōishi, Wakita, and Miyamoto explore some of its main dimensions. Wakita and McClain have an extremely useful book on the trade of Osaka, containing in

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

139

particular an article by McClain on the trade of the port.13 In English there are also very useful articles on the trade of Edo, on Osaka, and on the shipping along the Japan Sea coast of Japan (including Kaga).14 From Edo and Osaka, routes reached out to other ports. In particular Kaga and Echigo in the Hokuriku, both with fertile hinterlands, had a major rice trade to Osaka. They were, after Edo and Osaka, the most important single centers of coastal trade. By Genroku times (1688–1704) Kaga was providing 250,000 koku – roughly 20–25 per cent of Osaka’s intake of rice.15 In 1870, according to information from a British consul, 3586 vessels entered Niigata, the main port in Echigo.16 The most obscure sector of the trade is that to the coasts to the north of Edo. In some respects this is because detail, quantitative or otherwise, is so sparse for Edo, in contrast to Osaka. As the intake of rice is rather poorly documented at the Edo end, records of the han to the north are essential to providing some idea of the trade. Hayashi Reiko has shown that Edo itself had some redistributive role in sending goods arriving from the south to the north. For the coastal trade at large, there is no general study.17 However the difficult terrain of the Japanese interior meant that goods went by sea to a much greater extent than in countries of Europe or of Asia. A problem for such a study is that information is discontinuous, heterogeneous, and scattered, and it does not lead to an easy composite picture. It is also rarely quantitative. While conflicts of interest between han managers and buyers of their wares in Osaka were endemic,18 as can be seen from Mark Ravina’s account of Tokushima trade, the sources for these conflicts lack a statistical content. Prefectural and city histories are none the less useful. A source of particular value is the journal Kaijishi kenkyū ᾏ஦ྐ◊✲ (Studies in Maritime History), which has had articles over many years on both trade and shipping. The extraordinary role of Osaka as a nerve center of much trade and of financial support to the shogunate explains the importance of the bugyōsho ዊ⾜ᡤof Osaka, and also why the bugyō was either consulted or on occasion directed by the Edo machi bugyō ⏫ዊ⾜ acting for the shogunate. Edo officials seem to have been passive in documenting the inflow of goods to Edo, and were often content to rely on detail from Nagasaki (for foreign trade) or Osaka, where the bugyōsho acted as a powerful agency of the shogunate. The nexus between shogunate and Osaka bugyōsho is rather obscure today, given the shortage of documentary proof. For the few cases where real statistical information survives concerning the trade of Edo, the role of Edo officials

140

EARLY JAPANESE TRADE

in instituting the record keeping can be either identified or inferred. The major example of this, important in the context of shogun Yoshimune’s policy of protecting income of shogunate and samurai alike by either raising the price of rice, or by taking action against forestalling in the hopes that falling prices of other commodities would increase real incomes, can be seen in the unique exercise of recording trade in Osaka and Edo in eleven key commodities in 1724–1730, for which the central figure was the Edo machi bugyō Ōoka Tadasuke ኱ᒸᛅ┦ (1677–1751). The list of goods imported to Edo for 1856 reflects concern at a later date. However such an extraordinarily detailed exercise in 1856 would seem to imply that ready information in Edo itself on the city’s trade did not at the time exist, and hence an exercise conducted on a grand scale and in secrecy was called for. All ports, big or small, had their own hinterlands. With Japan’s mountainous spine and only short routes from the interior to ports, road carriage between ports and inland centers played a vital role. While the existence of these links is well known, for instance the so-called salt roads,19 it is rarely possible to document them in quantitative terms. However an estimate exists for Iida in the southern Japanese Alps, where in the course of a dispute over carriage rights a detailed count of movement of goods on packhorses over the year 1763 was made and has survived.20 The recording of trade depended on the functioning of bureaucracy. Administration was divided between the rōjū ⪁୰ (in effect a cabinet), the Kanjōsho ຺ᐃᡤ (Finance Office, mainly concerned with management of the shogun’s estates, but also handling economic issues), and the machi bugyō (magistrates or intendants of towns). The bugyō can be distinguished as kanjō bugyō ຺ᐃዊ⾜ (senior officials in the Finance Office) and the machi bugyō (urban magistrates). Bugyō, though occasionally already holding the rank of kanjō bugyō, were more typically men promoted from the ranks of the versatile and experienced metsuke ┠௜(inspectors), and machi bugyō were usually promoted at the end of a period of office to the grade of kanjō bugyō. Coastal trade was primarily a concern of the machi bugyōsho, and we rarely have evidence of circumstances in which Kanjōsho, rōjū, or shogun intervened directly. Under bugyō, the workhorses were the machidoshiyori ⏫ᖺᐤwho collected trade data from the wholesalers (tonya [or toiya] ၥᒇ). As far as trade was concerned only a small number of officials was necessary. This is in contrast with Europe where trade required a vast and decentralized bureaucracy, next to the army in size and importance as an arm of the state. The trade bureaucracy functioned to collect taxes, compile

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

141

statistics, survey all trade, coastal and foreign alike, and to crush evasion. In Japan, on the other hand, the groundwork for statistical information had not even to be done by the bugyō’s officials. The information was provided by the tonya or less often used by the guilds in lobbying the authorities.21 Certified by machidoshiyori, the details went to machi bugyō. Did they go further? They did in 1724–1730 when a run of low rice prices – the lowest in a century – threatened the solvency of shogun and samurai alike. The remarkable statistical exercise of 1724–1730 revolved around Ōoka, appointed in 1717 by Yoshimune as the machi bugyō of Edo. Given his long career (twenty years) in that post and the central role he played in implementing the bakufu’s rice policy,22 it seems certain that these statistics came to the personal attention of Yoshimune, who on other evidence was also Japan’s only statistically minded shogun.23 However, no specialized sub-grouping which had a statistical competence was ever created. The task seems to have depended on kanjō bugyō, whose own duties changed in a monthly rota, and machi bugyō whose role was often short-lived as in the short term they worked under a form of sankin kōtai ཧ໅஺௦, and in the longer term were usually replaced at fairly regular intervals by new officers. Such a pattern itself was likely to leave regular record keeping at risk. Statistical material was prized only in the context of a current problem. As a result, as far as can be judged, antecedent or historical information was hard to get. Under Arai Hakuseki ᪂஭ⓑ▼(1657–1725) during the reign of shogun Ienobu ᐙᐉ(1709– 1712), when longterm currency trends were a preoccupation, details of trade do not seem to have been readily available in Edo and had to be acquired from the bugyō in Nagasaki.24 Many years later, when Hayashi Fukusai ᯘ ᚟ᩪ (also known by his posthumous name Hayashi Akira ᯘ㡐, 1800/1–1859) and his team put together the Tsūkō ichiran ㏻⯟୍ぴ, they drew on heterogeneous sources and on a random corpus of books or nikki ᪥グin Edo, much of it the fruit of private copying. The Osaka bugyō did have access to statistical information, as the Abe report of 1842 and later contact in 1867 of a bugyō with a British official show. However, this information was not retained in any systematic fashion, and bugyō at times seem to have sought information with limited success. Thus, in a document drawn up for the bugyō by the nengyōji ᖺ⾜ྖ of an Osaka guild for ginned cotton, the information covered some but not all months of the two years 1766 and 1788, and even for the year immediately preceding compilation in 1797, some months were lacking.

142

EARLY JAPANESE TRADE

Edo is much more poorly documented than Osaka. Whether this is a random consequence of survival or a result of a more active role by Osaka bugyō (whose own records have had a poor survival history) is an unanswerable question. Post-1868, Katsu Kaishū ຾ᾏ⯚appears, on the evidence of his compilations, to have been able to gather little information on trade, inland or foreign, when at official instigation he compiled his collection of Tokugawa records, despite the fact that he had been a prominent official both before and after 1868. Independently of Katsu’s labors, officials in the finance ministry prompted the start of an effort to compile documents in a more rigorously academic fashion. The Nihon keizai taiten ᪥ᮏ⤒῭኱඾, in its origins a byproduct of this record-compilation exercise, has a single document with population figures (defective and imprecisely dated), and virtually no trade figures of consequence. The finance ministry, central though its role was to early Meiji trade statistics, at first repeated the casual practice of shogunal times. The first decade of its existence left what was regarded even at the time as a very scanty and confused amount of statistical information on trade. The documentation of the coastal trade raises questions about both the nature of record keeping and survival of records. The purpose of this paper is not to study trade; it is to examine the statistical and archival questions of its documentation. It is salutary at this stage to remember that the figures for Osaka trade in 1714 rest on three later copies of a 1903 copy (itself no longer in existence) which had been transcribed from earlier and now unknown records. For trade in 1766, we have only the vestige of the basic record, the mere aggregates of imports and exports, and we lack the detail which made the grand totals themselves possible. What there is, is a rather random survival of information in private more than public copies, probably a mere tip of an iceberg of data imperfectly accessible to any collector, even to a Meiji figure like Katsu. Trade presents a contrast with the population figures, which had formally to be reported to the rōjū, a process that increased the amount of copying and diffusion, with the result that at least part of the record became accessible to copyists, some of whom were serious, some, mere dilettantes. For coastal trade, focussed narrowly on an inner circle of officials in the machi bugyōsho of Edo and Osaka, paper circulated less and copying was narrower and less colorful. Sadly a dilettante interest in coastal trade had little opportunity to come into existence. Whether returns existed for other years but have not survived is a tantalizing

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

143

question, but it is ultimately unanswerable from the meager record left to us today. Japan is remarkable in having had statistics of coastal trade, quite apart from their uneven quality and limited survival. The first section of this paper is an examination of the obscure process of collecting such statistics from 1714 to 1868. As for the emergence of modern-style statistics of foreign trade in 1859–1882, the concern of the second section, the statistics, although different from figures of coastal trade, were compiled first in ports that with the exception of Nagasaki had no experience in foreign trade; at the outset the figures were collected by existing bugyōsho officials. For the first time in the history of Japanese statistics, the tonya had no role, and the bureaucrats had a new task, even if at first they themselves were unchanged. It also included entirely novel circumstance of collecting the customs revenue on foreign trade authorized by the treaties. Nagasaki itself fared poorly in the new circumstances. The port and its great office were mere shadows of their former selves, and Nagasaki was the slowest of Japan’s ports to adjust to the new demands in the first years. The process of creating statistical returns in those years illustrates both the limitations and the adaptability of Japanese officials. Insofar as it is a success story, which on balance one has to judge it, it also shows that the bureaucracy remained vigorous. Japanese bureaucrats continued to compile demographic statistics in the 1850s, and, despite having no real base of past training, created a successful run of trade statistics during the difficult years of the 1860s. This conclusion would also tend to support the argument that in a bureaucratic sense, at least, the shogunal administration was not near collapse in bakumatsu times. 2. Surviving Records and Their Use by Historians

For Edo, the very heart of Tokugawa power and its largest consumption center, remarkably little statistics survive of the trade of its port. There are good details for imports for two years only, 1726 and 1856, plus data for rice in one year, 1727. Otherwise just a few isolated figures remain available to us. Osaka, on the other hand, fared relatively better. Regrettably many documents that were still available c. 1900 no longer seem to exist, but two remarkable editorial projects in the first years of the new century ensured that some detail could survive. The first was the work from 1901 of the Ōsaka-shi Shi Hensan Gakari ኱㜰ᕷྐ⦅⧩ಀ, the history editorial section created by the Osaka

144

EARLY JAPANESE TRADE

city council and precursor of the future Osaka City Historiographical Center (Ōsaka-shi Shi Hensanjo ኱㜰ᕷྐ⦅⧩ᡤ). Borrowing documents to copy (the originals were returned to their owners), the staff of this editorial body provided the sources for the Ōsaka-shi shi, which appeared in five volumes of text (eight tomes) between 1911 and 1915. This is a key work in more senses than one: it was the first history of a major Japanese city and was based exclusively on documentary sources. The second archival project, dating from 1899, with no obvious direct interaction with the first, was launched by the Osaka Chamber of Commerce and Industry (Osaka Shōkō Kaigisho ኱㜰ၟᕤ఍㆟ ᡤ, hereafter OCCI). It collected a vast amount of archival material within a short span of years, the duration of which we now do not precisely know. Once stored in the offices of the OCCI, except in a few very exceptional circumstances, the documents became totally inaccessible to historians.25 This inaccessibility provided the motive for a project for publication, in a massive thirty-five volumes (Ōsaka shōgyō-shi shiryō ኱㜰ၟᴗྐ㈨ᩱ, Sources for the History of Osaka Commerce and Industry, hereafter abbreviated to OSSS) that appeared mostly in 1964, with the last few volumes appearing in 1966.26 A supplement (Bekkan kaisetsuhen sōmokuji ูᕳゎㄝ⦅⥲┠ḟ, hereafter referred to as OSSS Supplement 1966) with a series of editorial analyses and general list of contents was among the volumes released in 1966. The editorial work on the OSSS was carried out by a team of at least four historians who wrote the series of analyses within the volume of the contents of the series at large.27 Documents seem to have been copied out afresh (some five or six individual hands are readily identifiable, and the handwriting all appears to be postwar), then photocopied and published in sequence in thematic volumes. The individual documents were not numbered, however, and hence it is not always clear where the text of one document ends and another begins. The volumes also lack any indication of the source of individual documents, whether they are originals or just copies, or whether the copies archived in the chamber contain any marginal annotations that might suggest their origins.28 Three individuals played an especially important role in the study of the statistics of trade. Foremost was Kōda Shigetomo ᖾ⏣ᡂ཭ (1873–1954), a graduate of Tokyo University who had studied under Ludwig Riess (1861–1928), a German professor who was one of the father figures of modern Japanese historiography. Twenty-eight years old at his arrival in 1901 in Osaka,29 Kōda headed the project until finalization of the manuscript in his hand in 1909. The next year

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

145

he returned to Tokyo to begin a career of teaching and research at Keio University and Hitotsubashi University. The second was Yasuoka Shigeaki Ᏻᒸ㔜᫂(1928– ), a member of the editorial team working for the OCCI in preparing the OSSS. Yasuoka’s first contribution to the study of trade (published in 1960, well before the appearance of the supplement of 1966 on which he worked) included what purported to be a comparative table of imports for 1736, 1804–1830, and 1840. Many years later, in 1990, he published for the first time hitherto neglected aggregates of Osaka trade in 1766, drawn from the OSSS. The third figure is Ōishi Shinzaburō ኱▼ៅ୕㑻 (1923– 2004), whose contribution is twofold: a close analysis of figures for 1714, which though vaguely known had hitherto been obscure (these figures became readily accessible only in 1964, when OSSS, vol. 13, appeared), and a penetrating study of eleven commodities exchanged in Osaka-Edo trade in 1724–1730.30 One should add a further name, that of Miyamoto Matao ᐑᮏཪ㑻(1943– ), for his magisterial study of the rice trade of Osaka, which explored the rice trade as far as the original sources permitted.31 Table 1. Major Statistical Information on the Trade of Osaka and Edo* Port

Coverage (quantities recorded)

Additional information

1714

Osaka

exports and imports

values and aggregate totals

1724–1730

Osaka

11 export items to Edo

1726

Edo

11 import items from all Japan

1736

Osaka

imports

values given but not aggregated

1766

Osaka

no detail

aggregated values only survive

1804–1840

Osaka

16 import items

for rounded or isolated dates only

1856

Edo

imports

1856–58

Osaka

imports from Ezo

1858–64

Osaka

6 export items to Edo

1866

Osaka

20 import items

* Excluding kura rice (except for Edo 1726, 1856)

146

EARLY JAPANESE TRADE

Apart from the run of eleven selected exports from Osaka for 1724– 30, there are returns for the years 1714, 1736, and for 1766. Totals in Osaka of exports or imports or of both in 1714, 1736, and 1766 seem unprecedented among surviving data in that they clearly attempted to be comprehensive: commodities were both enumerated and valued, and gross totals were derived (at least for 1714 and 1766; arguably they were calculated for 1736 as well, but only a truncated document containing aggregate totals survived). Figures for the year 1714 may reflect the concern which led the bakufu to revalue the currency in 1715. Possibly they were intended to be a measure of the solvency of the Osaka business community, and hence of its ability to support daimyo and shogunal finances. There is of course a problem in that while official concern was with rice – the low price of which was depressing the income of state, daimyo, and samurai alike – the 1714 returns were by design not informative on the total rice trade. They included trade rice (nayamai ⣡ᒇ⡿) only, and omitted the much larger business in daimyo rice, or kura rice (kuramai ⶶ⡿, also described as bukemai Ṋᐙ⡿, sometimes called han rice).32 Kuramai was seen in different terms from trade rice, and was a relatively known quantity. A trawl of han sources affords further information at local level on rice shipments. Rice was doubly important: it engaged a large amount of shipping and in value was worth a half or more of the total for other imports in Osaka. Trade rice, although the quantity of it was relatively small compared with kura rice, was in value the single largest item in tables of imports. Sold commercially within han (and sometimes described as peasant rice, though it could originate in sales from daimyo stores in the han to local shippers), this rice was purchased by merchants and shipped on their own account to Osaka. It was mostly from the west of Japan and from the Hokuriku. What utility the precedent of the table for 1714 and the large amount of work which its preparation involved had for subsequent years is merely conjectural. Figures for 1736 could conceivably have reflected the perception of the economy that prompted Yoshimune’s currency debasement that year; the statistical compilation thus may be an exercise consciously modeled on the 1714 one. The figure for 1766 is even more obscure, and lacks an immediate obvious reason. All these documents were little known, as the meager paper trail for 1714, 1736, and more strikingly for 1766 suggests. They also totally lack accompanying evidence and are not supported by other sources setting out the reasons for attempting these ambitious compilations. It is likely that the

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

147

exercise was a rarefied one confined to senior officials, lacking a sense of an ongoing permanent administrative purpose and hence not widely known outside such circles. That, however, was manifestly not the case with the data for 1724–1730. Copies of the figures for those years survive in both Osaka and Edo, and the surviving documentation is informative on the purpose and on the conduct of the operation. The return of imports to Edo in 1856 is of similar character in that its remarkable detail bears out the delicate official purpose suggested by its title. For years other than those covered by Table 1, apart from the special case of rice, there is only scattered information for individual items, usually for isolated years. A comprehensive listing of the dates and sources for them for Edo and Osaka is set out in the Appendix. While oil may have ranked next to rice in importance, as Kōda observed, figures in both Osaka and Edo for the trade are few. Likewise for kuriwata ⧞⥥(ginned cotton), a major item in both exports and imports, outside the years set out in Table 1 above, figures for exports exist for only a few months of the years 1761, 1788, and 1796. Despite this paucity of surviving data, it seems to me likely that a count was regularly made of some or much of Osaka’s trade. In a report written in 1867, a British consular official who received details of imports of twenty items in 1866 from the magistrate of Osaka clarified the background: I found it impossible to ascertain even approximately the quantity of goods brought into Osaka by land and sea. Large quantities are, however, brought by sea, and of them some kind of statistical record seems to be kept, as will be seen from the following table, which was furnished by the governor, with the qualifying remark that the figures are, some of them, only approximate, and that of a few articles, including the somewhat important ones of silk goods, earthenware, sake, and miso, he had been unable, so far, to obtain returns.33

The existence of counts are hinted at also in the long report prepared by Osaka bugyō Abe and dated 9 April 1842, which cited some data for Osaka trade in the preceding four decades.34 For the years 1800–1840, Abe made statistical observations on the import at one time or another of sixteen commodities, either for specific dates or rounded periods. He was making a case at the behest of the city’s guilds for restraining competition in the hinterland of Osaka. The guilds had just been abolished, but their members, alleging that rural traders had reduced

148

EARLY JAPANESE TRADE

consignments to the city to their great detriment and seeking their own re-establishment, had convinced Abe to make a case for them. He did so by documenting what he claimed was a fall in trade from higher levels in earlier decades. The statistical argument, somewhat loose or imprecise, varied from one commodity to another. Some of the commodities to which he referred were of minor importance, and his figures are sometimes vague. He provided no data for the important commodity sake. In general Abe made a distinction between a higher level in the period ending 1817 (Bunka 14) and a lower level in the years immediately preceding his report of 1842. He compared data for 1840 or 1841 (and sometimes for runs of prior years) with earlier years that he identified as either 1817 or previous to 1817.35 In the case of cotton cloth, what he stated in 1842 was simply that in the ten years ended 1841 the annual import of cotton cloth did not exceed 3 million tan ➃(or ཯), as compared with figures of 8 million tan prior to that.36 Except for rice, the bugyō’s assertions are unverifiable from other evidence. The bugyō suggested that rice imports had fallen from 1,500,000 koku (and it is clear that he was referring to the level for years running up to 1817) to 1,085,000 koku for 1840, but other evidence contradicts a process of sustained contraction. The amount of rice recorded by the Dōjima Rice Exchange (Osaka Dōjima Beikoku Torihikijo ኱㜰ᇽ ᓥ⡿Ẇྲྀᘬᡤ) – and these are not figures of trade, but of stockage – reached a peak in the 1820s.37 The total fell only in the 1830s, during the harvest failures of the Tenpō era. In this way the bugyō’s assertion on the long-term trend in rice was far less well founded than at first sight his words might suggest. There may indeed have been no decline before the 1830s. Export trade was not a subject of discussion between the British official and the magistrate in 1867, and the consular report provides no information on exports. However, some surviving returns for six items in 1858–1864 suggest that figures may have been noted regularly. Five of these were the major exports from Osaka to Edo; for the sixth – rice – quantities from Osaka for Edo were either small or non-existent, and may suggest that if the practice of recording exports of eleven commodities from Osaka had not been continued in the structured form of 1724–1730, it did survive in vestigial form in the counting of a least five major Osaka commodities. The evidence lies in two sources, one an identified source, the “Ōsaka machiburegaki” ኱ᆏ⏫ゐ᭩, the other reproduced without indi-

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

149

cation of the source in both the modern OSSS and Ōsaka hennen shi. The two sources are incomplete, and the fact that their coverage differed suggests independent origins. Monthly returns for each commodity certified the number of vessels and the total quantity, over the signature of three machidoshiyori, prepared in the immediately following month. The commodities were sake, two categories of oil (with slight difference of nomenclature between the two sources), and ginned and woven cotton. Evidence in a document of 1797 for exports from Osaka in some months of the years 1761, 1788, and 1796 for one of these five commodities – ginned cotton (kuriwata) – might seem like an isolated trace of these statistics in a long intermittent period for which we have no other data. The recording of the number of vessels merits note. Recording seems to have been made of the number of ships arriving, their tonnage, and the han from which they came. The magistrate of Osaka in 1867 gave an “exact return of number and size of vessels and provenances for 1866” to a British diplomat.38 The list is identical in concept and layout to a few returns for the eighteenth century that have been published in modern times. The origin of such returns seems to lie in an order of 1727, and one may assume that the counts were made regularly every year – although some degree of uncertainty remains, as there are so few data for intervening years. While the image of a forest of ships’ masts is a dominant feature in contemporary pictorial representation, it is doubtful whether Consul J. F. Lowder, reporting on Hyōgo (Osaka) for 1868, was correct in observing “from a reliable source that the arrivals and departures of native junks average 300 daily.”39 As a count of both arrivals and departures, arrivals may logically be estimated to be half of that number. As the bar at the mouth of the port prevented fully laden vessels from entering, the figure is also more a count of lighters than of ocean-going vessels. It would however also include vessels of less than 200 koku, which had no occasion to unload their cargo into lighters in the anchorage and hence could proceed directly to the wharves. It is likely that any counts of goods on board were made at the port, the only effective way of taking account of the variety of craft involved: vessels with a capacity greater than 200 koku, which were lightened at the anchorage; lighters carrying goods from the anchorage; and craft carrying less than 200 koku, which were subject to no inspection at the anchorage. The question also arises of how or whether officials verified or amplified quantities, as opposed to

150

EARLY JAPANESE TRADE

the customary practice of reliance on returns from tonya. For rice, an intermittent contrast between precise numerical counts for trade rice and rounded totals for han trade might suggest that the tonya provided precise counts for trade rice only, and that han rice amounts were only roughly estimated. For those few years when there are precise numerical counts that seem to include both trade and han rice, a detailed count of the content of each vessel must have been either made directly or provided by the tonya. For the years 1724–1730, detailed counts were made in both Osaka and Edo – for rice and ten other commodities leaving Osaka for Edo, and for the same eleven items arriving from all parts to Edo.40 The tonya, wholesalers and undoubtfully also guild members represented by their tonya, feature in most returns. The obtaining of information through described by him as guilds is not itself surprising: the bugyō got their data on wages, for instance, through the guilds.41 It is less likely that monthly returns were compiled from direct filings by masters of vessels reporting on individual sailings, than that each tonya provided reports on his sailings in or out of the port. Almost certainly processing beyond a monthly level, if it occurred, was executed exclusively by bugyōsho officials. There is some modest clarification of this in the report by a British official in June 1867. Entertained in 1867 by all ten members of the tonya handling foreign goods (i.e., goods from Nagasaki, as Osaka was not yet a open port), he noted that while the tonya (which he said were 200 in number) were registered and under the supervision of the magistrate, “they carry on all their trading operations with little or no reference to him or his officers.” The tonya played a modest role in the statistical process in the understanding of this British offical. I endeavored to learn what were the amounts of their imports during the last eight years, but they assured me that they had no common books of the guild (sic) to which they could refer for such statistics, which could only be obtained by examining the private books of the members of the guild (sic).42

The trade returns for Edo in 1726 and 1856 include all rice (both trade rice and kura or daimyo rice) arriving in the city by water. For Osaka returns of the import trade include trade rice but exclude daimyo rice. Table 2 below presents all rice import figures known today from trade returns and other documents.

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

151

Table 2. Osaka Rice Imports (Coastal traffic) In man (units of 10,000 koku)

Year c. 1700 1714 c. 1716

Kura 118–141

Trade

112

28

c. 1747

140 OSSSb Mitsui archivesc

83–91

1736

Total Sources “Odaimyō ohatamotoshū goyōkiki” no kōa

22 90–118

Ōsaka-shi shi 1913 Odaimyōshū okurayashiki tsuki d rusui yakunin.

1766

141 OSSSe

1776

120 OSSS 1964c, pp. 101, 154

1777

86–113

121 Odaimyōshū okurayashiki tsuki d rusui yakunin. OSSS 1964c, f pp. 101, 154

1778

142 OSSS 1964c, pp. 101, 154

1779

114 OSSS 1964c, pp. 101, 154

1780

128 OSSS 1964c, pp. 102, 154

1817

150 Ōsaka-shi shi 1911g

1823–1828 1840

146–160

29–33

175–193 Suzuki 1938h 109 Ōsaka-shi shi 1911g

This table is a modified version of the table in Miyamoto 1972, 1988. a. “‘Odaimyō kata ohatamotoshū goyōkiki’ no kō” ࠕᚚ኱ྡ᪉ᚚ᪝ᮏ⾗ᚚ⏝⪺ࠖࡢ㡯, from “Gokinai Settsu Naniwa-maru” ஬␥ෆᦤὠ㞴Ἴ୸, in Kokka man’yōki ᅜⰼ୓ⴥグ (cited by Miyamoto). b. Though the figure in OSSS 1964a is quoted as 240,000 koku by Miyamoto, the total figure in that source is 282,792 koku. c. See note 43. The total is given as 830,000–910,000 by Miyamoto, 830,000–900,000 in Mitsui Bunko 1952, p. 2. d. Returns for 1747 and 1777 from “Odaimyōshū okurayashiki tsuki rusui yakunin narabini myōdai kuramoto kakeya yōkiki nayose” ᚚ኱ྡ⾗ᚚⶶᒇ㗤௜␃୺ᒃᙺேᖼྡ௦ⶶඖ ᥃ᒇ⏝⪺ྡᐤ, in “Settsu Naniwa-maru kōmoku” ᦤὠ㞴Ἴ୸⥘┠(Miyamoto). e. OSSS 1964a, p. 28 (incorrectly stated by Miyamoto to be from OSSS 1964c). The figure may be a composite one for kura and trade rice. See also OSSS Supplement 1966, p. 264. f. The figure of 1,210,000 is from OSSS, and differs from the kura estimate in “Odaimyōshū okurayashiki tsuki rusui yakunin narabini myōdai kuramoto kakeya yōkiki nayose.” It is either a further variant or a real count intended to include trade rice. g. The source quoted by Miyamoto (the Abe report of 1842) resorts to round years. The document claims that 1,500,000 was the level of imports up to 1817, not, as Miyamoto renders it, a figure for the years of Bunka and Bunsei (1804–1829). The figure for 1840 is for the year indicated by Abe.

152

EARLY JAPANESE TRADE

h. Two private archives in Suzuki, pp. 482–88, Hamamura Eizaburō-shi kyūzō ὾ᮧᰤ୕㑻Ặ ᪧⶶand Arioka Tahee-shi kyūzō ᭷ᒸኴර⾨Ặᪧⶶ. The dates of the documents are 1823 and 1828, admitting of a composite return for 1823–1828. Miyamoto gives them as for the Bunsei period as a whole (1814–1829).

Figures for trade rice (nayamai) which occur in general trade returns are minor if compared with estimates for kura rice from other sources. Where an isolated figure without context is given, it is not clear whether it is a kura count only or has incorporated also the figure for trade rice. That makes caution necessary when we use these statistics. A small number of returns of kura rice do exist, apart from trade rice given in the returns for 1714 and 1736, and apart from the gross figures with no identified provenances for 1766 and 1776–1780 reproduced in OSSS. One such return of unidentified origin, dating from ca. 1716, is in a Mitsui Bunko publication and is less than complete;43 another dates from Genroku times (see Table 2, note a); a third document contains returns for two widely separated years, 1747 and 1777 (see Table 2, note d);44 and a fourth return for the 1820s is actually a composite created by Suzuki from two private archives (see Table 2, note h).45 Where data for quantities of kura or trade rice exist, it is possible from readily available price information to value the imports of rice, and where aggregates for other trade exist, to add rice to the total. This is what Yasuoka has done for 1766. Kobayashi and Wakita employed the same procedure for Osaka imports in 1714, deriving the huge value of 456,510 kan for imports for that year.46 The usefulness of this large figure is limited, for purposes of comparison, because prices were grossly inflated in 1714; prices declined swiftly in the wake of the revaluation in 1715. The manuscript return for 1736 did not add up the component items, but Kobayashi and Wakita calculated a total 100,751 kan.47 As in the case of the1714 and 1766 returns, this basic return did not include kura rice. Osaka exports were significantly smaller than imports, even without taking into account the huge import of kura rice. At inflated prices exports were 95,799 kan in 1714, and at more normal prices, 76,218 kan in 1766. Some limited validation of the accuracy of aggregates of trade can be derived from the trade returns for 1724–1730. Hasegawa Akira 㛗㇂ᕝᙯand ShinbōHiroshi ᪂ಖ༤computed the annual average of six of the eleven Osaka exports for 1724–1730 on the basis of 1726 prices in Keizai shakai no seiritsu ⤒῭♫఍ࡢᡂ❧.48 The figures for five major commodities (ignoring the sixth item, a rather

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

153

trifling quantity of rice) come to 42,903 kan of silver, What stands out is the importance of kuriwata ⧞⥥, ginned cotton, which amounted to 19,766 kanme, or almost half the total. If the value of the many goods not included among the eleven commodities were known, almost certainly the grand total would be close to the count for 1766. Aggregates where they exist show how far short of imports the exports of Osaka fell. In 1714 imports (excluding kura rice) were 286,500 kanme, and exports 95,800 kanme. In 1766 imports were 103,734 kan and exports 76,218 kanme. Including and valuing kura rice and miscellaneous cereals, the value of imports in 1766 became 194,927 kanme. The gap between exports and imports, even without including kura rice, and a fortiori if it is included, illustrates the sheer magnitude of the deficit in the balance of trade of Osaka. The opposite side of this coin of course is that this emphasizes the scale of banking and exchange business conducted by merchant houses in Osaka. It was indeed their indebtedness which made a large number of han regularly move comparatively stable amounts of rice to Osaka, where they could utilize the market to liquidate or at least control their indebtedness. By the late eighteenth century, even faraway Tōhoku domains, which hitherto shipped little rice to Osaka, seem to have increased the amounts they sent. A four-fold distinction can be made in the figures. First, very full or elaborate returns existed for a few years, the 1714 return being the most complete. For 1736, a full list of imports may suggest that a matching list of exports has been lost and that aggregate values for both exports and imports are likewise missing. For 1766, aggregates survive, even if a breakdown does not; by chance, the totals were transcribed into a later summary document. Second, for a number of other years, some count existed. This is evidenced by the sixteen imported commodities of which bugyō Abe had some knowledge, by the 1866 count of twenty commodities given by later bugyō to a British diplomat in 1867, and by 1858–1864 vestiges of a regular count of six exports. The vagueness of statistical counts seems to be mirrored in the fact that the diplomat reported nothing for export statistics. The isolated scraps or blanks for other years hint at an appalling attrition of paper over time. Third, separate recognition has to be given to the remarkable figures for 1724– 1728 for both Osaka and Edo that were compiled in response directly to Edo instigation, and the return for Edo in 1856 whose origins are unclear but whose title and internal content illustrate official concern. Fourth and finally, the figures for rice, the largest single import in Edo and Osaka alike, merit separate consideration. Nayamai and kuramai

154

EARLY JAPANESE TRADE

alike were included in the gross import total in Edo, but in Osaka, they are separate; moreover, with some Osaka gross figures for rice, it is not clear-cut that they included trade rice. 3. The 1714, 1736, and 1766 Returns for Osaka

The 1714 return of Osaka imports and exports for both quantities and values is the fullest, best known and most commented on tabulation of trade figures for early modern Japan.49 It is the sole return to have been reproduced in English. Its authority is enhanced by the fact that copies survive in three locations. The 1714 figures first became accessible with their publication in 1964 in OSSS, vol. 13. Earlier, the existence of the 1714 return was known to a few, but they tended to be quite dismissive, noting flaws in its content. Kōda and Yasuoka, for example, knew of it and chose not to make wider use of it.50 Its publication led to the uncovering of more accurate versions in two sources, one in the Mitsui Bunko in Tokyo and the other in the Osaka Castle Museum (Ōsakajō Tenshukaku ኱㜰ᇛኳᏲ㛶; this copy was moved to the Ōsaka-shi Shi Hensanjo in 1969). These newly revealed sources were reported on by Ōishi in articles in 1964 and 1966, but until he incorporated them in a major study in 1975, other scholars seem to have continued to rely on the defective return.51 Curiously, while Kōda had dismissed the OSSS figures, the vastly superior copy in Osaka Castle Museum is actually in his hand. This copy, and the other two surviving copies of the 1714 return as well, originated in a version possessed by Uchida Gonzō that had been copied in 1903 for Tokyo Imperial University. Kōda’s copy was made in 1904 (as the date 1904 occurs elsewhere in the notebook into which he copied it). Another copy was made for Mitsui Bunko in 1913, and the OCCI made a copy at an indeterminate early date.52 The Uchida text is no longer extant, and the Tokyo Imperial University copy was destroyed in the earthquake and fires of 1 September 1923. By 1934, the highly capable but overworked Kōda himself (400 notebooks in his hand are preserved today) no longer recalled that he had made an accurate copy in 1904. This is less surprising than that he omitted this 1714 count from the Ōsaka-shi shi. His oversight, this shows, occurred not in the 1930s, but thirty years previously during his editorial work. It is evident that the Kōda and OSSS copies were made independently of one another. There are some very minor blemishes in the Kōda copy which do not feature in the OSSS copy. The OSSS copy,

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

155

moreover, despite doubts thrown on it by the few who saw it and the later emphasis by Ōishi on its defects, was not alarmingly defective. It is true that exports, if added up, came to a mere 63,736 kan, whereas the transcript itself gives the gross total as 95,799 kan. The transcribing however was not reckless. The table of imports in the OSSS text has few and very minor errors. In the case of exports, most of the gap between the transcribed components if added up and the transcribed gross total at the end of the table are accounted for by a copyist’s omission of four of five categories of oil. (In all probability this was a consequence of the transcriber having, after making the first entry for oil, inadvertently skipped across the other entries for oil in the text from which he was copying). These omissions apart, there were few errors of transcription; one was superficially more serious with the quantitative entry of 2,814,830 units of incense sticks (senkō ⥺㤶) repeated in the column for the value of goods as an absurd total of 2,814, 830 kan (a figure which if taken literally was many times the value of total imports). However, as an error of transcription it did not affect the independently-derived figure for the gross total value of trade at the foot of the return, which was correctly transcribed by the copyist. A more important point overlooked in modern commentary is that the trade figures in all three copies (Mitsui, Kōda, and OSSS) occurred within a more substantial document in which the trade figures were followed by a section of unrelated information under the heading of “Beikoku narabini ninzū yose” ⡿✐୪ேᩘᐤ. Identical in structure, the three documents contain the same miscellaneous material under the same general title in a small and identical compilation (in which admittedly the trade return is the main item).53 The fact that the version published by the OCCI in 1964 has an identical structure is not immediately evident in volume 13 of the OSSS series because the OCCI did not number documents within its volumes and the document started at the head of a page. The statistical return ends on the right hand side of sheet 14 and other material begins on the left-hand side under the heading “Beikoku narabini ninzū yose,” continuing on to sheet 15. Given the absence of a numbering of the documents, items under this heading might appear to be a separate document. (This is a small but striking illustration of the adverse consequences of not being able to refer to the original copies in the OCCI archives.) The kanji ୪(read as “narabini,” meaning “and”), which appears in the other versions, is missing here, but this is not a change of consequence.

156

EARLY JAPANESE TRADE

The 1736 return is more obscure than the 1714 return. Unlike the case of the 1714 return, we have no paper trail as to how the 1736 document was copied in modern times from earlier documents. There does not appear to be any evidence that the actual text used for vol. 1 of the Ōsaka-shi shi, where it first appeared, has survived. Yasuoka in 1960 and Yamaguchi in 1968 made the figures for 1736 more widely known.54 The 1736 figures indicate both values and the geographical provenances of the shipments. Quantities are lacking for some items, however, and for other items values are lacking (Yasuoka reproduced figures for twenty-eight items, twenty-five of them in descending order of value, Yamaguchi for twenty-five items). There is a total of 123 items or categories of goods. It seems likely that the transcription itself was made from an incomplete transcript of an earlier copy that was much more complete. It does not include any figures for exports; values are lacking for twenty of the 123 categories of imports, and no attempt was made, at any rate in the surviving copy, at adding up a grand total.55 While Yasuoka’s only firm figure for rice was 220,791 koku, given in his table of imports in 1736, in another table on Osaka imports for the years 1736, 1804–1830, and 1840, he expanded the figure for 1736 to one million koku.56 A figure of 800,000 koku for kura rice is not improbable, and a figure for 1,000,000 for total imports (kura and trade rice combined) is far from implausible.57 Cotton cloth can be distinguished by being given in units of measure (tan) for shiromomen ⓑ ᮌ⥥in 1736.58 The 1766 (Meiwa 3) aggregates are much less well known, and long languished almost unnoticed. General aggregates for this year were quoted neither by the authors of volumes 5 and 6 of Ōsaka-fu shi, which appeared in 1985 and 1987, nor by contributors to volume 1 of Nihon keizaishi (Hayami and Miyamoto 1988). It was not until 1990 that Yasuoka, who had worked on the editorial side of preparation of OSSS series, introduced the trade aggregates to the scholarly world in a short and very clear chapter in the volume on trade in the Shinshū Ōsaka-shi shi (1990, vol. 4).59 The reason the figures languished in obscurity for so long (and the reason why the return did not feature in the Ōsaka hennen shi series)60 is that the document containing the return sports a general and not informative heading, “Meiwa nenkan Ōsaka shotōkei” ᫂࿴ᖺ㛫኱ᆏㅖ⤫ィ (Miscellaneous Osaka Statistics of the Meiwa Period). Without reference to a brief statement buried in the analysis of the content of volume 13 in the later OSSS Supplement 1966, it can be easily glossed over. Hence it is not surprising that it apparently

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

157

has been consulted only by two scholars, very briefly by Miyamoto Matao61 for the figure for rice imports and more fully by Yasuoka. The document lacks a listing of individual commodities. It simply gives on a single sheet (sheet 28) the aggregate value of exports and imports, and, in two further entries, the quantities of rice and lesser cereals imported. The rest of the text of the document (sheets 29 to 32) simply contains more miscellaneous statistics. Using contemporary prices to value rice, miscellaneous cereals, and other items for which no values had been recorded in the return, Yasuoka was able to add the estimated value of rice and beans. He concluded that the figure for total imports should be raised from 103,734 kan to 194,927 kan. Whether this brief document is itself a mere copy of another document or a terse recital of gross figures from a fuller compilation has to be a matter of conjecture. I think it probable that the OCCI, which reproduced a great many full texts elsewhere in its series, had access only a now unknown summary in this instance. 4. 1724–1730 Listing of Eleven Commodities in Edo and Osaka Trade

The data for 1724–1730 are often reproduced in modern works. However, in contrast to the curiously isolated data for other years, figures for these years for eleven exports from Osaka represented a highly structured approach. That approach suggests that the compilers worked with a well-defined and ongoing purpose. Moreover, copies are somewhat more numerous than for other years, and have a closer official origin or association. A statistical urge to collect trade figures cannot be attributed uniquely to Yoshimune, who became shogun only in 1716, or to officials working under his direction, as the remarkable data for the year 1714 already illustrate what was probably at the time a statistical innovation. The figures for 1724–1730, intended to serve a real administrative purpose, were collected on a monthly basis and the detail was to be communicated to Edo with equal regularity. Furthermore, the compilation of data in Osaka was to be followed by a checking of vessels at their entry to Edo Bay, at Uraga. The original instructions were further elaborated with effect from the fifth month of 1727 (Kyōhō 12). Osaka officials were instructed to keep a count of all vessels entering and leaving the harbor. More detailed requirements for reporting by the tonya were set out from the seventh month in regard to vessels of 200 koku and upwards (in effect, if loaded, oceangoing vessels which could not ascend the shallow

158

EARLY JAPANESE TRADE

channels to Osaka proper): they were to provide to the machidoshiyori the details of vessels, burden, name of master, and whether or not the vessels had carried goods to Edo, and they were to submit these on a monthly basis.62 A count of vessels of 200 koku and upwards as late as 1866 suggests that this instruction was observed in Osaka to the end of bakufu rule.63 The figures for the years 1724–1730 first appeared, as was the case with so many other data, in Ōsaka-shi shi without indication of source. A fuller version with a monthly breakdown in Ōsaka hennen shi 1965 (vol. 26 pp. 334–41), though not identified, is probably from a document cited by Kobayashi and Wakita in 1973. While simply a copy entered in a larger compendium of documents, the compendium has a clearly defined scope and hence suggests that it is a relatively early copy. These two sources (one now lost) for eleven commodities shipped from Osaka over 1724–30 are supplemented by a surviving listing of eleven corresponding items landed in Edo for one year. Of this document there are two sources, the first a copy in Katsu Kaishū’s Suijinroku ྿ሻ㘓(for which he does not indicate a source);64 the second a copy contained in a compilation entitled Kyōhō tsugan ாಖ㏻㚷that was transcribed in 1886 from a volume in private possession.65 Thus the Kyōhō tsugan version is itself a copy of what was already a copy. These two versions, despite some discrepancies that are probably the result of copying errors, are in fact very close to one another. They must come at few removes from a common source. Ōishi thinks the 1886 document Kyōhō tsugan the better of the two.66 Two of the transcription variants were minor, the third one major. The quantity of sake given in Kyōhō tsugan is very obviously in error as 10,095,856 barrels; a more credible amount is the 795,856 barrels in Katsu’s Suijinroku, The really important difference is in regard to ginned cotton (one of the two textile items covered): it is accounted for in Kyōhō tsugan but omitted by Katsu. In place of ginned cotton, Katsu – or, as he was a very accurate transcriber, the source from which he drew his figures – lists sen 㖹, copper coins, among the eleven commodities.67 The account of eleven commodities entering Edo in 1726 was intended, as the words shokoku yori Edo e nyūshin ㅖᅜࡼࡾỤᡞ࡬ධ ὠ (“entering Edo harbor from various provinces”) in the document suggest, to include incoming trade from regions other than Osaka. We do not have instructions at the Edo end comparable to the full ones issued for Osaka for this recording exercise. Certainly it is plausible

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

159

that similar rules applied to goods entering Edo bay from the south and from the north alike. If there were not a comparable control of goods from the north of Japan, the whole purpose of the enumeration effort would have been negated. Given a diffuse trading pattern in the north of Japan with no center comparable to Osaka, Uraga was almost certainly the sole originating center of the recording of this traffic. As Uraga kept track of the final destinations declared for vessels entering the Bay, it should have been easy, with the help of the tonya who imported goods from the north, to compile statistics on vessels and quantities of goods and send those monthly to Edo. In the case of rice, of which Osaka shipped little, imports to Edo of 861,893 hyō were recorded in 1726. This figure covers only rice carried directly from Uraga to Edo, as noted on its arrival by sea. It omits rice landed elsewhere in the bay and then carried overland. Firewood and charcoal came almost exclusively from regions other than Osaka. In the case of sake, the quantity shipped from Osaka to Edo, 77,687 barrels, was large in absolute terms but comparatively small as a component of total imports into Edo of 795,856 barrels. Oil was important in both exports and in imports. In 1724–1730, the records of exports from Osaka show no quantities whatever for two goods, figures that we can take to be nominal for three other goods, and minute quantities for a sixth – rice – in all years but 1728 and 1729. From this we can suppose that there was in normal times a substantial export in only five of the eleven commodities. This might explain why recording in 1858–1864 of export trade from Osaka to Edo was confined to the same commodities. The plan for regular recording of the eleven commodities is contained in the bukka hikisage an ≀౯ᘬୗࡆ᱌of 1723. Having failed in efforts to raise rice prices, the shogunate at this stage endeavored to reduce prices of goods that recipients of rice incomes consumed. The objective was to raise the purchasing power of the shogunate and its servants. The order itself prohibited forestalling; creation of a statistical account of the movement of goods was intended as a means of identifying lags or abuses by withholding goods on their arrival in Edo Bay from the market (manipulating supply in order to manipulate prices).68 Details of shipments were to be sent every month to the machi bugyō in Edo from the beginning of 1724. For goods from the north, we have to assume that a return in some form was made by the relevant tonya for the northern trade at Uraga and again in Edo. Inspection of all vessels at Uraga would serve to keep track of quantities and destinations, and

160

EARLY JAPANESE TRADE

the returns made by tonya of vessels reaching their destination would identify abuses. Just as the Osaka bugyō made monthly reports to the Edo machi bugyō, constant communication between officials at Uraga and the machi bugyōsho must have been envisaged to make surveillance effective. It seems likely that the scrutiny in Edo took the form of matching monthly data for Edo from the tonya with the returns from Osaka and Uraga, and that the matching process itself was very general. No direct evidence seems to survive of the operation of the procedures at Uraga. The only concrete evidence as to how the control worked comes from the rice trade examined in an elaborate report kyonenjū tōchi e chakumai no oboe ཤᖺ୰ᙜᆅỤ╔⡿ࡢぬ(memorial of rice landed in the course of last year) in Kyōhō sen’yō ruishū beikoku no bu ாಖ᧝せ㢮㞟⡿✐ࡢ㒊 (Kyōhō collection of essential details: section for rice and other cereals) for 1727, drawn up in the seventh month of 1728.69 This source examines the trade in two stages. The first is the basic statistical return for rice recorded directly at the Edo market, amounting to 1,256,453 hyō for the year, a total made up of 835,681 hyō of samurai rice and 155,781 of trade rice, all on normal cargo vessels, plus a further 420,772 hyō on coastal vessels (in other words rice transshipped at Uraga from incoming vessels) that was allowed by the Uraga watch house to proceed unhindered, no doubt because of an undertaking to adhere to Edo as the ultimate destination of the rice. A second set of calculations followed. These were intended to arrive at a true statement of the gross traffic. A distinction was made between trade rice and rice on official account (bukemai, sometimes ryōshumai 㡿୺⡿). Han rice was less for consumption literally by samurai families than for sale. Some rice cargoes were recorded at Uraga and later transferred to the city overland from more distant landing points, and hence they were not included in counts of trade unloaded within the confines of the port. In Tokugawa Japan port boundaries were not defined in the Western sense. Convenience rather than legalities determined the extent of a port. In the case of Edo the main center was Shinagawa. The anchorage that served Shinagawa, however, was said in 1872 to be two or three miles offshore, and when goods were landed there they had to be carried five to eight miles further to the business center of Edo. The waterway from the anchorage to what the British official labeled Hama-goto (an erroneous rendering of hamagata ὾᪉, a general word for shore, not a specific place name) proceeded for some four miles with a depth of one and a half to three feet; within the mud flats, the actual channel used for navigation in many places had

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

161

a mere two feet of water at low tide.70 Effectively there was an advantage in using locations other than those served directly by the barges onto which goods were loaded in the crowded anchorage off Shinagawa, despite the extra costs of overland transport. The second stage – bugyō recalculations – touched on samurai rice and chōnin rice, resulting in a total of 1,377,118 hyō (2.5 hyo equals a koku). These figures resulted from an adjustment to the original calculation of samurai rice by the addition of 585,642 hyō first landed outside the port and then carried to the city, and an adjustment of the original figure for chōnin rice by the deduction of 44,205 hyō that had been landed in other places and did not reach Edo. When the quantity of rice loaded at Uraga onto coastal vessels (420,772 hyō) is added to this figure, the grand total amounts to 1,797,890 hyō.71 The shallow anchorage at Shinagawa was both a reason for unloading all or part of many cargoes at Uraga, and for vessels which did not do so to seek other points on Edo bay to unload. Some chōnin rice did not reach Edo at all. These statistics call our attention to the definition – or more precisely, lack of legal definition – of ports. In contrast with very rigid definition of ports and sub-ports in British or French practice, ports in Tokugawa Japan were diffuse entities. In Europe, the authorities collected fiscal charges on much trade, and this necessitated rigid definition of ports for all traffic; further, any landing outside the legally defined ports required the (rarely granted) prior approval of the revenue officers. In early modern Japan, as already observed, figures for commodities at every stage were returns made by the tonya or shipping wholesalers, not counts by customs officers in the Western sense. Of these there were, for example, 163 in Edo in 1726 and over 200 in Osaka in 1866. The returns made at the outset of voyages at Osaka and the later verifications in Edo in all probability reflect the authorities’ acceptance of quantities reported by the tonya. It is doubtful that there were independent inspections of actual cargoes. The major task of Uraga scrutiny was to superintend the transfer of goods from incoming vessels to coastal vessels. The machi bugyō’s office in Edo would be guided by returns from Osaka and Uraga as well as evidence gathered on the spot. Adjustments in Edo represented a reprocessing of existing paper, rather than a recount. As in Osaka, almost all han rice was carried on commercial vessels, and its entry was controlled by the rice tonya. Conversion into koku of statistics originally stated in hyō suggests consumption of the order of 720,000 koku in 1727. This might seem an understatement.

162

EARLY JAPANESE TRADE

Osaka, a smaller city, imported a million or more koku. Multiplying the estimated population of Edo by the estimated per capita consumption rate suggests that more rice may have been imported than was recorded. Ōishi suggests that there was daimyo rice that was not recorded at Uraga at all, and it should be added to the total.72 This seems unlikely, as toleration of such an inflow would have run counter to the exercise at large. Allowing for vagaries of statistical recording and of estimates of per capita consumption, figures of 700,000 koku need not be regarded as wildly out of line with the reality of Edo consumption of waterborne rice in the Kyōhō period. Arguably rice from the shogun’s own domains (tenryō ኳ㡿) carried on shogunal vessels was not included, but Ōishi did not broach that possibility. While rice could be on daimyo or shogunal account, its movement was almost exclusively effected by merchant vessels. A British consular official noted that of 1,967 junks with capacities of 200 koku and upwards entering the port of Osaka in 1866, a mere 124 were daimyo vessels.73 Shogunal rice would have been no different; the bulk would have been carried on commercial vessels. It is possible that the process of counting that yielded the statistics for the seven years 1724–1730 continued until 1736, the year when debasement was resorted to again in an effort to raise the price of rice. It is more likely, however, that it was abandoned at an earlier date, in 1731, when harvest failure radically changed conditions in the rice trade. Yet some elements of the system survived. In the case of Osaka, some traces still existed in 1858–1864, as we see in the enumeration of five to six commodities that had been the key items among eleven enumerated commodities in the 1720s. 5. The List of Edo Imports in 1856

It is striking how meager, apart from one year’s imports in 1726 (and rice imports for 1727), is the statistical information for Edo. In view of the paucity of data, the existence of very full figures of trade in Edo for one year in the 1850s (1856) is at first sight a puzzle, as it raises the question why, if statistics existed for one year, they do not survive or at least be known to have existed for others. A table with the 1856 data is printed in Tōkyō-shi shi kō ᮾிᕷྐ✏ without any indication of the source.74 The 1856 return itself however is very different from the small number of full or near full returns for Osaka. It is a very extensive document, giving not only quantities for 128 commodities (some of them with multiple subdivisions), but also details of provenance of

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

163

many goods, accompanied by more general observations. It thus differs from other surviving counts, which in presentation are very terse documents. The orderly presentation gives the impression of a specially devised exercise, not an ongoing one, intended to provide administrators with a picture of the nature and sources of supplies of food and of conventional necessities. Given the recent appearance of foreign fleets and the treaties already wrung from Japan, the document may have been prompted by fear of disruption of food supply by foreign navies and a need to estimate the scale of the problem that disruption might pose. This fear, as Fujita Satoru ⸨⏣ぬ argued more than twenty years ago, was a factor behind Mizuno Tadakuni’s Ỉ㔝ᛅ㑥 policy in the 1840s and especially in his relaunch of the effort to drain the Inbanuma ༳᪠἟marshes.75 Such concerns, long-lasting ones, might have been heightened by fear of social unrest in the wake of the terrible 1855 earthquake in Edo, but the fear of foreign threat and social unrest coinciding had haunted Japanese rulers since the 1780s. The title of the document reveals its special and secret purpose: Chōhōroku: Edo omote shoshiki funaunsō nyūshin okazuke chakunidaka mitsumitsu otazune 㔜ᐆ㘓 Ụᡞ⾲ㅖⰍ⯪㐠㏦ධὠ㝣㝃╔Ⲵ㧗ᐦࠎᚚᑜ (“Very useful record: Edo list of various ships carrying goods to its shores, with attached highly confidential enquiry into quantities landed”). The entry for rice is the first and the largest entry in the list, and compilers of the document made an attempt to calculate consumption.76 For other goods the document often cites not only locations from which supplies came, but also supplementary details. A return for imports to Osaka from Ezo for 1856–1858 can be viewed in the same context.77 It seems to exist only in an account in a relatively modern book, not in a surviving copy. The fact that it has no known antecedents might support the suggestion that it originated in Osaka bugyō fear that the large and important Ezo-Osaka trade was as vulnerable to blockade as the trade of Edo itself. Indeed disruption of trade from Ezo to Osaka would itself have been, as Japanese shipping hugged the coasts, an automatic consequence of the operation of foreign fleets in the approaches to Edo. 6. Comparison of Trade in Osaka and Edo in 1726 and Evidence of Changes in Edo Imports, 1726–1856

Although trade between Osaka and Edo is documented only for intermittent periods, the surviving records do afford the unique

164

EARLY JAPANESE TRADE

luxury of a comparison of the two cities in terms of trade in one year. That year is 1726. Once we deduct from total imports to Edo the figures for known shipments from Osaka, we have a fairly good idea of the extent of trade in these commodities to Edo from other parts of Japan, predominantly from the north. All movement of trade by sea carried risks, including loss of cargoes, long delays in arrival at destination (resulting in cargo reaching the point of arrival well into a following statistical year), and fraud. In Tokugawa Japan actual shipwrecks were remarkably few, because vessels often hove to on the shoreline at night (a practice that explains also why transport was by European standards very slow). However, as the machi bugyō report in 1728 on the rice trade in 1727 shows, some goods were landed elsewhere and then carried overland. There are inconsistencies in the data, then, and while these may not of themselves invalidate the statistics that have survived, comparisons should not be pressed too far. Table 3. Exports of Eleven Commodities from Osaka to Edo and Imports in Edo from All Japan, 172678 Item

Unit of measure

Osaka

Edo

Rice

hyō ಥ, bundle

3

861,893

Miso

taru ᶡ, barrel

---

2,898a

Sake

barrel

177,687

795,856b

taba ᮰, bundle

---

18,209,687

hyō ಥ, bundle

764

809,790

Lamp oil (mizu abura ỈἜ)

barrel

69,172

90,811

Fish oil (gyoyu 㨶 Ἔ)

barrel

---

50,501

Shōyu

barrel

101,457

132,829

ko⟠

12,171

36,135

ᮏ bolt[?]

98,119

82,019a

barrel

248

1,670,880

container weighing 15 kan

---

19,407e

Firewood (maki ⸄) Charcoal

Cotton cloth (momen ᮌ⥥) Ginned cotton (kuriwata ⧞⥥) Salt Copper coin

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

a b d e

165

Figure given by Katsu. In effect the figure given by Katsu has been adopted. c One ko equals 100 tan ➃. Omitted by Katsu Kaishū. Given by Katsu Kaishū only.

Given the survival of figures for Edo imports in 1856, it is also possible to compare the figures for that year with the Edo imports of 1726. Yamaguchi in 1968 contented himself with reproducing the 1856 figures for Edo.79 Hayashi Reiko carried this a stage further in 1969 in comparing the imports of Edo in both years.80 The conclusion drawn in Nihon keizai shi, vol. 1, which reproduced the table together with a commentary, is that trade expanded significantly between the two dates.81 However, while five important commodities (sake, cotton cloth, charcoal, shoyu, and miso) rose quite sharply, three others (salt, firewood, and oil) did not increase. The overall evidence seems too inconsistent to support a confident conclusion that commodity trade was generally expanding. It is hard to understand how the quantity of some conventional necessities increased greatly while others remained static. While the sharply falling imports of ginned cotton implied that a weaving industry around Edo was in contraction, it would also help to account for the rise in imports of cloth. Cloth illustrates the problems inherent in some of the figures; conversion of cloth measures poses difficulties. Nihon keizai shi, vol. 1, used a conversion rate of one ko equals 100 tan, the same rate as Katsu Kaishū had used in 1890. To make the matter comprehensible in European terms, we need to convert the tan. According to the Nelson dictionary, one tan equals ten yards,82 but according to McClain, a tan was larger, measuring 12.7 meters by 0.73 meters.83 There may be a problem also in the ko being variously said to be either 100 or 120 tan.84 The real problem may be not the tan itself but the conversion of ko into tan: The ko is an awkward concept as a measure, whether it is 100 or 120 tan. A figure of 100 tan is an impossible size for a bale of cloth, and hence raises a whole area of uncertainties as to what a ko represents precisely. The tan on its own is a safer counter, and incidentally is the one used by Abe in 1842. Total imports in 1856 were 80,168 ko, the equivalent of 80,168,000 yards (if we take one tan as the equivalent of ten yards, as Nelson does) or 101,813,607 meters (for a tan of 12.7 meters, as McClain has it). That this was the scale of trade might be indirectly suggested by earlier evidence of imports of 80 million tan to Osaka tonya in the early 1800s.85 Large figures such as this gain a degree of plausibility from the fact that all figures for cotton trade for Edo and Osaka are very

166

EARLY JAPANESE TRADE

substantial. Osaka tonya in the early 1800s handled 80 million tan of cotton annually. At a much earlier date, 1726, the imports of cloth from Osaka to Edo were a mere third of the total imports, hinting that much larger quantities were coming from other locations. However, the problem really lies in the fact that even if we take the lower of the two measures for cloth, an import of 80 million tan would imply a huge consumption, about 80 yards per head. This is an impossible figure for Edo on its own. Even for the Kantō as a whole with a population of 4.4 million in 1846, it would still mean an import of 18 yards per head. If a ko of 120 tan is used the figure rises to 21.6 yards. Two further questions arise. First, the statistics may be at fault as an effective count of trade, especially in the case of the 1856 count; we know nothing about the origins of that count or the methods employed to produce it. Second, it is possible that cloth was reshipped from Edo for the Tōhoku and other destinations. Hayashi does not state this directly for cloth, but for imports at large, she advances the general proposition that Edo was a staging point for re-exports.86 Combining the population of Kantō and the northeast (using data from Sekiyama’s study of population),87 a population of 7 million still means an overall consumption, for Kantō and Tōhoku combined, of eleven yards or fifteen meters per head. Though high, these numbers are not wildly unrealistic; possibly Edo tonya in cloth offered more flexible transfers, especially to lesser centers along the coasts to the north of Edo, than did Osaka tonya shipping direct. The data for 1726 might be seen to make the high figure of 1856 imports more plausible. Imports in 1726 were already high: 36,135 ko, which would amount to 36 million yards or more. The 1726 figures come from a recording process about which we know something, and hence poses fewer problems than the 1856 data. 7. Rice

Rice, as the Edo machi bugyō’s study in early 1728 suggests, transcended the other ten commodities reported in trade figures for 1724–1730. It was in a category of its own as the major foodstuff of Japan. Hence, it is hardly surprising that in Osaka, the key rice market, special attention was given to recording rice either in terms of storage or of trade. Volumes of rice storage and rice trade approximate each other, in Osaka counts, as relatively little that came in from the hinterland was consumed in the city, and some of the imported rice was redistributed into the surrounding countryside. It is not always clear to what extent

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

167

separate estimates of stockpiling and crude trade counts were simultaneously made. The only continuous data for rice are figures compiled by the Torihikijo or Rice Exchange for rice stored at year-end. Over a long period of 143 years from 1724 to 1867, they are complete except for a mere five years (1728, 1736–1737, 1743, and 1791). The numbers were published by the Dōjima rice merchants’ exchange in 1912 in compliance with an order by the city council.88 As many of the records had been lost at the time of the Meiji Restoration, the volume was based on copies or extracts from miscellaneous sources, including supplementary information from records held by citizens whose families were probably once associated with the nakama and tonya. It is not clear to what extent the records consulted were originals (as opposed to copies), or whether the statistics were reconstituted from earlier tabulations that had been based on scattered sources. Strictly speaking, a measure neither of kura annual turnover nor of trade itself, these figures are nevertheless the best indicator we have of trends in trade. If expressed in averages (usually ten-year averages), at a peak they reached 2,300,000 hyō over the decade of the 1820s, or 920,000 koku.89 At such levels they must have reflected a high proportion of the total trade, at any rate from 1744 onwards. The sharp rise in the two decades preceding 1744 suggests a change in statistical coverage or the ending of some kind of limitation rather than a dramatic change in trade. From 1744 to the 1790s (Kansei times) the figures remain at a very stable level. From 1798 (Kansei 10) onwards over the long term they fluctuate. A rise in the 1790s and in the early decades of the new century is consistent with what we know from other sources and with contemporary concern about very low prices over a period of more than two decades. The high point of storage seems to be the years 1819–1828 (Bunsei 2–11). The average is halved for the years 1829–1838 (Bunsei 12–Tenpō 9). A moderate upswing recurred in the 1840s, reversed by a decline in the 1850s to a nadir in 1859–1867 (Ansei 6–Keiō 3). In the final decades the figures mirror less food shortage than a wider business crisis faced by both the city and the daimyo who traded there. The reality of this crisis is evident in the concern among daimyo and their advisors in the 1860s over the very future of Osaka. If Osaka was contracting, it was not so much because of problems in the supply of rice generally as of a decline in Osaka’s place in the network of trade that linked Japanese regions with each other. This showed up in a relative weakening of Osaka finance houses. For the

168

EARLY JAPANESE TRADE

daimyo, any impairment in the status of the port threatened the continuance of the long-standing ease of borrowing from their creditors. If 3 million koku of rice was imported in 1856 in Edo – although, as demonstrated above, we cannot be certain of that figure – it already reflected a shift in the center of gravity of trade from Osaka. Such a change would correspond to the known strengthening of the ryō region centered on Edo and the weakening of the silver region centered on the western han and Osaka. As a result, in the 1860s, as well as a political crisis, there was a growing economic cleavage between east and west. That was very evident in the words of Saigō Takamori to the British diplomat Ernest Satow.90 Moreover, as Osaka rice turnover was high to the very end of the 1820s (according to the Torihikijo figures and also in the estimates of the rice trade in the mid-Bunsei years made by Suzuki, who reckoned that imports from the Tōhoku had risen to 20 per cent of the total), Osaka was still a booming center on the eve of the Tenpō food crisis in the 1830s. The origins of a shift in the business center of Japan may arise in the 1840s, rather than in the 1830s; by the forties, the trade should have been returning to normal. The emergence of change sometime in the two decades of the 1830s and 1840s understandably gave rise to deep concern in western Japan. The old pattern had existed for a century and a half, from Genroku times. The rise in Tōhoku rice sent to Osaka in Bunsei times is not in itself surprising considering how central the business advice and finances of Osaka houses were to the Tōhoku daimyo. But it mirrored a new instability in Osaka trade, as it attracted custom from han whose commodity business was with Edo as well as Osaka, and which could easily revert to Edo. Kura returns, by virtue of the fact that they are given within lower and upper bounds, and usually in rounded figures, do not seem to be statistical counts in the literal sense. These returns also contain lists for each han, with extensive subsidiary details. On the other hand the trade figures – when they exist, as in 1714 and 1736 – do seem to imply strict quantitative terms.91 The figures for 1766 and more particularly for 1776–1780, for which there is no supplementary information, present a challenge to interpretation. They imply, superficially, at least, a precise count. We lack, however, any indication as to whether the numbers are for kura rice only or for total trade.92 Atypically, for the year 1777 there are two separate counts, plainly illustrating the problem and the difficulty of reaching a conclusion with confidence. In a table of mere totals of rice imports in the five years 1776–1780 (in OSSS 1964c; see Table 2,

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

169

above) we find a figure of 1.21 million koku for 1777; this contrasts with the estimate for kura rice that was stated as falling within upper and lower bounds of 0.86 million to 1.13 million koku in 1777. It is possible that often counts were rough and ready, but that on occasion, the counting technique that yielded a more precise count for trade rice was extended also to kura rice.93 This may have been the case in 1777, for example. The bugyōsho did have some direct involvement in monitoring the trade in rice that went beyond passive acceptance of estimates from the tonya. According to Suzuki, officials took note of the contents of vessels at the entrance to the Kizu ᮌὠand Anji Ᏻ἞rivers.94 This could have been in keeping with an instruction of 1727, which called for taking a precise count as one of the procedures for monitoring the trade in rice and other commodities between Osaka and Edo. Both tonya returns and monitoring by bugyōsho officials existed, yet it is possible to conceive of year-to-year or period-to-period variation in the exactness of their counting, with calculations fuller on some occasions than others. If a close record was made of the number of vessels of above 200 koku at the offshore anchorage (because they presumably carried the bulk of the trade and transferred part of their cargoes to lighters in order to be able to continue their own journey upstream), it would be easy to visualize how a tight survey of vessels in the anchorage could be turned into a count of the contents of vessels and of a wider double check at entry to the port, especially as smaller craft (both coastal and oceangoing) were not subject to control at the anchorage.

Surviving counts of trade seem mostly to be privately made copies, rather than official copies of the original documents. Private circulation would explain also why compilers either had less than complete access, or simply failed to make a full copy, or in doing so made errors of transcription. For his calculations for the 1820s, Suzuki drew on two manuscript accounts that were separately held in private hands and therefore assumed to be copies. The fact that one of the few documents of kura rice has information for two years as far apart as 1747 and 1777 illustrates the somewhat random accessibility of surviving information. Even information accessible to the bugyō, especially if he was surveying a long period, could be relatively limited. The example of bugyō Abe making observations in 1842 on trends over several decades illustrates this. The variance between the trends he discerns and those that can be inferred from the Dōjima Rice Exchange figures suggests that either the bugyō had limited access to information or he rather too readily came to an impressionistic conclusion.

170

EARLY JAPANESE TRADE

8. Archival Issues

For both population and coastal trade (as opposed to the foreign trade of Nagasaki), as already observed, statistics for early modern Japan that have survived are almost solely in private collections. There is one very real difference in trade figures compared with demographic data, however. In marked contrast to the lack of tabular recording or presentation of population figures, a tabular or near-tabular approach did exist for trade. However loose his mode of analysis, bugyō Abe in 1842 wrote about Osaka trade in a manner in which population had never been discussed. Surviving copies rarely seem to take the exact form of the original documents concerned with trade, or even the form of versions made close to the date of the original. The survival of freshly made copies even in Meiji and later times was – and remains – hazardous. Copies noted in the early decades of the twentieth century have since disappeared or at least cannot be traced. What is now the best known of all statistical returns of trade – that for 1714 – was more or less unknown till 1964. Outside its latter-day printed format, it exists only in three manuscript copies, of which the earliest is the Kōda copy of 1904. The text held or made by Uchida is now unknown, and the copy made for Tokyo Imperial University in 1903 – from which Kōda, Mitsui and the OCCI transcribed copies – was destroyed in 1923. The Uchida document was thus copied four times that we know of (Tokyo Imperial University, Mitsui, Kōda, and OCCI). If Uchida was himself a copyist and not a collector, we are looking at a story embracing six copies (i.e., including as well as Uchida’s lost text the now unknown document or documents from which Uchida himself worked). Details of trade for years other than 1714 rest on an even more slender basis. The 1736 return for Osaka is now known in printed form from the old Ōsaka-shi shi alone; that for 1766 is known only in a six-line entry of unknown origin, accessible only in the transcription included in OSSS in 1964. The 1724–1730 material, because of its high administrative purpose, fared better. It exists in two sources at the Osaka end (a summary copy of unknown provenance in the Ōsaka-shi shi and a text copied into an early compilation of documents), and at the Edo end in two versions for a single year (in transcripts made in the Meiji era) plus a copy of a machi bugyō reworking in early 1728 of the data for rice in 1727 (which as it is incorporated into a compilation of administrative documents is clearly a copy, probably one made at a comparatively early

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

171

date). Even the great return for Edo 1856 is simply a copy (with one glaring error of transcription), whose transcriber offered no guidance to its context. For rice, apart from gross figures for 1766 and 1776–1780 (surviving only in OSSS and of unknown origin), the huge kura totals (with their upper and lower bounds) are known for a mere five years – one year from a Mitsui modern publication that gives no indication of the sources of its information, three years in two compilations which copied earlier documents, and the fifth year a composite table intended to correspond to the 1820s constructed by Suzuki from two documents that were in private archives when he wrote. Even the Dōjima Rice Exchange figures, uniquely almost a complete run for 147 years, were compiled a century ago from now unaccounted-for documents after the older papers already had been lost or scattered. For statistics on the bakumatsu years, we are unsure whether nowexisting copies are a random survival or a source providing evidence of a reinvigoration of administration. This is an important uncertainty, because it casts doubt on the assumption made by some scholars that the absence of statistics is proof of a breakdown of Tokugawa administration. Population statistics illustrate the problem more graphically than trade ones. Bugyō Isshiki’s personal compilation of population figures for Osaka down to 1856, surviving in his own hand,95 prompts the question as to whether it was just a curiosity of his individual interest or part of a wider, invigorated statistical approach. Were there no population censuses at national level after 1846, as has been argued, or have the copies just got lost? For the population of the cities of Osaka and Edo, does the paucity of data merely reflect the vagaries of survival, or is it a sign of administrative breakdown? For Edo in the 1860s an isolated population return exists for 1867. In the case of Osaka, three contemporary sources including the Isshiki table fail to bring the population data beyond 1862. Yet the British Parliamentary Papers (hereafter BPP) reveal that as a result of a conversation with the bugyō of the city, the young British diplomat Algernon Mitford in 1867 was provided without fuss with full details of a census for 1866 that was identical in format to earlier returns.96 What we know of earlier censuses comes from a flotsam and jetsam of copies, often made much later than the date of the census to which they relate. If a sexennial census were taken in 1852, 1858, and 1864 (the final Tokugawa-era count occurred in 1864), there may have been too little time for a diffuse copying process to have achieved a high probability of survival of the data. Katsu Kaishū illustrates the problem we

172

EARLY JAPANESE TRADE

face. He obtained a good deal of information on earlier censuses from miscellaneous sources, but for the nineteenth century, he found data on only two censuses. It seems not to have helped him that he had been a high official of both the shogunate and the new Meiji government, that people who had experienced the period in question were still living, and that he was acting on an express government commission when he sought the data. In a curious way, the difficulties he encountered were similar to those the Hayashi had to deal with several decades earlier in compiling Tsūkō ichiran. In Katsu’s case, the task was to reconstitute the story of the final decades of Tokugawa administration, and in the case of the Hayashi, it was to bring together the records on foreign relations in anticipation of an external challenge. In both instances the response was not – probably it could not have been – “to look at the files.” Rather it was to gather information together from highly miscellaneous sources. In contrast to census information, which has often survived in very miscellaneous, sometimes unorthodox or whimsical sources, trade data has survived only in narrower and less colorful compilations. One explanation for the general shortage of figures is that they circulated only within the narrow ambit of the bugyōsho of Edo and Osaka. While tonya provided information on a monthly basis to shogunal officers, it seems they often regarded the information as transitory. They were under no obligations to hold a permanent record. I remarked earlier on what the BPP noted in the case of one tonya in 1867: there was a deficiency of “common books . . . for such statistics.”97 Some tonya may have possessed some information. But the guild for traders in ginned cotton seems not to have been able to provide to the bugyō of Osaka in 1797 more than discontinuous figures for exports covering some months but not all of the three years 1761, 1788, and 1796.98 Events of the bakumatsu years made it seem imperative to learn more about trade. The external crisis that forced Japan, under the guns of foreign navies, to concede some access to foreigners suggested the importance of mastering details of the scale of the provisioning problems. This was the reason for the great return for Edo in 1856. Aware of the vulnerability not only of trade routes from the Tōhoku to Edo, but also of Osaka’s sea approaches, bakufu officials came to a heightened appreciation of the strategic significance of Osaka. The evidence for reinvigorated record keeping in Osaka is not conclusive, if we consider that the survival of more evidence for the 1850s and 1860s than in preceding decades might be a mere consequence of random chance. The sources for demography and trade, however, do not support a case for a

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

173

breakdown of Tokugawa administration. Quite the contrary, the paper trail, although it may be a little thin, suggests that statistical recording continued to function; perhaps there was innovation, even. Whatever the vagaries of survival – typical enough of Tokugawa documents – the important point is that some statistical competence existed. When after the opening of the ports to trade in 1859 there was pressure on Japanese officals from foreign consulates and merchants for statistical information, the officials were rather quickly able to create statistics of the new foreign trade. Acknowledgments: I am greatly indebted to Professor Saitō Osamu, Hitotsubashi University, Tokyo, and Prof. Katsuta Shunsuke, Gifu University, for assistance in getting documents and for help with bibliographical enquiries and difficult passages. I am also indebted to two anonymous referees for much appreciated suggestions on restructuring the paper. Other obligations are acknowledged in the footnotes. I am alone responsible for the errors which remain. Appendix

Statistical Returns of Exports and Imports for Edo and Osaka, 1714– 1866 Year 17141

Port Coverage Osaka exports, 95 items imports, 119 items2

Source(s) i. Ōsaka Shiryō Hensanjo 1904 ii. Mitsui Bunko 1913 iii. OSSS, vol. 13, pp. 4–14

1724– 1730

Osaka exports, 11 items to Edo

i. Ōsaka-shi shi 1913, pp. 650–51 (source unidentified, annual totals), ii. Ōsaka hennen shi 1978 (breakdown by month3

1726

Edo

imports 11 items from all Japan

i. Katsu Kaishū 1890, pp. 238–39 (source unidentified) ii. Kyōhō tsugan ாಖ㏻㚷vol. 2.12, 1884 iii. Mikan zuihitsu hyakushu ᮍห 㝶➹ⓒ✀, vol. 17 5

1727

Edo

imports, rice

Report from Edo machibugyō to rōjū Kyōhō 12.7.12 6

1736

Osaka imports, 123 items

Ōsaka-shi shi 1913, pp. 770–797

174

1766 1761, 1788, 17969

EARLY JAPANESE TRADE

Osaka aggregate values of exports, and rice imported (quantity) Osaka exports, ginned cotton (total noted as destined for Edo, Kanto, and northern Japan)

“Meiwa nenkan Ōsaka shotōkei” in OSSS, vol. 13, p. 28 8 “Kuriwata kaitsugidonya kakihikae” ⧞⥥㈙ḟ㛛ᒇ᭩᥍in “Kuriwata shōgyō ki” ⧞⥥ၟᴗ グin Ōsaka hennen shi 1978, pp. 341–45

1804– 1840

Osaka Imports, 16 items at various dates in period10

Report by Osaka bugyō Abe 1842, in Ōsaka-shi shi 1911, pp. 639–86

1817– 1822

Osaka lamp oil shipped to Edo (average)

“Mizu abura ikken” ỈἜ୍௳, in Ōsaka hennen shi 1978, pp. 345–47

1817– 1826

Edo

oil imports from Osaka (average)

Kobayashi and Wakita 1973, p. 318

1827– 1832

Edo

oil imports from Osaka (average) 11

Kōda 1934b, p. 186

1832

Edo

oil imports (also gives quantities from Osaka)

Kōda 1934b, p. 186

1833

Edo

oil imports (also gives quantities from Osaka)

Kōda 1934b, p. 186

1833

Osaka

oil exports (figure likely to be understated)

“Abura yosedokoro goyōdome” Ἔᐤᡤᚚ⏝␃Ōsaka hennen shi 1978, pp. 347–48

c.1841

Edo

round figure for oil imports Report by head (gyōji) of (lists name of main centers Nihonbashi honfunechō tonya, of supply) general observa- 1841 Kōda 1934b, p. 180 tion made in 1841

1851, Edo 1852 (10 mos.), 1859

oil imports from Osaka

Kōda 1934b, p. 188

c. 1864

Osaka

Exports of oil

“Keiō gannen kokusō shiryō” ៞ ᛂඖᖺ✐಴ྐᩱin Kobayashi and Wakita 1973, p. 318

1856

Edo

imports, 126 items (with miscellaneous additional detail)

“Chōhōroku”12

1856– 1858

Osaka

imports from Ezo

Ōsaka-shi shi 1914, pp. 839–4013

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

175

1856– 186414

Osaka

6 items to Edo, 6 items 15

i. Ōsaka hennen shi 1978, pp. 348–50 16 ii. “Osaka machiburegaki” ኱ᆏ⏫ ゐ᭩cited by Takekoshi Yosaburō ➉㉺୚୕㑻, Nihon keizai shi ᪥ ᮏ⤒῭ྐ, vol. 4, pp. 718–1917

1866

Osaka

Imports, 20 items

BPP, vol. 4, p. 274, report on Osaka / Hyōgo

1

2

3

4

5

㸴

7 8

9

10 11 12

13

14

On the tables for 1714, see Ōishi 1964, pp. 2–31; Ōishi 1966, pp. 107–124, and Ōishi 1975, pp. 140–67. The table has also been reproduced in Ōsaka hennen shi 1978, pp. 305–13 and Ōsaka hennen shi 1965, pp. 199–208. Full title of tables: “Shōtoku yon kinoeuma nenjū shokoku yori Ōsaka e kitaru” ṇᚨᅄ⏥༗ ᖺ୰ᚑㅖᅜ኱ᆏỤ᮶and “Shōtoku yon kinoeuma nenjū Ōsaka yori shokoku e unkō” ṇ ᚨᅄ⏥༗ᖺ୰ᚑ኱ᆏㅖᅜỤ㐠⾜. Ōsaka-shi shi. A breakdown on a monthly basis was published in 1978 in Ōsaka hennen shi 1978, pp. 334–41, and more briefly in Ōsaka hennen shi 1965. It also included data for the month of January 1731. The source is not identified in this volume, but it is probably the document cited by Kobayashi and Wakita 1973, p. 110, as “Ōsaka yori Edo e tsukizuki sekisōdaka yose oboe” ᚑ኱ᆏỤᡞỤ᭶ࠎ✚㏦㧗ᐤぬ, drawn from a collection entitled “Tenmagumi sōdoshiyori Satsumaya Jinbē ‘hikaechō’” ኳ‶⤌᝷ᖺᐤ⸃ᦶᒇோර⾫ࠕ᡿ᖒࠖ. Stated by Ōishi Shinzaburō to be from a shahon now in Shiryō Hensanjo which ends with the words “Meiji jūkyūnen shigatsu Kazoku Mizuno Tadahiro zōsho o utsusu ᫂἞༑஑ᖺᅄ᭶ ⳹᪘Ỉ㔝ᛅᘯⶶ᭩ࣤ෗and also gives the names of copyist and proof-reader. Ōishi 1998, p. 73, and footnote 13, pp. 101–102. From a series of volumes on Edo customs and manners compiled by Mitamura Engyo ୕⏣ ᮧ㬇㨶(1870–1952) (Ōishi, 1998, footnote 13, pp. 101–102). A modern edition in twelve volumes was published in 1976–1978. “Kyōnenjū gotōchi chakumai no oboe” ཤᖺ୰ᚚᙜᆅ╔⡿ࡢぬin “Kyōhō sen’yō ruishū, beikoku no bu” ாಖ᧝せ㢮㞟⡿✐ࡢ㒊. Ōishi 1998, pp. 79–80. Also reproduced in OSSS 1964, p. 20–27, and in Ōsaka hennen shi 1978, pp. 321–31. Not reproduced in Ōsaka hennen shi 1978. Detail reproduced in Yasuoka 1990, p. 16. Yasuoka simply atttributed the volume to the Ōsaka Shōgyō Kaigisho (OCCI), without indicating that it had been published in OSSS, vol. 13. 1766 (stated to be figures for 6 months, but only 5 are given), 1788 (9 months), 1796 (9 months). The source is cited as “Kuriwata kaitsugidonya kakihikae” ⧞⥥㈙ḟၥᒇ᭩᥍, a document of which the original was dated Kansei 9.6.2 by the offical responsible for that year, prepared for the bugyō. Used by Yasuoka 1960 in creating his table of imports for 1736, 1804–30, 1840. Average attributed by Kōda to Narahara Kenjurō. “Chōhōroku: Edo omote shoshiki funaunsō nyūshin okazuke chakunidaka mitsumitsu otazune” 㔜ᐆ㘓Ụᡞ⾲ㅖⰍ⯪㐠㏦ධὠ㝣㝃╔Ⲵ㧗ᐦࠎᚚᑜ. Tōkyō-shi shi kō 1926, pp. 13–99. The table is reproduced in Ōsaka hennen shi 1978, pp. 331–34 which indicates ໭ᾏ⏘Ⲵཷ ၥᒇ⤌ྜἐ㠉ྐ, 3 as the source. The source is also cited in the introduction to the OCCI Supplement 1966, p. 1. This seems to indicate that an early book, rather than a manuscript, is now the only known source for the data. “Ōsaka machiburegaki” ኱ᆏ⏫ゐ᭩covers the second through fifth months of 1858, the eighth month of 1860 to the first month of 1861, and 1864.1.8–1864.2.1 (sic). Ōsaka hennen shi covers the second to seventh months of 1858, the eighth month to twelfth month of 1860, the first month of 1861, the eighth to twelfth months of 1864, and the first month

176

15

16 17

EARLY JAPANESE TRADE

of 1865. The figures available in Ōsaka hennen shi 1978, pp. 348–50 up to the seventh month of Ansei 5 (1858) are drawn from OSSS 1964a, vol. 13, pp. 90–97; and from the eighth month of 1860 from OSSR 1964b, pp. 191–206. After the detailed information for the early dates, the data as given in Ōsaka hennen shi 1978 become summary, with monthly gross totals only. For one of the six commodities, rice, quantities are small. For each commodity the quantities are certified montly by three officals (machidoshiyori) in Osaka. See note 14. Figures in Ōsaka hennen shi are different in detail from those by Takekoshi. Quoted in Kobayashi and Wakita 1973, p. 318.

REFERENCES

Ackroyd 1979 Joyce Ackroyd. Told around a Brushwood Fire: The Autobiography of Arai Hakuseki. University of Tokyo Press, 1979. BPP British Parliamentary Papers. Area studies, Parliamentary Papers, Japan, ed. W. G. Beasley, 10 vols. Shannon: Irish University Press, 1971–1972. Cullen 2003 L. M. Cullen. A History of Japan, 1582–1941: Internal and External Worlds. Cambridge University Press, 2003. Cullen 2006 ———. “Tokugawa Population: The Archival Issues.” Japan Review 18 (2006), pp. 129–80. Flershem 1964 Robert G. Flershem. “Some Aspects of Japan Sea Trade in the Tokugawa Period.” Journal of Asian Studies 23:3 (1964), pp. 405–16. Fujita 1987 Fujita Satoru ⸨⏣ぬ. Bakuhansei kokka no seijishiteki kenkyū: Tenpō-ki no chitsujo, gunji, gaikō ᖥ⸬ไᅜᐙࡢᨻ἞ྐⓗ◊✲㸸ኳಖᮇࡢ⛛ ᗎ࣭㌷஦࣭እ஺. Azekura Shobō, 1987. Hayami and Miyamoto 1988 Hayami Akira ㏿Ỉ⼥and Miyamoto Matao ᐑᮏཪ㑻. Keizai shakai no seiritsu ⤒῭♫఍ࡢᡂ❧. Vol. 1 of Nihon keizai shi ᪥ᮏ⤒῭ྐ. Iwanami Shoten, 1988. Hayashi 1969 Hayashi Reiko ᯘ௧Ꮚ. “Kinsei chūkōki no shōgyō” ㏆ୡ୰ᚋᮇࡢၟᴗ. In Ryūtsūshi ὶ㏻ྐ, ed. Kodama Kōta ඣ⋢ᖾከand Toyoda Takeshi ㇏ ⏣Ṋ. Vol. 1. Yamakawa Shuppansha, 1969, pp. 187–210.

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

177

Hayashi 1994 ———. “Provisioning Edo in the Early Eighteenth Century: The Pricing Policies of the Shogunate and the Crisis of 1733.” In Edo and Paris: Urban Life and the State in the Early Modern Era, ed. James L. McClain et al. Cornell University Press, 1994, pp. 211–33. Hirakawa 1994 Hirakawa Arata ᖹᕝ᪂. Kinsei chūkōki no shōhin ryūtsū to chiikikan funsō: Uraga bugyō to Edo bugyō no funsō shori ㏆ୡ୰ᚋᮇࡢၟရὶ㏻࡜ᆅᇦ 㛫⣮த㸸ᾆ㈡ዊ⾜࡜Ụᡞዊ⾜ࡢ⣮தฎ⌮. In Yamamoto Hirofumi ᒣᮏᘯᩥ, ed. Kindai kōtsū seiritsushi no kenkyū ㏆௦஺㏻ᡂ❧ྐࡢ◊✲. Hōsei Daigaku Shuppankai, 1994, pp. 48–73. Hotta 2007 Hotta Akio ᇼ⏣ᬡ⏕. “Kōda Shigetomo to Ōsaka-shi shi” ᖾ⏣ᡂ཭࡜ ኱㜰ᕷྐ. Ōsaka shunjū ኱㜰᫓⛅127 (Summer 2007), pp. 66–71. Katsu 1890 Katsu Kaishū ຾ᾏ⯚. Suijinroku ྿ሻ㘓1890, part 5. In Katsu Kaishū zenshū ຾ᾏ⯚඲㞟, 10. Keisō Shobō, 1978. Kobayashi and Wakita 1973 Kobayashi Shigeru ᑠᯘⱱand Wakita Osamu ⬥⏣ಟ. Ōsaka no seisan to kōtsū ኱㜰ࡢ⏕⏘࡜஺㏻. Osaka: Mainichi Hōsō, 1973. Kōda 1934a Kōda Shigetomo ᖾ⏣ᡂ཭. “Kabu nakama” ᰴ௰㛫. In Kōda Shigetomo chosakushū ᖾ⏣ᡂ཭ⴭస㞟, vol. 2, Edo to Ōsaka Ụᡞ࡜኱㜰. Chūō Kōron Sha, 1972, pp. 191–208. Orig. pub. 1934. Kōda 1934b ———. “Abura” Ἔ. In Kōda Shigetomo zenshū, vol. 2, Edo to Ōsaka Ụᡞ ࡜኱㜰. Chūō Kōron Sha, 1972, pp. 180–190. Orig. pub. 1934. McClain and Wakita 1999 James L. McClain and Wakita Osamu, ed. Osaka: The Merchants’ Capital of Early Modern Japan. Cornell University Press, 1999. McClain 1999 James L. McClain. “Space, Power, Wealth, and Status in Seventeenth-century Osaka.” In Osaka: The Merchants’ Capital of Early Modern Japan, ed. James L. McClain and Wakita Osamu, pp. 44–79. Cornell University Press, 1999. Matsuura 2007 Matsuura Akira ᯇᾆ❶. Edo jidai tōsen ni yoru Nicchū bunka kōryū Ụᡞ᫬ ௦၈⯪࡟ࡼࡿ᪥୰ᩥ໬஺ὶ. Kyoto: Shibunkaku Shuppan, 2007.

178

EARLY JAPANESE TRADE

Mitsui Bunko 1913 Mitsui Bunko ୕஭ᩥᗜ. Shahon 1913 from Tokyo University 1903 transcript of an earlier shahon possessed by Uchida Gonzō. Mitsui Bunko 1952 Mitsui Bunko ୕஭ᩥᗜ(Nakai Nobuhiko ୰஭ಙᙪ), ed. Kinsei kōki ni okeru shuyō bukka no dōtai ㏆ୡᚋᮇ࡟࠾ࡅࡿ୺せ≀౯ࡢືែ. Nihon Gakujutsu Shinkōkai, 1952. Mitsui Bunko 1989 ———. Kinsei kōki ni okeru shuyō bukka no dōtai ㏆ୡᚋᮇ࡟࠾ࡅࡿ୺ せ≀౯ࡢືែ. Tōkyō Daigaku Shuppankai, 1989 (new edition of Mitsui Bunko 1952). Miyamoto 1985 Miyamoto Tsuneichi ᐑᮏᖖ୍. Shio no michi ሷࡢ㐨. Kōdansha, 1985 (repr. 2002). Miyamoto 1972 Miyamoto Matao ᐑᮏཪ㑻. “Kinsei chūkōki no Ōsaka ni okeru ryōshu mai ryūtsū: Shohan Ōsaka nobosemai daka no kentō” ㏆ୡ୰ᚋᮇࡢ኱ 㜰࡟࠾ࡅࡿ㡿୺⡿ὶ㏻㸸ㅖ⸬኱㜰Ⓩ⡿㧗ࡢ᳨ウ. Kokumin keizai zasshi ᅜẸ⤒῭㞧ㄅ125:6 (1972). Miyamoto 1988 ———. Kinsei Nihon no shijō keizai ㏆ୡ᪥ᮏࡢᕷሙ⤒῭. Yūhikaku, 1988. Miyashita 1997 Miyashita Saburō ᐑୗ୕㑻. Nagasaki bōeki to Osaka: yunyū kara sōyaku e 㛗ᓮ㈠᫆࡜኱㜰̿㍺ධ࠿ࡽ๰⸆࡬. Osaka: Seibundō, 1997. Nagasaki-ken shi 1985 Nagasaki-ken shi taigai kōshō hen 㛗ᓮ┴ྐ ᑐእ஺΅⦅. Ed. Nagasakiken Shi Henshū Iinkai. Yoshikawa Kōbunkan, 1985. Nagazumi 1987 Nagazumi Yōko Ọ✚ὒᏊ. Tōsen yushutsunyūhin sūryō ichiran 1637–1833 ၈⯪㍺ฟධရᩘ㔞୍ぴ1637–1833. Sōbunsha, 1987. Nelson 1962 Andrew Nathaniel Nelson. The Original Modern Reader’s Japanese-English Character Dictionary. Rutland, Vt. and Tokyo: Charles E. Tuttle Company, 1995. OSSS 1964a Ōsaka shōgyō-shi shiryō ኱㜰ၟᴗྐ㈨ᩱ, vol. 13. Osaka Chamber of Commerce and Industry (OCCI), Ōsaka Shōgyō Kaigishō, 1964.

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

179

OSSS 1964 b ———, vol. 20. Osaka Chamber of Commerce and Industry (OCCI), 1964. OSSS supplement 1966 ———. Bekkan kaisetsu hen sōmokuji ูᕳゎㄝ⦅⥲┠ḟ(Supplementary volume, analysis, general list of contents). Osaka Chamber of Commerce and Industry (OCCI), 1966. Ōishi 1964 Ōishi Shinzaburō ኱▼ៅ୕㑻. “Kyōhō kaikaku ki ni okeru Edo keizai ni tai suru Ōsaka no chii: Kyōhō kaikaku ki ni okeru shijō kōzō ni tsuite” ா ಖᨵ㠉ᮇ࡟࠾ࡅࡿỤᡞ⤒῭࡟ᑐࡍࡿ኱ᆏࡢᆅ఩㸸ாಖᨵ㠉ᮇ࡟ ࠾ࡅࡿᕷሙᵓ㐀࡟ࡘ࠸࡚. Nihon rekishi ᪥ᮏṔྐ 191 (April 1964), pp. 2–31. Ōishi 1966 ———. “Shōtoku yonen Ōsaka ishutsunyū shōhinhyō ni tsuite” ṇᚨᅄ ᖺ኱㜰⛣ฟධၟရ⾲࡟ࡘ࠸࡚. Gakushūin daigaku keizai ronshū Ꮫ⩦ 㝔኱Ꮫ⤒῭ㄽ㞟3:1 (1966), pp. 107–124. Ōishi 1975 ———. Nihon kinsei shakai shijō kōzō ᪥ᮏ㏆ୡ఍♫఍ᕷሙᵓ㐀. Iwanami Shoten, 1975. Ōishi 1995 ———. Shōgun to sobayōnin no seiji ᑗ㌷࡜ഃ⏝ேࡢᨻ἞. Kōdansha, 1995 Ōishi 1998 ———. Kyōhō kaikaku no shōgyō seisaku ாಖᨵ㠉ࡢၟᴗᨻ⟇. Yoshikawa Kōbunkan, 1998, Ōsaka Shiryō Hensanjo 1904 Ōsaka Shiryō Hensanjo. Kōda shiryō no. 161. Shahon by Kōda Shigetomo 1904 from Tokyo University 1903 transcript of a shahon possessed by Uchida Gonzō. Ōsaka Dōjima beishō enkaku 1912 Ōsaka Dōjima beishō enkaku ኱㜰ᇽᓥ⡿ၟἢ㠉(containing records of Ōsaka Dōjima Beikoku Torihikijo. Osaka: Beikoku Torihikijo, 1912. Ōsaka hennen shi 1965 Ōsaka hennen shi ኱㜰⦅ᖺྐ, vol. 7, 1965. Osaka: Ōsaka Shiritsu Chūō Toshokan

180

EARLY JAPANESE TRADE

Ōsaka hennen shi 1978 ———, vol. 26. Osaka: Ōsaka-Shiritsu Chūō Toshokan, 1978. Ōsaka-shi shi 1913 Ōsaka-shi shi ኱㜰ᕷྐ, vol. 1. Osaka: Ōsaka-shi Sanjikai, 1913. Ōsaka-shi shi 1914 ———, vol. 2. Osaka: Ōsaka-shi Sanjikai, 1914. Ōsaka-shi shi 1911 ———, vol. 5. Osaka: Ōsaka-shi Sanjikai, 1911. Ōsaka-shi shi henshū no hyakunen 2002 Ōsaka-shi shi henshū no hyakunen ኱㜰ᕷྐ⦅㞟ࡢⓒᖺ. Osaka: Ōsakashi Shiryō Chōsakai, 2002. Ravina 1999 Mark Ravina. Land and Lordship in Early Modern Japan. Stanford University Press, 1999. Saitō 1998 Saitō Osamu ᩪ⸨ಟ. Chingin to rōdō to seikatsu suijun ㈤㔠࡜ປാ࡜⏕ άỈ‽. Iwanami Shoten, 1998. Saitō and Shinbō 1989 Saitō Osamu ᩪ⸨ಟ and Shinbō Hiroshi ᪂ಖ༤. Kindai seichō no taidō ㏆௦ᡂ㛗ࡢ⫾ື. Vol. 2 of Nihon keizaishi ᪥ᮏ⤒῭ྐ. Iwanami Shoten, 1989. Sekiyama 1958 Sekiyama Naotarō 㛵ᒣ┤ኴ㑻. Kinsei Nihon no jinkō kōzō: Tokugawa jidai no jinkō chōsa to jinkō jōtai ni kansuru kenkyū ㏆ୡ᪥ᮏࡢேཱྀ ᵓ㐀: ᚨᕝ᫬௦ࡢேཱྀㄪᰝ࡜ேཱྀ≧ែ࡟㛵ࡍࡿ◊✲. Yoshikawa Kōbunkan, 1958. Shinshū Ōsaka-shi shi 1989 Shinshū Ōsaka-shi shi ᪂ಟ኱㜰ᕷྐ, vol. 3. Osaka: Ōsaka-shi Shi Hensan Iinkai, 1989. Shinshū Ōsaka-shi shi 1990 ———, vol. 4. Ōsaka : Ōsaka-shi Shi Hensan Iinkai, 1990. Suzuki 1938 Suzuki Naoji 㕥ᮌ┤஧. Tokugawa jidai no beikoku haikyū soshiki ᚨᕝ ᫬௦ࡢ⡿✐㓄⤥⤌⧊. Orig. pub. Ganshōdō, 1938; repr. Kokusho Kankōkai, 1977.

STATISTICS OF TOKUGAWA COASTAL TRADE AND BAKUMATSU

181

Takahashi 2003 Takahashi Nanako 㧗ᶫ⳯ዉᏊ. “Kōda Shigetomo no keizaishi kenkyū to sono shiryō: Hitotsubashi daigaku fuzoku toshokan zō Kōda bunko o chūshin ni” ᖾ⏣ᡂ཭ࡢ⤒῭ྐ◊✲࡜ࡑࡢྐᩱ̿ᶫ኱Ꮫ௜ᒓ ᅗ᭩㤋ⶶᖾ⏣ᩥᗜࢆ୰ᚰ࡟. Keizai shiryō kenkyū ⤒῭ྐᩱ◊✲ 33 (March 2003), pp. 29–43. Takase 1966 Takase Tamotsu 㧗℩ಖ. “Kaga han no kaiun seisaku to Fushiki no funadonya” ຍ㈡⸬ࡢᾏ㐠ᨻ⟇࡜అᮌࡢ⯪ၥᒇ, Kaiji-shi kenkyū ᾏ ஦ྐ◊✲7 (1966), pp. 58–74. Tatemoto 1968 Masahiro Tatemoto. “The Correction of Official Import Statistics, 1874– 1887, for the Depreciation of Silver,” in Masao Baba and Masahiro Tatemoto, “Foreign Trade and Economic Growth in Japan: 1858–1937.” In Economic Growth: the Japanese Experience since the Meiji Era, ed. Lawrence Klein and Kazushi Okawa. Homewood, Ill.: R. D. Irwin, 1968, pp. 183–84. Tōkyō-shi shi kō 1926 Tōkyō-shi shi kō ᮾிᕷྐ✏. Kōwan hen  ‴⠍, vol. 3. Tōkyō Shiyakusho, 1926. Reprinted Kyoto: Rinsen Shoten, 1978. Wigen 1995 Karen Wigen. The Making of a Japanese Periphery, 1720–1920. University of California Press, 1995. Yamaguchi 1968 Yamaguchi Kazuo ᒣཱྀ࿴㞝. Nihon keizai-shi ᪥ᮏ⤒῭ྐ. Vol. 12 of Keizaigaku zenshū ⤒῭Ꮫ඲㞟. Chikuma Shobō, 1968. Yamamoto 1994 Yamamoto Hirofumi ᒣᮏᘯᩥ. Kindai kōtsū seiritsushi no kenkyū ㏆௦஺ ㏻ᡂ❧ྐࡢ◊✲. Hōsei Daigaku, 1994. Yasuoka 1960 Yasuoka Shigeaki Ᏻᒸ㔜᫂. “Ōsaka no hatten to kinsei sangyō” ኱㜰ࡢ Ⓨᒎ࡜㏆ୡ⏘ᴗ, in Nihon sangyō-shi taikei ᪥ᮏ⏘ᴗྐ኱⣔, no. 6, Kinki chihō hen ㏆␥ᆅ᪉⦅, ed. Chihō-shi Kenkyū Kyōgikai ᆅ᪉ྐ◊ ✲༠㆟఍. Tōkyō Daigaku Shuppankai, 1960, pp. 108–126. Yasuoka 1990 ———. “Zenkoku shijō no keizai henka” ඲ᅜᕷሙࡢ⤒῭ኚ໬. In ᪂ ಟ኱㜰ᕷྐShinshū Ōsaka-shi shi, vol. 4. Osaka: Ōsaka-shi Shi Hensan Iinkai, 1990, pp. 5–19.

182

EARLY JAPANESE TRADE

せ᪨ ᚨᕝᮇἢᓊ஺᫆࠾ࡼࡧᖥᮎ࣭᫂἞ึᮇᾏእ㈠᫆ࡢ⤫ィ ࣝ࢖࣭M࣭࢝ࣞࣥ ஧㒊ᵓᡂࡢᮏㄽࡢ୺ࡓࡿ┠ⓗࡣࠊᚨᕝᮇ࠾ࡼࡧ᫂἞ึ㢌15ᖺ 㛫࡟࠾ࡅࡿ᪥ᮏࡢ㈠᫆࡟㛵ࡍࡿ㈨ᩱホ౯࡛࠶ࡿࠋ➨୍㒊ࡣᮏ ✏࡛࠶ࡾࠊ➨஧㒊ࡣ Japan Review, No. 22 ࡟ᥖ㍕ࡢணᐃ࡛࠶ ࡿࠋ⌧Ꮡࡍࡿ᪥ᮏἢᓊ஺᫆㸦࠶ࡿ࠸ࡣᾏእ㈠᫆࡟ᑐࡍࡿゝⴥ ࡜ࡋ࡚ࡢ ࠕ ෆ 㝣 ࠖ ஺ ᫆ 㸧 ࡢ⤫ィ࡟ࡣὀ┠ࡍ࡭ࡁࡶࡢࡀ࠶ࡿࠋࣚ ࣮ࣟࢵࣃ࡟࠾࠸࡚ࡣࡇࢀ࡟ẚ⫪ࡍ࡭ࡁࡶࡢࡣ࡞࠸ࠋ኱ᆏ࡟㛵 ࡍࡿࡶࡢ࡜ࡋ࡚ࡣࠊከᒱ࡟ࢃࡓࡿ㍺ධရࡸὀពࢆせࡍࡿභရ ┠࡟㛵ࡍࡿᩘᏐࡀṧࡉࢀ࡚࠸ࡿࡀࠊᜍࡽࡃࡑࢀࡽࡣ኱ᆏ࠿ࡽ Ụᡞ࡬㐠ࡤࢀࡓရ≀࡛࠶ࢁ࠺ࠋ኱ᆏ ࡢ⥲✚Ⲵ㧗ࢆ♧ࡍ1714 ᖺࠊ1736ᖺࠊ࠾ࡼࡧ1766ᖺࡢ⤫ィࡣࠊ௒᪥࡛ࡣ㠀ᖖ࡟㈗㔜࡞ ࡶࡢ࡛ࠊࡇࢀ௨እ࡟⥲ィࢆ♧ࡍᩘᏐࡣṧࡉࢀ࡚࠸࡞࠸ࠋᾘ㈝ ࡢ୰ᚰ࡛࠶ࡿỤᡞ࡟㛵ࡋ࡚ࡣࠊ㍺ฟࢆ♧ࡍグ㘓ࡣ࡞࠸ࠋ၏୍ ▱ࡽࢀࡿỤᡞἢᓊ஺᫆ࡢᩘᏐࡣࠊ1723ᖺ࠿ࡽ1730ᖺ࡟ࢃࡓ ࡿࠊ11ࡢရ≀ࡀ኱ᆏ㸦㍺ฟ㸧࡜Ụᡞ㸦㍺ධ㸧࡟グ㘓ࡉࢀ࡚ ࠸ࡿࡶࡢ࡜ࠊ኱ࡁࡃࡑࡢໃ࠸ࢆ┒ࡾ㏉ࡋࡓ1856ᖺࡢࡶࡢࡢ ࡳ࡛࠶ࡿࠋᾏእ㈠᫆࡟ࡘ࠸࡚ࡣࠊ1859ᖺࡢ㛤 ௨᮶ࠊࡑࡢ⤫ ィࡣ⣡ရ᭩࡞࡝࡟ᇶ࡙ࡁᖥᗓᙺே࡟ࡼࡗ࡚┤᥋グ㘓ࡉࢀ࡚࠸ ࡿࠋࡑࢀ௨๓ࡣࠊᙺேࡓࡕࡣἢᓊ஺᫆ࡢၥᒇࡸ⤌ྜ࡞࡝࡟ࡼ ࡗ࡚ᥦ౪ࡉࢀࡓ㈨ᩱ࡟౫ᣐࡋ࡚࠸ࡓࠋ

Source: Japan Review, No. 22 (2019), pp. 59–101

8 Post-1859 Foreign Trade

Statistics of Tokugawa Coastal Trade and Bakumatsu and Early Meiji Foreign Trade PART 2: TRADE IN BAKUMATSU AND EARLY MEIJI TIMES

™

Evaluation of archival sources on Japanese trade in the bakumatsu period (1853–1867) and in the first fifteen years of the Meiji period is the primary purpose of this two-part essay, of which Part 1 appeared in Japan Review, 21 (2009). After the opening of the ports in 1859, statistics were directly compiled by bakufu officials from traders’ invoices. Previously, officials had relied on data furnished by wholesalers (tonya) for coastal trade. The notional parity of a gold yen and a silver yen in the closed gold standard put in place in 1871 broke down in the open market for the Mexican dollar, the accepted unit of account for foreign trade in the newly opened ports. It was only as gold disappeared and Japan’s currency system became de facto a silver one that a conversion emerged in national accounts of totals from dollars (or silver yen) into gold yen at close to a one-for-one rate. Arguably, Japanese officials coped well in the early Meiji years with the challenge of creating Western style statistics, even if variant grand totals emerged in the conversion of dollar figures into gold yen in the uncertain monetary conditions. Archival information on coastal trade in earlier Tokugawa times rested on copies. Those copies are few and remote from the originals. This is true also for the new foreign trade of the 1860s. Even under central direction 183

184

EARLY JAPANESE TRADE

from 1869, only a few copies, mostly drafts rather than final versions, are known for the new aggregates (from 1869, in gold ryō; from 1871, in gold yen). The basic dollar totals forwarded regularly from individual Custom Houses to the central authorities, the key raw material for converting dollar grand totals into gold yen until the 1880s, are known almost exclusively from figures in the British Parliamentary Papers. LEARNING THE ART OF FOREIGN TRADE STATISTICS

If Japanese officials in the bakumatsu years had wished to maintain their trade accounts in the old style without being importuned by others to change, it would have been impossible to do so. Coming from societies already accustomed to the collection and use of copious statistics, representatives of Western nations sought to document Japan in statistical terms. Japanese officials found themselves badgered not only by consular officials but also by foreign merchants. The first foreign Chamber of Commerce was constituted in Nagasaki 㛗ᓮ in June 1861, and prominent among its objectives was the aim “to compile and publish a statement of trade, and otherwise assist to make the resources of the country generally known.” Chambers of Commerce were formed in other open ports, as well, and sought information from the Japanese officials, duplicating to some extent what Western consular officials were doing. Customs management lacked a precise institutional identity in the early years of open ports. After the Dutch-Japanese treaty of 1857, the Kaisho ఍ᡤ, which had handled foreign trade in Nagasaki, at first exercised this role, but lost its managerial function in controlling Western trade. Western diplomats from the outset referred to the Japanese office where customs matters were administered as the “custom house,” a usage that implied a highly specialized office manned by career specialists, close to the water’s edge. The first term used in Japanese to denote “custom house” was unjōsho 㐠ୖᡤ, which by decree of the Finance Ministry (Ōkurashō ኱ⶶ┬) in 1872 (11th month) became zeikan ⛯㛵. In Nagasaki in 1863, the fromer kaisho was restyled unjōsho, and a decade later the designation duly became zeikan. At the outset of the Meiji era, the unjōsho were responsible to the Foreign Affairs Agency, which in 1869 finally became the Foreign Ministry (Gaimushō እົ┬). Foreign Ministry reporting of foreign trade was surrendered in 1871 to the Ministry of Finance. Rutherford Alcock, who arrived in Japan in 1858 as the first consul general of Great Britain, complained in March 1860 that the want of a system, combined with the desire of individual merchants for

STATISTICS OF TOKUGAWA COASTAL TRADE

185

secrecy, rendered it impossible for foreign representatives to get details of trade.1 He had noted in the preceding November that the Japanese themselves admitted that they lacked both a system and knowledge of how to transact business.2 Many Western consular officials were like Alcock in having a background in China, where foreigners ran the customs departments in the ports, and they had little sympathy for the Japanese or understanding of their predicament. A passing post-1868 reference to the contrasting circumstances in Japan and China hints at a British lack of confidence in statistical collection that was not managed by Europeans. This lack of confidence was most explicit in the comments at Hyōgo/ Osaka රᗜ ኱㜰by Gower, the official most hostile in those years to Japanese management. He maintained in 1872 that the defects in the statistics were “likely to be repeated in future unless the Japanese government adopt, for a few years at least, a foreign inspectorate to protect their commercial revenues, as well as the interests of honest merchants.”3 The judgments of men who had minimal background in Japanese affairs at the time of their appointments tended to be unduly harsh. The British made no allowance for the fact that Japanese officers were making the adjustment from a situation in which no taxes had been levied on the transit of goods to one in which taxes were collected; and from an informal statistical framework which was a passive and intermittent collation of data, provided by wholesalers or tonya ၥᒇwith little central guidance, to one in which these officers of the bakufu were creating statistical compilations from raw returns, of which they were for the first time the originating collectors. For the Japanese, the problem was in part one of imposing a new statistical discipline on officials who had been trained in the old school of administration. Outside Nagasaki, it was rather a case of creating functionaries with no previous experience of foreign trade in any fashion. In the first decade of open ports, the accounting operation was an appendage of the office of bugyō ዊ⾜magistrate), and each office was independent of scrutiny or direction from the center for its routine working. In the 1860s, less than wholehearted cooperation is to be explained more by reluctance to yield to unwelcome and pushy foreigners than by opposition to new work routines. Though consuls quite quickly got access to the books, they were initially unsuccessful in getting access to statistics of total trade, whether compiled Japanese style on a monthly basis or added up at year’s end. Nagasaki presented the problems more acutely than Kanagawa ⚄ዉ ᕝ. As late as November 1860, a British representative complained that details of trade could not be obtained at the Custom House “owing to

186

EARLY JAPANESE TRADE

the want of method, dilatoriness and procrastination of those officials.”4 In 1859–1860, when Custom House officials withheld statistical estimates from consular staff, foreign diplomats attempted to make sense of the actual registers of “applications” for permits, which Japanese officials reluctantly allowed them to consult. The report of the consul at Kanagawa for the second half of the year 1859 was based on examination of such Custom House records, and on what he described as “books kept at this office.”5 British officials sent clerks to the Custom House to copy out particulars from permits, or they paid a customs official a monthly fee to supply them with the details.6 They remarked that the “applications,” a vital source of trade detail filed by merchants in order to acquire the permits authorizing exports or imports of goods (“entries” in Western customs jargon), were readily accepted at the Custom House, even if carelessly filled in. Given access to some details of Custom House paperwork but unable to obtain overall figures of trade, consular officials balked at attempting an independent assembling of the data and calculation of grand totals. For Nagasaki, the crude calculations for the latter half of 1859 were based on figures from Shanghai, through which most of Nagasaki’s trade was routed.7 For 1860, the Nagasaki return was again based on figures of imports from Japan recorded at the Shanghai Custom House, plus an estimate of exports from Nagasaki to Hong Kong and England, which were valued at 463,760 dollars, and converted into sterling at 5s.0d. to the dollar.8 In early 1860, Alcock recorded his belief that the trade of the three open ports for the first six months after opening amounted to £1 million sterling.9 Figures quoted for Kanagawa for the second half of 1859 in a report of 26 April 1860 were based partly on the Shanghai port books.10 In the course of 1860, British officials at Kanagawa were still without regular access to Japanese Custom House counts; they recorded detail month after month, but did not express clear-cut totals for accounting periods, whether monthly or longer.11 The low opinion held by British officials of the work of the Custom Houses was greatly colored by their access to the applications for permits to export or import. They do not seem to have been aware that grand totals of trade were being withheld from them, and hence they saw things simply in terms of a lack of systems. In retrospect, we might judge that the details of their registers could be faulted, but Japanese Custom House officials in the first years of open ports probably executed their work more fully and more rapidly than frustrated British representatives ever realized.12

STATISTICS OF TOKUGAWA COASTAL TRADE

187

Only the faintest traces of the early work of Custom House officials now exist. The only thorough study of the pre 1868 sources is that by Ishii in 1944.13 He was familiar with British Parliamentary Paper figures, and with M. B. T Paske-Smith’s published work.14 He also gathered sparse Japanese data, and exercised a real critical faculty in attempting to resolve the contradictions in much of his data. The dismissive comment by Sugiyama Shinya ᮡᒣఙஓthat “for the period 1859–67 Ishii Takashi ▼஭Ꮥhas given his own estimates, but his figures seem to have been put together in a random fashion from different sources which he considers appropriate rather than according to a coherent system” is not only unfair but inaccurate.15 The comment entirely ignores the scarcity and heterogeneity of data, and the problems posed thereby. It is also not the view of others. Baba and Tatemoto regarded Ishii’s as the most reliable examination of the figures.16 Ishii had quite carefully worked out his figures, first constructing a table for the years 1859–1867, albeit with some gaps, based on a short note published in 1895 by Kawai Toshiyasu Ἑྜ฼Ᏻ.17 Kawaii’s tables are a one page communication, and do not warrant comment, except that his tables are an indication that BPP papers were becoming more widely known in Japan in the 1880s and 1890s.18 Ishii then put together a second table presenting figures compiled by Paske-Smith in the 1920s from sources that he described as “obtained at the Custom House by H. B. M. Consuls.” These are from the archives of one or more British consular or diplomatic offices in Japan. A third table, confined to Yokohama data for 1859–1867, was originally published under the auspices of the Yokohama Chamber of Commerce and Industry (Yokohama Shōgyō Kaigisho ᶓ὾ၟ ᴗ఍㆟ᡤ) without any indication of its origins.19 To avoid any confusion, it should be noted at this point that, though the foreign trade was conducted almost from the outset at Yokohama, the British consulate at first remained at Kanagawa, the original site envisaged for foreign trade. It continued moreover to be designated by the British as the Kanagawa consulate long after it had moved to Yokohama itself. Kawai Toshiyasu stated in his article that he drew figures from consular reports to the British Foreign Office. Not content to accept all of Kawai’s numbers, Ishii supplemented those with data from other sources for years in which grand totals were lacking in the BPP, and for years in which there were variants for ports. The BPP figures were themselves becoming known in Japan by the 1880s. Serial data for all three open ports for the years 1860–1867, with a few gaps, were printed in the fourth issue of Kokka gakkai zasshi ᅧᐙᏥ᭳㞯ㄅ(15 June 1887),

188

EARLY JAPANESE TRADE

and although no source was identified, the data were quite clearly drawn from British consular reports.20 In transcription, errors inevitably slipped in. There are three deficiencies or errors in Kokka gakkai zasshi data, two for 1863 and one for 1865. Paske-Smith himself made a mistake for 1865, giving a slightly rounded figure of 1,560,800 dollars when, in fact, the number should read 560,788.21 The points we should not miss are these: Ishii’s figures vary on occasion from figures in Kawai, BPP, and Paske-Smith, and some of the figures of Paske-Smith differ from those in the BPP.22 The most disconcerting challenge facing Ishii, perhaps, was that the BPP give a range of figures, not a single figure, for Yokohama in the year 1863. He opted for the high totals of 10,554,012 dollars for exports, twice the amount of a variant figure in BPP and of the figure given by Paske-Smith.23 He also opted for 3,244,584 dollars for imports, disregarding without discussion a higher figure for imports of 3,474,749 dollars that appeared in Table 3 of his own volume.24 He left aside the very high exports total of 13,749,985 dollars that Kawai had offered; how that total was arrived at is unknown. However, it is likely that there was an error of transcription in which a “1” was erroneously inserted in front of a figure of 3,749,985 (fairly close to 3,704,484, the figure in Ishii’s Table 3), thus inflating the total to 13,749,985. The results of Ishii’s work in his Table 5 are the best available for the subject.25 With the exception of the year 1863, export figures from 1862 onwards tend to accord in all sources, and not just in Paske-Smith. The real divergence is in imports. There are marked differences in the figures for Nagasaki and Yokohama in the table in the modern Yokohama-shi shi ᶓ὾ᕷྐ.26 Imports to Yokohama in 1860–1862 as recorded in Yokohama kaikō 50 nen shi ᶓ὾㛤 ஬༑ᖺྐare even higher.27 The timing of consular access to grand totals of trade can be dated roughly for Kanagawa. In or before March 1861, Consul Howard Vyse had access to “declared” figures for 1860 (that is, the values declared at the Custom House and entered in its statistical record).28 In Nagasaki, the problem of access lingered on. Though figures for 1860 and 1861 later became known to the officials there, the absence of figures for 1861 suggests that a breakthrough in official communications did not occur until 1863. Finally, the BPP report for the year 1862 gives Nagasaki numbers denominated in taels, a money of account used by officials in Nagasaki. According to the BPP report, the figures were of Japanese origin and they had come quickly into the hands of consular officials.29 Alcock showed in his book, The Capital of the Tycoon: a Narrative of Three Years Residence in Japan published in 1863, how things

STATISTICS OF TOKUGAWA COASTAL TRADE

189

had improved. With figures in his possession either when he left Japan in March 1862 or coming into his hand afterwards, he was able to give figures for Nagasaki in 1861 “according to the returns made up on the spot” (i.e., from the Custom House) and for Yokohama in 1860 and 1861.30 For 1859, Ishii used what he identified very baldly as bugyōsho records (bugyōsho no kiroku ዊ⾜ᡤࡢグ㘓) for Nagasaki,31 and he also quoted a contemporary Japanese source for Hakodate ⟽㤋.32 For the years 1860 and 1861, he was able to drawn on two Japanese sources for Hakodate.33 For 1863, he quoted an export figure of 5,116,634 dollars for Yokohama from a Japanese source, slightly less than the better known BPP figure of 5,134,185 dollars.34 The fact that Japanese totals were becoming accessible in the early 1860s is also suggested by the evidence which Paske-Smith cited six decades later. Drawing on what appear to be older consular sources in Japan, he was able to give rounded figures for Nagasaki in 1859 and 1860, and for Yokohama in 1860. It seems probable that these figures came into consular possession as British relations with the Custom Houses improved, although too late to feature in reports filed by Alcock and his contemporaries in the early 1860s.35 The regularizing of information sharing by Custom Houses with consulates can be traced easily enough. Whatever the vagaries of access in the case of Nagasaki figures for 1859–1861, figures for 1862 were received and reported promptly. The consul’s problems in converting the 1862 totals from taels and the fact that figures for 1862 are missing from Paske-Smith’s source hint, however, at the confused situation of the time. This obviously continued the following year, as the report for 1863 was submitted only on 3 January 1865, and in response to a formal request from the legation in Edo. The tenor of the report implies that the data were already in the consulate. A consul, who was new to Nagasaki, commented that he was not “able to form a correct idea of the real value of trade during that year [1863], which the enclosed returns only represent in part,” giving the impression that he had failed to take any action on the return which was already in the office.36 In contrast to the situation in Nagasaki, information for Kanagawa (Yokohama) was provided in a timely manner. The consular report of 31 January 1863 gave figures for 1862 and retrospectively quoted the data for both 1860 and 1861, showing how the communication of statistical totals from Custom House to consular station had been established on a firm basis at Kanagawa.37 Prompt provision of data to the British was repeated a year later for the 1863 figures.38

190

EARLY JAPANESE TRADE

In Nagasaki, it looks as if the problems resided wholly or in part in the consulate. The confusion over currency and the incomplete coverage of the data in the return for 1862 and the fact that the report for 1863 was not received in Edo until 1865 suggest as much.39 The transition to regular reporting was achieved in Nagasaki only in January 1865 when the 1864 report was conveyed to British officials. By contrast, the change for the better was signaled in Yokohama from the time of the report for 1862. Hakodate reports began to become regular from the time of return of data for 1864. While a return for 1861 for Hakodate appears in BPP early in 1862, the return for the port for the year 1862 is simply a list of imports and exports without a grand total.40 Later, the report of 19 February 1864, which provided the 1863 figures, also added up the export figures for 1862.41 Finally, the 1863 figures in their prompt receipt (on 1 January 1864!) which were followed later by punctual data for 1864 and 1865 set the pattern of future smooth statistical reporting.42 Consular contact was primarily with the Custom House officials. It is less clear how much officials had to call on the bugyō in person for intervention in the early years. Bugyō goodwill was certainly forthcoming after mid decade. This can be seen in the trouble taken by the bugyō of Osaka, not yet a port, to provide up to date population census figures and also remarkable details of the port’s coastal trade for 1866 (See Appendix 2). These 1866 data, the sole data for several decades on the port’s general import trade, seem to contradict the widespread assumption of a bakumatsu breakdown in bureaucratic standards.43 They also serve to underline that, while Western style statistics demanded an entirely new approach, there existed enough of a working statistical awareness to facilitate a response to new challenges. From the time of receipt of the Nagasaki data for 1864, aggregated totals for all the ports could be compiled. They were incorporated into the first general British consular survey of trade, for the year 1869, which appeared in a report dated 1 March 1870.44 In Nagasaki, returns in taels were the norm in the early 1860s and returns in ichibus in the mid 1860s, but thereafter returns for all three open ports were in dollars. Even after 1873, when good official Japanese reports began to appear, British reports continued to be based primarily on the returns to the legation from individual consular stations. The reason for this was essentially the hybrid nature of Japanese currency. The Japanese central returns of trade converted dollars into gold yen, in keeping with the bookkeeping requirements of the nominal national yen gold standard of

STATISTICS OF TOKUGAWA COASTAL TRADE

191

1871. However, as the dollar remained the accepted unit of account of the ports (apart from a few early years in Nagasaki), the British returns were simply observing the commercial reality for Japanese officials; the dollar was the currency in which the accounts given by the Custom Houses to consular officers were denominated. Although a majority of transactions was conducted in local currencies, the aggregation of taxes and of gross totals of trade was presented in dollars alone. Before 1868, no national trade totals (that is, aggregation of all returns from the ports) contemporary with the year of account exist; this suggests that gross national aggregation came into existence only with the emergence in 1869 of the ryō and, later, the gold yen. A characteristic of the 1860s is the existence of variant figures. Variants in BPP figures for Hakodate in several early years or in the BPP figures for Nagasaki for 1864 can be regarded as reflections of imperfections or confusion in the early stages of regular reporting, but it is harder to explain why they exist also in some of the figures of the second half of the 1860s.45 A few of these variations are of substantial proportions. Export figures in Paske-Smith and Yokohama-shi shi are identical for 1866 at 14,100,000 dollars, but they give different import figures: 11,430,000 in Paske-Smith and 11,735,000 in Yokohama-shi shi.46 For 1867, exports and imports in BPP are larger than figures in both Yokohama-shi shi and Paske-Smith.47 In both pre and post Restoration sources, different gross totals may reflect one or more of several facts: they may have been drawn from incomplete original, or intermediate, workings of figures; they may or may not include bullion movements or coastal trade to or from open ports; and errors of transcription may have occurred. The preface to the first modern monthly reports in 1873 noted that the details of earlier statistics had not been “carefully prepared.”48 The only sure fact is that variations exist; beyond that, one enters the realm of speculation. The published consular reports provide the best guide. They are the sole series containing the original dollar denominations trade declared by the Custom Houses. This feature distinguishes them from records compiled using later official conversions from dollars into gold yen in Tokyo. It seems prudent to adopt the view of the Yokohama-shi shi volume on early trade statistics: the figures in the BPP are to be preferred not only for the early years, but for the entire period up to 1884.49 However, the two most substantial cases of contradiction in the early 1860s prove on closer examination of the sources to be superficial. In the Nagasaki figures for 1863, the problem was confusion in converting from taels to dollars, a confusion that had already existed for the

192

EARLY JAPANESE TRADE

1862 figures (see Appendix 3, A). In the case of the Yokohama figures for 1863, the confusion arose solely from a badly structured report (see Appendix 3.B). Dollar totals in the main text of the consular report on that year are the “declared values” at the Custom House, but the sterling totals are not conversions of the declared values; they are conversions into sterling based on higher gross dollar totals that had been created in consular calculations set out solely in the appendix to the consular report. The figures were dramatically inflated by a consular revaluation of the price of silk. The consular calculating operation was not made clear in the report, where the figures appear to be sterling equivalents of the declared values. In other words, there is neither a serious error in the figures, nor a gross confusion by the consul, but a contradiction, apparent rather than real, caused by the lack of an unambiguous statement drawing attention to the different basis of the figures in the appendix.50 In a paradoxical way, these artificial divergences in the returns for Nagasaki and Yokohama actually validate the basic soundness of the figures. As they got access to more figures, British representatives in Japan became more critical, and their complaints became more substantive. In essence, their criticism was three fold. The first point related to permits accepted by Custom Houses and the accuracy of the quantities stated on them. My own judgment of this type of complaint is that, while defects could indeed be detected, consuls exaggerated the extent to which the returns were defective. The second criticism was that Custom Houses too readily accepted declarations by merchants containing undervalued prices. British observers often noted undervaluation of prices in the Custom House as a failure of the statistics. The consequences of revision could be dramatic, as in 1863 in Yokohama, where silk was the dominant export, and the understated price was revised by the British in their return of trade for the year. Sugiyama asserts that all goods were undervalued, and that imports were undervalued more than exports.51 This weakness did exist, in my view, but it was not a fundamental statistical problem, as some modern writing has argued. There are modern commentators who failed to make a clear distinction between routine underestimations in entries declared at the Custom House (inevitable in an ad valorem system), and a more technical issue relating to conversion problems when gold currency was converted at a fixed and artificially high value on silver set by government fiat to be observed by Custom House officers in converting prices in gold into prices in silver. This latter point constitutes the third criticism that occurs in the contemporary consular comment, and it is echoed in the modern literature.

STATISTICS OF TOKUGAWA COASTAL TRADE

193

There was never a problem in invoices for exports which were declared in silver, the currency of the ports, nor for imports coming from silver currency countries, where any conversion was a straightforward operation between two currencies backed by silver. It arose only for imports from gold currency areas. In that case, the problem arose not for reasons of exchange, as market rates of exchange were well established and known in the ports, but because Japanese officials were tied to converting invoices in gold currencies at a fixed rate. For purposes of conversion of invoice prices in gold currencies, the Custom Houses adhered to a valuation of 5.90 taels per silver dollar, the rate that existed at the time of opening of the ports.52 Although this rate remained fixed in the Custom House calculations, by 1863 the open market exchange rate had shifted to 3.37 taels to the dollar. Expressed in different terms, the amount of sterling (a gold currency) that could be acquired with one silver dollar on the open market fell from between 5s.0d. and 5s.6d. to 4s.6d. In the market, then, but not in Japanese Custom House accounting, the value of silver currency had declined against that of gold currency. While aggregates for imports from silver currency areas were realistic, since they reflected market rates, total imports were deflated by understated figures for goods from countries with gold-based currencies. If gold was undervalued against silver in the exchange rate, conversion from gold to silver by a simple process of arithmetic automatically deflated the total value. The situation may be illustrated by a simple hypothetical example of an invoice for £1000 sterling, official exchange rate of 7s.6d. sterling to the dollar, and current market rate of 5s.0d. Conversion at the official rate would yield a dollar total of merely 2,664 dollars, while conversion at the market rate, used by merchants in their private accounting, would amount to 4,000 dollars. The problem in the early 1860s was a temporary one, and it was protracted by Alcock’s reluctance to agree to change, as the rate was advantageous to foreign traders in settling accounts at the Custom House.53 The Tokugawa authorities in 1865 accepted a proposal from foreign countries that gold prices of imports should be converted at real exchange rates, with the result that overvaluation did not occur in the statistics of 1865 and 1866.54 IMPROVEMENTS IN STATISTICS FROM THE MID 1860S

British representatives may have had doubts about the quality of the registers of permits or of the figures, but they conceded at a very early stage that progress had been made as far as access to data was concerned.

194

EARLY JAPANESE TRADE

Already in 1863, a new British official, Winchester, implied in an otherwise censorious report on Nagasaki that ready access to the records was at last now established.55 The report for Hakodate for the same year found Japanese officials obliging, although two years later the British were alleging that the methods of the Hakodate Custom House were irregular.56 Starting with the reports for the year 1865 (submitted in early 1866), consuls were able to forward annual returns of trade for the open ports to the British legation in Edo with very little delay. Indeed, reports from consular stations with figures for the preceding year were often dated 1 January of the following year. This rapidity, astonishing by European standards, was made possible by the fact that, repeating the pattern of Tokugawa times, monthly summaries were circulated to interested parties or, if not circulated, at least accessible. To arrive at a cumulative total for the year, Japanese officials had only to add figures for the final month of the year to the total for the previous months, which was already known. Consular officials seem to have relied on the annual figures from the Custom Houses, and either not to have received the monthly returns or to have ignored them. In observing in the spring of 1873 that the annual Kanagawa total for 1872 had been made from the aggregation of the monthly totals received at the consular station, the consul seems to have been noting the fact as exceptional.57 A swift aggregation of figures within Custom Houses suggests an absence on the part of Japanese officials of any process of revision or checking. The existence and retention of monthly figures would seem to account for the fact that, despite the loss of the records of the Custom House at Yokohama (Kanagawa) in late 1866, returns existed for much of the year. In applying to the Yokohama bugyō for “the usual Custom House returns” in early 1867, the Kanagawa consul was informed that they had been destroyed in a fire the previous November. The same fire had destroyed the records of the Chamber of Commerce, and no return was possible for 1866.58 In contrast to Japanese officials who possessed information for the preceding months, the Chamber of Commerce and consulate seemed totally bereft of figures. The consular station was unable to provide figures for 1866 to the legation in Edo, but after combing through the consular archives more than sixty years later, Paske-Smith was able to cite figures (presumably for the first ten months of 1866). This suggests that monthly returns had been received, but ignored at the time. In May 1868, a British consular officer took the liberty to pen an unprecedented adverse comment on the Kanagawa statistics. Writing up the report for the year 1867, having

STATISTICS OF TOKUGAWA COASTAL TRADE

195

had to use the 1865 figures as a reference point in the absence of figures for 1866, he remarked contemptuously: Although the desire always has been to accept the returns of the Custom House, such as they were, it was found impossible to do so for 1865, those furnished being found, after having been made up, so much below any reasonable estimate that they had to be thrown aside as worthless, and information sought in other directions.59

Turnover of consular personnel may be relevant. Myburgh, who wrote the report on 1865, had taken up duty only on 4 January 1866, and may not have had sufficient time to make himself conversant with the figures; at any rate, he said little. Fletcher, the author of the hostile critique of the 1865 figures, was also a new man, and he supplemented his adverse comments with the self-justifying remark that “it is only lately that returns came into my hands and afterwards they had to be arranged into some sort of system.” Neither man appears to have had easy familiarity with, let alone, mastery of, the figures that they commented on. The absence of any partial figure for 1866 shows that the consul was unaware of the existence of monthly returns, and suggests little statistical engagement with the Custom House officials. In applying to the Custom House for a return for 1866 – thereby exposing an attitude of passive reliance on Japanese officers that would have been unnecessary for a consul who kept in regular contact – the British representative was unaware that the November 1866 fire had destroyed all the records of the Custom House. Her Britannic Majesty’s consular officers’ expressions of discontent raise as many questions about the officers themselves as about the Japanese in these years. Criticism by British officials was directed primarily towards the process of handling the permits (which incorporated the detail from merchants’ formal applications for them, and hence sanctioned the movement of goods), whether in terms of valuation of goods or deficiencies in the quantities recorded. It was claimed that the Nagasaki Custom House returns for 1862 recorded exports of 554 piculs of silk, whereas an account maintained by “mercantile agencies” reported 967 piculs.60 As for prices, teas were regarded as undervalued in the Nagasaki returns, and British consular reports made upward adjustments to arrive at higher figures. Over-ready acceptance by the customs officials of the entries submitted at the Custom House by merchants was set out at length by Winchester at Nagasaki in his report on

196

EARLY JAPANESE TRADE

1862.61 As duties were on an ad valorem base, there was a real incentive for undervaluing the amounts, and the British believed local merchants were taking advantage of the innocence or inexperience of the Japanese officials. This was particularly the case for imports, with which many of the Custom House officers were unfamiliar; under valuations for exports were less marked. Undervaluation was of course never to go away entirely in a system in which goods were taxed ad valorem. By 1880, the view was that both imports and exports were undervalued though this was by now a minor issue. Exports tended to include prime cost only and shippers tended not to add in the extra cost incurred between purchase and shipment (that is, prices in practice were not on an f.o.b. (free on board) basis, and hence including all costs up to that point).62 Whatever the deficiencies in Japanese figures, however, it is worth noting that the criticism is based on a new-found sophistication, which was emerging even in Britain only in and after the 1850s. In a longer time perspective, it is not productive to damn the figures; that merely hinders our progress toward understanding the economic phenomena of which they are indicators. As remarked earlier, modern writers have been confused by a distinction, unrealistic in statistical terms, made by nineteenth century British officials between the Custom House statistics of imports and calculations made primarily by the Chambers of Commerce (but widely quoted by consular offices) of what they described as the “disposal” or sale of the goods. Few contemporary consular and trade officials were troubled by the confusion that has afflicted some historians and commentators. The estimates of disposals rested on official Japanese data of individual commodities adjusted by intelligence – perhaps not uniformly of high quality – supplied by consular officials, individual merchants and, above all, the Chambers of Commerce. The distinction was between customs figures, seen as a record of goods landed and warehoused, and returns emanating from the Chambers of Commerce and intended to record the “disposals” by importers to buyers. Disposals consisted of sales for the year, based on imports, adjusted by change in inventory on hand during the same period. As late as 1882, Sir Harry Parkes, the British minister who was a driving force behind the statistical work of his officials, described the customs figures as “a record of importation and not of consumption.”63 This ignores the fact that the Custom House figures measured in concrete terms the transfer of goods between countries, whereas the consular and Chamber of Commerce officials in the open ports put their energies into a bookkeep-

STATISTICS OF TOKUGAWA COASTAL TRADE

197

ing exercise of their own devising that was susceptible to subjective manipulation of Custom House data. We can either take on trust their confidence in the merit of the estimates they produced, or we can have doubts. There are reasons for doubting the adequacy of the information from trader associations, and there is no evidence of British consular staff, at least, sharing statistical information with foreigners other than the Americans. It has been suggested that the statistics first reported as consular figures owed as much to the Chambers of Commerce as to the consuls. Sugiyama has expressed the sweeping view that “the method of compilation is not clear.”64 However, the methods of compilation are not in doubt. The Chambers of Commerce and consular officials alike relied on the primary data from the Custom Houses, and for some of their observations they drew on trade information collected by individual merchants. Upon gaining access to the Custom House records, chamber officials or consular officials sometimes simply re-aggregated the quantities for individual commodities, and also adjusted prices. These data rarely included gross figures for trade. In Osaka, which became an open port in 1868, a decade later than the other ports, the concept of “disposals” was more persistently pursued than in other ports. 65 Estimates of disposals were turned into figures intended to represent total disposals for the year. To a lesser extent, data for exports also were adjusted, using criteria of foreigners’ specification, to yield annual totals. However, even in the case of Osaka, a careful reading of the annual consular reports makes clear the different basis of these calculations, and the situation was well understood by most of those involved. As discussion in the main body of annual reports shows, the principal value of disposal figures was to convey estimates of actual sales of individual commodities within a twelve month period, and the reader was referred to the appendices for tables from the Custom House. Where the current year was compared with preceding years, or where retrospective tables were given, the totals were invariably the figures provided by the Custom Houses. Put baldly, the Chamber of Commerce data are by no means statistics in the proper sense of the word. Even in Hyōgo/Osaka in the 1870s, where the emphasis on disposals was particularly strong, composite aggregates of imports were intended to be simply estimates of disposals, not substitutes for the official figures themselves. From 1873 onwards, the reader of British reports was referred for guidance on the overall trade of Hyōgo/Osaka to the statistics appended to the reports drawn from printed returns from the Custom House.66 A not unim-

198

EARLY JAPANESE TRADE

portant reason for calculations by Chambers of Commerce was that the Custom House did not distinguish in its published totals between the various countries until the 1880s. Before that decade, the distinction was between vessels in terms of the flags of different nations, but this was not helpful, as vessels very often carried the goods of other nations.67 EARLY MEIJI STATISTICS

What happened in the immediate wake of the Meiji Restoration in terms of central direction of trade accounting is obscure. Modern accounts in Japanese simply refer to the role of “the competent authorities” (kantoku kanchō ┘╩ᐁᗇ).68 By 1872, the Ministry of Finance had established full blown supervision, but the stages by which that supervision developed are not clear. There is no suggestion that central control had already emerged in 1868. The first steps in centralized management occurred as early as 1869 and 1870, when the Foreign Ministry seems to have compiled overall figures for the ports. The first return by the Ministry of Finance was for 1871.69 Comments by consuls, particularly on the quality of the completed permit forms, began to become more positive. As early as 1870, Consul Lowder observed in his report on Kanagawa trade in 1869 that “a marked improvement is observable . . . [T]hese returns may, therefore, be looked upon as more correct than those which have hitherto been supplied to this consulate.”70 Consular comment on the year 1871 at Kanagawa recognized the “improved administration” afforded by “the governor and superintendent of Customs.”71 The report on the year 1872 conceded a close correspondence between Custom House figures and Chamber of Commerce calculations in the case of exports of “exceptional prominency,” and no “great discrepancy” in the Customs figures for imports.72 New forms issued from the customs were coming into evidence in 1872, and officers in the ports were now under a compulsion to accept central guidance.73 The first bilingual reports from January 1873 were from the “Imperial Customs, Yokohama.”74 In the preface to monthly reports for the following year, the head of the office is identified as Matsukata Masayoshi ᯇ᪉ṇ⩏, Commissioner of Internal Revenue, at the Ministry of Finance.75 Reference is also made in 1874 to a Statistics Division, which implies that statistical work had now become a specialized office within the Ministry of Finance. In 1874 the consul in Osaka, dissatisfied with figures on the spot, sought and got figures from the Imperial Customs in Yokohama.76 This

STATISTICS OF TOKUGAWA COASTAL TRADE

199

was an episode which showed that the Custom House in Osaka was answerable to a central authority, even if in this case no better figures were made available. The consular report published in 1873 on the Nagasaki trade of 1872 noted the use of many new forms by officials, as well as a better conduct of business in a technical sense, notwithstanding the delay in the issue of permits: “It is a matter of general congratulation that greater care is now exercised in collecting the duties than used formerly to be the case.”77 The Ministry of Finance, in contrast to the bugyō whose functions were manifold, was devoted exclusively to the raising of revenue and, of course, the control of expenditure to meet the pressing needs of a cash-strapped state. Unlike the Foreign Ministry, whose initial role had been stop gap, the Finance Ministry had officers of its own in the ports. Officials trained in the old school of bugyōsho work with their local loyalties were replaced by new personnel and were closely supervised by a central administration whether they were new or old officers. Under the old system of administration, instructions had been minimal. The only instance of instructions being issued is one within the framework of the 1866 negotiations between the bakufu and foreign diplomats, which set out a modest schedule of 53 items for exports and 83 for imports, plus short lists of goods which were duty free or prohibited.78 However, by 1872 the schedule of exports and imports had expanded to embrace 288 export and 1019 import items.79 The schedule not only listed items but also set out details, the better to illustrate how to treat the items for revenue collecting purposes. Such a schedule, applicable to all officials, was essential for effective revenue collection and for statistics as well. The scope of the trade figures was also widened in 1873 to include a broad range of official purchases. While purchases of firearms and ammunition were included in the statistics at least from 1868, the figures were said to be below the actual amounts. The purchase of ships remained excluded, though the scale of these purchases both before and since 1868 was well known, and commented on by consular officers among others.80 But from 1873, the year in which the Ministry of Finance took effective charge of the compilation of data, official purchases were estimated in full: the figures were as low as 797,395 dollars in 1873, and as high as 3,475,277 dollars in 1875.81 From 1879, official purchases were no longer identified separately within the tables. With all these changes, in trade as in other areas, the statistical age had well and truly begun in Japan. The pattern of monthly returns, well established in pre-Meiji times, continued uninterrupted through bakumatsu and the early years

200

EARLY JAPANESE TRADE

of Meiji.82 It does not appear that there was a fast and hard rule as to whether consuls received monthly returns or only annual returns. In the problematic case of Kanagawa in 1866, the consul seems to have counted on an annual return and either did not receive, or disregarded, monthly returns. From 1873 onwards, the monthly figures provided the basis for the Ministry of Finance’s regular annual, and even half yearly, returns.83 From January 1873 onwards, copies of the monthly returns are fairly complete, and the few surviving documents are part of a now lost or dispersed but once large corpus.84 They served as the source for the monthly breakdown from 1873, published in later official documents.85 In the bureaucratic process of making statistical knowledge known within the administration, printed documents appeared at least as early as 1869. In the bakumatsu years, the British consuls were already receiving details of trade in the individual open ports, but the first BPP reference to tables for all Japan in 1873 indicates an immediate awareness of the new Japanese reports. From January 1873, the preface to the monthly returns declared that “beginning with the first month of the present year, a notice will be prepared here of exports, imports, values, quantities, and shipping, conveniently arranged for general information and published under the title of ‘Tables of imports and exports at all the ports.’”86 The figures were all the more accessible by their publication in movable type, which replaced the cramped and rather difficult presentation in the woodblock printing of earlier years. The publication of tables in movable type was also accompanied by a general tightening up in presentation. Figures for 1873, in the first of the new annual returns took up 140 pages, and were diffusely structured with each port given separately. Later tables, showing the ports within a general tabular presentation, typically took up a mere seventeen to twenty three folios, and the new sophistication was reflected in the fact that folios soon ceased to be described as pages. From 1873 onwards, the figures are final versions and create no internal confusion. For earlier years the data appear to be drafts, and are far from problem free. The new presentation, together with the use of the English language, would seem to be consequences of recruiting English-speaking foreigners in the early years of the Ministry of Finance, and the Statistics Division of that ministry is probably one of the fruits of the change. In typical fashion, however, Parkes’ attitude was begrudging. Though we may assume with confidence that he knew of the new monthly reports, his immediate response was a call for a quarterly return.87

STATISTICS OF TOKUGAWA COASTAL TRADE

201

The inclusion of figures from 1874 retrospectively in several compilations in the 1880s by the Ministry of Finance has led to a more formal status being accorded to them than to earlier figures. However, the figures for 1873 have clear cut internal evidence of their authority, and indeed are particularly informative on what was taking place within the Ministry of Finance. Reports in the 1870s were, however, not public documents; they were strictly speaking internal documents, made available to certain parties as a favor. This fact is reflected in the wording of the BPP, which refers to being “furnished” with the reports. The Ministry of Finance’s first formally published report concerned the trade of 1882, and was the prelude to a series of annual Ministry of Finance returns, appearing under the title of Dai nihon gaikoku bōeki nenpyō ኱᪥ᮏእ ᅜ㈠᫆ᖺ⾲published by the Customs Bureau (Kanzeikyoku 㛵⛯ᒁ; later, Shuzeikyoku Daiichibu ୺⛯ᒁ➨୍㒊, and variants).88 From the first issue, the reports also included retrospective summary tables of export and imports in gold yen dating back to 1868. From the 1884 report onwards, a table was included for imports and exports of bullion and specie from 1872. The reports were immediately recognized by the British legation. The first report was referred to in the legation’s own report, filed on 15 July 1883, which stated that much of its analysis was based on the figures in it.89 For the early post-1882 years, surviving copies of the reports are few and are to be found only in the Finance Ministry Archive (Ōkurashō bunko኱ⶶ┬ᩥᗜ), the Prime Minister’s Office Bureau of Statistics Library (Sōrifu tōkeikyoku toshokan ⥲⌮ ᗓ⤫ィᒁᅗ᭩㤋), and the National Diet Library. This is a remarkable illustration of the poor survival rate of printed and manuscript documents, even in the conditions of early Meiji Japan. That the reports no longer appeared with the alarming speed characteristic of reports in the 1860s reflected a greater degree of serious final processing of figures.90 The first woodblock printed figures of exports and imports for the years 1869–1872 are now available in modern publication.91 These figures were compiled under the aegis of the Foreign Ministry (1869, 1870) and Ministry of Finance (1871, 1872). They are, however, but incomplete drafts. For 1869, the surviving data are for quantities only, not monetary values. The figures for 1870 are not aggregated for the first half of the year, and only incompletely for the second half of the year. Figures for 1870 are valued in ryō, though the presence of a few figures in unconverted dollars supports the hypothesis that this document was not a finished one. For the year 1871, the first year of Ministry of Finance management, totals were aggregated into gold yen, as

202

EARLY JAPANESE TRADE

of course they were for 1872.92 It is very clear from the sparse surviving evidence that there was an ongoing loss of data through late bakumatsu and early Meiji years. Statistical reports were printed from woodblocks only for a short time. Monthly, half yearly and annual figures were set in movable type from January 1873, in English as well as Japanese, and in modern – Western – tabular fashion. Close to this date, perhaps beginning earlier but certainly terminating in June 1874, an effort was made to gather retrospective runs of figures for preceding years. There is a woodblock printing of a retrospective run of export quantities from 1866 to 1873 by the Customs Office, dated June 1874. There also exists a manuscript document with a run of values as well as quantities in a remarkably neat and uniform hand, stopping short in 1872. The absence of corresponding runs of tables for imports, reflects the confusion or loss of data characteristic of the pathetically small and defective corpus of surviving returns before 1873. The fact that the retrospective runs of figures for exports could be compiled up to 1872 and 1873 respectively shows that some data were to hand, and the tables are especially intriguing in hinting at some access to records preceding 1868, in other words for 1867 and 1866. (The latter is of course the year for which records for Kanagawa were lost).93 The Ministry of Finance observed in January 1873: “It cannot be said that the details of those statistics [of recent years] were carefully preserved.” The dearth of extant sources serves as a modern footnote to what the Ministry of Finance said in 1873. Yamaguchi and Ōuchi writing in 1968 drew on figures in the Ministry’s archives for 1869 and 1870, and in Tokyo University Economics Department for 1871 and 1872. Duplicates of figures for 1870–2 exist also in the Japanese National Archives (Kokuritsu kōbunshokan ᅜ❧බᩥ᭩㤋), though its published calendar does not identify the sources from which they were acquired. The best guide to the early statistics is Yokohama-shi shi: Shiryō hen and a series of photocopies in the Yokohama Kaikō Shiryōkan ᶓ ὾㛤 ྐᩱ㤋.94 There are trade returns in the National Diet Library ᅜ❧ᅜ఍ᅗ᭩㤋 (though lacking one for the year 1869) in microfilms, made in the 1960s from the Japanese National Archives and other sources.95 The Ministry of Finance’s own archives were destroyed in the 1923 earthquake. Microfilms which the Ministry has deposited in modern times in the National Diet Library, and which it claims to comprise all the early material in its possession, originated in post 1923 deposits in the Ministry, the Matsukata collection, and two collections deposited by the Matsuo ᯇᑿand Shōda ຾⏣families respectively.96

STATISTICS OF TOKUGAWA COASTAL TRADE

203

The Matsuo and Matsukata collections, thin for the 1870s, do not contain trade figures. The absence of statistics in the Matsukata deposit in the Ministry of Finance or in the published Matsukata papers is particularly to be noted, as Matsukata was head of the Customs Bureau in the 1870s.97 The sole case of a contemporary trade return (quantities in the second half of 1870) lies in the comparatively recent deposit in the National Diet Library of the papers of Inoue Kaoru ஭ୖ㤾, Vice Minister of Finance 1871–3, and the holder of many later offices of state.98 Two modern writings, one in 1962 and the other in 1968, refer to the existence of a return of trade figures for 1869 in the Ministry of Finance archives.99 As the Ministry of Finance archives are not now open to the public, the survival or current status of this very rare return is not known. The first year of Meiji, 1868, is even more sparsely documented: Yamaguchi and Ōuchi even speculate as to whether a return for that year had been compiled at all. In fact, one was attempted, and a return for the year survives in Kanagawa-fu nisshi ⚄ዉᕝᗓ᪥ㄅ (Kanagawa administration daily register), though it is inevitably only in the form of an entry into the daily register.100 An area of uncertainty is whether bullion and specie were included in trade totals. A modern (1935) study, Nihon bōeki seiran ᪥ᮏ㈠᫆ ⢭ぴ, suggests that exports for 1868–1871 in the yen versions include movements of gold and silver (which would impede comparison with other years and with the BPP figures in general).101 The facts are far from clear, however. The totals of Japanese trade in the BPP are drawn from the Custom House data, and positively exclude currency movements. Certainly, Japanese and foreigners alike were aware of movements of precious metals and specie, and consular comment was accompanied by some figures. The BPP report of 1870 referred to heavy shipments of coin and bullion from Japan in 1869,102 and for the following year the report suggested that outflow of coin and bullion in 1870 was of the order of ten million dollars.103 Some statistics of bullion and specie existed and, as early as 1863, figures of movement in and out of Kanagawa were recorded.104 For 1871, Hyōgo/Osaka exports of treasure were estimated at 2,258,654 dollars by the Custom House, and at 5,019,011 by the Chamber of Commerce.105 Kanagawa consulate also reported figures for treasure for the same year.106 For 1873, details for all the open ports were reported.107 While general totals including gold and silver movements were given in BPP, consular officials did not supply figures regularly or in standardized form. This suggests that information was not always to hand.

204

EARLY JAPANESE TRADE

It seems equally likely that knowledge on the Japanese side was imperfect. Revealingly, in 1882 the Ministry of Finance began a series of annual volumes of statistics on foreign trade (Dai Nihon gaikoku bōeki nenpyō), featuring a retrospective run of trade figures back to 1868, but the first data for bullion and specie to be published appeared not in the volume for 1882, but two years later, and went back only to 1872. A note to the table in the 1884 Dai Nihon gaikoku bōeki nenpyō stated tersely, “The Statistic (sic) before year 1871, being uncertain, is excluded.”108 The dilemma facing contemporary editors may be seen from the tables of exports and imports for 1869–1872 published by Yamaguchi and Ōuchi in 1968. There, bullion and specie are recorded for one year (1872) only, but duty free goods (in which category bullion and specie were included ) were for some reason declared “unclear” (fushō ୙ヲ) in the concluding summary table, and hence do not feature in the grand total for imports and exports. The 1872 count without the inclusion of bullion and specie in commodity tables at 16,847,033 yen, was higher than the 16,056,388 yen in the retrospective table compiled in 1874. For imports, two gold yen counts and the dollar count are much closer together, seeming perhaps to cast doubt on the presence of a distorting coin and bullion total. From 1873, study of these issues becomes easier, but care is still required with crude totals. Cursory analysis of Finance Ministry returns in Meiji 1873–6, for instance, appears to suggest that bullion and coin were included in the grand totals for imports and reexports in three of four years. However, the complexity of figures for the early years is underlined by the wide divergences in 1868–1871 between totals in dollars and in gold yen (the latter by definition retrospective calculations for these years), and by a puzzling inconsistency between the profile of totals of exports and imports in gold and silver prices. Data for the year 1872 illustrate the problem. There are four separate counts for this year: one in dollars in BPP, one in gold yen from a contemporary Ministry of Finance table, one of exports from tables for seven years from 1866 compiled in 1874 by the Ministry of Finance, and one in the first (1882) volume of the Dai Nihon gaikoku bōeki nenpyō. For exports, all three gold yen counts fall far short of the BPP dollar count, which is known not to include bullion and specie movements. Problems of definition affecting the treatment of specie and bullion apart, misreadings in transcription and inconsistencies in aggregations may have compounded the errors in calculation of early statistics. Differences between figures might at first sight seem to support the

STATISTICS OF TOKUGAWA COASTAL TRADE

205

observation by Sugiyama that alternative consular and official estimates continued to exist in early Meiji times, but the differences might also have arisen from prosaic errors. As returns from the Custom Houses to the central offices of the Ministry of Finance up to the end of 1884 were stated in dollars, a possibility of error resides in the Ministry of Finance conversions of the dollar totals into gold yen in early Meiji, when documents were confused and conversion rates between dollar and gold uncertain. This is particularly the case for the table of exports for 1866–1872 in gold yen prepared by the Ministry of Finance in 1874; construction of the table involved retrospective conversions from dollar totals for 1866 and 1867 and from ryō totals in 1868–1871. THE TRANSITION TO THE YEN

The currency system at national level in Tokugawa times had been in theory bimetallic (and even tri metallic, as there were three metals, gold, silver and copper). Silver had been the currency of Osaka and Nagasaki, and gold was the basic standard of Edo itself. The introduction in 1869 of a new gold ryō marked a stage forward in creating a formal national unit of accounting. The new ryō had no direct effect on other metals in circulation: it simply made existing Edo accounting practice into a national one. Two years later, in mid 1871, new regulations replaced the ryō by a gold yen as domestic unit of account.109 While in line with best new practice, it created a gold standard in internal payments. From the outset, the new system was a hybrid with a gold yen as the unit for internal payments, and a silver unit (the new silver yen in place of the Mexico dollar) intended to be confined to transactions within the open ports. The silver unit, the Mexican dollar, which was by informal agreement from 1859 the accounting unit of foreign trade, became in 1871 the de jure unit of foreign trade with the introduction of a silver yen (of the same weight as the Mexican dollar). Given the monetary history of Japan and ready access to silver in east Asia, there were good practical reasons for silver becoming the de facto unit after 1859, and under the currency edict of 1871 it became in that year the de jure unit. The Mexican dollar was long the currency of account in international settlements in East Asia; its role in underpinning international business was enhanced by Hong Kong’s issue of a coin identical in weight. Much of the early trade with the West after 1859 was routed through China, a circumstance which helps to explain why a majority of the international residents in Nagasaki already in the

206

EARLY JAPANESE TRADE

1860s and in Yokohama in the mid 1870s (the first years for which we have counts for the port), were Chinese.110 Japan faced the dilemma of being situated firmly in the silver currency region of East Asia and at the same time of having two Western countries as major trade partners. One of these, Britain, had long been on a gold standard, and the other, the United States, had been on it from 1873. With regard to exports, Japan operated the foreign trade of its open ports before and after 1871 either in Mexican (silver) dollars or in claims convertible into dollars, so that no problems of conversion arose when exports were declared at the Custom House in silver. Likewise in the case of imports, the conversion at the Custom House of prices on invoices from silver currency countries involved a purely routine conversion. Only invoices of imports from gold currency areas such as Britain presented a problem, but they accounted during the 1870s for two thirds of the import trade of Japan. Conversion between gold and silver currencies was not an intrinsically difficult problem. It became problematic only because the conversion of invoices at a fixed rate, abandoned in 1865, was in effect resurrected by the creation of a fixed rate of exchange between silver and gold units in the 1871 currency edict. Some words of explanation are necessary at this point on two distinct though overlapping issues. The gold ryō of 1869 had been a unit of account not physically embodied in a coin, and hence no rate of conversion between ryō and dollar had been settled. However, the new gold yen of 1871 created a problem in that it was accompanied by a formal silver standard confined to the open ports. A gold coin weighing 25.72 grains was the standard for the new gold yen; it was equal in value to a silver yen of 416 grains, that is the weight of the Mexican dollar. The mint ratio or exchange rate between the two metals at 1 to 16.2 reflected the international rate of exchange of 1871. However, a fixed ratio of exchange could only have worked in a closed system (a situation with which the 1871 act in creating a separate currency for the open ports had tried rather simplistically to grapple). There were two consequences to the 1871 edict. The first, a serious material one, was that gold was increasingly either exported or hoarded, since it could be purchased at the overvalued parity of silver in 1871, and as gold disappeared silver became a de facto standard beyond the ports. The second consequence, discussed below, was primarily a bookkeeping one, as customs officers observed the fixed exchange rate between gold and silver settled in 1871 in converting invoices for the purpose of levying duties from gold currency countries. In doing so the totals

STATISTICS OF TOKUGAWA COASTAL TRADE

207

were artificially depressed by a persistent fall in the value of silver on world markets undermining from the outset the rationale of a fixed rate, and the sum total of the invoices by definition understated the real value of the trade. The change of government in December 1867 had brought on apprehensions of currency destabilization. Uncertainty as to whether the new government would last combined with delay by the government in introducing regulations to meet its currency needs, and gave rise to serious confusion. Good coin, whether gold or silver, was in the short term either hoarded or exported and, adding further to confusion, some of the exports were on government account. As a result, inferior and even unfamiliar coins, new and depreciated in metal, came into circulation, along with paper.111 Good quality silver and gold grew scarce as both were hoarded and the latter was sent abroad, and rates of exchange between coins fluctuated wildly. In the volatile market of 1871, for instance, the price of the nibu ஧ศ㔠(an alloy of gold and silver) varied by up to 30 per cent.112 New coins provided for by the regulations of 1871 were slow to appear, and as late as 1872, the gold and silver yen had not reached Nagasaki. Increasingly, internal settlements were made in paper; the high discount on paper in the short term was less a reflection of doubts about paper than of uncertainty about the worth of the coins, which were their ultimate guarantor of value. Paper in fact began to gain ready acceptance quite quickly, and eventually circulated at par or even at a premium. There were two sharply contrasting currency markets, an internal one characterized by uncertainly and fluctuations, and a foreign exchange market in which silver traded freely against gold in the ports. Without any undue problems, inflows and outflows of coin reflected either short term speculative purposes, or the transfer in metal one way or other of the net balance in total transactions on current and capital account. The peculiar coexistence of two media of account, a silver unit for the ports and a gold unit for domestic transactions, has been responsible for the belief that it posed a special or complex statistical problem in calculating aggregates of Japanese trade. There is, however, no reason why the existence of two currency zones of itself should present difficulties, and the undervaluation of the total value of trade would not automatically follow from the fact of operating in two currencies. The initial aggregating of invoices in different currencies in the open ports would be a matter of elementary currency conversions, translating invoices that were not expressed in silver into silver terms at market rates. Conversion between two metals was not of itself the root problem.113

208

EARLY JAPANESE TRADE

The processing of statistical returns from dollar totals into the new yen should not have occasioned a problem, in theory, at any rate from 1872. For the years up to 1871, in retrospective calculations the fact that dollar and gold yen totals of foreign trade differed substantially reveals that there had been some other problem, not arising from the simple existence of two currencies. In 1868–1871, totals in dollars exceeded gold yen (as set out in the Ministry of Finance table for 1882) in three of four years for exports, and in one of four years for imports. These figures originated in dollars converted at the time for the first three years into ryō totals (and retrospectively into gold yen), and revised once more in or before 1882. More puzzling at first sight than variations from year to year between totals in dollars and in gold yen is the lack of consistency in the movement of converted totals as between exports and imports. The explanation is that there was strong Japanese demand for dollar coins because they could be used in re coining a deficient supply of coin. One result was that foreign buyers of Japanese goods, benefiting from the demand for silver in Japan, could dispose of their dollars at a premium. Japanese buyers of foreign goods might be able to pay in coin or paper money, but poor coin and paper were heavily discounted when they were accepted. It seems most likely that the Ministry of Finance made separate calculations using different terms for exports and imports, taking account of the quite different payment conditions for exports and imports respectively. After 1871, conversion from silver dollar totals to yen within Japan should have been at parity, at least in theory. In practice, this did not apply to the 1872 figures. The 1872 figures for imports exhibit near concordance among different estimates (a dollar total and two somewhat lower gold yen figures). On the other hand, for 1872 exports, the dollar estimate exceeded yen estimates by the largest margin recorded for either exports or imports in any year in 1868–1872. The discrepancies may be accounted for by the peculiar circumstances that caused bullion of all sorts to be in short supply in the early Meiji years. Given the demand for silver bullion to create new issues of both silver ichibu and alloys of silver and gold, foreigners imported dollars expressly to take advantage of a Japanese premium on silver dollars. The premium was supported by the siphoning off of the dollar into the hands of currency speculators or into purchases at the mint that opened in 1871. Regarding the differences between yen and dollar aggregates in the statistics, Sugiyama observes, “After 1873, there is a fair degree of consistency between the two series of figures. I have been unable to find

STATISTICS OF TOKUGAWA COASTAL TRADE

209

a reasonable explanation of the continuing discrepancy in the period 1869–1872.”114 However, he seems to have seen the problem as one of differing systems of recording figures rather than one of conversion rates. The difference between dollar and yen totals for the years 1868– 1870, and equally for 1871 (when the official gold-silver yen parity was defined only in mid year) and 1872 reflect several circumstances. For one thing, a widening medley of currency offered in payment for imports made general calculations for converting annual totals of trade problematic. Another factor was that the market placed a premium on dollars as a means of payment for exports, despite the existence of an official fixed rate for gold and silver yen. This benefited those who could pay dollars for their purchases. Demand for the dollar had been weak as late as 1866, but thereafter it usually commanded a premium.115 Maintenance of a one-for-one parity was effectively impossible, and the evidence suggests that Japanese officials in Tokyo recognized this. In the more orderly monetary situation of 1875, and in contrast to much higher premiums in prior years, a slight premium was officially declared for the silver yen: 100 silver yen exchanged for 101 gold yen.116 Movement toward parity of yen and dollar was already evident in 1873–1875, though the momentum was not maintained in 1874, and wavered somewhat for several subsequent years.117 The Introduction to the Yokohama-shi shi: shiryō hen of 1962, brief but by far the best analysis of the issue, observed that the figures were substantially the same from 1877.118 Yet some divergence still lingered on. The significance of 1879 as a watershed in trends, however, is easily explicable. In 1878, Japanese authorities established a silver unit of currency equal to the dollar for general use (not just in open ports), and in 1879, they declared the parity of the dollar and silver yen. Earlier conversion of the dollar and silver yen into gold yen fluctuated in tandem, and official attitudes to conversion are totally obscure. By the end of the 1870s, the Mexican dollar was still imported, but it had for all practical purposes disappeared from day-to-day transactions. As Custom House returns continued to be given in silver, the British consuls continued to tote up aggregates of trade from the returns from the individual Custom Houses; consular officials did not immediately accept the new central returns in gold yen created by the Ministry of Finance counts. Copies of the original returns from the Custom Houses, which would have been in dollars, are not known to have survived. Only with the consular report in early 1886 on the trade of 1885 did the British representatives switch from figures in dollars. The

210

EARLY JAPANESE TRADE

consular reports included in the BPP, thereafter, gave totals in sterling, converting on the basis of the rate for dollars or silver yen as quoted for the purchase of drafts drawn by merchants or bankers on London. One silver dollar was equated to one yen from 1885. In previous accounts, the extent and duration of monetary confusion has been greatly exaggerated – particularly the duration. Misunderstanding has been reinforced by a view of paper money as one of its causes. In fact, the trend towards stability in conversion, already apparent in 1872, was itself a consequence of a growing preference for paper that facilitated stabilization of payments within Japan, and simplified problems of settling on conversion rates. Contrary to statements that paper currency continued to depreciate, paper changed hands at parity with silver or even at a premium once the confusion of the early Meiji years passed.119 The inflation following the Satsuma ⸃ ᦶ rebellion of 1877 was in no sense a proliferation of earlier monetary problems, but rather a straightforward case of a general inflation caused by soaring expenditure and the sudden expansion in the issue of paper at that time. After 1881, it would be gradually reversed by the famous Matsukata deflation. The large discount on paper in the wake of the 1877 rebellion is not germane to the exchange problem. It is a textbook case of an inflated issue of paper in a budgetary crisis, and its effects are thus more easily comprehensible than the political and monetary complexities of the first years after the Restoration. In understanding the currency problems of these years, two separate issues have to be distinguished. The first is the declining value of silver in the wake of the fall in the price of silver bullion relative to the price of scarcer gold metal; this was a consequence of the increase worldwide in mining output. The Japanese government’s 1871 attempt to establish a fixed rate was thus progressively more unrealistic. The second issue concerned conversion of gold prices into silver prices in import invoices. A fixed conversion rate was settled by administrative fiat, as had been the case up to 1865. In regard to the first of these two issues, Japan was on the way to becoming de facto a silver currency area, through the operation of Gresham’s law, which holds that bad money drives out good. A new silver coin issued from 1871 to 1877, intended to rank with the Mexican dollar and to enlarge the supply of silver currency for a growing foreign trade, was confined to the ports.120 In 1878, a dollar coin identical to the Mexican dollar was made valid in all public and private payments throughout Japan, and in 1879 it was formally decreed the equivalent

STATISTICS OF TOKUGAWA COASTAL TRADE

211

of the Mexican dollar. Circulating beyond the confines of the ports, it represented a late stage in currency reform, as Japan, although de jure bimetallic, came de facto to use silver as the medium of exchange. The scarcity of gold had threatened to undermine a monetary system in which gold was the basic standard for internal trade. Silver inflow was sizable in 1882–1884, and movements of gold either way had become small and few. In 1883, interest on kinsatsu 㔠ᮐ, a form of government bond, originally issued in both silver and gold was made payable solely in silver. This was an implicit admission by the Meiji government that the country had moved to a silver standard.121 Gold was not abandoned in a formal sense, but the 1878 authorization of national circulation of a dollar coin meant that, for all practical purposes, a silver standard was in place, and would prevail until the Japanese government officially adopted the gold standard in 1896. Over the course of the 1870s and 1880s, silver purchased progressively less gold. The relationship was commonly expressed in the amount of sterling acquired by the dollar in drafts drawn on London. It could also be expressed, although more rarely, in terms of the relationship between the gold yen and the silver yen, which was beginning to replace the dollar as the currency standard of Japanese ports. In gold yen-silver yen quotations, the purchasing power of the gold yen became progressively greater after 1880. Gold yen were at a premium throughout 1882, for example, as quotations averaged about 92 1/2 gold yen for 100 dollars.122 Exchange quotations for gold yen in terms of silver yen rather than the dollar, departing from their formal parity, were a new bookkeeping convention to express actual transactions in the ports where the silver dollar exchanged against foreign gold currencies, and signified that the silver yen (despite or because of its parity with the dollar) was beginning to acquire an identity in its own right. The second complex issue in this story of the two metals was the fixed rate used for the conversion of gold prices in import invoices into silver prices. After 1871, there were two exchange rates for conversion of gold yen into silver dollars, the open market rate determined solely by currency transactions in the ports, and a fixed rate for gold standard currencies solely for Custom House transactions. For sterling the rate was settled at 4.88 dollars to the pound sterling (or 4s.1d. for the dollar). Conversions at this rate became unrealistic in the long term. The dollar, worth 4s.1d. in sterling as late as 1875, was only 3s.8d in 1884, and by 1887 it was down to 3s.2d.123 These rates are evidence of

212

EARLY JAPANESE TRADE

the strengthening of the British pound; in the early 1860s, a Mexican silver dollar had been equivalent to 5s. or 5s.6d. So long as the government-fixed rate of gold-silver yen exchange was in place, it was beneficial to traders. It was of no consequence as far as the settlement of daily private trade accounts was concerned, but when settling payment of ad valorem duties traders declared transactions at rates that greatly overvalued silver compared with actual market conversion amounts. Japanese government adherence to the fiction of an unchanging gold-silver rate thus entailed a loss of revenue to the treasury. In bookkeeping, when invoices figures were aggregated, statistics of imports were understated. The anomaly was finally ended in the last quarter of 1888 when imports and exports were valued in identical fashion on a basis intended to reflect the market prices of silver. A year later, the results were set out clearly in the British consular report for 1888: For the first nine months of the year 1888, every pound sterling of the original cost of all English and American articles of import was entered in the customs returns as 4 dol. 88 c., whereas, throughout the whole of the year 1889, each pound was entered at values varying each quarter, but the average of which for the whole year was 6 dol. 46 c.124

In the aggregates for 1888, the figures for the first three quarters were adjusted to conform to the new basis introduced for the fourth quarter. The sharp rise in the total value of imports over 1887 is nominal, not real; it reflects the adjusted exchange calculations of the Custom House. From 1888, totals were by definition realistic. When the fixed rate for conversion of gold into dollars was abandoned later in 1888, the Ministry of Finance did not recalculate past figures (apart from figures for the first three quarters of 1888). Some scholars, notably Tatemoto and Sugiyama have taken it on themselves to revise the figures for the years prior to 1888 in order to take account of past systematic understatement in the figures for imports. Superficial differences between published runs of figures could be taken to support the argument that separate compilations of figures were created in the ports or, alternatively, that great confusion existed in processing data. Runs of data in dollars and gold yen respectively might appear to support the existence of distinct or separate figures. Thus export figures are presented in the Yokohama-shi shi in dollars for both Yokohama trade and national trade,125 but more commonly, as in Baba and Tatemoto (1968),126 Sugiyama (1988), and other recent works,

STATISTICS OF TOKUGAWA COASTAL TRADE

213

they are in gold yen. The Yokohama-shi shi figures down to 1884 were drawn from BPP, and thereafter from the official Nihon gaikoku bōeki nenpyō.127 Scholarly recalculation to take account of undervaluation of imports was first executed by Tatemoto, and later by Sugiyama.128 This is not a straightforward exercise. It is relatively easy from the mid 1870s as gold and silver units of account were close to parity. As Sugiyama noted, it is fraught with serious problems between 1869 and 1873, given differences in the figures. For this reason, Sugiyama attempted major recalculations only for the years 1874–1887. He did also, however, make an effort also to revise the figures for imports and exports for the years 1860–1868.129 He posited that 1869 was a “normal” year, and for 1860–1868 he converted dollars into yen on the basis of the margin of excess of yen for dollars, shown by comparison of the two measures for 1869 (namely, the BPP and Ministry of Finance figures). If this is a correct interpretation of what he says he did, his methods are untenable. They could only end in an arbitrary result. It is contentious, to say the least, to regard 1869 as in any sense “normal.” The relationship between various figures in 1869 (or in any other year) is a product of a number of factors about which we know too little.130 THE MERITS OF JAPANESE STATISTICS

Over the 1870s, British consular acceptance of the Japanese figures became ever more positive. Even if the Nagasaki consul complained of “a difficulty in getting reliable statistics” in his report in early 1874, he conceded that “the better regulations now existing at the Custom House render this not so insuperable an obstacle as formerly.”131 While the consular report on Kanagawa for 1878 found that the Custom House data did not agree in all respects with Chamber of Commerce estimates, it declared them “tolerably correct.”132 Some problems were now of a different order, and arose from deficiencies in Chamber of Commerce data. Thus, the Chamber of Commerce in Kanagawa enumerated silk in bales in its 1880 accounts, while customs figures reckoned trade by weight. The Chamber of Commerce rate of conversion of bales into piculs underestimated the weight of bales.133 Osaka, an open port only from 1 January 1868, nine and a half years later than the other ports, seems to have been out of step with this story of improvement, at least in its first years of foreign trade. It is hard, however, to draw a line between objective weaknesses in the

214

EARLY JAPANESE TRADE

Custom House and the idiosyncrasies or limitations of Abel A. J. Gower, the consul from 1869 to 1873. For 1868, Acting Consul J. F. Lowder had simply reported the Custom House figures without comment.134 For 1869, however, Gower provided figures based on “private information” then, inconsistently, relied mainly on “the very imperfect Custom House return” for 1870.135 From reluctant acceptance of the Custom House returns for 1870, he progressed to outright hostility to Custom House data for 1871 and 1872, and he assembled figures of his own from the applications for permits in those years. His doubts were fuelled by figures for disposals compiled by the Chamber of Commerce. The divergences between the Custom House figures and the Chamber of Commerce calculations led him, in February 1872, to go so far as to declare the “worthlessness of the Japanese returns.” He dismissed the applications for permit forms as improperly filled in, to the extent that “the labor spent in endeavoring to compile correct Tables is little better than lost.”136 His view was no less strong in the following spring, when his report on the trade of 1872 contained the following phrase critical of Japanese officials: “The Custom House authorities [have not] yet adopted any proper system of preparing the returns themselves”.137 Compiling his report for the year 1873, he applied to the Ministry of Finance customs office in Yokohama for returns, apparently because of dissatisfaction with those received locally, and lamented that the employment of foreigners (presumably in Yokohama) had failed to raise the statistical standard.138 Gower’s failure to appreciate that estimates of disposals and Custom House returns had fundamental differences deepened his dismissive attitudes. The fact that, unlike other consuls, he did not use the actual term “disposals” in his writing suggests that he was oblivious to deeper statistical distinctions. His shallow understanding, and his helplessness implicit in his reliance on the Custom House data “principally” for 1873 – despite repeated strictures on them – contrast with the measured and competent assessment of the figures for 1874 and 1875 by his successor, Acting Consul (later Consul) A. A. Annesley.139 Whatever the early deficiencies, things were steadily put to rights by the Osaka Custom House. Proper attention to the Application for Permit forms was, of course, a vital base for statistical progress. A tightening in administration seems to be acknowledged in the report for 1872, even by the choleric Gower. The consul, while complaining of the absence of a system for preparing returns, noted “a marked improvement” in the quality of the filling in of the application forms

STATISTICS OF TOKUGAWA COASTAL TRADE

215

themselves.140 Gower did not attach weight to the fact that the relationship of Custom House and Chamber of Commerce figures was not consistent from one year to the next. For the year 1871, he merely noted that aggregates based on Application for Permit forms fell short of the Chamber of Commerce figures, while in the following year the aggregation of Custom House figures for imports exceeded the Chamber figures.141 He did not dwell on the contrast, but contented himself with dismissively attributing the difference in the 1872 figures to the omission of small parcels from Chamber of Commerce figures. This was facile and even evasive. The report on the trade of the year 1874 by the new consul provided both Custom House and Chamber figures, and demonstrated an appreciation of the fundamental differences in the two statistical concepts. A positive (if understated) note appeared in the Osaka consul’s report on the following year (1875), which remarked on the “tolerably relative accuracy” of Custom House figures.142 His report for 1876 conceded the “fair correctness” of the Custom House figures for imports, where the statistical divergence with Chamber of Commerce numbers was largest.143 Differences between Custom House returns and Chamber of Commerce estimates of “disposals” (which have led to modern confusion that there were two statistical counts) were occasionally noted in Kanagawa, e.g., in the report for 1872,144 but consular reports from Kanagawa, apart from the special case of 1863, were based virtually without qualification on the Custom House figures. “Disposal” was taken as a statistical concept primarily of use for illustrating the progress of individual trades. Consuls who followed Gower in Osaka were very much alive to the issue, but the concept of disposals continued to hold a prominence in reports from Hyōgo/Osaka, which it did not in other reports. This seems to have been a consequence of the emergence by 1871 of a very vigorous Chamber of Commerce, which devoted a lot of its energies to calculations of trade. The attitudes of consular officials in Hyōgo/Osaka were also colored by problems of transshipment, that is the arrival in any open port by coastal shipping of goods for export, or the dispatch by coastal shipping of goods from abroad. The transshipment issue was not confined to Hyōgo, but it did loom larger there than elsewhere because the port was located at a central point on the long coastline between Nagasaki and Yokohama. The problem of transshipment was not in statistical terms a specifically Japanese one. It was universal, solved in Europe at an early date only by acceptance of the practice of recording goods shipped

216

EARLY JAPANESE TRADE

between domestic ports at the first point of landing or last point of departure. Transshipment was first highlighted by the exports of tea, for which the issue arose as soon as Osaka became an open port: the Custom House counts of exports of tea in Osaka and Hyōgo proved to differ widely from the figures compiled by the Chamber of Commerce for 1868–1871. The Custom House figures notably exceeded the Chamber’s calculations even after adjustment for omissions in the Chamber’s figures in 1869 and 1870 was taken into account, and fell below them in 1868 and 1871. The erratic pattern was determined by the statistical fortunes of tea, depending on whether it was counted as a direct shipment from Osaka to the United States or as a transshipment via Yokohama. Export totals in the Custom House returns for 1868 and 1871 greatly exceeded Chamber of Commerce estimates or counts (made with access to the record of permits in the Custom House), and an Osaka consular official observed that “it is possible the customs may include them in the shipments to the latter port [Yokohama], while the Chamber properly includes them in the direct shipment to New York.”145 He was speculating that Osaka Custom House clerks recorded Osaka goods destined for transshipment from Yokohama as domestic (coastal) trade items, not as export trade items. The inconsistency between years would suggest a variable practice at Osaka, with tea sometimes included as an export and on other occasions treated as a coastal shipment. In late bakumatsu times and in early Meiji years, Custom House officers kept an account of both direct imports and exports and of intratrade between the open ports. Thus the Kanagawa (Yokohama) consular report for 1869 gave totals of trade for both categories. Imports of 24.3 million dollars included 17.4 million dollars of direct imports and 6.9 million dollars of foreign goods transferred from other open ports. Exports were 14.5 million dollars: 11.5 million dollars of exports originating in the ports and 3.0 million dollars originating in transfers from other open ports.146 The scope for double counting thus appeared in early statistics. In the case of imports, goods transshipped at Yokohama to Osaka/Hyōgo in 1873 and 1874 appeared as fresh imports in Osaka figures, “and have thus been entered twice in the Custom House returns.”147 Parkes in Tokyo noted in 1874 the need to keep apart the figures for transfer trade and foreign trade.148 This was notoriously a problem for the early figures for Osaka. In the case of exports, the anomalies that had led to double counting seem to have been winkled out in time. In 1879, tea shipped in Osaka to Yokohama for the United

STATISTICS OF TOKUGAWA COASTAL TRADE

217

States was not counted at the Custom House in Osaka: the consular officer, noting that the customs returns were not a complete account of tea passing through the port, obtained an estimate from the American consulate of the amounts where the ultimate consignees were in the United States.149 In the case of imports, we can say that foreign goods arriving from Japanese ports were not included in Hakodate returns, at least on the evidence of Hakodate figures for goods from Nagasaki and Yokohama in 1864 and from Yokohama again in 1881.150 Nor, it appears from evidence in 1881, did Osaka totals record the growing inflow of foreign textiles arriving via Yokohama.151 CONCLUSION

The statistical achievements of the 1860s and 1870s were real. The early years were slowed by the fact that the first statistics were collected by bugyōsho officials. Effective administration had to await centralized supervision, which was instituted finally in 1873. Japanese statistical achievement in the 1870s was not confined to trade statistics. A modern style population census did not take place until 1920, for economic reasons, but already in the 1870s, officials very cogently made the case for a census. For the short and medium terms, improvements in the system of household registration of Tokugawa times made it possible to go without one. The lament of British officials in 1873 that “[i]n Japan, the science of statistics is in its infancy” was disproved by the rapidity of advances in the 1870s;152 over the decade, statistical deficiencies were addressed. Of course, statistical information was of value only to those who had ready access to the figures. Limited access was a point of adverse British comment as late as 1877, but this defect too was addressed.153 For trade, well prepared publications began to appear from 1873 onwards in a form easy for Westerners to assimilate. These were not formal publications. Regular annual publications of trade statistics would debut in 1882. Early Meiji publications containing trade figures, moreover, were but one element in a much larger array of official publications, akin to those of Western countries. Most were in Japanese, though some also in English. By the late 1880s, the British Parliamentary Papers document the circulation to newspapers of summaries of a year’s trade at the outset of the following year.154 There are two ways of looking at the trade statistics of bakumatsu and early Meiji Japan. One is the unrealistic and culturally arrogant one of damning them because they fell short of Western standards. The other

218

EARLY JAPANESE TRADE

is to see them from the outset as a success. The task was to collate the new detail collected in the Custom Houses into statistics. Those Custom Houses themselves represented an institutional innovation. After 1868, this decentralized work had to be brought into a centralized structure, producing in the process more effective and accessible figures. The process of centralization, tracing its origins back to 1859, required a whole sequence of events: first, a response within the bugyō offices of the ports in collecting duties and in assembling the detail into statistics of trade which we can date to 1859–1867; second, imposing centralized direction on this work in 1868–1871 and, from 1871, of doing so under the aegis of the Ministry of Finance; and finally, streamlining procedures and publication in the period 1873– 1882. The major achievement was that the work approached Western standards within about fifteen years from the opening of ports in 1859. Trade figures have to be respected on their own terms. Japan presents an unusual problem in that it had to adapt to a pattern of financial settlement introduced by Westerners in the 1850s that made the dollar the unit of account and the most common medium of exchange in the open ports, despite the fact that gold was the currency of Edo (and from 1871, the unit of account for all Japan). Japan also had to cope with the fall in value of silver, a vitally important medium of exchange. Japan was not alone in having to deal with such problems, but as its foreign trade burgeoned, and as it found itself involved in two currency regions, it had a hard row to hoe. All early commodity statistics are deficient for balance of payment purposes, whether in prices or in omission of categories of goods. For calculating the commodity component of the balance of payments, it is legitimate to attempt to adjust them. In the case of the United Kingdom, for instance, this was done to great effect by A. H. Imlah in his Economic Elements in the Pax Britannica.155 However, such calculations have to be confined to specific purposes with caveats about their own limitations, if applied to wider ends. If they are quoted not as estimates for a well defined purpose but treated as a critique of the statistics themselves, then they result in some misunderstanding of the problems and an underestimation of the achievement itself. Where, as in Japan’s case, there is a belief that several sets of trade statistics exist, new attempts to calculate statistical series for commodities pose a danger of adding to the confusion. The gathering of statistical information and its processing into aggregates of trade were more systematic and successful than critics charged. Japan’s success in creating modern statistics was quite impres-

STATISTICS OF TOKUGAWA COASTAL TRADE

219

sive. The fact that consular officials over the 1870s increasingly conceded the merits of the figures illustrates this. Chamber of Commerce calculations were never seen as alternative to the Japanese statistics; nor was there a basis for serious criticism of these statistics by the late 1870s. Indeed, it is ironic that, although the BPP are the best and most convenient source for trade statistics up to the mid 1880s, the comments in the BPP up to the 1870s have been responsible for a very negative view of the data at large. This is not least owing to the failure of modern commentary, Japanese and English alike, to understand the significance of the “disposals” concept to which contemporary English diplomats and consular staff had attached so much importance. The statistics are, of course, not perfect. The lack of c.i.f. (cost of insurance and freight included) and f.o.b. (free on board) definitions marks them off from other figures. However, European trade figures also adjusted to demands for greater precision about definitions only late in the day. Here it is important to bear in mind the defective valuation of imports to Britain to 1854, highlighted in Imlah’s study. Even after 1854, the Board of Trade in London at first valued imports and exports simply by canvassing dealers for price information. Declared values (that is, values as declared at the time of making entry at the Custom House) became the practice in Britain only from 1868, but they had been the norm in Japanese statistics from 1859.156 Moreover, given the deficiencies in Japanese statistics in recording ships, it is salutary to recall that in Britain, a ship exporting country, ships sold abroad were not included until 1899, and figures of imports of coin and bullion were collected only from 1858.157 If we regard f.o.b. and c.i.f. price definitions as a refinement of very recent origin, and not as a canon for judgment of mid nineteenth century statistics, Japanese figures were collected and totals aggregated efficiently by the late 1870s. From 1875, the earlier omission of the major part of purchases on official account was remedied.158 Criticism on c.i.f. and f.o.b. grounds was imprecise in the 1860s and 1870s because the concept was a new and unfamiliar one internationally. While some relevant observations on the score of valuations were made in the early 1860s, the main concern was one about disposals, a concept that ran counter to the entire object of statistics of international trade. By the 1890s, it was less a question of remedying defects than of adopting more advanced statistical concepts that had become widely recognized only in the immediately preceding decades. This article has examined the Japanese capacity to collect statistical information and retain a record of it. A study such as this should be

220

EARLY JAPANESE TRADE

placed in the context of the archival history of Japan. Part of that is the story of destruction by fire or war, but the absence of records is more often due to the fact that records were not retained. The diffuse copying of the records by or for officials was both a product of such a system and, at the same time the means by which some material for the years before 1868 has survived in random fashion. It is striking, given the nature of the system, that it was private (or at least unofficial) copying, usually into manuscript notebooks (shahon ෗ᮏ), that ensured that many statistics survived. As proof of how fragile the source base is, we need merely point out that the actual counts in dollars for the years 1859–1884 are known only from BPP. The British reporting shows evidence that its authors sometimes consulted the Americans, who retained records of their own on trade. The British sources are the best known, largely because of the magnificent series of consular reports in BPP, and their prompt publication. As early as the 1880s, they provided data for the first Japanese scholars of foreign trade, and their figures have often been repeated in later studies both in Japan and abroad. The Japanese trade of other Western countries, except the United States, was relatively small, but a consultation of their consular sources, manuscript or published, should in all probability add to the story of relations between officialdom and the new communities of the open ports, and perhaps to our understanding of the evolution of the data themselves. The sheer capacity of the Japanese administrative system to function through the bakumatsu and early Meiji years is noteworthy. The ability of the system to function in crisis, as in the 1850s and 1860s, is important; it is a measure by which to assess bakumatsu society and polity in general. It suggests we need to qualify widely held assertions of administrative breakdown. It has been assumed that the absence of statistical information itself points to administrative failure in the 1850s. The capacity, however, was there, as indicated by the fact that the British diplomat Mitford got a full summary of the census of Osaka for 1866 even though the last known figures from the Japanese side are for 1862.159 Trade figures existed in the 1850s and 1860s, and an ability to innovate is illustrated in particular by the figures for Ezo ⼎ ዀimports to Osaka in 1856–1858 or for Edo in 1856. As statistics are a creation of administration, their existence, even if we can document them only imperfectly, shows that the Japanese administration remained in working order.

STATISTICS OF TOKUGAWA COASTAL TRADE

221

APPENDIX

Appendix 1: The Paske-Smith Figures for Exports and Imports in the Open Ports

The much quoted figures compiled by Paske-Smith for his book in 1930, together with other data principally drawn both directly and indirectly from British sources, are the basic figures around which Ishii in 1944 built his tables of the first decade of Japanese trade in the open ports. It is a matter of interest, therefore, to seek the identity of the sources Paske-Smith drew on in 1930, especially given some wide variations in his figures from those in the BPP. (Paske-Smith does not appear to have consulted parliamentary papers at large, though from internal evidence at least one paper in either published or draft form was available to him.) An answer to the question is all the more pressing as early consular records for Japan are not listed in the 1963 Guide to the Contents of the Public Record Office, part 2 (1973), p. 142. Details appear only in the very recent on-line listing of records. They also have high list numbers, which seems to confirm that their listing is comparatively recent. The records seem to have been “repatriated” to London at the outset of 1942. For Yokohama, local consular records exist from the last quarter of 1923 and again from 1945, although the earlier records were destroyed in the earthquake. From Nagasaki, they run from the outset of the mission to 1937. None now exist for Hakodate, and no pre Second World War records exist for Osaka/Kobe. It would seem then that Paske-Smith would have had easy local access to statistical material in the 1920s. The abrupt end of the Nagasaki records in 1937 suggests that, perhaps, records to that day had already been transferred to Tokyo for unknown reasons before the war crisis. The absence of consular records for Hakodate and for Osaka might appear speculatively a result of misfortune in their transfer, or of the sheer impossibility of moving records at short notice, from remote locations to Tokyo in 1942. Apart from consular archives, early archives of the embassy (and former legation) itself were already in London in the FO 242 series. This series, though principally one of correspondence with London and of miscellaneous correspondence, proves on inspection to have incoming consular letters to Edo (Tokyo) from 1858–9 for Hakodate, from 1859 for Nagasaki, from 1862 for Kanagawa, and from 1868 for Hyōgo. While these transferred series in general are held to be very defective, the FO 262 series for the Japanese volumes consulted seems more or less

222

EARLY JAPANESE TRADE

complete in some but not in other years. Effectively, the incoming letters from consulates touch fleetingly on other issues, while dealing mainly with establishment matters and in fascinating detail with the working of extraterritoriality (the legal functions exercised by consular stations). Thus, while consular archives no longer survive for any consular post except Nagasaki prior to 1923, the FO 262 series serves as surrogate consular archives up to a point. The Kanagawa consular office was unaffected by the great fire in 1866 in Yokohama (though the records of the Chamber of Commerce were destroyed). Remarkably, and as a sign of how little regular statistical contact officials actually maintained with the Custom Houses, the consul became aware of the destruction of all of the records of the Yokohama Custom House only when he applied on 26 March 1867 for trade figures for 1866 (Consul to minister, Tokyo, Kanagawa, 3 April 1867, FO 262/133, f.75). The incoming correspondence includes the draft annual reports, which appear with very minor change in the published BPP. There is however surprisingly little reference to statistical issues in the incoming letters, which would suggest that they took up little of the routine time of busy officers, and would also explain their unease, faced with a demanding annual chore, in handling statistical matters and at times for their very testy comment. The only true early consular archives prior to 1923 (i.e. archives maintained at a consular station) are those of Nagasaki, which are extensive though not complete for the early decades but fuller in these years than for later times. Documents of particular interest are two copy letter books, one containing the text of letters in Dutch, almost all to the “governor” of Nagasaki for June 1859–December 1862 (TNA, FO 796/18); the other to local Japanese officials 3 Jan.1863–29 Oct.1866 (TNA, 796/25). In addition to letter books for copies of outward letters from Nagasaki, there survive in broken fashion volumes of loose letters, variously incoming or outgoing, some in draft form (i.e. prior to their translation into Japanese) and some incoming letters in Japanese. In the present context, given the tensions displayed in the reports in BPP over access to or adequacy of statistical information, the most striking point is the high handedness of the language often employed by consular officials. The Nagasaki records are a remarkable illustration of a consular station at work. However, reference to trade is sparse, arising from individual cases concerning customs duties (an issue lacking in the surrogate FO 262 series). On infrequent occasions, the extraterritorial issue

STATISTICS OF TOKUGAWA COASTAL TRADE

223

involved customs or trade regulations in regard, for example, to breaking bulk outside the official opening hours of the Custom House. Hakodate, in particular, was a place of little trade (with accounts in the early years of often uncooperative or allegedly irregular behavior by Custom House officials often taking pride of place in reporting). However, it had an importance as whalers sailed in northern waters, and it was also a useful watching post for observing Russian intentions in the region. A palpable fear of Russian encroachment can be seen in the letters to the legation in Edo, a subject of even more concern to the Japanese themselves, of course. In 1862, in the course of a long letter on the Russian problem, the British consul observed that “the highest Japanese functionary trembles before the lowest Russian adorned with a white or yellow cap-band” (TNA, FO 262/44, fll.21–34 Enslie to legation, 2 Sept. 1862). A general impression from the records is of a well organized and efficient consular service maintaining close ties with the legation. Appendix 2: Return of Osaka Coastal Trade for 1866

The Osaka bugyō told a British official in 1867 that some figures for coastal imports, though they did not cover all commodities, existed, and he duly provided details of 20 commodities.160 They are reproduced here as they have not been adverted to in modern works. Coastal Imports in Osaka, 1866

Rice Other cereals and vegetables Oil Soy Salt Charcoal Wood Seaweed Dried bonito Teas Paper Floss silk Vegetable wax Raw silk Lead (Japanese) Iron

311,258 koku 64,535 koku 23,932 koku (sic) 11,000 tubs 916,270 bags 1,253,880 bags 921,540 bundles 24,900 bundles 63,560 kuamme (3,972 piculs) 104,470 kuamme (5,779 piculs) 134,00 bundles 22,500 bundles 47,290 packages 3,300 kuamme (206 piculs) 800 piculs 31,500 bundles

224

EARLY JAPANESE TRADE

Leaf tobacco Cut tobacco Mats (fine) Mats (rough)

9,250 bundles 13,760 boxes 15,000 packages 105,000 packages

The small number of items – a mere 20 – shows how incomplete the return was. While the governor had noted some omissions (such as sake, a major item in coastal trade), what is more striking is the total absence of cloth. The figure for rice is higher than in earlier trade returns, suggesting that it included kuramai ⶶ⡿(rice on domain or daimyo account, normally not included in trade figures of Osaka). The relatively modest total in the wake of the bad harvest of 1865, which would depress the overall level of trade in rice in 1866, is consistent also with the low number of ships entering the port in the course of 1866. Appendix 3: Problems in BPP Figures for Nagasaki and Yokohama, 1863

Up to mid decade a recurrence of problems in one or other port was a feature of statistical reporting, but in two cases, Nagasaki (1862 and more particularly 1863) and Yokohama (1863), figures present an uncommon degree of confusion. A. Nagasaki Trade Figures for 1863 Nagasaki Trade in 1863 BPP* (in dollars) exports

1,388,071

Paske-Smith† Japanese return۬ (in dollars) (in taels) 925,000

3,470,182

602,000

3,552,967

(925,381) imports

1,421,885 (947,458)

*

†

Figures from BPP, vol. 4, p. 121; figures in parentheses from BPP, vol. 4, pp. 205–206 (adding totals for British and foreign vessels). Figures of 1,388,071 dollars and 1,421,885 dollars, differing from those given in his table on p. 303 are also cited elsewhere by Paske-Smith 1930, p. 206. Table from Paske-Smith 1930, Appendix 1, p. 303.

۬ Returns in taels, from BPP, vol. 4. pp. 143–47; conversion of taels to dollars at variable rates.

STATISTICS OF TOKUGAWA COASTAL TRADE

225

Import figures of 947,458 dollars and export figures of 925,381 and the variant export figure of 1,388,071 dollars and import figure of 1,421,885 dollars seem to be a consequence of confusion caused by different rates of conversion from taels to dollars. Returns in taels by definition originated with Japanese officials rather than in a consular attempt at counting. The conversion rate of 5.85 taels to the dollar given by Paske-Smith for the trade of 1863 is erroneous. It is suspiciously close to the fixed rate of 5.90 taels that was used for valuing invoices in gold currencies (while the market rate had sunk to 3.370 in 1863). The tael was an archaic measure used in the international trade of east Asia, and was going out of fashion in Nagasaki. In 1865, and for some years thereafter, figures in Nagasaki were valued in ichibus, an actual coin and not an archaic method of valuing in a notional money of account. The amounts in ichibus were then converted into dollars at the market rate. As late as February 1863, the British consuls did not have independent knowledge of the conversion rate of the tael.161 In the case of the calculation for 1862, the presence of both taels and dollars had already led to confusion. In addition to a fixed rate for converting imports from gold currency countries, both fluctuating market rates and an accelerating fall in the market value of silver were possible sources of confusion. If conversion is done at a high rate of 5.90 taels to the dollar, imports of 3,552,967 taels in 1863 became 602,198 dollars. (This is by coincidence, or otherwise, Paske-Smith’s low figure for imports.) A total of 925,381 dollars for exports, if a rendering of a count of 3,470,182 in taels, would suggest a tael exchange of 3.75, a figure quoted in 1865 as an average value of the tael in 1863 and 1864.162 A total of 1,388,071 dollars, on the other hand, seems unduly high as it implies for 1863 a rate of less than 3 taels to the dollar. Thus, the likely figures are 925,381 dollars for exports (a figure close to that in Paske-Smith’s table), and 947,458 for imports. The higher figure of 1,421,885 dollars given in one BPP version requires a very low tael rate. The figures of 925,381 dollars for exports and of 947,458 for imports given in a table for the years 1863–65 for all three open ports would seem to confirm the validity of these figures, even though they may have been incorporated, perhaps with the wisdom of hindsight, into a report from Kanagawa on 28 March 1866.163 B. Yokohama Trade Figures for 1863

The figures for Yokohama in 1863 presented Ishii with more problems, and he spent more time on them than on figures for any other

226

EARLY JAPANESE TRADE

year. Finally, he opted for a high total of 10,554,012 dollars for exports rather than the alternative of only half that figure; and for imports he opted for 3, 244,584 dollars.164 In making this choice, he disregarded without discussion a higher figure for imports of 3,474,749 dollars that had appeared in his own Table no. 3.165 Here the appearance of confusion is much greater than the reality. The problem lay in the fact that for that year officials presented both the “declared values,” that is the figures as computed by the Custom House, and adjusted imports on criteria of their own (adjusted to account for information on prices, for example). Defective valuations were the most easily identified deficiency in the statistics. Silk was the dominant item in the exports of Yokohama; even in terms of declared values, it accounted for 4,127,340 dollars, or 80 percent of total exports of 5,134,184 dollars. In the declared or official values, silk was valued at 210 dollars per picul; in the appendix to the Kanagawa consul’s report it was valued at 450 dollars apiece. Nowhere is it stated that the figures had been revalued. Figures in dollars tables were declared to be at “the valuations accepted thereon at the Custom House,” but there was no indication in the text that figures in sterling arose from a conversion into sterling of data in the appendix for imports at revised valuations. Figures for exports also were adjusted. The table below sets out the various estimates for the confusing data for this year. Value of Yokohama Trade in 1863 in BPP Reports+ Consul’s report

Alternative estimate (appendix of consul’s report)

Dollars

Sterling

Dollars

(Sterling)*

Exports

5,134,184

2,638,503

10,554,022

(2,638,505)

Imports

1,595,170

811,146

3,244,589

(811.147)

+

BPP, vol. 4, pp. 87–101, especially pp. 87, 94–95. This table is obviously drawn from the same source as PaskeSmith’s account (Paske-Smith 1930, pp. 205–206).

*

Actual sterling totals are not given by the Kanagawa consul in the appendix; they are as cited in the report, resulting from conversion of dollar totals at 5s.0d. to the dollar.

The tabular presentation above identifies the cause of the apparent confusion in the BPP. The dollar and sterling totals in the consul’s report, viewed in the context of the report itself, would imply a conversion rate

STATISTICS OF TOKUGAWA COASTAL TRADE

227

of 1.9 in the case of both imports and exports. This cannot be the base of conversion, as the consul was aware of a conversion rate of 5s.0d. to the dollar, and quoted it several times. In other words, the figures are independent of each other. In the case of the alternative estimate, the figures in the appendix to the report were given in a series of separate tables for ships of various nations that were not summed up into grand totals for exports and imports. If added up and converted at 5s.0d. to the dollar, the dollar total yields the sterling figures which appear in the consul’s report. Thus, while dollar totals in the report are the “declared values” at the Custom House, the sterling totals in the consul’s report are not conversions of the Custom House declared values, but conversions into sterling of higher gross totals in dollars created by the consular revaluation process. In the body of the consular report (if not in the appendix), those totals seems to be identified quite erroneously as sterling equivalents of the declared values. In other words, there is neither a serious error in the figures, nor a gross confusion by the consul, but an apparent rather than real contradiction caused by the lack of an unambiguous statement of the different basis on which the figures in the appendix were compiled.166 The lack of an actual gross total for component detail in the tables in the appendix made it easier not to advert to the error. Acknowledgments

I am greatly indebted to Professor Saitō Osamu, Hitotsubashi University, Tokyo, and Prof. Katsuta Shunsuke, Gifu University, for assistance in getting documents and for help with bibliographical enquiries and difficult passages. I am also indebted to Professor Jim Baxter of the International Research Centre for Japanese studies for his editorial work on the first stages of revision of this article, and to two anonymous referees for much appreciated suggestions on restructuring the paper. Other obligations are acknowledged in the footnotes. I am alone responsible for any errors which remain. REFERENCES Abbreviations

BPP: British Parliamentary Papers; KBS: Kokuritsu Kōbunshokan; NDL: National Diet Library; YKS: Yokohama Kaikō Shiryōkan

228

EARLY JAPANESE TRADE

Alcock 1863 Rutherford Alcock. The Capital of the Tycoon: a Narrative of Three Years Residence in Japan, 2 vols. Longman, Roberts, and Green, 1863. Baba and Tatemoto 1968 Masao Baba and Masahiro Tatemoto. “Foreign Trade and Economic Growth in Japan: 1858–1937.” In Economic Growth: The Japanese Experience since the Meiji Era, ed. Lawrence Klein and Kazushi Okawa. Yale Economic Growth Center, 1968. BPP British Parliamentary Papers. Area studies, Parliamentary Papers, Japan, ed. W. G. Beasley, 10 vols. Irish University Press, 1971–1972. Dai Nihon gaikoku bōeki nenpyō 1990 Dai Nihon gaikoku bōeki nenpyō: Meiji 15, 16, 17 ኱᪥ᮏእᅜ㈠᫆ᖺ ⾲㸸᫂἞ 15, 16, 17.Ōkurashō, 1882–1883, 1884; repr. Tōyō Shorin, 1990. Cullen 2006 Louis M. Cullen. “Tokugawa Population: The Archival Issues.” Japan Review 18(2006). Cullen 2009 ———. “Statistics of Tokugawa Coastal Trade and Bakumatsu and Early Meiji Foreign Trade, Part 1: Coastal Trade in Tokugawa Times,” Japan Review 21(2009). Foreign Trade Returns , 1866–72, 1866–73 KBS Kokuritsu Kōbunshokan ᅜ❧බᩥ᭩㤋National archives, Tokyo. 187– 222. a. Kakkō yushutsu buppin hyō 1866–1873 ྛ ㍺ฟ≀ရ⾲ woodblock print; b. Kakkō yushutsu buppin kinryō daika hyō 1866–1872 ྛ ㍺ ฟ≀ရ᩹㔞௦౯⾲manuscript. Foreign Trade Returns, 1870–72 KBS Kokuritsu Kōbunshokan, Returns of foreign trade. 1870–72 (187–247 i, returns, second half of 1870; 187–247 ii, returns, two half years 1871; 187–247 iii, returns 1872: for full titles see below under Foreign Trade Returns, 1866–1880 YKS). Foreign Trade Returns 1866–1877, NDL National Diet Library YD-212 (microfilm), Parliamentary documents and official papers room. reel 1( filmed 1964), Exports 1866–72, 1866–73, Exports and imports, 1870 (in two parts), 1871 (in two parts), 1872; Reel 2 (filmed 1963), Exports and imports, 1873, 1874,1875 (in two parts for each year), year ended June 1876, half year ended Dec. 1876, year ended June1877.

STATISTICS OF TOKUGAWA COASTAL TRADE

229

Foreign Trade Returns, 1866–1880, YKS Yokohama kaikō shiryōkan Yokohama archives of history 1866–1880 ᶓ὾㛤 ㈨ᩱ㤋1866–1880 (14 vols of photocopies of returns of foreign trade variously in woodbook printing and movable type). Vol. 1, a. Kakkō yūshutsu buppinhyō 1866–1873 ྛ   ㍺ ฟ≀ရ⾲(Statistics Bureau of Ōkurashō 1874) woodblock print. b. Kakkō yūshutsu buppin kinryō daika hyō 1866–1872 ྛ ㍺ฟ≀ရ᩹㔞௦౯⾲ manuscript; Vols. 2, 3, Kakukaikōjō yūshutsunyū buppindaka ྛ㛤 ሙ㍺ฟධ≀ ရ㧗 Total exports and imports of every open port, half yearly 1870, 1871; vol.4, Kakukaikōjō yūshutsunyū buppindaka hyō ྛ㛤 ሙ㍺ฟධ ≀ရ㧗⾲ Tables of total exports and imports of every open port, year 1872; vol. 5, Dai Nihon kakkō yūshutsunyū buppin nenpyō ኱᪥ ᮏྛ ㍺ฟධ≀ရᖺ⾲ Yearly table for exports and imports of every port of empire of Japan (first bilingual annual report with original title in Engish of “Annual return at all the ports of Japan 1873”); vols. 6, 7, 8, Nihon gaikoku bōeki hannenhyō ᪥ᮏእᅜ㈠༙᫆ᖺ⾲ Half-yearly tables of Japanese foreign trade, half yearly 1875, 1879, 1880; vol. 9 and 10, bilingual monthly returns 1873; vol. 11, bilingual monthly returns Jan. 1874–May 1876*; vol. 12, 13, 14, bilingual monthly returns 1877– 1880* (* above: NDL YD-212 ( microfilm reel 2 )has variously yearly or half-yearly returns for years 1873–1877). Frost 1970 Peter Frost. The Bakumatsu Currency Crisis. East Asian Research Center, Harvard University Press, 1970. Howe 1999 Christopher Howe. The Origins of Japanese Trade Supremacy. Hurst and Co., 1999. Imlah 1958 A. H. Imlah. Economic Elements in the Pax Britannica: Studies in British Foreign Trade in the Nineteenth Century. Harvard University Press, 1958. Inoue Kaoru kankei monjo NDL Inoue Kaoru kankei monjo ஭ୖ㤾㛵ಀᩥ᭩National Diet Library Kensei Shiryōshitsu ᅜ❧ᅜ఍ᅗ᭩㤋᠇ᨻ㈨ᩱᐊ. Inoue Kaoru monjo kankei mokuroku 1975 NDL Inoue Kaoru kankei monjo mokuroku ஭ୖ㤾㛵ಀᩥ᭩┠㘓National Diet Library Kensei Shiryōshitsu, 1975. Ishii 1944 Ishii Takashi ▼஭Ꮥ. Bakumatsu bōeki shi no kenkyū ᖥᮎ㈠᫆ྐࡢ◊ ✲. Nihon rekishigaku taikei 4 ᪥ᮏṔྐᏛ኱⣔㸲. Nihon Hyōronsha, 1944.

230

EARLY JAPANESE TRADE

Itani 1931 Itani Zen’ichi ⊦㇂ၿ୍. “Bakumatsu no shakaikeizaiteki igi” ᖥᮎࡢ♫ ఍⤒῭ⓗព⩏. Shakai keizai shigaku ♫఍⤒῭ྐᏛ1:1 (1931). Katsu 1967 Katsu Kaishū ຾ᾏ⯚. Kaigun rekishi ᾏ㌷Ṕྐ. In Katsu Kaishū zenshū ຾ᾏ⯚඲㞟, vol. 8. Hara Shobō, 1967. Katsu 1978 ———. Suijinroku ྿ሻ㘓(1890), part 5. In Katsu Kaishū zenshū 10. Keisō Shobō, 1978 Kawai 1895 Kawai Toshiyasu Ἑྜ฼Ᏻ. “Bakumatsu no bōekigaku” ᖥᮎࡢ㈠᫆㢠. Tōkeigaku zasshi⤫ィᏛ㞧ㄅ109 (1895). KBS Kokuritsu Kōbunshokan, See Foreign Trade Returns, 1866–1872, 1866– 73 KBS, and 1870–72. Matsukata 1899 Masayoshi Matsukata. Report on the Adoption of the Gold Standard in Japan. Government Press, 1899. Matsukata Masayoshi monjo mokuroku 1954 NDL Matsukata Masayoshi monjo mokuroku ᯇ᪉ṇ⩏ᩥ᭩┠㘓 National Diet Library Kensei Shiryōshitsu, 1954. Matsukata Masayoshi kankei monjo 1981 Matsukata Masayoshi kankei monjo ᯇ᪉ṇ⩏㛵ಀᩥ᭩ vol. 3Daitō Bunka Daigaku Tōyōkenkyūjo ኱ᮾᩥ໬኱Ꮫᮾὒ◊✲ᡤ, 1981 Matsuo ke monjo mokuroku NDL Matsuo ke monjo mokuroku ᯇᑿᐙᩥ᭩┠㘓. National Diet Library Kensei Shiryōshitsu. Meiji zaisei shi 1971 Meiji Zaisei Shi Hensankai ᫂἞㈈ᨻྐ⦅⧩᭳, ed. Meiji zaisei shi ᫂἞㈈ ᨻྐvol. 7. Reprint by Yoshikawa Kōbunkan, 1971. (Orig. pub. 1904.) Nagasaki ken shi 1985 Nagasaki ken shi: Taigai kōshō hen 㛗ᓮ┴ྐ㸸ᑐእ஺΅⦅. Ed. Nagasaki ken shi Henshū Iinkai. Yoshikawa Kōbunkan, 1985. NDL National Diet Library, Tokyo. See Foreign trade returns 1866–1877, NDL; see also Inoue Kaoru kankei monjo, Inoue Kaoru kankei monjo mokuroku,

STATISTICS OF TOKUGAWA COASTAL TRADE

231

Matsukata Masayoshi monjo mokuroku, Matsuo ke monjo mokuroku, Shōda ke monjo mokuroku. Nihon bōeki seiran 1975 Tōyō Keizai Shinpōsha ᮾὒ⤒῭᪂ሗ♫, ed. Nihon bōeki seiran ᪥ᮏ㈠ ᫆⢭ぴ. Tōyō Keizai Shinpōsha, 1975 (first ed. 1935). Nishikawa, Odaka, and Saitō 1996 Nishikawa Shunsaku すᕝಇస, Odaka Kōnosuke ᑿ㧗↥அຓ, and Saitō Osamu ᩪ⸨ಟ, eds. Nihon keizai no nihyaku nen ᪥ᮏ⤒῭ࡢ200ᖺ. Nihon Hyōronsha, 1996 Paske-Smith 1968 M. Paske-Smith. Western Barbarians in Japan and Formosa in Tokugawa Days, 1603–1868 (Orig. pub. 1930). Paragon Book Reprint Corp., 1968. Ravina 1999 Mark Ravina. Land and Lordship in Early Modern Japan. Stanford University Press, 1999. Shōda ke monjo mokuroku NDL Shōda ke monjo mokuroku ຾⏣ᐙᩥ᭩┠㘓National Diet Library Kensei Shiryōshitsu. Sugiyama 1988 Shinya Sugiyama. Japan’s Industrialisation in the World Economy, 1859–1899: Export Trade and Overseas Competition. Athlone Press, 1988. Tatemoto 1968 Masahiro Tatemoto. “The Correction of Official Import Statistics, 1874– 1887, for the Depreciation of Silver,” in Masao Baba and Masahiro Tatemoto, “Foreign Trade and Economic Growth in Japan: 1858–1937.” In Economic growth: the Japanese Experience since the Meiji Era, ed. Lawrence Klein and Kazushi Okawa. Economic Growth Centre, 1968. TNA The National Archives, Kew, London. FO 262, legation Edo/Tokyo: correspondence with London and consulates, 61volumes 1859–1875: Nagasaki, Kanagawa, Hyōgo/Osaka/ Kōbe, Hakodate, Niigata; FO 796, consulate Nagasaki, 56 volumes 1859–1873: letter books, loose letters. Umemura and Yamamoto 1989 Umemura Mataji ᱵᮧཪḟ and Yamamoto Yūzo ᒣᮏ᭷㐀. Kaikō to ishin 㛤 ࡜⥔᪂ Vol. 3 of Nihon keizai shi ᪥ᮏ⤒῭ྐ. Iwanami Shoten, 1989

232

EARLY JAPANESE TRADE

Yamaguchi 1968 Yamaguchi Kazuo ᒣཱྀ࿴㞝. Nihon keizai shi ᪥ᮏ⤒῭ྐ Vol. 12 of Keizaigaku zenshū ⤒῭Ꮫ඲㞟. Chikuma Shobō, 1968. Yamaguchi 1990 ———. “Dai Nihon gaikoku bōeki nenpyō ni tsuite” ኱᪥ᮏእᅜ㈠᫆ ᖺ⾲࡟ࡘ࠸࡚ . In Dai Nihon gaikoku bōeki nenpyō, Meiji 15, 16, 17, comp. Ōkurashō. Repr. Tōyō Shorin, 1990. Yamaguchi and Ōuchi 1968 Yamaguchi Kazuo ᒣཱྀ࿴㞝and Ōuchi Tsutomu ኱ෆຊ. Meiji shonen no bōeki tōkei ᫂἞ึᖺࡢ㈠᫆⤫ィ. Tōkyō Daigaku Shuppankai, 1968. Yamazawa and Yamamoto 1979 Yamazawa Ippei ᒣ⃝㐓ᖹ and Yamamoto Yūzō ᒣᮏ᭷㐀. “Bōeki to kokusai shūshi” ㈠᫆࡜ᅜ㝿཰ᨭ. Chōki keizai tōkei 㛗ᮇ⤒῭⤫ィ14, Tōyō Keizai Shinpōsha, 1979. Yasuoka 1990 Yasuoka Shigeaki Ᏻᒸ㔜᫂. “Zenkoku shijō no keizai henka” ඲ᅜᕷሙ ࡢ⤒῭ኚ໬. In Shinshū Ōsaka-shi shi ᪂ಟ኱㜰ᕷྐ, vol. 4. Ōsaka shi Shi Hensan Iinkai, 1990. YKS Yokohama Kaikō Shiryōkan. See Foreign trade returns, 1866–1880, YKS Yokohama-shi shi 1959 Yokohama-shi shi ᶓ὾ᕷྐ, vol. 2. Yokohama shi, 1959. Yokohama-shi shi: shiryō 1962 “Nihon bōeki tōkei 1868–1945” ᪥ᮏ㈠᫆⤫ィ1868–1945. Yokohama-shi shi: Shiryō hen, vol. 2, ᶓ὾ᕷྐ㸸㈨ᩱ⦅஧Yokohama shi, 1962. Yokohama-shi shi 1963 Yokohama-shi shi, vol. 3, pt. 2. Yokohama shi, 1963. せ᪨ ᚨᕝᮇἢᓊ஺᫆࠾ࡼࡧᖥᮎ࣭᫂἞ึᮇᾏእ㈠᫆ࡢ⤫ィ 3DUWᖥᮎ࣭᫂἞ึᮇ࡟࠾ࡅࡿ㈠᫆ ࣝ࢖࣭㹋࣭࢝ࣞࣥ ᖥᮎ࠿ࡽ᫂἞15 ᖺ࡟࠸ࡓࡿ᪥ᮏࡢ㈠᫆࡟㛵ࡍࡿ㈨ᩱホ౯ࡀࠊ ࡇࡢ஧㒊ᵓᡂࡢㄽᩥࡢ୺ࡓࡿࢸ࣮࣐࡛࠶ࡿ㸦Part 1 ࡣ Japan

STATISTICS OF TOKUGAWA COASTAL TRADE

233

Review 21 ࡟ᥖ㍕ࡉࢀࡓ㸧ࠋ1859 ᖺࡢ㛤 ௨㝆ࡢ⤫ィ㢮ࡣࠊ㈠ ᫆ࡢ⣡ရ᭩࡞࡝ࢆᖥᗓࡢᐁྣࡀ┤᥋཰㞟ࡋࡓࡶࡢ࡛࠶ࡿࠋࡑ ࢀ௨๓ࡣࠊᐁྣࡓࡕࡣἢᓊ஺᫆ࡢ⤌ྜࡸၥᒇ࡞࡝࡟ࡼࡗ࡚ᥦ ౪ࡉࢀࡓᩘᏐ࡟౫ᣐࡋ࡚࠸ࡓࠋ㛢㙐ⓗ࡞㔠ᮏ఩ไᗘ࡟࠾ࡅࡿ ෇㔠㈌ࡸ෇㖟㈌ࡢᖹ౯ࡣࠊ᪂ࡋ࠸㛤 ሙ࡟࠾ࡅࡿእᅜ㈠᫆ࡢ ࡓࡵࡢᶆ‽ⓗ࡞㏻㈌༢఩࡛࠶ࡿ࣓࢟ࢩࢥ࣭ࢻࣝࡢᕷሙබ㛤࡟ క࠸ 1871 ᖺ࡟ᔂቯࡋࡓࠋ㔠ࡀὶฟࡋ᪥ᮏࡢ㏻㈌ไᗘࡀ஦ᐇ ୖ㖟ᮏ఩ไ࡜࡞ࡗ࡚ึࡵ࡚ࠊࢻࣝ㸦࠶ࡿ࠸ࡣ෇㖟㈌㸧࠿ࡽ෇ 㔠㈌࡬ࡢ᥮⟬ࡀ༢୍Ⅽ࣮᭰ࣞࢺ࡟㏆࡙࠸ࡓࠋ᫂἞ึᮇࡢ᪥ᮏ ࡢᐁྣࡓࡕࡣࠊ୙Ᏻᐃ࡞㏻㈌ไᗘࡢ୰࡛ࢻࣝ࡜㔠ࡢ஺᥮࡟⏕ ࡌࡿ⥲ィࡢ㣗࠸㐪࠸࡟ࡶ࠿࠿ࢃࡽࡎࠊすὒⓗ࡞⤫ィࡢᵓ⠏ࢆ ヨࡳ࠺ࡲࡃฎ⌮ࡋࡓ࡜ࡶ⪃࠼ࡽࢀࡿࠋᚨᕝึᮇࡢἢᓊ஺᫆࡟ 㛵ࡍࡿグ㘓㢮ࡣ෗ᮏ࡜ࡋ࡚ࡋ࠿ṧࡗ࡚࠸࡞࠸ࡀࠊࡑࢀࡽࡣ ᩘࡶᑡ࡞ࡃࠊཎᮏ࡜㣗࠸㐪ࡗ࡚࠸ࡿሙྜࡶ࠶ࡿࠋྠࡌࡇ࡜ࡀ 1860 ᖺ௦ࡢᾏእ㈠᫆࡟࠾࠸࡚ࡶゝ࠼ࡿࠋ1869 ᖺ௨㝆ࡢ᪂ᨻ ᗓ࡟ࡼࡿᣦᑟୗ࡛ࡶࠊᮍ᏶ᡂ࡞࣓ࣔࡢࡼ࠺࡞ഹ࠿࡞෗ࡋࡀ᪂ ࡋ࠸㏻㈌ὶ㏻㔞ࡢ⥲㢠࡜ࡋ࡚▱ࡽࢀࡿ࡟ࡍࡂ࡞࠸㸦୧࡟㛵ࡋ ࡚ࡣ1869 ᖺ௨㝆ࠊ෇࡟㛵ࡋ࡚ࡣ1871 ᖺ௨㝆㸧ࠋᇶᮏⓗ࡞ࢻࣝ ⥲㢠ࡣಶࠎࡢ⛯㛵࠿ࡽ୰ኸᙜᒁ࡟㏦ࡽࢀࠊࡑࡢ㘽࡜࡞ࡿᇶ♏ ࢹ࣮ࢱࡣ1880 ᖺ௦ࡲ࡛ࡢࢻࣝ࡜෇㔠㈌࡜ࡢ᥮⟬࡟ᇶ࡙࠸ࡓࡶࡢ ࡛ࠊⱥᅜ㆟఍㈨ᩱ㸦BPP㸧࡟࠾࠸࡚ࡋ࠿▱ࡽࢀ࡚࠸࡞࠸ࠋ

Source: Japan Review, No. 25 (2013), pp. 33–65

9 Archives

Japanese Archives: Sources for the Study of Tokugawa Administrative and Diplomatic History1 ™

Japanese record keeping contrasted starkly with more abundant Western archives. The poor state of shogunal records was compounded by losses in the Restoration and in the 1923 Tokyo earthquake. Domain records often fared much better, though they are often in runs of nisshi and nikki with information entered from documents later discarded. The papers of final decision making, and in general of rōjū and bugyō, were personal to the office holder. Survival of the records of daimyo who had served as rōjū depends on the integrity of domain archives. Access to the private property of former daimyo was prompted by their wish to preserve a fair record of the role of the great domains in bakumatsu times. The survival of papers of bugyō and lesser officials, unless in domain archives, was only secured by retention by heirs or random passage through the hands of collectors, private copyists or booksellers. Administration functioned through the circulation of copies, in turn often copied into either official compilations such as the Tsūkō ichiran, or private ones. In the 1860s, a pressing need for access to recent and current records led to the compilation in two stages of the Tsūshin zenran. While the diffuse holding of papers had posed the initial problem, it also provided the solution. In Osaka, a floating mass of miscellaneous paper was the source base for two compilations in the late Meiji period, one published in 1911–1913, the other only sixty years later in the 1960s. The first printed compilations 234

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

235

were four by Katsu Kaishū under official support in 1889–1893. The first volume of Bakumatsu gaikoku kankei monjo followed in 1911, itself in part made possible by the Tsūshin zenran. The Dai nihon ishin shiryō, its origins traceable to daimyo commitment in the 1880s, and heavily dependent on domain sources, finally appeared from 1938. A Ministry of Finance series, Nihon zaisei keizai shiryō (1921–1925), reveals how little material it had inherited in 1871. THE ARCHIVAL EVIDENCE OF TOKUGAWA JAPAN

Despite gaps in surviving records, substantial administrative archives have been a feature of the state in Japan as in the West. Overall, archives in Western states have been more substantial, better defined and above all more continuous, and as an unbroken corpus go back farther in history. Kikuchi Mitsuoki ⳥ụග⯆, president of the Kokuritsu Kōbunshokan ᅜ❧බᩥ᭩㤋, points out: Japan has a history of placing relatively little importance on the concept of collecting documents and saving them for future generations . . . At first sight, the public archives in Japan are very much smaller than in the West. Despite the fact that Japan is such a unique nation, perhaps specifically because it is such a unique nation, recorded documents have been far from continuous.2

There are two issues. First, the restricted scale or extent of Japanese archives; secondly, discontinuity: records of high policy and decision making are a disjointed run of notebooks or shahon ෗ ᮏ , survivors from a high rate of loss over time. However, there is a further issue. These observations refer to shogunal or national archives. The domain records, where they survive, are at times in long runs, though often they were discarded once detail was entered into diaries or registers. Nevertheless, the correspondence retained by a number of politically active domain has proved key to the study of the story of the bakumatsu shogunate. Some of it, moreover, was the correspondence which daimyo serving as rōjū ⪁୰took with them when they returned to their home domain. Japanese administration during the Edo period was shared between the shogun and the daimyo. At the outset, no corpus of directly retained clerks for higher administrative tasks existed. Delicate issues of foreign relations were handled by Zen monks, first Saishō Shōtai す➗ᢎඟ and then Ishin Sūden ௨ᚰᓫఏ.3 These monks centered on Tenryūji ኳ㱟ᑎ in Kyoto soon lost their central administrative role.

236

EARLY JAPANESE TRADE

Thereafter, they were retained by the shogunate solely as specialists in the Chinese language to oversee the Tsushima ᑐ㤿 links with Korea. After the early prominence of Hayashi Razan ᯘ⨶ᒣ (1583–1657) as adviser and drafter of documents, the Hayashi family likewise had a modest and merely scholarly role until the novel foreign crisis in the 1790s won them a new role. At that stage, Hayashi Jussai ᯘ㏙ ᩪ (1768–1841) was regularly consulted on policy issues, and Hayashi Akira ᯘ㡐 (1801–1859), directed the compilation in the late 1840s and 1850s of a huge collection of diplomatic precedents. As hereditary family head, he was titular leader of the team which negotiated with Commodore Perry in 1854. The presence of dedicated permanent staff ensured that both within the shogunate and in the domain the routine functions of accounting, and of the legalities of relations between shogun and daimyo, acquired an archival existence. In contrast, great affairs of state had no permanent administrators. As a result, intimate administrative documents (correspondence between rōjū and daimyo, and of rōjū and bugyō ዊ⾜with immediate subordinates) were de facto personal working papers. Where they survived, they did so in the hands of descendants, and in the case of rōjū office holders, in the archives of former daimyo after 1868. Copies acquired a life of their own. W. G. Aston’s account in 1873 of events in Ezo in 1806–1807 was based on access to “a collection... comprising the private correspondence of officials on duty in Hakodate, together with proclamations and other official documents.”4 Even more striking was the “little manuscript book” which later he purchased at a book stall in Tokyo, on which he based an account of the Phaeton Incident. While itself a copy of the official diary, it was actually the personal copy of the diary’s compiler, Tokuzaemon ᚨᕥ⾨㛛, incorporating reflections of his own.5 Such shahon passed in random fashion through the hands of dealers in old papers, and have been acquired even by offical purchase from booksellers. As recently as 1980, six documents with correspondence of the 1650s between the shogunate and a merchant on reopening trade with south east Asia were found in a junk shop.6 The concept of original is little used in Japan. The terms teihon ᐃ ᮏ (original or authentic text), genkō ཎ✏ (manuscript), genpon ཎᮏ (original text), gensho ཎ᭩ (original document) are overlapping and ambiguous, referring to the quality or authenticity of the text rather than to status as a holograph. The term shahon (manuscript in the sense of copybook) by definition signifies a copy. Likewise the phrase tome ␃or tomesho ␃᭩, frequently used for a file, implies a copy

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

237

(usually of a number of items entered into a single shahon). The English word original in its katakana version is the safest word to denote originating documents. Copies, by definition lacking signatures and seals and, frequently, names for the signatories or an indication of seals, are generally bereft of a date of copying and of a name for the copyist. Sometimes traced on paper resting on an earlier text, tōsha ㅞ෗, they look earlier than they actually are. Shahon, copied at the time, or recopied in subsequent generations, survived in a random sequence of passing though multiple hands. For example, a text by Hayashi Jussai, acquired by the Shiryō Hensanjo ྐ ᩱ⦅⧩ᡤ in 1910, is recorded, according to a very cursive note at the back of the shahon, as being first in Ōkōchi bunsho ኱Ἑෆᩥ᭩, before passing into other hands in a trail of later names.7 An acquisition in recent times by the Shiryō Hensanjo, Roshiajin toriatsukai tedome 㟢 すளேྲྀᢅᡭ␃(a major source for Matsudaira Sadanobu ᯇᖹᐃಙand events in 1792), is said to be from the library of the Kuwana Matsudaira ᱓ྡᯇᖹ.8 However, according to the statement on the last page of the third of three copybooks, it was made in 1915 from a book said to have been in the possession in 1821 of Count Matsuura Atsushi ᯇᾆཌ. The location of the work in 1821 and the description of a rank (non-existent in Tokugawa times) might imply that the 1915 copy was made from a more recent copy, which retained the date 1821, rather than from a copy actually made in 1821. A shahon entitled On kakitsuke narabi ni hyōgi dome ᚚ᭩௜୪ホ ㆟␃ is a fascinating illustration of the ambiguous status of individual documents. It is a shahon into which were transcribed documents from 1791 to 1824, but it gains its real value from eight letters written in 1825 by officials on the debate over maintaining the uchiharai policy (of firing on foreign vessels approaching the coast). Held by three successive owners (the third seal that of the Meiji historian Naitō Chisō ෆ⸨⪬ཽ), it came into the possession of Tenri Central Library in 1931. Fujita speculates that the copybook may actually be that of the four bugyō, with whose letters it terminates.9 This he puts forward as “a high possibility,” but there is no way of telling. The opening line Bunsei hachi toridoshi gogatsu itsuka shakuyō ᩥ ᨻඵ㓀ᖺ஬᭶஬᪥೉⏝, untypically explicit for this period, states it is a copy of a borrowed document, hence made self-consciously and perhaps privately. Fujita’s scholarly quest to enlarge the documentary basis for the policy and thought of Matsudaira Sadanobu (1759–1829) on the

238

EARLY JAPANESE TRADE

exclusion of contact with foreigners well illustrates the modest dimensions of the archival base. The relevant texts are all copies, transcribed into documents of widely different archival origin in four scattered locations: Tokyo, Kyoto and two libraries in Hokkaido. (i) Ezochi ikken ikensho sōan ⼎ዀᆅ୍௳ពぢ᭩ⲡ᱌, Kitami Shiritsu Chūō Toshokan ໭ぢᕷ❧୰ኸᅗ᭩㤋, acquired in 1976 from a second hand bookseller.10 (ii) Roshiajin toriatsukai tedome 㟢すளேྲྀᢅᡭ␃, containing the text of Ezochi onsonae ikken ⼎ዀᆅᚚഛ୍௳, now in the Shiryō

Hensanjo, a copy from 1915 of an earlier shahon of 1821 (referred to above).11 (iii) San bugyō hyōgisho ୕ዊ⾜ホ㆟᭩, a shahon transcribing earlier documents, now in Hokkaidoritsu Toshokan ໭ᾏ㐨❧ᅗ᭩㤋, from the papers of the Fukuyama daimyo, Abe Seishō 㜿㒊ṇ⢭, rōjū in 1817–1823. (iv) Ezo byōgi ⼎ዀᘁ㆟, a shahon in two volumes in Kyōto Daigaku Bungakubu ி㒔኱ᏛᩥᏛ㒊, transcribing a metsuke kakidome ┠ ௜᭩␃, containing reports by a metsuke and a gakumonjo official, sent to Matsumae on the arrival in Ezo in 1792 of Laxman. It also contains an account of the rebuttal by Sadanobu, rōjū leader at the time.

The Ishin corpus of archives was diminished by a fire in 1873 of some of its more precious books and papers recently transferred from the Momijiyama Bunko (central shogunal repository) to the fire-prone honmaru, and much later in 1923 in the earthquake which destroyed the library of University of Tokyo, as well as the archives of several government ministries.14 Fortunately, the Shiryō Hensanjo, accommodated today in the University of Tokyo, was then situated at another site, and thus escaped fire or destruction. The papers of the Nagasaki bugyōsho ዊ⾜ᡤ, from the scale of the survivals, illustrate better than any other shogunal corpus the character of Japanese records. They are, to start with, remarkable because of the office’s good fortune in escaping fire after 1663.15 This helps explain why the Nagasaki hankachō ≢ ⛉ᖒ(criminal department records), the greatest surviving run of judicial documents in Japan, is uninterrupted from 1666 to 1867.16 In Osaka, massive losses were partly offset by wholesale circulation of copied versions, the source in later times for the two great modern compilations of Osaka material, the Ōsaka-shi shi ኱㜰ᕷ

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

239

ྐ in 1911–1914,17 and the Ōsaka shōgyō shi shiryō ኱㜰ၟᴗྐ㈨ ᩱ, compiled by the Osaka Chamber of Commerce and Industry ኱㜰ၟᕤ఍㆟ᡤ (OCCI) at the outset of the twentieth century (but

published in 1963–1966).18 The copies on which these two collections drew are unaccounted for: the documents once copied were returned to their owners, and are not identified in the final published compendia.19 The abandonment in 1868 of the Nagasaki records and later protracted recovery of many of the papers of the huge office with near 1,000 personnel underlines dramatically the uncertain fate records faced in 1868. The relative completeness of several categories of surviving records for the two decades preceding 1868 suggests that they may have been retained, by design or accident, in 1868. Other records, however, had a checkered career. Most famously the hankachō, handed over to the police in 1871, were later disposed of to a dealer in old papers.20 They were recovered only in late Taisho years, after an alert from a collector of Tokugawa documents, and the intervention of the city librarian and the prefectural governor.21 Many records passed into private hands including those of booksellers before final rescue by serious collectors.22 A process of recovery began in the combination of two men with scholarly interests, Kanai Toshiyuki 㔠஭ಇ⾜ (1850–1897), from 1886 head of a Nagasaki city ward, and Yakushiji Kumatarō ⸆ᖌᑎ⇃ኴ㑻(1863–1929).23 The records, passing through the ward office (kuyakusho ༊ ᙺ ᡤ ), finally reached the Nagasaki city museum. Kanai was the author of works on Nagasaki institutions, and Yakushiji personally donated 30 satsu of major interest. Post 1915, Koga Jūjirō ྂ㈡༑஧㑻 with high school teacher Mutō Chōzō Ṋ⸨㛗ⶶ, Fukuda Tadaaki ⚟⏣ᛅ᫛㸦a disciple of Koga’s), and Watanabe Kurasuke Ώ㎶ᗜ㍜ were active collectors. Lodged in 1954 and 1964 respectively in the Nagasaki Kenritsu Toshokan, the Koga and Watanabe deposits accounted for 4,000 items. A donation of 17,000 items and 109 satsu, a huge haul of which little is known, had already been made in 1943 by the Fuji ⸨family (the items bearing the stamp Fuji kizō ⸨ᐤ㉗).24 The Koga, Watanabe and Fuji collections amount to 21,000 items, a sizeable segment of the 48,000 items relating to foreign contacts in the Nagasaki Rekishi Bunka Hakubutsukan 㛗ᓮṔྐᩥ໬༤≀㤋.25 Material which found its way to the City Museum and the Kenritsu Toshokan included both items abandoned in the bugyōsho in 1868, and shahon already in private circulation from an earlier date.26

240

EARLY JAPANESE TRADE

The nikki of the Nagasaki interpreters of Chinese, from 1663 to 1715 are a remarkable illustration of uncertain survival. Copied by 㔠஭ಇ⾜ Kanai Toshiyuke, from 1886 head of a Nagasaki city ward , seven satsu continued after his death in 1897 to be held in the ward; the remaining two being held by the ྥ஭Mukai family, by tradition serving as head of the (Chinese ) Confucian Temple. 27 It is not clear why the Mukai held these copies. They may not even have held the original documents from which copies were made . Perhaps transcribed later than the other seven, Kanai may have counted on Mukai negotiating an exchange of the copies for the originals. The likelihood is that at the time individual nikki were in the hands of interpreter families or their descendants , which would help to explain why the originals have failed to survive. Documents still come to light from unsuspected locations such as the two great emaki recording the Rezanov embassy in 1804, of whose origins nothing is known.28 Copies of two documents in relation to Ranald MacDonald, a crew member of an American whaler detained in Nagasaki in 1848, acquired from an unknown source by a collector of old papers, are now deposited in the library of the Literature Department in Kyushu University in Fukuoka.29 The survival of documents in the hands of officials’ descendants is widespread. These were either working copies or the results of later transcribing. A small number are originals. Thus, the voluminous record in twelve satsu of the Rezanov mission by the officials in the Saga బ㈡watch house in Nagasaki, in the family of the han elders, were deposited in modern times in , too, an the Isahaya ㅋ ᪩ public library.30 For the visit of the official Saga account came down in family possession.31 CONTEMPORARY EFFORTS TO KEEP TRACK OF PAPER

Administrative cadres above the level of accounting clerks and paper keepers were few. Given reliance on a relatively fixed tax on land, absence of a system of indirect taxation, and the delicate relations between domains and shogunate, the domains made no contribution to shogunal expenses, occasional arbitrary demands apart. In other words, as a result of a confined revenue base, the shogunate could afford only rudimentary institutions, other than for collecting the rice levy and for audit, justice and ceremonial. In Europe by contrast, by the end of the sixteenth century, war on a grand scale had already necessitated heavy taxation, which radically changed the elementary administration of the monarch living off his own resources into states in which soaring

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

241

administrative costs were defrayed from taxation. European countries in the sixteenth century already had rapidly developing bureaucracies, headed by a council of ministers. While ministers constituted a cabinet or council in some form, they were substantially independent in the conduct of the affairs of their own ministry. Ministers did not necessarily hold office for long periods, but a new minister inherited already established procedures, and permanent subordinate officials. Under their guidance, the order determined for archives was faithfully followed. An obligation to maintain papers, both in ministries and within subordinate bodies, was taken seriously and deficiencies in the state of the papers—inadequate storage or space, poor state of preservation— merited reports, and a call for remedial action. When their duties ended, ministers took away with them much correspondence in which they had engaged, but they respected the integrity of the central administrative corpus of papers: the key documents of decision making. Japan lacked designated archives intended to preserve systematically the paper record of administration. Survival depended on a wide diffusion of copies, which more than matched the attrition suffered over time by individually held paper. The Ikoku nikki ␗ᅜ᪥グwith transcripts of correspondence is the sole source for the foreign policy of the early shogunate. Copies compiled later rested ultimately on transcripts first made by Ishin Sūden, a Gozan monk and adviser to the early shoguns. Sūden does not appear to have retained his originating copies, or at least they had already disappeared by the end of the century. Arai Hakuseki ᪂஭ⓑ▼, reviewing foreign policy, had a further copy made from Sūden’s text in 1713, a copy that only in the 1790s entered the Momijiyama Bunko.32

If retained archives were few, copied documents were the essential corpus in both the archival culture and office management of Tokugawa Japan. Contemporary accounts by foreigners of meetings with Japanese officials refer to secretaries sitting in the background taking notes. Officials made sketches also. This is particularly striking in the case of the mission of the Russian ambassador, Nikolai Rezanov (1764–1807), seeking in 1804 to open trade with Japan. Officers, men, ship, gifts for the shogun were sketched, and detail from them was reproduced in further copies, both contemporary and later.33 These copies provided the source for the depiction of men and dress in the first of the two great emaki or scrolls made in Nagasaki in 1804–1805. The pattern of note taking by officials other than the principals themselves existed into the early 1860s when it was termi-

242

EARLY JAPANESE TRADE

nated in the interests of confidentiality: “It would have been all right if [the note taker] had simply kept notes, but it often happened that he would tell others and spread things secret around, which created problems.”34 This urge to communicate information privately almost certainly refers to written rather than verbal communication, a fact itself accounting for the distinctive lack of identifying features, names or seals, and also for the often whimsical or vague titles of shahon. The metsuke were rather generously provided with clerks and copyists (some 100–200), but in a very loose framework: “Each would [make] use [of them] as he pleased.” Resulting too from the existence of accessible copies, full-blown compendia were put together by officials either serving or retired: two were compiled by men who had seen service in Nagasaki; a third was by an official in a small Kyushu han. The first is in the writings of ኱ ⏣ ༡ ␇ Ōta Nanpo (1749–1823), who before becoming a minor bugyōsho official, collected information on the Russian situation, to which he continued to add after retirement. He described his work ironically as that of shokusan gaishi ⻎ᒣእྐ (amateur or unofficial historian).36 His surviving manuscripts on foreign relations, Enkai ibun ἢᾏ␗⪺and Kaibō kiji ᾏ㜵グ஦, have little on 1804 itself, though as one of the officials overseeing the Russian residence ashore at Umegasaki ᱵ ࢣ ᓮ (a tiny promontory close to Dejima and the Chinese yashiki), he made the acquaintance of Rezanov and developed a respect for him.37 The thirteen volumes (originally fifteen) of Enkai ibun in the National Diet Library (from a private collection) and the less complete runs in four other archives, all lack volume 6, which must have held the data he had assembled for 1804.38 The Nagasaki kokon shūran 㛗ᓮྂ௒㞟ぴ, a very rich quarry of documents, was compiled on retirement in 1790 by Matsuura Tōkei ᯇᾆᮾ῱ (1752–1820), superintendent of the Chinese yashiki.39 Only recently brought to light is a compilation of documents by Yoshida Hidefumi ྜྷ⏣⚽ᩥ(1760–1832), a middle ranking Kurume ஂ␃⡿samurai, covering events in the years 1771–1812.40 It was a private compilation, and the argument that descendants fearfully kept it secret is fanciful. The family later left Kurume for Nagasaki, which explains how it came into the possession of Fujiyama ⸨ᒣprimary school in 1915. The Nagasaki fūsetsugaki 㢼ㄝ᭩(documents with details received from the Dutch on external affairs) are few in Nagasaki, visibly copies of copies, and may even not have been among the documents dispersed in 1868. The most important single source is a volume of 244

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

243

folios for the years 1660–1763 compiled by the Nakayama ୰ ᒣ intepreter family.41 Moreover, the Tsūkō ichiran ㏻⯟୍ぴ (hereafter TKIR), completed in 1853 and 1856, despite its status as a collection of diplomatic documents, contains very few. The preliminary wording preceding a run of fūsetsugaki from 1686 to 1703 suggests they were drawn from a single shahon.42 Texts continue for the Hōei ᐆ Ọ period (1704–1709). At the outset of the section, there is a warning that they are not complete from the Kyōhō period onwards. There are fūsetsugaki for only three Kyōhō years (1717, 1718, 1720). Thereafter, the sole fūsetsugaki are for 1781, 1783, 1805.43 Despite this thin resource, and the alleged secrecy of the reports, copies are scattered across Japan, and the number of known copies continues to grow.44 The Tōsen fūsetsugaki ၈⯪㢼ㄝ᭩ fared somewhat better, in that several compilations, recopied among themselves, are now in the Kokuritsu Kōbunshokan and the Shimabara ᓥཎmunicipal library.45 The great compilations of late Tokugawa and early Meiji times were themselves copied well into the Meiji era. There are eleven versions of the vast Tsūkō ichiran, surviving in whole or in part.46 Another great bureaucratic exercise, two huge series of bakumatsu documents, Tsūshin zenran ㏻ಙ඲ぴ (1864) and Zoku Tsūshin zenran ⥆㏻ಙ඲ ぴ (1879), were themselves recopied at later dates in the early twentieth century.47 For a lesser institution, the reconstitution for a book published in 1912 of the lost or scattered records of the Osaka rice exchange, Ōsaka Dōjima Beikoku Torihikisho ኱㜰ᇽᓥ⡿✐ྲྀᘬᡤ, was based on papers in private possession, not now accounted for.48 A startling detail in the trail of lost records is the story of the Osaka trade return for 1714. The three modern manuscripts rest on a copy made in 1903 for the University of Tokyo library, from the document of either an obscure copyist or collector, which no longer exists; the 1903 copy was destroyed in the 1923 earthquake.49 The story of Tokugawa population figures (whether for Osaka, Edo or Japan) is no different. They rest on often scrappy and incomplete documents, themselves copies of uncertain origin, from earlier sources.50 In many cases, the printed version alone now survives. A striking case is two censuses, the fullest surviving text of any Tokugawa census records, made by a later copyist, whose identity is obscure; the texts also contain easily identifiable transcribing errors.51 Population figures for Tokyo are few, largely from Katsu Kaishū’s Suijinroku ྿ሻ㘓and from Kōda Shigetomo’s ᖾ⏣ᡂ཭work of modern times; for

244

EARLY JAPANESE TRADE

Osaka, none beyond 1862 are cited in Japanese sources, though Mitford, a young British diplomat in 1867, was given the figures for 1866.52 A hikitsugi mokuroku ᘬ⥅┠㘓 (a catalogue prepared by an official for his successor) of 1850 for the Nagasaki Oheyabu ᚚ㒊ᒇ㒊 is the most complete guide to the contents of the Oheyabu of any Tokugawa office. The count is modest—a mere 1407 satsu and 659 fukuro ⿄ (bags).53 The fact that documents before 1780 are scattered through the listed records suggests the Oheyabu papers were an eclectic archive, relating primarily to relatively recent business, and perhaps kept together by bureaucratic inertia, once business was dispatched. The 1850 listing did not include the fūsetsugaki, hankachō, or the shūmon aratamechō ᐀㛛ᨵᖒ(all of which survive in small or large numbers). While it has been suggested that there were other archival stores within the bugyōsho, they appear to have been in the direct possession of officers in the various administrative areas, and were administratively and physically remote from central management of records. That too might explain why the hankachō, by transfer to the police perhaps simply by decision of judicial officers in default of general action, escaped the chaos which affected most of the records in 1868. The absence of fūsetsugaki arises from the fact that the records were held by the Dutch interpreters. Fūsetsugaki in any number in a single location (together with documents in the original Dutch or in translation) survive only in two family collections. One is the huge Motoki collection in two separate locations.54 The other is the Nakayama collection, now in Shīboruto Kinenkan ࢩ࣮࣎ࣝࢺグᛕ㤋 (SMH).55 The Dutch and Chinese interpreters maintained their own series of annual nikki. Nikki survive for only two years for the Dutch, in contrast to the more substantial survival of Chinese.56 Much less is known of papers in offices in Edo itself. A surviving summary count for 1723 of records in the Kanjōsho ຺ᐃ᭩, the central accounting office, which had its store at Takebashi ➉ᶫ, was 94,200 satsu. But scarcely anything is known of the archives of the Hyōjōsho ホᐃᡤ, with its legal and judicial functions and its own store at Ōtemon ኱ᡭ㛛. The picture is blank also for the two offices of the machi bugyō ⏫ዊ⾜with responsibility for civic administration, and the jisha bugyō ᑎ♫ዊ⾜ , in charge of the huge temple and shrine domains in Edo and beyond. Ogyū Sorai Ⲷ⏕ᚂᚙ(1666–1728), neoConfucian scholar and critic of public issues in the 1720s, with the Hyōjōsho no doubt in mind, was highly critical of verbal precedents and failure to keep records:

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

245

No office should fail to keep records of business. At present, it is the general practice to deal with business on the basis of precedents committed to memory. It is entirely due to the lack of records of business that the officials are vague and ignorant of the duties of their offices.58

In Meiji, a man who had served both as metsuke and machi bugyō in bakumatsu times recalled that rules had earlier been less detailed than they later became.59 It seems in general true that the higher the matter, the more informal or casual the handling of the paperwork.60 Two repositories, however, had some role for papers from the central administration: the Momijiyama Library within Edo Castle and the Hayashi Library Shōheikō ᫀᖹᰯ̿located from 1690 at the Confucian temple in Yushima ‫‮‬ ᓥ— from 1797 known as the Shōheizaka Gakumonjo ᫀᖹᆏᏛ㛛 ᡤ. Both the Momijiyama Archive and the Hayashi Library existed from the time of Tokugawa Ieyasu. The former became more directly identified with the shogunate, while the latter retained the character of an independent service provider.61 Momijiyama acted as a repository for records of the goyōbeya ᚚ⏝㒊ᒇthough as far as one can judge not as a general administrative archive. It was primarily a library of books in Japanese and Chinese, and its holdings were limited. Of 100,000 satsu in a count for the early Meiji period, 73,000 were texts in classical Chinese, only 26,000 in Japanese. Of the latter, moreover, half were in history.62 In addition, according to Fukui, it tended to avoid deposits other than Japanese woodblock books, and held only a small number of valuable items.63 However, it had held early Tokugawa records, sensitive documents relating to Korea and to the Ryukyus, and the texts of treaties (though apparently not the Tokugawa nikki). Thus, to take two examples, the Ikoku nikki, preserved in the Hakuseki Shozōsho ⓑ▼ᡤ ⶶᡤentered it in the Kansei period; in 1858, the TKIR, soon after its completion, was lodged there. Most of the documents it held are now unaccounted for.64 Its confined archival role is well illustrated in the fact that rōjū working papers did not enter it. The shahon made for or by bugyō and metsuke, and constituting the mass of paper in the higher reaches of shogunal administration, entered it haphazardly if at all. The papers of Kondō Jūzō ㏆⸨㔜ⶶ(1771–1829), for instance, highly sensitive because of the still unsettled frontier with Russia, were held not at one point but at several centers.65 Remarkably obscure in the eighteenth century, the profile of the Hayashi family grew in the early nineteenth century. Its change of for-

246

EARLY JAPANESE TRADE

tune was due to the reforms in the 1790s, which turned it into an official academy. As a result of the urgent need to understand novel foreign problems from the 1790s onwards, its scholarly role expanded and its library grew.66 Its head for the early nineteenth century, Hayashi Jussai, enjoyed enormous prestige, and ranked on a par with kanjō bugyō for official consultation. By 1825, in the library there were more than 1300 bu and 7000 kan.67 Its catalogues also make it possible to estimate the areas of new concern. Among forty three bu on more current matters, there were six bu on Russian relations, six on foreign relations from early to recent times, ten on Chinese vessels, and significantly in a further seven bu dealing with the arrival of foreign ships, the Phaeton Incident held the central place.68 Books in the Momijiyama Bunko provided the nucleus of the Imperial Library and of the later Ueno Library (subsequently Kokuritsu Kokkai Toshokan ᅜ❧ᅜ఍ᅗ᭩㤋). While in the Restoration foreign books are said to have been transferred to Momijiyama from the Bansho Shirabesho ⶽ᭩ㄪᡤ (founded in 1856 as successor to the banshowage goyōgakari ⻅᭩࿴ゎᚚ⏝᥃of 1811), the story is more complicated.69 Some apparently passed into Tokugawa hands, and were taken to Shizuoka, where Yoshinobu, the last shogun, lived in retirement. That explains how at a later date they finally came to rest in the Shizuoka Kenritsu Toshokan 㟼ᒸ┴❧ᅗ᭩㤋.70 THE RECORDS OF THE DOMAINS (HAN)

Domain archives covered first and foremost the records of formal correspondence with the shogunate and other daimyo and a vast array of accounting records. The latter (some relating to samurai; others to tax details) to a degree furnished information on the size and composition of population. The archives of 17 domains have been examined in detail by Kasaya Kazuhiko ➟㇂࿴ẚྂ.71 Accounting categories apart, the formal correspondence with shogun and rōjū, in essence the records defining the legal relationship between individual domain and the shogunate, was rich and well preserved. The formal correspondence had in effect the status of legal documents, some of it the counterpart to the now lost shogunal records, defining the very existence of the domains in legal terms. Some categories were retained in the original, others discarded after detail was entered in nikki (diaries) or nisshi ᪥ ㄅ (registers). The shūmon aratamechō were regularly discarded after a short interval (since they were voluminous), but summary details

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

247

were entered into nisshi.72 Where actual copies survived, it was mostly in the hands of descendants of village headmen. The story of population returns for the domains for the six-year shogunal censuses is similar: the returns were not retained, but summary figures were, as in Okayama, entered in domain registers.73 The sheer number of domain ensured that if misfortune befell the records of some, others enjoyed a kinder fate. Many years ago, John Whitney Hall conducted a seminal study of the archival records of the Ikeda ụ ⏣ family (daimyo of Okayama domain), which had been presented to Okayama University in 1949.74 More recently Kasaya Kazuhiko has shown that a large number of domains retained a substantial corpus of records. Vicissitudes experienced in their history mean that they have been held in several locations, whether daimyo family descendants, city or prefectural archives, and on occasion Tokyo archives such as the Monbushō itself. Domain archives are often impressive in scale. Okayama has 60,000 satsu, the small domain of Matsushiro ᯇ ௦ no less than 30,000. Equally impressive is the fact that individual series have long uninterrupted runs: Okayama from 1648, Hirosaki ᘯ๓from 1661, Morioka ┒ᒸfrom 1664, Matsushiro from 1667 and 1686, and Fukui from 1686.75 As noted by Hall for the Okayama records, “In nearly every case the mobile papers were copied in whole or in part in record books of one or another of the domain officials. These record books, variously called chō, tome and nikki, are without question the most important of the operational records for the historian.”76 Retention in the form of data entry into registers was true even at the level of the papers of the office of daimyo and of the domain Council. These registers could be voluminous. Thus, the otemoto nikki ᚚᡭチ᪥グcontains a detailed record of the daimyo’s private and public activities with 780 volumes covering the years 1663–1875.77 For the Council “only a fraction of this paper work has been retained in its raw form.”78 In addition, a summary record of domain activities at large was made for each year, with 200 volumes covering the period from 1673 to 1894, “compilation having gone on even after the abolition of the domains.”79 There was, according to Hall, for each daimyo a collection of correspondence. He does not expand on this, though Kasaya gives a brief listing of the categories.80 Han records have two striking characteristics: a widespread nonretention of originals and an ongoing process of copying detail from documents not retained into nikki and nisshi. Hall’s study of Okayama is the story of domains at large. Among the domain, Tsushima ᑐ㤿

248

EARLY JAPANESE TRADE

holds a special position partly because it was charged by the shogunate with the supervision of relations with Korea; partly because survivals are substantial, originating in three administrative centers, Fuzhou ⚟ ᕞ, Pusan and Edo, and in six actual locations.81 Some were held until recent times in the store house of the Sō ᐀ family temple; others are in Korea. While the records illustrate the fragmentation and losses that are commonplace, they also provide evidence not available elsewhere. Thus, a letter from a shogunal minister in Edo to the Tsushima daimyo clarifying foreign policy exists only in uncatalogued papers in Seoul.82 Domain records in the main do not have extensive informal or political correspondence. However, domains, whose daimyo had served as rōjū, took their papers with them at the end of their period of service. In addition, the politically conscious daimyo of Mito (a sanke house) and Satsuma (a tozama daimyo connected by marriage to Ienari ᐙᩧ, shogun from 1787 to 1837), compensated for their exclusion from office under Tokugawa convention by assertive policies: the result was a high degree of order in their policy papers and correspondence. In the dearth in Edo of records of high policy, Hagiwara Yutaka ⴗཎ⿱, archivist of the Gaimushō in the 1880s , drew on domain records to create a more complete record of foreign affairs: the family papers of six daimyo families who had provided rōjū (including the heirs of Mizuno Tadakuni Ỉ㔝ᛅ㑥, Abe Masahiro 㜿 㒊 ṇ ᘯ , Hotta Masayoshi ᇼ⏣ṇ╬and Ii Naosuke ஭ఀ┤ᙌ), the archives of the Owari ᑿᙇ Tokugawa house, the papers of the Shimazu family of Satsuma, and of twenty-five other daimyo houses. This work yielded correspondence with metsuke, kanjō bugyō, gaikoku bugyō and also with the bugyō for Hakodate, Uraga, Shimoda, Nagasaki.83 Kasaya’s survey has noted some evidence of political correspondence in domain records, but survivals appear to constitute isolated sources rather than long runs. For Fukui domain, he noted Mito ke kankei shorui Ỉᡞᐙ㛵ಀ᭩㢮㸹for Hikone, Bakumatsu kaibō kankei ikken shiryō ᖥᮎᾏ㜵㛵ಀ୍௳ྐᩱ and fūbungaki 㢼⪺᭩; for Okayama, Bakumatsu no shoka raikan ᖥᮎ ࡢㅖᐙ᮶⩶㸹and for Uwajima, a large amount of incoming correspondence in the 1850s (more than 1000 items from Tokugawa Nariaki and Shimazu Nariakira, daimyo of Mito and Satsuma respectively).84 The range of surviving informal correspondence in domain records appears to have varied from little or none to substantial amounts. In the very extensive archives of the Sanada family of Matsushiro, the amount is small. However, more than 500 satsu of nikki recording correspondence with the Edo karō, from 1684 to 1871, illustrate an

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

249

active connection.85 There are also papers relating to the exercise of the office of rōjū by Sanada Yukitsura ┿⏣ᖾ㈏ (1791–1852), the one daimyo of the domain to serve in that role.86 Otherwise, the only trace of political interest lies in eleven tracts on Ezo affairs and five for the years 1854–1858.87 Domain archives, in contrast to shogunal archives, include holograph correspondence. There is a dilemma in explaining how, amid good record keeping, “political” correspondence is often fragmented or confined to bakumatsu records. Given the practice of daimyo to store records in fireproof kura, the storage of these records can hardly have differed from domain records at large. The contrast with cases such as Mito and Satsuma suggests that in less politically aware domains, this category of record may have been isolated from the formal administrative records, and a pattern of benign neglect over time, or some degree of privacy, may have put their survival at risk. Mito is an outstanding case of large-scale survival of records, many of them papers in Nariaki’s own hand. The Mito archives, held in a library run by family descendants, have a rich array of documents. As drawn on for the Dai Nihon ishin shiryō ኱᪥ᮏ⥔᪂ྐᩱ(hereafter DNIS), they include the Mito daimyo proposal in 1854 for building large vessels and in 1858 papers on defence.88 The latter source includes a letter of Nariaki to Hotta in 1858, the year of his resignation, which in turn was copied into two Hotta records.89 Easy access by archivists at a relatively early date to the other great collection, the Shimazu records, is impressive, resulting from the interest of Shimazu Hisamitsu ᓥὠஂග (1817–1887), regent to the last Satsuma daimyo, in vindicating Satsuma’s past political stance.90 This culminated in the year after his death in the founding of the Shidankai ྐ ㄯ ఍ intended to ensure a fair account of domain policy of bakumatsu times.91 As later catalogued when the records came into the possession of the Shiryō Hensanjo, they consist of 11,700 items in general papers, and 6,500 shahon. From 1720 to 1868, there are 1,732 letters of one or more sheets, not texts transferred into shahon.92 Shimazu records in the DNIS, may be summarized under three headings, correspondence of named daimyo, papers dealing with the Ryukyus (for Satsuma a very sensitive issue); and documents from successive karō of the Edo yashiki of the domains under the style of Kagoshima han Edozume karō jō 㮵ඣᓥ⸬Ụᡞワᐙ⪁≧. Mito and Satsuma apart, other collections held by immediate heirs of daimyo or their descendants in Meiji were important for the story

250

EARLY JAPANESE TRADE

of Japanese politics as recorded in DNIS. An example is the archives of the post 1868 head of the Kuroda family of Fukuoka, Kuroda Nagashige 㯮⏣㛗ᡂ. They were the source of Edo goyōjō Ụᡞᚚ⏝≧ from which the Fukuoka han zaifu karō shokan ⚟ᒸ⸬ᅾᗓᐙ⪁᭩⩶in 1858 was reproduced. Fukui domain was one of the great dissident domains of bakumatsu times. Its records include correspondence with Mito.94 The Gōdō haku nyūsō hiki ྜྠ⯧ධ┦⛎グcontains Fukui Matsudaira Yoshinaga shuki ⚟஭ᯇᖹ៞Ọᡭグwith correspondence on daimyo views in 1854 on foreign vessels.95 FOREIGN POLICY AS SEEN IN THE RECORDS

Except for bakumatsu times, when han records are often the source, policy in earlier times is elusive. There is a contrast for example between the comparatively well documented Phaeton Incident of 1808, and the thin record from all sources for other events. The interpretation of thought or policy poses problems, given the small amount of documentation of decision-making individually and even more collectively. While the immediate purposes of foreign exclusion are clearly set out in the 1630s, doubt has been cast on its intended permanence. It has been suggested that the English monarch’s marriage to a Portuguese princess accounted for the rejection of an English démarche in 1673. Access would have been open in later decades had the English sought it. On this telling, radical exclusion for all save the existing Dutch and Chinese dates only from the 1790s.96 Analysis of the source basis may be more helpful in our moving to conclusions than somewhat speculative deductions based on slight evidence. The Japanese response on the occasion of the visit of the English vessel the Return in 1673 seemed to confirm the exclusion policy, and frequent later copying of the statement is consonant with an unchanged policy in the following century. In another instance, the sweeping uchiharai decree of 1825 may have been less a break in policy than a response to a new situation, arising from the helplessness revealed by the Phaeton visit to Nagasaki in 1808, and fear in the early 1820s of a recurrence. It does not appear to relate to the older problem of the Russian presence in the Ezo islands. In regard to early seventeenth century foreign trade, TKIR has little information, Dutch and Portuguese trades apart. It noted that an earlier account, which the TKIR cited (Koshūki ྂ㞟グ, kan 253, p. 354) lacked a list of vessels under license. Observing the want of

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

251

records, it cites a document from 1614 with details of licences to Dutch vessels. The basic source remained the frequently copied Ikoku nikki. Relations with Spain or England for these years are obscure: the best documentation for the early English trade lies in letters by Englishmen who were servants of the English East India Company. The visit of the Return in 1673 is poorly documented in the TKIR, a single kan drawing on eight shahon and a Tokugawa nikki.97 Two modern studies provide a wider appraisal.98 Both were aware of the account in the Kokon shūran compiled by Matsuura, and reproduced its text. Machin’s account provided three other texts on the visit, and details of two lesser recordings. There is, moreover, a little known document in the Nagasaki University Economics department library on the departure of the Return.99 Meagre though the surviving sources are, reporting over the next century has significance beyond its purely archival dimension. As the account not only documented the English request for trade but also made clear in convoluted fashion that the question would not be reopened on the death of the Portuguese consort of Charles 11, it amounted in fact to a confirmation of a sakoku or “closed country” policy.100 For the visit of the English warship the Phaeton to Nagasaki in 1808, the Hayashi devoted five kan to the topic, citing nine sources.101 Moreover, they referred to a number of family records and letters from Nagasaki, none of which proved to add anything new.102 Untypically, Nagasaki is relatively rich in surviving documents of the episode. These include two long documents relating to the visit of the reforming bugyō dispatched to Nagasaki immediately after the debacle, the second one declared by Katagiri to be of fundamental importance.103 There is also a document in the Nagasaki University Economics Department Library, descriptive of the events, starting with a letter of Henrik Doeff. This document, prepared for the bugyō, also provides a rare case of a name on the title page, Kure Tokutarō ࿋ᚨኴ㑻, presumably a copyist acting for the new bugyō.In 1972, in a list of seven papers Katagiri had three bearing directly on the episode (including one from Kyushu Daigaku).105 He also edited the fullest of these, primarily on the new defense arrangements for Nagasaki bay.106 In recent times, eleven documents relating to the Phaeton (apparently originating with Nagasaki bugyōsho officials, interpreters, Kurume officials and private individuals in Nagasaki, or in communication with the Saga authorities) have come to light in a compendium by a Kurume official: the last two, Ara’ara oboegaki ⢒ࠎ ぬ᭩and Nagasaki yadai zokuzokuhō 㛗ᓮᒇ௦⥆ࠎሗ, are relatively long and detailed.107

252

EARLY JAPANESE TRADE

A document used briefly in the TKIR is the Nagasaki ontsukai shoyōbeya nikki 㛗ᓮᚚ౑ᡤ⏝㒊ᒇ᪥グ from the Nagasaki goyōbeya, a text which would appear to correspond to a “copy of the official diary kept in the Government House at Nagasaki,” purchased in the 1870s by Aston in Tokyo. The compiler’s name is Tokuemon according to Aston, or in Japanese sources ᚨᕥ⾨㛛 Tokuzaemon.It covers the three days of the Phaeton’s visit, highlighting indecision and ineffectiveness. It thus provides a direct account of the debacle, which led to the urgent dispatch of Magaribuchi Kai no kami to Nagasaki with a brief to reorder the defenses. This task, recorded largely by daikan ௦ ᐁTakagi Sakuzaemon 㧗ᮌసᕥ⾨㛛resulted in the new arrangements which were, in Katagiri’s words, a radical reform of defense in Nagasaki.109 In consultation with the Dutch, tighter procedures were introduced to reduce the risk of a vessel arriving under false colors. Direct supervision by the bakufu replaced complacent reliance on domain responsibility for the overview of the defense arrangements, so inadequate in 1808. Basic decision-making was now firmly in the hands of the bugyō and his daikan. The presence of English whalers on the coasts in the early 1820s raised questions as to the deeper significance of their appearance: the earlier 1808 incident is implied in the wording of the uchi harai decree in 1825. If so, debate over the significance of the whalers arises from the problem of distinguishing between whaling visits in general and more suspect sightings. These might imply exploration by foreigners (in these years exclusively English) in preparation for visits by hostile warships. The presence of whalers off Uraga and off Mito, south and north respectively of Edo Bay, underlined the vulnerability to attack of Edo and its vital supply routes. In other words, there was a material problem of real strategic significance in the mid-1820s. A lack of worry about a Russian menace is confirmed in the fact that Edo authorities restored the Matsumae family to their domain in 1822, leaving defense in their hands. The archival detail on 1808 is conspicuous. By contrast, for the Rezanov incident 1804, there are sixteen miscellaneous shahon (in kan nos. 275–83). The Hayashi account drew heavily on Nagasaki shi zokuhen 㛗ᓮᚿ⥆⦅, the second part of a detailed and well informed compendium (less likely than the first part to have been complied on bugyō order).110 Reliance on the aforementioned Kankai ibun for Russian relations as a source is surprising.111 While it has information on 1804, and illustrations drawn from sketches made at the time, it is

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

253

entirely derivative. Its use is evidence of the poor information in Edo in the 1850s on the Russian visit. With a total of nineteen kan (nos. 297–315), the information for the protracted Golownin Crisis (1811–1813) seems superficially very rich. However, the major single source is Sōyaku Nihon kiji 㐼གྷ᪥ᮏ ⣖஦ (Golownin’s narrative), used extensively in fourteen of the nineteen kan.112 In the closing lines in the first kan dealing with the incident, Golownin’s journal is announced with the bald statement that in conjunction with Japanese sources it would be used in the rest of the account.113 It was reproduced virtually in its totality. The eighteen remaining sources, all shahon, were highly miscellaneous, and only Ezochi hikki ⼎ዀᆅ➹グand Seihokuroku 㟹໭㘓were cited frequently. The sources at large are thus uninformative on high policy. There is very brief recognition of the visit from Edo of Takahashi Kageyasu 㧗 ᶫ ᬒಖ on a fact finding mission, intended to inform decisionmaking in Edo. He had been accompanied by Baba Sajūrō 㤿ሙబ ༑㑻 who remained in Ezo (later translating the Sōyaku Nihon kiji). A mere two kan (nos. 313–14) contain texts in which Baba, or Murakami Teisuke ᮧୖ㈆ຓ(so prominent in Golownin’s account) feature among the signatories. The resolution of the crisis is not known in documents in Japan. Only in visits in recent times to the archives in Petrograd have researchers from the Shiryō Hensanjo obtained a copy of the formal reply supplied to the Russians, and which terminated the affair.115 How the Hayashi accessed information is not clear. The extent to which they relied on the Momijiyama Bunko, on their own records or on shahon accessible in the circle around them, cannot be determined. Hayashi Akira ᯘ㡐and his team consulted the Momijiyama records. They edited a supplement Jūtei goshoseki rairekishi 㔜ゞᚚ᭩⡠᮶ Ṕᚿto the catalogue Jūtei goshoseki mokuroku 㔜ゞᚚ᭩⡠┠㘓, compiled between 1814 and 1836.116 However, given the limited range of material in the repository, they must have relied extensively on other sources. Thus while the Nagasaki shi zokuhen could well have been in the repository, for events in the 1830s they also consulted in the 1850s, a copy of that work possessed by the Togawa ᡞᕝfamily. As there is no Togawa material in the surviving manuscript versions of the Nagasaki shi zokuhen, Togawa Yasuzumi ᡞᕝᏳΎ (1787–1868) must have later transcribed into a personal copy the records of the delicate mission.117 Hayashi Akira’s access to material is particularly interesting for 1854, the year of Perry’s second visit, because TKIR was compiled by

254

EARLY JAPANESE TRADE

1856 and on the Japanese side he had presided at the negotiations with the Americans. Some of the approximately fifty shahon the Hayashi drew on for their account of 1854 are obviously substantial items across the entire year; others were quoted more rarely, and the fact that as many as three shahon are sometimes cited together as a source suggest that there is much overlap. Hayashi’s sources support the impression that at the time a great many shahon circulated, and that there was no central controlling record. Rich cases like that of Togawa for the 1830s apart, the absence of sustained correspondence of bugyō or rōjū leaves the central issues obscure. There is evidence neither of the reasons for delay in 1804 in communicating decisions to Nagasaki, nor of the tortuous course of decision making in Edo. The same problem arises in the case of Golownin’s captivity. Later in 1821–1822, remarkable though the reversal of policy was, there is no explicit evidence of the reasons behind the return of Matsumae in 1822. Until something is known of the advice on uchi harai given to rōjū in 1825, the reasons that influenced rōjū in their final decisions can only be guessed at. RŌJŪ, SENIOR BUREAUCRATS AND THEIR PAPER WORK

In the Edo period, the structure of Japanese administration had evolved little beyond the tradition of rulers living off their own estates. At the apex, the shogunal administration depended on the rōjū, selected from fudai daimyo, serving in monthly rotation. They had neither a personal core of officials answerable to them individually, nor a fortiori the tangible features of separate buildings (ministries). The Hyōjōsho apart, narrow specialization was not envisaged. Problems as they arose were handled by ad hoc redeployment. This explains two striking features of Japanese bureaucracy: first, how officials were detached to carry out specific tasks and how several of the officials could be combined to make very delicate reports; second, the smallness of the staff resources explains also why officers of low rank on occasion exercised high responsibility. Kondō Jūzō, though holding the rank of shihai kanjō ᨭ㓄຺ᐃ, was made responsible for surveying Ezo after a short period of service in Nagasaki.118 Daikan on occasion also found themselves in a defense role, as was famously the case with Egawa Tarōzaemon Ụᕝ ኴ㑻ᕥ⾨㛛(1801–1855).119 The kanjō bugyō themselves, no more than the rōjū, commanded workers who served permanently under them.120 Officials at large were

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

255

primarily low grade, underworked civil servants employed to collect revenue, supervising its expenditure, and as clerks maintaining the paperwork of these operations. In the Kanjōsho, the largest administrative arm with a total staff of about 700, senior figures were few. The number of kanjo bugyō was about eight, supplemented by two kanjō ginmiyaku ຺ᐃྫྷ࿡ᙺ and ten to fourteen kumigashira.121 The kanjō bugyō were sometimes supplemented for special tasks by Hyōjōsho metsuke (who numbered sixteen to twenty).122 Rōjū and officials worked in close proximity to one another. The goyōbeya was a complex of rooms rather than literally a central workspace. The wakadoshiyori room was in the central position within the goyōbeya. On one side were the offices of the rōjū and beside them the kanjōsho, bugyō point of contact with the rōjū; on the other side were offices of the kanjō bugyō, machi bugyō, jisha bugyō and metsuke. There were also several spaces known as osoba shū ᚚഃ⾗. The metsuke with no space allocation, individual or collective, may have met there on occasion or in the wakadoshiyori beya. For critical problems, the rōjū came to seek reports from the “three bugyō,” in effect groups for individual cases made up of one or more officials from kanjō bugyō, machi bugyō and jisha bugyō. The involvement of several officials simultaneously added to the necessity of multiplying copies of relevant documents for briefing purposes. By the 1860s, growing pressures on officials’ time and attention meant they gradually acquired more autonomy. The metsuke, investigating officers exercising distinctive duties, supplemented the kanjō bugyō in general concerns of the shogunate. In very delicate negotiations (with foreigners), metsuke were automatically present. With the creation of the new office of gaikoku bugyō, the metsuke ceased to appear at negotiations in the 1860s.123 The informal correspondence of rōjū (often with powerful daimyo) and of kanjō bugyō on non routine issues remained personal to the office holder. In other words papers, personal to the man, followed him. Toby in his pioneering study of the shogunate’s Korean diplomacy noted the problems caused by the fact that “the rōjū did not keep records of its deliberations,” and that “the only evidence we have is inferential.”124 For Abe Seishō, Fukuyama daimyo and rōjū from 1817 to 1823, surviving papers on the Russian question suggest how a rōjū was briefed.125 As the Nagasaki bugyō spent every second year in Edo, the personal nature of their papers was if anything further enhanced. In their year in Edo, they effectively played for Nagasaki affairs the role of ambassador: they had long and usually cordial meetings with the Dutch dur-

256

EARLY JAPANESE TRADE

ing their regular Edo visits. The record of Izawa Masayoshi, formerly bugyō in Nagasaki (17 April 1842–8 March 1845) later a metsuke and an Edo machi bugyō, suggests the fate of documents. On his final departure from Nagasaki, he took with him letters from Edo he had received during his period of office. In 1891, Yamaguchi Naoki ᒣཱྀ ┤Ẏ(1830–1895), a retired official and an acquaintance of Izawa, recollected that they had included letters in the hand of Mizuno Tadakuni on important affairs of state: “He (Mizuno) sent him (Izawa) detailed instructions in his own hand. There was a bamboo basket full of these letters. Izawa showed them [to me].”126 The contrasting pattern in the survival of records of Abe and Hotta, the senior rōjū between 1845 and 1857 shows how survival can vary. For Abe Masahiro, a rōjū of exceptional interest because he headed the administration from 1845 to 1853, the major survival is in Mito, in copies of correspondence made, it is said, by Tokugawa Nariaki in his own hand, into a series of five shahon, the Shin Ise monogatari ᪂ఀໃ≀ㄒ. It contains the correspondence between Nariaki and Abe. It also has correspondence from Shimazu Nariakira. With Abe’s correspondence missing in Fukuyama, a copy was made in eight satsu for the family from the Mito record.127 It now rests in the Seishikan Kinenkan ㄔஅ㤋グᛕ㤋in Fukuyama.128 By contrast, documents of Hotta Masayoshi enjoyed a happier fate. In early Meiji times, they were held by the family. The archives of Hotta’s son and heir, Hotta Masatomo ᇼ⏣ṇ೔, held Masayoshi’s diary. The domain archives also held items such as Sakura han tsuchi no e shū బ಴⸬ᠾᖺ㞟. Some copies were lodged later in the National Archives. One shahon, described as a Masatomo family record, Hotta Masatomo kaki ᇼ⏣ṇ೔ᐙグ, originally given to the history section of the Council of State (Dajōkan Sei’in Rekishika ኴ ᨻᐁṇ㝔Ṕྐㄢ), later passed to the Naikaku Bunko.129 Another, described as Hotta Masayoshi kiroku ᇼ⏣ṇ╬グ㘓, under the title Jōyaku shokan wage ᲄ ⣙᭩⩶࿴ゎ contains a very substantial twelve kan dealing with the American negotiations, Amerika shisetsu taiwasho ள⡿฼ຍ౑⠇ᑐヰ᭩is also in the Naikaku Bunko.130 Lacking an indication in the catalogue of a source, it may have reached the archives in a similar relay. NEW ADMINISTRATIVE STRUCTURES FROM 1858 AND NEW ORDER IN RECORD KEEPING

The Hayashi account for 1854 is the first occasion for which there is a glimpse of the Japanese archival working record in detail. Admin-

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

257

istrative changes were sweeping from 1858 with a radical alteration of record keeping as a counterpart to the changes. The pace was set by the gaikoku bugyō, in charge of foreign affairs and also of defense matters, resulting in the biggest administrative change since the time of Ieyasu.131 In embryonic form, the posts were thus the first step not simply in creating a modern foreign office but Western style ministries in general. The collective action of three grades of bugyō no longer existed. The role of metsuke also changed. They had regularly met as a group, and added their seals to documents. Informal consultation Ozashiki hyōgi ᚚᗙᩜホ㆟of metsuke and bugyō also took place with the parties meeting in a room that happened to be empty.132 The older pattern of bugyō consulting metsuke thus continued. So entrenched was it that even gaikoku bugyō at first worked under this constraint.133 But pressure of events meant that metsuke participation in foreign affairs declined. Everyone was busy, and those in charge of foreign affairs couldn’t look into other matters, because they had to travel back and forth to the capital and Kobe and Osaka to be present at receptions of the bugyō. They had hardly any spare time so general matters were left untouched. But at the time I [Yamaguchi Naoki] started as metsuke we did do such things.134

Metsuke attendance at high level meetings with foreigners ceased two or three years before the Restoration in 1868.135 There was concurrently a pressing two fold need: to recover copies of missing documents, and to keep order among current papers for the daily exchange of correspondence with diplomats in Edo, Kanagawa, Nagasaki and Hakodate. Much is obscure about the loss and recovery of paper. As early as 1864, the gaikoku bugyō had noted the need for order, and in 1865 efforts started to recover copies of missing documents.136 As copies circulated, it appears that what was everybody’s business had been nobody’s business, and that the occurrence of fires and the demand for copies has led to overstatement of actual loss. The diffusion of papers was the real problem, though equally it made possible a highly successful recovery operation. The Hayashi were still relevant in the mid 1860s. At the end of 1864, fūsetsugaki and other documents were transferred to them.137 But their specific proposals were rejected in 1867.138 The Hayashi had a small staff; they were also cast more in the scholarly mode of storage and

258

EARLY JAPANESE TRADE

study than of actively dealing with current issues. Radical new thinking was already emerging. A Gaikoku Goyōjo እᅜᚚ⏝ᡤ, a proper office for the gaikoku bugyō, was established where monthly meetings took place.139 In 1867, the Shomotsu goyōshutsu yaku ᭩≀ᚚ⏝ฟᙺ with a small staff, completed copying of records for the years 1859–1863. In total, the compilation contained 320 satsu, stored in six boxes.140 While the master copy seihon ṇ ᮏ was placed in Momijiyama, another copy was lodged at Shōheizaka; a further one was retained by active officials in the field.141 Efforts to continue the process for the subsequent years collapsed, but were later undertaken between 1871 and 1879. They resulted in the Zoku Tsūshin zenran ⥆㏻ಙ඲ぴin 505 kan containing 1,869 satsu.142 We do not know from whom the compilers uncovered the documents or how the originals were disposed of. The overall exercise amounted to 200,000 manuscript pages.143 This huge collection was published in printed form only in 1989.144 The combined Tsūshin zenran (first and second series) is, if measured in terms of the modern printed page, four times the scale of the TKIR. THE POST-1868 TRANSFER OF TOKUGAWA ARCHIVES TO THE NEW REGIME

The state of affairs in Edo in 1868 reveals that the Dajōkan ኴ ᨻ ᐁ had no archival policy. The fact that in 1869 the Shōheizaka and the Momijiyama Bunko were envisaged as the nuclei of universities shows that two years after the Restoration archives were still not a foremost concern. While for Momijiyama the university idea was short-lived, its management under five successive sub-agencies of the Dajōkan, was at least close to the centre of government. Real clarity came only when a Dajōkan Bunkokan ኴᨻᐁᩥᗜ㤋was decreed in 1884 with responsibility for archives.145 For Shōheizaka, the idea of a role as university endured till 1875. Lacking the protective mantle of the Dajōkan, changes had been less reassuring: it passed under the tutelage of the Monbushō, then two years later came briefly under the scrutiny of the Hakurankai Jimukyoku ༤ぴ఍஦ົᒁ, before being transferred in 1876 to the Naimushō. Though in early Meiji years new government institutions assumed responsibility for Tokugawa bodies, their discharge of that role was far from clear or reassuring. They took from the two existing archives records relating to their area of interest. The Gaikoku Jimukyoku እ ᅜ஦ົᒁ, the last of several names for a new agency for foreign affairs preceding the Gaimushō, received records from the Tokugawa family

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

259

in 1869.146 In particular, papers relating to Korea and the Ryukyus were moved from Momijiyama to the Naimushō; also to the Kunaichō. The Tokugawa family took over books, but also some manuscripts, including a copy of the Japanese translation of Golownin’s Narrative, which is not included in lists of known copies.147 The Hayashi collections were dismembered, and went to the new ministries. Seals or internal evidence in records later deposited in 1885 in the Naikaku Bunko often reveal Hayashi origins.148 The 1885 deposits were substantial but selective, the Naimushō continuing to regard Korea and the Ryukyus as sensitive, even retaining some pre-1868 material. Given the vulnerability of the Ryukyus to foreign incursion from the 1840s until formal annexation in 1879, many records, even pre-1868 ones, were transferred to the Kunaichō and Naimushō. An example from the Naimushō cited on a number of occasions in DNIS is Izena Pechin unjō shokan ఀ᫝ྡぶ㞼ୖ᭩⩶ for 1846 from Ryūkyū Ōkoku Hyōjōsho nikki ⌰⌫⋤ᅜホᐃᡤ᪥グ.149 The Gaimushō deposited mostly older records or reports, often duplicates. Scarcely any correspondence featured in its deposits. Storage added to the hazards of early survival. In 1875, the Naimushō opened a converted rice warehouse for Shōheizaka records in Asakusa ὸⲡ. In 1884, they were moved to a site at the Wadakura mon ࿴ ⏣ ಴ 㛛 ; later, they were transferred to a repository in Ōtemon. Momijiyama records at this stage held, within a former bookstore in the precincts of the imperial palace in 1891, also were transferred to Wadakura mon.150 The archives of other institutions were less fortunate. The Hyōjōsho records first went at least nominally to the Shihōshō ྖἲ┬; then, for a time after the creation of the new archive in 1884 to the Dajōkan Bunko; in 1895, to the Law Department of Tokyo Imperial University and, in 1904, to the Imperial University library.151 The jisha bugyō records were transferred successively to the Naimushō, and the Imperial University library. The machi bugyō papers, transferred to the city administration in Tokyo were ultimately moved also to the of the Imperial University library.152 All this material was lost in 1923.153 The modern assumption is that significant quantities of pre-1868 records had been transferred to ministries at their establishment and, retained there, were lost along with like deposits in Tokyo University Library in 1923.154 Even Fukui presumes that material was transferred to the Ōkurashō, and was then lost in 1923.155 This view is not tenable for the Ōkurashō and the Naimushō, which covered a wide range of interests. The Ōkurashō, well run after 1880,

260

EARLY JAPANESE TRADE

was guilty in its early years of bureaucratic neglect of archives. Ōkuma Shigenobu ኱㝰㔜ಙ(1838–1922) gave serious attention from 1878 to collecting pre-1868 records. Little material from the Ōkurashō went to the Naikaku Bunko in 1884–1885, suggesting it had inherited little in 1868, and exercised limited responsibility. Its keeping of new records like foreign trade statistics from 1868 to 1876 was also poor.156 Ōkuma’s decision to collect material was prompted by awareness of the absence within the ministry of earlier records. At Ōkuma’s direction, “necessary” work was undertaken on the finances and economy of Tokugawa times from the sources.157 The history of Edo times, which the Ministry compiled, was published in part on the eve of the 1923 earthquake. The remaining volumes, already in final draft or in text with the printers escaped nemesis, appearing in the two years after the earthquake. The series amounted to some 13,000 pages in ten volumes.158 The lack of material in the Ōkurashō is shown also in the fact that, though Katsu Kaishū worked in the 1880s with Okuma’s successor Matsukata Masayoshi ᯇ᪉ṇ⩏, his famous Suijinroku was drawn from miscellaneous and inferior sources. Because of their political sensitivity, the Naimushō held some material beyond 1885. It does not appear, however, to have been substantial, and it too was lost in 1923. What survives does so in the form of transcripts made in yet another great copying exercise, launched under the aegis of the Monbushō from 1911, and in the DNIS published some three decades later. CONCLUSION

Publication in modern printed form of records began with Katsu’s pioneering work in the 1880s. Apart from the unique and irreplaceable Suijinroku, his other works were volumes on army, navy and foreign policy respectively. For foreign policy his Kaikoku kigen 㛤ᅜ㉳ཎwith documents from the 1840s reflects the wide diffusion of copies in the senior circle of active or retired officials. But it is little cited in modern calendars of records. The source base in records surviving into Meiji was poor for the years prior to 1859. The first volume of the Bakumatsu gaikoku kankei monjo ᖥᮎእᅜ㛵ಀᩥ᭩ (BGKM) for Kaei 6 (1853), 6th and 7th months for instance, cites the TKIR on 71 occasions, reflecting the thin source base. As in TKIR’s use of Golownin’s account, the Japanese record in the BGKM was interspersed with Hawk’s official account of the Perry expedition. But for the years from 1860, the twentieth century

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

261

compilers of the BGKM were able to rely principally on the Gaimushō together with material relating to the bugyōsho of Kanagawa, Nagasaki and Hakodate.161 The BGKM contrasts with the DNIS, intended to be the political history of the Restoration and which, for want of central shogunal records, perforce relied heavily on the rich but uneven domain record. That venture grew out of the politically inspired proposal in 1890 by Kaneko Kentarō 㔠Ꮚሀኴ㑻for an editorial board of national history. The compilation of the DNIS, a task launched in 1911, could count on the good will of daimyo descendants. Publication (of the 4,215 kan of copied materials) was not envisaged until, in 1935, the government decided for political reasons to publish it. By way of contrast, for the years from 1859 the Gaimushō series became increasingly close to a Western style series of diplomatic documents. The records reconstituted from transcribed copies in the hands of individual officers cover both high policy, and the daily humdrum interaction of officials and foreigners. It provides a remarkable tribute to the vitality of Japanese administration. The response to the challenge presented by a permanent and unwanted foreign presence supports Kasaya’s observation that Japanese officials “possessed the facility to discern quickly the most suitable way to overcome problems, and were therefore able to respond to the perilous situations they faced by reforming themselves through a process of trial and error.”163 REFERENCES

Ansei ninen 2001 Ansei ninen yorozukichō Ᏻᨻ஧ᖺⴙグᖒ. Nagasaki Kenritsu Toshokan, 2001. Aston 1874 W.G. Aston. “Russian descents in Saghalin and Itorup in the years 1806 and 1807.” Reprinted in Collected Works of William George Aston, vol. 1. Ganesha Publishing, 1997, pp. 20–29. Aston 1879 W.G. Aston. “H.M.S. ‘Phaeton’ at Nagasaki in 1808.” Reprinted in Collected Works of William George Aston, vol. 1. Ganesha Publishing, 1997, pp. 107–120. Auslin 2004 Michael R. Auslin. Negotiating with Imperialism: The Unequal Treaties and the Culture of Japanese Diplomacy. Harvard University Press, 2004.

262

EARLY JAPANESE TRADE

Beerens 2000 Anna Beerens. “Interview with a bakumatsu official: a translation from Kyūji shimonroku.” Monumenta Nipponica 55:3 (2000), pp. 369–98. Beerens 2002 Anna Beerens. “Interview with a bakumatsu official: a translation from Kyūji shimonroku (2).” Monumenta Nipponica 57:2 (2002), pp. 173–204. BGKM Bakumatsu gaikoku kankei monjo ᖥᮎእᅜ㛵ಀᩥ᭩ 1855–1860. Dai Nihon komonjo ኱᪥ᮏྂᩥ᭩, 43 vols. Tokyo Daigaku Shiryō Hensanjo ᮾி኱Ꮫྐᩱ⦅⧩ᡤ, 1910–1991. BPP British Parliamentary Papers. In Area studies: Parliamentary Papers, Japan, ed. W.G. Beasley. Irish University Press, 1971–1972. Cullen 2003 Louis Michael Cullen. A History of Japan, 1582–1941: Internal and External Worlds. Cambridge University Press, 2003. Cullen 2004 Louis Michael Cullen. ‘“Sakoku,’ Tokugawa policy, and the interpretation of Japanese history.” Transactions of the Asiatic Society of Japan, 4th series, vol. 18 (2004), pp. 17–31. Cullen 2006 Louis Michael Cullen. “Tokugawa Population: The Archival Issues.” Japan Review, no. 18 (2006), pp. 129–80. Cullen 2009 Louis Michael Cullen. “Statistics of Tokugawa Coastal Trade and Bakumatsu and Early Meiji Foreign Trade, Part 1: Coastal Trade in Tokugawa Times.” Japan Review, no. 21 (2009), pp. 183–223. Cullen 2010 Louis Michael Cullen. “Statistics of Tokugawa Coastal Trade and Bakumatsu and Early Meiji Foreign Trade, Part 2: Trade in Bakumatsu and Early Meiji Times.” Japan Review, no. 22 (2010), pp. 59–102. DNIS Dai Nihon ishin shiryō ኱᪥ᮏ⥔᪂ྐᩱ1840–1858, 1st series 7 vols; 2nd series 5 vols; 3rd series 7 vols. Ishin Shiryō Hensan Jimusho, 1938–1943. Doi 1997 Doi Ryōzō ᅵᒃⰋ୕. Bakumatsu gonin no gaikoku būgyō: kaikoku o jitsugen saseta bushi ᖥᮎ஬ேࡢእᅜዊ⾜: 㛤ᅜࢆᐇ⌧ࡉࡏࡓṊኈ. Chūō Kōronsha, 1997.

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

263

EDUN Library of Economics Department, University of Nagasaki (i) Visit of the Return (under the date Enpō 1.6.28) 511M181 Enpō gan’nen ushi rokugatsu nijūhachi nichi Nagasaki bugyō Okano Magokurō dono… egeresusen junpū shidai defune ōsetsukerare sōrō ni tsuki men onwatashi sōrō kōjōgaki. ᘏᐆඖᖺଢ଼භ᭶஧༑ඵ᪥㛗ᓮᚚዊ⾜ᒸ㔝Ꮮ ஑㑻Ẋ㸬㸬㸬࠼ࡅࢀࡍ⯪㡰㢼ḟ➨ฟ⯪⿕௮௜ೃࢽ௜㠃ᚚΏೃཱྀୖ᭩. (ii) Phaeton Incident 1808 511M157 Ikokusen nyūtsū ikken ni tsuki oyakusho e Oranda meshidasare Magaribuchi Kai no kami sama ontazune no utsushi Kure Tokutarō ␗ᅜ ⯪ධ㏻୍௳஧௜ᚚᙺᡤỤ㜿⹒㝀⿕ྊฟ᭤ῡ᭤⏥ᩫᏲᵝᚚᑜࡢ෗࿋ᚨኴ㑻 ( October 1808).164 (iii) Rezanov Embassy 1804 407 M49 Roshia shisetsu torai hottan 㟢すள౑⠇Ώ᮶Ⓨ➃. Egoshi and Urakawa 2009 Egoshi Hiroto Ụ㉺ᘯே and Urakawa Kazuo ᾆᕝ࿴⏨. Ikokusen torai zakki ␗ᅜ⯪Ώ᮶㞧グ. Nagasaki Bunkensha, 2009. Enomoto 2004 Enomoto, Maki. “Just for the Record.” The Japan Journal (November 2004). Fujita 2005 Fujita Satoru ⸨⏣ぬ. Kinsei kōki seijishi to taigai kankei ㏆ୡᚋᮇᨻ἞ྐ࡜ ᑐእ㛵ಀ. Tokyo Daigaku Shuppankai, 2005. Fukui 1980 Fukui Tamotsu ⚟ ஭ ಖ . Naikaku Bunko shoshi no kenkyū: Edo bakufu Momijiyama Bunko bon no kōshō ෆ㛶ᩥᗜ᭩ㄅࡢ◊✲: Ụᡞᖥᗓ⣚ⴥᒣᩥ ᗜᮏࡢ⪃ド. Seishōdō Shoten, 1980. Fukui 1983 Fukui Tamotsu. “Kaisetsu” ゎㄝin Edo bakufu hensanbutsu Ụᡞᖥᗓ⦅⧩ ≀. Yūshōdō Shoten, 1983. Hall 1968 J.W. Hall. “Materials for the Study of Local History in Japan Pre-Meiji Daimyo Records.” In Studies in the Institutional History of Early Modern Japan, ed. J.W. Hall and M.B. Jansen. Princeton University Press, 1968, pp. 143–68.

264

EARLY JAPANESE TRADE

Hamada 1986 Hamada Giichirō. ὾⏣⩏୍㑻. Ōta Nanpo ኱⏣༡␇. Yoshikawa Kōbunkan, 1986 (revised 1963 edition). Hankachō (i) Hankachō: Nagasaki bug yōsho hanketsu kiroku ≢⛉ᖒ: 㛗ᓮዊ ⾜ᡤุỴグ㘓 1666–1867, 11 vols, ed. Morinaga Taneo ᳃Ọ✀ኵ. Hankachō Kankōkai, 1958–1961. (ii) Oshioki ukagai shū: Nagasaki bugyōsho kiroku ᚚ௙⨨ఛ㞟: 㛗ᓮዊ ⾜ᡤグ㘓, 2 vols., ed. Morinaga Taneo. Hankachō Kankōkai, 1962. (iii) Nagasaki bugyōsho kiroku kuchigaki shū 㛗ᓮዊ⾜ᡤグ㘓ཱྀ᭩㞟, 3 vols., ed. Morinaga Taneo. Hankachō Kankōkai, 1963–1964. (iv) Nagasaki daikan kirokushū 㛗ᓮ௦ᐁグ㘓㞟, 3 vols., ed. Morinaga Taneo. Hankachō Kankōkai, 1968.

Harada 2007 Harada Hiroji ཎ⏣༤ோ. “Kaidai” ゎ㢟in Nagasaki bugyōsho kankei shiryō 㛗ᓮዊ⾜ᡤ㛵ಀ㈨ᩱ, no. 1. Nagasaki Rekishi Bunka Hakubutsukan, 2007, pp. 269–80. Hayami 2001 Hayami, Akira. The Historical Demography of Pre-Modern Japan. University of Tokyo Press, 2001. Hayami 2008 Hayami Akira ㏿Ỉ⼥. “Kinsei Nihon no jinkō kōzō to hendō” ㏆ୡ᪥ᮏ ࡢேཱྀᵓ㐀࡜ኚື. Nihon Gakushiin Kiyō ᪥ᮏᏛኈ㝔⣖せ62:3 (2008). Honma 2000 Honma Sadao ᮏ㛫㈆ኵ. Kenritsu Nagasaki Toshokan kyōdo shiryō ni tsuite ┴❧㛗ᓮᅗ᭩㤋㒓ᅵ㈨ᩱ࡟ࡘ࠸࡚. Kiyō ᓮ㝧(2000), pp. 39–45. Ikeda 1973 Ikeda Shirō ụ⏣ྐ㑻. “Phaeton gō jiken to Saga han” ࣇ୍࢚ࢺࣥྕ஦௳ ࡜బ㈡⸬. Rekishi Ṕྐ305 (1973), pp. 120–25. Ikoku nikki Nakamura Tadashi ୰ ᮧ ㉁ et al. Eiinbon Ikoku nikki: Konchiin Sūden gaikō bunsho shūsei ᙳ༳ᮏ ␗ᅜ᪥グ: 㔠ᆅ㝔ᓫఏእ஺ᩥ᭩㞟ᡂ. Tokyo Bijutsu, 1989. Imaizumi 2011 Imaizumi Yoshiko. “The Making of a Mnemonic Space: Meiji Shrine Memorial Art Gallery 1912–1936.” Japan Review, no. 23 (2011), pp. 143–76.

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

265

Ishii 1974 Ishii Takashi ▼஭Ꮥ. Katsu Kaishū ຾ᾏ⯚. Yoshikawa Kōbunkan, 1974. Ishii 1998 Ishii, Yoneo. The Junk Trade from Southeast Asia: Translations from the Tōsen Fūsetsugaki, 1674–1723. Institute of South East Asian Studies, 1998. Itazawa 1937 Itazawa Takeo ᯈἑ೺ኵ. Oranda fūsetsugaki no kenkyū 㜿⹒㝀㢼ㄝ᭩ࡢ◊ ✲. Nihon Kobunka Kenkyūjo, 1937. Iwao 1979 Iwao Seiichi. Oranda fūsetsugaki shūsei, a Collection of the World News Presented Annually by the Dutch Factory at Deshima to the Tokugawa Shogunate, 1641–1858, 2 vols. Japan-Netherland Institute and Hōsei Rangaku Kenkyūkai, 1979. Kasaya 1998 Kasaya Kazuhiko ➟㇂࿴ẚྂ. Kinsei buke monjo no kenkyū ㏆ୡṊᐙᩥ᭩ ࡢ◊✲. Hosei University Press, 1998. Kasaya 2000 Kasaya, Kazuhiko. The Origins and Development of Japanese-Style Organization. International Research Center for Japanese Studies, 2000. Katagiri 1972 Katagiri Kazuo ∦᱒୍⏨. Sakoku jidai taigai ōsetsu kankei shiryō 㙐ᅜ᫬௦ ᑐእᛂ᥋㛵ಀྐᩱ. Kondō Shuppansha, 1972. Katō 2000 Katō, Yūzō. “The opening of Japan and the Meiji restoration, 1837–72.” In The History of Anglo-Japanese Relations, vol. I. Palgrave, 2000, pp. 60–86. Katō 2004 Katō Yūzō ຍ⸨♸୕. Bakumatsu gaikō to kaikoku ᖥᮎእ஺࡜㛤ᅜ. Chikuma Shobō, 2004. Katsu 1890 Katsu Kaishū ຾ᾏ⯚. Suijinroku ྿ሻ㘓, parts 1–5. Ōkurashō, 1890. In Katsu Kaishū zenshū ຾ᾏ⯚඲㞟, vols. 6–10. Keisō Shobō, 1974–1978. Katsu 1893 Katsu Kaishū. Kaikoku kigen 㛤ᅜ㉳ཎ, 4 parts. 1893. In Katsu Kaishū zenshū ຾ᾏ⯚඲㞟, vols. 15–18. Kōdansha, 1973–1975. KBS Kokuritsu Kōbunshokan ᅜ❧බᩥ᭩㤋 National Archives (formerly Naikaku Bunko ෆ㛶ᩥᗜ)

266

EARLY JAPANESE TRADE

A.

Nagasaki shi zokuhen 㛗ᓮᚿ⥆⦅(1780s to 1839)

(i) 176–95; (ii) 176–97; (iii) 176–96; (iv) 176–113; (v) 176–103 B. Kankai ibun ⎔ᾏ␗⪺(i) 185–107 (16 satsu); (ii) 185–113 (10 satsu); (iii) 185–130 (16 satsu); (iv) 271–7 (16 satsu); (v) 271–5 (6 satsu) C. Rezanov Embassy 1804 (i) 185–227 Hyōryū jin Roshia wa ⁻ὶே㟢すளヰ; (ii) 185–274 Garakisen no ki ಂ⨶᪁⯪அグ; (iii) 185–261 Roshia torai ikken 㟢すளΏ᮶୍௳ Kizaki 2005 Kizaki Hiroyoshi ᮌᓮᘯ⨾. Kinsei gaikō shiryō to kokusai kankei ㏆ୡእ஺ ྐᩱ࡜ᅜ㝿㛵ಀ. Yoshikawa Kōbunkan, 2005. Kōbe mokuroku 1997 Kōbe Shiritsu Hakubutsukan kanzōhin mokuroku: Bijutsu no bu ⚄ᡞᕷ❧༤ ≀㤋㤋ⶶရ┠㘓: ⨾⾡ࡢ㒊13; monjo ᩥ᭩V. 1997. Kokon shūran a. Nagasaki kokon shūran meishō zue 㛗ᓮྂ௒㞟ぴྡ຾ᅗ⤮ (compiler Matsuura Tōkei ᯇᾆᮾ῱, 1752–1820). Nagasaki Bunkensha, 1975. b. Nagasaki kokon shūran 㛗ᓮྂ௒㞟ぴ (compiler Matsuura Tōkei). Nagasaki Bunkensha, 1976. Kondō Jūzō Kondō Jūzō ㏆ⶶ㔜ⶶ. (i) Kondō Jūzō Ezochi kankei shiryō ㏆⸨㔜ⶶ⼎ዀᆅ㛵ಀ㈨ᩱ. 3 vols. Tokyo Daigaku Shuppankai, 1984–1989. (ii) Kondō Jūzō Ezochi kankei shiryō: fuzu bessatsu ㏆⸨㔜ⶶ⼎ዀᆅ㛵ಀ ㈨ᩱ௜ᅗู෋. Tokyo Daigaku Shuppankai. 1993. Kunaichō Shoryōbu Kunaichō Shoryōbu ᐑෆᗇ᭩㝠㒊. (i) (ii) (iii) (iv)

Togawa ke nikki kiroku ᡞᕝᐙ᪥ググ㘓. vols. 4–10 (1827–1839). Kiyō Nagasaki ⣖せ㛗ᓮ(vol. 14, 1837). Nagasaki tochūki 㛗ᓮ㏵୰グ(vol. 15, 1838). Nagasai tochū zakki 㛗ᓮ㏵୰㞧グ(vol. 16, 1838). Kutsukake 2007

Kutsukake Yoshihiko 2007 Kutsukake Yoshihiko ἅ᥃Ⰻᙪ. Ōta Nanpo ኱⏣༡␇. Minerva Shobō, 2007. MacDonald 1848 Ranald MacDonald, interrogation of, in 1848. Shahon nos. 139 and 140 in Matsuki Collection. Center for the Study of Kyushu Cultural History, Department of Literature, Kyushu University.

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

267

Machin 1978 Roger Machin. Experiment and Return: Documents concerning the Japan Voyage of the East India Company, 1671–3. Richard Cocks Society, 1978. Matsukata 2007 Matsukata Fuyuko ᯇ᪉෤Ꮚ. Oranda fūsetsugaki to kinsei Nihon ࢜ࣛࣥࢲ 㢼ㄝ᭩࡜㏆ୡ᪥ᮏ. Tokyo Daigaku Shuppankai, 2007 Matsuura 2010 Matsuura Rei ᯇᾆ⋹. Katsu Kaishū ຾ᾏ⯚. Chikuma Shobō, 2010. McEwan 1962 J.R. McEwan. The Political Writings of Ogyū Sorai. Cambridge University Press, 1962. Murakami 1899 N. Murakami. Diary of Richard Cocks, cape-merchant in the English factory in Japan, Tokyo, vol. 2. Sankōsha, 1899. Nagasaki chōsa hōkoku 1997 Nagasaki Ken Kyōiku Iinkai 㛗ᓮ┴ᩍ⫱ጤဨ఍, ed. Nagasaki bugyōsho kankei monjo chōsa hōkokusho 㛗ᓮዊ⾜ᡤ㛵ಀᩥ᭩ㄪᰝሗ࿌᭩. Nagasaki ken bunkazai hōkokusho 㛗ᓮ┴ᩥ໬㈈ሗ࿌᭩, no. 131. Nagasaki Ken Kyōiku Iinkai, 1997. Nagasaki shi seihen 1973 Nagasaki shi seihen 㛗ᓮᚿṇ⦅(also styled Nagasaki jitsuroku taisei: Seihen 㛗ᓮᐿ㘓኱ᡂ: ṇ⦅), ed. Morinaga Taneo ᳃Ọ✀ኵ. Nagasaki Bunkensha, 1973. Nagasaki shi zokuhen 1974 Nagasaki shi zokuhen 㛗ᓮᚿ⥆⦅(also styled Zoku Nagasaki jitsuroku taisei ⥆㛗ᓮᐿ㘓኱ᡂ), ed. Morinaga Taneo. Nagasaki Bunkensha, 1974. Ms versions; see under KBS and NDL. Naikaku Bunko hyakunen shi 1986 Kokuritsu Kōbunshokan ᅜ❧බᩥ᭩㤋, ed. Naikaku Bunko hyakunen shi ෆ㛶ᩥᗜⓒᖺྐ. Kyūko Shoin, 1986. Naikaku mokuroku (i) Naikaku Bunko kokusho bunrui mokuroku ෆ㛶ᩥᗜᅜ᭩ศ㢮┠㘓, 3 vols. Naikaku Bunko, 1961–1962 (2 vols. plus index vol. Revised version, 1974–1976). (ii) Bunsetsu ᩥ⠇, vol. 1. (iii) Naikaku Bunko shōshi ෆ㛶ᩥᗜᑠྐ, vol. 2.

268

EARLY JAPANESE TRADE

Nakada 1985 Nakada Masayuki ௰⏣ṇஅ. Egawa Tan’an Ụᕝᆠᗡ. Yoshikawa Kōbunkan, 1985. Nakada 1998 Nakada Masayuki. Nirayama daikan Egawa shi no kenkyū 㡞ᒣ௦ᐁỤᕝẶ ࡢ◊✲. Yoshikawa Kōbunkan, 1998. Nakamura et al. 1989 Nakamura Tadashi et al., eds. Eiinbon Ikoku nikki: Konchi’ in Sūden gaikō bunsho shūsei ᙳ༳ᮏ␗ᅜ᪥グ: 㔠ᆅ㝔ᓫఏእ஺ᩥ᭩㞟ᡂ. “The ikoku nikki: Konchiin Sūden’s Compilation of Foreign Correspondence.” Translated by James B. Lewis, in Ikoku nikki 1989. Nakao 1997 Nakao Hiroshi ௰ᑿᏹ. Chōsen tsūshinshi to Tokugawa bakufu ᮅ㩭㏻ಙ౑ ࡜ᚨᕝᖥᗓ. Akashi Shoten, 1997. NDL National Diet Library ᅜ❧ᅜ఍ᅗ᭩㤋. A. Kankai ibun ⎔ᾏ␗⪺MSS (i) 862–1; (ii) 130–27; (iii) W338N2. B. Nagasaki shi zokuhen 㛗ᓮᚿ⥆⦅MSS 319–3 (4 satsu). C. Enkai ibun ἢᾏ␗⪺MSS (Ōta Nanpo MSS) (i) ms no. 181–54, 4 satsu containing kan nos. 3, 7, 9, 13. (ii) ms no. 83–10, 1 satsu containing kan no. 2. (iii) ms no. 14–8, 15 satsu containing 13 of the original 15 kan. Nihon zaisei keizai shiryō Nihon zaisei keizai shiryō ᪥ᮏ㈈ᨻ⤒῭ྐᩱ, 10 volumes. Ōkurashō Zaisei Keizai Gakkai, 1921–1925. NRBH Nagasaki Rekishi Bunka Hakubutsukan 㛗ᓮṔྐᩥ໬༤≀㤋 (Nagasaki Museum of History and Culture). A. Phaeton Incident 1808 (i)

16–1–3 Ikokusen torai no setsu osonae muki ikken ni tsuki ryōke tasshi ukagai tome: Magaribuchi Kai no kami sama ni gozaikin ␗ ᅜ⯪Ώ᮶அ⠇ᚚഛྥ୍௳࡟௜୧ᐙ㐩ఛ␃: ᭤ῡ⏥ᩫᏲᵝᘥᚚᅾ ໅ (14 February–July 1809, 365 folios). Bunka gonen shōgatsu to Bunka rokunen rokugatsu.

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

(ii)

269

Fuji 14–186 Ikokusen torai no setsu onsonae taii onkakitsuke: Magaribuchi Kai no kami sama ni gozaikin ␗ᅜ⯪Ώ᮶அ⠇ᚚഛ ኱ពᚚ᭩௜: ᭤ῡ⏥ᩫᏲᵝᘥᚚᅾ໅ (July 1809). (See Katagiri 1972 for edited version.)

(iii) Watanabe 14–245 Igirisusen torai ikken ࢖ࢠࣜࢫ⯪Ώ᮶୍௳ (4 November 1808, 7 folios). (iv) Fuji 14–126 Shobasho shohō ninzū warikae sōrō omomuki mōshiwatashi onkakitsuke ㅖሙᡤㅖ᪉ேᩘ๭᭰ೃ㊃⏦Ώᚚ᭩௜ (July 1809, 17 folios). (v) 13–392 Tasshigaki Fueton gō jiken Matsudaira Hizen no kami naikan den no in 㐩 ᭩ࣇ࢚ࢺࣥྕ஦௳ᯇᖹ⫧๓Ᏺෆ㛵ఏஅඔ (14 November 1808). (vi) 390–20 Nagasaki bugyōsho Matsudaira zusho no kami sama onwatarima hissha kakidome no utsushi 㛗ᓮዊ⾜ᡤᯇᖹᅗ᭩㢌ᵝ ᚚᗘ㛫➹⪅᭩␃அ෗. (vii) 310–63 Ikoku torai no ki. Nagasaki bugyōsho Matsudaira zusho no kami sama onwatarima kakidome no utsushi ␗ᅜΏ᮶அグ( 㛗ᓮዊ⾜ᯇᖹᅗ᭩㢌ᵝᚚᗘ㛫᭩␃෗) (1808). B. Rezanov Embassy 1804 (i) 13–579 Bunka gan’nen Roshia sen Nagasaki e raichō ikken ᩥ໬ඖᖺ ࣟࢩ࢔⯪㛗ᓮ࠼᮶ᮅ୍௳. (ii) 14–2–3 Roshia sen nyūshin yori shuppan made no kiroku Bunka gan’nen ninen 㟢すள⯪ධὠࣚࣜฟᕹ㎾グ㘓ᩥ໬ඖᖺᘨᖺ (1804–5, 156 folios). (iii) Fuji 13–84 Bunka gan’nen ne no kugatsu muika nyūshin Oroshia sen ikken ᩥ໬ඖᖺᏊ஑᭶භ᪥ධὠ࠾ࢁࡋ࠶⯪୍௳(17 folios). (iv) 13–244 Roshia sen fūsetsugaki 㟢すள⯪㢼ㄝ᭩(9.10.1804, 88 pages). C. Fūsetsugaki (i) 14–119–1 Oranda fūsetsugaki 㜿⹒㝀㢼ㄝ᭩(1827–1856, 144 pages). (ii) Fukuda 14–116 Fūsetsugaki 㢼ㄝ᭩(1843–1845). OCCI See OSSS 1963–1966. Ōhori 2007 Ōhori Satoshi ኱ᇼဴ. “Goaisatsu” ࡈ࠶࠸ࡉࡘ. In Nagasaki bugyōsho kankei shiryō 㛗ᓮዊ⾜ᡤ㛵ಀ㈨ᩱ, no. 1. Nagasaki Rekishi Bunka Hakubutsukan, 2007. Ōsaka Dōjima beishō enkaku 1903 Ōsaka Dōjima beishō enkaku ኱㜰ᇽᓥ⡿ၟἢ㠉. Ōsaka Dōjima Beikoku Torihikisho, 1903.

270

EARLY JAPANESE TRADE

Ōsaka-shi shi Ōsaka-shi shi ኱㜰ᕷྐ, 5 vols. Ōsakashi Sanjikai, 1911–1914. Ōshima 2000 Rezānofu and trans. Ōshima Mikio ኱ᓥᖿ㞝. Nihon taizai nikki 1804– 1805 ᪥ᮏᅾ᪥グ1804–1805. Iwanami Shoten, 2000. OSSS 1963–1966 Osaka Shōkō Kaigisho ኱㜰ၟᕤ఍㆟ᡤ, ed. Osaka shōgyōshi shiryō ኱㜰ၟ ᴗྐ㈨ᩱ, vols. 1–35. Ōsaka Shōkō Kaigisho, 1963–1966. OSSS supplement 1966 Osaka Shōkō Kaigisho, ed. Ōsaka shōgyōshi shiryō: Bekkan kaisetsu-hen sōmokuji ኱㜰ၟᴗྐ㈨ᩱ: ูᕳゎㄝ⦅⥲┠ḟ. Ōsaka Shōkō Kaigisho, 1966. Ōta Nanpo zenshū Hamada Giichirō ℈⏣⩏୍㑻, ed. Ōta Nanpo zenshū ኱⏣༡␇඲㞟, vol. 19. Iwanami Shoten, 1985. Roshia torai roku 1994 Roshia torai roku 㟢すளΏ᮶㘓. Isahaya kyōdo shiryō sōsho ㄴ᪩㒓ᅵྐᩱྀ ᭩III. Isahaya Kyōdo Shiryō Kankōkai, 1994. Sanada ke monjo mokuroku 1979 Sanada ke monjo mokuroku ┿⏣ᐙᩥ᭩┠㘓. In Shiryōkan shozō shiryō mokuroku ྐᩱ㤋ᡤⶶྐᩱ┠㘓, no. 28, pp. 373–400. Tōkyō Kokuritsu Shiryōkan, 1979. Sasama1965 Sasama Yoshihiko ➲㛫Ⰻᙪ. Edo bakufu yakushoku shūsei Ụᡞᖥᗓᙺ⫋㞟 ᡂ. Yūzankaku Shuppan, 1965. Schodt 2003 Frederik L. Schodt. Native American in the Land of the Shogun. Stone Bridge Press, 2003. Seishikan Kinenkan 1993 Seishikan Kinenkan shozōhin zuroku ㄔஅ㤋グᛕ㤋ᡤⶶရᅗ㘓. Alumni Association, 1993. Sekiyama 1957 Sekiyama Naotarō 㛵ᒣ┤ኴ㑻. “Kansei jūnen oyobi Bunsei gonen no kunibetsu jinkō: Shin shiryō no hakken.” ᐶᨻ༑ᖺཬᩥᨻ஬ᖺࡢᅜูே ཱྀ: ᪂ྐᩱࡢⓎぢ. Keizai riron ⤒῭⌮ㄽ37 (1957), pp. 61–74. Shimazu ke monjo mokuroku (i) Shimazu ke monjo mokuroku: Kurourushi nuribako ᓥὠᐙᩥ᭩┠ 㘓: 㯮⁽ሬ⟽. Shiryō Hensanjo, 1997.

JAPANESE ARCHIVES: SOURCES FOR THE STUDY OF TOKUGAWA

271

(ii) Shimazu ke monjo mokuroku ᓥὠᐙᩥ᭩┠㘓II (Shiraki monjo tō zen ⓑᮌᩥ᭩➼඲, 4,511 ten). Shiryō Hensanjo, 1999. Shiryō Hensanjo (i) 2051.9/77. Roshiajin toriatsukai tedome 㟢すளேྲྀᢅᡭ␃. “Shahon,” 3 satsu. (ii) S0051–1. Roshia shisetsu Rezanofu raikō 㟢すள౑⠇ࣞࢨࣀࣇ᮶⯟ (2 picture scrolls). (iii) 4171.08–14. Hayashi Jussai shokan ᯘ㏙ᩪ᭩⡆. (“Shahon” from Ōkōchi bunsho ኱Ἑෆᩥ᭩). Shizuoka Kenritsu Chūō Toshokan 1970 Shizuoka Kenritsu Chūō Toshokan Aoi Bunko 㟼ᒸ┴❧୰ኸᅗ᭩㤋ⵇᩥ ᗜ. Edo bakufu kyūzō yōsho mokuroku Ụᡞᖥᗓᪧⶶὒ᭩┠㘓. Shizuoka Kenritsu Chūō Toshokan Aoi Bunko, 1970. Shizuoka Kenritsu Chūō Toshokan 1996 Onko chishin: Shizuoka Kenritsu Chūō Toshokan shozō no kichōsho shōkai   ᨾ▱᪂: 㟼ᒸ┴❧୰ኸᅗ᭩㤋ᡤⶶࡢ㈗㔜᭩⤂௓. Shizuoka Kenritsu Chūō Toshokan, 1996. SMH Shīboruto Kinenkanࢩ࣮࣎ࣝࢺグᛕ㤋(Siebold Memorial Hall). (i)

Roshia senchū nikki 㟢すள⯪୰᪥グ (October 1804–April 1805, ed. Orita Takeshi ⧊⏣Ẏin Narutaki kiyō 㬆⣖せ).

(ii)

14–2–93. Fūsetsugaki 㢼ㄝ᭩. 1660–1763.

(iii) 14–2–91. Fūsetsugaki 㢼ㄝ᭩. 1830. Sugimoto 1986 Sugimoto Tsutomu ᮡᮏࡘ࡜ࡴ (with Shimura Hiroyuki ᚿᮧᘯᙉ). Kankai ibun: Honbun to kenkyū ⎔ᾏ␗⪺: ᮏᩥ࡜◊✲. Yasaka Shobō, 1986. Tanaka 1998 Tanaka Masahiro ⏣୰ṇᘯ. Kindai Nihon to bakumatsu gaikō bunsho hensan no kenkyū ㏆௦᪥ᮏ࡜ᖥᮎእᩥ᭩⦅⧩ࡢ◊✲. Shibunkaku Shuppan, 1998. Tanaka 1996 Tanaka Takeo ⏣୰೺ኵ. Zenkindai no kokusai kōryū to gaikō bunsho ๓㏆௦ ࡢᅜ㝿஺ὶ࡜እ஺ᩥ᭩. Yoshikawa Kōbunkan, 1996. TKIR Tsūkō ichiran ㏻⯟୍ぴ (1566–1825), 8 vols. Kokusho Kankōkai, 1912– 1913.

272

EARLY JAPANESE TRADE

TKIR zokushū Tsūkō ichiran zokushū ㏻⯟୍ぴ⥆㍴ (1825–1856), 5 vols. Seibundō, 1968–1972. Toby 1991 Ronald P. Toby. State and Diplomacy in Early Modern Japan: Asia in the Development of the Tokugawa Bakufu. Stanford University Press, 1991. Tō tsūji nichiroku Tō tsūji kaisho nichiroku ၈㏻஦఍ᡤ᪥㘓, 7 vols. Dainihon kinsei shiryō ኱ ᪥ᮏ㏆ୡྐᩱ. Tokyo Daigaku Shuppankai, 1984. Tsūshin zenran (i) Tsūshin Zenran Henshū Iinkai ㏻ಙ඲ぴ⦅㞟ጤဨ఍, ed. Tsūshin zenran, first and second series ㏻ಙ඲ぴṇ࣭⥆⦅, 60 vols. Yūshōdō Shoten, 1983–1988. (ii) Tsūshin Zenran Henshū Iinkai ㏻ಙ඲ぴ⦅㞟ጤဨ఍, ed. Tsūshin zenran: sōmokuroku, kaisetsu ㏻ಙ඲ぴ⥲┠㘓࣭ゎㄝ. Yūshōdō Shoten, 1989. Yamamura 1974 Kozo Yamamura. A Study of Samurai Income and Entrepreneurship: Quantitative Analyses of Economic and Social Aspects of the Samurai in Tokugawa and Meiji, Japan (East Asian). Harvard University Press, 1974. Yasutaka 2010 Yasutaka Hiroaki Ᏻ㧗ၨ᫂. Kinsei Nagasaki shihō seido no kenkyū ㏆ୡ㛗ᓮ ྖἲไᗘࡢ◊✲. Shibunkaku Shuppan, 2010.

Source: Japan Review No. 31 (2017), pp. 69–104.

10

The Nagasaki Trade of the Tokugawa Era: Archives, Statistics, and Management ™

The study of Tokugawa-period trade policy poses problems because of the poor survival of archives. Rōjū took their records with them when they vacated office, as did bugyō in Nagasaki. Trade records transmitted to Edo had a poor survival rate. In contrast, the records of the Kaisho (trade office) and of interpreters in Nagasaki were remarkably well maintained up to the Meiji Ishin. We know less about the process of loss in 1868 than we do about the effort of a small number of individuals to recover records. In Japanese sources, trade statistics—apart from originals for 1709–1714 (wrongly said to be Edo files)—survive in mere scraps for both the Chinese and Dutch trades. As a consequence, the archives of the Dutch factory on Dejima are not only a complete run for the Dutch trade, but even with gaps compensate in part for the loss of records of the trade with China. Sakoku did not imply intent to reduce trade. It reached its peak in 1661, and thereafter the shortage of silver and copper successively posed problems. The Dutch trade receded from the 1690s. The Chinese trade by contrast recovered briefly in the 1690s and the early years of the following century, partly through the presence of a lobby favoring imports, partly by some upturn in the copper supply. Nagasaki’s prosperous days were, however, behind it by the 1720s. A recovery from the end of the eighteenth century was not broad based, but simply a burgeoning exchange of marine products for medicinal products. The Nagasaki authorities, seeking a quasi-monopoly of this trade for the port, had long sought to eliminate Satsuma from much of it. The evasion of restrictions was exaggerated by facile assumptions about the extent of smuggling. 273

274

EARLY JAPANESE TRADE

INTRODUCTION

Japanese trade is a story of trade buoyed up by an abundant supply of silver in the first sixty years of the seventeenth century, followed by progressive contraction, especially after 1715, although towards the end of the eighteenth century, there was an upturn in Chinese trade. This story has also to be seen in the context of the sakoku policy from the 1630s. One thing is clear, however, from the tenor of the statistical evidence. The Japanese were not actually seeking policy reasons to reduce the volume of trade at that time or even to lower the ceilings set to trade either in 1685 or in the much less happy circumstances of 1715. Trade continued to grow for several decades to a peak in 1661. On the Japanese side it was imports and not exports that provided the drive for promoting trade. The key commodity was silk, in which domestic production was deficient in both quantity and quality. The rise in incomes among daimyo and upper samurai amid the prosperity of a stable and increasingly un-warlike Tokugawa regime accounts for the sharp rise in silk imports. Japanese weaving itself was already of a high standard; the weakness lay in the inferior quantity and quality of raw silk. But quite quickly the pattern began to change, with raw silk holding up better than cloth in imports. In other words, the domestic industry was already maturing, in time virtually terminating imports. The change in silk was to be part of a process over two centuries of silent but profound change in domestic production: it laid the basis in the open economy of the 1860s and 1870s for a dramatic expansion of exports of both tea and silk. For foreigners, the lure of Japan lay in silver, important in money supply in many countries. On a world scale, Japan’s output of silver was large, and expansion of mining provided in the short term what seemed a painless means of payment. For Europeans, lacking domestic supplies, they had to acquire silk in China and Southeast Asia, or in the case of the Dutch, in default of supplies, on occasion by plundering vessels laden with silk for Japan. The revamped bugyōsho (governor’s administration, located in offices on four sites) of 1633, with power transferred from local elite figures to shogunal officers dispatched from Edo, put the five decrees on sakoku into effect in Nagasaki. The move of the Dutch from Hirado to Nagasaki in 1641 was primarily prompted by the simple urge of concentrating all foreign trade in a single port under the watchful eye of a reinforced bugyōsho.

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

275

Seventeenth-century Japanese trade was an exchange of silk for silver, gold, and later copper; silk accounting for 70 percent or so of the value of imports. Japanese early expansion attracted vessels from three European nations in addition to the Portuguese already using Nagasaki as a base since 1570. It is impossible to quantify this early trade, not least because of its many channels: Japanese red-seal ships, Portuguese vessels, Dutch and Chinese and fleetingly English and Spanish. Chinese vessels traded at many locations along the coast of Kyushu in contrast to Europeans tied ab initio to either Hirado or Nagasaki. In a recasting of control of trade, all foreign trade narrowed down by the end of the 1630s to Dutch and Chinese traders, with Nagasaki becoming the sole center of foreign trade. The control of the trade lay firmly in the hands of shogunal officials, aided by a new breed of interpreters employed directly by them, and not as in the past by the foreign traders. One result was that meaningful statistical totals for commodities began to appear from 1648. This paper is a study of archives, and of the overall statistical profile of Japanese trade over two centuries. Many of the sources, and indeed most quantitative sources, have failed to survive in Japan. Contracting from a peak in 1661, meager evidence suggests stagnation from the 1670s. But there was an upturn in the 1690s and the following decade, and the case was even argued for increasing imports. Chinese trade was later to stage a recovery in the late eighteenth century. Vessels were fewer, but cargo values significantly higher. The policy of the rōjū is far from clear in the absence of rōjū archives. Even papers by officials such as the surviving papers of Arai Hakuseki ᪂஭ⓑ▼ (1675–1725), advisor to two shoguns, 1709–1716, are all too rare.1 The older view of sakoku was that it was an isolationist and reactionary policy. It was well summarized as late as 1970 by the American historian Harootunian, for whom the uchi harai rei (fire and repel order) of 1825 was “little more than a tired restatement of Tokugawa isolationship, which revealed the incapacity to see beyond the immediate implications of events.”2 The work of Iwao in the 1950s and 1960s, followed by Tashiro Kazuo’s writing which provided a coherent economic and diplomatic account of Tsushima and the Korean trade, laid the basis for reinterpreting sakoku policy.3 In some ways Iwao’s is the more influential reinterpretation. He claimed that sakoku was directed against existing Catholic countries, and that a reopening of English trade might have occurred had England sought it in later times when, after the reign of Charles II, the problem of a Portu-

276

EARLY JAPANESE TRADE

guese consort no longer arose. In the absence of archival evidence, however, Iwao not only adverted to a Dutch effort to foment Japanese unease about the English vessel, the Return, seeking trade in 1673, but assumed that it accounted for the rejection of the English request.4 The case was taken a step further in 1984 in the influential State and Diplomacy in Early Modern Japan by Ronald Toby. He followed Iwao’s arguments closely and, as a research student, had benefitted by contacts with Tashiro, at the time himself a doctoral candidate.5 Toby argued that “the possibilities of the system were far more openended, more manifold, than what has been visible in the received vision of the Tokugawa past.”6 He saw Matsudaira Sadanobu, leading senior councillor in 1787–1793, as taking advantage in 1793 of the absence of contacts other than with the Dutch and Chinese for a century and a half, to create an argument that the Russian request for trade was precluded by “ancestral law.”7 Much discussion of the issues was to take place in Japan from the 1970s to the 1990s. For Arano Yasunori, in the subtlest contribution to the debate, the central measure in the sakoku policy of the 1630s was the prohibition on Japanese going abroad. The shogunate’s long-standing policy was to avoid being drawn further into the troubles of East Asia (the reason for prohibiting Japanese settlement overseas). Contemporaries did not, on his argument, see the issue as one of either opening or closing the country, and the aim was to conserve orderly relations with Japan’s neighbors. Only from the 1790s onwards, faced with novel Western appearances, did a clear-cut idea of a closed society begin to take a forceful shape.8 It is commonplace for modern accounts to refer to four portals (yottsu no kuchi ⚃̥̯⎋) of external trade: Nagasaki itself, Korea (via the island of Tsushima), the Ryukyus (for Chinese goods to Nagasaki or Satsuma), and Ezo. This concept is a somewhat optimistic assessment of the reality and extent of “foreign” trade. Its significance is weakened by the arguments of Nakamura and others about an overstatement of Tsushima trade.9 As for Ezo, the absence of a frontier in the chain of islands to the north of Japan suggests that the trade was a domestic one, which for well over a century expanded little. There are two practical challenges in the statistical study of trade. One is that notional limits to foreign trade set in 1685, while intended to define firm ceilings to exports of silver, proved flexible in regard to the actual size of imports, once payment was not in silver. The second and related problem is Satsuma’s trade with the Ryukyus, for which ungenerous notional limits and efforts to confine its shipments to

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

277

sales in Nagasaki were set in the 1680s.10 These were never observed by Satsuma and, for political reasons, were de facto unenforceable by the shogunate. This situation makes it necessary to look at allegations of smuggling, which in the absence of reliable documentary evidence, are sometimes too readily made. ARCHIVES AND LOST DOCUMENTS

There is a paradox in early Japanese statistics. They were abundant in the mid-seventeenth century, when they were few in Europe. But though European trade statistics were slower to appear by as much as two or three generations, when they did appear they survived, in contrast to Japanese figures. Trade data, when passed from Nagasaki to Edo, had a poor survival rate.11 By contrast, the sources held in Nagasaki remained complete until after virtual total loss in 1868. Compilation from Japanese sources of either a full or a partial record of the Dutch and Chinese trades is now impossible. A problem does not arise for the Dutch trade as the Dutch sources compiled in Dejima provide a remarkably detailed record. For the Chinese trade, were we to depend on Japanese sources, apart from isolated satsu for 1709–1714 and a private notebook of a merchant in 1804, we would be almost entirely in the dark.12 The statistics, in so far as they can be assembled, otherwise come from the Dutch records. Good record keepers though they were, the Dutch did not consistently transfer statistical information from the jonken boekjes (booklets recording cargoes on Chinese junks) into the dagregister office diary or daily record kept by the opperhoofd (head of the factory). From the 1640s, the dagregister records much detail of the import trade because figures for silk supplied by their competitors were of vital interest to the Dutch. The export trade on the other hand was recorded perfunctorily in the dagregister, and references to copper begin to recur frequently only from the 1680s. For imported cargoes, a separate record for 1652–1657 was a precursor of the jonken boekjes of later decades.13 Early practice in relation to recording exports is not clear, but in December 1689 the “possible cargoes” of eighteen outgoing junks were noted in a jonken boekje.14 With figures entered in jonken boekjes there was no compelling reason to enter precise figures in the dagregister itself. But for that reason the later loss of the jonken boekjes resulted for historians in an irreparable loss for the years up to 1706 when their compilation seems to have ceased.

278

EARLY JAPANESE TRADE

Dutch access to information seems to have remained problemfree until a novel clampdown was imposed from 1682 on information on the Chinese trade in the face of a worsening crisis in the supply of silver and, post-1684, an abrupt rise in the number of Chinese vessels. In 1689 the Chinese were corralled in a Tōjin yashiki ⒸġṢġ⯳ġ㔟ġ, (often referred to by the Dutch as “the Chinese island”), an enclave of 229 by 133 meters surrounded by a wall and four watch towers, a temporary home every year for over two thousand individuals in unhygienic and difficult circumstances. The Japanese interpreters of Dutch now lacked a ready flow of information from the interpreters of Chinese. Nevertheless, an interpreter, Motoki Tarōemon 㛔㛐⣒恶⎛堃攨, remained an informant to a greater or lesser degree, until his death in 1695. In noting in October that without him the rumored amount of copper on seventeen vessels could not be confirmed, the dagregister seems to hint that his services had remained useful.15 In the circumstances of the time he did not have a replacement. But a trickle of information from various, usually lesser, sources was possible in part because the junior Dutchmen, many of whom had remained for years in Dejima, spoke some or much Japanese.16 The monumental work by Nagazumi to reconstitute the Chinese trade from the dagregister suggests unintentionally that the Dutch were more poorly informed than they were. Much of the recorded information was too vague for use in her scholarly approach.17 Intermittent information, however, continued to be received. On one occasion, uniquely in the reporting of trade, the dagregister noted in November 1696 after five junks had left, that “secretly we manage[d] to get the little book in which their sales had been recorded.”18 The jonken boekjes continued in existence: on an occasion in November 1703 the opperhoofd recorded that “the junks booklet and the diary give different numbers.”19 The information flow deteriorated very sharply for 1708–1718. It is tempting to assume that a ragged supply of intelligence led to abandonment of the jonken boekjes (never mentioned in the dagregister after 1706), and these now obsolete documents were lost with the passage of time. An abrupt upturn in 1718 was made possible by the fact that a Dutch interpreter was brother to the chief interpreter of the Chinese island. In February 1718, Ichijirōzaemon, the chief interpreter, provided a bill of lading for nine outward junks.20 Thereafter the information came from his brother. This situation lasted till July 1727

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

279

when the opperhoofd was told by an interpreter, “On pain of corporal punishment the governor had forbidden the Japanese to supply us with information concerning Chinese exports and imports.”21 Two years later an interpreter, when asked for a price list, replied, “The governors had forbidden the servants of the Chinese island to disclose the prices.”22 A flow resumed in 1732–1738. Interpreters, like Japanese officials at large in financial penury, privately borrowed money from the Dutch. That may explain some of the information flow. But by 1740 the new regularity of the flow must have reflected an easing in policy. The informants were either rapporteur interpreters (nenban tsūji ⸜ġ䔒ġ忂ġḳġ), or interpreters well established enough to later reach that rank. Information was also usually received in writing, and was without fail and in full recorded in the dagregister. Thereafter the only interruption for the remainder of the century was in exports only for 1754/1755. In October 1755 the interpreters, when asked for details of cargoes in and out, informed the Dutch that “the interpreters of the Chinese island have been forbidden to inform them about the cargo (sic) of the junks.”23 Edo government and Nagasaki bugyōsho retained within their archives few policy documents. Rōjū as a matter of course took them away with them at the end of their period of office. And Nagasaki bugyō seem to have done likewise. While statistics appear to have survived very well in Nagasaki, they were less secure in Edo. Comprehensive runs of figures from 1648 were furnished on at least two known occasions (at Hakuseki’s request in 1708 and on a rōjū order in 1719); perhaps two of many such requests. They imply either a lack of material already in Edo or simply poor awareness of what was held there. In the great assemblage of documents from several centuries collected in the late 1840s and 1850s, the Tsūkō ichiran 忂ġ凒ġᶨġ奏ġ(TKIR), launched by a team engaged by shogunal order in the late 1840s, there is a gap for the Dutch trade after 1670 and for Chinese trade after 1672 despite their consulting the Hakuseki records. The TKIR lacks runs of figures for other items apart from listing until 1718 gold for the Dutch.24 For copper, despite its importance as the dominant export, there is a remakable lacuna in records for three and a half decades. The TKIR has no details of copper. Katsu Kaishū ⊅㴟凇, with an official brief in the 1880s, despite past service as an official both in Tokugawa and Meiji times, was able to provide litttle data on trade at large. In the case of copper, for the Dutch trade there were gaps in his figures, which themselves appear to have been drawn from isolated sources.25 Contrary to

280

EARLY JAPANESE TRADE

belief that pre-1868 trade figures had survived in the finance ministry of later times and were lost in the earthquake in 1923, the ministry had in fact inherited little from the Kanjōsho.26 The surviving and unique original satsu on the Chinese trade of Nagaski in 1709–1714 are not as Yamawaki stated and Nagazumi and Nakamura repeated, papers from central archives in Edo.27 They are original Nagasaki documents, collected into the sort of maverick compilation typical of the Edo era. In this instance they were collected by an unknown individual into twenty booklets, the satsu for individual cargoes (of which six original texts were created and signed by six individual officers for each cargo). Some of the twenty booklets containing the satsu were later lost and some individual satsu are missing within the surviving booklets. They were later bound into four kan by an unknown individual.28 Unfamiliar with the satsu, he omitted a key word from their title.29 An identical omission occurs in a separate document, Nagasaki goyōdome 攟ġⲶġ⽉ġ䓐ġ䔁ġ, compiled from several sources to fill the many gaps in 1711 in vessels numbered from 17 to 54.30 Though there were several individuals involved in compiling this document (in the writing there are several hands), repetition of the omission suggests that a sole person may have been responsible for both compilations. The two compilations, one binding originals together, the other a straightforward transcription of documents, finally reached the National Archives. The first was from a very unlikely source (Nōshōshō kyūzō 彚ġ⓮ġ䚩ġ㖏ġ哝ġ, the former archives of the Agriculture and Trade Ministry), the other from an unidentified one. Further illustrating the random pattern of dispersal, Kaisho Ể㇨ġ repertories (hikae mokuroku ㍏䚖拚) of its records for the year 1719 ended up in the Ōmura Municipal Library.31 Finally, 163 files dealing with individual cases or jiken in the Ansei ⬱ġ㓧ġ period form a small part of the Koga zōsho ⎌ġ屨ġ哝ġ㚠, a collection assembled by Koga Jūjiryō ⎌屨⋩Ḵ恶ġ (1879–1954), one of the main Nagasaki collectors of documents of Tokugawa times.32 While no primary record of post-1715 licences to Chinese vessels survives, there is also in the Koga collection a shahon containing them.33 Its status— whether a copy in official archives in 1868 or an earlier private copy—is not clear. The records of the interpreters, for whose bureaucratic effectiveness the Dutch had a high regard, had also been well kept. In 1704, the interpreters were reported as looking at their archives as far back as 1681.34 In May 1744, the opperhoofd somewhat dramatically recorded

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

281

that the interpreters had been ordered “to draw up an exhaustive report about all the goods the company had imported during the last twenty-nine years. Toeksemon (sic) told me that fourteen clerks were busy day and all night copying their records concerning our imports.”35 For the interpreters of Dutch there is now scant material apart from what has been handed down in two interpreter families, the Nakayama family (now in the Siebold Museum in Nagasaki) and the Motoki 㛔ġ㛐ġfamily (shared between a museum in Nagasaki and an art gallery in Kobe). For the Chinese trade, though a substantial quantum survives for both Kaisho and interpreters, it is a small part of what once existed.

Figure 1. Chinese vessel unloading into flat-bottomed boats (sampan). Vessels were moored in the open bay close to shore; speedy unloading was important to admit replacement of the cargo with ballast to stabilize vessels in what were at times stormy waters. (Courtesy of Nagasaki Museum of History and Culture.)

Ironically, we know more about the recovery of material in and beyond Meiji times than about the losses in 1868.36 Apart from large batches mainly of bakumatsu papers, records in Nagasaki seem to have been either abandoned haphazardly or even given away at the time. The clearest single illustration is the primary series of hankachō 䉗ġ䥹ġⷛġ(criminal investigation records) complete from 1666 to 1867. No longer serving a purpose, as the bugyōsho office had ceased to exist, they were given to the police, who, likewise finding no use for them, sold them to an antiquarian dealer. The first and pioneering

282

EARLY JAPANESE TRADE

savior of records was Kanai Toshiyuki 慹ġḽġὲġ埴ġ(1850–1897), a ward official who saved many documents.37 Held in his ward office at his death, they remained in city possession to finally find a place in the City Museum (opened in 1941). Other collections were to find a home in the prefectural library, founded in 1912. Koga himself in 1919 gave it two satsu of the hankachō; on his death, his own collection of papers went to it. Documents collected by Watanabe Kurasuke 㷉形⹓庼, a local historian and writer, were deposited as late as 1964. The vulnerability of papers is well illustrated in the story of the surviving nikki for 1663 to 1715, which post 1868 may have ben held by the families or deseendatns of the former interpreters of Chinese, They had a close association with the Seidō 俾ġ➪ġ(the Confucian templea under the care of the Mukai ⎹ġḽġfamily, hereditary heads of the Seidō), which may explain how the Mukai were later to hold copies of two nikki. By the time Kanai began his salvage work, only ten volumes of nikki for the period survived. Presumably because he failed to acquire the originals, they were copied in 1886–1889. One was later lost. From the ward office (and successor city offices) seven finally reached the City Museum. For reasons which are obscure two copies were held by the Mukai, passing in 1934 to the prefectural kyōiku kai 䚴ġ㔁ġ做ġỂġ, then in post- war years to the prefectural library and finally to the museum in 1959.38 TRADE: COMPOSITION AND STRUCTURE

The composition of trade is clear. Sugar gradually replaced silk as the major import. Figures are not readily or, indeed, at all available for many items. But many of them served a vital need, enjoying official encouragement. Ginseng (ninjin Ṣġ⍪ġ, especially from Korea), much sought after for its medical properties, and sandalwood (kō 楁ġ ), the source for fragrances used in domestic and formal settings, were outstanding items, and to them should be added many other pharmaceuticals.39 The importance of this trade was already recognized in loans granted to Chinese shipowners under Tsunayoshi 䵙ġ ⎱ġ (1680–1709) and which were a subject of repayment years later under the frugal Yoshimune ⎱ġ⬿ġ(1716–1745).40 In April 1738, with the supply of Chinese medicines short and the good will of the Dutch being sought to encourage the Chinese traders in Batavia to respond to Japanese needs, the opperhoofd of the Dutch factory

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

283

in Dejima noted his offer to supply these goods (were the Dutch allowed to import them): “I concluded that the Japanese cannot do without the Chinese imports.”41 Two features have an interest out of the ordinary, even if hardly of consequence in volume terms. The first arises from the concern felt by the Nagasaki authorities about books in Chinese, leading to their listing or inspection on arrival.42 The second is the analysis in recent times in remarkable detail by Professor Ishida Chihiro 䞛ġ䓘ġ⋫ġ⮳ġ from the Dutch sources of the categories of kenjōhin 䋖ġᶲġ⑩ġgifts by the Dutch for the shogun; shinmotsuhin 忚ġ䈑ġ⑩ġ, gifts for high officials; and atsuraemono 娪ġ䈑ġ , goods ordered by or for the shogun and by Edo and Nagasaki circles.43 The latter goods included cannons ordered by the famous artillery expert Takashima Shūhan 檀ġⲞġ䥳ġⶮġ (1798–1866), a subject studied in much detail by Ishida. Exports on Dutch vessels were in a very narrow range, but camphor was sought after and at times the supply fell well short of Dutch demand. It was, next to metals, their major export. Unlike Europe, a multilateral trade, in which surplus earnings in trade with one country were already helping to bridge a gap between imports and exports in other areas of trade, did not exist. The absence of such a structure underlies the thinking of Japanese officials. While their emphasis in political terms was more on imports than exports, they saw the figures for exports and imports as broadly identical. In Japan imports and exports were siamese twins, and trade was effectively a single protracted operation spread over months. At the end of the process, there was neither surplus nor deficit to carry foward, with the exception until the 1680s of an amount of silver denominated as tsukaisuteginࠉ㐵ᤞ㖟import income by the Chinese and Dutch not converted into exports, and subject to approval available for other purposes and notably the living expenses of the Chinese).44 It was effectively the balance of trade. Originating in retained income from the proceeds of imports, it should, in theory at least, if substracted from imported commodity figures, give us the total amount of exports. With the termination of tolerance for tsukaisutegin in a silver crisis in the 1680s, the Chinese had to cover their expenses in other ways, including bringing funds with them for this purpose. Gross trade figures, given the extremely high value of the main items (silk and silver), greatly overstate the physical size of the trade

284

EARLY JAPANESE TRADE

in Nagasaki. A small tonnage contrasts with a huge tonnage in Osaka’s coastal traffic in rice or in Europe with the massive trade in basic goods such as minerals, wood, wine, and grain. The 29,314 kanme ㈏┠ of exports by Chinese for 1661, if converted into sterling currency, amounted to £685,068.45 This figure is one and a half times the exports in 1665 of Ireland, a country with good statistics for the 1660s and, through its colonial status, a highly developed trade. Yet it was, in terms of tonnage, a small traffic carried in 1661 on a mere thirty-nine vessels. During the Ching or Manchu challenge to the rule of the Ming dynasty that terminated only in 1684, the Ching dynasty prohibited foreign trade with the result that for several decades traders were Ming loyalists rather than traders from Ching-controlled districts, and disproportionately from Chinese communities in Southeast Asia and Siam. With the final triumph of the Ching in 1684 admitting a resumption of trade, the number of ships from the central reaches of the Chinese coast rose sharply from 1685. The peak was 192 in 1688.46 In response, the Nagasaki authorities limited incoming vessels to seventy (later briefly raised to eighty), though many of these vessels were refused permission to land their cargoes. For concrete information about individual vessels, the sole source is the fūsetsugaki 㢼ㄝ᭩ submitted from the 1640s by vessels on their arrival in Nagasaki.47 Declarations provide some commentary on cargoes and the difficulties in sales.48 While for vessels from China silver had always been the most sought-after item, the statements by the Chinese of Batavia in 1685 made a point of stressing that their return cargoes had been in goods, not silver.49 The fūsetsugaki do not provide information on merchants, who must have been aboard as passengers. They remain a shadowy group.50 They were probably small-scale operators, their operations somewhat augmented by petty speculations by sailors. The issue of new Japanese licences in Chinese from 1715 posed a problem for Chinese merchants who had traded in earlier years but had not been in Nagasaki at the time of their issue in 1715. It affected “all together about fifty merchants, some of whom have traded to Japan for several years, as well as sailors . . . and are now experiencing financial hardship.”51 Licences were issued to vessels while in Nagasaki, and on their next voyage had to be produced for permission to enter the port; without them vessels were turned away.

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

285

Tonnage figures for individual vessels are rare.52 However, according to Souza, junks were usually 120 to 220 tons burthen, small vessels or wankans 30 to 150 tons.53 Japanese ships, which of course were prohibited from trading overseas, were limited to a capacity of 500 koku (roughly seventy-five tons). Fisscher, a warehouseman in Dejima 1821–1829, noting that vessels from Ezo were the largest, observed that they were capable of taking a cargo of sixty tons while leaving generous room for passengers and crew.54 The Ryukyu trade was free from this restriction. In the 1790s, a British officer had noted in Naha “twenty large junks” at anchor, from two to three hundred tons.55 Vessels from the Ryukyus and Satsuma were said in the 1860s to have been somewhat short of twenty in number.56 For some purposes the value or at least estimates of cargoes are more useful. For cargoes in the 1710s, the most common estimate was 200 kanme, and in the 1720s even lower.57 In contrast, the value of goods on the nine vessels arriving in 1803 was 5,800 kanme, or an average of 644 kanme per vessel.58 Such cargo values were a new norm: the figure is even close to the value for cargoes on the annual Dutch vessel. The Chinese trade had become more ordered, and individual cargoes more valuable (with the continued rise in pharmaceutical products) and varied (including from 1763 silver from China), and the merchants fewer and more prosperous. While Dutch trade had contracted to about 1,000 kanme by the end of the century, Chinese trade in 1803 or 1804 was significantly higher than it had been in the 1720s. Exports other than metals, once a weakness of trade, expanded to fill the gap and were central to the trade. Regular, almost daily official harassment of the Chinese was no longer a feature, though the authorities could act decisively in 1825 or 1835–1837 in the face of problems. TRADE STATISTICS59

From 1648 figures for trade existed (the starting source of the figures on metals later supplied to Hakuseki).60 With the exception of export figures for 1648–1672 (from the TKIR), figures are few.61 There is a run in Iwao for 1690–1700, and in Yamawaki for 1704– 1711 (source not directly indicated but on the evidence of references elsewhere apparently from the diaries of the Chinese interpreters).62 In Nakamura, there are imports for 1715– 1726 from a shahon, Shinpai kata kiroku ⅁㛔ᾉ 䇴㕡姀拚.63

286

EARLY JAPANESE TRADE

Table A. Exports, Chinese and Dutch, including silver exports (averages). kanme

kanme

Chinese vessels years (average)

vessels exports of which (no) silver

Dutch vessels balance*1

vessels exports of which (no) silver

1648–1654

49.4

10,694

5,199

2,293

6.6

6,054

5,139

1655–1661

48.6

17,786

12,671

1.074

7.6

6,767

5,008

1662–1668

36.4

12,734

8,322

2,167

8.4

8,719

4,940*2

1669–1670

37

12,825

345

3,021

5.5

10,665

—

*1. Tsukaisutegin. Silver retained by Chinese (and Dutch) from import income. Ōta includes it in the export total; Iwao correctly excludes it. The Dutch balance, while deducted from Dutch exports above, is not particularized in the table. *2. Average of six years.

Table B. Composition of combined Chinese and Dutch exports (averages).*1 years (average)

Total*2

silver

1645–1654

16,634

1655–1661

24,551

(in kanme)

gold*3

copper coins*4

merchandise*5

10,338

229

6,296

17,679

341

6,872

*6

1662–1668

21,453

12,557

(7,769)

434

5,563

1669–1670

23,989

345

17,330

530

6,314

Chinese vessels only 1671

11,815

950

5,931

3,934

1672

11,729

8,964

9

2,756

*1. Slight variances between the grand total and totals for China and Holland in Table 1 above. *2. Net of Chinese and Dutch tsukaisutegin. *3. Value of gold in kanme. *4. Figures for 1662–1668 and 1669–1670 include exports by the Dutch, not available for earlier years. Though figures appear in Ōta’s tables, they are excluded from his grand totals. *5. Merchandise excluding metals apart from copper. *6. Average for three years, 1666–1668. Individual years were as follows: 1666: 2292 kanme; 1667, 4,322 kanme; 1668, 16,784 kanme. A sum of 29 kanme was recorded for 1664.

The kan itself was a unit of weight, and silver passed hands in coins or bars in terms of its weight (a kan of silver weighed 3.75 kilograms); kanme signifies silver money of account (with a variation in

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

287

its purchasing power in proportion to commodity prices as they rose or fell). For commodity exports in an upsurge from 1658, especially by Chinese vessels which continued into the mid 1660s, a total for Chinese and Dutch trade combined of 36,982 kanme in 1661 was the peak.64 The exports of silver, comprehended as a commodity within the total, peaked in the same year at 31,313 kan. These years were the high-water mark of the trade, both Dutch and Chinese. The silver exported was currency standard chōgin ᶩ戨ġ (silver bars with purity of 80 percent). Exports in 1650 were 6,827 kan of chōgin plus haifukigin 䀘⏡戨ġ (pure silver) and gindōgu 戨忻℟ġ (silverware). The amount of tsukaisutegin was 3,178 kan, and it remained substantial at least into the early 1670s. As silver exports were at that stage reduced and virtually ended in the mid-1680s, tsukaisutegin (silver balances or at least entitlements to silver, which in the case of the Chinese could also provide a cover for exporting silver) likewise ceased to exist. The composition of exports underlined the vulnerable nature of the trade. Commodity trade in items other than metals were as little as a mere quarter of the total. Changes over the 1660s in regulating exports of silver, copper, and gold reflected the falling output of metals. 65 Given a sense of crisis, exports of silver permitted to the Chinese were halted for almost three years before resuming at lower levels than in the past. In the years 1669–1671, silver to Chinese traders (now the only permitted outlet for silver) averaged a mere 547 kan. The short-term termination of silver exports was compensated for by dramatically increased exports of gold. In the very short term, that made it possible to maintain total exports by the Chinese and Dutch at a high level. CONTRACTION AND RECOVERY: THE COURSE OF TRADE CA.1672–1708

The need from the 1660s to conserve for home use the declining output of silver finally accounted for a notional ceiling to the value of trade in 1685 (6,000 kanme as the ceiling in silver money of account, for the Chinese trade, and 3,000 kanme for the Dutch). Already from 1668 the Dutch were no longer permitted to ship silver, but on the other hand were allowed a generous ceiling in gold (50,000 ryō, the equivalent of 3,000 kan of silver).66 The celebrated limit of 6,000 kan applied both to silver (the metal itself in kan weight),

288

EARLY JAPANESE TRADE

and also to the total value of trade as measured in silver money of account (kanme). The original ceiling for 1685 rested pragmatically on the fact that exports of silver to China were around 6,000 kan in 1682–1684. However, given a persistent quest for silver and a shortage of the metal, exports of silver became nominal well before 1697. With silver progressively reduced, the ceilings lost their original significance, becoming mere orders of magnitude to guide the management of trade. For the two decades from 1672, a shortage of trade figures makes some conjecture unavoidable. Exports of silver, for which figures do survive, averaging 6,184 kan in 1672– 1684, were roughly a quarter of the level of the 1660s. A crude estimate of trade can be made for 1690–1694. Table C. Exports to China 1690–1694 (average). (in kanme)

*

copper

4,123

non-metal exports*

2,443

grand total

6,556

There are no figures for the total of non-metal exports in 1690–1694, but a rough estimate is possible if we calculate the figure for 1663–1672 and then proceed to assume that it remained constant in later years. While copper and non-metal products were combined in a single rubric in the surviving trade abstract for 1663–1672, the known figures for copper, once converted into kanme averaging 798 kanme, when deducted from a total for copper and non-metal products of 3,241 kanme left a residue of 2,443 kanme as an estimate of non-metal exports in 1663–1672.

A value of 5,910 kanme for the China trade in 1693 is a confirmation of sorts of this crude arithmetic.67 In other words, post-1672 trade was probably static. Despite post-1672 stagnation, trade acquired a real momentum in the second half of the 1690s. A policy from 1697 of the bartering or exchange of copper (shiromono gae ẋġ䈑ġ㚧ġ ) intended to encourage imports and guaranteed a supply of copper within a ceiling set at 5,000 kanme.68 However, this facility itself accounted for a mere 42,000 piculs in 1697, well short of the total exports of 89,081 piculs of copper.69 In a table by Yamawaki for the years 1704–1711, exports were close to 12,000 kanme in two years, and above 12,000 in another two of the five years from 1704 to 1708. These high figures in turn explain a high level of silk imports before 1709.

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

289

Table D. Trade with China, 1704–1717.70

*

Exports to China*

Imports

kanme

kin

Exports

Barter copper

Raw silk

1704

12,524

5,000

84,250

1705

7,625

2,670

38,525

1706

12,430

4,763

44,460

1707

11,859

4,100

70,970

1708

11,220

4,260

81,830

1709

4,561

2,720

23,850

1710

7,163

2,825

23,859

1711

4,794

600

50,276

Export figures include the barter trades in copper and in marine products.

To 12,000 kanme should be added a figure for exports to Korea (in some ways simply an extension of the traffic in goods to and from China), negligible in the past but in 1684–1710 averaging a subsantial 2,977 kanme.71 If, with much optimism, the fluctuations in the supply of copper from Osaka and Sakai are disregarded, a figure of 15,000 kanme makes it easy to see how political and commercial pressures in favor of imports of silk, ginseng, and medicines, all seen as vital, prevailed. Effectively this lobby ruled out the alternative strategy of simply cutting imports. The lobby, encouraged by the sharp rise in exports of copper in the late 1690s, even pressed for an increase in permitted imports.72 In 1698, Chinese exports of copper were 60,824 piculs and in 1708, 66,040 piculs or 7,204 kanme and 7,862 kanme respectively. A crude picture of the expanded export trade would emerge as follows. Table E. Estimated exports in 1698 and 1708. (in kanme)

1698

1708

Copper to China

7,204

7,862

Other products*1

2,500

2,500

Dutch exports*2

2,228

1,541

Total

11,932

11,903

290

EARLY JAPANESE TRADE

*1. This is a round figure for exports. The barter trade other than in copper was set at a value of 2,000 kanme, and averaged 2,389 kanme in 1710–1712. It was mainly in marine products. Exports appear to have been in two categories, tawaramono ᾝġ䈑ġ (bagged goods) and shoiro mono 媠ġ刚ġ䈑ġ(various goods). Tawaramono were iriko 䃶㴟滈ġ(sea cucumber extract), hoshika ⸚毗ġ(dried sardines), fukahire 氞毕ġ(shark fins); shoiro mono were goods such as shiitake 㢶勠ġ (mushrooms) and marine goods such as surume 残(cuttle fish) and konbu 㖮ⶫġ(tang). *2. An import figure in the absence of export figures, hence an imperfect replacement.

The buoyant trade in wares to and from China contrasted with the Dutch trade, which was static or falling after 1700. If the Tsushima trade (ca. 3,000 kanme) is added to the figure for exports to China the short-term buoyancy of Chinese demand is all the more evident. THE ROLE OF THE KAISHO

An increasingly complex policy was operable only through the establishment in 1698 of the Kaisho (translated as the expressive Dutch word, geldkamer or cash office).73 This office provided the machinery in Nagasaki to execute a closer scrutiny of what was already a highly managed trade. The emphasis in the office’s work was on accounting operations, not on the physical transfer of goods. An opperhoofd was to observe in 1717 that, “It is a strange way of doing business. First we have to sell the goods and only after the goods have been sold they are inspected by the merchants.”74

Figure 2. Dejima Island and structures (warehouses, residences, and the office for the interpreters). The entrance and exit by the bridge was guarded all day everyday. (Courtesy of Kobe City Museum)

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

291

Silver had not been prohibited in the letter of the law to the Chinese in 1685. For the thirteen years from 1685 to 1697 exports of chōgin (silver metal of currency standard) averaged a mere 229 kan.75 In 1687 the figure was 465 kan; in 1693 it was a mere 2.5 kan of silver in a total 65 kan mainly of gindōgu silverware.76 Exports of silverware averaged 77 kan in 1685–1697. From 1699, total silverware was limited to 100 kan a year, and its export was finally prohibited from 1708. From 1699, the export of silver itself was limited to minute quantities per vessel. For four years in 1718–1725, for which actual figures exist for some vessels, it was of the order of 2 kan per vessel, and in one year, 1718, quantities were somewhat more generous, in one instance actually being around 12 kan. From 1733, no more than 950 monme (that is, slightly less than one kan) per vessel was permitted. Between 1709 and 1762, shipments averaged 23 kan. Shipments were totally stopped in 1763, and Japan itself became an importer of silver.77 The office ran into difficulties in its management of the copper trade. There was the unceasing challenge to ensure a supply of copper to meet Chinese demands, eased only by a fall in Dutch exports of copper. The concept of an orderly barter of copper in exchange for imports was breaking down and, for want of copper, Chinese cash surpluses began to emerge. The barter traffic was recorded in separate files, that is, not entered in the general files. (In the 1709–1714 returns, the files are extant only for Dutch cargoes.) The consequence shows in Yamawaki’s count from the surviving satsu of barter copper exports on Chinese vessels for 1711: a mere 17,547 piculs as against the actual Chinese exports that year of 42,579. THE COPPER SUPPLY, 1698–1715, AND THE SHINREI OF 1715

The export of copper peaked in 1697 and 1698 at 89,081 and 90,202 piculs respectively.78 It then fell sharply in 1699 and 1700 to the lowest level since 1694. In the first decade of the new century, quantities were to oscillate wildly. In February 1700, unprecedented permission had to be conceded to the Chinese “to leave large balances in Japan and settle these next year because there is not enough copper to export.”79 To keep the balance down, the authorities in some desperation conceded permission to export silver. The amount for the year was 1,085 kan.80 Much, although not all of it, may have been in the form of silverware. The Dutch in February made a note of “secret” information that the Chinese were being permitted to buy silverware

292

EARLY JAPANESE TRADE

for want of copper.81 However, there was a remarkable upturn in total copper exports in 1704–1710 to figures of between 64,000 and 74,000 piculs. Chinese exports in 1708 were second only to the peak figure of 1696. The year 1708 was the last buoyant year. In early 1709, copper was already in insufficient supply and the last two junks of the season left belatedly in March.82 Total exports of copper by both the Chinese and Dutch fell in 1711 to 52,578 piculs and in 1712 even more sharply to 37,701.83 In June 1714, some vessels were again said to have been in Nagasaki for ten months. They were finally given permission to depart, leaving some of the income from sales behind them—in a repetition of what had occurred in 1700—to be used for the purchase of ink fish in the following year’s sales.84 Strict adherence to the ruling is seen in the Chinese exports of copper in 1715 of a mere 7,637 piculs. In contrast, Dutch exports at 11,500 piculs conformed closely to their level of the preceding year. When the crisis in copper had to be faced, the strength of the political lobby that favored imports can be seen in the terms of the Shōtoku Shinrei 㬋ġ⽛ġ㕘ġẌġ or new decree of 1715 regulating trade.85 An impasse between optimists and pessimists was mirrored in the wavering of the ceiling level set in the first years, and in high exports. Opposition to cutbacks existed even among the rōjū, and was ended only on Yoshimune’s ascent to office in 1716.86 The new decree was, in essence, a response to the copper problem, which had replaced the silver crisis of an earlier generation. The bureaucratic and discredited shiromono gae was terminated, and replaced by a simple quantitative restriction of copper to 45,000 piculs. It was not in the circumstances ungenerous: the figure was almost identical to the level of exports in 1713 and 1714. Chinese vessels, however, were henceforth to be limited to thirty, under a new and tight licencing system (with licenses granted only to vessels already in the trade); Dutch vessels to two. The promises of trade ceilings being maintained and of a reduced but relatively generous ceiling for copper were one matter, but reality was another. The protracted time Chinese vessels still had to remain in port in 1718–1720 was an indication of reality.87 The symbolic notional ceiling for the Chinese trade was reduced from 6,000 kanme to 4,000 kanme for the year 1720. More concretely, the ceiling of 45,000 piculs of copper was reduced to 30,000 piculs (20,000 for the China trade, 10,000 for the Dutch).88 Aside from an interlude in the early 1740s, an attempt was made until 1791, though

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

293

with variable success, to adhere to the 1720 ceilings. It was the regulations in 1719–1720 that finally marked a change from a policy which favored foreign trade to one of contracting its scale. The average value of imports in the 1720s was the lowest on statistical record.89 The nominal ceilings were gradually reduced, replaced in 1791 by still lower figures, which then remained unchanged until the time of the opening of the ports in 1859. In 1791, the nominal ceiling of Chinese trade was set at 2,740 kanme, the Dutch trade at 700 kanme.90 The number of vessels was limited to ten Chinese vessels and one Dutch. In the case of copper the effort by the authorities to honor the promises of 1720 was finally abandoned, and copper to the Chinese and the Dutch alike entered a decidedly downward trajectory. In 1839, Chinese exports of copper, now a minor constituent of the export trade, were a mere 920 kanme (that is, less than 8,000 piculs). POST-1720 EVOLUTION OF TRADE

In contrast to an earlier story of success in maintaining trade at, or even above, the level of 1685, the trend from 1720 was firmly downwards. Japan had little in quantity to offer in the short term that foreigners wanted other than metals. Post-1690, copper remained the crucial element in maintaining a Chinese and Dutch presence. As for imports of silk, they began to fall as early as the 1660s and contracted sharply in the early 1700s. Even imports through Tsushima, now the main channel for silk, contracted. Sugar, already an import and becoming the main one when silk waned, was a relatively low value product and did not attract merchants to Nagasaki as silk had done. Nor did it provide the profit that silk did at its peak. The ceilings for Nagasaki trade do not include the permitted ceilings for trade with Korea (1,000 kanme) and the less clear or changing ceiling for the Ryukyus (variable but not above 1,250 kanme).91 The Tsushima trade was favourably singled out, notably by its access to silver, when silver exports from Nagasaki had virtually ended, and in 1710–1714 specially minted silver coins made to the former standard were provided. The reason for this favor was almost certainly the fact that this trade was tightly controlled by a small milieu in Tsushima and a factory in Pusan, and the belief that abuses at the two ports could be monitored or prevented more easily than in Nagasaki. Atypically, trade figures are known because of the survival of an account book into which an unknown official of the Wakan ῕ġ棐ġ copied in 1716 details

294

EARLY JAPANESE TRADE

of the trade for 1684–1711.92 The scale and importance of the Tsushima trade has been a subject of controversy mainly over the accuracy of calculations, but the general outlook is not in doubt.93 While raw silk imports in the Korean trade with Tsushima for a time exceeded those of Nagasaki, they were in sharp decline in the first decade when the silk import trade at large was declining.94 It is not in doubt that the general trade of Tsushima never came near to equaling the Nagasaki trade. It flourished for a decade either side of 1700; it later tapered off and in time altogether withered. The Satsuma trade unfortunately lacks documentation comparable to that of the Tsushima trade. Its profile is obscure, and its scale rests on a belief, lacking in quantitative terms, of widespread smuggling. The fact that copper had been necessary to attract the Chinese and Dutch for a long time had protected copper from further decline. As far as the Dutch trade is concerned, the most obvious cutback was in gold, which was reduced to a negligible amount, and in camphor, much sought after but where their demands often were not met. The Dutch trade entered slowly into a permanent decline. In contrast the Chinese trade staged a recovery, at first shadowy because of an absence of statistics but very evident in trade figures for 1803 and 1804. The trade had acquired a new dynamic. Though the number of ships was small, the value of cargoes was much higher and the export trade, laggard apart from metals in the past, shared in the dynamism. The trade in real value was now well above the post-1791 notional ceiling of 2,740 kanme. Exports were 7,345 kanme in 1804 and 7,034 kanme in 1839.95 In something of a parallel to the pre-1715 shiromono gae, imports of pharmaceuticals were encouraged provided the outlay was matched by exports of marine products.96 The cutbacks in copper and in silk imports combined with the limitation in 1715 of Dutch vessels to two a year was the decisive feature in accounting for the reversal of the city’s fortunes. The complement aboard (on a single vessel up to one hundred men or more), in combination with the long delay in getting return cargoes, resulted for decades into the eighteenth century in a transient population of several thousand Chinese in Nagasaki. With its China trade gradually reduced, the fears of the Dutch abandoning Dejima often recurred, and in the years in which either one or both Dutch vessels failed to appear, poverty was rife. The port’s population, at a peak in 1696 of 64,524, was halved by 1789. Officials and their dependents accounted for about a third of the population. In other words, it was a company

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

295

town, and from the early eighteenth century, the reduced traffic added to the penury characteristic of all direct employees of the shogunate. Adjusting to altered circumstances over the eighteenth century, dynamism finally reemerged in the trade to China. With vessels from the southern areas in Southeast Asia declining, trade now centered on ports in the central reaches of the coastline. By the 1760s this region (Nanjin, Ningbo, and Shanghai) accounted for 90 percent of shipping.97 This contraction was in part a consequence of a direct involvement by the Chinese government in securing its supply of copper. Management was entrusted by the Ching government to the Bō 勓ġ family who, from 1735 over several generations, became responsible for securing copper on behalf of the Ching government.98 Hao has described it as a soshikika 䳬䷼⊾(systemization) of trade.99 Much on the trade was on contract. The Bō family was provided with loans from the Ching authorities with this object in view. Control at both ends (in China for copper, in Nagasaki for pharmaceuticals) gave the trade a new stability. An added support was the import of silver on the account of the shogunal authorities from 1763.100 Imports linked to the ever-pressing demand for medical/pharmaceutical goods ensured that far from trade stagnating, it began to expand. The buoyancy may explain why Satsuma tapped into the Chinese trade of Nagasaki. A few of the Chinese vessels licensed to trade in Nagasaki unloaded irregularly on the coasts of Satsuma with the consent of the Satsuma authorities. SMUGGLING: HOVERING CHINESE VESSELS ON THE COASTS AND UCHI HARAI.

Smuggling looms large in the story of trade. The secretive nature of smuggling, readily assumed in modern writing to have been large, has usually led in Europe and Japan to a great overestimation of its scale. Japan lacked the high duties which, in Europe, led to a concentrated and highly organized international business in a small number of commodities. It has been argued that the reduction of official trade from the 1710s provided a fillip for its expansion in Japan.101 There was both small-scale activity and larger ventures. Smallscale activity was driven by the sheer number of Chinese vessels present in Nagasaki in the late 1690s and early-eighteenth century, and by the problems both in marketing a surfeit of imports and in securing either goods or metal for the return journey.

296

EARLY JAPANESE TRADE

With no high duties to evade, Japan’s concerns in the Genroku about smuggling centered on the outward smuggling of silver. The administration was obsessive to the point of hysteria often over minute quantities, with savage penalties for Japanese who became partners of Chinese. The stay of the Chinese was worsened by a harsh and arbitrary administration, long delays, and an uncertain outcome to their demands for return cargoes. Some official actions operated outside the realm of the judicial processes recorded in the hankachō. At times the Chinese were required to strip naked before they boarded their vessels. The “island” was subject to searches, and when vessels were being loaded, streets in the proximity were closed and policing was redoubled. For the Chinese, many of whom were small or marginal traders, some mere sailors, a demand for silver reflected less a large secret trade in silver imagined by officials than an often fatal hope of concealing minute quantities as the reward for petty speculations ashore. The opperhoofd in his dagregister often expressed the opinion that a profitless trade was driving the Chinese in desperation to resort to smuggling. A frenzy in punishing smugglers in the wake of the closure of silver exports was recorded in Kaempfer’s account of his stay as medical doctor in Dejima in 1690–1692.102 This changed in time as ships became fewer and minor abuses became less pervasive. When a skipper was allowed to board his junk in 1738 without his person being searched, the opperhoofd noted that “The Chinese are treated better than before.”103 By the middle of the century the travails of the Chinese had begun to lose their former painful prominence in the dagregister. ⃫ġ䤬ġ era

Figure 3. “To the left of the fort on the summit of the forest in a location called Kosezaki are dwelling houses; proceeding from that point and facing them the residence for the Chinese stands out [Recent miscellaneous notes]”. (TKIR zokushū, vol.1, p. 136.)

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

297

Official concern about larger ventures had already emerged before 1715. In I714 in the wake of a request by Osaka merchants, the shogunate issued orders to the bugyō of cities and to daimyo. The arrest of some smugglers in Osaka led to forty-one citizens of Nagasaki being named as accomplices, and later, after three days of house searches in Nagasaki, seventeen arrests were made.104 Smugglers required access to established distribution channels. They neither bartered nor bought goods for their return journey. Hence they depended on contact with businessmen with capital or at least cash. Some of the ringleaders were said by the Dutch in 1713 to be individuals who had settled in Shimonoseki and Osaka. For these reasons, Chinese vessels hovering (hyōryū 㺪ġ㳩ġ, “drifting on the coast”) were rarely to be found beyond the coasts of Kyushu from Nagasaki to Kokura, coincidentally also close to the established routes of Chinese and of the much rarer Korean vessels.105 Where goods appeared in Osaka or in northern Japan, as far afield as Kaga ≈屨ġ or Matsumae 㜦ġ⇵ġ , they originated in Kyushu. In general, goods legal and illegal alike were moved by established coastal shipping, its scale measured somewhat in a report in June 1713 of some thirteen vessels with “Chinese goods” being then in Osaka.106 Inevitably, Satsuma vessels, active participants in coastal traffic, were drawn into it. While in 1772, on the evidence of the hankachō, goods were loaded into vessels which had travelled from the north of Japan,107 Satsuma vessels equally travelled far afield. One of the attractions of engagement was the use directly or indirectly of Chinese goods as payment for the marine products of Matsumae, especially iriko sea cucumber. The exchange centered on the province of Echigo 崲ġ⼴ġ, especially at Niigata 㕘ġ㼇ġ. The shipwreck in 1835 of a Satsuma vessel loaded with Chinese medicines was much commented on in official reports. While bugyō belief in 1835 has an exaggerated tone, it saw an exchange of Chinese goods for marine products as responsible, via the Ryukyus, for an alleged decline in the quality and quantity of marine goods arriving on Nagasaki for its own China trade.108 Chinese vessels hovering off the coast of Kyushu aimed to land goods at particular locations, counting on contact with local merchants. There were two official responses to this challenge. The first was to demand identification of the Nagasaki origins of Chinese goods both in Kyushu and in the final market destination. As early as 1714 the opperhoofd recorded that, “Even if the Chinese do succeed in smuggling goods into the country, it has become impossible for the merchants to

298

EARLY JAPANESE TRADE

sell them since they have to say from whom and where they bought them.”109 Preventative measures had some teeth. In Osaka in 1718, two thousand pieces of smuggled loincloth and three thousand Persian fabrics were identified and confiscated.110 Preventative action was even more important at the local level in Kyushu. As early as 1717 in Saga and Chikuzen, the buying and selling of Chinese goods were under observation. Of the merchants in Kokura ⮷ġᾱġ only one was licensed to buy Chinese goods and from not more than one of four designated Nagasaki houses.111 Saga was in 1763 prohibited from handling Chinese goods with the exception of pharmaceutical products.112 The second step was to ensure within Kyushu a vigorous coastal watch when hovering vessels were spotted. The policy of uchi harai developed in response to their presence in the region.113 The domains mainly concerned were Kokura and Fukuoka. The shogunate itself in 1718 dispatched a metsuke 䚖ġẀġ̒ġ from Edo to Kokura to direct the chasing off of Chinese vessels.114 The opperhoofd was prompted to observe that, “The smuggling trade must be very profitable since the Chinese do not seem to be afraid of the Japanese musket balls.”115 In 1726 when two vessels off the coast at Shimonoseki refused to depart, the Japanese opened fire, killing several crew.116 As for the scale of the traffic, the market for most goods was finite. The exception was the insatiable market for pharmaceutical products. Traders, too, both in Nagasaki and in Osaka, claimed that slow sales were caused by a flood of goods. The sparse evidence suggests, however, that ventures were episodic rather than sustained, and that smugglers had to count on established intermediaries. In other words, unlike Europe where trade depended on the irresistible appeal in high-tax regimes of a handful of smuggled consumer goods (passing through parallel markets), smugglers in Japan did not have the benefit of a network of their own. The Ryukyus and Satsuma stood in a special position. Unlike the Gotō ḼġⲞġ islands, sometimes seen as a haven for smugglers, Satsuma did not openly welcome Chinese smugglers. It had a lucrative market in Nagasaki, and had no wish to undermine its trade at large by free and ready access for vessels to its shores in rivalry with its own shipping. The balance implicit in the bakuhan taisei (system of shogunate and domains sharing administration of Japan) is relevant. The Shogunate and domains shared an interest in coastal protection and the domains of Kyushu followed a policy of uchi harai. Ships were fired on and, as often reported in the dagregister, vessels arrested off

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

299

Satsuma’s coasts were brought into Nagasaki by escorts. In contrast to this harmony about hovering vessels, once one turns to regulation of the trade of the Ryukyus, the interest of shogunate and Satsuma diverged. The shogunate, especially under the long reign of Ienari ⭞ġ㔱ġ (1787–1837), valued a good relationship with Satsuma more than did the officials in Nagasaki with an exclusive interest in the port. From the outset, in the seventeenth century Satsuma’s Ryukyu trade had been limited by order (eased on occasion when Chinese goods were in short supply) with the aim of preserving a near monopoly of trade for Nagasaki, and excluding most Ryukyu goods from direct sale by the domain to other markets.117 Regulations had set ceilings for individual goods rather than for an overall valuation of the trade. Given variations over time and some uncertainty over the total value, the safest figure remains that of Kaempfer in 1691 as a working estimate. In 1810, the number of Chinese goods that Satsuma could trade even on these terms had been a mere eight items. The notional figure of permitted trade was set at 1,720 kanme in 1825 (with specified limits for each of sixteen items).118 In 1810, in regard to a question as to the coverage of permitted Chinese medicines, opposition by the Kaisho was overruled by the shogunate.119 However, permitted trade even in the categories of 1825 excluded many medical items. PROBLEMS IN INTERPRETING THE CHARACTER OF SMUGGLING

Satsuma is associated in the modern literature with smuggling, at first inwards to the domain and as a second stage outwards from the domain to Japanese markets. There are two aspects to this. The first is the scale of the Ryukyu trade, and more specifically the extent to which the trade as a whole was more substantial than the amount actually landed in Nagasaki (often somewhat below the permitted level). This leads to the second aspect, namely the extent to which there was on its coastline a smuggling of goods into Satsuma by Chinese vessels. In some studies, a large smuggling trade on the coasts of Satsuma is taken for granted, as for instance those by authoritative historians like Miyamoto Matao ⭖ġ㛔ġ⍰ġ恶ġ and Hayami Akira 忇ġ㯜ġ圵ġ or Ichimura Yūichi ⶪ㛹䣸ᶨġand Ōishi Shinzaburō ⣏䞛ヶᶱ恶, who in brief words saw the coast of Satsuma as frequented by smugglers.120 The picture is not different in monographs on trade. Yamawaki alleged that Chinese vessels were numerous in the many islands on the western coastline of

300

EARLY JAPANESE TRADE

Satsuma and that the domain provided Chinese interpreters.121 Uehara, drawing on official concern in 1835, saw an open-ended Chinese traffic “in the many islands off Satsuma [where] Chinese vessels were able to conduct a lucrative trade.”122 The problem that this poses can be seen at its starkest in Sakai’s article, which assumes at one and the same time an official tolerance of smuggling on the coasts and, in contradictory fashion, a need by smugglers to conceal their goods: This activity could have been stamped out had the Satsuma government so desired. It is probable that the government was a silent partner in the trade, though this would be difficult to establish. Satsuma’s own seclusion policy, which kept out strangers and hampered the activities of bakufu agents, no doubt served to provide the necessary security for the operators along the coast.123

Satsuma on the record of its history did not welcome Chinese vessels hovering on its coasts. The question is, therefore, whether the welcome in Satsuma for some vessels which broke their China-Nagasaki run to visit Satsuma extended to other Chinese vessels, and whether these latter vessels were numerous.124 Given a place for two or three vessels, which broke their route to Nagasaki, it seems likely that there were Chinese vessels coming direct from China, which enjoyed similar permission. But with a closely monitored trade, a jealous protection of its own commercial interests, and harsh production monopolies, an unmanaged trade is highly improbable. In other words, the activity may have been both known and closely policed. If smuggling by Chinese vessels on the coasts of Satsuma was as wholesale as suggested in modern accounts, calculations based on the number of vessels in the Ryukyu-Satsuma fleet would result, with the addition of Chinese vessels, in either an impossibly large quantity of smuggled goods or else a greatly underutilized fleet. Traditions of concealment of smuggled goods in locations on the coast point not to the silent toleration suggested by Sakai, but to clandestine activity intended to evade domain restrictions.125 A rising trade in marine products financing an equally burgeoning trade in pharmaceutical goods led over time to a significant direct exchange in the northwest of Honshu of marine products from Ezo for pharmaceutical products from Satsuma or from Nagasaki. The shogunal policy had long been to keep Satsuma out of this circuit.126 This

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

301

concern may also explain why statistics of the quantities of pharmaceuticals imported in Nagasaki, which had lapsed in 1735, resumed in 1820.127 The shogunate took over Niigata in 1843, in part acting in response to a report of 1841 of an estimated six Satsuma vessels a year in Niigata.128 Claims of depression or of difficulties in 1835–1837 emanating from Nagasaki may have been special pleading in favor of the Kaisho’s business.129 But there may have been a contrast between intervals within the 1830s. While imports in Nagasaki in 1839 were 20 percent above the level of 1804 (carried on eight vessels compared with eleven in 1804), there is evidence of unsold goods in the later years of the decade.130 The uncovering in 1835 of some of the vessels licenced for the Nagasaki trade as visiting Satsuma is not surprising. The marketing in Nagasaki of Chinese wares coming from the Ryukyus gave Satsuma merchants regular business links in Nagasaki to an extent enjoyed by no other domain. The domain’s yashiki had its officials; daimyo sometimes visited either Dejima or the Dutch in their inn when in Edo on the hofries (an association famous in the long friendship of the daimyo Shimazu Shigehide Ⲟġ㳍ġ慵ġ尒ġ [1745–1833]); and Satsuma merchants had long resided in Nagasaki. From 1810, the Ishimoto family settled there, and in 1822–1835 Ishimoto Heibee 䞛ġ㛔ġ⸛ġℝġ堃ġ , domain agent in Nagasaki, selling commodities on domain accounts and remitting payments to the domain, had a direct interest in the trade. In 1835, he even passed into shogunal service.131 An overlapping profile of the trade of the two ports inevitably meant that merchants were likely to find shared interests. OFFICIAL MEASURES IN THE 1830S

A vigorous campaign conducted against smugglers in 1825 had a local context. A total of fifty-three cases appeared in the hankachō between the second month of 1825 and the third month of 1826 as against token action at other times, namely an average of below three cases a year from 1718 to 1862. It was still being followed up in 1827. Events developed rapidly in the 1830s. The arrival of two Chinese vessels in Nagasaki in ballast in the autumn of 1834 was an awakening.132 The bugyō in 1835 in the third month ordered the implementation of tight supervision.133 The routine responsibility for this fell on the metsuke Togawa Yasuzumi ㇠ⶅ⬱㶰ġ (1787–1868), already in Nagasaki and promoted on the spot to the rank of bugyō in 1836. At the outset of 1836 the authorities became aware of a Chi-

302

EARLY JAPANESE TRADE

nese vessel which visited Kataura 䇯ġ㴎ġat the end of 1835 and shortly afterwards entered Nagasaki.134 Togawa had made a series of notes on the pattern of the shipping at Nagasaki from the outset of the century, adding further detail for vessels in the Bunka 㔯⊾ġ and Bunsei 㔯㓧ġ periods, and for Tenpō ⣑ġᾅġ years noting the numerical identification from their last visit to Nagasaki of vessels dallying in Satsuma.135 However, Kataura apart, another source, Kindai zakki 役ġẋġ晹ġ姀ġ, cited in the same pages of the TKIR zokushū, contained a sketch of a Chinese vessel anchored seawards of a Tōjinkan ⒸġṢġ棐ġġ residence for Chinese at Kosezaki ⮷ġ㿔ġⲶġ.136 Though the TKIR is usually a negligible source for action against smuggling, for 1835–1837 the account was enriched by access by the compilers of the TKIR zokushū to a copy of the widely transcribed Nagasaki shi zokuhen 攟ġ Ⲷġ⽿ġ䵂ġ䶐ġinto which Togawa had copied commercial intelligence and correspondence. In the mid-1830s the Nagasaki authorities pursued a carrot-andstick policy. On the one hand, punishment of Chinese for criminal offenses had been extended in the late 1820s to branding on the arm;137 a number of Chinese were jailed, including, as the Chinese shipowners claimed in petitions, innocent ones as well as the guilty. On the other hand, the atmosphere was open enough for the Chinese shipowners to petition collectively about the effects of bugyōsho actions on their trade, and petitions were listened to attentively and sometimes favorably. Even if some ships were excluded from trade and shipowners punished, the overall impact of official policy may have been benign. There was a remarkable temporary intervention by the authorities to buy goods the merchants were unable to sell. This may have been recognition of the adverse effects of policy on trade and an effort to mop up unsold goods, which might otherwise be smuggled. This was first taken in response to a request from the Chinese merchants, apparently in 1835, and a request for a further year was acceded to in 1836, in the fourth month.138 That the policy may have been extended for several further years is suggested by the fact that the figure for exports fell short of imports of 9,217 kanme in 1839, leaving a balance of 2,183, and suggests that sales may have been sluggish. Held in gin satsu (paper money convertible into silver), it was used for several purposes, including defraying costs of the Chinese community in Nagasaki.139 The many reports in the TKIR papers have no reference to Chinese vessels literally hovering on the Satsuma coasts. There was, however,

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

303

repeated concern about Chinese goods introduced to Satsuma and filtered to Osaka and other trading centers.140 The concern led finally to an order prohibiting all Satsuma trade in Chinese goods in 1839.141 This was in some senses a futile gesture as in the absence of direct action by the bugyōsho in the domain or on its borders it was unenforceable. The 1839 order was reversed in 1846 when trade in sixteen commodities was restored.142 If Abe Masahiro 旧ġ 悐ġ 㬋ġ ⻀ġ , the new leader, made the change to improve relations with Satsuma (in part because of the common threat of the novel presence in the 1840s of some Europeans in the Ryukyus), the concession itself points to the problems of reconciling the conflicting interests of the Kaisho and of Satsuma (sometimes denoted by the name of the domain’s head town, Kagoshima). While permitted exports were restored to the former number, the ceiling set at 1,200 kanme was well below the former ceiling of 1,720 kanme. QUANTIFICATION OF THE TRADE

The trade from the Ryukyus to Satsuma is impossible to assess other than very tentatively in quantitative terms. At the peak, a fleet of some twenty vessels may serve as some measure of general capacity. Some twenty cargoes each valued at 200 kan would have amounted to 4,000 kanme. Deducting 1,000 kanme for sales in Nagasaki, the balance of goods destined for sales in the domain or beyond its boundaries could have been 3,000 kanme. Allowing for a number of imponderables making for either plus or minus adjustments, a trade of 3,000 kanme is very credible. No data are available in domain archives on the trade of Satsuma for 1808–1839.143 Matsui’s suggestion—which appears to be derived from revenue data— that “after 1830” the exports from Satsuma were 5,000 kanme is not clear enough to be convincing.144 However, it would tend to lend support to higher rather than lower estimates. Satsuma exports, whether 3,000 or 5,000 kanme, would in relative terms have made official apprehensions understandable.145 The domain’s legitimate traffic in Nagasaki, plus unquantified and, indeed, unquantifiable traffic to other locations gave Satsuma a real weight. The basic problem is the absence of documentation of the unquantified trade from the Ryukyus to Satsuma. Even Satsuma business in Nagasaki is poorly documented in surviving Kaisho records, and only from a late date. In 1847, it was 947 kanme, in 1848 1,337 kanme, and 951 kanme in 1849.146 In 1858 it was 2,265 kanme, and in

304

EARLY JAPANESE TRADE

1859 1,346 kanme.147 In the eighteenth century, reference in the dagregister to vessels to and from Naha underlined the regularity of the traffic. In 1739, for instance, four from Naha were noted. And at year end, seven sailed to Naha. In 1734, ten vessels were noted from Naha, a total that might suggest in 1734 a trade of 1,000 to 2,000 kanme (assuming that the value of cargoes lay between 100 and 200 kanme). As the trade of vessels from China to Nagasaki averaged 4,000 kanme in the 1720s, Satsuma’s supply to the Kaisho, even if we cannot be sure of its precise value, would have been a significant addition to Kaisho business. Moving back along the supply chain, the trade in goods of Chinese origin from the Ryukyus to Satsuma has no statistical documentation. For the trade between China and the Ryukyus in both imports and exports there are data only for the years from 1821 to 1873.148 These statistics are uncharacteristically complete (presumably reflecting Satsuma’s close scrutiny of the trade even in Naha) though in modern accounts the figures warrant no comment on their archival significance. Exports to China from the Ryukyus, limited in number, reflected the poverty of the Ryukyus (and indirectly of Satsuma itself ). By far the largest item in volume terms in exports was kaigan sai 㴟ġ Ⱡġ厄ġ(seashore plants and edibles), running at 2,000 to 3,000 piculs. As they would take up no more than 180 tons of shipping space, they suggest the underlying limitations of the trade, and indeed of Satsuma’s capacity to pay for imports. Imports were varied, but medical goods (yakuzai 啔ġ㛸ġ) and textiles were the sole significant categories. Medical (or pharmaceutical) products, demand for which in Nagasaki repeatedly outran supply, were the real strength of Satsuma’s trade. Amid fluctuations, where goods came from China to the Ryukyus, by reverting to but not exceeding preceding peaks, the pattern suggested a steady rather than expansive profile. BAKUMATSU TRADE OF NAGASAKI AND SATSUMA

From the late 1840s, Nagasaki trade wilted. From 1846 to 1851, the number of Chinese vessels in Nagasaki was down to between four and six a year, and in the 1850s as few as one, two, or three. In 1858, there was a single vessel. A table for 1858 seems to suggest imports worth 3,792 kanme and exports of 5,214 kanme. The superficially substantial figures appear to relate to two-way speculative movements in currency or precious metals and no longer mirror a commodity trade. Nagasaki benefited from English trade from 1859. The consular reports from

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

305

Nagasaki noted trade from Shanghai worth 1,104,061 dollars (the equivalent of approximately 7,360 kanme) for the first half of the year. However, as the port did not open to trade until 1 July, it must have referred only to China traffic. That, combined with a lack of vessels, suggests speculation in currency as the main feature in transactions. A figure of 870,436 dollars for the second half of the year covers the first months of open trade, and would have included trade in English as well as Chinese hands. Rutherford Alcock, British consul general, noted in June 1859 that there were already fifteen British residents and fifteen foreign vessels in the harbor.149 Less is known of Satsuma. The domain’s trade continued in the 1860s and to a degree in stable quantitative terms if the statistics of trade between the Ryukyus and China are regarded as a proxy for its trade. In the shelling by the British of Kagoshima in 1863, five junks from the Ryukyus in the harbor were destroyed.150 The past history of arbitrary levies imposed on Osaka merchants helped to sharpen the fears Saigō Takamori expressed on October 1867 to Ernest Satow about the implications of the shogunate proposing to take over control of the new trade in prospect for Osaka.151 Though in a precarious state in the 1850s, Nagasaki’s trade increased in the 1860s. At the outset of the decade, at least as measured by the Shanghai statistics (the focal point for British trade at the outset of the open ports), its trade exceeded that of the newly opened port of Kanagawa. This reflected the initial advantage of Nagasaki as an established center, and the consequent settlement of several merchants, especially Thomas Glover, arriving in 1859, first dealing in tea and from 1864 establishing close ties with dissident domain.152 Nagasaki’s poor immediate hinterland handicapped it in relative terms. As it made more sense to buy near the sources, foreign trade moved to Yokohama (the successor location to Kanagawa) and after 1868 also to Osaka/ Hyogo. While Kagoshima’s trade appeared to hold its own in the 1860s (on the evidence of the figures for China-Ryukyu trade), the import of medical goods failed to return to the high 1861 figure. In other words, it may have begun to lose ground in the wake of open trade from China to other ports. The Chinese were quick to become the most numerous foreigners in the ports. Satsuma and the Ryukyus alike figure little in the commercial story of early Meiji. Satsuma’s small ocean-going fleet probably counted for less in Satsuma leadership of the future Meiji navy than the realpolitik of ensuring that defence was shared in a rough and ready way between the two great rivals, Chōshū and Satsuma.

306

EARLY JAPANESE TRADE

CONCLUSION

Japanese trade had continued to expand for several decades in the wake of the introduction of sakoku. Later, when silver was virtually prohibited, political pressures in support of maintaining and even expanding imports existed for a time. The contraction of the Dutch trade, very real by the end of the 1690s, was the beginning of the long-term trade decline of Nagasaki. The pattern of the China trade is more nuanced. Exports to China rose in the 1690s and in the following decade, but thereafter trade fell and stagnated. But by the end of the eighteenth century, cargoes of greater value on a much-reduced number of vessels represented a new vigor. Imports in 1804 were 83 percent above the average of the 1720s. Demand conditions favored Satsuma’s irregular trade as much as they did Nagasaki’s legal trade. Nagasaki’s exports to China of 7,345 kanme in 1804, combined with estimated Satsuma/Ryukyu exports of 3,000–4,000 kanme, would give a figure in excess of 11,000 kanme, close to the relatively high exports to China a century before in 1704–1708.153 But the comparison is not quite like with like. The Tsushima trade, if added to the 1704–1708 total, would have raised the total for the base years. Later, for the nineteenth century Satsuma trade, we have to rely on a crude estimate, which must have included shipments to other parts of Japan, as well as trade with China. Japan’s trade, moreover, had become increasingly an exchange of marine products for pharmaceutical imports. Satsuma and Nagasaki were in sharp competition with one another for imports and exports alike in what may have been difficult years in the late 1830s. Perhaps significantly, exports from Nagasaki to China, resting on surviving figures for a mere two years at 7,035 kanme for 1839, were somewhat below the level of 1804. Uchi harai (fire on and repel), a measure originally intended to deal with smugglers, was directed against Western vessels from the 1790s to the early 1840s. From the 1840s, it was to take second place to wider defence preoccupations. Influenced by the rise in the number of foreign vessels on Japan’s coasts, fear of the risk of disruption to the vital trade in rice from the northern domains to Edo led finally to the remarkable study of Edo’s intake by sea of goods for the year 1856.154 Defence concerns centered on the approaches to Edo. While Noell Wilson has seen complacency recur in Nagasaki in the decades after the Phaeton incident, sustained investment in Edo Bay and its approaches began in the 1840s.155 By 1853 there were one hundred can-

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

307

nons in batteries around Edo Bay, though Admiral Perry’s men had a dismissive opinion of them. Effective cannons required high-quality cast iron. This need was met by small blast furnaces in several domains in the 1850s, producing annually around three hundred tons of goodquality cast iron.156 The cannon in Edo Bay came from Saga.157 While Japanese cannons lost artillery duels with foreign warships in the strait of Shimonoseki in 1863 and 1864 and in Kagoshima in 1863, the cannons inflicted real damage on the enemy, notably so in the latter and most serious instance. If, in the 1820s, shogunal officials had divided over uchi harai, daimyo did so in the 1850s over opening ports to trade. Acquiescence in 1858 was for many daimyo a means of buying time, intended from a position of later-acquired strength, either to end or change the treaties. As for the new trade, an ongoing difference in mint ratio (that is, the relative price of silver and gold) between the outside world and Japan, together with a hybrid currency, introduced an extended period of problems, some simply in bookkeeping terms, but some very real.158 On the commodity front, sakoku has often been seen as having had a high cost in opportunities foregone. Overblown accounts of ikki (rural unrest) have been taken as proof of impoverishment caused by a closed society. But Japan in the final decades of sakoku was already a food surplus region.159 Its isolation moreover had not prevented its silk industry from developing dynamically; tea production, in now specialist regions responding to a growing domestic taste, had become larger and more efficient. Silk and tea, both sold to Americans, were to be the backbone of foreign trade. While exports went to the United States, imports were drawn in more widely. If one made a counterfactual argument of opening the ports in, say, 1848, the external markets would not have been there. But in extending its frontier to the Pacific, the American market trebled from 1859 to 1900 (from a population of 23.2 million to 76.2 million). Foreign trade in 1848 would have been modest in outcome; later, resting on a fast-spreading railway system, it led to a vast market and a changed course of history for both Japan and the broad north Pacific. ACKNOWLEDGEMENTS

I am indebted in innumerable ways for assistance and support to Professor Katsuta Shunsuke of the University of Tokyo, and for information on Nagasaki over three decades to the late Tsukahara Sueko,

308

EARLY JAPANESE TRADE

almost seventy years ago the kanbun teacher of my deceased friend Professor Matsuo Tarō of Hosei University. Without their help this paper would have been imperfect and indeed perhaps might not have taken shape. I am indebted also to Dr. Honma Sadao of Nagasaki Kenritsu Toshokan, Professor Kasaya Kazuhiko of Nichibunken, two anonymous referees for invaluable comments and suggestions, and Dr. John Breen for his patient editorial guidance. APPENDIX

Glossary of Weights and Measures kin 㕌: unit of weight, 160 monme, 0.6 kg or 1.32 lb. kan 屓: unit of weight, 1,000 monme, 3.75 kg or 8.27 lb. kanme 屓䚖: silver (coin or metal) by weight as a measure of value. Silver coins were assessed by weight, their purchasing power fluctuating in proportion to the commodity price level and to demand for silver relative to other coins, especially the gold ryō. koku 䞛: unit of capacity, approx. 180 litres or 5 bushels. Weight of a koku of rice approx. 150 kg or 330 lb.

monme ⊩: unit of weight, 1,000 to the kan and kanme.

picul ΌͧΣ: pikuru, unit of weight, 60.48 kg or 133 1/3 lb . Weight often quoted in kin (100 kin to a picul). ryō ᷉: Round-figure conversion of silver kan into gold ryō at 17.2 ryō to a kan (see section 5 of this paper). An official exchange rate, as opposed to the market rate, of 58 monme to the ryō. The official rate was set from time to time by shogunal decree (See Cullen 2003, pp. 73, 76). ton: maritime ton: gross tonnage (measure of total capacity), net tonnage (gross tonnage less capacity reserved for crew and passengers), and deadweight tonnage (weight of cargo). Japanese maritime tonnage calculated in koku (see above). Japanese figures are probably of net capacity or deadweight tonnage. Net capacity and deadweight cargo of a sailing vessel probably differed little for a cargo of rice.

REFERENCES

Ackroyd 1979 Joyce Ackroyd. Told around the Bush Fire: The Autobiography of Arai Hakuseki. University of Tokyo Press, 1979. Arano 1994 Yasunori Arano. “The Entrenchment of the Concept of ‘Sakoku Seclusion.’” Acta Asiatica 67 (1994), pp. 83–103. Arano 2005 Yasunori Arano. “The Formation of a Japanocentric World Order.” International Journal of Asian Studies 2:2 (2005), pp. 185–216.

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

309

Bodart-Bailey 1999 Beatrice M. Bodart-Bailey. Kaempfer’s Japan: Tokugawa Culture Observed. University of Hawai‘i Press, 1999. BPP British Parliamentary Papers. Area Studies Parliamentary Papers, Japan. 10 vols. Ed. W. G. Beasley. Irish University Press, 1971–1972. Broughton 1804 W. R. Broughton. A Voyage of Discovery to the North Pacific Ocean. 2 vols. Cavell & Davies, 1804. Cullen 2003 L. M. Cullen. A History of Japan 1582–1941: Internal and External Worlds. Cambridge University Press, 2003. Cullen 2009 Louis M. Cullen. “Statistics of Tokugawa Coastal Trade and Bakumatsu and Early Meiji Foreign Trade. Part 1: Coastal Trade in Tokugawa Times.” Japan Review 21 (2009), pp. 183–223. Cullen 2010 Louis M. Cullen. “Statistics of Tokugawa Coastal Trade and Bakumatsu and Early Meiji Foreign Trade. Part 2: Trade in Bakumatsu and Early Meiji Times.” Japan Review 22 (2010), pp. 59–101. Cullen 2013 Louis M. Cullen. “Japanese Archives: Sources for the study of Tokugawa Administrative and Diplomatic History.” Japan Review 25 (2013), pp. 33–65. DDM 1700–1740 The Deshima Diaries: Marginalia 1700–1740. Eds. Paul Van der Velde and Rudolf Bachnofer. Japan-Netherlands Institute, 1992. DDM 1740–1800 The Deshima Diaries: Marginalia 1740–1800. Eds. Leonard Blussé, Cynthia Vialle, Willem Remmelink, and Isabel van Daalen. Japan-Netherlands Institute, 2004. DDR 1986–2010 The Deshima Dagregisters: Their Original Contents. 13 vols. Leiden Centre for the History of European Expansion, 1986, 1987, 1989, 1990 (bis), 1991, 1993, 1994, 1996, 1997. Institute for the History of European Expansion, 2001, 2005, 2010.

310

EARLY JAPANESE TRADE

Hellyer 2009 Robert I. Hellyer. Defining Engagement: Japan and Global Contexts, 1610 –1868. Harvard University Press, 2009. Ichimura and Ōishi 1995 Ichimura Yūichi ⶪ㛹䣸ᶨġ and Ōishi Shinzaburō ⣏䞛ヶᶱ恶. Sakoku yuruyaka na jōhō kakumei 挾⚥͇̫͌̌ͅね⟙朑␥. Kōdansha, 1995. Ishida 2004 Ishida Chihiro 䞛䓘⋫⮳. Nichiran bōeki no shiteki kenkyū 㖍嗕屧㖻̯⎚䘬 䞼䨞. Yoshikawa Kōbunkan, 2004. Ishida 2009 Ishida Chihiro. Nichiran bōeki no kōzō to tenkai 㖍嗕屧㖻̯㥳忈̩⯽攳. Yoshikawa Kōbunkan, 2009. Ishii 1998 Yoneo Ishii. The Junk Trade from Southeast Asia: Translations from the Tōsen Fūsetsugaki, 1674–1723. Institute of Southeast Asian Studies, 1998. Ishikawa 1967 Shigeru Ishikawa. Economic Development in Asian Perspective. Kinokuniya Bookstore, 1967. Iwao 1953 Iwao Seiichi ⱑ⯦䓇. “Kinsei Nisshi bōeki ni kansuru sūryōteki kōsatsu” 役ᶾ㖍㓗屧㖻̬敊͌̚㔘慷䘬侫⮇. Shigaku zasshi ⎚⬎晹娴ġ LXII, 16 (1953), pp. 1–40. Iwao 1963 Seiichi Iwao. “Re-opening of the Diplomatic and Commercial Relations between Japan and Siam during the Tokugawa Period.” Acta Asiatica 4 (1963), pp. 1–31. Iwao 1976 Seiichi Iwao. “Japanese Foreign Trade in the 16th and 17th centuries.” Acta Asiatica 30 (1976), pp. 1–18. Jansen 2000 Marius B. Jensen. The Making of Modern Japan. Harvard University Press, 2000. Kagoshima kenshi 1940 Kagoshima kenshi 渧⃸Ⲟ䚴⎚. Vol. 2, Kagoshima ken, 1940. Katsu 1976 Katsu Kaishū ⊅㴟凇. Suijinroku ⏡⠝拚ġ 1890. In Katsu Kaishū zenshū ⊅ 㴟ℐ普, vol. 9. Keisō Shobō, 1976.

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

311

Kobata 1993 Kobata Atsushi ⮷叱䓘㶛. Nihon dōkōgyōshi no kenkyū ᪥ᮏ㖡㖔ᴗྐ̯䞼䨞. Shibunkaku Shuppan, 1993. Lewis 2003 James Bryant Lewis. Frontier Contacts between Chōson Korea and Tokugawa Japan. RoutledgeCurzon, 2003. Matsui 1975 Masato Matsui. “Shimazu Shigehide 1745–1833: A Case Study of Daimyo Leadership.” PhD dissertation, University of Hawai‘i, 1975. Matsui 2015 Matsui Yōko 㜦ḽ㲳⫸. “Dejima to kakawaru hitobito” ↢Ⲟ̩̌̌͐͌Ṣ ˦. In Nichiran kankeishi o yomitoku ᪥⹒㛵ಀྐ͓͉̩̀̐, ed. Matsukata Fuyuko 㜦㕡⅔⫸. Rinsen Shoten, 2015, pp. 146–80. Matsuura 2007 Matsuura Akira 㜦㴎䪈. Edo jidai Tōsen ni yoru Nicchū bunka kōryū 㰇㇠㗪 ẋⒸ凡̬͉͌㖍ᷕ㔯⊾Ṍ㳩. Shibunkaku Shuppan, 2007. Miyamoto and Hayami 1988 Miyamoto Matao ⭖ġ㛔ġ⍰ġ恶ġ and Hayami Akira 忇ġ㯜ġ圵ġ . Nihon keizai shi 1: Keizai shakai no seiritsu, 17–18 seiki 㖍㛔䳴㶰⎚ġ 1Ļġ 䳴㶰䣦Ể̯ㆸ 䩳ˣ17–18ᶾ䲨. Iwanami Shoten, 1988. Miyashita 1997 Miyashita Saburō ⭖ᶳᶱ恶. Nagasaki bōeki to Ōsaka: Yunyū kara sōyaku e 攟Ⲷ屧㖻̩⣏旒Ļġ廠ℍ̌͊∝啔̹. Seibundō, 1997. Nagasaki kenshi 1985 Nagasaki kenshi taigai kōshō hen 攟Ⲷ䚴⎚⮦⢾Ṍ㶱䶐. Yoshikawa Kōbunkan, 1985. Nagazumi 1987 Nagazumi Yōko 㯠ġ 㥻ġ 㲳ġ ⫸ġ . Tōsen yushutsunyūhin sūryō ichiran 1637– 1833 Ⓒġ凡廠↢ℍ⑩㔘慷ᶨ奏ġ1637–1833. Sōbunsha, 1987. Nakai 1988 Kate Wildman Nakai. Shogunal Politics and the Premises of Tokugawa Rule. Harvard University Press, 1988. Nakajima 2007 Nakajima Motoki ᷕⴳ⸡崟. Koga Jūjirō: Nagasaki gaku no kakuritsu ni sasageta shōgai ⎌屨⋩Ḵ恶Ļġ 攟Ⲷ⬎̯䡢䩳̬̖̖̠̓䓇㵗. Nagasaki Bunkensha, 2007.

312

EARLY JAPANESE TRADE

Nakamura 1988 Nakamura Tadashi ᷕ㛹岒. Kinsei Nagasaki bōeki no kenkyū 役ᶾ攟Ⲷ屧㖻 ̯䞼䨞. Yoshikawa Kōbunkan, 1988. Nakamura 2000 Nakamura Tadashi. Kinsei taigai kōshō shi ron 役ᶾ⮦⢾Ṍ㶱⎚婾. Yoshikawa Kōbunkan, 2000. Nihon zaisei keizai shiryō Nihon zaisei keizai shiryō 㖍㛔屉㓧䳴㶰⎚㕁. 10 vols. Ōkurashō Zaisei Keizai Gakkai, 1921–1925. Ōba 1967 Ōba Osamu ⣏ġ⹕ġ僑ġ. Edo jidai ni okeru Tōsen mochiwatashisho no kenkyū 㰇ġ㇠ġ㗪ġẋ̬̋̒͌Ⓒ凡㊩㷉㚠̯䞼䨞. Kansai University Institute of Oriental and Occidental studies, 1967. Ōba 1974 Ōba Osamu, ed. Tōsen shinkō kaitōroku, Shimabarabon Tōjin fūsetsugaki, Wappu tomechō Ⓒ凡忚㷗⚆つ拚ĭġ Ⲟ⍇㛔ⒸṢ桐婔㚠ĭġ √䫎䔁ⷛ. Kansai University Institute of Oriental and Occidental Studies, 1974. Ōta 1992 Ōta Katsuya ⣒䓘⊅ḇ. Sakoku jidai Nagasaki bōeki shi no kenkyū 挾⚥㗪ẋ 攟Ⲷ屧㖻⎚̯䞼䨞. Shibunkaku Shuppan, 1992. Ōta 2000 Ōta Katsuya. Nagasaki bōeki 攟Ⲷ屧㖻. Dōseisha, 2000. Peng 2015 Peng Hao ⼕忈. Kinsei nisshin tsūshō kankei shi 役ᶾ㖍㶭忂⓮ġ 敊Ὢ⎚. Tōkyō Daigaku Shuppankai, 2015. Sakai 1964 Robert K. Sakai. “The Satsuma-Ryukyu Trade and the Tokugawa Exclusion Policy.” Journal of Asian Studies 23:3 (May 1964), pp. 391–403. Satow 1921 Sir Ernest Satow. A Diplomat in Japan. (Reprinted by ICG Muse, Inc., 2000.) Shimada 2006 Ryuto Shimada. The Intra-Asian Trade in Japanese Copper by the Dutch East India Company. Brill, 2006.

THE NAGASAKI TRADE OF THE TOKUGAWA ERA: ARCHIVES

313

Souza 1986 George Bryan Souza. The Survival of Empire: Portuguese Trade and Society in China and the South China Sea, 1630–1754. Cambridge University Press, 1986. Suzuki 2004 Suzuki Yasuko 懜㛐⹟⫸. Kinsei Nichiran bōeki no kenkyū 役ᶾ㖍嗕屧㖻̯ 䞼䨞. Shibunkaku Shuppan, 2004. Suzuki 2012 Yasuko Suzuki. Japan-Netherlands Trade, 1600–1800: The Dutch East India Company and Beyond. Kyoto University Press, 2012. Takeno 1956 Takeno Yōko 㬎ġ 慶ġ 天ġ ⫸ġ . “Satsuma Ryūkyū bōeki to bōeki shōnin Ishimoto ke no kankei” 啑㐑䎱䎫屧㖻̩屧㖻⓮Ṣ䞛㛔⭞̯敊Ὢ. Kyūshū keizai-shi ronshū ḅⶆ䳴㶰⎚婾普ġ2 (1954), pp. 93–115. Tashiro 1976 Tashiro Kazuo. “Tsushima Han’s Korean trade, 1684–1710.” Acta Asiatica 30 (Feb. 1976), pp. 85–105. Tashiro 1981 Tashiro Kazuo 䓘ẋ␴䓇. Kinsei Nicchō bōeki shi no kenkyū 役ᶾ㖍㛅屧㖻 ⎚̯䞼䨞. Sōbunsha, 1981. TKIR 1912–1923 Tsūkō ichiran 忂凒ᶨ奏. 8 vols (1566–1825). Kokusho Kankōkai, 1912– 1923. TKIR zokushū 1968–1972 Tsūkō ichiran zokushū 忂凒ᶨ奏䵂廗. 5 vols (1825–1856). Seibundō, 1968–1972. Tōban kamotsuchō 1969–1970 Tōban kamotsuchō Ⓒ嚖屐䈑ⷛ. 2 vols. Naikaku Bunko, 1969–1970. Toby 1991 Ronald B. Toby. State and Diplomacy in Early Modern Japan: Asia in the Development of the Tokugawa Bakufu. Stanford University Press, 1991. Tōtsūji Kaisho nichiroku 1984 Tōtsūji kaisho nichiroku Ⓒ忂ḳỂ㇨㖍拚. 7 vols. Ed. Tokyo Daigaku Shiryōhensanjo. Tokyo Daigaku Shuppankai, 1984.

314

EARLY JAPANESE TRADE

Uehara 1981 Uehara Kenzen ᶲ⍇ℤ┬. Sakoku to han bōeki: Satsuma han no Ryūkyū mitsubōeki 挾⚥̩喑屧㖻Ļġ啑㐑喑̯䎱䎫⭮屧㖻. Yaedake Shobō, 1981. Van Dam 1927 Pieter Van Dam. Beschryvinge van de Oostindische Compagnie. Martinus Nijhoff, 1927. Vol. 1, Inleiding, pp. IX–XXXIII. Van Overmeer Fisscher 1833 T. F. Van Overmeer Fisscher. Bijdrage tot de kennis van het japansche rijk. J. Muller, 1833. Wilson 2015 Noell Wilson. Defensive Positions: The Politics of Maritime Security in Tokugawa Japan. Harvard University Press, 2015. Yamamoto 1995 Yamamoto Hirofumi Ⱉ㛔⻀㔯. Sakoku to kaikin no jidai 挾⚥̩㴟䤩̯㗪 ẋ. Azekura Shobō, 1995. Yamawaki 1960 Yamawaki Teijirō Ⱉ傯㡗Ḵ恶. Kinsei Nicchū bōeki shi no kenkyū 役ᶾ㖍ᷕ 屧㖻⎚̯䞼䨞. Yoshikawa Kōbunkan, 1960. ———— Yamawaki Teijirō. Nihon no Tōjin bōeki 㖍㛔̯ⒸṢ屧㖻. Yoshikawa Kōbunkan, 1964. ———— Yamawaki Teijirō. Nukeni: Sakoku jidai no mitsu bōeki ㉄匟Ļġ挾⚥㗪ẋ̯⭮ 屧㖻. Nihon Keizai Shinbunsha, 1965. ———— Yamawaki Teijirō. “Kaidai” 妋柴. In vol. 2 of Tōban kamotsu chō Ⓒ嚖屐䈑 ⷛ. Naikaku Bunko, 1970. Yao 1998 Yao Keisuke ℓ䘦⓻ṳ. Kinsei Oranda bōeki to sakoku 役ᶾ旧嗕旨屧㖻̩挾 ⚥. Yoshikawa Kōbunkan, 1998. Yasutaka 2010 Yasutaka Hiroaki ⬱檀⓻㖶. Kinsei Nagasaki shihō seido no kenkyū 役ᶾ攟Ⲷ ⎠㱽⇞⹎̯䞼䨞. Shibunkaku Shuppan, 2010.

NOTES ™

INTRODUCTION 1

2

3 4 5

6

7

8

9

10

11

12

13

14 15

16

17

18

Fujita Satoru, ᖥ⸬ไᅜᐙࡢᨻ἞ྐⓗ◊✲ኳಖᮇࡢ⛛ᗎ࣭㌷஦࣭እ஺ Bakuhansei kokka no seiji shi-teki kenkyū: tenpō ki no chitsujō, gunji, gaikō (Studies in the politics of bakumatsu Japan: public order, defence, foreign policy in the Tenpō period ),Tokyo, 1987. For details of statistical and archival evidence pointing to a robust functioning of both society and economy. see footnote 5. Ooms, Herman, Tokugawa Village Practice: Class, Status, Power, Law (Berkeley, 1996). On the emperor’s place in the constitution, see Fujita Satoru, Kinsei seiji- shi to tennō (Tokyo, 1999). For summaries of five papers from the Japan Review in this volume, see pp.xxviii–xxxi. The full texts of the papers are in chapters 6 to 10. Cullen, Texts of papers from the Japan Review, paper no. 4 on archives,.summary , pp.xxx– xxx,i, chapter 10, especially pp.241–3. 㟢すளΏ᮶㘓, ㄴ᪩㒓ᅵྐᩱྀ᭩ III. Roshia torai roku, Isahaya kyōdo shiryō sōsho III, Isahaya Kyōdo Shiryō Kankōkai, 1994. Account of the Russian mission to Nagasaki in 1804. There are very short accounts in the introductions to vol. 1, 1–7 and vol 7, 109–114 of Tōtsuji nichiroku, Dai nihon kinsei shiryō, Tokyo daigaku shuppankai (1984). Kōbunshokan (Naikaku Bunko )Tokyo/ The documents are reproduced inTōban kamotsuchō ၈⻅㈌≀ᖒ (Tokyo, 1969 [Vol. 1], 1970 [Vol. 2]); another document, the Nagasaki goyōdome 㛗ᓮᚚ⏝␃, also bears no indication of the original author(s), nor of where it was held before it was transferred to the National Archives. A study of sources by the Kenkyōiku i-inkai, which lists the Seidōas a point of origin for many surviving documents, does not state that the nikki were part of their holdings; see Nagasaki bugyōsho kankei monjo chōsa hōkoku sho (Nagasaki, 1997), 198. Translated from Latin into German and English, the appendix in time became famous via Shizuki Tadao’s Japanese translation of 1801. On both Andō Shōeki and the interest of Norman and Murayama in his writings, see Cullen, A History of Japan, 1582–1941, 116, 117n, 125,, 284, 284n. Cullen, Louis, Book Review: Ikki: Social Conflict and Political Protest in Early Modern Japan, by James W. White (Ithaca, N.Y., 1995), The Japan Foundation Newsletter, 24 (1996), 21–22. Borton, Hugh, Japan’s Modern Century (New York, 1955). Toby, Ronald, State and Diplomacy in Early Modern Japan (Stanford, 1991; 2nd edition), 11, 10–15,24,242; see also the Preface from the original 1984 edition, (Princeton, New Jersey, 1984), xxvii-xxxvii. Ravina, Mark,To Stand with the Nations of the World: Japan’s Meiji Restoration in World History (Oxford, 2017), 56. Sir Ernest Mason Satow Papers, PRO 30/33/15/7, The National Archives, Kew, diary entry for September 22 1882, 10. Nagasaki Museum of History and Culture,14–93–2, fūsetsugaki 1851–1857 with text of five betsudan fūsetsugaki plus one for 1850 (the texts for 1852 and 1853 are also summarised in

315

316

19

20

21 22

23 24 25

26 27

28

29

30

31

32

33

34

35 36

EARLY JAPANESE TRADE

114–119–1) For the remarkable account of Edo incoming trade in 1856, see chapter 7, Japan Review paper on coastal trade, especially pp.162–3. See also Shimamura Motohiro, ‘Information on western naval matters and their realities, as derived from Dutch special reports’ [English translation of Japanese title], in Katagiri Kazuo, ed., Nichiran kōryū shi: sono hito, mono, jōhō (Kyoto, 2002), 454–474. Cullen, Louis, ‘Sakoku, Tokugawa policy, and the Interpretation of Japanese History’, Transactions of the Asiatic Society of Japan, 18 (2004), 17–31; Katō Yūzō, ‘The Opening of Japan and the Meiji Ishin’, in The History of Anglo-Japanese Relations, 1 (2000), 60–86; see also Katō Yūzō, Bakumatsu gaikō to kaikoku (Tokyo, 2004); Auslin, Michael R., Negotiating with Imperialism and the Culture of Japanese Diplomacy (Harvard, 2004). Sekiyama Naotarō, Kinsei Nihon no jinkō kōzō: Tokugawa jidai no jinkō chōsa to jinkō jōtai ni kansuru kenkyū ㏆ୡ᪥ᮏࡢேཱྀᵓ㐀: ᚨᕝ᫬௦ࡢேཱྀㄪᰝ࡜ேཱྀ≧ែ࡟㛵ࡍࡿ◊✲ (Tokyo, 1958), 95–96. There is a useful summary of its work in Fujita Satoru, Kanjō bugyō no Edo jidai (Tokyo, 2018). Totman, Conrad, The Collapse of the Tokugawa Bakufu, 1862–1868 (Honolulu, 1980), 259, 293–295, 369–370. Fukui Tamotsu, Naikaku bunko shoshi no kenkyū (Tokyo, 1980), 44. Totman, The Collapse of the Tokugawa Bakufu, 197. The Deshima Dagregisters : their original tables of contents, vol. X, 1790–1899, ed.Cynthia Viallé and Leonard Blussé , Leiden, Institute for the History of European Expansion, 1997, p. ix, and appendix 4, pp.187–189. See also The Deshima Diaries: Marginalia 1740–1800, ed.Leonard Blussé , Cynthia Viallé, Willem Remmelink. Isabel Van Daalen , Japan-Netherlands Institute, Tokyo, 2004. The fire in April is not recorded in the main text, but in the introduction (pp.9–11, 29–35) in verbatim extracts for 9 June from the diary of the bookkeeper , and on 21 and 24 April by Pogedt, acting for the opperhoofd absent on the hofreis. I am indebted to Prof. Katsuta Shunsuke for these extracts. Totman, The Collapse of the Tokugawa Bakufu, -294. Kenkyōiku i-inkai, Nagasaki bugyōsho kankei monjo chōsa hōkoku sho (Nagasaki, 1997), 198. Nagasaki ken shi henshū i-inkai, Nagasaki ken shi: taigai kōsho hen, Vol. 4 (Tokyo, 1985), 892–893; Cassel, Kristoffer, Grounds of Judgement: Exterritoriality and Imperial Power in Nineteenth-Century China and Japan (Oxford, 2012), 95. The Motoki papers are in two locations. The first is Nagasaki Kenritsu Rekishi Hakubutsukan, their history summarized in Chapter 9, Japanese archives, page 244, note 54. The papers in Kobe are listed in Kobe shiritsu hakubutsukan kanzōhin: bijutsu no bu, 1997. The two collections, though substantial, are very incomplete which explains why a hope of more being found lingered for long. Mitani Hiroshi, in his Meiji ishin to nashonarizumi: Bakumatsu no gaikō to seiji hendō (Tokyo, 1997) has argued that the 1880s marked the origins of modern Japanese nationalism. Satow Papers, PRO 30/33/11/10, f. 47, The National Archives, Kew, letter from Satow to John Harington Gubbins, June 2 1918. See a letter from Matsukata to a new prime minister Terauchi quoted in Hara Matsukata Reischauer, Samurai and silk: a Japanse and American heritage, Tokyo 1987, 144–5 Satow Papers, PRO 30/33/11/18, f. 65, letter from Satow to Lord Reay, January 30 1921; PRO 30/33/11/14, letter from Lord Reay to Satow, January 17 1921. Grew, Joseph C., Ten Years in Japan, contemporary record drawn from the diaries and private and official papers of Joseph. Grew, Unuted States ambassador to Japan 1932–1942 (London, 1944), 27–28. It can be seen also in the tone and timing of Grew’s Ten years in Japan. A unique account of language instruction and work in translating ‘hundreds of thousands of letters written by ordinary Japanese living under foreign occupation’ is provided by the littleknown America’s Japan: The First Year, 1945–1946 by Grant K. Goodman (New York, 2005). This is a translation of a Japanese-language text written by Goodman and co-author Hideo

NOTES

317

Kobayashi, Amerika no Nippon gannen 1945–6 (Otsuki shoten,Tokyo, 1986). Under the supervision of Prof. J.W. Hall, Goodman wrote a PhD thesis, published as The Dutch Impact on Japan, 1600–1853 (Richmond, 1967). Sadao Oba’s The ‘Japanese’ war: London University‘s WWII secret teaching programme and the experts sent to help beat Japan, Japan Library, Folkestone, 1988, Is of particular interest in illustrating linguistic and practical problems at length.

CHAPTER 1 1

2

3

4

5

6 7

8

9

10

11 12 13

See L.M. Cullen, A history of Japan 1582–1941: internal and external worlds (Cambridge, 2003), pp. xi, 4, 14–5, 159–164,185, 317–9, for some comments on the issues. For a not dissimilar view, see Mizutani Mitsuhiro, “The hard-dying myth of Edo misrule”, Japan Echo, Vol. 30, No. 3 (June 2003), pp. 53–7. The site is on the hillside immediately below the Kenritsu Toshokan. Part of it was occupied by the city’s fine Shiritsu Bijutsukan, which has been demolished, and which has been replaced by a new gallery on another site. ᚚ⏝㒊ᒇᘬ⥅┠㘓Nagasaki bugyōsho kankei monjo chōsa hōkoku sho, Nagasaki ken bunka-zai hōkoku, dai 131 shū [Reports of investigations into the Nagasaki bugyō office, Nagasaki-ken, cultural reports, collection no, 131], Nagasaki-ken Kyōiku Iinkai, pp 9–14. Tō-tsūji kaisho mokuroku (published by Shiryō Hensanjo in 7 volumes, Tokyo, 1955–68). They go down only as far as the years of Shotoku. Later records have not survived. The Motoki records are in Nagasaki Shiritsu Hakubutsukan, and Kobe Shiritsu Bijutsukan. G.M. Golownin, Recollections of Japan (London, 1819), pp. i–xxxix. Narrative of the expedition of an American squadron to the China seas and Japan: performed in the years 1852, 1853, and 1854.... by order of the Government of the United States, compiled from the original notes and journals of Commodore Perry and his officers, at his request and under his supervision by Francis L. Hawks (Washington, 1856), 2 vols. One of the very rare appreciations of this is to be found in Katō Yūzō, “The opening of Japan and the Meiji Restoration, 1837–72”, in The history of Anglo-Japanese relations, Vol. 1, The political-diplomatic dimension, 1600–1930 (Basingstoke, 2000), ed. Ian Nish and Yōichi Kibata, pp. 60–86. This account is in the Motoki archives in the Nagasaki rekishi bunka hakubutsukan (Nagasaki Museum of History and Culture). It has not been mentioned or reproduced in recent work on Titsingh, but well merits attention for its insight into Titsingh’s outlook, and intellectual approach to senior Japanese officials. P. F. von Siebold, Nippon: Archiv zur Beschreibung van Japan (originally published in Leiden, 1835, reprinted Tokyo, 1965). The integral work is still available only in its original German. H. Doeff, Herinneringen uit Japan (Haarlem, 1833). The Deshima dagregisters, Vol. v, 1720–1730, ed. P. van der Velde (Leiden, 1990), p. 39. G.M. Golownin, Narrative of my captivity in Japan in the years 1811, 1812 and 1813 (London, 1818), 2 vols. This is distinct from his Recollections of Japan (London, 1819), a broader account, though in the main drawing on his stay, and on the information gleaned in his conversations widi Murakami Teisuke (or, as Golownin described him, “Teske”).

318

14 15

EARLY JAPANESE TRADE

See Cullen, A history of Japan, p. 152, footnote. Hankachō: Nagasaki bugyōsho hanketsu kiroku 1666–1867 [Nagasaki bugyō office: register of Nagasaki bugyō decisions in criminal matters], 11 vols. (Nagasaki Kenritsu Toshokan, 1959–1961).

CHAPTER 2 1

2

3

4

5

6

7 8

9

10 11 12 13

14 15 16 17 18 19 20 21 22

23 24

Marius B. Jansen, The Making of Modern Japan, Cambridge Massachusetts, Harvard University Press, 2000,p.81. DDR, vol X (1790–1800), p.59. The “Marginal Notes”, extensive summaries of the content of the dagregisters have been edited in two series, firstly The Deshima dagregisters: their original contents (subtitle omitted from later-edited volumes), 1641–1670, 1680–1800, 13 vols, Leiden Centre (Later Institute) for the history of European expansion, 1986–2010 (hereafter abbreviated to DDR); and secondly in fuller form in The Deshima diaries: Marginalia, JapanNetherlands Institute, 1992, 2004, 2 vols, 1700 -1740, 1740–1800 (hereafter abbreviated to DDM). L. M. Cullen, A History of Japan 1582 – 1941: Internal and External worlds, Cambridge, Cambridge University Press, 2003, p.156. Relatio XIV, Regium Japaniae, optima ratione, abgressu civium exterarum gentium ingresu, et communione, clausum (The kingdom of Japan on good grounds closed to the exit of its citizens and entry of or communion with foreign peoples), pp.478–502. See Cullen, “Gulliver in Japan”, Eighteenth-century Ireland, vol. 28 (2013), p.175. Isabel Tanaka-Van Daalen, “Communicating with the Japanese under Sakoku: Dutch letters of complaint and attempts to learn Japanese”, in Large and Broad: Dutch Impact on Early Modern Asia, ed. Yoko Nagazumi, Tokyo, The Toyo Bunko, 2010, p.120. Beatrice M.Bodart-Bailey, Kaempfer’s Japan: Tokugawa Culture Observed, University of Hawaii Press, 1999, p.193. Quoted in Tanaka-Van-Daalen, p.106n. Hendrik Doeff, Recollections of Japan, translated by Annick M. Doeff, Trafford, Canada, 2003, p. 2; J.F. van Overmeer Fisscher, Bijdrage tot de kennis van het japansche rijk, Amsterdam, J. Müller, 1833 p.92. G.F. Meijlan, Japan:vorgesteld in schetsen over de zeden en gebruiken van dat ryk, byzonder over de ingezetenen der stad Nagasaky. Amsterdam, M. Westerman, 1830, pp.36–7. His comments on the Japanese language are set out on pp.117–24. DDM 1700–40, p.305. DDM 1700–40, p. 360, referring to a letter of the opperhoofd’s. Tanaka-Van Daalen, p. 110. Beatrice M.Bodart-Bailey, Kaempfer’s Japan: Tokugawa Culture Observed, University of Hawaii Press, 1999, p.187–8. Tanaka-Van Daalen, pp. 112, 114. Tanaka-Van Daalen, p.117, 118. DDR, vol. 1, 1680–90, pp.33,34. DDR, vol. 1, 1680–90, p.34. Van Daalen, pp. 118–9. DDR, 1780–90, p.72. DDR, 1780–90, introduction, p.vi. DDR, vol. 1,1680–90, p.32;.DDM 1700–1740, p. 493. T. Screech (ed.), Japan Extolled and Decried: Carl Peter Thunberg and The Shogun’s Realm, 1775–1776, London, Routledge, 2005, p.184. DDR, vol. 2, 1690–1700, p. 94. DDR, vol 2, 1690–1700, p.189.

NOTES

25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

44 45 46 47 48 49 50 51

52

53 54 55

56 57 58

59 60 61 62 63 64

65 66 67

319

DDR, vol, 2,1690–1700, p. 79. DDR, vol 2, 1690–1700l p. 96. DDR, vol .2, 1690–1700, pp.131, 140. DDR, vol. 1, 1680–90, p. 42.. DDR, vol. 2, 1690–1700, pp. 94–6., 106.; DDM 1700–1740, p.139. DDR, vol,2, 1690–1700, p.133. Jansen, pp. 80–1. DDM 1700–40, p.34. DDR, vol. 2, 1690–1700, pp. 78, 179. Bodart=Bailey, pp.239–40. DDM 1700–40, p. 87, 14 March 1707. DDM 1700–40, p.105. DDM 1700–40, p.139. DDM 1700–40, p.175. DDM 1700–40, p.185. DDM 1700–40, p. 222. DDM 1700–40, p. 326. DDM 1700–40, 1 May, 1730, p.368. Cord Eherspacher, “Johan Georg Keyserling (1696–1736): A German Horseman at Nagasaki and Edo”, Crossroads: a Journal of Nagasaki History and Culture, no.2 Summer 1992, pp.9–25. DDM 1700–40, p. 360. DDM 1700–40, pp. 481 (several refs.), 482 (do)., 486.(do). DDM 1700–40, pp. 478. DDR, vol.2, 1690–1700, pp. 21, 29. Bodart-Bailey, p. 364. Bodart-Bailey, p. 413. DDR, vol. 3, 1700–10, p.7. Opperhoofeden and staff of the VOC factories in Japan. wolfgangmichel.web.fc2.com/serv/ histmed/dejimasurgeons.html.Accessed 18 May 2017. DDM 1700–40, p. 173. The reference to his being “famous” may be due to the fact that he knew at least some Dutch, which was uncommon at this stage DDM 1700–40, p.299. DDR 1690–1700, pp.40–1, 50–1, 66. DDM 1700–40, p.392. I am indebted for help with this passage to Professors Katsuta Shunsuke and Makihara Shigeyuki of Tokyo University, and for Kyushu dialect of the period, Professor Kimura Naoki of Nagasaki University. Doeff, p.3. DDM 1700–40, p.429. R.H. Hesselink, “A Dutch New Year at the Shirando Academy”, 1 January 1795, Monumenta Nipponica, vol. 50, no. 2 (summer, 1795), pp.205–23. Fisscher, pp.314–5. DDM 1740–1800, p. xxx. DDM 1740–1800, p.xxxvi, DDR, vol. 10, pp.ii-iii. They later served as governor-general in Java and Bengal respectively. Cullen, 2003, p.133. Jan Cock Blomhoff, The Court Journey to the Shogun of Japan: From a Private Account by Jan Cock Blomhoff, ed. F.R. Effert. Leiden, Hotei Publishing. C.2000. His collection was purchased by William 1 in 1826 for the Koninklijk Kabinet van Zeldzaamhedenn (Royal Cabinet of Rarities). Meijlan, pp.117–20. Meijlan, pp.36–7. Officials with widespread supervisory and reporting duties under the governor.

320

68 69 70 71 72 73 74 75 76 77

78 79 80 81

82 83

84 85 86

87 88

89 90 91

92 93

94

95

96

97

EARLY JAPANESE TRADE

Bodart-Bailey, p.362. Blomhoff, p. 86. Cullen, 2003, p.130. Blomhoff, p. 86. Fisscher, pp. 273–4. Bodart–Bailey, p.153. DDM 1700–40, p.117. Fisscher, p.274. Named in Lequin, p.83. Lequin, pp. 85–92. There is an English translation in T. Screech, The Secret Memoirs of the Shogun: Isaac Titsingh and Japan, 1779–1822. Routledge, 2006, pp.202–16. Frank Lequin, p.201 (from letter numbered PC 204). Meijlan, p.124. Lequin, p. 130. Lequin, p.77–8. The introduction has reference to his work with interpreters in collecting words. Lequin, p. 129–30, 228. Lequin, p. 101. Titsingh’s method of studying Chinese is suggested at the foot of p. 195. It seems to rule out study as kanbun, i.e. classical Chinese as written by Japanese. This conclusion is reinforced by the fact that his tutors in the Chinese language were Chinese, not Japanese. Fisscher, p.93. Fisscher, p.307. The one reference is, in a fashion of the time among rangakusha, of giving a Western name to the daimyo of Nakatsu, second son of Shimazu Shigehide. Doeff, p.86. Lequin, p.222. Siebold sought to ignore the work of others. Lequin has observed that there is little mention of Titsingh in Siebold’s Nippon, published from 1832 onwards. (p.224). Cullen, 2003, pp.131–2. Doeff, p. 100. Nagasaki museum of history and culture,14–93–2, pp.261–596, copies of reports for 1850– 1855, lacking diplomatic and details of the circumstances in which the individual reports were requested or provided. Blomhoff, p. 11. Katagiri Kazuo, “Bakumatsu ni okeru ikokusen Ōsetsu to oranda-tsūji Baba Sajūrō, Kaiji-shi kenkyū, no.10 (April 1968), pp.1–36. On Sajūrō’s involvement from1818 in interviewing whalers, see Katakiri, Oranda tsūji no kenkyū, Tokyo, Yoshikawa kobunkan, 1985, pp.363–7. On his questionnaire and English-Japanese short vocabulary, see also Cullen, 2003, p.153. The incident has been much written about. The book by Kure Shūzō (1865–1932) in 1926 is the most complete account. Now translated into German as Philipp Franz Von Siebold: leben und werk, Tokyo, Deutsches Institut fur Japanstudien, 2 vols., 1995. Frederik L. Schodt, Native American in the land of the shogun: Ruonald MacDonald and the opening of Japan, Berkeley, Stone Bridge, 2003, pp.254–262. Francis L Hawks and Mathew Galbraith Perry, Narrative of the expedition of an American squadron to the China seas: performed in the years 1852, 1853 and 1854, under the command of M.C. Perry…, Washington, A.O. Nicholson, 1856, vol.1, p. 340. Townsend Harris, The complete journal of Townsend Harris, revised 1930 edition, Tokyo, C.E. Tuttle, 1959, p.208.

CHAPTER 4 1

The most accessible edition is Kaempfer’s Japan: Tokugawa culture observed, ed. Beatrice M. Bodart-Bailey, Honolulu, University of Hawaii Press, 1999. It is supported by an impressive

NOTES

2 3

4 5

6

7

8 9 10

11

12 13 14

15

16 17

321

critical apparatus, but omits the appendices of the original history. On Sir Hans Sloane’s acquisition of the manuscript and his role in the translation of the manuscript, see Beatrice M. Bodart-Bailey and Derek Massarella, The Furthest Goal: Englebert Kaempfer’s Encounter with Tokugawa Japan, Folkestone: Japan Library, 1995. Gulliver’s Travels: a critical study, Princeton University Press, 1923, pp. 70, 169. Reprinted in Japan in eighteenth-century English satirical writings, ed. Takao Shimada, Tokyo: Edition Synapse, 2007 The works of Dr Jonathan Swift, vol. VIII, pt. 1, collected by Deane Swift, Esq. Titles readily identified from the catalogues of major libraries. The introduction to Kaempfer’s Japan, p. 7, suggests however a total of 12 editions in the decade from 1727. R. Markely, The Far East and the English imagination, Cambridge University Press, 2006, p. 141. Gulliver’s Travels, ed. Herbert Davis, Oxford, Basil Blackwell, 1959), p.79. The suggestion by Maurice Johnson, Kitagaki Muneharu, Philip Williams in Gulliver’s Travels and Japan: a new reading, Kyoto, Amherst House, Doshisha University, 1977, pp. 2, 12 that Swift had also made use of the early-seventeenth account by the Englishman Will Adams, when briefly there was an English trade with Japan, gains support from Swift’s strictly speaking inaccurate account of an English master and his vessel returning in the early eighteenth century from Japan. Gulliver’s Travels, p.203. Gulliver’s Travels, pp. 193, 215. The full title is Amoenitatum exoticarum politico-physico-medicarum fasciculi V: quibus continentur variae relationes, observationes & descriptiones Rerum Persicarum & Ulterioris Asiae, multi attentione, in pergrinationibus per universum Orientem, collectae ab auctore Engelberto Kaempfero, Limgoviae, 1712. It is difficult to argue that Swift saw the work in Dublin, as the cryptic misunderstanding of Japan in Part 1, written in Dublin, would suggest otherwise. Kaempfer’s Japan, pp.44–5. The passage later discovered is the Bering Strait For contemporary and later knowledge of Tartary, see Marc C. Elliott, “The limits of Tartary: Manchuria in Imperial and National Geographies,” The Journal of Asian Studies, 59, no.3 (Aug., 2000, pp. 603–646. Kaempfer’s Japan, p.46 Gulliver’s Travels, p.130. John Smith, a contact of Molesworth’s, who opened a book shop in Dublin in 1724 and whose impressive catalogue in 1726 contained books “newly arrived from England, Holland and France” could conceivably have been a source of information on the prospective publication of Kaempfer. On Smith, see M.A. Stewart, “John Smith and the Molesworth circle, “Eighteenthcentury Ireland, vol. 2 (1987), p. 102. Maurice Johnson, Kitagaki Muneharu, Philip Williams, Gulliver’s Travels and Japan: a new reading, p. 12. Markely, p.255. I am indebted to Professor Bodart-Bailey for drawing my attention to this work as the first published statement of Kaempfer’s knowledge of Japan.

CHAPTER 5 1

2 3

The paper was presented at the public session of the conference on Japan in the age of the Pax Britannica and Pax Americana at the Kokusai niihon bunka kenkyu senta (International research centre for Japanese studies), Kyoto in September 2002. G.M. Golownin, Recollections of Japan (London, 1819), pp.32–33. Now available in a 4-volume translation under the title The iwakura embassy, 1871–73: a true account of the ambassador extraordinary and plenipotentiary’s journey of observation through the United States of American and Europe, (Matsudo, The Japan documents, 2002, Chiba).

322

4

5

6 7 8

9 10

11

EARLY JAPANESE TRADE

On the evolution and interpretation of the policy see L.M.Cullen, A history of Japan 1582–1941: internal and external worlds (Cambridge, 2003)„pp. 146–48, 170, 173, 288, 307, 310. As for the uchi harai policy itself, the actual incidents need a close study taking n account of all the circumstances. Reprinted by Kennikat Press, Pert Washington, N.Y., 1970. Noel Perrin, Giving up the gun: Japan’s reversion to the sword, 1543–1879 (Boulder, Colorado, 1979) See Cullen, op.cit., p. 4; and the penetrating article by Katō Yūzō, “The opening of Japan and the Meiji Restoration, 1837–72”, In The history of Anglo Japanese relations, vol. 1, The politicaldiplomatic dimension, 1600–1930, ed. Ian Nish and Kibata Yoichi (Basingstoke and London), 2000), pp. 60–85 Public Record Office, London, Satow papers, 30/33/11/8, Gubbins to Satow, 22 January 1917. On Sun Yat-Sen and Japan, see Marius Jansen, The Japanese and Sun Yat-Sen, (Harvard, 1954), reprinted Stanford, 1970. This was reflected also in debate on international trade in opium, a subject of major concern in the League of Nations in the mid-1920s. See Goto-Shibata Harumi, “The international opium conference of 1924–25 and Japan”, Modern Asian studies, vol. 36, no. 4 (2002), pp. 981–983

PART II CHAPTER 6 1

2 3

4

5 6 7

8 9 10 11 12 13 14 15 16 17 18

I am greatly indebted to Professor Saitō Osamu. Without his advice on sources and help in getting copies of some of them, as a non-demographer I would not even have attempted the topic. I am indebted to the staff of the Osaka-shi shiryō hensanjo for bringing the papers of Kōda Shigetomo to my attention, and to Professor Saitō and Dr Takagami Shinichi for help in tracing them. Ms Takahashi Nanako, Hitotsubashi University Library, was very helpful during my examination of them. I am indebted also to Dr. Katsuta Shunsuke for much help with proper names and place names, and advice on difficult passages of text and on some problems of identifying sources. I am alone responsible for the errors that remain. For some comment on these issues, see Cullen 2003, pp 3–5, 160–164. Sekiyama 1958, pp. 97–8, 108. Katsu’s work appeared in 1890. Sekiyama quotes the work as 1887, which was the date of Katsu’s preface. However, Sekiyama apart, the year 1890 is universally cited. Suijinroku 1890a, pp. 289–332, “Jinkō oyobi kokudaka no bu.” It appears in part 1 of Suijinroku. Suijinroku itself takes up five substantial volumes of Katsu Kaishū zenshū (vols. 6 to 10). The final part of Suijinroku contains some data on the town population of Edo (reproduced in ibid., vol. 10, pp. 230–238) and of Ezo (ibid., pp. 323–354), Droppers, pp. 253–284. Inoue 1904. Honjō 1972. Honjō 1935 provides the best early account in English of the statistics and of their limitations. Sekiyama 1958. Minami 1978. Takahashi 1955a. Takahashi 1955b; Takahashi 1962. Sekiyama 1958, p. 97. White 1992. Ōhashi 2000, pp. 69–99. Ibid., p. 70. Hayami 2001, pp. 43–4. Ibid., p. 28. Hanley and Yamamura 1977, p. 42.

NOTES

19 20 21 22 23 24 25

26 27 28 29 30

31

32 33

34 35 36

37 38 39

40

41

323

Hayami 2001, p. 36. Cornell and Hayami 1986, p. 318. Ibid., p. 326. Hanley and Yamamura 1977, p. 348, based on Sekiyama 1958, pp. 33–34. Cornell and Hayami 1986, p. 318. Takagi 2004, p. 9. Even in much more evolved administrative societies such as eighteenth-century Ireland, collection of indirect taxes or excise and of the hearth tax (a tax on households) became relatively static. In the case of the hearth tax, collectors worked from registers. It was easy to take the existing registers as the base of their work (quite apart from the issue of officials accepting bribes), rather than to resurvey in radical fashion. Hayami 2001, p. 60. Ibid., p. 28; Cornell and Hayami 1986, p. 313. Cornell and Hayami 1986, p. 313. Hayami 2001, p. 30. The fullest data are in the table in Mito-shi shi 1968, p. 70. Figures are available for town population and for otera jisha ryō (temple and shine domains) only for very isolated years. Tables in other publications are less complete reproductions of the same data. For purposes of comparison it should be noted that the figures for 1834 and 1864 are inclusive of town and temple population. From as early as Genroku 10, counts of cattle and horses existed. Such counts seem to imply that by that time Mito counting had evolved from shūmon aratame chō into ninbetsu aratame chō, and certainly from 1721 the surviving village reports themselves were denominated as ninbetsu chō. See Mito-shi shi 1968, pp. 70–74. Hayami 2001, p. 25. From p. 26 it is clear that the surviving records are those of Kōriyama. This would also explain the unchanging sex ratio which has sometimes been noted for figures for han population. For births, deaths and migration, see Takahashi 1962, pp. 58–60, 61–63, 136–7, 150. Takahashi 1962, p. 136. Hayami, in referring to figures from the population registers for a later date—early Meiji times—notes that “vital statistics such as births and deaths can sometimes be difficult to accept as numerical values, so it is dangerous to calculate the fertility and mortality rates directly from these data. Since the figures for both births and deaths seem to have been under-estimations, calculating the percentage would result in extremely low figures for some regions.” Hayami 2001, p. 61. Hayami 2001, pp. 79–82. Takagi 2004, p. 6. The available Mito and Hitachi figures however do not necessarily always follow an identical definition of coverage of population. The surviving data for Mito han are of rural population only, with the exception of Tenpō 5 when they include monks and town population. The change in that year may account for the contrast in trend between Hitachi figures and Mito figures for that year. The contrast would seem to imply that the returns for the kuni were of rural population only, a fact which is consistent with the evidence that the surviving aggregates for Mito han for most years are of rural population. However this conclusion, with its fragile base, is itself tentative. A grouping denominated as Mutsu in the shogunal return contained the figures for the kuni of Mutsu, Iwake, Iwashiro, Rikuzen, and Rikuchū, and a further grouping denominated as Dewa combined the two kuni of Uzen and Ugo. The use of the terms Mutsu and Dewa in the census returns seems to have been based on an archaic and earlier definition from as early as the fifteenth century of the kuni of Mutsu and Dewa. At that stage the north of Honshū was greatly underdeveloped compared with Tokugawa times. Its territorial units later changed, but for some purposes the use of the archaic provincial units persisted. Hanley and Yamamura 1977, pp. 48–49.

324

42 43 44 45

46 47 48 49

50

51 52

53 54 55 56 57 58

59 60 61 62 63

64

65 66 67 68 69 70

EARLY JAPANESE TRADE

Ibid., p. 151. Honjō 1935, pp. 145–158; Hanley and Yamamura 1977, pp. 39–68, 182–3. Hayami 2001, p. 44. For the source see note 41. Hanley and Yamamura also suggest that Bizen returns to the shogunate excluded towns. See note 39 above. Hanley and Yamamura 1977, pp. 149, 150, 151, 152. For such use of language, see e.g., Takahashi 1962, pp. 4, 6, 174, 189. Fuller information arising from han calculations survives for the years 1683, 1712, 1752 and 1790. See below. Takahashi 1962, pp. 3, 174, 189. The first nisshi or zassho is from 1644, but the first population count dates from 1653, the next from 1680 and with much greater detail from 1683. The nisshi are supplemented by 118 satsu of gozai fudome (records of temporary absences from han) from 1674 to 1865. Combining the two categories there are in all 310 volumes. A full list of the nisshi is given in Takahashi 1962 (pp. 4–6). A small number of other documents from 1784 to 1881 exists (listed in Takahashi 1962, pages 6–7). Mori 1934, pp. 75–6. Takahashi 1962, pp 195–200. Takahashi himself did not say that two sets of figures exist. He simply made the point that the nisshi were an indirect source of population data, i.e., that they were drawn from other documents (Takahashi 1962, p. 6). However, the fact that only the higher-bound figures from the nisshi appeared in his tables, could lead, along with the imprecise language and the very general or vague descriptions of sources, to other conclusions. The volume of Iwate-ken shi, appearing a year later, which alone gives both lower and upperbound figures, at no point elaborates directly on the significance of the fact. Hanley and Yamamura 1977, p. 150. Ibid., p. 154. Ibid., p. 150. Mori 1934, pp. 75–6. Takahashi 1962, p. 4, 6. The sources are indicated in Iwate-ken shi 1963, pp. 643,659. It is noted than the source materials for the history of Iwate-ken, entitled Ryōnai jinkō korui chō (p. 659) or Nanbu ryōnai jinkō chōsa (p. 643), were drawn from the han nisshi. There is no analysis in notes to the table of the distinction between the two figures for households. It could be argued of course that it can be deduced from the table (though that is asking a lot of the reader). However, even if that were the case, it leaves undiscussed the abrupt appearance of two figures for total population in and from 1756, Iwate-ken shi 1963, pp. 641–643, 655–659. Ibid., pp. 622, 632. Ibid., pp. 660–1. Ibid., pp. 676–7. For the full details of the 1790 count, see pp. 677–684. Ibid., p. 685. For town population, see also Hanley and Yamamura, pp. 151–2. Apart from town population, mining communities were involved. Takahashi 1962, pp. 199–200. Further figures with very minor variations are given in parentheses. Data for earlier years are in Takahashi 1962 and in Iwate-ken shi 1963. Takahashi 1962, pp. 191–95 Iwate-ken shi 1963, pp. 653–4 Takahashi 1962, p. 174. Hanley and Yamamura 1977, pp. 148–50. Cullen 2003, pp. 100–101. It was less sharp than suggested in Cullen 2003, p. 101, in which I drew on a less complete table of population counts in Mito-shi shi 1984, p. 523. For the fuller table, see Mito-shi shi, middle series 1968, p. 70. Takahashi 1955b, pp. 173–74, has a table with figures for various years in 1697–1822. Mito figures for 1834 and 1864 include town population, and are thus

NOTES

71 72

73 74

75

76

77

78

79

80 81 82 83 84

325

not directly comparable with preceding ones. This further hints at the somewhat erratic and incomplete survival of population counts for Mito. Hayami 2001, p. 30. See table in Mito-shi shi 1968, p. 70. It should be noted that counts are missing for two shogunal census years, Kan’en 3 (1750) and An’ei 9 (1780). Seya and Toyosaki 1973, pp. 198–99. It is unlikely that there was a rise in fief holders. It is more likely that the decline reflected an increase in holdings by collateral branches of the ruling family, which put their rice production capacity beyond the reach of the central authorities of the han. The figure for 1834 is 242,939. The figure of Genji 1 (1864) is 274,908, but this figure may be erroneous and the true figure should read 244,908. Mito shi-shi 1968, p. 71. The figures were edited in the past from mss in the Kōchi Prefectural Library that were destroyed in the the Second World War. Takahashi 1962, p. 3. An error renders the year Tenpō 5 (1834) as 1822. The figures are also reproduced in Takahashi 1955b, pp. 229–35. Figures from Aizu’s Matsudaira daimyo family in “Aizu kasei jikki.” See Takahashi 1962, pp. 3–4, and data reproduced in graph between pages 170 and 171. Takagi 2004, pp. 17, 128. Takahashi gives total figures for the following years. 1685

816,061

1742

816,061

1756

594,637

1764

648,486

1786

596,282

1801–4

639,070

1825

687,802

1827

697,046

1828

699,334

Takahashi 1962, pp. 44, 53. There are figures with minor variations on pages 42 and 52. If we deduct approximately 200,000, the estimated fief population, Takahashi’s figures are close to those given by Takagi for gunkata. Figures for twenty-three years are also reproduced on a graph in Takahashi 1962 between pages 170 and 171. The falls are not necessarily due solely to excess mortality; increased mobility was likely as people, at least the younger, able-bodied, or simply the more desperate, temporarily moved within or even out of the han. This would also help to explain the relative rigidity in the number of households as opposed to inhabitants. The conclusion by Saitō Osamu that “a close look at the evidence reveals that it was not necessarily because famine heightened mortality levels but because it tended to further reduce fertility whose background levels were already low” (Saitō 2005b, p. 24) should also be borne in mind. It would resolve some of the problems that otherwise arise in reconciling contemporary comment and population trends in Tokugawa Japan. Cullen 2003, p. 100. Hayami 2001, p. 50. The population of the kuni at each census is given in Sekiyama 1958, pp. 137–39. Hayami 2001, p. 50. In the Tōhoku the Mutsu aggregate alone comes close to the Mito trend. While one could simply regard the figures as statistically suspect, it seems that the Mutsu data like those for Mito in the Kantō represent some juggling with the range or coverage in population figures. That may reflect concern over the pattern of the distribution of land between fiefs and kura

326

85

86

87

88 89

90 91 92 93 94 95 96 97

98

99 100

101 102

103 104 105 106 107

EARLY JAPANESE TRADE

or domain. The Nanbu han data hint at the type of statistical challenge faced by officials. It is doubtful if the data point to an ongoing demographic crisis. The internal figures for Sendai and Nanbu han, both component regions of Mutsu no kuni, point to an underlying stability. Hitachi is all the more interesting because of its very nonconformist profile in the mideighteenth century. Climate alone can hardly have been a factor. Tōhoku population did not fall between 1828 and 1834. The fall in Hitachi population in 1834 is at variance with a stable pattern in the Tōhoku and in some but not all of the kuni in Kantō. Mito apart, Hitachi consisted largely of tenryō. The presence of tenryō lands did not guarantee a stable statistical profile (see note 87 below). Kazusa, despite being largely tenryō, fell more sharply than any han in Kantō in 1834; its huge rise in 1846 (an increase of about a quarter on the earlier return) had no parallel in Japan; and it then recorded one of the three sharpest rises in Kantō in 1846–1872. Its highly unstable pattern suggests a problem of data, or more probably of arbitrary changes in presentation by shogunal officials. In contrast to both fudai and tozama, who had their permanently resident officials in the han, tenryō lands were administered by daikan whose base was in Edo. Tenryō lands are usually seen as having been administered in a more benign way that han territories. It may also have been on occasion a case of benign statistical neglect. Problems or uncertainties in dealing with fief holds and possibly with “honored” fiefholders (fiefholders who held a fief in name only and were paid from the daimyo’s kura) may have been a factor in accounting for the vagaries. McEwan 1962, pp. 94–5. On this problem, see Cullen 2003, p. 99, 127. Even Lafcadio Hearn, so favorable in almost all of his comments on Japanese life and of course with experience of school-teaching, observed the importance of developing “the mathematical faculty. At present this is the weak point, hosts of students being yearly debarred from the more important classes of higher study through inability to pass in mathematics.” Hearn 1984, p. 271. Hayami 2001, pp. 43–4. Deshima dagregisters 1990, p. 39. See Ōsaka-shi shi 1990, p. 198. Suijinroku 1890b, pp. 306–7. Katsu also gave a figure for Nagoya in 1692. Sekiyama 1958, p. 232. Ibid., pp. 235–37. See table in Kōda 1972 between pp. 248 and 249. For the guesstimates, see Sekiyama 1958, p. 228. They are very broad indeed, and lack any firm statistical basis. However they are the only working figures possible. They also appear to imply that all shogunal retainers resided in Edo. This is a simplification, although it is not unfounded, in the sense that they mostly operated from an Edo base. Broadly speaking, many officials, e.g., daikan, must have spent much of their time outside Edo. However, their families remained in permanent residence in Edo. Kato and Toyama 1984, pp. 193–96. The figures are taken from “Fuji-ke monjo shoshin ni ninzū kazoeru” ⸨ᐙᩥ᭩᭩ಙ࡟ேᩘᩘ࠼ࡿ. Hayami 2001, p. 54. Household figures for years from 1832 are reproduced in Kōda 1972, table between pages 248 and 249. Figures for some of the earlier years up to 1743 are given in Suijinroku 1890b, pp. 230–34. The complete table is in Ōsaka-shi shi 1990, p. 199. Tenpō sen’yō ruishū, Kōda, 1972, p. 262. This is a collection of copies of documents which circulated between the machi bugyōsho and kanjōsho, now held in National Diet Library. Sekiyama 1958, pp. 231–32. Kōda 1972, pp. 244–65. Ibid., pp. 245–46. Ibid., p. 246. Suijinroku 1890b, vol. 10, pp. 227–38.

NOTES

108

109 110

111 112 113 114 115

116 117 118 119 120

121 122

123

124

125 126

127 128 129

130

131 132

133 134

327

The data for 1713 are not given in Kōda’s table, as he limited his table to the period from 1721, the year which marked the first national census. See Kōda 1972, p. 262. See note 102 above. The other is the return for all Japan which Minami Kazuo dated to 1840 (see note 163). See also note 102 above for a return for the city of Edo on its own in 1840. While the 1840 reports are the last surviving figures drawn from the sextennial census exercise as such, population data, as drawn from the population registers which exist for Edo for 1841 and for seven later years. From Suijinroku 1890b, pp. 227–38. Suijinroku 1890b, p. 237. This is recognised by Kōda, 1972, p. 265. Ōsaka-shi shi 1990, p. 198. This page has an account of vagaries of other estimates. Ōsaka-shi shi, vol. 1 (1913), pp. 370–71, 483–83, 602, 880–81; vol. 2 (1914), pp. 107, 180– 01, 546, 758–59. The sources, though sometimes not identified in the original Ōsaka-shi shi, are identified specifically in the modern Ōsaka-shi shi 1990, pp. 198–200, as “Gyokurosō” ⋢ 㟢ྀ, “Matsudaira Iwaminokami-dono ohatsuiri ni tsuki sashidashitaru oboegaki” ᯇᖹ▼ ぢᏲẊᚚึධࢽ௜ᕪฟᚚぬ᭩, “Gojōdai goshihaisho yorozu oboe” ᚚᇛ௦ᚚᨭ㓄ᡤ୓ ぬ, “Nanboku ryōmachibugyō rensho kakiage” ༡໭୧⏫ዊ⾜㐃⨫᭩ୖ, and “Chihōyaku tekagami” ᆅ᪉ᙺᡭ㚷. Described by Kōda 1972, p. 244. So described, but obviously the “Chihōyaku tekagami” referred to elsewhere. Ōsaka-shi shi 1990, pp, 198–201. Ōsaka hennen shi 1969, pp. 278, 510–11. I am indebted to Ms. Takahashi Nanako for much help when I consulted this document, also for information on Isshiki and for confirmation that the handwriting was his. For wider information on the collection see Takahashi 2003, pp. 29–43. See Cullen 2003, pp. 76–77. Kōda’s account of the figures from the document in the reprint of his article on Edo population is somewhat coy. While one of the four series of subsidiary counts is for the eta population, and is so described in the original Ōsaka-shi shi, it is designated elliptically as “isson” in Kōda’s article (Kōda 1972, p. 244). The population for the city is the addition of the eta population to the figure for Sango; the other three townships are not counted as part of the population of Osaka proper. Data for 66 years are reproduced in Ōsaka-shi shi 1990, p. 199 plus one one year, 1689, in Ōsaka hennenshi 1969, pp. 278–79. Some data were given in the first Ōsaka-shi shi, vol. 1, pp. 482–84, 602–03, There are also some data for two years, 1689 and 1703, in two documents whose demographic detail are reproduced in Ōsaka hennen shi 1969, pp. 278, 510–511, and for 1681 in Suijinroku 1890b, p. 307. On the use of the 1689 date for for comparative purposes, see Saitō 2002, p. 146. Kōda 1972, p. 262. They include Ezo. The Ryukyu Islands are however excluded. This further illustrates the fact that the Ryukyus, despite multiple ties, were outside the Japanese polity. See Cullen 2003, p. 50. Excluding however the returns by Ōta of the population of a mere ten kuni for 1721. See note 155 below. Some tabular presentation of currency and trade data existed. I am preparing a paper on the statistical data for trade and their circulation in Tokugawa times. Inoue Mizue 1904, pp. 127–129. He regarded the document as relating to a census of Kaei 5 (1852). See note 139 below. For a more recent source with the same concerns which also influence interpretation, see Kawaguchi 1996, pp. 151–168. Golownin 1819, pp. 221–2. Ōta 1976, pp. 579–583.

328

135 136

137 138

139 140

141 142

143

144 145 146

147 148

149

150

EARLY JAPANESE TRADE

Suda Akiyoshi 1954. The data are also reproduced in Takahashi 1955a, pp. 92–93. The copy containing the data is itself from the 1830s, It is not chronologically the earliest surviving document with a breakdown: the data for two others censuses (1750 and 1756) were reported in a source at a date (1770s) closer to the actual date of census taking. The data are attributed in Sanka manroku to a named wakaitoshiyori and kanjō bugyō in Kyōhō 7-en 8-gatsu. (This information is not contained in Suda’s article, but, acquired from a private communication from Suda, it appears in Sekiyama 1958, p. 110, note 13, and also p. 101.) The data for sex distribution are lacking and, while the population of individual kuni is given a grand total for Japan is not given. However, the figure, when the kuni data are aggregated, is identical to that given by Ōta. The title of the text containing “Kokuchū ninzū kokudaka no koto” ᅜ୰ேᩘ▼㧗ࡢ஦as reproduced in the Sanka manroku also includes reference to the census of Tenpō 5 (1834) While now missing, the implication is that the document must have been compiled in the wake of the 1834 census. For source, see note 134. Katsu Kaishū in his summary listing of census totals, in regard to 1732, added the comment that Ishikawa Sōjirō (a Kanjōsho official of late bakumatsu times) had data for the kuni, and that according to a “certain book” (not identified by Katsu) the data had been reproduced in a further book called Burin inken roku Ṋᯘ㞃ぢ㘓. In modern times Sekiyama, who eventually saw copies of the latter book in the University of Tokyo library, noted that it contained only stories about Tokugawa warlords, and had no demographic data. Sekiyama’s conclusion that Katsu was mistaken is itself perfectly reasonable (Sekiyama 1958, pp. 102–103). But, as elsewhere Katsu gave full figures where they were available—for 1804 and 1846—and his comments suggest that he was simply recording his awareness of the existence of a fuller source which was not accessible to him, and that the error in regard to Burin inken roku may have occurred in the unidentified “certain book.” In other words he was anxious to record that a source covering all the kuni appeared to exist. The case may imply his thoroughness rather than any carelessness or direct error on his part An implication too is that in post-1868 times at least two sources had survived for the 1732 census: a source for the total population given by Katsu (as in all his census figures for Japan as opposed to his figures for Edo on its own its documentary origins unidentified), and a further document which he did not see, and which for that reason he uncharacteristically appeared to identify. Yokoyama has a variant figure (Sekiyama 1958, p. 102). Ibid., p. 97 and footnote 6, p. 110. According to Sekiyama, the editor drew on information both in a paper in Asakusa Bunko and in a book by Kurokawa Mayori. Ibid., p. 103. On the source, see also note 141 above. He also gives data for 1756. See under 1756 in table, and note 144 below. Suzuki gives no indication of sources. The formal census introduction, names of officials and kokudaka are not given. Sekiyama notes that the details in general correspond with the figures in Kanchū hisaku, but the source he believes to be of independent origin (Sekiyama 1958, p. 103). However as Kanchū hisaku was published in Meiji 12, there remains the possibilty that Suzuki’s text is a copy. On variants in figures for this year, see Sekiyama 1958 pp. 103–4. Ibid., p. 104. This report contains the kokudaka of all kuni. While containing the normal formal introduction, it lacks the signatures of ōmetsuke and kanjō bugyō. Ibid., p. 104. Matsuura 1978. The figure was quoted in Sekiyama 1958b, p. 111, note 18, from an earlier edition edited by Yoshikawa Hanshichi and published by Kokusho Kankō Kai (1910–11), vol. 3, p. 334. According to Sekiyama, the information was probably acquired by the daimyo’s intimacy with a shogunal daikan (ibid., p. 105), but this seems to be supposition. See notes 148 and 149.

NOTES

151

152

153

154

155

156

157

158

329

For a description of this document, see Sekiyama 1957, pp. 61–74. Sekiyama uncovered it in the Monbusho archives. It was originally in the possession of the Machida family in the village of Mimase, Saku gun, Shinano province. There is no indication of the source from which the family obtained the data. Some internal errors may suggest that it was already a copy from another source, itself not only a copy but possibly already a defective one. It has names of officers and gives the kokudaka. Minami Kazuo simply notes the location of this document, under the title Shokoku ninzū chō, as Kokuritsu Shiryōkan (Minami 1978, p. 165). Variant figures are given by Komiyama. These were later accepted by Honjō Eijirō. But this is, according to Sekiyama, a mistaken dating to 1804 instead of 1816. Sekiyama 1958, p. 106. First published in Takimoto 1930a, pp. 179–188, it has been the subject of later commentary by Sekiyama and Minami. There has been much confusion over its dating. The date has been identified in Sekiyama 1958, p. 106, as 1804 as it corresponds to the period of service of the two named officials. As well as the names of the officials the report for this year also contains the kokudaka. Minami has described the source as a Meiji copy, with the title “Shokoku ninzu chō,” held in in the Naikaku Bunko, originating from the Shūshi-kyoku ಟྐᒁcollection (Minami 1978, p. 165). As a collection from Tokugawa times, this suggests a possible origin at that time and hence that it may be at one or more removes a copy of an official copy. Not a return for 1822, but a figure derived from a return recording changes in population in the 1828 census. See note 157 below. Not a return for 1822, but a figure derived from a return recording changes in population in the 1828 census. See note 158 below. See note 151 above. The document gives the kokudaka as well as the population for each kuni, the formal introduction and names of the officers. Takimoto in Takimoto 1930b drew the data from Tokugawa rizai kaiyō, published by the Ministry of Finance. These are some summary data attributed in the Tokugawa rizai kaiyō to a source entitled “Chōkaiki.” On Sekiyama’s suggestion that they originated in the Ministry of Finance, see note 158 below. As the Nihon zaisei keizai shiryō 1922b was compiled under the auspices of the Ōkurashō, Sekiyama observed that the ultimate origin was the Kanjōsho (Sekiyama 1958, p. 107–8 and note 23, p. 111–112).). He also suggested a common origin in the Kanjōsho for the figures in the “Ninbetsu aratame okanjōsho chōsho jō no utsushi” in Nihon zaisei keizai shiryō 1922b and the “Chōkaiki” version in Takimoto 1930b (see note 157 above). An origin directly in the Kanjōsho itself may be ruled out by losses by fire in 1859 in the more central of its two stores. It should be noted that Nihon zaisei keizai shiryō does not profess to draw on sources originating in the Kanjōsho itself. The volumes are the fruit of a work of compilation from daimyo and private sources, begun in 1878, by some ten or more collaborators (See Nihon zaisei keizai shiryō 1922a, foreword by Kuroda Hideo 㯮⏣ⱥ㞝, head of the banking division of Ōkurashō, and preface by Takimoto Seiichi). While the meagre data available in the two published sources Nihon zaisei keizai shiryō and Nihon keizai taiten, vol. 54 (national total, totals for women and men, and two totals, one for the aggregated population of kuni whose population increased and one for those kuni whose population decreased between the census years 1822 and 1828 which in turn yield the net national increase) would support Sekiyama’s inference of a common source for the versions in “Chōkaiki” (Takimoto 1930b, p. 392) and “Ninbetsu aratame okanjōsho chōsho jō no utsushi” (Nihon zaisei keizai shiryō 1922b, p. 1193), they both are very remote from being fair copies of a reliable document (which make even less plausible the implication of an origin in the Kanjōsho. While the very uncommon mode of giving the change in population (separate totals for those kuni showing an increase over the preceding census year and for those kuni showing a decrease in the same interval) suggests a common origin, the fact that the documents give different gross totals for the population of Japan suggests that they are separate transcriptions, with independent errors made in the process of transcribing of a now unknown common document. Whether that document was itself a copy of an earlier document, or a compilation by someone with access to figures for two census years, is a matter of pure speculation. The general implication

330

159

160 161

162

163

164

EARLY JAPANESE TRADE

is however that the brief and defective summary accounts in Takimoto 1930b and Nihon zaisei keizai shiryō 1922b rest not on a document in the Kanjōsho but on copies made at a remove from the Kanjōsho and found in a trawl of daimyo and private records. While the original data would have been official, the documents used in preparing the compilation were already private ones. It should be added that Takimoto superintended the publication of the final Nihon zaisei keizai shiryō, and that his Nikon keizai taiten is an offshoot of his work on Tokugawa sources. The whole operation was thus the counterpart to the work that Katsu Kaishū conducted. For the dating of this document, see Sekiyama 1958, pp. 107–8. Inoue wavered between the dates of 1828 (Bunsei 11) and 1853 (Kaei 5), for these data, but as they are identical to figures for 1828, they are, in Sekiyama’s view, properly census figures for that year. The date written in the title of the document is not Kaei 5, but Kaei 6, which could be taken as the year of transcription, rather than of a census. Inoue reproduces totals at national and kuni level for men and women; but he does not give the kokudaka, text of a formal introduction, or the names of the officers. Sekiyama noted that the 1828 dating appears to be confirmed from a work published in Meiji 14, Gajikan gasho ᡃ⮬หᡃ᭩, reproducing a text entitled “Bunkyō kō jitsuroku” (Sekiyama 1958, pp. 107–8 and note 24, p. 112). For further details, see ibid., pp. 107–108 and footnote 25 (p. 112). It contains the text of the formal introduction to the census, the names of the two officers, and the kokudaka (ibid., p. 108). Komiyama, who first introduced the document, gave a total only (Sekiyama 1958, p. 108). For the fuller census details, see Shinozaki 1917, pp. 14–15. The report gives figures for men and women and also the kokudaka for the kuni. It does not give the formal introduction. In the source as outlined by Shinozaki Ryō the ōmetsuke’s name occurs; the impression is that Shinozaki’s data are drawn from a fuller source, which is likely to be the same source from which the text reproduced by Takahashi 1955b, p. 333–343, was drawn. (See note 164 below.) Takahashi noted that in Shinozaki’s article the figures for men and women for eight kuni differed slightly, if added up, from the gross total for men and women as given for each of those kuni; the source in his view was therefore different from that for the text he reproduced. However, it is more likely that a simple error or misreading was made in transcription in these instances. The author, Takahashi Bonzen, reproduced a full text in Takahashi 1955b, pp. 333–343 from a document in the National Diet Library. As this copy came from the old Ueno toshokan, which contained many documents surviving in early Meiji times, it is likely to have been a copy held by some official in the final years of the Tokugawa period or in the very early Meiji years. It contains the formal introduction, names of the officers, and, for the kuni, detail of both kokudaka and population. This copy lacks a date in the heading or title of the document. However, within the corpus of information the actual dating is cited. For fuller details, see Minami 1978, pp. 164–8. The copy includes the names of the officials, and the kokudaka.

CHAPTER 7 1 2 3

4 5 6

See Nagazumi 1987 for the most informative account of the data drawn from Dutch sources. Cf. Nagasaki-ken shi 1985, pp. 553–58. Best illustrated today in the celebrated Port Books in the Public Record Office, London, which cover the sixteenth to eighteenth centuries. Based on data in Cullen 2003, pp. 41–47. Nagasaki-ken shi 1986, pp. 789–90; Matsuura 2007, p. 267. Cullen 2003, p. 44. The calculations are based on the assumption that Japanese currency was 80 per cent pure. If the proportion were higher, Japanese values converted into sterling would be higher, but my conclusions would not be radically altered.

NOTES

7 8

9

10 11

12 13 14 15 16

17 18 19 20 21

22 23 24 25

26

27

28 29 31 32

331

Cullen 2003, p. 84. Hirakawa 1994, pp. 48–73. The quantity were said to be 2,600,000 (sic) to 700,000 hyō ಥ, or bags, of dried sardinesᖸ㫇, though the quantity fell later (p. 49). 260 man ୓seems to be a misprint for 26 or 20 man. Vessels had a carrying capacity as high as 1000 koku (though most in fact were much smaller than this figure). Cullen 2003, p. 83, and note 51. There is a problem in regard to shipping capacity. Measures of burden underrated carrying capacity. Ships of up to 1000 or more koku engaged in the coastal trade. Hence depending on a count of either 500 koku or 1000 koku per vessel, the rice trade required 3000 to 6000 vessels. The higher figure seems unlikely. Miyashita 1997, pp. 250–67. McClain and Wakita 1999. Hayashi 1994, pp. 211–33; McClain 1999, pp. 44–79; Flershem 1964, pp. 405–416. Takase 1966, p. 61. BPP, vol. 5, p. 101. A figure of 3340 vessels is given in BPP, vol. 4, p. 615. The difference may be accounted for by the latter figure being given for a Western calendar year whereas the former figure is for a Japanese calendar year. But see Yamamoto 1994. Ravina, 1999, pp. 158, 159, 161–62, 168–75, 182–86. Miyamoto 1985. Wigen 1995, p. 56. The words tonya ၥᒇ (wholesaler) and kabunakama ᰴ௰㛫 (literally, stock association) crop up repeatedly and almost interchangeably in modern literature. In administrative documents of the Tokugawa period, the word tonya was the most common term used in relation to statistical formalities. To become a member of a guild it was necessary to purchase a kabu, or share, which conferred the rights of guild membership on its holder. See BPP, vol. 4, pp. 624–25, “Memorandum by Mr Aston on the commercial system of Osaka.” Aston gave the number of guilds as 200, a figure which, for guilds as opposed to tonya, seems in error. Ōishi 1995, pp. 166–71. Cullen 2006, p. 150. Ackroyd 1979, p. 247. In 2002, the OCCI scaled down the operations of its library, which in any event does not seem at earlier dates to have given ready access to the archives. This has copper-fastened the situation. Professor Ōishi Shinzaburō ኱▼ៅ୕㑻 once commented that this library was extremely inconvenient for consultation (“Etsuran ni hijō ni fuben de atta” 㜀ぴ࡟㠀ᖖ࡟ ୙౽࡛࠶ࡗࡓ) (Ōishi 1966). I am very grateful to Ms. F. Moriwake of the OCCI for much assistance in regard to the records of the Chamber. The publication project arose out of the association between a senior managing director of the OCCI, Satoi Tatsusaburō 㔛஭㐩୕㑻, who was interested in the Chamber’s historical materials, and Miyamoto Mataji ᐑᮏཪḟ (1907–1991), a professor of history in Osaka University, who already had close links with OCCI. Miyamoto had written an account of the Chamber in 1955, and he treated it again in a book published in 1963. The preface by the President of the Chamber in the OCCI Supplement 1966, itself a fanfare for the completion of the project, contains a warm note of thanks to Miyamoto, and from the internal evidence, the eleven-page introduction was written by Miyamoto. Sakudō Yōtarō స㐨ὒኴ㑻, Yasuoka Shigeaki Ᏻᒸ㔜᫂, Iwahashi Masaru ᒾᶫ຾, and Takashima Masaaki 㧗ᔱ㞞᫂. There is a brief account of the work of compilation in OSSS Supplement 1966, pp. 7–8. Hotta 2007, pp. 66–71; Ōsaka-shi shi henshū no hyakunen 2002. 30 Ōishi 1964, 1966, 1998. Miyamoto 1972, 1988. However, OSSS, vol. 13, p. 14 has a figure for daimyo rice of 1,123,070 koku and for zakkoku 㞧✐(minor grains) of 72,895 koku for 1714.

332

33

34 35

36 37 38 39 40

41 42

43

44 45

46 47

48 49

50

EARLY JAPANESE TRADE

BPP, vol. 4, p. 274, Sydney Locock, “Report on the Ports of Osaka and Hyōgo,” 10 June 1867. This report (pp. 271–84) is a most informative one. “Governor” was what the British called the bugyō, elsewhere in this article called “magistrate.” Ōsaka-shi shi 1911, pp. 639–86. In nine cases it was possible for Yasuoka to select comparable figures from the 1736 return to create a profile of Osaka trade for 1736, 1804–1830, and 1840 (Yasuoka 1960, p. 120). In other words the dates or spans of time, as they arise, are often approximations, and Yasuoka somewhat simplistically but up to a point defensibly based his approach on extending the varying periods chosen by the bugyō to the entire period 1804–1830 as a basic period of relatively higher trade levels (perhaps the fact that this gathered together neatly the years of the entire Bunka and Bunsei periods may have also been a factor in the choice) and settling on the year 1840 as the bugyō’s date for measuring the decay of trade. Ōsaka-shi shi 1911, p. 651. Ōsaka Dōjima beishō enkaku 1912. BPP, vol. 4, pp. 275–76. BPP, vol. 4, p. 356. It would have been relatively simple to expand an exercise of this sort into a count of Osaka imports. However, the respective roles of bugyō officials and tonya, and the question whether officials sometimes demanded fuller returns or more precise counts from tonya, remain matters of conjecture. Uncertainty remains also with regard to whether returns for different categories of vessel (e.g. rice on vessels above 200 koku, rice on lighters and on smaller vessels sailing directly to the wharves) might on occasion be further scrutinized by bugyōsho officials, as had been the case in Edo in 1727. See Saitō 1998. BPP, vol. 4, p. 276–78, Sydney Locock, “Report on the Ports of Osaka and Hyōgo,” 10 June 1867. Miyamoto attributes the figure to ୕஭ᩥᗜ, ㏆ୡᚋᮇ࡟࠾ࡅࡿ୺せ≀౯ࡢືែ, pp. 2–5 without giving a date for the volume. There are two editions of this work, one in 1989 (Mitsui Bunko 1989) and an earlier one in 1952 (Mitsui Bunko 1952). The main text is identical but the preliminary matter is different and the information on pp. 2–5 has been omitted. The original edition has estimates of han rice, in some cases in a spread between two figures rather than a single figure, and they add up to the gross total of 830,000–900,000 koku quoted in the text. The source for the figures is not indicated. The text also made a tentative suggestion, based on these figures for kura rice and for trade data in the 1736 statistics that imports in this period were in the region of 1,000,000 koku (kura rice about 800,000 koku, and trade rice 200,000 koku). Miyamoto 1972, p. 54. Suzuki 1938, pp. 482–88. Original sources were two: for 1823, “Ōsaka kurayashiki tsuki narabini kokudaka nobose mai tsuki” ኱ᆏⶶᒇ㗤㝃ᖼ▼㧗Ⓩ⡿㝃, and for 1828, “Shodaimyō keizu rokudaka Ōsaka kaimai roku” ㅖ኱ྡ⣔ᅗ⚘㧗኱ᆏᘔ⡿㘓. Both were preserved in private archives in Osaka. Kobayashi and Wakita 1973, p. 108. Ibid., p. 109. Their table lists only thirty-one commodities, of which twenty-six include values. Their total however is based on an enumeration not only of these twenty-six values but of other goods the values of which are not set out in their table. In Hayami and Miyamoto 1988, p. 232. To avoid any confusion it should be noted that the new Shinshū Ōsaka-shi shi ᪂ಟ኱㜰ᕷྐ reproduced the 1714 return on two occasions, one in a full form for imports only together with the 1736 return (Shinshū Ōsaka-shi shi 1990, pp. 12–13), and the other in a somewhat truncated form for both expots and imports in Shinshū Ōsaka-shi shi 1989, pp. 506–507. A statement by Kōda in an 1934 article that the 1714 return enumerated 91 items and gave a sole category of oil makes it clear that he had this return in mind (Kōda 1934a, p. 191). The book itself was first published in 1934, and the references to trade figures in the article remain

NOTES

51

52

53

54 55

56

58

59

60

61

62 63 64 65

66

67 68 69

333

unchanged in later editions. Kobayashi and Wakita 1973 (pp. 106–107) noted that Kōda had not presented the return in the Ōsaka-shi shi, but was aware of its existence when he wrote the paper “Kabu nakama” in 1934. Their citation of p. 304 as the page on which Kōda referred to the 1714 return is to the 1934 edition, not to the unchanged version in the later Kōda zenshū. See also Yasuoka 1960, p. 114. Ōishi 1964, 1966, 1975. For examples of reliance on the defective return, see Yamaguchi 1968, pp. 60–61, and Kobayashi and Wakita 1973, p. 107. Probably the date of the OCCI copy is close to 1903 or 1904, as the document was becoming known. These observations are based on study of the copy in the Mitsui archives in Tokyo and of the copy in the Ōsaka-shi Shi Hensanjo. I am greatly indebted to Ms Otsuki Yōko in Mitsui archives and to Ms Yoshida Hiroko in Ōsaka-shi Shi Hensanjo for much assistance in the course of examining both records. I am indebted to Mr Uchida Masahiro and Mr Yataka Kōno for assistance on a later visit in Osaka. Yasuoka 1960, p. 115; Yamaguchi 1968, p. 60. While the return is usually presented without addition of figures, Kobayashi and Wakita added up the figures for which values are given to give a total of 100,751 kan (1973, p. 109). Their table lists only thirty-one commodities, for which twenty-six include values. Their enumeration however includes also the values not set out in their table. Yasuoka 1960, p. 123. He offered the total of 1,000,000 koku without explaining its derivation. 57 He stated in his text that the amount of kura rice was 800,000 to 900,000 koku, and that the total import of rice was generally 1 million koku (p. 114). White cotton cloth is rendered variously as shirotewata ⓑᡭ⥥ or, more commonly, shirokiwata or shiromomen ⓑᮌ⥥. Kiwata is ambiguous: it can mean raw cotton and as such is identified in written records by its weight. Kiwata as cloth is identifiable in the records by being counted in tan, a unit of length. A source of confusion is that Yasuoka’s table of 1736 trade gives shirotewata only, whereas in a later table in the same article, without advertence to the fact, Yasuoka adds 32,723 tan of shimamomen ⦤ ᮌ ⥥ or striped cotton cloth to 1,178,391 tan of shiromomen, making a total of 1,211,154 tan (Yasuoka 1960). Hayashi 1969 (p. 193), accepted Yasuoka’s total in 1736 of 1,210,000 tan. Yasuoka 1990, p. 16. Yasuoka simply attributed the volume to the Ōsaka Shōgyō Kaigisho without indicating that it had been published in OSSS 1964A. Volume 26 was published in 1978, but the table was reproduced from a much earlier volume in the same series. Miyamoto 1972, p. 56. Miyamoto Matao attributed it in a footnote to OSSS 1964c. While that volume has some statistics on rice for later years, the 1766 figure for rice occurs in OSSS 1964a. Ōsaka-shi shi 1913, p. 651. BPP, vol. 4, p. 275, report by Locock on Osaka and Hyōgo, 10 June 1867. Katsu 1890, pp. 238–39. Ōishi 1998, p. 73, and footnote 13, pp. 101–102. According to Ōishi, the shahon is now in the Shiryō Hensanjo. It ends with the words “Meiji jūkyūnen shigatsu kazoku Mizuno Tadahiro zōsho outsusu” ᫂἞༑஑ᖺᅄ᭶⳹᪘Ỉ㔝ᛅᘯⶶ᭩ࣤ෗(a copy of notes from the writings of [nobleman] Mizuno Tadahiro in April of Meiji 19) and also gives the names of copyist and proofreader. There is also, according to Ōishi, a third or more modern copy of Kyōhō tsugan in mikan zuihitsu hyakushu ᮍห㝶➹ⓒ✀, vol. 17 (Ōishi 1998, footnote 13, pp. 101–102). This is part of a series of volumes on Edo customs and manners compiled by Mitamura Engyo ୕ ⏣ᮧ㬇㨶 (1870–1952), an amateur researcher of Edo manners and customs. A modern edition in 12 volumes was published in 1976–1978. Ōishi 1998, p. 74, has a table of the variations. See Ōishi 1998, p. 72. Ōishi 1998, pp. 79–80. report from Edo machi bugyō to rōjū, 13th year, 7th month, 17th day.

334

70 71

72 73

75 76

77

78

79

80 81 83 84

85 86

87 88

89 90

EARLY JAPANESE TRADE

BPP vol. 5, p. 62, vice-consul Dohmen to Adams, 15 February 1872. This total is different from Ōishi’s total (Ōishi 1998, p. 83). In arriving at his grand total, Ōishi seems not to have taken into account the deduction of 44,205 hyō from the original chōnin figure of 155,946 hyō. Ōishi 1998, p. 83. See also detail in map on p. 82. BPP, vol. 4, p. 275, report by Locock on Osaka and Hyōgo, 10 June 1867. 74 Tōkyō-shi shi kō 1926, pp. 13–99. Fujita 1987, pp. 248–253. There is a very large mathematical error in the document which confirms both that it is a copy and that an error of transcription occurred. Having noted that daily consumption was 7500 koku, the total for annual consumption is given as 700,000 koku. For a year of 365 days in fact the consumption would amount to 2,717,500 koku. Elsewhere the document notes that employment categories other than samurai and chōnin (e.g., fishermen), and other uses of rice (making sweet goods) would raise the consumption to above 3 million koku. While such a large estimate is debatable, it shows that the author of the original text was fully aware of a total well in excess of 700,000 koku. As the document almost invariably employs round figures, the most plausible way in which the error arose was that the total consumption was rounded by the original author to 2,700,000 koku and that the “2” was omitted by a transcriber (a sort of error which was painfully common in transcriptions in Tokugawa times). The figures appeared first in the old Ōsaka-shi shi 1914, pp. 839–40. The table is reproduced also in Ōsaka hennenshi 1978, pp. 331–34, where Hokkaisan niuke tonya kumiai enkaku shi ໭ ᾏ ⏘ Ⲵཷၥᒇ⤌ྜἢ㠉ྐ3 is cited as the source. The source is also cited in the introduction to the OSSS Supplement 1966, p. 1, which would seem to suggest that both Ōsaka-shi shi, vol. 1 and OSSS drew on it. In other words, an early book rather than a manuscript seems to be the only known source for the data. For details of imports see Ōishi 1998, p. 74. The 1726 details are reproduced as part of a table by Hayashi 1969, p. 192. For the sake of completeness, Katsu’s curious figure for zeni is included in the table in the text above, but is almost certainly an error in the text that he consulted (not known to us today), rather than one committed by Katsu himself. The Edo figure for 1856 is from Tōkyō-shi shi kō 1926. Yamaguchi’s figure (p. 61) for cotton cloth at 7,909,364 tan is somewhat different from Hayashi Reiko’s figure of 80,168 ko (Hayashi 1969, p. 192), and the difference would be larger still if converted into tan at Hayashi’s rate of 120 tan to the ko. Hayashi 1969, p. 192. Hayami and Miyamoto 1988, pp. 253–55. 82 Nelson 1962, p. 229 (no. 817). McClain 1999, p. 63. The total import would be still higher by Hayashi Reiko’s conversion rate of 120 tan to the ko. Hayashi 1969, p. 193. Ōsaka-shi shi 1911, p. 651. Hayashi 1969, p. 195. The pattern of silk exports from Kyoto is taken as an illustration of how widespread was the distribution of textiles (pp. 195–7), although from its internal detail, that pattern has little direct relevance to redistribution from Edo. It simply shows that 37.7 per cent of the total distribution was consigned to Edo. However the absence in the table of a region to the north east of Edo would imply that part of the Edo total must have been reshipped to the Tōhoku. Sekiyama 1958 Ōsaka Dōjima beishō enkaku 1912. The Council decision to require publication of the records and statistics of the exchange (taken in 1903) is reported in an unnumbered page in the front matter of the volume. The history of the project is narrated very briefly on behalf of the Torihikijo, p. 1. Miyamoto 1972, p. 57. Miyamoto 1988, p. 135. Cullen 2003, p. 200.

NOTES

91

92 93

94 95

96 97 98

335

Suzuki 1938, p. 490, quoting Ōsaka-shi shi, vol. 4, p. 596. In the return constructed by Suzuki on the basis of two documents from the 1820s, the trade figures were both rounded and given within upper and lower bounds, in the manner of kura returns. Miyamoto 1972, p. 54, furnishes an example of a complete individual entry. OSSS 1964c, p. 154. Elsewhere in the same volume identical data are presented together with data for exports (pp. 101–102). Exports were much smaller and for Osaka were more positively a trade activity rather than a daimyo-conducted business. The combination of exports and imports in the table would appear to confirm that the import data for the years 1776–1780 with their precise numerical count to the apparent last koku were intended to be comprehensive. Suzuki 1938, p. 490, quoting Ōsaka-shi shi, vol. 4, p. 596. Uniquely for Tokugawa-era demographic data, Isshiki’s figures are presented in serial or tabular form. Cullen 2006, p. 156. BPP, vol. 4, p. 273. Ibid., p. 278. Ōsaka hennen shi 1978, pp. 341–45. The details are an extract from a record entitled kuriwata kaitsugi toiya kakiage hikae ⧞⥥㈙ḟ㛛ᒇ᭩ୖ᥍ that survived because of its inclusion in a wider compilation, kuriwata shoki ⧞⥥᭩グ. Minor textual omissions within the extract could either be the fruit of careless transcription, or reflect existing defects in the document from which it was extracted. The title of the extract seems to hint at tonya origin, though it is certified by the signature of a guild nengyōji. This is a rare instance of a document illustrating the interplay of tonya and guild.

CHAPTER 8 1

2 3 4 5 6 7

8 9

10

11

12

13 14

BPP, vol. 1, p. 194, Alcock, Edo, 6 March 1860. A paucity of interpreters was also noted as adding to problems (BPP, vol. 1, p. 165, Alcock). BPP, vol. 1, p. 165, Edo, 21 Nov. 1859. TNA, FO 262/230, f.103, Gower, Hyogo, 29 Feb. 1872. BPP, vol. 4, p. 29. BPP, vol. 4, p. 21. For an interesting summary of the situation, see Paske-Smith 1930, p. 203. BPP, vol. 4, pp. 19, 20. The Nagasaki exports for 1859 of 870,436 dollars were not cited by Ishii. (Ishii 1944). There is also an estimate for exports from Nagasaki for the first half of the year. As the port of Nagasaki was open only from July 1859, this figure is simply a return from the Shanghai Custom House for the external trade conducted by the Chinese from Japan for the period. BPP, vol. 4, pp. 29–31. See also Paske-Smith 1930, p. 205. BPP, vol. 1, p. 194, Alcock, 6 March, 1860. In a letter of 26 April, he described it more ambiguously as the trade of “less than 12 months” (BPP, vol. 4, p. 18). Alcock seems to have drawn this conclusion primarily on the basis of the evidence he had for Nagasaki. BPP, vol. 4, 26 April 1860, pp. 20–21. The total was 584,262 dollars, consisting of 461,386 dollars to Britain but including some coastal trade, and 122,876 dollars to foreign countries. These figures were not cited by Ishii in his classic study (Ishii 1944), no doubt because of their nature as external estimates not generated in Japan itself. BPP, vol. 4, pp. 23–29, Consul Vyse, Kanagawa, undated report for half year ending 31 Dec. 1860. For an amusing but patronizing and possibly apocryphal story of the entry at a Custom House, apparently in Nagasaki, of two tigers, see Alcock 1863, vol. 2, pp. 385–86. Ishii 1944. Paske-Smith 1930, p. 303. Montague Bentley Talbot Paske-Smith was His Britannic Majesty’s Vice-Consul in Osaka 1920–1925, Consul in Nagasaki 1926–1930, and Consul in Dairen

336

15 16 17 18

19 20

21 22

23 24 25 26 27 28

29

EARLY JAPANESE TRADE

1931. Some of his data differ from those in the BPP. While widely quoted, their sources have never been analyzed. See Appendix 1, The Paske-Smith figures for exports and imports in the open ports. Sugiyama 1988, p. 47. Baba and Tatemoto 1968, p. 164 Kawai 1895, p. 4. His figures were later used extensively by Itani Zen’ichi. BPP were available in later bound volumes and, in modern times, papers for Japan have been brought together in the Irish University Press series of 10 volumes. Parliamentary papers were first published and circulated as individual documents, and at the time it was to these that Japanese and others alike had access. Ishii 1944, p. 38. From an English language text entitled “Foreign trade of Japan: a statistical survey,” and separately paginated (Nihon bōeki seiran, 1935, p. 16). This table is missing from what is otherwise a broadly similar Japanese-language version preceding it in the same volume. Ishii 1944, exercising due discrimination, adopted the latter figure in his Table no. 5. Paske-Smith did not give figures for Hakodate, hence his totals, which are for Kanagawa and Nagasaki only, cannot be compared with totals for all three open ports. Paske-Smith 1930, p. 303. Ibid. There is a still higher figure of 3,701,084 dollars in Yokohama-shi shi 1959, p. 548. Ishii 1944, pp. 50–51. Yokohama-shi shi 1959, p.548. As quoted in Ishii 1944, p.38. BPP, vol. 1, pp. 285–87, Vyse, 19 Feb. 1861. Vyse converted the totals into sterling. These are the only figures he cited: £198,000 for imports and £824,000 dollars for exports. As printed, the report has a footnote reference to an “annexed table and returns” which had not yet been received in London. There is a slightly edited and undated copy of this report in BPP, vol. 4, pp. 31–32. Though contrary to treaty intent, business was conducted almost from the outset at Yokohama (which British officials at first claimed was a new Dejima ฟ ᓥ). The consular station itself, reflecting British opposition to the Japanese decision, continued for some considerable time to describe its reports as coming from the consular station at Kanagawa. The alarming magnitude of the consul’s “approximate estimate” for 1862 of 2.5 million dollars for imports and 3 million dollars for exports was simply caused by the consul in his confusion giving too low a value in taels to the dollar; his conversion inflated the dollar sum (Nagasaki consul, 18 February 1863, BPP, vol. 4, pp. 60–67, Winchester). The consul, not conversant with the going exchange rates (BPP, vol. 4, p. 63), quoted an improbable exchange rate of 1.80 taels to the dollar for the export trade on the authority of the Portuguese consul. Winchester took the dollar for imports to be equal to 1 ½ taels (p. 61); on what authority Winchester relied, we do not know. To add to the confusion, elsewhere in the report the British consul quoted 3.50 to 4.00 taels as the valuation of the dollar on the open market (p. 61). He also gave a rate of 5.50 taels to the dollar for what seems the fixed rate employed by the Custom House for converting gold invoices into dollars (p. 61). This was a rate both below the actual rate of 5.90 at the time and the conversion rate of 5.85 tael to the dollar in later years. A rate of 5.90 taels (or 5.85) would appear to be inapplicable to exports at large (as they were in silver prices), and also to imports from silver currency areas. For these areas of trade, Custom House valuations of the dollar were probably at, or close to, market valuations. Conversion of the gross total of trade in taels to dollars on the basis of 5.85 or 5.90 taels to the dollar (a rate much higher than an open market rate around 3.50 taels), would by definition result in an unduly deflated gross total in dollars. In addition, as the gross total consisted of imports in silver currency and of a gold currency component already deflated by a high official rate for the dollar, a gross total converted at a high rate of 5,90 taels would include a component in effect deflated twice over. The consular confusion over taels and dollars, which was not spotted

NOTES

30

31

32

33

34

35

36

37

38 39

40

337

in Kawai’s study, was recognized by Ishii, who used a rate of 3.75 taels to the dollar to calculate total exports of 800,000 dollars and total imports of 667,000 dollars. For Nagasaki, Alcock’s figures were exports of 1,000,317 dollars (£203,000 sterling) and imports of 669,261 dollars (£140,000 sterling). (Alcock, vol. 2, p. 387.) Alcock’s figures were higher than Paske-Smith’s, but the two versions are not grossly out of line. Alcock’s data for 1861 seem the more solid. Alcock’s figures for Yokohama appear in Alcock 1863, vol. 2, p. 384. His figures for 1861 were £558,948 sterling for exports and £307,981 sterling for imports. The figures were derived apparently from dollar totals converted at 4s.2d. to the dollar. His figures for 1860 (exports of £823,812 sterling and imports of £197,023 sterling) are identical to Vyse’s (see note 28), if allowance is made for some rounding by Vyse. The pagination of Alcock’s book, as given in Ishii, is from the American edition. His sources are identified in note 3, referring to p. 43 as Shokanjō sono hoka jūkomi ㅖ຺ ᐃ඼እ⤧㎸ The source is identified in footnote no. 4, referring to page 43, as Kakkoku shokandome man‘en gannen ྛᅜ᭩⩶␃୓ᘏඖᖺ Referring to p. 44 of the text, footnote 6 gives as its authority Isen sho kakitsuke ␗⯪ ㅖ᭩௜, and footnote 9 (p. 58) gives for authority Nezu Masashi ⚲ὠṇᚿ, “Bunkyū gannen Rokan Posadoniggu no Tsushima senkyo ni tsuite” ᩥஂඖᖺ㟢Ⰴ࣏ࢧࢻࢽࢵࢢࡢ ᑐ㤿༨ᣐ࡟ᑵ࠸࡚, Hō to keizai ἲ࡜⤒῭2:2. Footnote 18 (p. 58) referring to text on page 46 is based on “Burenwarudo hōkoku” ࣈࣞࣥ ࣡ࣝࢻሗ࿌, cited in a paper by Itani (Itani 1931). The author and work are identified from Ishii, p. 42, footnote 8. While Paske-Smith’s table has rounded figures for 1859 and 1860, it entirely lacks 1862 data for Nagasaki. It also has the curiosity of returns for Yokohama for 1866, which do not feature in official consular reports. Almost certainly these figures are an aggregation of monthly figures for the first ten months of 1866, although it is an open question whether the monthly returns existed in the Custom House at the time, or came into consular possession at a later date. Strikingly, Paske-Smith refers to a source other than the printed reports in the parliamentary papers, though for 1863 his evidence seems to suggest that he had access to the printed report on the trade of that year. On Paske-Smith’s sources, see Appendix 1: The Paske-Smith figures for exports and imports in the open ports. BPP, vol. 4, p. 142, 3 Jan. 1865, Gower, forwarding return for 1863. Gower seems to have had some difficulty with statistics at large, and his inadequacy became more apparent in his service in Osaka in the early 1870s. BPP, vol. 4, pp. 48–49, 31 Jan. 1863. Ishii’s text appears to give Alcock’s figures in taels for the year, with the dollar total coming from a conversion of the tael total into dollars. However, it is clear that the consular report was Ishii’s source, though he omits to mention the BPP. A retrospective table for 1860–1864 appears in the report for 1864, dated 24 March 1865 (BPP, vol. 4, p. 105). BPP, vol. 4, pp. 94–95, March 1864. The 1863 report does not seem to have been formally presented to London, and was not printed in BPP. The figures for 1863 can be gleaned only obliquely: in retrospect from figures in taels for that year which are given alongside the figures for 1864 in the 1864 report made early in the following year, and in dollars in tables comparing the trade of the three open ports. These numbers appear in Kanagawa reports on 21 April 1865 and 28 April 1866 (See appendix B). The report for Nagasaki for 1865, dated 15 Jan. 1866 (BPP, vol. 4, p. 186) referred to a double report for 1863 and 1864. This was not literally the case; rather there was an interval of fifteen days between the two reports (3 Jan. 1865 and 18 Jan. 1865). The temporal separation of the reports explains why the very late report for 1863 was in effect suppressed, and only the report for 1864 entered into the official domain. BPP, vol. 4, pp. 37–38, 41–44. This was picked up by Kawai Toshiyasu and thence entered by Ishii in his Table no. 5. No figure is given for imports for 1862, which could be either

338

41 42 43 44

45

46 47 48 49 50

51 52 53 54

55 56 57 58

59 60 61

62 63 64

EARLY JAPANESE TRADE

an omission or simply a case of the minute quantity of imports not being considered worth reporting. BPP, vol. 4, pp. 77–86. BPP, vol. 4, pp. 153–168, 1 Jan. 1865; BPP, vol. 4, pp. 174–84, 1 Jan. 1866. Cullen 2009, pp.192, 210. BPP, vol. 4, p. 494–500, Parkes, Edo, 31 March 1870. The report included not only totals of foreign trade, but slso the amount of trade in foreign goods between Japanese ports. The report observed (p. 494) that, as some of the foreign trade was intended for transshipment, an addition of the data on foreign trade and local trade would entail double counting. The following report for 1870 appeared under the date of 21 April 1871. BPP, vol. 4, pp. 637–42. Hakodate exports for 1863 are given variously as 269,050 dollars (BPP, vol. 4, p. 80); 266,134 (BPP, vol. 4, p. 167); 167,025 (BPP, vol. 4, pp. 205–206); and 148,712 (BPP, vol. 4, p. 121). There are Nagasaki exports and imports for 1864 of 1,159,892 dollars and 1,316,897 dollars in BPP, vol. 4, p. 188, and exports of 1,739,838 dollars and imports of 975,435 dollars for the same year in BPP, vol. 4, p. 121. Ishii opted for the Paske-Smith figure for imports. Yokohama-shi shi, 1959, p. 548. Foreign trade returns, 1866–1880 YKS, vol. 9, Jan. 1873. Yokohama-shi shi: Shiryō hen, 1962, Hanrei. Significantly or otherwise, perhaps because the compiler detected the confusion in the report on 1863 and sought to avoid its messy implications, a retrospective table of exports from Yokohama for 1860–1872 “for such of the preceding years as the consulate records furnish me with” gave no data for 1863. (BPP, vol. 5, p. 141, 31 March 1873.) Consular compilers of figures who scrupulously based retrospective figures on declared values alone also omitted 1863 from all subsequent runs of trade figures. Sugiyama 1988, p. 48. According to BPP, vol. 4, p. 145, the rate was 5.85 taels in 1863. Frost 1970, p. 31. See BPP, vol. 4, p. 186. Ishii converted the tael on the basis of 3.75 taels to the dollar, a rate quoted in the BPP as the average for 1863–1864, as against conversion in consular reports into dollars at a rate of 5.85 taels to the dollar. While this latter conversion rate is almost identical to that of gold invoices followed by the Custom Houses, it would appear to be inapplicable to exports or to imports for silver currency areas. For these areas of trade, Custom House valuations of the dollar were probably at, or close to, market valuations, and a dollar rating of 5.85 or 5.90 taels would have the result of further deflating totals in dollars, which were already deflated in Custom House figures based on gold currency invoices. Following a decision in 1865 to give prices in actual currency, totals for 1865 and 1866 were given in ichibus, the silver coin most widely available in Nagasaki, instead of taels at the rate that had prevailed in 1859–1860. BPP, vol. 4, pp. 60–67, Winchester, 18 Feb. 1863. BPP, vol. 4, p. 73, Porter, 7 Jan. 1864; pp. 175–76, Vyse, 1 Jan. 1866. BPP, vol. 5, p. 141, 31 March 1873. BPP, vol. 4, p. 240, Consul Myburgh, Kanagawa, 3 April 1867. Some accounts using figures of non consular origin give figures for 1866 (for an instance, see Yokohama-shi-shi, 1959, p. 548), which may actually be for the first ten months of the year (as remarked above). As I observed in Part 1 of this article (Japan Review 21), British reports identified the bugyō, or magistrate, as “governor.” BPP, vol. 4, p. 326, Consul Fletcher to Parkes, Kanagawa, 31 May 1868. BPP, vol. 4, p. 63, Winchester. See, for example, the case of adjustments to the price of tea in Nagasaki exports in 1860 (PaskeSmith, p. 204). BPP, vol. 6, pp. 676–77, Report for Hyōgo/Osaka for 1880. BPP, vol. 7, p. 101, 31 July 1882. Sugiyama 1988, p. 45.

NOTES

65

66

67 68 69 70 71 72 73 74 75

76 77 78 79 80

81 82

83

84

85

86

87

339

Osaka preceded Kobe ⚄ ᡞ as the early consular station. From the start, the shallow bar at the mouth of the Yodo ᾷriver made it impossible for larger Western vessels to enter, and Hyōgo, with deep water (though no shelter from south east winds), became the center of virtually all the trade. This prompted the government quickly to abandon Osaka in favor of Hyōgo as the harbor for foreign trade. The foreign settlement of Hyōgo grew, and by the end of 1870, as a result of the pace of house building, the two separate communities of Hyōgo and Kobe had effectively become one. Only at a much later date, in the wake of great engineering works conducted under the supervision of Dutch engineers, did Osaka itself become an ocean going port. Consular reports until 1885 were presented under the joint name of Hyōgo/ Osaka. The advent from January 1873 of statistics very clearly presented in modern printing and bilingual format may be a factor in the emergence of a more positive acceptance of the statistics. Older reports in wood block printing and in somewhat confused presentation are far from easy to follow. BPP, vol. 5, p. 426, Parkes, 30 Aug. 1874. Umemura and Yamamoto 1989, p. 193. Yamaguchi and Ōuchi, 1968, pp. i, ii. BPP, vol. 4, p. 390, 10 March 1870. BPP, vol. 5, p. 77, Consul Robertson, 1 May 1872. BPP, vol. 5, p. 141, Consul Robertson, 31 March 1873. BPP, vol. 5, p. 174. Nagasaki, report for 1872, Consul Flowers, 26 Feb. 1873. Foreign trade returns, 1866–1880 YKS, vol. 9, trade of year 1873. Foreign trade returns, 1866–1880 YKS, vol. 11, monthly returns from Jan. 1874 to May 1876, especially for first months, for which there is more detail than for later months. BPP, vol. 5, p. 344. 30 June 1874. BPP, vol. 5, 174, Consul Flowers, 26 Feb. 1873. Meiji zaisei shi ᫂἞㈈ᨻྐ1904, pp. 194–98. Ibid., pp. 246–323. An estimate of purchases of vessels in 1862, 693,000 dollars, is contained in BPP, vol. 4, p. 48, consul Vyse, Kanagawa, 31 Jan. 1863. Incidental reference is common in other years (e.g., BPP, vol. 4, pp. 61, 87, 89, 100, 141, 186–87). On the Japanese side, details of purchases are set out by Katsu Kaishū (1967, pp. 442–54). These later figures have been added to Custom House data in modern accounts to give a more comprehensive return of imports. See BPP, vol. 7, p. 109, table of imports 1867–1881. Paske-Smith’s reference to paying Custom House clerks a monthly fee to supply returns is an oblique reference to this practice (Paske-Smith 1930, p. 203). As well as annual figures for years ended December, totals may also have been compiled for years terminating in June, given the existence of half-year figures. At any rate, returns exist for years ended June 1876 and June 1877, and some round totals suggest that such returns were also compiled for other years. With a gap for the second half of 1876, monthly returns for 1873–1880 are in Foreign trade returns 1866–1880, vols. 9–14, YKS. The monthly figures from 1873 are tabulated in Nihon bōeki seiran, 1935, pp. 666–670. The monthly breakdown from the customs department of the Ministry of Finance appeared from 1895 in Tōyō keizai shinpō ᮾὒ⤒῭᪂ሗ. See the preface to Nihon bōeki seiran, 1935, pp. 1–2, and also “Japan’s foreign trade, past and future,” p. 41ff. in the same volume. The latter section of 43 pages is preceded by a substantially similar but not identical 48 pages in Japanese under the title “Waga kuni no bōeki” ᡃࡀᅜࡢ㈠᫆. Both sections have a separate pagination from the bilingual main text, entitled “Naichi oyobi Karafuto gaikoku bōeki no bu” ෆᆅཬᶟኴእᅜ㈠᫆ࡢ㒊, which contains the corpus of statistical data. The statement is repeated in many of the later months of 1873. Foreign trade returns 1866– 1880, vols.9–10, YKS. BPP, vol. 5, p. 439, Parkes, 26 Sept. 1874.

340

88

89 90

91 92 93

94

95

96

97

98

99 100 101

102 103

104 105 106 107 108 109

EARLY JAPANESE TRADE

Modern copies of the series have been made, and are widely available in libraries throughout Japan. Hitotsubashi University ୍ᶫ኱Ꮫ, despite its origins as a higher school of commerce of Meiji times, does not have originals for the early years: it has them for later years. The years 1882–4 are reprinted in Dai Nihon gaikoku bōeki nenpyō: Meiji 15, 16, 17. They are preceded by a useful introduction by Yamaguchi Tetsuo entitled “Dai Nihon gaikoku bōeki nenpyō ni tsuite” ኱᪥ᮏእᅜ㈠᫆ᖺ⾲࡟ࡘ࠸࡚. BPP, vol. 7, p. 194, 15 July 1883. Useful contemporary retrospective reports are Dai Nihon gaikoku bōeki taishō hyō ኱᪥ᮏ እᅜ㈠᫆ᑐ↷⾲(1893), and Nihon gaikoku bōeki yonjūrokunen Taishō hyō: Meiji gannen yori Taishō ninen ni itaru᪥ᮏእᅜ㈠᫆ᅄ༑භᖺᑐ↷⾲㸸᫂἞ඖᖺᐤ⮳኱ṇ஧ᖺ. On the latter publication, see Honsho shotōkei shiyōjō no chūi ᮏ᭩ㅖ⤫ィ౑⏝ୖࡢὀព, p. 45 (Nihon bōeki seiran 1935). Yamaguchi and Ōuchi 1968. See the tables in Yamaguchi and Ōuchi 1968. Foreign trade returns 1866–1880, vol. 1 YKS, two documents, copies of returns in KBS (see bibliography under Foreign trade returns 1866–72, 1866–73): Kakkō yushutsu buppin kinryō daika hyō ྛ ㍺ฟ≀ရ᩹㔞௦౯⾲with figures for 1866–1872 and Kakkō yushutsu buppin hyō ྛ ㍺ฟ≀ရ⾲, with figures for 1866–1873. The latter document indicates clearly that it was compiled by the Ministry of Finance’s Statistics Bureau (Tōkei ryō ⤫ィᑅ) in 1874. Yokohama-shi shi: Shiryō hen 1962. The most convenient location for these figures is in foreign trade returns 1866–1880, vols.1–14 YKS, which contain summary export figures to 1873, and full returns from 1870 to 1880. These are photocopies, and as the pages are sometimes faded, the source from which they are copied is not always easy to identify. Foreign trade returns 1866–1877, NDL microfilm. Summary figures from 1866 and full returns from 1870. Kanji and rōmaji titles are similar to those in bibliography under Foreign trade returns 1866–1880 YKS. Matsukata Masayoshi monjo mokuroku ᯇ᪉ṇ⩏ᩥ᭩┠㘓, 1954 NDL, Matsuo ke monjo mokuroku ᯇᑿᐙᩥ᭩┠㘓NDL, Shōda ke monjo mokuroku ຾⏣ᐙᩥ᭩┠㘓. The Shōda collection for a much later period is not relevant in the present context. “Kōshaku Matsukata Masayoshi kyō kanki” ೃ∖ᯇ᪉ᨻ⩏ཀᐶグin Matsukata Masayoshi kankei monjo ᯇ᪉ṇ⩏㤋㛵ಀᩥ᭩. There is, however, in vol. 3, pp.1–2, a retrospective table of foreign trade for 1868–1885. The microfilms in the National Diet Library contain a few papers relating to Matsukata’s work in customs reform in the 1870s. Inoue Kaoru kankei monjo ஭ୖ㤾㛵ಀᩥ᭩ NDL, , file 691, no. 5, 1870 (Meiji 3). The first six months of the year are in a woodblock printing, styled “Kaku kaikōjō yushutsu buppin daka” ྛ㛤 ሙ㍺ฟ≀ရ㧗. There is a full calendar of the Inoue deposit in Inoue Kaoru kankei monjo mokuroku 1975 NDL. Yokohama-shi-shi: Shiryō hen in 1962 and Yamaguchi and Ōuchi in 1968 Yokohama-shi-shi: Shiryō hen in 1962, introduction (unpaginated, fourth page). Nihon bōeki seiran 1935, p. 18. The table of exports of coin and bullion in the same volume is accompanied by a note stating that “figures of coin and bullion exported or imported up to 1871 are unavailable.” This seems based on the judgment that, as separate returns for coin and bullion are available from 1872, coin and bullion were included in earlier gross trade figures. BPP, vol. 4, pp. 637–42. Outflow of coin and bullion in 1869 was ten million dollars. Report for Kanagawa, 31 March 1870, BPP, vol. 4, pp. 495–96. BPP, vol. 4, p. 87, March 1864. BPP, vol. 5, p. 23. Osaka/Hyōgo, 29 Feb. 1872. BPP, vol. 5, p. 82, report for Kanagawa, 1 May 1872. BPP, vol. 5, pp. 429, 432, 30 Aug. 1874. Yamaguchi 1990, p. 32. See Sugiyama 1988, p. 48; Nishikawa, Odaka, and Saitō 1996, p. 175.

NOTES

110

111

112 113

114

115 116

341

In Nagasaki, the total of foreign residents for various dates between 1862 and 1869 came to a cumulative figure of 2,290. Of these, 1,391 were Chinese. (Nagasaki ken shi 1985, p. 848.) Consul Winchester in the early 1860s referred to a total of 1,800 Chinese. On the Chinese in Nagasaki, see also Report on the foreign trade of Nagasaki for 1870, BPP, vol. 4, pp. 586–87. For Kanagawa, Chinese were more than half the total (excluding Americans) in 1874 and 1875. (Reports on the foreign trade of Kanagawa for 1874 and 1875, BPP, vol. 5, pp. 544, 644.) In Hyōgo in 1877, Chinese were slightly less than half the total. (Report on the foreign trade of Hyōgo for 1877, BPP, vol. 6, p. 297.) For counts of Westerners and Chinese in open ports and in Tokyo in 1876, 1880, and 1885, see Report on the foreign trade of Japan for 1885, BPP, vol. 8, p. 55. For currency debasements and depreciation in 1869 in terms of the dollar, see BPP, vol. 4, pp. 442–43, Report from Nagasaki, 31 Jan. 1870. Observations in other years are equally helpful. BPP, vol. 5, p. 37, Nagasaki report for 1871, 31 Jan. 1872. Sugiyama is confused in stating as follows: “During the period 1871–1887, imports from gold standard countries were converted in terms of Japanese gold yen, while those from silver-standard countries were calculated in terms of Japanese silver yen. Despite the changing difference in value between gold and silver, however, a simple total was made from these figures. It is therefore necessary to convert the figures so that they stand wholly on the basis of silver yen” (Sugiyama 1988, p. 47). Sugiyama seems to suggest here that aggregates of silver yen and gold yen were crudely added together to arrive at the grand total for imports. As the grand totals in Japanese statistical returns were already in silver dollars (including the conversion at fixed rates of sums on gold currency invoices), it would have been impossible for the process to occur as Sugiyama implies. Less than half of imports were expressed originally in silver, and the balance in the values of gold standard countries. The problem of converting gold into silver was in no way an inherent problem of bimetallic currency. The issues that arose were twofold. The first, in the short term a very real problem, was that while the official parity between gold and silver yen was fixed, silver traded freely in the open ports; recognizing that silver could trade at a discount or premium, Japanese officials converted the Custom House returns, already expressed in dollars, into gold yen at a variable rate. The other problem, a mere bookkeeping one and in the short term minor, was that the gross returns from the Custom Houses were made up for invoices in silver currency of unadjusted figures and, for invoices in gold, of sums converted into silver dollars at a rate fixed in 1871, rather than at the current market exchange rate. If clear recognition of the two conversion processes is made, confusion should not arise. In the process of calculating totals, gold invoice figures were converted into dollar figures at the time of their presentation at the Custom House at the fixed 1871 rate. At a much later date (and in some cases retrospectively for earlier years), officials in Tokyo converted the Custom House dollar figures into gold yen, taking account of variation in the actual exchange rates. Sugiyama 1988, p. 48. Sugiyama regards the problem of comparing dollar and yen totals as arising from the problems of aggregating totals of trade expressed in silver and gold respectively. However, the problem for the Ministry of Finance arose not from that challenge but from the much more difficult task of retrospective calculation for the earlier years, converting the growing medley of mainly depreciated coins. Sugiyama postulates that the exchange rate for dollars for exports in 1869 was 1.12 yen, and for imports, 1.20 yen (Sugiyama 1988, pp. 239–40, n. 44). These ratios do not reflect actual exchange rates, but they are a mathematical consequence of prior conversion. His statement of an annual average exchange rate of 1.24 yen per dollar in 1869 simply adds to the complexity of arguments based on such data. In the case of the bill rate on London, the dollar traded fairly consistently in 1869–1871 at 4s.6d. BPP, vol. 4, p. 260. See “Memorandum on Japanese currency,” Oct. 1893 in BPP, vol. 9, p. 92. This is a very clear report on the Japanese currency over preceding decades.

342

117

118

119 120

121

122

123

124 125 126

127 128 129

130

131 132 133 134 135 136 137 138 139 140 141

EARLY JAPANESE TRADE

Exports of bullion were exceptionally large in 1874. (BPP, vol. 5, pp. 573–74, report for 1874, 31 Aug. 1875.) Divergences between dollars and gold yen in both exports and imports in 1877 and 1878 were larger than those for 1876 or, for that matter, 1873. Although from 1879 yen and dollar totals for both exports and imports were very close to each other, the sole case of yen and collar figures being identical was for imports in 1882. Import figures, however, varied slightly from the emerging close general correspondence in 1884. Introduction (unpaginated). The figures are characterized here as “almost the same” (hobo onaji ࡯ࡰྠࡌ). Howe 1999, p. 141. From 1878 this, which never proved as popular as the Mexican dollar in the open ports (the problem of securing acceptability for it had been anticipated in its slightly heavier weight at 420 grains than the Mexican dollar), was no longer coined. See “Memorandum on Japanese currency,” Oct. 1893, in BPP, vol. 9, p. 92. Probably the clearest account of the Japanese currency is contained in Matsukata 1899, pp. ii–vii. Except for May 1882, when gold yen were quoted at 95 for 100 dollars, “the average quotation during the rest of the year having been about 92 1/2” is a fair summary of the course of exchange. (Report on Hyōgo and Osaka for 1882, BPP, vol. 7, p. 134, 7 May 1883). Average rates per pound sterling were: 3s.8d. in 1883; 3s.8d. in 1884; 3s.6d. in 1885; 3s.4d. in 1886; 3s.2d. in 1887. A table from 1874 of sterling per yen and dollar per 100 yen is in Yamazawa and Yamamoto 1979, p. 256. Whether these figures were based on a dollar rate or a yen rate is unclear. However, as divergence would have been small between the dollar and the silver yen for most of these years, the exchange would be broadly similar in both cases. BPP, vol. 8, p. 489, 10 June 1890. Yokohama-shi-shi 1963, p. 196. Statistical appendix by Masahiro Tatemoto, “The correction of official import statistics, 1874– 1887, for the depreciation of silver” in Baba and Tatemoto 1968, pp. 183–84. Yokohama-shi-shi 1963, p. 196. Baba and Tatemoto 1968, p. 164 and Sugiyama 1988, pp.46–8. Sugiyama’s data are based on Paske-Smith’s estimates and on a rejection of Ishii’s work. As Sugiyama was primarily concerned with estimating the balance of payments, he included additions for the purchase of steamships. In regard to earlier figures, undervaluation was not a problem in the figures for 1866–1868. Fixed conversion rates for gold into silver had been abandoned in 1865 in favor of market rates (though they reappeared in 1871). Among the factors that affected the statistics we have for 1869 are these: exchange rates, later retrospective conversions by Ministry of Finance officials, capital movements (which may or may not have been included by the officials who compiled the statistics), and the possibility that inter-port figures (for trade between treaty ports) were incorporated into figures supposedly for foreign trade only. BPP, vol. 5, p, 407, Consul Flowers, 28 March 1874. BPP, vol. 6, p. 446, 15 June 1879. BPP, vol. 6, p. 705, Kanagawa, 20 June 1881. BPP, vol. 4, pp. 356–71. BPP, vol. 4, p. 411, 25 Jan. 1870; BPP, vol. 4, p. 551, 28 Jan. 1871. BPP, vol. 5, p. 12, Consul Gower, 29 Feb. 1872. BPP, vol. 5, p. 120, 4 April 1873. BPP, vol. 5, p, 344, 30 June 1874. BPP, vol. 5, p. 545, 16 June 1875; BPP, vol. 5, p. 595, 12 June 1876. BPP, vol. 5, p. 120, Gower, 4 April 1873. BPP, vol. 5, pp. 12, 24, 29 Feb. 1872; BPP, vol. 5, p. 120, 4 April 1873. Gower’s poor grasp is very evident in his reports. Though aware of the concept of “disposals” (a fact confirmed in his statement that in his compilations he followed instructions from Tokyo), his failure ever to use the term suggests that he was poorly equipped to make comparisons of Custom House and Chamber of Commerce counts. The treatment of coastal trade to and from other

NOTES

142 143 144 145

146

147 148 149 150 151 152 153 154 155

156

157 158

159 160 161 162 163

164 165 166

343

treaty ports also raises difficulties. For 1871 and 1872, Gower added figures for the coastal trade to and from treaty ports to counts from Custom House data of direct foreign trade, but it is not clear how coastal trade was treated in Chamber of Commerce data. (The question of transshipment, discussed above, is relevant in this context.) For the foreign trade of Osaka, the questions of disposals for imports, differences in pricing of commodities, and the treatment of coastal trade make any judgmental comparisons of the merits of Custom House figures highly problematic. BPP, vol. 5, p. 545, 16 June 1875; BPP, vol. 5, p. 595, 12 June 1876. BPP, vol. 6, p. 85, Hyōgo, 28 June 1877. BPP, vol. 5, p. 141, 31 March 1873. BPP, vol. 3, p. 118, “Report on the production of tea in Japan,” Edo, 11 Oct. 1872, contained in “Report by acting vice-consul Wilkinson on trade of Hyōgo and Osaka.” BPP, vol. 4, p. 497, report on the trade of Japan for 1869, 31 March 1870. See also BPP, vol. 5, pp. 427–28, report on the trade of Japan for 1873, 30 Aug. 1874. In the return for 1869s, there are errors within the totals in the final column of the table, and the grand total should read 24.3 million, not 24.1 million. BPP, vol. 5, p. 572, report on the trade of Japan. Edo, 31 Aug. 1875. BPP, vol. 5, p. 438, Parkes, 26 Sept. 1874. BPP, vol. 6, p. 580, report on Hyōgo and Osaka, 31 March 1880. BPP, vol. 4, p. 156; vol. 7, p. 11. BPP, vol. 6, p. 676, Hyōgo, May 1881. BPP, vol. 5. p. 224. Report on the present educational system of Japan, 30 Nov. 1873. See BPP, vol. 6, p. 14, Report by Mr. Mounsey on the finances of Japan, March 1877. BPP, vol. 8, pp. 221–2, 28 Jan. 1888. Imlah 1958. Imlah showed that if the value of commodities was recalculated from the unchanging and increasingly unrealistic valuations applied to exports by earlier statistics, Britain was running not a surplus as was thought, but a large deficit in its balance of trade. This circumstance underlay the huge and unsuspected scale of the balance of payments surplus from “invisibles”, which was beginning to emerge, and to finance a growing outflow on capital. Ibid., p. 44. Declared values existed for exports from 1798. However, they were often overlooked in favor of the unchanging fixed official price totals which continued to be compiled in contemporary analysis of trade. Trade returns in official prices continued to be printed up to 1869. Ibid., p. 44. The earlier omissions in Japanese statistics were exaggerated; figures for official purchases were included. However, they were underestimates and did not include steamships. Cullen 2006, pp. 155–56. BPP, vol. 4, p. 274, report on Osaka/Hyōgo. See notes 29 and 54. Paske-Smith 1930, p. 204. BPP, vol. 4, pp. 205–6. The gross returns for exports and imports can be made up by summing the separate tables for trade on British and foreign vessels respectively. Ishii, pp. 46–48. A still higher figure of 3,701,084 is given in Yokohama-shi shi 1959, p. 548. The omission of figures for this year from later retrospective tables by consular and legation offices is an implicit recognition of the problems created by the 1863 report See also footnote 50.

CHAPTER 9 1

I am indebted to Professors Satō Osamu, Katsuta Shunsuke and Kuwajima Hideki for help on many occasions; to Professor Hoya Tōru of the Shiryō Hensanjo for photocopies of much of the two picture scrolls; over many years to Dr. Honma Sadao of the Nagasaki Kenritsu

344

2 3 4 5 6 7 8

9 10 11 12 13 14

15 16 17 18 19

20 21 22

23 24 25 26 27 28 29

30 31 32 33

34 35

36

EARLY JAPANESE TRADE

Toshokan; and to Professor Kasaya Kazuhiko of Nichibunken for advice. Two anonymous referees made important suggestions, and I was the beneficiary of many patient editorial criticisms from John Breen. I am indebted also to the late Kato Eileen in Tokyo, and to Tsukahara Suekko in Nagasaki for information and books, and for enquiries made on my behalf in archives in Kyushu. Enomoto 2004, p. 30 Toby 1991, p. 235. On the role of the Zen monks, see Nakao 1997 and Tanaka 1996. Aston 1874, p. 20. Aston 1879, pp. 107–120. Toby 1991, p. 10. Shiryō Hensanjo (iii). 4171 08. Shiryō Hensanjo (i). 2051.9/77. It contains a copy of Sadanobu’s Ezochi onsonae ikken ⼎ዀ ᆅᚚഛ୍௳, said by Fujita to be of first rate importance (Fujita 2005, pp. 188, 309. See also p. 167, and note 51). Fujita 2005, pp. 240–41. Fujita 2005, pp. 309–310. Shiryō Hensanjo (i). 2051.9/77. Fujita 2005, p. 25, note 5, and p. 51. Fujita 2005, pp. 32, 188. The hommaru of the Castle , formerly the residence of the shoguns, served for the emperor also until the disastrous fire of 1873. See Kizaki 2005, p. 51. Hankachō (i). There are also three sets of supporting documents, though these are not complete, in Hankachō (ii), (iii) and (iv). Osaka-shi shi. OSSS 1963–1966. For the history of this project, OSSS, “Supplement 1966” is indispensable. See also Cullen 2009, pp. 190–91 and note 25. The 400 notebooks of the editor of the Ōsakashi shi, Kōda Shigeru, are in the Ōsaka Hensanjo. See Cullen 2009, p. 197. Yasutaka 2010, p. 152. Honma 2000, p. 39. E.g. NRBH (iii) Igirisusen torai ikken (4 November 1808) has several stamps, one that of a shōya ᗉᒇ. Harada 2007, pp. 269–71. Honma 2000, pp. 40–43. On Kanai, see also Harada 2007, p. 270. Ōhori 2007, Introduction. Unpaginated. Nagasaki chōsa hōkoku 1997. Tō tsūji nichiroku, vol. 1, pp. 1–7; vol. 7, pp. 109–114. Roshia shisetsu Rezanofu raikō 㟢すள౑⠇ࣞࢨࣀࣇ᮶⯟. Shiryō Hensanjo (ii). .See MacDonald’s own later Narrative, in Schodt 2003, pp. 257–62, and Schodt 2003, p. 395, note 17. “Maegaki,” Roshia torai roku 1994 (unpaginated). An official account transmitted through the Kuramachi family of domain elders. Ikoku nikki; and especially Nakamura et al. 1989. See shahon in NRBH B. Rezanov Embassy 1804, and KBS C. Rezanov Embassy 1804, and also a very striking one in EDUN (iii) Rezanov Embassy 1804. This latter shahon, 407 M49, while mentioned on page 25 of Nagasaki chōsa hōkoku 1997, is missing from items listed in the same volume on pp. 187–88. Beerens 2000, p. 388. Beerens 2000, p. 381, also note 44. The higher graded among them seem also to have drafted documents (p. 382). Ōta Nanpo zenshū, vol. 19, p. 685. See also Kutsukake 2007, p. 224. Kutsukake can be supplemented by Hamada 1986, pp. 195–97.

NOTES

37

38

39

40 41

42 43 44

45 46 47 48 49 50 51 52 53 54

55

56 57

58 59 60

345

Ōta Nanpo zenshū, vol. 19; NDL C. Enkai ibun mss (i), (ii), (iii). The tone of exchanges between Rezanov and Japanese is well conveyed in his diary. See Ōshima 2000. NDL C. Enkai ibun (iii). Ōta Nanpo zenshū, vol. 19, pp. 6–7, 9. For full details of the copies, see Ōta Nanpo zenshū, vol. 19, p. 666. As further proof of Ōta’s interest in the Russians, kan 2 of the six kan Kaibō kiji ᾏ㜵グ஦, copies of which are in University of Tsukuba ⟃Ἴ኱Ꮫ and Hakodate City Central Library ภ㤋ᕷ୰ኸᅗ᭩㤋, depict Russian ships and dress for 1807. Kokon shūran a, b. In the introductory remarks of the two volumes, there are contradictions in the dating of Matsuura’s career. Egoshi and Urakawa 2009. See SMH (ii) 14–2-93. Though the catalogue gives the date of 1830, the entries are in several hands, suggesting they were made at various earlier dates. Another document, SMH (iii) 14–21, in both Japanese and Dutch, seems to represent the process of translating an individual fūsetsugaki from Dutch into Japanese. See also NRBH C. Fūsetsugaki (i) and (ii). No. (i) is a run of transcripts from 1827 to 1856 made by a single copyist. Matsukata (2007, pp. 300–303) lists details provided by the Dutch for the compilation by tsūji of fūsetsugaki and the original texts drawn up by the Dutch, of the betsudan fūsetsugaki ูẁ㢼ㄝ᭩for 1834–1859. TKIR, vol. 6, kan 247, pp. 264–88. TKIR, vol. 6, pp. 286–302 (kan 248). Itazawa 1937; Iwao 1979. See especially “The leakage and transcription of Oranda fūsetsugaki and Public opinion,” in Iwao 1979, vol. 1, pp. 21–24. Ishii 1998, pp. 6–7. Kizaki 2005, pp. 173–76. Tanaka 1998, pp. 145–71, Cullen 2009, p. 206. Cullen 2009, pp. 197, 208. Cullen 2006, pp. 159–60, 162–63; Hayami 2008. Sekiyama 1957. See also Cullen 2006, p. 161. BPP, vol. 4, p. 273. See also Cullen 2010, p. 209. Nagasaki chōsa hōkoku 1997, p. 14. At some stage the papers were separated, and till recently the hope lingered that more might turn up. However, the fact that surviving documents in Dutch are almost all confined to the last two decades of the eighteenth century suggests the loss or separation began early. The papers now in Nagasaki were rescued in 1912 by a now unknown party following the intervention of a student or an apprentice of Motoki Shōzō ᮏᮌᫀ㐀. They then passed to an alumni group, were acquired by the city, and ultimately by the city museum (See a brief note in the printed catalogue of the former City Museum (“Shuroku bunsho shiryō narabi ni kakushu bunko (bunsho) no shōkai” ྲྀ㘓ᩥ᭩ྐᩱ୪ࡧ࡟ྛ✀ᩥᗜ>ᩥ᭩@ࡢ⤂௓). The Kobe papers contain sixty seven catalogue entries of Japanese texts (from 1751 to 1856), and 131entries of Dutch texts from the 1780s and 1790s (see Kōbe mokuroku 1997). Only eight of the seventeen sections in the Nagasaki papers are listed in Nagasaki chōsa hōkoku 1997, pp. 38–40, 245–62. They were at some stage crudely stitched together by someone who did not comprehend their content. Some further Motoki items entered the City Museum in Nagasaki (nos. 840–1, 840–2, 840–5, now in NRBH) through other channels. SMH (i). The papers were presented to the Siebold Memorial Hall by a family descendant living in Kyoto in 1988 (Nagasaki chōsa hōkoku 1997, p. 27). They contain about 1000 items and are particularly significant since in part they relate to the Siebold Incident. The nikki for one year has been published as Ansei ninen 2001. Fukui 1980, p. 139. Observations made by Ōta Nanpo in 1800 on the archives are noted on the same page. McEwan 1962, pp. 94–95. Beerens 2002, p. 175. Cullen 2003, p. 58.

346

61 62 64

65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

86 87 88 89

90 91 92 93 94 95

96

97 98

99 100 101

EARLY JAPANESE TRADE

The history of archival institutions is summarized in Naikaku mokuroku (iii), and in a chart in Naikaku mokuroku (ii), p. 7. Naikaku mokuroku (iii), p. 4. 63 Fukui 1980, p, 137. Fukui 1980, p. 137. See Kondō Jūzō kankei bunken kaidai ㏆⸨㔜ⶶ㛵ಀᩥ⊩ゎ㢟in Fukui 1980, pp. 311–36. The published texts of his papers are in Kondō Jūzō (i), (ii). Kizaki 2005, p. 106; Fukui 1980, p. 22. It had suffered from fires in 1657 and in 1772. Naikaku mokuroku (iii), p. 5. Kizaki 2005, p. 107. For a breakdown of records, see Kizaki 2005, pp. 115–16. 68 Kizaki 2005, p. 122. Naikaku mokuroku (iii), p. 4. On sources relating to Takahashi Kageyasu 㧗ᶫᬒಖ, Bansho wage goyō and the Tenmonkata ኳᩥ᪉, see Fukui 1983. Shizuoka Kenritsu Chūō Toshokan 1970, 1996. Kasaya 1998, pp. 40–145. Kasaya provides a full analysis of other formal documents. Hayami 2001, pp. 79–80. Hall 1968, p. 163. Hall 1968. Kasaya 1998, pp. 23–38. Hall 1968, p. 156. Hall 1968, p. 158. Hall 1968, p. 159. Hall 1968, p. 166. Hall 1968, p. 157; Kasaya 1998, p. 29. Toby 1991, pp. 261–62. Toby 1991, p. xxxi. Tanaka 1998, pp. 384–86. Kasaya 1998, pp. 27–29, 32. Kasaya 1998, p. 26. Kasaya 1998, p. 26. It was highly unusual for a tozama daimyo to serve as rōjū. Sanada ke monjo mokuroku 1979, pp. 373–79. It contains more than 30,000 items and 1,800 satsu of nikki from the 1730s onwards. DNIS, 2nd series 2, vol. 4, pp. 298–300, and DNIS, 3rd series, vol. 6, pp. 479–81. Nariaki was a very active correspondent in 1858. The titles of the two Hotta shahon are Hotta Masatomo kaki ᇼ⏣ṇㄽᐙグ and Hotta Masayoshi gaikoku kakari chū shorui ᇼ⏣ṇ╬እᅜ᥃୰᭩㢮. Imaizumi 2011, pp. 148–49. Imaizumi 2011, p. 149. Shimazu ke monjo mokuroku (i), (ii). DNIS, 3rd series, vol. 7, pp. 598, 715, 785, 798.2. Mito ke kankei shorui Ỉᡞᐙ㛵ಀ᭩㢮. See Kasaya 1998, p. 27. Some of the individual letters to and from Matsudaira Yoshinaga ᯇᖹ៞Ọin DNIS were gathered from other sources, illustrating how han records from several locations helped complete the surviving source base. See DNIS, 2nd series, vol. 4, pp. 38–82 and 498–99, 589–90. Toby 1991, pp. 10–14, note; pp. 241–42; Fujita 2005, pp. 7–10. This interpretation originated in a rather forced argument by Iwao Seiichi in 1963. See Cullen 2003, p. 49. TKIR, vol. 6 (kan no. 253). Murakami 1899, vol. 2; Machin 1978. The Murakami volume also has the text of Delboe’s diary of the visit of the Return. EDUN (i). See also Nagasaki chōsa hōtoku 1997, p. 187. Cullen 2003, p. 49. TKIR, vol. 6, kan nos. 256 to 259, plus kan no. 260 dealing with the new bugyō, Magaribuchi Kai no kami Kagetsugu ᭤ῡ⏥ᩫᏲᬒ₞, who arrrived in the ninth month.

NOTES

102

103

104 105 106 107 108

109 110

111

112 113 114 115 116 117

118

119 120 121 122 123 124 125 126

127 128 129

130 131 132

347

TKIR, vol. 6, p. 409 (kan no. 256). See also p. 421 (kan no. 257), and pp. 438, 443, 445 (kan no. 259). Ryakusho ␎᭩, a summary account by Takaki Sakuzaemon, is referred to explicitly in kan no. 259 (vol. 6, p. 445). NRBH A. Phaeton Incident (i), (ii); Katagiri 1972, p. 11. The integral text of NRBH A. Phaeton Incident (ii) appears in Katagiri 1972, pp. 73–182. The introduction (pp. 10–17) should also be consulted. The name of the copyist recurs in one other document dealing with the episode (NRBH A. Phaeton Incident [vi]). EDUN (ii). Katagiri 1972, pp. 10–11. Katagiri 1972, pp. 71–182. Egoshi and Urakawa 2009, pp. 210–31. TKIR, vol. 6 (kan no. 259), p. 436. There appears to be an error with ྑappearing in place of ᕥin the surname. Katagiri 1972, p. 12; NRBH A. Phaeton Incident 1808 (ii). See Tsukuba University site for Nagasaki shi zokuhen, http://www.tulips.tsukuba.ac.jp/ limedio/dlam/ B1132580/1/mokuji/3709.pdf (accessed 17 March 2013). There are five versions in KBS A. Nagasaki shi zokuhen, (i) to (v); and a complete version also in NDL B. Nagasaki shi zokuhen. Kankai ibun included maps, sketches of crew members, the two leaders, dress and the warship surrounded by Japanese craft. KBS B. Kankai ibun (i) to (v)㸹NDL A. Kankai ibun mss (i), (ii), (iii). Much the superior one in artistic quality is KBS B. (i) 185–107. Sugimoto 1986 is the best modern edition. The text was edited by Shimura Hiroyuki ᚿᮧᘯᙉ, with an introduction and afterword by Sugimoto Tsutomu ᮡᮏࡘ࡜ࡴ. TKIR, vol. 7 contains kan nos. 297–306, and vol. 8, kan nos. 307–315. TKIR, vol. 7, kan no. 297, p. 396. TKIR, vol. 8, kan no. 315, p. 118. Fujita 2005, pp. 140–42, 155. Fukui 1980, pp. 315, 326–28. Fukui has some more general comment on p. 144. TKIR zokushū, vol. I. First dispatched as a metsuke to Nagasaki in the 1820s to prevent smuggling, in 1836 he was promoted to bugyō rank. Togawa’s copy is not included in the papers lodged in the Kunaichō by the Togawa family in the 1880s (see Kunaichō Shoryōbu [i] to [iv]). The deposit consists of office diaries and three satsu of largely topographical information on Kyushu. On Togawa’s career in Kyushu, see also Cullen 2003, p. 48 and note 77. Recognition of his service with promotion to the rank of kanjō ginmiyaku ຺ᐃྫྷ࿡ᙺ, came only in his fortieth year in 1810. Egawa’s later role in superintending defenses of Edo Bay is another instance of the pattern. Nakada 1998, and Nakada 1985. Sasama 1965 and Yamamura 1974 both provide useful background information. Yamamura 1974, pp. 20–22; Kasaya 2000, p. 132. The number of metsuke may at times have been higher. See Beerens 2000, p. 381, note 43. Beerens 2002, p. 389. Toby 1991, p. 165. See above p. 238. Beerens 2002, p. 190. h t t p : / / w p 1 . f u c h u . j p / ~ s e i - d o u / re k i s i - s i r y o u / 0 0 2 4 6 t o k u g a w a - n a r i a k i - s h i n ise/00246tokugawa-nariaki-shinise.htm. Accessed, 22 October 2011. Seishikan Kinenkan 1993. Naikaku mokuroku, vol. 1, p. 588. Ansei nenkan bōfu Bitchū no kami bakufu rōjū gaikoku kakari kinyaku chū shodaimyō kenpakusho utsushi Ᏻᨻᖺ㛫ஸ∗ഛ୰Ᏺᖥᗓ⪁୰እᅜ᥃໅ ᙺ୰ㅖ኱ྡᘓⓑ᭩෗. It is used in DNIS for a letter of Tokugawa Nariaki in 1854. America shisetsu taiwasho, Naikaku mokuroku (i), vol. 1, p. 590. Tanaka 1998, p. 53. See also Doi 1997. Beerens 2000, p. 389.

348

133 134 135 136 137 138 139 140 141

142 143 145

146 147 148 149 150 151 152 153

154 155 156 157

159 160

161

162 163

164

EARLY JAPANESE TRADE

Beerens 2000, p. 389. Beerens 2000, p. 382. See also p. 381. Beerens 2000, p. 388. Tanaka 1998, pp. 30–31. Tanaka 1998, p. 33. Relations with the Hayashi are summarized in Tanaka 1998, pp. 30–42. Tanaka 1998, pp. 49–50, 54, 56–57, 60. Tanaka 1998, p. 77. For details of the contents, see Tanaka 1998, pp. 135, 138–39. Tanaka 1998, pp. 80, 144–45. For details, and an account of copies made of the compilation, see Tanaka 1998, pp. 143–79. Tanaka 1998, pp. 231, 412. “Maegaki,” p. 2, in Tsūshin zenran (ii). 144 Tsūshin zenran (i). It was renamed a year later after the creation of cabinet government as Naikaku Bunko. Naikaku mokuroku, p. 4; Naikaku Bunko hyakunen shi 1986, p. 30 Tanaka 1998, p. 66. Cullen 2003, pp. 152 note, 196 note, 197. See also Shizuoka Kenritsu Chūō Toshokan 1970. Cullen 2003, p. 319. See also Kizaki 2005, p. 135 and Fukui 1980, p. 142. DNIS, 1st series, vol. 1, pp. 341, 557, 559. Naikaku mokuroku (iii), pp. 4–5. Fukui 1980, p. 138. Fukui 1980, p. 140. However, a small amount of material from all these institutions, some 6,000 items, found its way to the National Diet library; about 70 percent are machi bugyō records. See, for example, Kasaya 1998, p. 22. Fukui 1980, p. 140. Cullen 2010. Nihon zaisei keizai shiryō, vol. 1, p. 1. 158 Nihon zaisei keizai shiryō. Katsu 1890. Katsu 1893. Biographies afford only a brief glimpse of his research even in the case of the fullest account (Matsuura 2010, pp. 669–75), See also Ishii 1974. For diplomatic reasons, the Gaimushō was not involved in the publication of this volume. See editorial statement in BGKM, vol. 43 (1991), editing letters for 1860. Apart from Hakodate, reestablished in 1854 as a bugyōsho, the survival into modern times of archives from the newly opened offices is negligible. Imaizumi 2011, pp. 146–48, 150. Kasaya 2000, p. 166. Others too have commented on Japanese success in negotiations: Katō 2000; Katō 2004; Cullen 2004, p. 21; and Auslin 2004. Modern protective cover of document errors in giving 9th year in place of 5th year.

CHAPTER 10 1 2 3

4 5 6 7 8 9

Ackroyd 1979; Nakai 1988. Towards Restoration: the Growth of Political Consciousness in Tokugawa Japan, quoted in Cullen 2003, p. 307. Tashiro’s work was first summarized in an article in Acta Asiatica 1976. Tashiro 1981 is his major work on the topic. Iwao 1963, pp. 30–31; Iwao 1976, p. 16. Toby 1991, pp. xxvii, xxxvii (preface to original 1984 edition). Toby 1991, p. 11. Toby 1991 (reprint of 1984 edition), pp. 10–15, 24, 242. Arano 1994; Arano 2005. Nakamura 2000, pp. 173–91; Lewis 2003, pp. 96–98. Diplomatic ties—tsūshin ㏻ಙ—of course survived.

NOTES

10

11 12

13

14 15

16

17 18 19 20 21 22 23

24 25

26

27

28 29

30 31

32

33 34

349

There is some uncertainty about low and varied early figures for the size of the permitted trade. It is as low as 120 kan in Tashiro (1976, p. 91). The problem rests in the distinction between the tribute trade to China and the private trade. Kaempfer’s estimate was 125,000 tael (or 1250 kanme), and may serve also as a working figure for the eighteenth century. Bodart-Bailey 1999, p. 228. See Cullen 2013 for a survey of Edo, daimyo, and Nagasaki records. For the 1709–1714 returns, see “Kaidai” by Yamawaki (1970), twelve pages at the end of volume 2 of Tōban kamotsuchō (1970). For the manuscript of 1804 by the merchant Murakami under the title Sashidashi chō ᕪฟᖒ, see Yamawaki 1964, pp. 196−203. Nagazumi 1987, p. 6. Nagazumi drew the details of trade in 1652–1657 from AJ 823 Staten houdende opgave van goederen door Chinese junken an Nagasaki angevoerd in the State archives in the Hague DDR 1986, vol. 1 (1680–1690), p. 83. DDR 1987, vol. 2 (1690–1700), p. 60. Many of the Dutch resided for years in Dejima. The slaves of the Dutch, brought by individual Dutchmen as their servants, may have spent the rest of their lives in the factory, even after their owners had left. That may explain why they acquired a good command of Japanese, about which a question was posed at the shogunal reception of the Dutch in 1684. DDR 1986, vol. 1 (1680–1690), p. 32. For a useful account of the duration of stays and responsibilities of some of the members of the factory, see Matsui 2015, pp. 151–57. Nagazumi 1987. DDR 1987, vol. 2 (1690–1700), p. 79. DDR 1990, vol. 3 (1700–1710), p. 59 DDM 1700–1740, p. 222. DDM 1700–1740, p. 326. DDM 1700–1740, p. 368. DDR 1993, vol. 7 (1740–1760), p. 308. The figures appear in TKIR 1912–1923, vol 4, kan ᕳ160 and 161. See also footnote 62. For the Dutch, he had figures for 1698, 1715, and a run for 1760–1839 (1775, 1776, 1777, and 1820 missing), and for the Chinese trade for 1755–1839 (1820 missing). Katsu does not indicate sources, but the sequence of information suggests that he drew on three shahon ෗ ᮏfor his Dutch data, and two for his Chinese data. He also had isolated and imperfect data for Chinese trade for 1688, 1698, 1742, 1746, and 1749 from a single shahon. From a further three isolated sources, he appears to have drawn figures for 1749 (a duplicate figure), 1765, and 1791. Katsu 1976, pp. 3–60 (from part four of Suijinroku). The absence in Meiji times of surviving trade data for earlier years is confirmed in the huge Nihon zaisei keizai shiryō ᪥ᮏ㈈ᨻ⤒῭ྐᩱ in ten volumes, assembled before the destruction of the Ōkurashō in the 1923 earthquake. Yamawaki 1970, pp. 1, 11; Nagazumi 1987, p. 6; Nakamura 2000, pp. 192–96. Published as Tōban kamotsuchō, 1969–1970, 2 vols. Kamotsuchō ㈌≀ᖒin place of kamotsu aratame chō ㈌≀ᨵᖒ. His prefixing the term Tōban ၈⻅to the title of the collection, however, is correct as it included both Dutch and Chinese shipping. The Tōban kamotsu aratame chō had a complete run for the first sixteen vessels; thereafter gaps emerged in the listing. By perusing footnotes in Nakamura 2000, pp. 201, 203, 207, 237, the catalogue numbers in the Ōmura library can be identified as 102–1, 103–2, 103–3, 103–4, 103–5, 103–12, and 103–13. Nakamura 2000, p. 228, footnote 109. On Koga, see Nakajima 2007, p. 37. At first a teacher in a middle school, he returned to Nagasaki to devote his life to collecting documents and writing, in a close association with the Kenritsu Toshokan. Hao 2015, pp. 37 and 53, note 11. DDM 1700–1740, p. 62.

350

35 36 37 38 39 40 41 42 43 44 45 46 47

48 49 50

51 52 53 54 55 56 57

58 59 60 61 62

63

63 65 66

67

68 69

70 71 72

EARLY JAPANESE TRADE

DDR 1993, vol. 7 (1740–1760), p. 72. There is a very general but all too rare account in Yasutaka 2010, pp. 148–52. For details, see Cullen 2013, p. 37 Tōtsūji Kaisho nichiroku 1984, vol 1, pp. 1–7; vol. 7, pp,;101–14.. Hellyer 2009, pp. 87, 117–20. DDM 1700–1740, p. 465. DDM 1700–1740, p. 479. Ōba 1967, p. 67. Ishida 2009. See trade details for 1659, 1686, 1693, and 1804 in Yamawaki 1964, pp. 35, 72, 103, 206. Valued at £23.37 per kan (Cullen 2003, p. 41, note 51). Ishii 1998, p. 10. Tōsen Shinkō kaitōroku ၈⯪㐍 ᅇ᝚㘓, Shimabarabon Tōjin fūsetsugaki ᓥཎᮏ၈ே㢼ㄝ ᭩, and Wappu tomechō, ๭➢␃ᖒ; Ōba 1974. The first and third items are single kan or volume sources. The second—the Shimabara fūsetsugaki—consists of thirty-seven volumes, Shimabara fūsetsugaki included. Tōsen fūsetsugaki from all sources amount to seventy-seven volumes. Ishii 1998, pp. 6–7. Ishii 1998, p. 56. Ishii 1998, pp. 211–13. Matsuura 2007, pp. 191, 193. Ishii 1998, p. 243. The licences were suspect to the Chinese authorities at the outset, and were seized by them for a considerable period of time before being returned to their holders. On tonnage, see the “Glossary of Weights and Measures” in the Appendix. Souza 1986, p. 133. On tonnage, see also Ishii 1998, pp. 3–4. Van Overmeer Fisscher 1833, p. 251. Broughton 1804, vol. 1, p. 239; BPP, vol. 5, p. 681. BPP, vol. 5, p. 681. Nakamura 1988, pp. 344, 347. Nakamura 1988, p. 433. A memoir in 1837 assumed a value of 490 kan for cargoes. Hellyer 2009, p. 136. See the Appendix for a glossary of weights and measures. The figures covered gold, silver, and, from 1663, copper, exported from 1648 to 1708. Ackroyd 1979, p. 24. TKIR 1912–1922, vol. 4. The figures are reproduced in Ōta 1992, pp. 93–95. They are also in Iwao 1953, p. 22. Iwao 1953, p. 19; Yamawaki 1964, p. 106. The figures are reproduced in Ōta 1992, pp. 93–95. They are also in Iwao 1953, p. 22. Iwao’s figures are only superficially different, with tsukaisutegin excluded from his table. Ōta has expressed doubts about the figures (Ōta 2000, pp. 152–53). The erratic values from year to year, in some years wholly out of proportion with any conceivable level, suggest strongly that they are to be disregarded. Nakamura 1988, p. 347. The total net of tsukaisutegin. Cullen 2003, p. 42. Round-figure conversion of silver kan into gold ryō at 17.2 ryō to a kan (at an official exchange rate of 58 monme to the ryō). Yamawaki 1964, p. 103. On shiromono gae, see Ackroyd 1979, pp. 242–43, and for a fuller account, Ōta 1992, pp. 345–64. Price calculated on the basis of the data for 1697 in Yao 1998, p. 88. They give 8.4 piculs of copper to a kan of silver. Yamawaki 1964, p. 106. Tashiro 1976, p. 90. Ackroyd 1979, pp. 242–43, 249.

NOTES

73

74 75 76 77

78

79 80 81

82 83 84 85

86 87 88

89 90 91 92 93

94 95

96 97

98 99

100

101 102 103 104 105 106 107

351

The locus classicus is Nakamura 1988, pp. 390–422. See also Nagasaki kenshi 1985, pp. 581–98. DDM 1700–1740, p. 215. Yamawaki 1964, pp. 57, 214. Yamawaki 1964, pp. 57, 72, 103. This paragraph draws on Yamawaki 1964, pp. 57, 213–15, and on the trade returns in Nagazumi 1987. There is a convenient table of exports in kin ᩹from the outset of the copper trade up to 1715 in Kobata 1993, pp. 695–97. DDR 1987, vol. 2 (1690–1700), p. 135. Yamawaki 1964, p. 57. DDR 1990, vol. 3 (1700–1710), p. 5 The exceptional silver exports for the year 1700 included 371 kan in silverware (Yamawaki 1964, p. 57). DDM 1700–1740, p. 110. Kobata 1993, p. 687. DDM 1700–1740, p. 175. Nakai 1988, pp. 106–107, 111–12; Ackroyd 1999, p. 249. On the decree, see Nakai 1988, pp. 109–14; Yamawaki 1964, pp. 140–47. Nakai 1988, pp. 113–14. Nakamura 1988, pp. 348–51. There is an almost complete absence of statistics of copper exports from Japanese sources for the second half of the 1710s and the following three decades (see footnote 26 above). In the few incomplete figures given in Katsu (1976), a count of ten vessels and twenty thousand piculs for 1746 is probably close to actual exports. Nagazumi (1987) has counts of Chinese vessels for some years, with 16,590 piculs in 1725, and another of twelve vessels and 17,415 piculs in 1745 that are probably close to the actual figures. Her statistics for 1724 and 1725 suggest that many Chinese vessels were rationed to 760 piculs; from the 1750s, the figures provided by Katsu suggest totals of twenty thousand piculs or less. There was no gap in Dutch figures. For them, see the comprehensive accounts by Shimada 2006 and Suzuki 2012. Nakamura 1988, p. 347. Nakamura 1988, pp. 372–75, provides details of ceilings from 1715 to 1848. See footnote 10. Tashiro 1976, p. 87. Lewis 2003, pp. 96–98. See figures in Tashiro 1976, p. 89 and in Yamawaki 1964, p. 229. The fullest recent account is in Nakamura 2000, pp. 173–91. Yamawaki 1964, pp. 206, 208. In 1839, 2,093 kanme of imports were retained, that is, not expended on exports. Hellyer 2009, pp. 80–81, 84, 97, 121–24, 173–74, 178, 182, 183. Shimada 2006, p. 25. For a table of the Bō family succession, see Hao 2015, p. 181. Charts illustrating the nature and evolution of the new organization of trade are in Hao 2015, pp. 245–46. For details of both ships and houses engaged in the silver trade, see Hao 2015, pp. 259 and 266. In peak years, imports exceeded 1,000 kan, and in the two years of 1801 and 1802 were around the 2,000 kan mark. From 1804, they fell very sharply. Nakamura 1988, p. 447. Nakamura 2000, pp. 146, 148; Hao 2015, p. 15. Bodart-Bailey 1999, pp. 221, 222, 227, 391, 393–95, 397, 435, 437, 438. DDM 1700–1740, p. 481. DDM 1700–1740, p. 176. “Hovering” is a technical term formerly of English customs usage. DDM 1700–1740, p. 161. Nakamura 2000, p. 151.

352

108 109 110 111 112 113 114

115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134

135 136

137

138 139 140

141 142

143 144 145

EARLY JAPANESE TRADE

Yamawaki 1964. See the long quotation on pp. 269–71. DDM 1700–1740, p. 181. DDM 1700–1740, p. 223. Nakamura 2000, p. 148. Nakamura 2000, p. 152. Yamamoto 1995, p. 159. See also Wilson 2015, pp. 67–93. DDM 1700–1740, p. 222. The metsuke’s name was Watanabe Geki. Wilson 2015, pp. 80–83, 84–86, 89. DDM 1700–1740, p 223. DDM 1700–1740, p. 317. Yamawaki 1964, pp. 266–69. Uehara 1981, pp. 197, 209–10. Yamawaki 1964, p. 271. Ichimura and Ōishi 1995, pp. 50–51; Miyamoto and Hayami 1988, pp. 163–64. Yamawaki 1965, pp. 93–95, 100. Uehara 1981, p. 211. Sakai 1964, p. 402. Bōnotsu, a center at an earlier date, is a suspected case. See Hellyer 2009, pp. 46, 130. Kagoshima kenshi 1940, pp. 758–62. Hellyer 2009, pp. 86–88, 132. Miyashita 1997, pp. 250–67. Hellyer 2009, pp. 139–40. Hellyer 2009, pp. 134–64. Yamawaki 1964, pp. 206, 208. Sakai 1964, pp. 399–400. TKIR zokushū, vol.1, p. 164. From reports/commentary, 1835.3. TKIR zokushū, vol I, p. 162, 1835.3; p. 186, 1835.7. TKIR zokushū, vol.1, p. 130, referred to on the first page of kan no. 8 in this volume. Kataura is sited on the tip of a promontory (in the modern district of Kasasa-chō ➟ Ἃ ⏫ in Minami Satsuma-shi ༡ࡉࡘࡲᕷ) on the southern end of a long curve, on the external coast of the western arm of Kagoshima bay. From Yahoo Maps. TKIR zokushū, vol. 1, pp. 130–48. TKIR zokushū, vol. 1, p. 136. The location has not been identified with certainty but may be Kozechō ᑠ℩⏫ in Ichiki kushikino shi ࠸ࡕࡁ୵ᮌ㔝ᕷ, Kagoshima-ken 㮵ඣᓥ┴ at the northern end of the curve of coast referred to in note 134. This location, if correct, is within the immediate hinterland of Kagoshima town itself, and is not the island suggested in some modern writing. TKIR zokushū, vol. 1, p. 183. Japanese practice involved limited legal action against Chinese in the past. They were usually expelled or prohibited from returning, and jailing was uncommon. They were said not to be subject to physical punishment (Hao 2015, pp. 150–51), though in fact the Dutch recorded the torturing of Chinese in 1718 (DDM 1700–1740, p. 222). TKIR zokushū, vo1. 1, pp. 337–38. 1836.4.12. Yamawaki 1964, p. 208. See the long reports/commentary on trade, TKIR zokushū, vol. 1, 1835.3, pp. 162–86; 1835.7 to 1835.11, pp. 186–202; 1836.4 (also with correspondence from earlier dates), pp. 320–58. Details of bugyōsho dealings with shipowners are continued in later pages of the volume. Yamawaki 1964, pp. 272–75; Uehara 1981, p. 250. Sakai 1964, p. 398; Yamawaki 1964, pp. 272–75. Nakamura 1988, pp. 501–504 has details for the permitted items from 1825 to 1846. Matsui 1975, p. 244. Matsui 1975, p. 246. Figure for 1803 in Nakamura 1988, p. 433.

NOTES

146

147 148

149 150 151 152 153

154 155 156 157 158 159

353

Uehara 1981, p. 266. Matsui has taken the amount of the actual trade to have been at its permitted ceiling (2015, p. 246), which was far from being the case. Nakamura 1988, p. 508. Uehara 1981, pp. 271, 272. Figures for fifteen years. There are fuller tables for every tenth year from 1821 to 1871 in Kagoshima kenshi 1940, pp. 765–72. BPP, vol. 4, p. 13. BPP, vol. 2, p. 109. Satow 1921, p. 179. Jansen 2000, p. 316. Satsuma sales to the Kaisho are disregarded to avoid double counting as they were in all probability reexported to China. Cullen 2009, pp. 205–206. Wilson 2015, pp. 135–70. Cullen 2003, p. 171. Wilson 2015, pp. 162–68. Cullen 2010, pp. 74–79 Ishikawa 1967, pp. 94–95, 98–100.

GLOSSARY ™

bakufu

Literally tent or campaign administration. Used as a term of derision by opponents of the shogunate in the 1860s. Adopted later, arguably inappropriately, to denote Tokagawa era administration at large.

bakuhan taisei

Modern term denoting shared power between shogunate and domains.

bakumatsu

Somewhat elastic term to denote final years or decades of the Tokugawa dynasty

bugyō

General time for senior administrators (of hatamoto status) of shogunate. Recruited from, or promoted to kanjō bugyō grade. The term sometimes rendered as magistrate or governor. The French term intendant is a happier version.

Dagregister

Office diary of the Dutch factory in Dejima

daikan

Middle-ranking Kanjōsho officials (156 in number) responsible for raising revenue in the tenryō lands scattered as far afield as Kyushu (where the town of Hita managed the relatively small tenryō income of the island).

daimyo

Territorial rulers, of either fudai or tozama status

dollar

Mexican dollar . From 1878 a Japanese minted dollar was issued and made valid in 1879 in all public and private payments in Japan . A mint ratio of gold and sliver in 1871 settled inflexibly had the result that gold which held its value was hoarded or exported, and in return substantial imports of silver turned Japan de facto though not de jure inio a silver currency country. The somewhat unreal currency situation in Japan - disappearing gold and silver replacing it even in the Edo region ( where traditionally gold had been at the hearth of the currency) was reflected in the silver yen being close to par or at par with the gold yen or even commanding a premium.

fudai

Fudai daimyo were variously descendants of domain families who had ffought on the side of Tokugawa Ieyasu in the battle of Sekigahara in 1600 or hatamoto promoted to fudai rank in later times. From the fudai were recruited the rōjū or ministers of the shogunate,

fūsetsugaki

Reports of external news made on arrival of Chinese and Dutch vessels in Nagasaki

Gaimushō

Foreign ministry

han

Term used to denote territory of a daimyo family. Within the Edo period it was a relatively late usage, replacing ryōgoku. In recent times in English “domain” rather than “han” has been favoured to describe the territorial unit. 355

356

EARLY JAPANESE TRADE

hatamoto

Directly retained samurai of shogun. While the upper limit of their income was 10,000 koku, they mostly had modest incomes measured in hundreds of koku. From them were recruited the senior administrates of the shogunate.

hofreis

The visits to Edo by the opperhoofd of the Dutch factory at Dejima to make his obeissance to the shogun. Formerly made annually, bit much less frequently in last 100 years of shogunate.

Hyōjōsho

Judicial office of shogunate

Hankachō

Register of decisions in criminal cases in Nagasaki, 1666- 1867

(Meiji) Ishin

Imperial restoration in 1868 , also describing post-1868 decades

Jisha bugyō

Officers responsible for supervision of territories of temples and shrines (collectively holding substantial property and a large resident population). They were of fudai rank. As many as four shared the work. Given their high rank, they often were included as members of commissions handling major affairs of state.

Ikki

Rural unrest or protest

Kaisho

Nagasaki Finance office

kan㈏

Unit of weight of 1,000 momme or 6 kin of silver (8.27 lb or 3.75 kg.)

kanᕳ

volume, in the sense of bound or stitched collection of papers (on occasion very large)

kanme㈏┠

kanme indicated silver as a money of account. It passed by weight, assessed on a weighing scale, its purchasing power determined by the fluctuating price of silver as a metal.

kanjō bugyō

Officers heading the Kanjōsho (8 in all), and as machi bugyō charged with administration of Edo and of other shogunal cites which peaked at a total of11. Each city had two bugyō, in Edo sharing administration and in other major cities serving in a rotation including one year in Edo.

Kanjōsho

Office responsible for collecting the revenue of the shogunate and for auditing its expenditure

koku

Equivalent of five bushels of rice, the notional amount of rice to feed one individual for a year. Koku also used as a measure of tonnage of vessels.

kuni

Archaic term for provinces, which had lost administrative significance. However, as they had a precise geographical boundaries, the term had its uses (notably in population returns which were converted from a domain basis to a kuni one).

kura

Literally rice store or warehouse. Kuramai (kura rice) or bukemai (samurai rice) referred to rice stored on domain account in contrast with nayamai or rice on merchant account.

metsuke

Officers originally holding primarily a judicial or police role, but in latter times widely entrusted on an ad hoc basis with major responsibilities. Many were promoted in time to grade of bugyō.

Monbushō

Post-1868 word forMinistry of Education

GLOSSARY

357

Momijiyama

Location of the archives of the shogunate in Edo Castle . The major loss was by fire in 1873 of records moved from the Momijiyama to the Castle’s hon maru (central precinct or shogunal residence)

nikki

Regularly maintained office diary, standard in Japanese administration

nishi

Registers recording events in domain administration

Ōkurashō

Ministry of Finance

o-metsuke

Senior rank of two grades of inspector. Often involved in major issues (an o-metsuke was signatory of the report on population census held every six years)

opperhoofd

Head of the Dutch factory in Nagasaki, described by the shogunate as “Oranda kapitan”

rōjū

Counselors, usually 4 or 5 in number, serving as cabinet of shogunate

ryō

Gold coin (koban) and also a money of account valuing cash transactions the medley of coins minted in different periods. Its standing was made nation-wide in 1869, and in 1871 it was replaced by a new coin the gold yen. Trade figures in (silver) dollars from ports were converted at successive dates into ryō and gold yen in Yokohama and later Tokyo.

satsu ෉

A counter for books; can also mean volume or book (though more in the sense of a file than of a bound volume)

shahon෗ᮏ

Manuscript copy (implying a transcript of letters or documents)

Shoheizaka Gakumon(jo)

Term used from 1797 to denote Hayashi Academy in its new function as official school of shogunate. Its head had the title of daigaku no kami.

shiromono gae

Commodities marked for compulsory barter in export trade in lieu of payment in silver

tael

A money of account long used in eastern Asia to evaluate the currencies of the region, and hence familiar also to traders in Nagasaki.The first aggregates of Nagasaki’s new foreign trade from 18 59 were in tael There was some confusion or uncertainly among the first consular officers in conversion rates of tael into dollars.

tenryō

Directly managed lands of shogun, i.e lands not ceded to fudai or to hatamoto.

tonya ၥᒇ

As a singular noun signifies a trader, typically a wholesaler, also as collective noun a group of tonya sharing a common purpose. A British consular officer in Osaka in 1867 seeking information from a tonya of 10 members (one dealing in foreign goods, i.e. goods from Nagasaki) noted the total number of tonya groups as 200. Translation of tonya as guild is misleading. Such entities were a shogunal device to raise funds from traders grouped for the purpose into a relatively small number of units (in Osaka in 1867 twenty).

tozama

Domains which either opposed Tokugawa Ieyasu at Sekigahara in 1600, or stood aside. As outsiders they did not participate in later Tokugawa government.

TKIR

A modern abbreviation for Tsūkō ichiran (survey of foreign intercourse), in two parts, the first completed in 1853, the second in 1856

tsūji

Interpreters of Dutch (oranda tsūji ) and of Chinese (tōtsūji)

uchi harai

Policy of firing on and driving off foreign vessels

INDEX

™

Abe Masahiro, Fukuyama daimyo, senior rōjū, 248, 256, 303 Abe Seishō, Fukuyama daimyo, rōjū, 238, 255 Abe Shōzō, Osaka machi bugyō, 138, 141, 147, 148, 153, 169, 170 Alcock Rutherford, British consul-general, 184, 185, 188, 193, 305 Andō Shōeki, physcian and political philosopher, xxv Arai Hakuseki, 141, 241, 275, 285 Arano Yasunori, historian, 276 Armenault, Daniel, opperhoofd, 33 Aston, W.G., British consular service, 38, 43, 236, 252 Baba Sajūrō, junior interpreter, 12, 15, 36, 40, 42, 253 Blomhoff, Jan Cock, opperhoofd, 29, 34, 35, 42 Bodart-Bailey, Beatrice M., historian, 9 Boockestijn, Pieter van, opperhoofd, 31, 32 Butenheim, Henrich von, opperhoofd, 27 Borton. Hugh, historian, xxv Broughton, W.R., British naval captain, 12 Brunt Wm, 1697, Dejiima, lnowledge of Japanese, 26 Brunt William, 1809, Titsingh letter to him in Amsterdam, 37 Caron, François, opperhoofd, Hirado, Dejima (1639–41), 21 Chamberlin, Basil Hall, japanologist, 43 Chassé, Petrus Theodorus, opperhoofd, 33 Chikuzen domain, 298 China, trade, 62 xxxvi-xxxvii, 68, 277, chapter 10 passim Opium and Arrow wars, 60 Boxer rebellion, 64 Chinese in Meiji ports, 205–6 Bō family commissioned by Ching authorities, 295

See also Shanghai, Canton, Hong Kong, Nagazumi Chōshū, xx vii, 62, 65, 81 85, 305 De Laver, Rogier, opperhoofd, 32 Doeff, Hendrik, opperhoofd, 10, 20, 30, 31, 33, 34, 37, 39–40, 251 158, 159, 159 Domberg, Dejima expulsion, 18, 20, 23–4, 38–9 Droppers, Garret, early population, study, 77 Edo, 28, 108–112, 114–115, 134, 135, 137–142, 143, 145, 147, 148, 150, 153, 157, 158, 159, 160, 16 1, 162, 168, 170–172, 1 72, 205, 257, 306–7 archives 245 fires, xxxiii-xxx1v, 122 rice consumption (1727, 160–2 imports (1726), 158–9, 164; mports(1856), 143, 162–3, 171, 172; Osaka-Edo trade, 143, 145–6, 148, 150, 152, 157–8, 159, 170 population, 107–8, 108–9, 110–112 Egawa Tarōzaemon, daikan, 254 Esserd, understeward, Hrado and Dejima, 21, 22 Ezo (modern Hokkaido and northern islands), 8, 11, 12, 19, 138, 163, 220, 249, 250, 276, 300. See also Kondō juzō Feith, Arend Willem, opperhoofd, 33 Fisscher, J.F. Van Overmeer, warehouseman, Dejima, 20, 24, 32, 33, 34, 36, 39, 40, 285 Fuji family, donor Nagasaki archives, 239 Fujita, Satoru, historian, xviii, 163, 237 Fukazawa Yukichi, propagandist and founder Keio University, 67 Fukuda Tadaaki, Nagasaki collector of documents, 239 Fukui doman, 250 359

360

INDEX

Fukui Tamotsu, author on archives, 245, 259 Fukuoka domain, 250, 298 Glover, Thomas, merchant, Nagasaki, 305 Golownin, Vassely, Russian naval captain, xxviii, 6, 9, 13, 15, 56, 57, 125, 253, 254, 259, 260 Gotō Sōzaemon, head of Kaisho, 37 Grew Joseph, Amlerican diplomat, xxxix, xl Gubbins, J.H., British consular service, 67 Gozeman, taken prisoner by Phaeton, 41 Hagiwara Utaku, Gaimushō archivist 1880s, 248 Hakodate, 189, 190, 191, 194, 217, 221, 223, 236, 248, 257, 261 Hall, J.W., historian. xviii, xxv, 247 Hanley, Susan and Yamamura Kozo, historians, 80, 87, 91, 95 Hansa(Fansa), wholesaler tailor, Nagasaki, 26, 27 Hara, Kei, prime minister 1918–21, 68 Harris, Townsend, American consul Shimoda, 42 Hasegawa, Akira and Shinbō Hiroshi, Osaka exports, 1724- 1730, 152 Hawaii, 63, 67 Hawks, Francis L., compiler of report on Perry expedition. 7, 10, 260 Hayami Akira, demographer, xix, 78 . 83, 96, 99, 100, 141, 253, 299 with Miyamoto Matao, co-editor of Nihon keiza-shi (vol 1), 156 Hayashi Reiko, historian, 139, 165 Hayashi family, academy (Shōheikō) moved in1691 to Yushima Seidō becoming from 1797 Shoheizaka gakumonjo, 11, 76, 112, 172, 245–6, 251, 252, 253 257–8 Akira (in life time Fukusai), xxi, 15, 141, 236, 253 -4 Jussai, xxi, 237, 246 decline of house and later dispersal of archives, 257–8, 259 Hemmij, Gijbert Hendrik, opperhoofd, 33 Hesselinks, historian, R.H, 9 Hirado, xxvi, 20, 21, 22 24, 66, 274, 275 Hirōsakai domain, archives, 247 Hokkaido, 238 Homoed, Gijbert van, opperhoofd, 33, Hong Kong, 186, 205 Honjō, Eijirō, historian, 77, 78, 87

Hosokawa, Hiroyo, 1883 population data, 76, 121 Hotta Mayasoshi, Sakura daimyo, senior rōjū, 248, 249, .256 Hyōgo1, Hyōgo (Osaka), 149, 185, 197, 203, 215, 2i6, 305 See also Osaka Ichirōzaemon, head interpreter(Chinese), Nagasaki, 28, 278 Ienari 11th Tokugawa shogun, xxvi, 248, 299 Ii Naotsule, daimyo Hikone, senionr roju 1858, xxvii, 248 Ikeda, daimyo family, (Okayama), archives, 247 Imlah, A.H., economist, 218, 219 Inoue Kaoru, vice-minister of Finance, 203 Inoue, Mizue, 1828 census, 77, 121 Ishida Chijirō, study of Japanese foreing trade, 183? 283 Ishii Takeshi, 1944 study of foreign trade, 187, 188, 189 Ishikawa, Sōjirō, population figures reputedly from a Kanjōsho official, 123 Isshiki, Osaka bugyō, 75, 113–115. 121, 124, 171 Ishin Suden, Gozan monk, 235, 241 . See also Tenryūji temple Ishimoto Heibei, Satsuma agent, Nagasaki, 301 Iwao Seiichi, historian, 275, 285 Izawa Masayoshi, former Nagasaki bugyo, 256 Jansen, Marius, historian, 18, 26 Kaempfer, Englebert, physician Dejima, xxvi, xxiv-xxv, xxvi, 6, 7, 19, 22, 27, 29, 30, 31, 32, 35, 48–9, 52, 53, 54, 60, 296, 299 Kaga domain, 85, 139 Kamchatka peninsula, 51, 52

1

When Osaka opened as a port for foreign trade in 1868 Hyogo’s deep water ensured that from the outset it (as later Kobe as well) served as port for the foreign trade of Osaka . Until 1885 British consular reports were presented under the joint name of Hyōgo and Osaka. See p. 339, footnote 65 .

INDEX

Kanagawa,2 185, 186, 188, 189, 194, 198, 200, 202, 203, 213, 215, 216, 221, 222, 226, 257, 261, 305. See also Yokohama. Kanai, Toshiyuki, ward head Nagasaki, xxiiixxiv, 239, 240, 282 Kaneko Kantarō, proposal for a national history, 261 Kantō (region), 99, 100, 100–101, 102, 166 Kasaya, Kazuhiko, historian, xxi, 246, 247, 248, 261 Katagiri Kazuo, historian, 251, 252 Kataura, Satsuma smuggling location, 302 Katō Yūzō, historian, xxviii Katsu Kaishū, record compilations, xxiii, 75, 76, 77, 107, 110, 111, 112, 115, 117, 118, 121, 123, 124, 142, 158, 163, 171–2, 235, 243, 260, 279 Kawaii, trade figures from 1859, 187 Kawakatsu Heita, historian, 65 Keijser (recte Keysering, Johan Georg,) horseman, 28 Ketelaar, physician, Dejima., 30 Kikuchi Mitsuoki, president National Archives, xx, 235 Kinki (region), 99, 100, 101, 102, 108 Kōda Shigetomo, historian, xix, 110, 113, 114–115 i44–5, 147, 154, 155, 170, 243 Koga Jūjirō, Nagasaki érudit, 239, 280 Kokura, port in Kyushu, 298 Komiya Yasusuke, early Meiji demographic studies, 76, 121, 122 Kondō jūzō,. survey of Ezo, 4, 245, 254 Korea, 64, 67, 68, 70, 245, 248, 255, 259, 275, 276, 282 289, 293, 294, 297. See also Tsushima Koriyama, village in Nihonmatsu domain, 83 Kuchiki Masatsuna, daimyo Fukuchiyama, 35, 37, 38 Kume Kunitake, secretary Iwakura mission, 59 Kure Tokutarō, named copyist of document, 251 2

Removal from Kanagawa to Yokohama was decreed in mid-1859. Though merhants lost no time in moving to a deep-water site, the British consular station long remained at its old location, though its reports were of Yokohama trade.

361

Kurile islands, 4, 11, 12 Kuroda, dainyo, Fukuoka domain archives, 250 Kuze Hirotami, Nagasaki bugyo. 37, 38, 40 Kuwana Matsudaira, library, 237 Labrousse, Ernest., French hiistoran, 4 4 Lapérouse, comte de, Jean-Francois, de Galaup, French naval officer, surveyor of north Pacific, 12 Laxman, Adam, Russian trade mission in Ezo, 1792–3, 19 Liaotung peninsula. 63 MacDonald, Ranald. Infiltration into Japan (1848), 13, 240 Machida family, possessor of copies earlier census results, 120, 122 Machin, Roger, three texts of visit of Return (1673), 251 Magarbuchi kai no kami, post-Phaeton inspector of Nagasaki defences, 252 Mahan, Alfred, American naval strategist, 63 Makino Nobuaki, Japanese diplomat, 69 Marsden. William, Amsterdam, Titsisngh’s 1809 letter to, 37 Maruyama Masao, 7–8, 59. See also Norman, E.H. Matsudaira Sadanobu, senior rōjū, xviii, xxvii. xxviii, 10, 11, 66, 103, 124, 237, 238, 276 Matsukata, Masayoshi, xxix, 202, 203 “Matsukata” deflation, 210 archival deposit, 202, 203 Matsumae, domain and daimyo family, 61, 252, 254, 297 Matsuo family, archive deposit, 202, 203 Matsura, Atsushi, reputed holder of a 1821mansicrspti, 237 Matsuura, Tōkei, compiler of Kokon shūren, 242, 251 Matsushiro domain, archives, 247 McClain, J., article on Osaka trade, 139, 165 witah Wakita Osamu, study of Oaaka 138 Meijlan, Germain Felix, opperhoofd, 20, 29, 34 -35, 37–8 Mensingh Harmanus (Hermanus), book keeper in 1690s, opperhoofd ( 1705–6, 1707–1708, 1709–1710), 25, 27 Minami Kazuo, 1840 census, 77, 123 Mito domain, 81, 82, 83, 86, 87, 89, 90, 96–7, 101, 248, 249, 250, 252

362

INDEX

Mitford, Algernon, British diplomat, 171, 220, 244 Miyamoto Matao, xxx, 145, 151, 156, 157 Mizuno Tadakuni, xviii, xxvii, 46, 163, 248 Mogami Tokunai, explorer of northern islands, 8, 19 Molesworth, Robert, viscount, politician and writer, 53 Mori Kahee, 1934 article on population, 91, 92, 125 Moriyama Einosuke interpreter of English, 13, 42 Motoki, interpreter family, Nagasaki, xxxvii, 5, 25, 244, 281 Motoki Tarōemon long standing informant to Dutch, 25, 278 Mukai, Seidō (Confucian temple), Nagasaki. xxiii, xxiv, 240, 282 Murakami Teisuke (“Teske”), interlocutor of Golownn, 9, 253 Mutō Chōzō, high school teacher, collector of papers, 239 Matsui Masato, estimate of Satsuma exports in 1830s, 303 Mutsu kuni, 86, 89, 90 Nagazumi Yōko, Chinese trade, 278, 280 Nakasaki, 5, 10, 12, 21, 16, 31, 39, 41, 61, 84, 105, 107, 119, 123, 143, 135, 136, 137, 138, 143, 150184, 185, 186, 189, 190, 191, 194, 195, 205, 213, 222–3, 224–5 239, 242, 243, 248, 251, 255–6, 257, 248, 254, 261, pp.273–308 passim Diary of Saga watchouse (1804), 240 English ship Return at Nagasaki (1673), 250–251, 276; and Phaeton (1808), 12, 36, 61, 240, 246, 250–251, 252, 306 fires, xxxiv-xxxv For Russian mission 1804, see Rezanov. Niigata, 139, 297, 301 Nakamura Tadashi, historian, 276 Nakayama, interpreter famliy, Nagasaki, xxxvii, 243, 244, 281 Nanbu (Morioka) domain, 82, 83, 84, 85 85, 87, 89–90, 91–96, 97, 98 archives, 247

Nishiyama Genbun, compiler of Kanchū Hisaku, 118, 119, 120 Nitobe Sengaku, population figures, 92 Norman, E.H. Canadian diplomat and historian, xxv, 7, 8 Northern Pacific, early geographical knowledge of, 50, 51–52 Ogyū Sorai, poilitical philosopher, 103–4, 244 Oishi Shinzaburō, xix, 138, 145, 154, 158, 162 and IchimuraYuichi, 299 Okayama domain, archives, 87, 247 Okuma Shigenobu, Saga politician, proponent of archival reform, founder of Waseda University, xxii, 260 Ooka Tadasuke, Edo machi bugyō, 140, 141 Ooms, Herman, historian, xviii Osaka, 75, 76, 105, 107, 108, 109–110, 112– 115, 134, 135, 136, 137–142, 142, 143, 145, 146, 147, 148, 149, 150, 152, 153, 156, 157, 158, 159, 160, 161, 162, 163, 168, 171, 172, 173–6, 190, 197, 205, 215, 217, 220, 221, 238–9, 297, 298, 303, 305 Dōjima rice exchange, 148, 167, 169, 171, 243 Osaka Chamber of Commerce, 144, 239 Osaka-shi Shi Hensanjo, 143–4, 1 54 Osaka goods to Edo, 143, 148, 152, 157–8, 159, 164–5 imports and exports (1714,) 146, 154–6 imcomplete figures ( 1736, 1766), 146, 156–157 rice imports, 151, 161–162, 166–169, 173–4 coastal import of 20 commodiites, 1866, 223–4 population, 107–8, 109, 112–114, 124 See also Hyōgo, Hyōgo/Osaka ōta Nanpō, collected population details, also as a minor Nagasaki official met Rezanov, 119–20, 121, 242 Outhorm, Cornelis van, opperhoofd, 27 Parkes, Sir Harry, 9, 67, 196, 216 Paske-Smith, M.B.T., British consular service Japan (1921–1931), 187, 188, 189, 191, 194, 221- 225 passim Perrin, Noel. Tokugawa era demilitaritsation 65 Perry, Matheew C., Coimmodore American fleet (1853, 1854), 14, 42, 59, 65 253, 307 Philippines, 63

INDEX

Ravina, Mark, historian xxvi, 139, Reede tot de Parkelaar, Johnn Frerrk van, opperhoofd. 33- 4 Rezanov, Nikolai, Nagasaki mission 1804, xxxi, 3, 240, 241, 242, 252–3 Riess Ludwig, historiogtapher, xix, 144 Romberg, Hendris Caspar, opperhoofd, 33, 34 Rostow, Walt, historian, 57 Ryukyus, trade with Satsuma, xxxi, 49, 136, 245, 249, 259, 276–7, 285, 293, 297, 299, 300, 301, 303–303–4, 305. See also Shimazu; Satsuma (Kagoshima) domain Saga domain, 298, 307 Saga watchhouse in Nagasaki (1804 diary), 240 Saigō Takamori, leading Satsuma official, 56, 168, 305 Saishō Shōtai, Gozan monk, 235. See also Tenryūji temple, Sakai, outport for larger vessels for Osaka, 137 Sakai, Robert K., comments on smuggling, 300 Sakhalin, 4 Sanada, daimyo of Matsushiro, 248–9 Satow, Sir Ernest, 9, 38.]43, 56, 168, 305 Satsuma (Kagoshima) domain, 81, 85–6, 89, 90 99–100, 248, 273 298, 299–301. See also Ryukyus; Shimazu Scheuchzer, J.G., translator from German of Kaempfer, 53 Sekiyama, demographer, xxxii, 75, 77, 80 Sendai domain, 80–81, 82, 83, 85. 87, 89, 90, 97–8, 99 Shanghai, xxxvi, 186, 295, 305 Shantung peninsula, 63, 68, 69 Shidehara Kijūrō, (moderate) foreign minister 1920s, 70 Shimabara, revolt 1637–8, 22, 79 Shimazu, 85, 99 Shigehide, 36, 37, 301 Nariakira, 248, 256 Hisamitsu, regent, xxii, 249 See also Satsuma (Kagoshima) domain Shimoda, American consular station, 42, 248 Shimonoseki, narrow strait between Kyushu and Honshu, 62, 297 Shinagawa, shoreline landing in Edo for barges carrying goods from deeper waters. 160, 161 Sidotti, Giovanni Battista, Italian jesuit, 25, 36 Siebold von, 8, 10, 19, 31, 38, 40, 42 Six, Dutch resident, Dejima, 28

363

Shōda famly, archive deposit, 202 Sloane, Sir Hans, Kaempfer ms, 48 Sōraisaki, Satsuma coastal location of Tōjinkan (Chinese residence), 302 Suenaga Tokuzaemon, compiler of diary of visit by Phaeton 1808, 252 Suenaga Tokuzaemon, rapporteur 1738, 29 Sugiyama, Shinya, bakumatsu and early Meiji trade figures, 187, 192, 205, 2o8. 212, 213 Sun Yat-Sen, xxxix, 67 Suzuki, Naoji, trade counts for rice, 169, 171 Suwa domain, Shinono kuni, 84 Swift, Jonathan, author of Gulliver’s Travels, 48–55 passim Takagi Masao, study of Sendai domain population, 80–1 Takagi Sakuzaemo, daikan, report of postPhaeton inspection, 2 52 Takahashi Bonsen, demographer, xxix, 77, 80, 91, 92, 95, 125 Takahashi Kageyasu, first head of Office for translating foreign books, 37, 42, 253 Takashima Shūkan, gunnery expert, 2 83 Tanaka-Van Daalen, Isabel, Dutch knowledge of Japanese, 19, 23 Tanuma Okitsugu, xxvi, xxvii Tartary, 51, 52 Tashiro Kazuo, sakoku trade, 275 Tatemoto Masahiro, adjusted Meiji trade figures, 212, 213 with Baba Masao, study of foreign trade. 212 Temple, Sir William, ambassador to Holland 1680s, 49 Tenryūji temple, Kyoto, seat of Gozan Zen buddhist sect, specialists of Chinese language and hence diplomacy, 235. See also Ishin Suden, Saishō Shōtai (Gozan monks) Thunberg Carl Peter, Dejima physician, 19, 25, 31, 32, 57, 60 Titsingh, Isaac, opperhoofd, 9, 20, 23, 24, 25, 29, 33, 34, 36, 37–39, 41 1779–80, 1781–82 (with war-time extension for 1782–3) short autumal stay 1784 on the annual vessel (to and from Batavia ) Togawa Yasuzumi, bugyo, 253, 254, 301, 302 Tomiinaga, Niffei(sic), apprentice interpreter, 26 Toby, Ronald, historian, 255, 276 Tōhoku, 57, 79, 84, 99, 100, 101, 138, 153, 166, 168, 172

364

INDEX

Tokugawa Nariaki, 59, 248, 249, 2 56 “Mito-sama”, 35 Yoshinobu, 15th and final Tokugawa shogun, xxvii, 246 Tokuzaemon, compiler of diary of visit by Phaeton 1808, 252 Tokuzaemon, rapporteur 1738, 29 Tōsaburō , rapporteur, 29 Tsunayoshi, 5th Tokugawa shogun, xviii, xxvi, 282 Tsushima domain, 236 247–248, 276, 290, 293, 294, 306 archives of Sō family 248 Wakan, factory in Pusan (Korea), 293 Uchida Gonzō, trade figures 1714 (original or copy), 154. 170 United States, relations with East Asia, 59, 62, 63, 64, 65, 68, 69, 70–72. See also Hawaii, Hawks F.L., Perry, Philippines Uehara Kenzen, Satsuma smuggling, 300 Uraga, outport to Edo, 137, 159, 160, 161, 162, 248. 252 Visser, Joosete, junior merchant Dejima, 26 De Vos, Pieter, senior merchant, Dejima, opperhoofd, 1697–8, 1699–1700., 26 De Vries, Dutch explorer, 17th century, 51 VOC, Verenige Oost-Indische Compagnie, 18, 23 Waardenaar Willem, opperhoofd, 33 Wagemans, physician Dejima, 30 Wakita Osamu With Kobayashi Shigeru, 152, 158 Watanabe Kurasuke, Nagasaki collector of documents, 282

White, James, demogapaher, 78 Wijs, Lodewijs, deputy opperhoofd, 26 Wilson, English captain, compilation of Japanese words, 38 Yakushiji Kumatarō, collector of Nagasaki records, 239 Yamagato Bantō, Osaka economic adviser to domains, 138 Yamaguchi Naoki, acquaintance of retired Nagasaki bugyo, 256, 257 Yamaguchi Kazuo, 18th century trade figures, 156, 165 With Ouchi Tsutomu, early Meiji trade statistics, 202, 203, 204 Yamawaki Teijirō historian of trade with China, 280, 285, 288, 299 Yamashita, Shigatami, 1889 article on population, 110, 111, 112 Yasuoka Shigeaki, eighteenth-century Osaka trade statistics, xxx, 145, 152, 154 156, 157 Yokohama, 136, 187, 188, 189, 190, 192, 194, 198, 202, 206, 212, 214, 215–6, 216, 221, 224, 305. Yokohama fire (1866), xxxv, 222 trade figures (1863), 225–227 headquarters of early Customs adminisration, 198 See also Kanagawa Yokoyama, Yoshiyo, variant population figure, 76 Yoshida Shōin, Chōshū teacher, 59 Yoshida Hidefumi, private compulation by Kurume domain official, 242 Yoshimune, 8th Tokugawa shogun, xxvi, 10, 11, 76, 80, 89, 104–105, 116, 141, 146, 157, 282 Yoshio Chūjirō, interpreter, Nagasaki, 1740, 29 Yuzuki Jūzō and Yasuzō Yorie, 1930 article on Edo population, 110