FA Cop 2. SLAVERY IN CHINA DURING THE FORMER HAN DYNASTY 206B.C.— A.D.25 BY CLARENCE MARTIN WILBUR SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN THE FACULTY OF PHILOSOPHY COLUMBIA UNIVERSITY ANTHROPOLOGICAL SERIES FIELD MUSEUM OF NATURAL HISTORY VOLUME 34 JANUARY 15, 1943 PUBLICATION 525 u, or »uw iiB. Return this book on or before the Latest Date stamped below. University of Illinois Library Jk!'. -6 1^3^ m 27 1 7i m. -im Llol — H41 ,J^ \ '-irW ^ Olr: ■--•v^ J (IS H- ^^--..-a;^,.. v->V.' >.a " f\^^ CHINA DUUING THE FuRMKU HAN DYNASTY, CA. 100 B.C. Publications OF FIELD MUSEUM OF NATURAL HISTORY ANTHROPOLOGICAL SERIES Volume 34 THE LIBRARY OF THE SEP 5 1944 UNIVERSITY OF ILLINOIS f.W NATURAL ^\ ^ HISTORY >•) \SWlCAGJ CHICAGO, U.S.A. 1943 SLAVERY IN CHINA DURING THE FORMER HAN DYNASTY SLAVERY IN CHINA DURING THE FORMER HAN DYNASTY 206B.C.— A.D.25 BY CLARENCE MARTIN WILBUR SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN THE FACULTY OF PHILOSOPHY COLUMBIA UNIVERSITY THE LIBRARY OF THE SEP 5 1944 UNIVERSITY OF ILIINOIS ANTHROPOLOGICAL SERIES FIELD MUSEUM OF NATURAL HISTORY VOLUME 34 JANUARY 15, 1943 PUBLICATION 525 PRINTED IN THE UNITED STATES OF AMERICA BY FIELD MUSEUM PRESS 0-^ FA i^e^^ CONTENTS PAGE List of Illustrations 9 Preface 11 PART I Abbreviations Used in Part I 16 List of Rulers During the Former Han Dynasty 16 I. Han History and Society 17 Decline of Feudalism and Beginning of Empire 17 The Han Dynasty 20 First Phase: Founding, Consolidation, and Recuperation .... 20 Second Phase: Expansion and Depletion 22 -<^ Third Phase: Gradual Economic Decline 25 Fourth Phase: Last Minute Reform, and Collapse 27 1/ Distribution of Population, and Urbanization 30 Local and Central Administration 31 Class System 33 Commoners 33 The Bureaucracy 35 The Nobility 36 Consort Families 38 The Lowest Social Classes 40 Fluidity of Social Position 41 n. Historical Sources, and Definition of Terms 50 Historical Sources 50 Nature and Deficiency of Sources 53 N Advantages of Sources 55 V Method of Using Sources 58 "J Problems of Definition 60 . .Tj^lave Terminology of the Former Han Period 64 ^ in. Enslavement 72 vv Enslavement of Criminals and Their Families 72 Distinction Between Convicts and Slaves 80 Enslavement Because of Economic Distress 85 Legality of Sale into Slavery 88 Illegal Forced Enslavement 90 Importation of Foreign Slaves 92 \> IV. Were Prisoners of War Enslaved? 98 'Y Records of Captives 99 Differentiation Between Surrendered and Captured Enemy .... 102 S ■ (J\' Typical Reports of Wars and Captures 103 Presumptive Evidence for Enslavement of Captives 109 Proportion Enslaved, and Disposition of Others 114 >^ 7 8 CONTENTS PAGE V. Acquisition, Hereditary Slavery, and Manumission 118 Acquisition of Slaves 118 Buying and Selling Slaves 121 Hereditary Slavery 126 Manumission of Government Slaves 129 Manumission of Private Slaves 134 VI. Status of Slaves 140 Customary Attitude Toward Slaves 140 Slaves in Criminal Law 146 Legal Rights of Masters over Slaves 152 Slaves in the Courts of Law 156 "Mixed Marriages," and Status of Children 158 VII. Slave Owners and Numbers of Slaves 165 Types of Slave Owners 166 Numbers of Slaves Individually Owned 169 Total Number of Slaves 174 VIII. Service Functions of Private Slaves 178 Duties of Domestic Slaves 178 Treatment and Position of Domestic Slaves 184 Slaves as Instruments of Power 187 IX. Productive Employment of Private Slaves . 195 Evidence and Presumption of Slaves in Agriculture 195 Relation Between Slave and Free Labor 203 Factors That Produced Landless Free Labor 204 Use of Land Owned as Investment 210 _, Probable Unimportance of Slaves in Agriculture 215 tK'f Use of Slaves in Manufactures and Commerce 216 X. Functions of Government Slaves 221 Government Enterprises and Labor Supply 222 Hypothetical Spheres of Government Slave Work 226 Service Duties of Government Slaves 227 Productive Employment of Government Slaves 232 XI. Synthesis 237 Evidences of Historical Development 237 Former Han Slavery Generalized 240 Reasons for Arrested Growth 244 PART II Notation and Methods of Translation 255 Abbreviations Used in Part II 256 Documents: Numbers 1-138 258 Appendix: Abstracts of Lesser Documents 469 Bibliography 473 Index 482 FACING PAGE LIST OF ILLUSTRATIONS PLATES 1. Mortuary figurine of a servant or slave of the Han Period . 178 2. Scenes of upper-class life during the Han Period, showing services often performed by slaves 186 MAP China during the Former Han dynasty 1 PREFACE The economic history of Greece and Rome, and of our own nation before 1860, cannot be understood without a knowledge of slavery. China, too, had a slavery system, but it has never been adequately described by Western historians. Noting many similarities between ancient China and the classical world, modern Chinese scholars have recently devoted considerable attention to this subject, and their research has fostered an ardent dispute concerning its impor- tance in successive periods. How important was slavery in ancient China? Is the social and economic history of that country, or of any of the periods in its development, also unintelligible without a clear picture of slavery and its function there? Chinese slavery may be studied either extensively or intensively. The first method would seek to describe the institution throughout a period longer than that from Homeric times down to the middle of the nineteenth century. More than a hundred years ago Edouard Biot used the extensive method in his "M^moire sur la condition des esclaves et des serviteurs gag^s en Chine." Toni Pippon used the same approach in his "Beitrag zum Chinesischen Sklaven- system," pubHshed in 1936. The present work employs the intensive method, concentrating entirely upon the first period for which native historical literature allows a detailed examination of the system in its historical and economic setting. Chinese slavery did not originate during the Former Han dynasty (206 b.c.-a.d. 25), but it expanded rapidly at that time. Slaves probably then achieved their greatest numbers in proportion to the total population, and the period is the first in which it is possible to suppose, on the basis of historical texts, that slavery had an important place in Chinese economy. This study of Chinese slavery is divided into two sections: Part I seeks the solution to two general questions. What was the nature of Chinese slavery in Former Han times? What were the position and function of slaves in Han society and economics? Since these problems relate to a particular epoch in China's develop- ment, the period itself must be described from the point of view of its unfolding history, its society, and its general economic system. The success of the answers depends upon the adequacy of the available source material, the way it is used, the precision with which the term "slave" is defined, and the degree of identity that can be established between our usage and the terminology found in Chinese sources. The first question asks for a descriptive answer, and requires 11 12 PREFACE an analysis of the sources of slaves, acquisition by owners, the slave trade, hereditary slavery and manumission — everything that might be called the "life cycle" of slaves in the abstract. The position of slaves in Former Han society is illustrated by their status, both customary and legal. Servile status depends to a considerable degree upon the kinds of people that owned slaves, the purposes for which slaves were owned, and the proportion of slaves to the total population. When we know what kinds of people owned slaves it is possible to ask why they owned them, that is to say, what functions the slaves fulfilled: whether they were important to their masters or to the state as producers of wealth — that is, as labor — or whether they were more important in other ways. The response to the two major questions raises a third. What were the place and function of the whole slavery system in China as a whole, during the Former Han period? But is it permissible to discuss China during two and a quarter centuries "as a whole"? What evidences are there of rigidity or of historical change in the slavery system? Answers to such queries should contribute to a solution of the controversy, namely: Was the society of the Former Han period a "slavery society," and was the economic system a "slavery economy"? Part II, which was prepared first, translates and annotates some 140 passages on slaves discovered in historical literature written during the Former Han period or shortly thereafter. When all the basic documents are placed in one section they retain their independent validity (except for that personal factor of translation), and therefore may be employed by any investigator without regard to the analysis of them in Part I. Aside from being the foundation of this book, the documents are an integral part of it, constantly referred to by number in substantiation of all descriptions and con- clusions. Many reveal, inter alia, fascinating details about life in the imperial palaces or patrician households, the political system of the dynasty, and the rise and fall of Chinese statesmen, generals, or relatives of the imperial house by marriage. For casual readers I venture to recommend document Number 14-, which poignantly describes the reunion of a young slave and his elder sister after she unexpectedly became Empress; Number 27, which tells how a chorus girl captivated Emperor Wu and ultimately became his consort; the revealing account of jealousy and intrigue in a royal household, detailed in Number 37; the report, in Number 55, of an imperial investigating commission which substantiated the claim that a PREFACE 13 rustic old dame was the maternal grandmother of Emperor Hslian; the lurid story of a pathologically jealous queen, recounted in Number 6Ji.; and document Number 75, the confidential report of a detective who had been commissioned to spy upon a man deposed from the imperial throne and suspected of plotting a coup d'Hat. There are many other passages of general interest, but perhaps the most revealing of all is Number 107, which gives an eye-witness description of the machinations of an imperial concubine. She so dominated Emperor Ch'eng that he weakly submitted to her wishes and destroyed his only sons, thus dying without an heir, and most calamitous of all, leaving no direct line to carry on his ancestral worship! Many people and institutions have assisted in the preparation of this book. I wish to record first my gratitude to Hollis Adelbert Wilbur and Mary Matteson Wilbur, my parents. At Columbia University Professor William Linn Westermann, himself an authority on ancient slavery, provided the first inspiration for the study, and his continued interest and critical advice have been very stimulating. Dr. Luther Carrington Goodrich, Executive Officer of the Department of Chinese and Japanese at Columbia, provided unflagging encouragement for my research, and has read the entire manuscript, making numerous suggestions for improve- ment. Professor Jan Julius Lodewijk Duyvendak, Director of the Sinologisch Instituut at Leiden and Visiting Professor of Chinese at Columbia, went through most of the translations with infinite care, correcting errors, and proposing many felicitous renderings. Two other friends of long standing in the same department. Dr. Cyrus Henderson Peake and Mr. Chi-chen Wang, have read the manuscript and made numerous useful suggestions. During several years of extensive correspondence. Dr. Homer H. Dubs, Professor of Philosophy at Duke University and translator of the Imperial Annals in the History of the Former Han Dynasty, has thrown light on a number of vexing problems concerning the China of those ancient days. He has meticulously read most of the manuscript, suggesting innumerable corrections and offering additional information from his storehouse of knowledge and notes. Likewise, during a period of more than two years Mr. Charles Y. Hu £fe^^, a gifted graduate student in the University of Chicago, worked with me on the translations and their analysis. It is also a genuine pleasure to express my gratitude to Dr. Charles Sidney Gardner, of Harvard University; to Dr. Herrlee 14 PREFACE Glessner Creel, of the University of Chicago; to Dr. A. W. Hummel, Chief of the Division of Orientalia of the Library of Congress; and to Mr. and Mrs. Carl Whiting Bishop of the Freer Gallery of Art. Mrs. John King Fairbank kindly made the delicate and accurate restoration of the rubbing which is reproduced on Plate 2. Mr. Robert Yule Mi^^, of Field Museum, Assistant in Archaeology, drew the map, and greatly improved the legibility of the sixteenth century copy of the contract for a slave, reproduced in document 83. Miss Rose Harris and Mrs. Anna Pfeiffer did the tedious work of typ- ing, and patiently checked and rechecked citations through several recensions. There is also my wife, who would prefer to see no mention of her part; and indeed there is no way to describe the extent of her aid and of my gratitude. Through the good offices of Mrs. Emily M. Wilcoxson, Librarian at Field Museum, I have been freely allowed to borrow books in Chinese and Japanese from the Division of Orientalia, Library of Congress, and from the Far Eastern libraries of Columbia University, the University of Chicago, and the Harvard-Yenching Institute. Some of the preliminary work on this book was done during after- hours when I was a Fellow of the Social Science Research Council; its support is gratefully acknowledged here again. By a happy arrangement, the American Friends of China, Chicago, enabled me to continue work with Mr. Charles Hu longer than would have been otherwise possible. To the American Council of Learned Societies I am indebted for a grant which provided the electrotypes of the Chinese texts printed in the second part of the book. Mr. Mortimer Graves, Administrative Secretary of the Council, aided greatly by working out the technical details. Dr. George A. Kennedy, of Yale University, supervised the setting of the Chinese type, which was made doubly difficult by the frequency of archaic forms. Finally, there is Field Museum of Natural History, whose Director, Colonel Clifford C. Gregg, allowed me to devote the greater part of the past year, and some of my Museum time previously, to this study. Dr. Paul S. Martin, Chief Curator of Anthropology, constantly encouraged me and, by lightening my routine tasks, allowed me more uninterrupted time for study and writing. Miss Lillian Ross, Associate Editor of Scientific Publications, has greatly improved this book in seeing it through the press. C. Martin Wilbur August 1, ISIfl SLAVERY IN CHINA DURING THE FORMER HAN DYNASTY PART I ABBREVIATIONS USED IN PART I BEFEO Bulletin de I'Ecole Frangaise d' Extreme-Orient BMFEA Bulletin of the Museum of Far Eastern Antiquities CHHP Ch'ing Hua hsiieh pao (Tsing Hua journal) CHS Ch'ien Han shu CLHP Chin Ling hsiieh pao (Nanking journal) HFHD The history of the Former Han dynasty by Pan Kii, translated by Homer H. Dubs HHS Hou Han shu HJAS Harvard journal of Asiatic studies JAOS Journal of the American Oriental Society JNCBRAS Journal of the North China Branch of the Royal Asiatic Society MH Les memoires historiques de Se-ma Ts'ien, translated by Edouard Chavannes MRDTB Memoirs of the Research Department of the Toyo Bunko MSOS Mitteilungen des Seminars fiir orientalische Sprachen SC Shih chi TP T'oung pao YCHP Yen Ching hsiieh pao (Yenching journal of Chinese studies) LIST OF RULERS DURING THE FORMER HAN DYNASTY B.C. Emperor Kao (Kao-tsu) 206-195 Emperor Hui 194-188 The Empress Dowager nee Lij 187-180 Emperor Wen 179-157 Emperor Ching 156-141 Emperor Wu 140- 87 Emperor Chao 86- 74 Emperor Hsuan 73- 49 Emperor Yuan 48- 33 Emperor Ch'eng 32- 7 Emperor Ai 6-1 A.D. Emperor P'ing 1-5 Wang Mang (Regent) 6-8 Wang Mang 9-23 Ascension to the throne usually occurred in the year preceding the official commencement of the reign. 16 I. HAN HISTORY AND SOCIETY^ Storming down from Tibet, arching into a great northward loop around the Ordos, racing south between Shensi and Shansi, the Yellow River finally cuts its way eastward through loess-covered hills to flow sluggishly across its delta, the rich north-China plain, to the sea. The lower basin of the river and the valley of its affluent, the Wei, are the arena of early Chinese history. Screened behind the Tibetan massif, which has deserts to the north and jungles to the south, China stands at the eastern end of the great Eurasian continent, her history the central history of eastern Asia. The deserts and steppes of Mongolia fix northern limits to the spread of China's economy of intensive agriculture. But southward lie no natural barriers, and century after century the Chinese pushed their way and spread their culture from the Yellow River, first to the Yangtze, then slowly southward to the sea. The southward expansion had only begun in early Han times. The political center of the Former Han empire was in the northwest, where the fertile valleys of the Wei and the Ching rivers served during most of the period as the main economic base. Blest with strong natural defenses, this region "within the pass" was populous, and rich in agricultural and grazing land. Set in the heart of an irrigated, mountain-surrounded plain, Ch'ang-an, the capital, kept its watchful eye upon the strategic eastward passes from the south bank of the Wei. Outside those passes were the regions which had been independent feudal kingdoms only a few decades before the Han dynasty commenced. Decline of Feudalism and Beginning of Empire History always begins mi medias res. The Han dynasty arose from the ruins of the Ch'in empire, which was originally only one of many states into which China was divided along feudal lines. The two centuries prior to the beginning of the Han empire saw a 1 This chapter is designed as a historical and societal framework for the rest of the book. Much that would be important from some other point of view has been omitted, and only primary historical trends and the essentials of the social system are discussed. In order that a conception of the period may be presented in broad outline and without interruption, the documentation has been placed in a long note at the end. There is no attempt to cover every statement with exact citation, but many assertions made here are more fully developed and better substantiated in later chapters. 17 18 SLAVERY IN THE FORMER HAN DYNASTY revolution in Chinese society. Shortly after 400 B.C. feudal govern- ment began to assume a regional quality, presaging empire. One evidence of this regionalism was the building of great walls which apparently attempted to define "forever" the maximum areas that could be ruled by feudal forms: walls not only between China and the slowly emerging steppe society on the north, but also interior barriers between the great states, between Wei and Ch'in, Ch'i and Ch'u, and between Ch'u and the smaller states of the middle Yellow River Valley. Within the great states a system of administration was developing which increased state power at the expense of the various vassal lords. Ch'in dominated the northwest, Ch'i lorded over a group of small states in the east, while Ch'u controlled most of the area in the south between the Yellow River and the Yangtze, and the coastal area northward to Shantung. Disunited, Chao, Wei, and a different Han, the succession states of the once powerful Chin in the north, were in danger of being rolled up from the west by Ch'in. Warfare was changing in both technique and objective. Mounted archers, trained infantry, and crossbowmen gradually replaced the clumsy feudal chariots surrounded by poorly armed serfs, and the objective was no longer to assert leadership among vassals, but to destroy defeated ruling houses and to absorb defeated states. Ch'in, the protagonist of this non-feudal way of fighting, succeeded little by little in consolidating adjacent areas, west, east, and south, into one ever-expanding empire. Among the internal reforms of Ch'in, giving it superiority over all rivals, was its change in the system of land tenure, ascribed to Shang Yang, to whom in the capacity of Chancellor was intrusted the rule of Ch'in state. In feudal China the owning of land had been almost a religious matter, and only nobles could hold it either in fee or in domain. The actual farmers were serfs working part of the land for the overlord and part for themselves. Shang Yang, who is stylized by formal Chinese history as the originator of changes going forward in many regions, is credited with permitting private ownership of land by the comm.on people, thereby shattering the foundations of feudalism and forwarding the subjugation of the nobility to an absolute monarch. By introducing direct taxation in kind, Ch'in helped to transform the actual farmers from serfs to free peasants, who no longer owed their lord stated amounts of labor, but owed only a tax to the state and perhaps rent to a land- lord. For administration and law-enforcement, families were grouped into fives and tens, the members of which were severally responsible HAN HISTORY AND SOCIETY 19 for the conduct of all the others. By direct taxation, emphasis on agriculture, and encouragement of immigration, Shang Yang en- deavored to increase the amount of surplus grain, and thus support a large professional army solely in the service of the state. Ch'in used its new military machine ruthlessly, encouraged its troops to slaughter defeated enemies, and gave bonuses according to the number of heads taken in battle. At the same time it slowly suppressed its great feudal families, and created a new but honorary aristocracy of fighters and people of wealth. As its old nobility was shorn, or as new territory was conquered, Ch'in established commanderies or prefectures whose officials were directly appointed by the Ch'in king. During the fourth and third centuries B.C., other states also were altering their political, economic, and social forms. A whole class of "political scientists," men like Shang Yang, traveled from court to court advising on statecraft and war. Ch'in seems to have had the edge on its rivals by an early start and by a more thoroughgoing economic reform which made its base more productive than any of the other great regions, with the possible exception of Ch'u. In war Ch'in had the great advantage of nearly impregnable natural defenses on the south and east, together with control of passes and head-waters leading into the territories of its rivals. The climax of the last two centuries before Han came during the reign of the First Emperor of the Ch'in dynasty, who ruled (246- 210 B.C.) first as King of Ch'in, then as the Emperor of a consolidated China. The breath-taking speed with which he mastered all the rival kingdoms is shown by the following chronology of conquest and annexation: 230 B.C., Han; 228, Chao; 225, Wei; 223, Ch'u; 222, Yen; 221, Ch'i — all of China conquered. Principally on the advice of his great minister, Li Ssu, the First Emperor crushed the old feudalism wherever he found it and established the Chinese imperial system whose main forms endured for two millenniums. He further encouraged private landowning and introduced a standard coinage. He tried to enforce a detailed and harsh criminal code, emphasizing legalism in place of the essentially feudal customary morality. He unified the forms of the written characters, and attempted to remove from circulation the ancient historical literature employed by scholars to oppose his reforms. He established an administrative system using a civil bureaucracy, and divided the country into thirty-six commanderies, laying the basis for the provincial and prefectural system of today. 20 slavery in the former han dynasty The Han Dynasty FIRST phase: founding, consolidation, and RECUPERATION The First Emperor enjoyed his empire for only a dozen years. After his death in 210 B.C. it burst apart. Rebelhon sprang up in southern Honan and spread quickly all over eastern China. It began as a mel^e of regional armies and rival generals, some of whom were members of the old aristocracy, others mere commoners who gave their local revolts the color of legitimacy by sponsoring various royal scions. In this respect the rebellion was a re-assertion of feudalism. But it was also a popular movement arising from the masses. As the rebellion progressed, the aristocracy proved inca- pable, while commoners fought their way to the top. The capture of the Ch'in capital late in 207 B.C. was the signal for a bitter struggle between Hsiang Yii, descended from famous generals of Ch'u, and Liu Chi, a former village official who turned bandit and then became a rebel leader. This struggle lasted until 202, when Liu Chi finally slew his rival, absorbed his army, and was made Emperor by his nobles and adherents. Thus the Han dynasty began. Liu Chi is known historically by his temple name, Kao-tsu, "the Eminent Founder," or by his posthumous title. Emperor Kao; and the dynasty officially begins in 206 B.C., the first year after the Ch'in dynasty fell. During nearly seven years of rebellion and civil war, rival armies had crossed and recrossed north China, conscripted troops, besieged towns, burned stores of grain, looted, and fought pitched battles. Disrupting agriculture and trade, they brought on cruel famines which in some areas cut the population in half. Emperor Kao therefore faced the colossal problem of organizing an empire, de- mobilizing and rewarding his troops, and returning his country to a productive, peace-time basis. Fortunately he had competent advisers for, though he was a good leader of troops and a shrewd politician, he had had no experience in governing. One of his first acts was to claim Shensi for himself and to establish there his capital. Other regions he awarded as kingdoms and marquisates to his best generals and to some of Hsiang Yii's leaders who had surrendered to him. This policy quickly spelled trouble, and Kao-tsu had to devote much of his reign to quelling revolts among seven kings not of the house of Liu. At his death, nine of his sons or relatives occupied kingdoms, while men from his native prefecture held practically all the important government positions. HAN HISTORY AND SOCIETY 21 Thus, while the early years of the dynasty witnessed a return to regionalism, with political administration in several important areas controlled by virtually independent nobles, it was not a return to Chou feudalism. Organization of society and developments in economics, especially land ownerships, prohibited that. In areas not controlled by kings, Kao-tsu set up commanderies and pre- fectures just as the Ch'in rulers had done. Furthermore, the new nobility was not descended from the old aristocracy, but was com- posed entirely of members of his own clan and of his followers, some of whom had risen from very humble station. During the course of the next century actual control over the whole country, including the feudal fiefs, was wrested from the nobility and lodged again with the central government. The first serious threat to Kao-tsu's imperial line came from his Empress nee Lii, a forceful and scheming woman who had con- siderably aided Kao-tsu in his conquests. She succeeded in getting her son made Heir-apparent, although he was neither the oldest nor the favorite son of the Emperor. She closely controlled Emperor Hui during his seven years on the throne, and after his death appointed a child for whom she ruled as Regent. When this child died she appointed another, asserted to be a son of Emperor Hui, but in reality from her own clan. Some members of the Lii clan had assisted in Kao-tsu's conquest and had been awarded mar- quisates. During her regency the Empress ennobled many others from her own clan, even making some of them kings; she placed one of her nephews in charge of the civil government, and another in command of the army. This conspiracy to steal the empire was crushed within six weeks after her death in 180 B.C., when Kao-tsu's relatives and old followers, supported by the army, massacred the entire Lii clan. Yet the fear that a consort family might usurp the throne haunted every ruler thereafter. Emperor Hui had no living male descendant. The Liu clan there- fore selected the oldest living son of Kao-tsu to be the next Emperor. Canonized as Emperor Wen, he gave China a long and model reign. It was under his successor. Emperor Ching, that there arose the second threat to the dynasty: the Rebellion of the Seven States, in 154 B.C. This was a contest between central control and the regional power lodged in a group of kings, all from the house of Liu. Both Emperors Wen and Ching had systematically limited the authority of their kingly relatives whenever a suitable opportunity arose. Thus, when the King of Ch'i died without a son in 164 B.C., 22 SLAVERY IN THE FORMER HAN DYNASTY Emperor Wen divided Ch'i into seven parts, placing one of the King's brothers over each district, thereby automatically weakening that important region. Emperor Ching reduced the territory of Chao, Ch'u, and Chiao-hsi, and planned to do the same with Wu, the richest and strongest of all. But the King of Wu rebelled, and was quickly joined by the Kings of Chao, Ch'u, and four divisions of the former Ch'i. The imperial armies quelled the revolt, but it cost thousands of lives, including those of the seven rebels, whose king- doms were abolished and made into commanderies. Thereafter, the imperial governm.ent methodically reduced the power of the kings by directly appointing for each noble a chancellor who actually governed the fief, and by dividing fiefs among all the sons of a deceased noble, thus finally settling for that dynasty the question of regionalism versus empire. The first seventy years of the d5masty was a period of rest and recuperation for the people, marred only by the revolts during Kao-tsu's reign, the Rebellion of the Seven States, and sporadic conflict with the Hsiung-nu. Most of the harsh, exacting laws of Ch'in were revoked. Empress Dowager nee Lli gave a good adminis- tration, and the long, frugal reign of Emperor Wen was a golden period for the common people. In 195 B.C. the basic tax on agricul- tural products was set at one part in fifteen, a reduction from the Ch'in tithe; in 167 it was entirely abolished, and in 156 it was restored at only half the earlier rate. Emperor Wen insisted on lighter corvee duties for the people; his edicts, and the memorials of such great statesmen as Chia Yi and Ch'ao Ts'o, show real solicitude for the condition of the farmer. Developments in irriga- tion brought increased yields, while an edict of 163 B.C., stating that the amount of farm land per person was greater than in ancient times, suggests that over-crowding of the land was not yet the problem it later became. Peace and reconstruction brought a general prosperity and an ever-increasing flow of taxes to the government. Ssu-m^a Ch'ien pictures economic conditions after the first six decades of peace in glowing terms, and reports treasuries and granaries bursting with unspent revenue. SECOND PHASE : EXPANSION AND DEPLETION Emperor Wu's long reign (141-87 B.C.) was a period of foreign wars and territorial expansion, but also of economic depletion and frantic experiment in government finance. Although a map of the Han Empire after Emperor Wu's wars shows nearly all of south HAN HISTORY AND SOCIETY 23 China and even part of northern Annam under Chinese control, this is deceptive. Between 135 and 110 B.C. the imperial armies did conquer the principal coastal states, and even parts of present Yunnan, by costly and sometimes ingenious military campaigns. But these wars were not the central effort, and China was not yet ready to absorb the south into its economic system. It is doubtful whether the government collected there more than enough taxes to support Chinese officials stationed in a few occupied towns. Chinese armies also conquered a fringe of southern Manchuria and northern Korea, but the colonies established there had little impor- tance in the empire's structure or internal economy. The south and northeast were the fringes; the central drama of Emperor Wu's reign was the wars against the Hsiung-nu on the north and northwest. Contemporary Chinese described the Hsiung-nu as war-like nomads inhabiting the steppes of Mongolia from Manchuria west into Chinese Turkestan. Their whole economy revolved about sheep and cattle, horses and camels, and the search for fresh pasture determined their seasonal migrations. During the Ch'in period they had been welded for the first time into a confederacy of tribes, each with its traditional grazing land, and all acknowledging the suzerainty of one ruler, the Shan-yil. Probably not numbering more than a few million, they were nevertheless a formidable enemy, masters of the powerful Asiatic bow, expert horsemen, and highly mobile. From the beginning of the dynasty, China had been subjected to frequent plundering raids by swift-moving Hsiung-nu cavalry units which penetrated deep into the frontier commanderies, looting, slaughtering, and kidnaping. During Emperor Wen's reign one such band of raiders actually came within sight of the capital. The first wars under the "Martial Emperor" were at least partly punitive, but there must have been other reasons, not all of which are under- stood. Lattimore has adduced one such cause of conflict, emphasiz- ing the wide marginal terrain supporting a people of mixed Chinese and Hsiung-nu culture who employed a mixed economy neither intensively agricultural nor entirely nomadic. The necessity of holding this border region within the Chinese political orbit, and the pull of this frontier region, led China into vastly expensive campaigns both into the steppe and into the desert. Emperor Wu was drawn by his early successes into larger and larger campaigns against the Hsiung-nu. Many of his best generals came from the frontier and were adept at border fighting. The 24 SLAVERY IN THE FORMER HAN DYNASTY Chinese armies often numbered from fifty to a hundred thousand cavalry with even larger infantry and supply columns. Equipping and provisioning such armies cost the government enormous sums, not to speak of hidden costs in forced labor for transporting provisions and military supplies. Rewards to victorious troops mounted into hundreds of millions of cash. The government furthermore fed and clothed surrendered Huns, on one occasion numbering 40,000, till they could be established in new colonies. Chinese armies wrested from the Hsiung-nu all the area of what is now southern Suiyuan, eastern Kansu, and Ningsia, within and adjacent to the great northward bend of the Yellow River, and also the lengthy Kansu corridor leading out to the "Western Regions" of Chinese Turkestan. This territory, though not agriculturally important, became a strategic part of the empire, populated, it is said, by more than 700,000 Chinese colonists who were moved there at government expense and mixed with surrendered allies of the Hsiung-nu. Between 129 and 119 B.C. China crushed the armies of the Shan-yil in the east, the west, and the center, but conflict still flared up sporadically until 51 B.C. By controlling the Kansu corridor the imperial armies were able to penetrate the Tarim basin, subjugating and loosely attaching various oasis kingdoms to the empire, till China dominated all of Chinese Turkestan politically. Chinese leaders became aware of other nations and other cultures to the west, of India, of Persia, and dimly of the Roman Orient. This warfare drained the abundant treasury with which the "Martial Emperor" had begun his reign. His advisers were fertile with schemes to replenish the coffers. One attempted method was the creation of eleven ranks of new nobility based on war chest donations ranging from 170,000 to 370,000 cash — the higher the donation the higher the rank. The central government exacted large contributions from the old nobility, and levied property taxes on merchants and speculators. Those who failed to report their total wealth had their fortunes confiscated, while informers were encouraged by the promise of one-half the confiscated property. By this device so many people were impoverished that the govern- ment, in addition to getting large amounts of cash, was temporarily embarrassed with less negotiable things, such as land, houses, and slaves. Several novel and sub-standard issues debased the monetary standards of the day, and then, because of a wave of counterfeiting, the central government instituted a monopoly on the coinage of money. Probably the greatest revenue-producing agency, however, HAN HISTORY AND SOCIETY 25 was the government monopoly of those indispensable commodities, salt and iron, reinstituted in 119 B.C. after the pattern of the Ch'in dynasty. The reign of Emperor Wu was a period also of intensive canal building, an activity designed primarily to improve transport of tax grain from regions not adequately tapped. A secondary purpose, but one which became increasingly important as mounting expenses cancelled the gains of better transport facilities, was to increase production in those areas from which the central government could economically collect its thirtieth part of the harvests. The greatest canal-digging enterprise was naturally in Shensi and nearby areas, from which the government could profit immediately. Cutting a canal from Ch'ang-an south of the Wei to meet the Yellow River, for example, greatly reduced the time and difficulty of that stage in the transit up the Wei which brought grain from Honan, Shansi, and regions farther east; but it also irrigated thousands of acres close to the capital. Other canals were dug, with increased produc- tion the primary aim. Furthermore, reorganization of the grain transport from the lower Yellow River basin quadrupled the yield from that area. Not till this improvement had made eastern China directly important to the capital did Emperor Wu tackle a twenty- year-old problem of flood control at Ku-tzu on the lower Yellow River. As the tempo of foreign wars decreased, the various financial schemes and the increasing flow of tax grain began to refill the treasury. Thus, by the closing years of Emperor Wu's long reign, China's political boundaries were roughly staked out. The central government firmly commanded the territory between the Yangtze and the Great Wall, and it controlled by military power other areas which only later were fully absorbed into the Chinese social and economic pattern; but the people were economically exhausted. THIRD phase: gradual economic decline The century from 87 B.C., which covers the reigns of Emperors Chao, Hsiian, Yiian, Ch'eng, Ai, and P'ing, is not easy to generalize. A second period of recuperation, with only sporadic foreign wars, it also developed into an age of great luxury for the upper classes and increasing poverty for the masses, of corruption among rulers and officialdom, and of the labor-pains of peasant revolt. Young Emperor Chao was supported by wise ministers, who noted and tried to ameliorate the suffering of the masses; during his reign the 26 SLAVERY IN THE FORMER HAN DYNASTY country began to recover from the costs of previous wars. Emperor Hsiian, who had grown up among the people, gave an enlightened rule. His reign was perhaps the zenith of the dynasty, the high point of Chinese diplomatic success, and a period of general economic stability. Early in the reign of his son. Emperor Yiian, the economic balance seems to have turned. This ruler and some of his ministers tried to halt the mounting expenses of government, to reduce the tax-fed bureaucracy, and to lighten the burdens on the people. r But there was a fundamental maladjustment between population / and land, which the economic system could not solve. Steady / growth in population began to crowd the land, offsetting all gains through reclamation and irrigation, opening of state lands to the poor, slow peopling of new marginal territory, and improvements in \ farming techniques. Theoretically Chinese farmers had an unlimited \ frontier southward, but practically this frontier was limited by the need for extensive drainage works, and government projects are not reported south of the Huai River Valley. Farming techniques were still far below what they later became. The use of iron tools and plowing with animal power were apparently not very widespread. As late as about 90 B.C. the government attempted to increase production by teaching selected local officials and outstanding farmers the methods of crop rotation, and by giving them sample tools designed separately for plowing, planting, and harvesting. Spread of this knowledge was apparently slow except near the capital and in northern state-supervised agricultural colonies. Furthermore, an important part of the best-irrigated land passed gradually into I the hands of nobles, officials, merchants, and gentry, who owned it for investment and rented it to share-cropping tenants. It was the landlords who benefited most by the low tax on agricultural produce. The "average" peasant, working a smaller plot than his ancestor early in the Han period, or paying half his produce to the town- dwelling landlord, had small reserves against endemic north-China famine. Successive droughts or sudden floods sent swarms of refugees trooping along the highways in search of food and work in unaffected regions. Concurrently the tax-fed bureaucracy grew in numbers, while the nobility, higher officials, and merchants vied with one another in luxury. This contrast was not the event of a single decade but a trend, already noted by ministers of Emperor Wu, which grew sharper and sharper as the dynasty reached its decline. During the weak reign of Emperor Ch'eng, four popular revolts, each starting with a few HAN HISTORY AND SOCIETY 27 scores of desperadoes, rapidly spread to dangerous proportions before they were quelled. They were the storm signals of disaster. A few farsighted ministers tried to persuade Emperors Ch'eng and Ai to enforce sumptuary laws and break up the landed estates. Where only Draconian measures could have sufficed, Emperor Ch'eng relied upon exhortations to his officials, while the ordinances drawn up for Emperor Ai were prevented from becoming law by the objec- tions of his affinal relatives and his favorite, Tung Hsien. The 'dynasty had reached that stage when the officials and the great landlords were indivisible, when the personal stake of those who administered the government and enjoyed its bounties compelled them to resist any reform that touched the substructure of their wealth. Tung Hsien is a singular example of that inner circle of imperial favorites and affinal relatives who strove to pile up family fortunes during a precarious heyday of power. During the brief reign of Emperor Ai, Tung Hsien rose from a mere court attendant to the rank of marquis, and held the office of Commander-in-Chief of the armed forces. After Tung Hsien was forced to commit suicide on the death of his imperial paramour, his property was confiscated and sold by the government for four billion three hundred million cash. In about five years he had acquired this fabulous wealth out of imperial grants, presents and bribes from officials, and the salaries and perquisites of office, all compounded by incomes from his investments and estates. FOURTH phase: LAST MINUTE REFORM, AND COLLAPSE The threat of consort families, first raised against the dynasty by relatives of the Empress nee Lii, was finally fulfilled by Wang Mang. His father's half-sister was the Empress of Emperor Yiian and mother of Emperor Ch'eng. The Wang family rose to great political heights during the reign of Emperor Ch'eng, but was eclipsed during the reign of Emperor Ai. When the latter died with- out an heir in 1 B.C., the Grand Empress Dowager nee Wang re- asserted her family's power, and called the most competent man of the family, Wang Mang, to take charge of the government. By skilful manipulation he continually augmented his power, first as Regent, then as "Acting Emperor," until he was strong enough to seize the throne in a.d. 9. Wang Mang is perhaps the most controversial figure in Chinese history. Some modern writers consider him a farsighted reformer, while the traditional view excoriates him as a political scoundrel 28 SLAVERY IN THE FORMER HAN DYNASTY who lusted only for personal power. No man is so entirely lofty or so completely base as Wang Mang is variously portrayed. He unquestionably committed a great political crime by overthrowing an imperial house impermanently. His "reforms" failed, and his dynasty was swept away by popular rebellion. His greatest misfor- tune, coloring all historical judgment, was to have the succeeding dynasty established by a member of the old Liu house, making it a continuation of the Han dynasty. Although his biographer quotes archival materials both favorable and damning, they are interlarded with caustic comments which unconsciously prejudice our judgment. Publicly writing a story of the dynasty Wang Mang overthrew, and writing it during the continuation of that dynasty, the historian could not possibly have given Wang Mang a favorable treatment even had he so desired. But judgment of Wang Mang's political crime must be made, not from the viewpoint of those who supplanted him, but by considering the quality of the last weak and infamous rulers whom he supplanted. Likewise, his "reforms" cannot be fairly judged by how wise or foolish, how humanitarian or avaricious, they now appear. They must be judged against the background of economic and social conditions of his own day. These conditions he neither created nor successfully changed, although some of his reforms temporarily dug at the roots of the problems. Only the holocaust of his downfall, which greatly reduced the population and broke up the landed estates, ameliorated the economic unbalance and gave the house of Han another two centuries of grace. When Wang Mang assumed the throne in a.d. 9, he issued a remarkable imperial order (translated in document 122). The introduction explains the reasons for instituting reform. He described how the Ch'in dynasty had made possible the accumulation of both land and offices in the hands of the same people, on the one hand by overtaxing the peasantry, on the other by abolishing the ancient communal land system. Not only did greed for wealth arise from this, but also the strong annexed the fields of the poor to such an extent that they did not even have enough land upon which they "could stand an awl." He also described the evils of slavery, and quoted the Book of History to prove that in antiquity only the government possessed slaves, and they were criminals. Admitting that the house of Han had reduced the tax on produce to one-third the Ch'in tithe, he called attention to the corvee service and extra poll-taxes levied on old and weak alike. Worst of all was the "usurious" rental charge, whereby tenants paid five parts in ten HAN HISTORY AND SOCIETY 29 when the tax on produce was only one part in thirty. The reforms he proposed were these: All fields were to be nationalized and there- after called "the King's fields"; private slaves were to be called "private retainers." Neither fields nor slaves could be bought or sold. Furthermore, land was to be equitably distributed. Each family comprising less than eight males but owning more than 102 acres was to distribute its excess land among its clansmen and peighbors. Conversely, those without land were to receive it according to the regulations. Anyone who dared to oppose the new law would be banished to the frontier. This was fundamental but it was not entirely novel, for elements of the reform had been openly advocated by Confucianists in the reigns of Emperors Wu and Ai. Introduced by executive decree, in an age when free ownership and free sale of land had long been legal, such a sweeping reformation could never have been enforced. To efi'ect it anywhere would require a bloody revolution. From the nobility and the highest officials down to the commoners, innumera- ble people were punished for refusing to obey, but the plan came to nothing in the end. Wang Mang was persuaded to repeal the law in three years. In the shadow of this failure all the other measures were mere tinkering. The alteration and debasing of the coinage enriched the treasury but caused great hardship. The "six state controls" included several well-tried monopolies and new types of taxation. The "controls" also attempted to stabilize prices of basic commodities by a novel method, and made loans available to the poor without interest or at low rates. These acts were primarily designed to raise revenue, but some of them did benefit the poor. Even Wang Mang's defamers admit that he worked tirelessly, but he could not enforce his grandiose schemes, which, furthermore, were continually per- verted by his officials. Every new law created a host of enemies. Expensive wars and famines produced a rebellion whose seeds had been sprouting long before Wang Mang entered public life. A rebellious movement called the Red Eyebrows first became active in Shantung about A.D. 18. A little later two pretenders from the house of Liu organized regional revolts. Strife followed the classic pattern. China fell apart into constituent regions all at war with the central authority and each ultimately opposing the other. By A.D. 22 all of eastern and southern China was lost to Wang Mang. Finally one rebelling army captured Shensi, invested the capital, and Wang Mang was killed in October, A.D. 23. 30 SLAVERY IN THE FORMER HAN DYNASTY One of the rival pretenders, who had earher been made Emperor, fixed his capital at Lo-yang, then in the next year moved it to Ch'ang- an. His two years were anarchy. The Red Eyebrows devastated eastern China, and several rival emperors established themselves elsewhere. Then the Red Eyebrows, who had become the worst sort of plunderers, captured Ch'ang-an and took the first pretender captive. It was their pillage and burning that destroyed the old capital beyond repair. The city is said to have burned for three months. It is hard to understand how any of the imperial archives survived, although it is said that some hundreds of cartloads of books were later taken to the new capital at Lo-yang. The ultimate victor among all the rivals was Liu Hsiu, who is lauded for his humanitarian reign and brilliant victories by the posthumous title "Emperor Kuang-wu." Although he accepted the throne in A.D. 25, he still had years of warfare ahead, both against the Red Eyebrows and against rival political areas in China. Not till A.D. 36, when he conquered Szechwan, did he really control all of China, a country that had suffered nearly two decades of brutal civil war. The census of A.D. 2 fists the population at 59,594,978. This was two decades before the death of Wang Mang. The earliest Latter Han census reported 21,007,820. This was in A.D. 57, almost thirty-five years after the death of Wang Mang, and following two decades of internal peace. Figures for the Latter Han period rose to more than 49,000,000 in A.D, 140 but never achieved the Former Han total. While Chinese census figures are always inaccu- rate in detail, this tremendous drop in the half-century between A.D. 2 and 57 eloquently bespeaks the leeching which relieved the fever if it did not cure the malady of China's ancient agrarian economy. Distribution of Population, and Urbanization What was the structure of Chinese government and society during the Former Han dynasty, and in what manner did people live? The census which recorded a population just under sixty millions shows a distribution very different from that of modern times. The basin of the Yellow River stands out darkly on a map of population density. A great black nebula centers between Loyang and Kaifeng, blanketing eastern Honan, southern Hopei, and western Shantung. West of this area there is a heavy concentration around the capital, near modern Sian, in the irrigated region between the Ching and Wei rivers. Northeast and southeast of the nebula HAN HISTORY AND SOCIETY 31 the distribution is fan-like, covering the northern coastal plain. South of the Yangtze is almost a vacuum, and so too are the northern and western frontiers. Most of the people were peasant farmers, settled in tiny hamlets and concentrated where wheat, kaoliang and millet could be grown most abundantly. These regions were southern Shensi in the valleys of the Lo, the Ching, and the Wei, and the Honan-Shantung segment of the Yellow River, bordered by southern Shansi and Hopei, and northern Kiangsu and Anhwei. Because Han dynastic history is concerned primarily with the metropolitan upper classes and locates occurrences with reference to administrative centers, it suggests a degree of urbanization which probably did not exist. With a low level of industrial development and a primarily agricul- tural economy, most "cities" were merely walled towns, focal points for administration, grain-storage, garrisons, services, handicrafts, and market trade, and probably did not have more than ten or twenty thousand residents. A few important cities there were. Lin-tzu, former capital of Ch'i, in the heart of silk-producing Shantung, bustled with processors and traders of silk. Yen, near Peking, was the emporium for trade in the north. Han-tan in modern Hopei, once capital of the state of Chao, and Yiian near modern Nan-yang in Honan, were centers of iron-smelting. The capitals of the commanderies of Pa and Shu, corresponding to Chungking and Chengtu in Szechwan, gathered the trade of the southwest, while Lo-yang maintained a certain importance as an old cultural center. All roads led, like the ribs of a fan, to the capital district, with a population of nearly two and a half millions. But the capital district was not a city; it was a congregation of fifty-seven towns, each with its tributary agricultural terrain, which together composed three commanderies. Ch'ang-an itself claimed a population less than a quarter of a million, and where individual figures exist no other Han metropolis (including surrounding farm land) is credited with more than three hundred thousand souls. Local and Central Administration The basic geographic unit of government was the prefecture or hsien, an area of agricultural land, villages, and towns, whose borders were seldom more than two days' walk from a central walled "city." The lowest extension of empire and the widest expansion of Chinese familism was the point where prefectural government, 32 SLAVERY IN THE FORMER HAN DYNASTY in the hands of a prefect or chief and his staff, met local government, typified by the San-lao, or "Thrice Venerable." Hamlets were often composed of single clans or groups of families, each with its clan or family elder. According to a decimal scheme, perhaps more theoretic than real, ten hamlets made up a commune, or t'ing, with a chief in charge of civil and military affairs, and ten fing composed a district (hsiang). Early in his reign the first Han Emperor, who himself had been chief of a t'ing, adopted the system of San-lao, whereby venerable men of cultivated personality, "able to lead the masses and do good," were selected, one in each district. Among the district San-lao one was chosen to serve as consultant for the Prefect and his staff. Somewhat equivalent to prefectures in function, but more military in nature, were the tao, or border marches, primarily inhabited by barbarians. Many prefectures were designated as marquisates or as estate-cities of female nobles, but they were apparently not governed directly by the nobles. Prefectures, border marches, marquisates, and estate-cities were grouped into commanderies or kingdoms that roughly corresponded in function to modern Chinese provinces, though they were smaller and much less populous. Each commandery was under an adminis- trator in charge of civil affairs, and a military governor. Kingdoms were actually independent of the central government early in the Han period. As the dynasty progressed, however, the kings, all of the house of Liu, were slowly shorn of political power, and their royal bureaucracies were controlled by chancellors appointed by the central government and approximating commandery adminis- trators in function and power. Toward the end of the dynasty, groups of commanderies and kingdoms were loosely joined together into thirteen chou, or provinces, but these were not administrative divisions so much as circuits under inspectors who reported to the central government on the administration and activities of officials. The "Treatise on Geography" in the Ch'ien Han shu gives popula- tion figures for each commandery and kingdom during the reign of Emperor P'ing, and lists all the constituent divisions by name. At the end of the dynasty there were 103 commanderies and kingdoms, 83 and 20 of each. They were made up of 1,314 prefectures and estate-cities, 32 marches, and 241 marquisates. The government of the empire centered in the capital district around the Emperor at Ch'ang-an. There resided the Chancellor, in charge of civil government, the Grand Marshal and Commander- HAN HISTORY AND SOCIETY 33 in-Chief, in charge of military affairs, the Grandee Secretary, who acted as a check upon the Chancellor, and many other high officials. There, too, were the many departments and bureaus that administered state finances, justice, affairs of the nobility and dependent or foreign states, agriculture and commerce, and public lands. There, also, was that part of the government more directlj'- connected with the imperial household, the dynasty as distinguished from the empire. Officials of various departments administered the affairs of the privy purse, the palaces, and the ancestral ceremonies, and cared for the needs of the imperial family. The capital district was a special administrative area, organized by commanderies and prefectures, and having officials and bureaus that were a part of the central government. By the time of Emperor Ai, the bureaucracy numbered some 130,285 officials, counting from junior clerks up through the Chancellor. Class System commoners In Han times, just as recently, the Chinese "large family," rather than the individual or the marriage group, was the basic unit of society from the Emperor down to the bottom of the social scale. The structure of society was already complex in Former Han times, but the class system was fairly simple as referred to by con- temporary writers. The mass of the people was called "commoners" ishu-jen), or "the people" {min), being thus differentiated from two higher classes, the officials and the nobility, and two lower groups, convicts and slaves. Commoners included all the peasantry, as well as artisans, shop- keepers, merchants and the like. They were the governed, who supported the state and enriched the upper classes by their taxes, their corvee labor, and their industry. Taxes of the common people varied greatly during the two and a quarter centuries of the Former Han epoch. They depended also upon occupation and place of residence — whether, for example, on the frontier, in a marquisate, or near the capital. The average adult commoner during most of the dynasty paid a poll-tax of 120 cash annually to the state and 63 cash for the uses of the imperial family, while children were charged 20 and 3 cash per year, respectively. Households in mar- quisates and kingdoms paid 200 cash to the overlord, but were relieved of the poll-tax and the tax on children. Farmers paid one- 34 SLAVERY IN THE FORMER HAN DYNASTY thirtieth of their crops; merchants paid double poll-taxes, heavy market dues, and assessments on their capital; artisans paid income taxes. Furthermore, common people were subject to four kinds of forced labor or corvee duty. The most important was labor for the prefectural government on roads, canals, embankments, buildings, and tombs, one month a year. Military duty, frontier duty, and service at the capital were usually commuted by payment of money for a permanent force. These taxes and services were a heavy burden, and supplied considerably more revenue than the govern- ment normally required. This made it possible for the Emperor, in his magnanimity, frequently to pass up one or another of the taxes to win the people's gratitude, and also to hand out gifts to widows, the aged, and the poor. Such devices were only a palliative, and thoughtful statesmen regularly urged reduction of the burdens on the people. Within the broad status group of commoners there were, of course, different economic and professional groups. Among these, merchants and the landed gentry were most important. Because of the pecul- iarly cellular structure of China's political and economic geography, in which each district or prefecture contained a walled city sur- rounded by tributary agricultural terrain, each prefecture fed itself and produced in its own trade center most of the goods needed for everyday life. Unification of north China in a single empire, abolition of most trade barriers, and the increasing use of coined money did, however, produce in Han times an inter-regional trade in certain essentials not everywhere produced, particularly salt and iron, and a luxury trade in fine silks, furs, bronze and gold, jade and pearls, lacquer, bamboo, exotic foods and spices, herbs and medi- cines. Besides merchandising, many fortunes of the Han period were based on mining, smelting, grain-dealing, and money-lending. The chief evidence of the growth of commerce is contained in memorials of statesmen deploring the fact, and in laws penalizing and disfranchising merchants. The prevalent economic philosophy held merchants to be the enemies of farmers, and trade the natural rival of agriculture. Popular metaphor described farming as the root or trunk, and manufacturing and trade as the end branches of the national economy. Emphasis on agriculture maintained what was fundamental, while trade drew people away from the basic pursuit. Merchants imposed on farmers, fleecing them of their produce at low prices and selling them finished goods at exorbitant rates; they lived luxuriously in towns and cities, while farmers HAN HISTORY AND SOCIETY 35 suffered privation and want on their land. This philosophy squared with the fact that a money economy, though not far advanced in Han times, was apparently upsetting the rural economy based on grain, and with the fact that many merchants also dealt in grain. They handled the farmer's surpluses in good years and advanced seed grain and food in bad. Organized, sophisticated, and in touch with crop conditions and prices elsewhere, local dealers had a tre- mendous advantage over the peasants, and there is little reason to doubt their depressing effect on rural economy, which many Han memorialists described. The government legislated against mer- chants and encouraged the agricultural class by pious edicts and fre- quent remission of taxes, because the financial structure of the state was built primarily upon taxes levied on produce. Furthermore, the policy-making bureaucracy came from the landed gentry, and the nobility acquired its principal revenue from taxes paid by the peasants. Thus, any group adversely affecting the taxpaying potential of the farmer imperiled both the state and the ruling class. Merchants were therefore heavily taxed, denied in theory the right to hold office, and above all forbidden to invest in farm land. In- directly, the government attacked commercial profits by monopoliz- ing salt and iron, experimenting with price-stabilizing granaries, and lending seed and food. The prohibition against merchants owning farm land, though never entirely effective, was crucial because there were no other important forms of capital investment. Land was the property to own. It brought security, influence, and regular income from tenants who paid — what seems to have been considered excessive — a half share of the harvests. Produce taxes fell lightly on landlords in comparison with their income. Money taxes, so heavy for poor peasants, were minuscule for the well-to-do, and anyone who could afford to pay the fee escaped corvee duty. Landed gentry and local officials naturally clung together. Officials were recruited partly from the landed gentry and when sent to the provinces to govern they found in the local landlords people of their own kind. The most important and most active line of cleavage in Han social structure was not that between the free and the slave, or between the nobleman and the commoner, but that between landlords, officials, and the nobility on the one hand, and small farmers, tenants, and laborers on the other. THE BUREAUCRACY Most government positions required a fair knowledge of the written language, and the better civil positions demanded a quoting 36 SLAVERY IN THE FORMER HAN DYNASTY familiarity with the classics. To learn the written language and the contents of ancient books required leisure and security, and these were made possible by landholding. Several roads led to the lower and middling offices. People of wealth other than merchants could buy the honorary ranks, which made them eligible for selection. High officials were privileged to propose candidates, and they naturally promoted their relatives or fellow townsmen. Scholars and men of special ability might recommend themselves by some clever scheme for government, or be recommended by commandery or prefectural officials. The government was constantly searching for local worthies and students, astute interpreters of omens and portents, military strategists and border fighters, or loyal underlings who smelled out plots against the throne. For the top positions only two things counted: special distinction in administration, which usually meant a long, slow climb, or affinal relationship to the imperial house. Salaries of important officials were generous. Men of the highest brackets received 9,000 cash and 72 hu (40 bushels) of grain monthly. From that figure the scale of payments descended by sixteen stages to junior clerks, who received only 8 hu (4.5 bushels) monthly. Higher officials and their families had special benefits aside from salaries, perquisites, and opportunities for graft : the right to introduce relatives for official position, and freedom from certain taxes. Families of some officials were exempted from implication in crimes committed by the official, and others could appeal directly to the Emperor for leniency, a favor otherwise reserved to members of the imperial clan. THE NOBILITY The aristocracy consisted of a titular nobility of eighteen ranks, and an enfeoffed nobility of several types. The first spanned the commoner and official classes. Lower honorary ranks were handed out rather liberally to officials and deserving plebeians, and all could be purchased. Imperial edicts celebrating enthronement, appoint- ment of an heir-apparent, selection of an empress, or some auspicious omen frequently also announced general grants of the first grade to heads of families or eldest sons. The principal advantage accorded those of the lower grades was reduction in sentence for crime; men in the ten higher brackets were exempt from taxes and corvee duty, and occasionally received money grants from the Emperor. This titular aristocracy, though appearing to be all-inclusive, probably comprised collateral descendants of noblemen, scholars, landed HAN HISTORY AND SOCIETY 37 gentry, local worthies, lesser officials, and people of some financial pretensions. It was, in other words, the upper middle class surround- ing the throne and the enfeoffed nobility. The enfeoffed nobility came from three sources: members of the imperial family by blood, the most valorous warriors and highest officials, and members of consort families or imperial favorites. Normally one son of each emperor was selected heir-apparent and the rest were appointed vassal kings. Ranking first in the nobility, they lived in their kingdoms away from the capital, and, during the early part of the dynasty, were rulers in fact, holding royal court, collecting taxes, controlling independent armies and civil administration. After the radical changes instituted by Emperor Wu, kings no longer ruled, and they only received the taxes from their fiefs; still, they were very wealthy, for they continued to own royal lands that brought rich income. Early in the dynasty, when a king died one son inherited the kingdom while others received separate marquisates. Emperor Wen was the first to divide kingdoms among a number of heirs (in this case brothers), each of whom became a king. Later still, title to the kingdom was given to only one son, while others received marquisates carved from the kingdom. The appointment of "marquises who were sons of kings" paralleled that of vassal kings. Only one son of such a marquis inherited the title; others continued on the books of the imperial clan, and had special privileges, but had to make their own living as landlords or business men. Vassal kings and marquises who were sons of kings all had the surname Liu. Daughters of emperors became princesses, receiving estate-cities and mansions near the capital. The title was not hereditary. Princesses generally married marquises of other surnames than Liu and their daughters could marry emperors. The second branch of the enfeoffed nobility came from generals who distinguished themselves in foreign or domestic wars, men who quelled revolts or uncovered plots against the throne, and enemy leaders who surrendered to China. At the beginning of the dynasty Kao-tsu and his followers had sworn a solemn oath that none but members of the imperial family would be made kings, and only men of valor would be given marquisates. The early marquises were men who helped Kao-tsu conquer the empire, and throughout the dynasty a good proportion of all marquises not sons of kings or members of consort families were military men. On the death of the title-holder, the marquisate was normally inherited by the 38 SLAVERY IN THE FORMER HAN DYNASTY oldest son and thus the noble line continued, though the estate was reduced by an inheritance tax. Originally the system of sending away from the capital all the male relatives of the Emperor seems to have been to hold important regions loyal to the dynasty, to compensate disappointed candidates for the throne, and to prevent their intriguing at court. The net result, after the vassal kings and their sons had been shorn of adminis- trative power, was that consanguineal relatives of the Emperor, his uncles, brothers, cousins, and nephews, had no part in ruling. The men of the Liu clan being disqualified, most of the high positions at court went to the Emperor's male relatives by marriage. It was men of the consort families who mostly comprised the third branch of the enfeoffed nobility. This phenomenon is one of the most interesting aspects of Han social structure, and it had far-reaching effects on the djmasty. CONSORT FAMILIES The ways in which various consort families rose to power differed in detail but appear to have followed a discernible pattern. The Emperor had in his palace many ladies of high rank and innumerable women of lowly status, as did the Heir-apparent also. Ladies were pushed into the seraglio by their powerful families, but lesser women were chosen from all parts of the empire in beauty contests. Some happened to be noticed and summoned by emperors on their travels. As various women bore children, especially sons, their male relatives were rewarded by positions at court. One son had to be chosen as Heir-apparent; it might be the oldest, the brightest, the Emperor's favorite child, or the son of the woman with the most powerful connections. When a boy became Heir-apparent his mother almost automatically became Empress, and then her family started its climb to power. The first step was usually appointment of her father, and fre- quently her brothers, to marquisates, often kuan-nei marquisates, which were of low grade but had the advantage of estates located near the capital. The Empress's close relatives slowly acquired important positions and succeeded in placing their clansmen in the bureaucracy. As sons of the Emperor by other women grew up they were sent off to kingdoms, which meant that their maternal relatives lost their chief access to court. Not every boy chosen Heir-apparent actually achieved the throne, however; in several cases a new selection was made, and the first boy deposed. Five emperors died HAN HISTORY AND SOCIETY 39 without living sons, so that collateral descendants had to be selected, giving great opportunity for intrigue by consort families. When an Emperor died, the Heir-apparent ascended the throne, his mother was made Empress Dowager, and the new Emperor's maternal grandfather and some of his uncles or cousins were often granted full marquisates and high positions in the government. If the ne.w Emperor was young, his mother and uncles had control over him, and, through him, of the government. If he was already mature, with several wives and sons, the contest for selection of the next Heir-apparent began once more. Several consort families related to the Dowager Empress, the actual Empress, and the wives of the Heir-apparent, might simultaneously hold high positions and in- trigue to perpetuate their influence. Empresses were often much younger than their spouses, and several even outlived their sons, thus helping to protect the family power. For example, Emperor Yiian's Empress nee Wang entered the court in 54 B.C., bore the Heir- apparent's first son (later Emperor Ch'eng) in 51, became Empress in 49, lived through the reigns of Emperors Ch'eng, Ai, and P'ing, and died in a.d. 13 at the age of eighty-four. Struggles of the consort families darken the inner political history of the dynasty, especially after the reign of Emperor Wen. Yet the system had its good points along with the bad. It constantly brought vigorous new blood into the nobility, and talent to high administrative positions. It prevented any one consort family from gaining enduring or exclusive control over the palace. The families currently in power had their whole stake in supporting the ruler. While for personal advantage they sought to manipulate him through his grandmother, his mother, or his consort, they had to maintain him in power. Few consort families attempted to over- turn the dynasty; it was members of the Liu clan who most often revolted, hoping to seize the throne. On the other hand, the system led to bitter palace intrigues, sometimes culminating in the assassination of imperial sons or favored ladies. Women wielded great power at court and their rivalries embittered the palace. Some imperial sons were purposely debauched to make them pliable weaklings. The worst evil arose from the fact that most consort families could expect no more than two or three generations of power. Therefore they hastily amassed great fortunes and invested in farm land, hoping to give their descend- ants security — a vain hope, for many a family was tricked into crime, or accused of lese majesty and stripped of titles and holdings. 40 SLAVERY IN THE FORMER HAN DYNASTY Among great names resounding through the dynastic history or echoing among the documents on slavery are many noble relatives of empresses or imperial concubines. Here are only a few: Lii Lu and Lii Ch'an, Tou Kuang-kuo and Tou Ying, T'ien Fen, Li Kuang-li and Li Yen-nien, Wei Ch'ing, Ho Ch'ii-ping and Ho Kuang, Shang- kuan Chieh and Shang-kuan An, Wang Shang, Shih Tan, Chang Fang, Wang Feng and Wang Yin and Wang Shang and Wang Ken, and, greatest of all, Wang Mang. The catalogue of noblemen in the Han history classes together those from consort families and high official position, and places military men and their like in a section by themselves. Frequently, however, the distinction between these three paths to enfeoffment is dubious, for affinal relatives often proved to be great generals or excellent officials before they became noblemen. Conversely, im- portant military or civil officials frequently succeeded in placing women from their families in the imperial seraglio and thus became relatives of the house of Liu by marriage. Sizes of estates given to marquises who were sons of kings are not recorded in terms of households. Marquisates of the other two sorts varied greatly in value. Some possessed less than a hundred households, while the greatest numbered twenty thousand. A sam- pling of one in ten shows 2,600 to be an average. From each household the marquis collected a fee of 200 cash annually. In terms of purchasing power, noblemen and high officials received princely incomes from their estates, investments, salaries, perquisites, and imperial grants and graft. Vying with each other to imitate the life of the palace, they sometimes had princely mansions with private parks, slaves and servants dressed in fancy silks, and many ladies in their concubines' quarters. Singers and dancers, acrobats and musicians entertained at their banquets. They watched cock-fights and bear-baiting, and raced dogs and horses for sport. Riding out in their handsome carriages, they were escorted by mounted retainers who officiously cleared the highways. Some noblemen patronized scholarship and the arts, while many high officials were men of distinctive culture and learning. China possessed a rich and complex culture, and the upper classes doubtless matched their Western contemporaries in luxury and sophistication. THE LOWEST SOCIAL CLASSES In the recognized divisions of Han society people of plebeian status were by far the most numerous. Above them were the HAN HISTORY AND SOCIETY 41 officials, numbering a hundred-odd thousand, and the real nobility, never more than a few hundreds, or a few thousands if their families be counted. Below the commoners were two large groups, convicts and slaves. Though the Han law code no longer exists, and was indeed lost by the sixth century, some of it has been reconstructed from edicts, recorded cases, citations in commentaries, and later codes based upon the great code of the Former Han period. Apparently it was very detailed. Toward the end of the first century of our era more than six hundred listed crimes involved the death sentence, and there was an almost infinite number of ways people could be sentenced to criminal servitude. The Chou and Ch'in punishments of bodily mutilation were theoretically abolished early in the Han period, and thereafter convicts served sentences varying in length from one to five years. Clad in felons' dress, often shackled and with shaven heads, sometimes even tattooed on the face, convicts worked out their terms in frontier guard duty and in building the Great Wall, transporting army provisions, constructing imperial mausoleums, toiling in state mines or government iron bureaus, and in many other ways. Convicts, during sentence, were much like some govern- ment slaves, but after the term was completed they were freed and became commoners. The number of convicts controlled by the government at any one time cannot be estimated, but there were always myriads, and on occasions upward of a hundred thousand. Slaves, both government and private, formed a distinct and recognized class in society; they are the principal subject of this study. Fluidity of Social Position Describing Chinese society according to the components recog- nized in Han times perhaps creates the impression of rigid stratifica- tion. This is the opposite of the facts. Not only was the commoner class so broad as to include most of the population, but also fluidity of social position, the negation of a caste system, was a prime charac- teristic of the times. Individuals and their whole families rose meteorically from the lowest rank to the highest, while others tumbled as precipitously to the bottom of the social scale. The founding of the dynasty was itself a great upheaval which brought commoners and even convicts to the top and submerged finally the remnants of the feudal aristocracy. Liu Chi, the founder of the Han House, was an uneducated peasant who became a village official and then turned bandit; his principal followers, who later 42 SLAVERY IN THE FORMER HAN DYNASTY became marquises, were nearly all commoners, many from his native district. During the whole dynasty, but particularly during the reign of Emperor Wu, men of ability rose from the lowliest origins to high official positions and entered the nobility. Many of the biographies of eminent people go back no farther than one generation to trace the family history or point out an illustrious ancestor. This is the more significant in a land where the cult of ancestors was of cardinal importance. The following biographical vignettes illustrate the ways in which people of lowly background rose without hindrance to the highest positions in government. Chi Yen, a descendant of the old aris- tocracy, was only an outrider for the Heir-apparent in the reign of Emperor Ching. He climbed through the ranks of the bureaucracy to be Administrator of part of the Capital District. He was a great champion of the common people, and Emperor Wu's most fearless critic. . . . Kung-sun Hung was a swineherd who first studied the classics at the age of forty, took top honors among a group of scholars examined by Emperor Wu, and rose because of his pliability to be Grandee Secretary, then Chancellor and a marquis. ... A vaga- bond, Chu-fu Yen, attracted the attention of General Wei Ch'ing about 134 B.C., became a Palace Grandee and was appointed Chan- cellor of Ch'i to keep his eye on the King. He successfully ex- pounded to Emperor Wu the clever scheme for weakening kingdoms by dividing them among all the royal sons. . . . Jen An was an orphan whose first positions were assistant thief -catcher, chief of a commune, and then San-lao. Entering the service of General Wei Ch'ing as a squire he met another squire, T'ien Jen. Both were poor and could not afford to buy the necessary paraphernalia for introduction to court, so the general grudgingly staked them and they distinguished themselves in the imperial audience. Jen An rose to be Inspector of a province, while T'ien Jen, because of his fearless denunciation of incompetent officials, became Assistant to the Chancellor. . . . Pu Shih, a shepherd, became successful enough to contribute liberally to Emperor Wu's war chest. In reward he was made an official and advanced to chancellorship of a kingdom and then to the position of Grandee Secretary. . . . Chu Mai-ch'en, coming from a poor family, cut firewood to support himself while studying. Because of his literary talents he became a Palace Grandee, then Administrator of a commandery in Chekiang. . . . Sang Hung-yang, the son of a shop-keeper, was drafted into the government because of his busi- ness acumen and given the job of provisioning Emperor Wu's HAN HISTORY AND SOCIETY 43 armies. Later he was that Grandee Secretary who, having helped to estabhsh certain state monopolies, defended the government's economic policies in the great debate of 81 B.C., immortalized in the Discourses on salt and iron. . . . Another successful scholar was the poor farm boy K'uang Heng, who indentured himself in order to study. Passing through many official positions, he became Tutor for the Heir-apparent under Emperor Yiian, then Grandee Secretary, Chancellor, and a marquis. . . . Wang Tsun, starting his career as a petty jailer, became a commandery Administrator, Major in the army. Censor, and Chancellor of a kingdom. Demoted to the rank of a commoner, he rose again to become Inspector of Morals in 33 B.C. All those men, and many others like them, climbed from obscurity mainly by personal ability. To round out the picture, here are longer sketches of two consort families that rose from the humblest origin. Emperor Wu's first Empress nee Wei was the daughter of a slave woman in the household of his older sister. At a banquet given by her, the Emperor spotted the girl singing and dancing in a chorus, was infatuated, took her into his palace, and then forgot her for more than a year. Once when he noticed her weeping, he "pitied" her and granted her his "favors." When she conceived, the Emperor summoned her older brother and her younger half-brother, Wei Ch'ing, to serve in the palace. Wei Tzu-fu bore three daughters and finally a son who was chosen Heir-apparent, as a result of which she was made Empress. By that time Wei Ch'ing had become a general. The next year he received a marquisate, and later saw his three sons ennobled. One of the greatest Han warriors, Wei Ch'ing was surpassed in his own day only by his brilliant nephew, Ho Ch'u-ping, the natural son of another of Wei Ch'ing's half-sisters. He, too, became a marquis in recognition of his military feats, and introduced to court his half- brother, Ho Kuang. When Emperor Wu died Ho Kuang was one of the three regents for the minor. Emperor Chao, to whom he married his granddaughter and whose government he dominated. When Emperor Chao died without issue. Ho Kuang engineered the selection of a successor. Then he led the coalition of ministers who petitioned the Empress Dowager (his own granddaughter, age fourteen or fifteen) to depose his unwise selection, and he helped pick the next ruler. Emperor Hsiian. For his services to the state Ho Kuang was richly rewarded, held the chief military position, and was awarded the largest estate of any Han marquis. His son and two grandnephews also became 44 SLAVERY IN THE FORMER HAN DYNASTY marquises and high officials, yet two years after his death the family was accused of intrigue and the poisoning of an empress and was stamped out. The second family of humble origin was related to Emperor Hslian's mother. Wei Tzu-fu's son was Heir-apparent for thirty- seven years, and had a grown son. A retainer of the Heir-apparent, sent to find female entertainers for his patron's household, secured a quintet of singing and dancing girls, among whom was Wang Weng-hsii. The daughter of a simple village couple, she had been taught her trade in the household of the younger son of a marquis, and then sold by him to a merchant. Entering the Heir-apparent's household, she became the favorite of his son, to whom she bore a male child who much later became Emperor Hsiian. In 91 B.C., when the child was only a few months old, there occurred a palace intrigue in which the Heir-apparent was accused of conspiring to kill Emperor Wu by black magic. The Emperor was sick and away from his capital, so the accusers were able to slay the Heir-apparent and his whole family. Only the infant son of Wang Weng-hsii was rescued by a loyal official and reared as a plebe- ian in the family of his grandmother, nee Shih. After Emperor Chao died without heir in 74 B.C., this forgotten great-grandson of Emperor Wu was raised to the throne at the age of seventeen. It then became essential to discover whether any relatives of the new Emperor's mother were still alive. After many disappointments, the commission found an old lady. Dame Wang, and her two sons, Wu-ku and Wu, and brought them to the capital in 67. The people of Ch'ang-an roared with laughter when these rustics rode through the gate, but a careful judicial investigation proved that the woman was indeed the Emperor's grandmother, and the two men his uncles. Dame Wang was made a baroness with an estate of 11,000 households, while the men became marquises, each with 6,000. Wu-ku's son rose from that humble background to become Grand Minister of Agriculture and General of Cavalry and Chariots. Wu's son, Wang Shang, inherited the marquisate in 52 B.C. and steadily advanced to the top position of Chancellor in 29. But Emperor Ch'eng was only his second cousin once removed, and already another Wang clan, related to Emperor Ch'eng's mother, was gaining power. Blaming an eclipse of the sun on Chancellor Wang Shang, and producing arguments to prove that his family wealth and power were so great as to menace the dynasty, the other HAN HISTORY AND SOCIETY 45 Wang family secured his removal from office in 25. After three days he "fell sick, spat blood, and died." Fall from glory was easier and often quicker than ascent. A characteristic of Han nobility was its impermanence of tenure. Convenient tables which assemble pertinent data about the nobility can be analyzed by rough statistics. Out of approximately 850 persons who were granted titles — vassal kings, marquises who were sons of kings, marquises of military merit and of consort families — the lines of more than a hundred are said specifically to have expired for lack of an heir, and at least as many more end suddenly for reasons no longer known. Lack of an heir can often be accounted for by death of the title-holder before he reached maturity, but since collateral descendants were appointed by imperial favor (and such cases were not included among the hundred) that does not explain the high figure. Such family dissolution is the more sur- prising in a polygamous society which put great emphasis on progeny. Almost half the titles, over four hundred, were lost because the holder committed some serious crime. Usually demotion was considered punishment enough; in some 230 cases the noblemen were dismissed and became commoners. In 45 cases, however, the guilty noblemen were sentenced to serve terms as convicts, and 99 others were executed, while a few committed suicide to escape that fate. Another 170 are simply said to have been dismissed, without mention of crime. Impermanence of tenure is most graphically shown by calcula- tions based upon figures given by Wu Ching-ch'ao. The average period for all the nobility was only 2.31 generations. During the lifetime of the appointee or at the time of his death, 41 per cent of all titles were lost; nearly 63 per cent had been lost by the end of the second generation; and 79 per cent no longer existed after the death of the original nobleman's grandson, if he had one. These calcula- tions are, of course, weighted by the fact that all noble lines were terminated during Wang Mang's reign, so that titles granted toward the end of the dynasty had only a brief course. However, not a single direct or collateral descendant of Kao-tsu's original marquises, the men who helped found the dynasty, had noble title by 86 B.C.; and in 62 Emperor Hsiian bemoaned the fact that descendants of this early aristocracy had fallen to the position of indentured laborers. If fluidity of social position typified the upper classes it also characterized the lower. Members of any societal group might become convicts at one stroke. Slaves were generally recruited from 46 SLAVERY IN THE FORMER HAN DYNASTY the commoner class, but sometimes from the nobihty itself. Con- versely, descent to those classes was no bar to becoming a commoner and even a nobleman. Imperial amnesties of convicts were frequent, and theoretically all convicts became commoners within six years. There were several ways in which slaves could achieve freedom, thus becoming commoners automatically. Thus the upper and lower classes were not rigidly fixed, and their members were incessantly absorbed into the great commoner group. To place this subject in proper perspective one adjusting observa- tion needs to be made. While there were no unsurmountable walls between the social classes, the number of commoners who ever left that class was trifling in proportion to the total group, which itself made up the bulk of the population. The educated people of the Han period fell heir to a considerable literature of philosophy, classical texts, poetry, history, and romance belonging to the Chou epoch, and important intellectual activities of the age were discovery, investigation, and annotation of the books which had been driven from circulation by the First Ch'in Emperor. Privileged academicians had access to the imperial library where some of this early literature was preserved. Competing private schools expounded various classical texts and developed the concepts of different Chou philosophers. The period also produced an extensive literature of its own, in history, poetry, political phi- losophy, military science, arts, divination, belles-lettres, and many other branches. Just as the dynasty was nearing its close, Liu Hsiang and his son Hsin made digests of the extant literature and listed some 600 authors whose works numbered 13,000-odd fascicles, Ssu-ma Ch'ien compiled the first general history of China, bring- ing the narrative well down into the reign of Emperor Wu. Other men extended Ssu-ma Ch'ien's chronicle, or wrote histories of par- ticular periods and subjects. Then, during the first century. Pan Ku picked up the research of his father and wrote a history of the dynasty just ended. Others treated special subjects relating to the Former Han period, while commentators, living only a century or two after Wang Mang, added bits of data here and there to explain facts and terms already growing obscure. Only a small part of Han literature remains. Some of it contains the information about slavery on which this book is based. Only those sources which have been employed and found most useful in preparing this introductory chapter, both as to facts and concepts, are here listed by order of subjects. HAN HISTORY AND SOCIETY 47 For decline of feudalism, unification, and the Ch'in empire see Henri Maspero, La Chine antique, Paris, 1927, pp. 361-425 (Book IV, The Warring States); J. J. L. Duyvendak, trans.. The book of Lord Shang, a classic of the Chinese school of law, Probsthain's Oriental Series, vol. 17, London, 1928, pp. 1-65 (introduction), and passim; Edouard Chavannes, trans., Les memoires historiques de Se-ma Ts'ien, 5 vols., Paris, 1895-1905 (cited as MH), vols. IV and V passim (sections on the hereditary houses after 400 B.C.), and II, pp. 58-246 (on Ch'in after 400, and Ch'in Shih-huang); Owen Lattimore, Inner Asian frontiers of China, American Geographical Society, Research Series, No. 21, New York, 1940, pp. 369-443 (on kingdom and empire in ancient China), and passim; Derk Bodde, China's first unifier, a study of the Ch'in dynasty as seen in the life of Li Ssu (2807-208 B.C.), Sinica Leidensia, vol. 3, Leiden, 1938. From the founding of the Han dynasty, the principal source is the Ch'ien Han shu [History of the Former Han dynasty], by Pan Ku and others (originally titled Han shu; later the prefix Ch'ien was added to distinguish it from the Hon Han shu or History of the Latter Han dynasty). The edition used is the imperial Ch'ien-lung edition (1739-46) of the twenty-four dynastic histories, reprinted at Shanghai in 1908 by the Chi ch'eng t'u shu kung ssu, which has been checked with the monu- mental Han shu pu-chu by Wang Hsien-ch'ien, printed in Changsha in 1900. The Ch'ien Han shu is abbreviated throughout this book as CHS. Chilan ("chap- ters") are cited by Arabic numerals; parts of chiian that are separately paged are cited by capital letters, A, B, C, etc., in sequence; page numbers follow chapter numbers, and are cited by Arabic numerals, with a and b for recto and verso. (References to the Shih chi (SC), Hou Han shu (HHS), and other dynastic histories are to the same edition, and the system of notation is identical.) CHS deals with this period, in ch. 1-5 (emperors); 24A, la-6b, and B, la-3a (economics); 31-52 (important persons); and elsewhere. The first five chapters of CHS, covering the period 209-141 B.C., have been translated by Homer H. Dubs (Baltimore, 1938; hereafter cited as "HFHD, vol. I." Four other volumes are promised). This covers the "first phase" of my historical account, and the introductions by Dubs to separate chapters were exceedingly useful. Parallel information appears in the Shih chi by Ssu-ma Ch'ien, translated in MH, vols. II, pp. 246-510 (on Hsiang Yii, Kao-tsu, and other emperors to Wu), and III, pp. 538-544 (economics). The reign of Emperor Wu, the "second phase," is covered in CHS, 6 (Emperor Wu); 24A, 6b, and B, 3a-8a (economics); 53-63 (important generals, statesmen, and literary figures); most of 94A, 95, and 96 (on foreign states); 97A, 5a-7b; and many other places. On the period in general, see MH, vol. I, introduction, pp. Ixii-cviii. On foreign wars see chapter IV, below, and CHS and SC references there cited; also Lattimore, op. cit., pp. 429-510. On the Hsiung-nu particularly, cf. J. J. M. de Groot, Die Hunnen der vorchristlichen Zeit (vol. 1 of his Chine- sische Urkujiden zur Geschichte Asiens), Berlin, 1921; A. Wylie, trans., "History of the Heung-noo in their relations with China," Journal of the Anthropological Institute of Great Britain and Ireland, vol. 3, 1873, pp. 401-451. On economic conditions, in addition to CHS, 24, cf. MH, vol. Ill, pp. 544-600, which translates SC, 30. On canal digging see Chi Ch'ao-ting, Key economic areas in Chinese history, London, 1936, pp. 80-86; MH, vol. Ill, pp. 520-537; M. S. Bates, "Problems of rivers and canals under Han Wu Ti (140-87 B.C.)," JAOS, vol. 55, 1935, pp. 303- 306. Third phase, "economic decline": CHS, 7-11 (emperors); 24A, 7a-8a, and B, 8a-b (economics); 68-87 (statesmen); 93, 4b-7a (Tung Hsien); 97A, 7b ff. and all of B (consort families). On population-growth, landlordism, and famines see chapter IX, below, and CHS references there cited. Also, Ch'en Po-yin, Chung- kuo t'ien chih ts'ung k'ao [An investigation of the Chinese land system], rev. ed., Shanghai, 1936, pp. 51-55 (hereafter cited by translated title) ; Wan Kwoh-ting, Chung-kuo t'ien chih shih {An agrarian history of China), vol. I, Nanking, 1933, pp. 82-88 (hereafter cited by its English title); Wan Kwoh-ting, "Liang Han chih chiin ch'an yiin-tung (The movement for equal land holdings in the Han dynasty)," CLHP, vol. 1, 1931, pp. 1-25 (see pp. 14-16); T'ao Hsi-sheng, Hsi Han ching-chi shih [An economic history of Western Han], Shanghai, ed. of 1935, passim, esp. pp. 43-73 (hereafter cited by translated title). For contemporary Han descriptions of economic conditions see Esson M. Gale, trans., Discourses on salt and iron, a debate on state control of commerce and industry in ancient China, chapters I-XIX, 48 SLAVERY IN THE FORMER HAN DYNASTY translated from the Chinese of Huan K'uan, Sinica Leidensia, vol. 2, Leiden, 1931; and Esson M. Gale, Peter A. Boodberg, and T. C. Lin, trans., "Discourses on salt and iron {Yen t'ieh lun: chaps. XX-XXVIII)," JNCBRAS, vol. 65, 1934, pp. 73-110. Fourth phase: On Wang Mang, CHS, 98 and 99; 24A, 8a-9a, and B, 9a-llb; and many others. CHS, 99, has been translated by Hans O. H. Stange, Die Monographie iiber Wang Mang (Ts'ien-Han-shu Kap. 99), Abhandlungen fiir die Kunde des Morgenlandes, vol. XXIII, pt. 3, Leipzig, 1938; see also his Leben, Personlichkeit und Werk Wang Mang's dargestellt nach dem 99. Kapifel der Han- Annalen, Berlin, 1934. I have also read, in manuscript, a forthcoming translation of CHS, 99A, by Clyde Bailey Sargent, Wang Mang: A translation of the official account of his rise to power as given in the History of the Former Han dynasty. On details of Wang Mang's reforms, and widely conflicting appraisals, see Hu Shih, "Wang Mang, the socialist emperor of nineteen centuries ago," JNCBRAS, vol. 59, 1928, pp. 218-230; and Homer H. Dubs, "Wang Mang and his economic reforms," TP, vol. 35, 1940, pp. 219-265. The rebellion against Wang Mang and the establishment of the Latter Han are described in CHS, 99C, and HHS, lA, and 41 ff. See also L. Wieger, trans., Textes historiques (Rudiments, vols. 10-11), 2 vols., Ho-chien Fu, 1903-04, vol. 1, pp. 732-770. Population figures for a.d. 2 are in CHS, 28B, 9a; for Latter Han, in HHS, 33, 8a-b. See also Wan Kwoh- ting, "Han i ch'ien jen-k'ou chi t'u-ti li-yung chih i pan (Population and land utilization in China, 1400 B.C.-200 a.d.)," CLHP, vol. 1, 1931, pp. 133-150 (pp. 138-142) (hereafter cited by its English title). Population distribution: idem, map, p. 142, based upon CHS, 28. Larger cities, Albert Herrmann, Historical and commercial atlas of China, Harvard- Yenching Institute, Monograph Series, vol. 1, Cambridge, 1935, pp. 20, 22 and 23; HFHD, vol. I, inset map; see also Map, this volume. On political divisions: CHS, 28 (for totals, 28B, 8b). Local and central administration, CHS, 19A (abstracted, MH, vol. II, pp. 514-533); HFHD, vol. I, p. 27, footnote 2, and p. 75. Also for any details, Hsi Han huiyao, ch. 31-43. (The Hsi Han hui yao is a classified compendium of information in CHS, compiled by Hsii T'ien-lin and completed a.d. 1211. Hereafter cited by Chinese title; page references to Commercial Press edition [Kuo-hsileh chi-pen ts'u,ng-shn], Shanghai, 1935.) The section on the commoner class draws on many parts of CHS; see references in chapter IX, below. See also Wu Ching-ch'ao, "Hsi Han ti chieh-chi chih-tu (The class system of the Western Han dynasty)," CHHP, vol. 10, 1935, pp. 587-629 (pp. 598-606) (hereafter referred to by English title). He assembles much useful informa- tion, as does the Hsi Han hui yao, ch. 47 and 52. On the attitude toward merchants see memorials of Chia Yi, Ch'ao Ts'o, and Tung Chung-shu in CHS, 24; and same, passim, for government action. Many modern Chinese writers have discussed merchandising and the Han attitude toward it, but T'ao Hsi-sheng, op. cit., has perhaps gone farthest in making it a central theme. For the bureaucracy, same references as for local and central administration; see also Wu Ching-ch'ao, op. cit., pp. 613-614. Concerning honorary ranks, cf. MH, vol. II, pp. 527-528; and Wu Ching- ch'ao, op. cit., pp. 614-615. The basic sources on the nobility are CHS, 14-18, which are Tables of noble houses, with important prefaces; see also SC, 17 (MH, vol. Ill, pp. 86-92). For vassal kings and marquises who were sons of kings cf. also individual biographies in CHS, 38, 44, 47, 53, 63, 80. CHS, 97, assembles important data on individual consort families and gives cross-references to separate biographies of more important ones such as Tou Ying and T'ien Fen (CHS, 52), Li Kuang-li and Li Yen-nien (61, 93), Wei Ch'ing and Ho Ch'u-ping (55), Ho Kuang (68), Wang Shang and Shih Tan (82), Chang Fang (59), and the relatives of the Empress of Emperor Yiian (98). These are only a few of many important affinal relatives. Sizes of estates were calculated from CHS, 16-18. References to criminals are widely scattered; see citations to CHS in chapter X, below. On the Han law code, cf. Ch'eng Shu-te, Chiu ch'ao lii k'ao [An investi- gation of the legal codes of the nine dynasties {Han through Sni)], Commercial Press one volume ed., 2nd ed., Shanghai, 1935 (hereafter cited by Chinese title), intro- HAN HISTORY AND SOCIETY 49 duction; and also references and citations in chapter III, below. On number of crimes involving death sentence see HHS, 76, 4b. The section on fluidity of social position is more or less original. Biographical vignettes: Chi Yen (CHS, 50), Kung-sun Hung (58), Chu-fu Yen (64A), Jen An and T'ien Jen (my document 39), Pu Shih (58), Chu Mai-ch'en (64A), Sang Hung-yang (24B, 5a), K'uang Heng (81), Wang Tsun (76). The careers of most of them may also be traced in 19B; there are many others like them. Full abstracts of the CHS accounts of all important persons are to be published by Dr. Dubs in his final volumes of Glossary and Onomasticon. The accounts of the two consort families, Wei and Wang, are based upon material translated and annotated in my documents 26, 27, 29, 70, 72; and 55, 80, 98, in which the basic sources are cited. Calculations on loss of noble title are based on CHS, 14-18, while numbers of generations are figured from Wu Ching-ch'ao, op. cit., pp. 612- 613. Ssu-ma Ch'ien himself noted the impermanence of noble title when he pointed out that in his own day (specifically the period 104-101 B.C.) descendants of only five of Kao-tsu's original nobles still enjoyed the title. Cf. MH, vol. Ill, p. 124. II. HISTORICAL SOURCES, AND DEFINITION OF TERMS Source material has different degrees of value depending not only upon its nature and authenticity, but also upon the subjects to which it is applied. The subject here considered is a lowly social group in ancient China — slaves — and the essential problems are social and economic. The sources available, on the other hand, were primarily designed to recount the political history of an empire and the activities of the ruling class. This inconsistency between sources and subject fundamentally delimits the investigation. In what respects are the sources inadequate, wherein are they strong, and how do they restrict this study? Historical Sources Most of the slavery documents translated in Part II come from the Ch'ien Han shu [History of the Former Han dynasty] by Pan Ku (and others), who wrote near the end of the first century of our era. Of four principal divisions in the history in its present form, the first or "Imperial Annals" deals in strictly chronological fashion with the official acts of emperors and the political history of the empire from 209 B.C. to A.D. 6. A classificatory principle underlies the second division, or "Tables," each of which contains terse but pertinent facts about members of the various classes of enfeoffed nobility and about the holders of the highest positions in the bureaucracy. Within each Table noble houses are arranged chrono- logically by date of appointment, and the inheritors of each title are traced through the several generations; appointments to all higher offices are also treated chronologically. These Tables might be made appendices in Occidental history. Chapter 21 begins the third division, containing the ten "Treatises." Each is a general monograph on ritual and music, jurisprudence, economics, astronomy, geography, literature, and so forth, but only the "Treatise on Economics" is important for material on slavery. The fourth division, the "Memoirs," is in seventy chapters and accounts for half the total work. It is by far the richest division in a history apparently conceived as a narrative of the deeds of great men. Starting with biographies of those who aided and opposed Kao-tsu, it proceeds chronologically to the end of the djmasty. Chapters 88 through 97 (58 through 68 in the division itself) depart from the chronological principle and give first place to classification. Thus, chapters 88 to 93, and 97, bring together important literati, 50 SOURCES AND DEFINITION 51 officials who were champions of the people, others excessively tyrannical, rich business men, wandering redressors of wrong, im- perial male favorites, imperial consorts and such of their relatives as were not granted individual biographies. Three important chapters, 94, 95, and 96, give historical accounts of China's relations with the Hsiung-nu, with the peoples of the southwest, southeast, and northeast, and with the peoples of the Western Regions, respec- tively. Logically following chapter 97, on consorts and their families, comes the biography of Wang Mang's aunt. Empress of Emperor Yiian, and her relatives; then the longest and penultimate chapter, on Wang Mang. Chapter 100 is the historian's "preface" and family history. Because of the book's cyclopaedic character and classificatory arrangement, facts about any event are widely scattered; for ex- ample, to study the campaigns against the Hsiung-nu during Em- peror Wu's reign one must read back and forth through the Annals of the period, the biographies of leading generals and policy-making statesmen, the "Memoir on the Hsiung-nu," and the "Treatise on Economics." Scraps of information appear in the Table of those marquises who acquired titles for military merit, the "Table of the Bureaucracy," the "Treatise on Geography," and in many other places. Indeed it is impossible to be sure of covering a problem merely by reading through what appear to be relevant chapters. Passages dealing directly with slavery were found in sixty-eight of the hundred chapters, some in the most improbable places, others in sections having many long and significant items. Only second to the Ch'ien Han shu is the Shih chi by Ssu-ma Ch'ien and his father Ssu-ma T'an, completed some time between 100 and 90 B.C., but containing important additions by Ch'u Shao- sun and others. For the first century of Han the two histories are parallel, each supplementing the other. Pan Ku was heavily in- debted to Ssu-ma Ch'ien for many of his chapters, while conversely the present text of the Shih chi contains sections which appear to have been copied back into it from the Ch'ien Han shu because they were lost in transmission. Nearly half the work deals with the pre- Han epoch, and since its references to slavery for the Han period are nearly all found in the Ch'ien Han shu, only its unique items have been translated, the rest being taken from Pan Ku's more comprehensive work, and textual variants noted. These two works begin the series of twenty-five (or twenty-six) Standard Histories which together cover all recorded Chinese history 52 SLAVERY IN THE FORMER HAN DYNASTY to 1912. The third of the series, the History of the Latter Han dynasty, written in major part by Fan Yeh of the fifth century (but based on many earher works), supphed a few of the documents, translated because they apphed to the Former Han period. The work also contained numerous data which have been used for comparison with and illustration of Former Han conditions. Later dynastic histories down to T'ang were also consulted for illustrative material, but they were used cautiously because of the increasing factors of elapsed time and historical change. During the half century after Wang Mang, Wei Hung composed the Han chiu i to record the governmental system of the period just ended. This book contains some of the most revealing items on slavery included among the documents. Another important source was the long and semi-humorous essay by Wang Pao, dated 59 B.C. The Discourses on salt and iron or Yen t'ieh lun, compiled by Euan K'uan during the reign of Emperor Hsiian (74-49 B.C.), contained one important passage and a number of minor references to slavery. This work is part of the Han Wei ts'ung-shu [Collection of books on the Han and Wei periods] ; other Han texts there included were consulted and used as reference, but very little material on slavery was found that could be included among the reliable docu- ments upon which this study is primarily based.' 1 These are the basic sources. Problems of textual criticism are not in the province of this book. All the works have passed the exacting examination of Chinese scholarship and (save Wang Pao) of Sinological scrutiny. Historical criticism has been attempted in the footnotes to those documents which presented special problems concerning the type of source material on which they were based, probable accuracy of statements made, and evidences of special bias. For the Ch'ien Han shu Dr. Dubs promises to supply in his introductory volume a translation of the historian's "Introductory Memoir," lives of the author and others who worked on the book, a discussion of the texts and their tradition, and a list of important commentators (HFHD, vol. I, p.ix). His introductions and appen- dices to the five chapters already published contain important historical criticism; see also items in my bibliography under Dubs. On Pan Ku and other members of the Pan family, the plan and sources of the Ch'ien Han shu, and a comparison between it and the Shih chi, see Lo Tchen-ying, Les formes et les methodes his- loriques en Chine: Une famille d'historiens ei son oeuvre, Paris, 1931. Another valuable work primarily concerned with Pan Ku's sister, who is supposed to have helped with the CHS, is that by Dr. Nancy Lee Swan, Pan Chao: Foremost woman scholar of China, New York, 1932. This also has some valuable notes (pp. 158- 161) on the composition of the History of the Latter Han dynasty. On the Shih chi the most authoritative Occidental textual and historical criticism is still that by Chavannes in his introduction to MH, vol. I, which contains chapters on the authors, the age in which they wrote, their sources, their critical method, and the later history of the text. For its precise discussion of Chinese historical method and native textual criticism, as well as for its convenience and wealth of reference, the work of Charles S. Gardner {Chinese traditional historiography, Cambridge, 1938) is unsurpassed. sources and definition 53 Nature and Deficiency of Sources "The Chinese," Gardner points out, ". . . conceive of the past as a series of concrete events and overt acts; and of history as a registration of them which should be exact and dispassionate, with- out any projection across the scene of the personaHty of the registrar, who must punctiliously refrain from garbling his presentation by his own perhaps imperfect appreciation of the true sequence of causation. It is the function of the Chinese historian to collect the facts and to subject them to a process of discreet filtering which may only suppress those of insignificant importance and present those of greater moment to speak for themselves without interference .... And accordingly, verbatim reproduction of the records of earlier historians, no matter how extensive, is to be regarded, not as plagiarism, but rather as the natural and reasonable process by which new histories of previously recorded events should be con- structed. [Chinese] historical writing ordinarily involves, not original composition of any considerable length, but compilation of choice selections from earlier works." ^ There is abundant evidence, internal and historical, that the authors of the History of the Former Han dynasty, adhering to the methods of Ssu-ma Ch'ien, used as their basic sources both state archives and existing books such as the Shih chi, themselves based upon archives. These sources were not employed merely as the foundation for a synthetic account, but many were copied in entirety or in large part. Thus, the Ch'ien Han shu is itself a vast storehouse of documents faithfully reproducing the selected originals, except in so far as errors have crept in through centuries of transmission. Em- bedded in it are imperial edicts and orders, long memorials on economic conditions, formal recommendations on state policy, reports of investigating commissions summarizing testimony of witnesses, records of trials and judicial verdicts, accusations pre- sented by individuals or groups against other officials, intra-bureau communications, records of administrative acts, memoranda on military campaigns, and other data, almost without end. This eclecticism gives the history a high degree of reliability in the fields it was intended by the historian to cover; the difficulty is that slavery was not one of the subjects considered of historical importance. There are many references to slavery in the copied archives and in other passages whose sources are no longer evident. These references have a fortuitous character; they were included because 1 Gardner, op. cit., pp. 69-70. 54 SLAVERY IN THE FORMER HAN DYNASTY they were integral parts of selected documents, or because they happened to be necessary elements of the larger narrative. As they are entirely unsystematic, they leave great gaps in the picture even when assembled and organized. For example, it is mere chance that we know that slaves could purchase their freedom. There would be no evidence of the practice in Former Han histories if one such case had not been mentioned in connection with the trial of a nobleman. Likewise, it is almost by accident that we know of the enslavement of a certain group of noble folk; we know because fourteen years later they were freed by an imperial edict which only happens to be recorded. Most historians write for their contemporaries rather than for future generations, and accordingly take much common knowledge for granted. Even writing for the future, one cannot predict what parts of his own culture will change so radically as to need explana- tion. Most references to slavery in the previous dynasty were self- evident when set down; to explain them would have been pedantic. Yet many of them soon became so obscure as to arouse dispute among Chinese commentators only a few centuries later, while today uncertainties of meaning abound. Two other general weaknesses arise from the nature of the sources. Because mention of slaves is incidental, the slaves appear chiefly in association with important people or events. The "important" people were those connected with the state (noblemen and high officials), or those who by their acts or accomplishments either influenced the course of national events or won some niche in the historian's hall of fame. About the slaves of these people we are comparatively well informed; but about slaves belonging to "unim- portant" people, those who were only somewhat wealthy, somewhat successful in business and scholarship, or somewhat important as administrators, there is little information. If the common folk had slaves, virtually nothing is known about them. This lack of information may greatly distort the picture of Han slavery, especially in regard to its extent and economic importance. In the second place, where slaves happen to be mentioned in matters of state concern there is a conspicuous lack of detail about them; for example, we read an imperial order freeing government slaves, but nothing indicates whether, or to what extent, the order was carried out. Presuming that it was at least partially executed, there is no answer at all to such natural questions as the effect of the order upon the slaves, the mode of establishing them in plebeian life, SOURCES AND DEFINITION 55 the requisite changes in records of bureau property, sorts of manu- mission papers, and the Hke. To give another example, there are numerous references to trials of noblemen who had ordered their slaves to murder people, or to do other unlawful acts, and in each case the verdict against the nobleman is reported. Not one of the cases, however, gives the slightest indication of what happened to the slave. This is not a weakness of the material on slavery alone; it is a characteristic of Han history, and perhaps of all formal history. Throughout the annals of the Former Han period there is an exasper- ating lack of information on all the intimate details of administration and economics. When we read of thousands of prisoners captured in any particular war, we search in vain for clear-cut evidence of their disposition, of their having been brought into China, or even for proof of their having been captured. There is no eyewitness description of prison camps, no account of triumphal processions, no report by an official who had actually inspected or counted a batch of captives. Were they merely prisoners on paper? Are the reports fabrications? Only by devious means can we learn that they were not. The Former Han history remains a curious mixture : archives copied in extenso, unexpectedly revealing important facts about administration, law, and society; biographies recounting the most intimate matters in the lives of the great; and large state- ments serenely floating in a vacuum. Advantages of Sources This deficiency, in the last analysis, should not be charged against the histories but against the attempt to use them for an end they were not designed to fill. Records of administrative routine, filled with passing references to government slaves, must have accumulated in piles and mountains in the archives of various bureaus. Clearly, they were too unimportant to encumber a grand history. Only archaeology normally reveals such inconsequential details. Already, for Han China, as for so many other ancient cultures, archaeology has unearthed a rich written record that fills some of the gaps left by native historians. Because of climatic conditions and the types of writing material used, documents of the Han and later periods have appeared most extensively at the periphery of the ancient empire, in the dry sands of Chinese Turkestan among ruined watch towers and settlements along the now desolate sections of the Great Wall. Mere scraps of inscribed wood and silk, bamboo and paper, these memoranda from the rubbish heaps reveal in fine detail the 56 SLAVERY IN THE FORMER HAN DYNASTY ordinary affairs of life in military encampments. Some documents discovered by Stein and Hedin, translated by Chavannes and Conrady, do mention slaves, but so fragmentarily as to disclose only that slavery existed also along China's northwest frontier at a time roughly corresponding with the Former and Latter Han periods.^ In 1930 the Sino-Swedish Scientific Expedition to Northwest China, organized by Dr. Sven Hedin, discovered "more than ten thousand" inscribed wooden slips, reportedly of Han date, near Estingol (Chii-yen) in Ningsia, while later many others were found by the Chinese archaeologist, Huang Wen-pi, in the Lop-nor region already made famous by discoveries at the ancient Chinese military station of Lou-Ian. These documents are said to include many references to slavery, and one bit of information on prices has already been published by Lao Kan.- Unfortunately the Chii-yen and new Lop-nor documents are still unavailable. They may not in the end prove very enlightening on problems of Han slavery; or they may by lucky chance supply some of those minute and informal details 1 Aurel Stein, Serindia, detailed report of explorations in Central Asia and westernmost China, 5 vols., Oxford, 1921, vol. 2, pp. 722-790; Edouard Chavannes, Les documents chinois decouveris par Aurel Stein dans les sables du Turkestan oriental, Oxford, 1913 (English translation of introduction in the New China Review, vol. 4, 1922, pp. 341-359); August Conrady, Die chinesischen Handschriften und sonstigen Kleinfunde Sven Hedin's in Lou-Ian, Stockholm, 1920 (Conrady's tran- scriptions are reprinted and corrected in the Bulletin of the National Library of Peiping, vol. 5, No. 4, July-August, 1931, pp. 25-64). Documents apparently mentioning slaves are the following: Chavannes, p. 81, No. 356; p. 94, No. 422; p. 95, No. 428 (might be of date 39 B.C.); p. Ill, No. 508(?) ; also a Chin dynasty document, p. 167, No. 770. See also Conrady, p. 81, No. 5.1; p. 97, No. 19.6; p. 107, No. 29.6. Conrady's documents are mostly later than Han. In several other of his documents the term "slave" seems to be part of a tribal name, as [Hsiung]-nu, p. 105, No. 27.2; or Shao-nu (of uncertain mean- ing), p. 104, No. 25.1; p. Ill, No. 33.1 (three times). 2 "Han tai nu-li chih-tu chi liieh (The system of slavery during the two Han dynasties)," Academia Sinica, Bulletin of the Institute of History and Philology, vol. 5, pt. 1, 1935, pp. 1-11 (hereafter cited by its English title), p. 2. The vicissitudes of these documents after discovery is almost melodramatic. In 1937 the following statement appeared in the June issue of the Quarterly Bulletin of Chinese Bibliography: "After several years study, the thousands of manuscripts on wood discovered by the Sino-Swedish Scientific Expedition to the Northwest have been transcribed. The wooden strips are now in Shanghai, where they are being photographed with a view to publication. "The manuscripts and other objects discovered by the expedition at Lob Nor have been studied by Huan Wen-pi, a member of the Expedition, whose study is ready for the press. It is reported that this study will be published by the Commercial Press, Shanghai." (English ed., vol. 4, No. 1, 1937, p. 66.) Almost at the moment that Bulletin reached America, China was invaded by Japan. During August and September great parts of Shanghai were destroyed, including the Commercial Press. What happened to the "wooden strips now in Shanghai''^ Had they been preserved for two millenniums only to be burned on SOURCES AND DEFINITION 57 which transmitted historical literature lacks so lamentably. At present the information on slavery in the Former Han period is much less than that for a period of equal length in Greece after the Persian Wars, and only a trifle of that for the contemporary period in Rome. In spite of their deficiencies the available sources for this study have at least two advantages. In the first place, the fact that so much of the Shih chi and Ch'ien Han shu consists of archives copied direct and verbatim means that references to slavery in those passages enjoy nearly the independent validity of excavated records. The high standard of integrity shown by Ssu-ma Ch'ien and Pan Ku in their use of sources, together with the fact that matters concerning slavery were generally incidental, minimizes the dangers of original distortion. Errors in transmission cannot be controlled absolutely, but fortunately textual criticism is the Chinese forte. ^ Whereas records were often copied into the histories intact, and whereas bits on slavery are always intact, most of the excavated material is badly mutilated and in part undecipherable. Furthermore, though dates are rare on such material, most of which cannot be identified more precisely than within a century, the archival rem- nants in the histories can all be closely dated, usually within a year and sometimes even to the day. This advantage of authenticity only slightly less than that of original records pertains to some 35 per cent of the texts, not counting many others obviously based on archives. When, for example, an the eve of publication? In September, 1940, another note in the Bulletin finally answered these questions: "In 1937, plans were well under way for their publication, but they were interrupted by the outbreak of the Sino-Japanese hostilities. All the plates were destroyed during the invasion of Shanghai in August, 1937. "Through the financial assistance of the Board of Trustees of the Indemnity Funds remitted by the British Government, its publication is now assured. During 1938-40 much time has been spent in the difficult task of photographing these records. The work of photographing having been completed, the Commercial Press is commissioned to publish this book on behalf of the Scientific Mission to North- western China. The plates alone will occupy over 600 pages and the book will be bound in the traditional Chinese style." {Quarterly Bulletin of Chinese Bib- liography, English ed., n.s., vol. 1, No. 3, 1940, p. 284.) No further report on this publication was received up to the time of the out- break of the war between the United States and Japan. The original Han documents, however, are known to be in safety. For information on the discoveries at Lop-nor see Folke Bergman (BMFEA, vol. 7, 1935, pp. 76-77) and two items by Huang Wen-pi and Ma Heng cited in his bibliography (p. 143). • See especially Gardner, op. cit., pp. 18-63, and references there cited. In the "Treatise on Economics," CHS, 24A, Pelliot discovered that one hundred characters differed as between a T'ang manuscript copy preserved in Japan and the modern text. Cf. BEFEO, vol. 2, 1902, p. 335 (also my J^5, footnote 4). 58 SLAVERY IN THE FORMER HAN DYNASTY official noted for his integrity is commissioned by the Emperor to investigate a certain suspicious ex-king and sends back a written report of his personal investigation, and in this quoted report casually mentions the ex-king's 183 slaves, that item may be con- sidered accurate. It is even more likely to be true because the investigator's letter of transmission mentions an appended list of the slaves and an invoice of palace property. Again, one of the long- est translated passages is the official report of a judicial commission, appointed after Emperor Ch'eng died without leaving an heir, to investigate the previous suspicious deaths of his two imperial sons. This was a most serious dynastic matter, and the officials cross- examined all employees of the palace who knew of the details. Witnesses are mentioned by name and identified by office. Some of them were slaves, and some of the facts revealed happened to concern palace slaves and their work. This report, left in its original form, is the raw material of history, skilfully woven into the chronicle by a compiler who knew by long habit when to let official documents speak for themselves. Nowhere is there any evidence that the historians themselves were interested in slavery as such. This is the second advantage of the sources. There is no abolitionist sentiment, no special plead- ing. Although the historians quote statements which show that some thinkers in the Han dynasty opposed slavery, the institution was apparently a matter of indifference to them historically. A subject of common knowledge (Pan Ku himself was a slave owner), slaves are mentioned only when entering naturally into the narrative. Reports are dry and matter of fact, and only occasionally show evidence of exaggeration. Thus, even in those numerous passages, particularly among the biographies, where we cannot determine the original sources, casual references to slaves have an exceptional reliability. Method of Using Sources How can transmitted literature possessing the inadequacies and advantages described be used most fruitfully in a study of slavery? To get the best results every reference from the basic texts pertaining to the period should be scrutinized, no matter how "inconsequential," and each should be placed in its historical setting. To counteract the Chinese categorical method of writing history, each reference should be dated (as it refers to slavery) and the whole group arranged chronologically. The result of using all references is a wider and surprisingly richer corpus than any other assembled; presentation SOURCES AND DEFINITION 59 of background places most of the documents in the setting essential for understanding; strict chronological arrangement has obvious historical advantages, but strangely enough is a unique feature of the present work.^ In the matter of analysis as little as possible was taken for granted about the system of Chinese slavery, and the documents, being assembled topically and compared, were allowed to speak for them- selves. Analysis itself, however, involves preconceptions; the materials had to be arranged in some manner to disclose logically the "essential" aspects of Former Han slavery. Questions about the natui'e and function of this slavery demanded formulation and some attempt at solution. Wherever possible, analyses and conclusions were tested in two ways: by comparing the Former Han data with similar data from later periods (particularly that of the adjacent Latter Han); and by studying the conclusions of modern Chinese scholars — conclusions which conflict among themselves but which all arise from a conceptual background somewhat different from that of an Occidental. The period studied intensively spans two and a quarter centuries, an elapsed time about equal to that between 1710 and 1940. It was approximately concurrent with the last two centuries of the Roman Republic, from the end of the Second Punic War through the founding of the Empire and part of the reign of Tiberius — a period, incidentally, of great development in the system of Roman slavery. In such a span of time great changes in the slavery system seem likely a priori. What justification is there, then, for taking items of information from various parts of the period and combining them in a description of slavery in the Former Han era? Does this not presuppose a static rather than an ever-changing situation, and does it not create a purely artificial something which never existed, as described, at any particular moment? - What reality would a des- cription of American slave-trading possess if it were synthesized from references dating: 1644 (a biography of a sometime slave), ca. 1646 (a description, in a geographical text, of an African slave port), 1669-72 (the description of slave markets in a speech in Parliament), 1691-1702 (a remark by a slave owner quoted in a physician's memoirs), 1726 (a biography of a slave owner), 1756-62 (the report 1 In the preparation of this book all discovered references to slaves were used, but about a quarter of them have now been placed in the appendix or used as footnotes of other documents. ^ No Chinese writer seems to have considered this deficiency in the method universally employed. 60 SLAVERY IN THE FORMER HAN DYNASTY of an investigating commission), 1787 (a slave contract), 1839 (an economic report), 1842 (a petition from a bm-eau chief to the presi- dent about the sale of public bondsmen), 1855 (a president's inaugural address condemning the slave trade), 1860 (a presidential executive order forbidding the sale of slaves), supplemented by many other items showing merely that slave-selling was continuous in that period? Because materials are too scanty for a detailed study decade by decade there is an obligation to try to detect and emphasize all evidences of historical change. Usually, therefore, references to each topic are treated chronologically, but the assumption of continuity is tested by noting whether the phenomena occur through- out later Chinese history, or better, in the dynasty immediately following the Former Han. When information on important topics is entirely lacking or quite inadequate the fact is pointed out. Unfortu- nately it is necessary to refer to a few of the documents repeatedly and ad nauseam. Actually, the analogy just made between Han China and North America from 1644 to 1860 is unjustified except as a warning. In ancient China changes in society certainly did not occur with any- thing like the rapidity that has characterized the Occident shortly before and during the Industrial Revolution. Probably social and economic developments were very gradual, and I believe the Former Han period is not too long to treat as a unit for a description of slavery if evidences of development are conscientiously sought out. Problems of Definition The fundamental problem of semantics remains. What do we mean by "slave"; and is the Chinese meaning of the terms translated as "slave" close enough to our meaning to approach identity? Words usually connote much more than they denote. And even if the Chinese terms translated as "slave" can be shown to possess in fact essentially the same meaning as our word, used precisely, the Chinese conceptual background of their terms must differ radically from our own today, as well as from that of peoples who practiced enslavement at other times and in other parts of the world. There is an overwhelming variety and dissimilarity in the aspects of that social institution which has been called "slavery," and which has been analyzed and described by ethnologists and historians. These disparities arise from differences in the social organization SOURCES AND DEFINITION 61 and economic system of the various societies that practiced it. Assuming that some core of identity does exist among various systems, the task of defining "slave" is so to emphasize the core that the "universal" qualities — the characteristics — will stand out. The extremes of difference are greatest among various primitive societies that have used the institution, and between slavery at a primitive cultural level and in more complex civilizations. In general, ethnologists have been much more interested than historians in defining the institution. H. J. Nieboer, who has concerned himself primarily with primitive slavery and has given the question of definition very shrewd analysis, first presents a definition of slave in the popular sense of the term as "a man who is the property of another, politically and socially at a lower level than the mass of the people, and performing compulsory labour." This he sharpens to "a man who is the property of another man, and forced to work for him." Finally he concludes that "slavery is the fact that one man is the property or possession of another." ^ Professor W. L. Wester- mann has given a somewhat similar definition: "Slavery is a system under which some human beings are chattels. Where this funda- mental legal and social fact does not exist another relationship between human groups has arisen which is not slavery." - The following dictionary definitions of slave in the specific rather than derivative sense are still not very precise: "One who is the property of, and entirely subject to, another person, whether by capture, purchase, or birth; a servant completely divested of freedom and personal rights." ^ "A person who is the chattel or property of another and is wholly subject to his will; a bond-servant; a serf." ^ "A person held in bondage to another; one held as a chattel; one whose person and services are under the control of another as owner or master; a thrall; a bondsman." In distinguishing between slave and serf: "A slave is the absolute property of his master and may be sold at will." '" 1 H. J. Nieboer, Slavery as an industrial system: Ethnological researches. 2nd ed., The Hague, 1912, pp. 5-9. - William Linn Westermann, "Athenaeus and the slaves of Athens," Harvard Studies in Classical Philology, Special Volume, Cambridge, 1941, p. 452, footnote 2. ' [The Oxford English dictionary], A new English dictionary on historical principles, vol. IX, pt. 1, Oxford, 1919, p. 182. * The Century dictionary and cyclopedia, rev. ed.. New York, 1911 (vol. IX), p. 5687. '" Webster's new international dictionary .... 2nd ed., Springfield, Mass., 1935, p. 2361. 62 SLAVERY IN THE FORMER HAN DYNASTY Finally, it might be useful to note a recent definition formulated for international agreement. Article 1 of the Slavery Convention, Geneva, September 25, 1926, states: "For the purpose of the present convention, the following defini- tions are agreed upon: (1) Slavery is the status or condition of a person over whom any or all of the powers attaching to the right of ownership are exercised." ^ Most definitions stress ownership of a human being, but some emphasize control of his services.- Ownership, however, is not necessarily vested in an individual, as some of these definitions imply. It may be vested variously in the state, in a group of individuals such as a family or tribe, or in some such organization or corporate body as a guild, lodge, temple, monastery, or stock company. Slaves must be in the category of property, however the society conceives it, and the abstract right to own humans must in some way be validated by that society. The right may be legally codified or simply recognized generally by customary usage. Individual owner- ship of individual slaves will be authorized in either way depending upon the society involved. A pragmatic test of true ownership is the legally or socially recognized right of an owner to transfer his slaves to another owner outright, by sale, barter, gift, or in some other way. This right may be shown to exist by concrete evidences of such transfer openlj^ and legally performed. This is a positive test, but it cannot be applied universally because certain societies forbid transfer of certain types of slaves who are otherwise owned as property. The right to destroy (enjoyed in regard to some sorts of property) is not an essential characteristic of slave-owning, being a good example of those attributes of slavery in which cultures vary. Control of services derives from ownership. It is secondary and less precise. Depending upon the culture involved it may be very ^ American journal of international law, vol. 21, Supplement: Official docu- ments, 1927, p. 174. Also International conciliation, Documents for the year 1928, No. 236, p. 13. The other definition is of the slave trade. 2 For example, Edward Westermark (The origin and development of the moral idea, 2 vols., London, 1906, vol. 1, pp. 670-671) takes the compulsory nature of the slave's relation to his master to be the chief characteristic. The nearest he comes to a definition is to say: "Slavery is essentially an industrial institution, which implies compulsory labour beyond the limits of family relations. The master has the right to avail himself of the working power of his slave, without previous agreement on the part of the latter." The weakness of this definition is that by emphasizing compulsory labor alone it does not exclude convict, corvee, or other types of forced labor which cannot be classed as slavery if that term is to have precision. SOURCES AND DEFINITION 63 mild or very strict. It involves the right of the owner to regulate the habits of his slave and to command performance of duties and separate acts, even against the slave's will, but only within the limits allowed by law or sanctioned by customary usage. This right to regulate and command implies the further right of enforcement by punishment or penalty imposed by the owner, or, for him, by some agency such as the state. Customary usage or law denies, however, the right of the owner to order the slave to do what is considered injurious to other people's property, other members of society, or society in general. Often the restriction extends also to acts injurious to the slave, and the right of punishment and penalty may be confined — but these latter limitations are not essential elements. A second characteristic is inherent and may be mentioned as the clue to the distinction between slaves and other groups in each society. It is the recognition in custom and/or law of a particular and distinct status for an individual because he is property.^ Against the background of this discussion the following definition of "slave" is offered tentatively: A slave is a person who is owned as actual property hy another person, group, corporation, or the state, whose services are therefore controlled, and who is accorded a distinct status as one of a group so owned and controlled. From this definition, that of "slavery" in reference to the individual, follows as: The condition of a slave; the fact of being a slave; and in reference to the phenomenon : The fact of slaves existing as a class in the community. There remains the second part of the question: Is the Chinese meaning of the terms translated as "slave" close enough to our meaning to approach identity? Chinese students of slavery in the Han period often give lists of the terms which they believe denote slavery, or rather, which fall within the Chinese expression nu-li 1 Certain discriminations, perhaps not adequately covered by this discussion, may be suggested. The term "serf" is usually distinguished from "slave" Ijy the fact that a master can require from his serf only legally specified services or dues, and by the fact that serfs are bound to the land and not to other men, and cannot be sold away from the land on which they work. Indentured ser- vants or bondsmen are distinguished by a closer definition of control, an agreed limit to period of service, and usually by restrictions on the right of transfer. Convicts for life should be distinguished even though they are completely subju- gated to the state, unless no distinction in status is made between them and slaves, or unless they may be transferred outright to another owner. In societies which conceive or formulate the relationship of parents to children (or of husbands to wives) as ownership, including right of transfer, the children would not be the slaves of their parents because the status group "children" is too broad, and because the status of an individual child would not derive from ownership but the other way around, ownership deriving from kinship. Other analogous situa- tions can be distinguished by some reflection. See, for example, "Slavery distin- guished from similar phenomena" in the article Slavery (Primitive), Encyclopedia of religion and ethics, vol. XI, p. 596. This is based upon Nieboer, op. cit., pp. 9 ff. 64 SLAVERY IN THE FORMER HAN DYNASTY chih-tu '^'^MM'^ Usually these lists include several terms which do not prove on close inspection to mean "slaves" as opposed to other groups in society; indeed, Chinese \\Titers dispute among them- selves which terms are admissible. There would appear to be two methods of establishing the basic terminology for the Former Han period: one, to consult Chinese dictionaries; the other, to observe the ways in which the terms are used contextually. The first has the disadvantage that the conceptions of dictionary compilers inter\'ene as extra, uncertain factors between the terms and the investigator. Furthermore, various senses are usually explained by use of synonyms which cannot be exactly identical, and which themselves have to be correlated with some imperfect English equivalent. Studying terms contextually, on the other hand, is comparable to the use of a simple algebraic formula in which the value of an unknown factor, X, is determined by its relation to several known factors. Terms suspected of being equivalent to the word "slave" are thus defined by their context. Neither method escapes the inherent uncertainties of translation, but the second involves fewer filtrations and allows us to search out Han period meanings from contemporary texts. Wherever possible, therefore, the second method has been employed. Slave Termlnology of the Former Han Period NU& Male Slave. — A few contextual evidences of males being sold and becoming 7iu, or of nu being sold, are as follows. Before 207 B.C.: "[Luan] Pu was kidnaped by someone and sold as a nu at Yen . . . ." (5.)- In 202 B.C.: "Chu Chia recognized him to be Chi Pu. He bought him and put him in a house among the fields. IThen he 1 The distinction is worth making although the expression may merely be a term coined to translate the Occidental concept. I do not know how early it appears in native literature. For examples of such lists see Liang Ch'i-ch'ao ("Chung-kuo nu-li chih-tu (System of slavery' in China)," CHHP, vol. 2, 1925, pp. 527-553 [hereafter cited by its English title,;, pp. 527-528), Ma Fei-pai ("Ch'in Han rhing-chi shih tzu liao [Source material on the economic history of Ch'in and Hani," pt. 6, "nu-li chih-tu [The slavery' system], " Shih Huo, vol. 3, No. 8, March 16, 1936, pp. 385-400 [hereafter cited by its translated title], pp. 585-586), Ma Ch'eng-feng (Chung-kuo ching-chi shih [An economic history of China], 2nd ed., vol. 2, Shanghai, 1939 [hereafter cited by translated title], pp. 246-247). See also Wu Ching-ch'ao, "Hsi Han nu-li chih-tu [The slavery system of the Western Han]," Shih Huo, vol. 2, No. 6, Aug. 16, 1935, pp. 264-270 (hereafter cited by translated title), p. 246. = Hereafter numbers in itaUcs refer to the translated documents numbered consecutively in Part II. In most cases the whole text and the footnotes add considerably to the part cited or quoted. SOURCES AND DEFINITION 65 enjoined his son, saying: 'In the field work be lenient with this nu; you must eat together.']" {8.) From a contract dated 59 B.C.: "... the gentleman Wang Tzu-yiian, of Tzu-chung, purchases from the lady Yang Hui of An-chih village in Chengtu, the bearded nu, Pien-liao, of her husband's household. The fLxed sale [price] is fifteen thousand [cash]. The nu shall obey orders about all kinds of work and may not argue." {83.) From an excavated fragment from Chii-yen, date uncertain: "Two young nu, price thirty thousand [cash]; a grown pei, price twenty thousand [cash]." * These items fulfill the pragmatic test of ownership. Slave. — Ver>' rarely nu seems to have a generic sense of "slave" rather than the specific sense of "male slave." The generic sense is usually supplied by the combination nu-pei. One example is from the "Treatise on Jurisprudence," in its description of criminal laws of the Chou period, and is also found in the present Chou U, of uncertain date and authorship: "As to nu, males went into criminal ser\-itude; females went into pounding dried grain. Xo one who had noble title, or was seventy, or had not yet lost the milk teeth became a nu." (i.)'- Other Uses. — Throughout Han texts, nu appears as part of the name of the people on China's northern frontier, the Hsiung-«?/. Sometimes nu clearly stands for them. Elsewhere it appears in place names, such as Lu-/^z/. and in the names or appellations of people, such as Chao Fo-nu, "Chao, the vanquisher of the [Hsiung]- nu." Such uses are, of course, not included in this study. I have found no other uses of nu in Han texts. Aside from these there is no evidence that nu signifies anything but "male slave" or "slave," actually or figuratively. PEI W- Female Slave. — Date 10-S B.C.: "[Wang Mang] once privately bought a serving pei. Some of his cousins heard rumors of it. 1 Cited by Lao Kan, op. cit., p. 2. = The Shuo wen chieh tzu, the great etymological dictionary presented to the throne in a.d. 121, in its definition of nu, says, "Xu-pci were all criminals in antiquity (. . . ■^'6"^^A.)i" and then quotes this citation from the Chou li. Among the various editions of the Shuo wen assembled in the Shuo wen chieh tzu ku lin, the most important for information on the term nu (pp. 5554-56) is that cited in abbre\'iation as ^1^ (the Shuo wen chieh tzu i cheng), which has nineteen double columns of quotations from early dictionaries, histories, and encyclopaedias, including some passages quoted above. It is a mine of lore from the Chi chin p'ien, Feng ssu t'ung, Ch'u hsiieh chi, and other early works, and is recommended as an example of the definitions of key slaverj' terms in the numerous Chinese encyclopaedias. 66 SLAVERY IN THE FORMER HAN DYNASTY [Wang] Mang thereupon said: 'The General of the Rear, Chu Tzu- yiian, has no sons . . . and so I bought her for him.' He immediately presented the pei to [Chu] Tzu-yua,n." (108.) 3 B.C.: "...the Empress Dowager [nee] Fu sent an internuncio to buy pei from various government bureaus, taking them at a low price; and she also took eight pei from the Bureau of the Chief of the Palace Guard. [Mu-chiang] Lung memorialized, saying: 'The price is too low. Please readjust it.' " (116.) Earlier evidences of sale are given below. Other Uses.— There are no uses of pei in place names, personal names, or tribal names so far as I have discovered. Prior to Han times the term pei was also used figuratively by women in speaking of themselves, but this has not been noted in texts referring to the Former Han period. NU-PEI Male and Female Slaves. — References to the compound term are numerous; herewith are only a few dealing with sale. In 202 B.C., an edict of Emperor Kao: "Those common people who because of famine have sold themselves to be people's nu-pei are all to be freed and become commoners." (9.) Ca. 120 B.C.: "[Ho] Ch'ii- ping liberally bought fields, houses, and nu-pei for [Ho] Chung-ju, and then left." (J(.2.) Because of a law proposed in 7 B.C. to limit the amount of farm land and nu-pei that might be owned by people in various classes "the price of fields, residences, and nu-pei depre- ciated." (109.) Quoting the inaugural edict of Wang Mang, A.D. 9: "Furthermore, [the Ch'in dynasty] established markets for nu-pei [putting humans into] the same pens with cattle and horses." Wang Mang proclaimed that neither land nor nu-pei could be bought or sold. As a result, "the people went so far as to weep in the markets and highways. Moreover, those who were tried for selling and buying fields, residences, and nu-pei . . . were in- numerable." (122.) Slaves. — There are no cases where a male is called a pei and only the one case quoted above (1) where females are called nu, in the generic sense. Wherever sex is shown males are nu, females pei. Therefore the term nu-pei is usually translated "male and female slaves." However, it probably has the generic meaning "slaves" in many instances where the reference is general. The legal and customary status of nu, pei, and nu-pei is an involved problem dealt with in considerable detail later in this work. SOURCES AND DEFINITION 67 T'UNG it or fti Youths. — This term is frequently used in the sense of "slave," but also merely in the sense of "child," especially "young boy." Therefore, it is always translated "youth," though usually it is con- sidered to mean "slave" in the passages accepted. It has the generic sense of slaves, both male and female, and also a specific sense, male, and sometimes, female. Ca. 200 B.C. : "Some of the people of Pa and Shu went out clandestinely for trade, taking their horses from Tse, t'ung and yaks from P'o; and because of this [trade] the people of Pa and Shu became prosperous and wealthy." (10.) 177-174 B.C.: "Nowadays people who sell t'ung dress them up in embroidered clothes and silken shoes with the edges all embellished, and put them into pens." (16.) 113 B.C.: "With many attendants [she intends] to go to Ch'ang-an [where they] will be made captives and sold to become t'ung-nu . . . ." (50.) Wang Pao's essay on the purchase of a nu, dated 59 B.C. and using nu several times, is entitled "The Contract for a T'ung." (83.) Meng K'ang, who lived ca. A.D. 180-260, says, "T'ung are nu-pei." (53, footnote 5.) Youth (male). — When Chi Pu was disguised as a slave he was sent to be sold with several tens of household t'ung, who were prob- ably males also (8). Likewise t'ung used as cavalry escorts were probably males (81, 112), this being a regular function of nu, as described in chapter VIII, below. When the expression "youth horsemen" arises. Yen Shih-ku, A.D. 581-645, explains it as "t'ung-nu horsemen" (112, footnote 7); again, "[They] made cavalry men of their t'ung-nu." (81, footnote 2.) Used thus in conjunction with nu, the term may be adjectival, or merely a compound. Shih Tan's wealth and extravagance during the period 33-15 B.C. are described: ". . . and t'ung-nu numbered by the hundred." (97.) The extrava- gance of Wang Mang's uncles is similarly noted: "In their women's quarters each had several tens of concubines, and their t'ung-nu were numbered by the thousand or hundred." (99.)- 1 The second form seems to be the original, and prior to Han times some distinction was made. In extant literature of the period the forms are used inter- changeably. The first occurs in fourteen of the translated documents, the second in six; in documents 2, lf.9, and 53, CHS uses the second while the equivalent SC passage uses the first. The Shuo wen makes only the second form correct in this sense. See footnote 2. 2 The Shuo wen (op. cit., pp. 1111-12), defining Vung in the second form (above), says: "Males who had undergone criminal punishment were called nu, nu were called t'ung, females were called ch'ieh j^^^'^H^^Hml^H^." Some commentators correct this: "Males who had undergone criminal punishment and become nu were called Vung |^^f|:^. ..." I have not found any Former Han dynasty cases of Vung specified as government slaves because of crime. 68 SLAVERY IN THE FORMER HAN DYNASTY Youth (female). — Wei Ch'ing's mother was a fung, and Yen Shih-ku says, "The word fung is a general term for pei-concubines." (26, and footnote 5.) Wang Mang's wife dressed so simply that other women calling on her mother-in-law, in 8 B.C., thought she was a fung or a servant (A17). TSANG ^ and HUO m These terms have a variety of meanings, even in reference to slavery, and are not translated. The first more often refers to males and the second to females, but the terms can be a compound referring to males, or to slaves generically. The terms appear in only one document which assembles the various definitions (11, and foot- notes). Meanings are here listed without attempt to unravel their interrelationships. (1) In southern and eastern China (of the Former Han period or earlier) "when cursing a nu, one says 'tsang'; when cursing a pei, one says 'huo.' " (2) In the northeast, "all plebeian males who mate with pei are called tsang; [plebeian] females who become wives of nu are called huo," This use may not mean slave; the status of the plebeians is not made clear. (3) Runaway nu are called tsang; runaway pei are called huo. The author of the Fang yen then summarizes by saying: "They are all abusive terms of diverse regions for cursing nu-pei." (4) Chin Shao (fl. ca. a.d. 275) is quoted as having said: "Tsang- huo are those defeated enemy who have been captured and made nu-li." (5) Wei Shao (a.d. 204-273) is quoted as having said: "When a good man takes a pei as wife and she bears a child, [the child] is called a huo; when a nu takes a good woman as wife and she bears a child, [the child] is called a tsang." This reverses the termi- nology for plebeians given in (2), but follows the sense of (1) and (3) that nu (here child of a nu) is a tsang, and pei (here child of a pei) is a huo. (6) Finally the Feng su fung i, supposedly by Ying Shao (ca. A.D. 140-206), is quoted as having said: "In the ancient institutes there were originally no nu-pei, and then those who committed offenses were the origin of it. Tsang-che are those who have under- gone the punishment of tsang, being seized and becoming govern- ment nu [a variant here has a nu-pei]. Huo-che are runaways who SOURCES AND DEFINITION 69 have been recaptured and become pei [variant, nu-pei]." The second part of this is a repetition of the second part of (3). Several terms are used to refer to special kinds of nu or pei, but they do not occur frequently in the basic texts. Definition of these comes, in the main, from direct statements by writers of the Latter Han period, or from commentators, rather than from con- textual evidence. TS'ANG-T'OU i^m "Green-head." — Wei Hung, who was active between a.d. 25 and 57, and who wrote the Han chiu i, on governmental practices during the Former Han epoch, says: "Government nu were selected to give [service] as writers and accountants. Those of [the rank of] Attach^ and below were ts'ang-t'ou, [wearing] blue-green turbans." (92.) A memorial dated ca. 3 B.C. says: "That ts'ang-t'ou and lu-erh should all be employed [as officials] and become rich, is not Heaven's intention." (118.) On this, Meng K'ang remarks: ". . . [people of the] Han period named nu, 'ts'ang-t'ou,' not pure black, in order to differentiate them from good people." The term was also applied to private slaves, as indicated by the report that when Ho Ylin, grandnephew of Ho Kuang, should have been attend- ing court he preferred "sending a ts'ang-t'ou nu up to court to pay the visit [in his stead]." (72.) The term ts'ang-t'ou was used shortly before Han times to desig- nate members of private armies who wore green kerchiefs or turbans around their heads. During Latter Han times it was apparently used quite specifically for male slaves. LU-ERH it^ "Hut-dweller." — This term is met infrequently, and, like ts'ang- t'ou, is a sort of sobriquet. Lu were the houses for servants and minor attendants in the imperial palace. In the commentary just cited, Meng K'ang adds: "The place where all those who served in the halls [of a palace or residency] lived was a 'hut' (lu). Ts'ang-t'ou who were serving attendants were therefore called lu-erh." {118, footnote 4.) KUNG-JEN ^A Palace-women.— Wei Hung twice speaks of kung-jen as a t5T)e of pei. "Kung-jen were selected from among palace pei in their eighth year or over. They waited upon [ladies of the palace from] the Empress on down." (91.) Also, "In the Inner Apartments [of the 70 SLAVERY IN THE FORMER HAN DYNASTY palace], Maidservants and Orderlies were all government pei selected in their eighth year or over. They dressed in green and were called kung-jen." {92.) Chao Fei-yen, the Empress of Emperor Ch'eng, "was originally a kung-jen of Ch'ang-an." Yen Shih-ku remarks about her, "Originally a kung-jen, she was presented to the household of the Princess of Yang-a. 'Kung-jen' was the name for government pei who were servants in the forbidden parts of the palaces." {101, and footnote 3.) However, kung-jen were not exclusively slaves, and there is one reference, dated 113 B.C., to a kung-jen in a king's palace, who was a man. This was Luan Ta, the magician {91, footnote 1). There is a group of words regularly used to modify the terms nu, pei, t'ung, and sometimes nu-pei. Kuan ^, "government"; ssu fi., "private"; kung ^, "palace"; chia ^, "household" or "family"; ta i^, "senior" or "elder"; chH t^x, "cavalry"; ts'ung ^, "attendant"; shih ^, "serving"; yii ^, "personal" ; /w #, "chamberlain"; t'u ^, "convict." A few of these terms occasionally stand alone, in combination, or with suffixes, in place of the terms they regularly modify. SHIH f# Serving. — Shih is used to modify pei in documents 79, 82, 108, and A20, the last dated about a.d. 9. Yiian Ang's Attendant Secretary, some time between 165 and 157 B.C., "had secret relations with [Yiian] Ang's shih-child." Ylian Ang "then presented him with the shih-che." After the first reference the commentator Wen Ying (fi. ca. a.d. 196-220) says definitely that she was a pei {23, and footnote 3). Some time between 164 and 154 B.C. a King summoned a physician to examine "all his girls and shih-che." The King said, "... I bought them in the common people's [market] place." Be- cause the King did not believe the diagnosis of one girl's condition, "he did not sell her at the [market] place for nobles." {21.) About 142 B.C. Wei Ch'ing's sister, the daughter of a household t'ung, was a shih-che in the same household {29). The following account equates shih-che with yii-che and yil-pei contextually. (The latter terms are discussed next.) About 125 B.C. the Queen of Heng-shan "had a shih-che, a fine dancer, to whom the King had granted his favors." She arranged for her step-son, named Hsiao, to have relations with the girl, and his brother told the King, "Hsiao had relations with the King's yii-che . . . ." Later, "Hsiao was tried for having had relations with the King's yil-pei," and was executed {37). SOURCES AND DEFINITION 71 Personal. — Yii modifies pei in documents 37, Jl^-S, and 85. In 82 B.C. Shang-kuan An, when drunk, would "have incestuous relations with his stepmother and various of his father's Ladies and shih-yii." Yen Shih-ku here makes the important statement, "The shih-yii were at the same time pei." {59, and footnote 3.) In 115 B.C. a marquis "was tried for having had relations with his father's yii-pei, and killed himself." The terse account of this trial in the Table of Marquises, CHS, 16, 3a, uses exactly the same words, but drops the one word pei, as though it were unessential; it appears, however, in the equivalent SC, 18, 3b {^8, and footnote 2). Chamberlain.— Fu occurs as a modifier of pei in documents 76 and 120, and much more frequently in the History of the Latter Han dynasty. Yen Shih-ku explains the term by saying, "Whenever it says fu-pei it means [one who] assists with the affairs of [her master's] clothes and bed." (120, footnote 2.) In only one instance has the term been found standing alone — in the quoted accusation that Wang Shang had "had intercourse with his father's /m." There Yen Shih-ku says "fu means fu-pei." (98, and footnote 5.) Instances in which these modifying terms stand alone, with suffixes, or in combination, in place of the word they modify are relatively rare. There is no way of telling contextually that they always mean slave. Therefore a few indeterminate references which add nothing to the body of information are not included among the documents in Part II. T'U^ Convict.— In several documents t'u precedes nu or nu-pei in such a way that it might be a modifier. T'u occurs innumerable times by itself. Several Chinese writers include it among component terms for their studies of Chinese (or Han period) slavery, while others use it without special comment. There appears to be no contextual evidence for the Former Han period that t'u were sold, bartered, given away, or in any other way transferred in a manner that would fulfill the pragmatic test of ownership. While t'u were indeed subjugated to the full extent of the word, there seems to have been a distinction in status between t'u, on the one hand, and nu, pei, or t'ung, on the other. This, however, is an involved subject, best treated in its appropriate place in the next chapter. III. ENSLAVEMENT When the Former Han period was at the height of its luxury- there were several hundred thousands to perhaps a million real slaves in China. Many of them passed in various ways from freedom into bondage. What were the methods by which free people became slaves? Which led to government slavery and which to private? Free people were enslaved for crime, sold because of economic distress, forced into bondage illegally, and imported from foreign regions for sale. Some prisoners of war were perhaps enslaved, but this is a complex question reserved for special discussion in the next chapter. Crime led always to government slavery, while economic distress and illegal bondage produced private slaves by the first sale. Both the government and private individuals acquired im- ported slaves. Enslavement of Criminals and Their Families Enslavement of criminals and the families of men executed for treason and rebellion has a long history as a typical Chinese mode of punishment. Many writers of the Han period state that this form of punishment was in use during late Chou and Ch'in times. The famous scholar and legal authority Cheng Hsiian (A.D. 127-200) explains in his commentary to the Chou-li that in ancient times males and females drawn into trial (i.e. related to people tried) were seized^ by the government as slaves; also, that contemporary male and female slaves were the descendants of, or the same as criminals of antiquity. His contemporary, Kao Yu, remarked in a commentary to the Ch'in dynasty book, the Lm shih ch'un ch'iu, that fathers and older brothers of criminals were tried and seized as slaves. The great dictionary Shuo wen, produced around A.D. 121, says that males who underwent criminal punishment were called male slaves (nu), while females who underwent criminal punish- ment were called female slaves {pei).~ 1 The term ^xA. literally "submerged into," is hereafter translated as "confiscated" when it refers to property, and as "seized" or "seized and enslaved" when it refers to people previously free. 2 There are several other similar references but these most definitely refer to pre-Han conditions, and illustrate the point that Han scholars believed criminals or members of their families were enslaved in pre-Han times. These commentators were men of real scholarship, having access to sources now lost, conversant with customs and laws which developed out of earlier ones, and using a language in 72 ENSLAVEMENT 73 From the end of the Han period until late into Ch'ing times laws specified enslavement as punishment for families of people guilty of crimes classed as treason and rebellion. Legal codes changed, the classes of crimes subject to this form of punishment varied, and occasional attempts to abrogate the system occurred. Perhaps it is impossible to prove for everj^ period that such laws were enforced, but the principle remained a basic feature of Chinese law during most of its recorded history. ^ A basic social and philosophic principle of family unity and mutual responsibility underlay this practice of enslaving a major criminal's relatives: the whole family was considered responsible for the acts of one of its members. The Chinese family was much larger than the simple marriage group of husband, wife, and children, and therefore it is hard to tell how many living generations and what degrees of relationship comprised the legally responsible family in each criminal situation. The gravity of the offense determined the number of family members and the classes of relatives held responsible. which legal terminology was still fairly close to that of late Chou and Ch'in times. In the works which they studied certain passages appear to refer to enslavement of criminals, but contextual evidence alone does not prove that they do. Since it is necessary anyway to fall back upon the explanations of the commentators, it is more objective to quote what they said than to try independently to interpret ambiguous passages in works bristling with textual and historical problems. The above quotations and others like them are conveniently assembled in Ch'eng Shu-te, Chiu ch'ao lii k'ao (Han lii k'ao, ch. 3), pp. 81-82. Cf. also Shen Chia-pen, Li-tai hsing-fa k'ao [An investigation into the history of the laws and punishments], "Fen k'ao," ch. 15, 3a-5b (hereafter cited by Chinese title); and YUan chien lei han, ch. 258 (Jen pu, ch. 17), 2a-b, and 16a. 1 For citations of the law, modifications of it, or examples of its practice during the period between Han and T'ang, cf. Ch'eng Shu-te, op. cit., pp. 239, 246, 271, 294, 305, 369, 406, 438, 471, and 520. For a discussion of the continuity of the law during T'ang and later periods, cf. Wang Shih-chieh, "Chung-kuo nu-pei chih-tu [The Chinese slavery system]," She-hui k'e-hsUeh chi-kan (Social Science Quarterly), vol. 3, 1925, pp. 303-328 (see pp. 307-308) (hereafter cited by translated title. This article has been translated by Toni Pippon, "Beitrag zum Chinesischen Sklavensystem," Mitteilungen der Deutschen Gesellschaft fiir Natur- und Volker- kujide Ostasiens, Bd. 29, Teil B, Tokyo, 1936, pp. 93-113). The law on enslavement in the T'ang Hi su i, ch. 17, section on thieving and robbery, may be translated as follows: "Those plotting rebellion or major crimes shall all be beheaded. Fathers, and sons over sixteen years old, shall be strangled. [Sons] fifteen years old or younger, mothers, daughters, wives, concubines, sons' wives and concubines, grandfathers, grandsons, older and younger brothers, older and younger sisters, and such others as pu ch'il [shall be enslaved], and property, fields, and houses shall be confiscated by the government. Men over eighty or incurably sick, women over sixty or incurably sick, shall all be excused." (Cf . Pippon, p. 100.) The T'ang lii su i is the earliest extant Chinese law code (or, more precisely, a commentary on a code now lost) and dates from a.d. 653-654 (cf. Paul Pelliot, "Notes de bibliography chinois II: Le droit chinois," BEFEO, vol. 9, 1909, pp. 123-152 [see pp. 124-125] and Jean Escarra: Le droit chinois, pp. 96-97). Various 74 SLAVERY IN THE FORMER HAN DYNASTY Some crimes, for example, were punished by executing three sets of relatives. Commentators disagree in their explanations of "the three sets of relatives." Chang Yen of the third century says that they were the criminal's parents, brothers, wife, and children. Ju Shun (fl. ca. 189-265) says that they were the members of his father's clan, mother's clan, and wife's clan. Other suggestions give his father, sons, and grandsons; or his father and father's brothers (including male cousins?), his own brothers (including male cousins?), and his sons and their brothers (i.e., cousins).^ Theoretically this punishment was abolished by an imperial edict of the Dowager Empress nee Lii in 187 B.C., yet in 164 B.C. when Hsin-yiian P'ing plotted rebellion he was exterminated, to- gether with his three sets of relatives. Again in 104 B.C. the relatives of Wang Wen-shu were tried and executed, probably to the jive degrees of relationship. ^ It is not unusual to read that a punishment was abolished and then to learn that it was practiced only a short time later. Emperor Wen abrogated all statutes and orders for arresting wives and children of criminals and punishing them also. Yet an apparently contradictory event occurred under his successor some twenty-five years later, when Chi K'uei-yiieh conspired to revolt and kill his father, Chi T'ung-chia. This was pronounced to be "treason and inhumanity." Ju Shun says that according to the Code in cases of "treason and inhumanity" the father, mother, wife, children, brothers, and sisters of the criminal should all be publicly executed. In spite of Emperor Wen's abrogation. Emperor Ching especially pardoned the father, together with his wife and children, "who should have been con- demned with him," and ordered that Chi K'uei-ylieh, together with his wife and children, should be sentenced "according to the law." ^ parts of it have been translated by R. Deloustal in his "La justice dans I'ancien Annam," BEFEO, 1909-13, and 1919-20, passim, which translates an Annamite code based upon, and in many ways similar to, the T'ang code. It may be pointed out here that the T'ang lii su i contains scores, if not hundreds, of references to slaves, either in special laws applicable to them, or in modifications of laws for cases in which slaves are involved. It is invaluable for a legal study of slavery in T'ang times. This multipUcation of slave legislation is in marked contrast to the paucity of (extant) similar legislation before T'ang times, especially during the Han period. 1 Cf. Ch'eng Shu-te, op. cit., p. 59, and Tz'u yiian. 2 HFHD, vol. I, pp. 193 and 260; for Wang Wen-shu see CHS, 90, 4b. Ch'eng Shu-te (op. cit., p. 119) quotes the "Hsing fa chih" of the Chin shu as saying that Hsiao Ho abolished the punishment of three sets of relatives and those drawn into trial {lien tso) when he took over the Ch'in code. 3 CHS, 4, 2b; and 5, 2a. Cf. HFHD, vol. I, pp. 233 and 313, respectively. ENSLAVEMENT 75 If whole families, including remote relatives of people guilty of particularly heinous crimes, might be executed, it is obvious that their enslavement was comparatively a mild form of punishment. The History of the Former Han dynasty records only a few instances of this form of enslavement, but they involved considerable numbers of people. Furthermore, they are probably typical of a constant series of criminal cases which were either not important enough to be reported, or were reported in such meager detail that the fate of the families was not specified beyond the ambiguous remark that they were punished "according to the law," The families of the leaders in the Rebellion of the Seven States in 154 B.C. were enslaved. This is the first case recorded for the Han period. It illustrates how the formal histories often lack crucial details for a study of slavery, for it is almost an accident that this instance is on record at all, though the rebellion was the most im- portant one that occurred during the dynasty proper. Although it was quickly suppressed, thousands died on both sides, one of the kings was beheaded, and the others committed suicide.^ The biographies of the seven rebellious kings do not indicate what happened to their families. The "Table of the Vassal Kings" shows in each case only that the ruler expired and that the royal line was terminated. Emperor Ching's edict in the "Annals" refers to the extermination of the kings, but also neglects to mention what happened to the families. It only tells that the Emperor "could not bear to apply the law" to Liu Yi (the son of a former King of Ch'u, and an uncle of the rebelling King) and some other imperial clansmen who had joined in the rebellion. Instead, he ordered all their names expunged from the register of the imperial house. He pardoned officials and people who had been coerced to join in the rebellion, although they should have been sentenced as accomplices. 2 Thus, no record of the events at that time tells what happened to the families of the leading rebels. Yet fourteen years later Emperor Wu "pardoned the families of [the leaders of the rebellion of] Wu, Ch'u, the Seven States, who had been condemned to the govern- ment." Ying Shao says explicitly in his commentary: "At the time of the rebellion . . . the wives and children of the leaders had been seized and made government slaves. Emperor Wu, pitying them, pardoned them and sent them all away." {30, and footnote 3.) 1 For a clear summary, cf. HFHD, vol. I, pp. 292-297. =! Biographies: CHS, 35, 7a; 36, 2a; 38, 2a and 4b; also SC, 106. Tables: CHS, 14, 3b-8a, passim. Edict of Emperor Ching: CHS, 5, 2b; cf. HFHD, vol. I, pp. 314-315. 76 SLAVERY IN THE FORMER HAN DYNASTY There is no reason to doubt that this amnesty occurred as recorded. Living in the century after the Ch'ien Han shu was written, Ying Shao simply amphfied and explained the statement. Chinese encyclopaedias and modern writers on slavery uniformly cite this early example of enslavement. For fourteen years these wives and children were government slaves, but if Emperor Wu had not pardoned them, or if the ten-word record had not been entered in his "Annals," it might not now be known at all that these people became government slaves. How often did the historians fail to record such enslavements in the appropriate places? Chapters 14 through 18 of the Ch'ien Han shu list in tabular form all the important noble families of the Han dynasty, with brief records of succession to and termination of the lines. At least fifty- three of these noble families were abruptly cut off because the title- holder rebelled or plotted to rebel, and was executed or committed suicide. Eliminating the leaders of the Rebellion, and the members of the Empress nee Lii's clan, which was entirely exterminated "without consideration of youth or old age," there are still some thirty other cases of rebellion in which there is no hint about the fate of the rebels' wives and children. It is reasonable to suppose that in some of these cases also the families became government slaves. A special case of enslavement of families occurred in 120 B.C. when Wang Wen-shu, one of Emperor Wu's most ruthless officials, was appointed Administrator of Ho-nei Commandery. On taking office he immediately arrested more than a thousand families among the tjrannical gentry of the commandery. "Tyrannical gentry" (^^f" or ^?^) usually means powerful plebeians who oppressed their neighbors, appropriated property, and accumulated large amounts of farm land which they rented. They were the economically powerful people who manipulated or defied the laws and were the actual bosses of their communities. Emperor Wu had a policy of suppressing these people. He quickly consented when Wang Wen- shu requested permission to execute the major criminals and their families and to enslave the families of the lesser offenders, who would be executed singly. Half the property was to be given to the victims of their oppression, while the government kept the rest. Here is a clear-cut example of enslavement as an alternative for execution, and a punishment one degree less severe. The passage closes with the grim sentence, "Blood flowed over ten or more li." US.) Only one document definitely reports enslavement of criminals themselves during the Former Han period. This was during the reign ENSLAVEMENT 77 of the usurper Wang Mang, in connection with his currency "reforms." Several times during his regency and reign, Wang Mang altered the monetary system, greatly to the disadvantage of all people who owned money. Many new types of coins were introduced, and their metal content was reduced. People were ordered to adopt and circulate new issues whenever Wang Mang altered the cur- rency, and were supposed to turn in their old bronze coins in exchange for new coins containing much less metal. There was tremendous opposition and many traders refused to accept the new coins at their face value. Almost anyone could coin money, for it was cast rather than struck. Therefore many people melted down their old coins and made new and lighter Wang Mang coins instead of turning in their old coins for exchange at great loss. Widespread counterfeiting greatly reduced the profits Wang Mang expected to obtain by depreciating the metal content of his money. ^ To enforce circulation of the new coins and to stamp out counter- feiting Wang Mang altered the monetary laws several times. In A.D. 9 he decreed that anyone who clung to the old five-chu cash or talked against the new currency would be banished to distant frontiers (122). Counterfeiters were executed. ^ Possibly their families were enslaved, for the next year Wang Mang increased the severity of his laws to include the five mutually responsible families (see p. 78) who would be tried along with the counterfeiter and be enslaved by the government (123). When these stringent measures did not stop the evil, he changed and lightened the law. Instead of being executed, those who privately coined money were to be seized with their wives and children and made government slaves; and officials or groups of mutually respon- sible five families who knew of such counterfeiting and failed to report it were to be punished with them. Instead of banishing people who opposed the new money, he sentenced plebeians to one year of punishment, and dismissed officials from office (131). Leniency was less effective than severity. Offenders were even more numerous. In A.D. 21 a great frost destroyed crops, and in 1 For a clear and detailed summary of Wang Mang's currency reforms, their disastrous economic effects, and the resulting wide-scale counterfeiting, cf . Homer H. Dubs, "Wang Mang and his economic reforms," TP, vol. 35, 1940, pp. 233-237. ' I have not found a specific Wang Mang decree that condemned counter- feiters to execution, but CHS, 24B, 11a, reports that the people who died for counterfeiting and were cast to the four bounds of the empire for opposing Wang Mang's new money could not be counted. Emperor Ching's law of public execu- tion for coining cash or making alchemistic counterfeit gold (cf. HFHD, vol. I, p. 323) was presumably still in effect in Wang Mang's time. 78 SLAVERY IN THE FORMER HAN DYNASTY consequence plebeians of eastern China resorted to counterfeiting. They and mutually responsible groups of families tried with them were seized and enslaved to the number of ten myriad. Males were transported in cage carts, while women and children walked with iron fetters dangling on their necks. When they arrived at Ch'ang-an these slaves were set to work in the Bureau of Mint {131, 132). An ironic touch, but thoroughly Chinese! The number reported enslaved need not be taken too literally, nor even perhaps the historian's grim statement that 60 or 70 per cent died of grief and suffering. What is noteworthy is that in this instance criminals, that is, the counterfeiters, were enslaved. Even more interesting are the references to enslavement of people not actually related to counterfeiters but held responsible for their crimes because they and the counterfeiters were members of the same groups of five families. Each family was held legally respon- sible for the conduct of four others — a great extension of the funda- mental principle whereby a family was responsible for the acts of its own members. Wang Mang did not originate this system. It had been advocated by Shang Yang in the fourth century before Christ, and may have originated earlier.^ Not merely a harsh legal measure to extend the terror of the law, it was a far-reaching system of political organi- zation based upon conditions of small, self-contained peasant hamlets and close-knit villages within towns. Schematically this administra- tive system consisted of five families in a neighborhood, five neighbor- hoods to a hamlet, five or ten hamlets in a commune, and ten communes to a district. A headman or elder in charge of each division was the focal point of administration and responsibility. Exact quinary or decimal units could not have been followed rigidly, but the principle was applied for purposes of taxation, corvee labor, military levies and self-defense, and the administration of laws. The state of Ch'in, and later the djniasty, charged each family in a neighborhood to denounce crimes committed by members of any of the other families; if certain crimes occurred all five families were implicated and tried together. The harsh application of that part of the system which made people responsible for the conduct of their neighbors was thoroughly repellent to the subjects of the Ch'in empire — intra-family respon- sibility was an integral part of Chinese social organization, but 1 Ch'eng Shu-te, op. cit., p. 82; J. J. L. Duyvendak, The hook of Lord Shang, p. 14 and footnote 7, pp. 57-59; Derk Bodde, China's first unifier, pp. 35, 166. ENSLAVEMENT 79 extension of legal responsibility beyond the family was apparently an artificial system imposed from above. The Han dynasty seems to have abandoned the objectionable part of the system while main- taining most of it for administrative purposes.^ Wang Mang merely re-established mutual legal responsibility among groups of five families in order to enforce his currency reforms. Thousands of people thereby became slaves in a.d. 21 because they or some of their neighbors were secretly coining money. Ten years later, after Wang Mang had been killed, the new Emperor Kuang-wu freed those who had survived the revolution and were still slaves. Writing finis to this mass enslavement, the imperial edict said: "Those officials and people who, during Wang Mang's time, were seized and became slaves not in accordance with the former laws, are all to be freed and made commoners." (135.) Orthodox enslavement of families of executed criminals (and perhaps of some criminals themselves) continued, however, during the Latter Han dynasty. An edict of a.d. 106, prepared for the infant Emperor Shang, said that many members of the Imperial House had been enslaved since a.d. 25, and ordered the unfortunates all to be freed and made commoners. In 110 Emperor An decreed: "All those who have been tried and banished to the frontiers for monstrous talking and other crimes since the Chien-ch'u [reign period (76-83)] are to be returned each to his original commandery; those seized by the government as slaves are to be pardoned and become commoners." ^ These passages are evidence of enslavement of criminals' families throughout the Han period. Any assertion about the frequency of the practice, or estimate of the number of people thus enslaved during the dynasty, would be without documentary foundation. Considering the nature of the historical sources it seems certain that there were many unreported cases. Such enslavement was always to the government in the first instance. 1 In 179 B.C. Emperor Wen "completely abrogated all statutes and orders for arresting wives and children and punishing them with [the criminals]." (CHS, 4, 2b; HFHD, vol. I, p. 233.) This certainly indicates that the extension of responsi- bility beyond the family or clan was already obsolete. ^ HHS, 4, 9a; 5, 4a. The wording of the edict of A.D. 110 leaves some doubt today whether the slaves (1) had been enslaved, rather than banished, for "mon- strous talking" and other crimes; whether they were (2) families of people banished; or whether they were (3) all those people seized and enslaved since the Chien- ch'u period for any reason. On "monstrous talking" cf. HFHD, vol. I, p. 193, footnote 2. 80 SLAVERY IN THE FORMER HAN DYNASTY Distinction Between Convicts and Slaves Several modern Chinese writers on Han slavery treat convicts {i!u ^) as though they were slaves.^ It is true that convicts were sentenced to hard labor, that they worked side by side with certain types of government slaves, and that both sometimes wore similar clothes, shackles, and marks of identity. But unless convicts were slaves in a legal sense, unless they were property of the government in the same way that slaves were, only confusion results from calling them slaves. Even at the risk of "beating the dead tiger" it is imperative to clear up this point. What are the facts? The earliest specific definition of t'u known to me is that by Wang Ch'ung, who lived a.d. 27 to ca. 97, and was a contemporary of Pan Ku, the author of the Ch'ien Han shu. He says: "[Those who] have undergone criminal punishment are called fu I^^JIi:^^." - This is so general that only equally general terms such as "convict" or "felon" convey the same meaning. The great etymological dictionary Shuo wen chieh tzu, presented to the throne in a.d. 121, offers no hint that t'u refers to enslavement, as indeed it would not unless that meaning were considered primary. However, none of the later commentators on the Shuo wen suggests the idea of enslave- ment or equates Vu with nuJ^ The only early writer who makes such an equation is Li Chi, who flourished ca. A.D. 200. He says: "A general term for male and female t'u was nu Ji:^^^^^^." {1, footnote 2.) However, Li Chi himself speaks (CHS, 8, la) of fu sentenced to one year. People sentenced to a single year of government servitude cannot be called slaves. If Li Chi is correct that nu was a general term for t'u — and the pre-Han context of 1 Liang Ch'i-ch'ao, "System of slavery in China," p. 532; Ma Fei-pai, "Source material on the economic history of Ch'in and Han," pt. 6, "The slavery system," p. 385; Wu Po-lun, "Hsi Han nu-li k'ao [An investigation of slavery in the Western Han]," ShihHuo, vol. 1, No. 7, March 1, 1935, pp. 275-285 (see p. 278) (hereafter cited by translated title) ; Lao Kan, "The system of slavery during the two Han dynasties," p. 8; Ma Ch'eng-feng, An economic history of China, vol. 2, p. 246. This usage is refuted by Wu Ching-ch'ao ("The slavery system of the Western Han," p. 246, footnote 1), who quotes in his support Wang Shih-chieh, who has made the most exhaustive study of Chinese slavery from the legal point of view. Wang points out that in enslavement for crime the basic principle seems to be that the slaves were the people mutually implicated, but not the criminals them- selves. It was a punishment for relatives of major criminals, arising from the Chinese family system (op. cit., p. 306; Pippon, p. 99). ^Lun heng, ch. 23 (sec. 3), 10b; cf. Alfred Forke, trans., Lun-Heng, 2 vols., London, 1907, Beriin, 1911, vol. 2, p. 378. The whole passage concerns t'u, yet gives no evidence of enslavement. 3 See Shuo wen chieh tzu ku lin, pp. 736-737. ENSLAVEMENT 81 his commentary should be remembered — it tends rather to diffuse the meaning of nu than to sharpen the meaning of t'u. Several passages contextually associate the terms t'Uy and 7m or nu-pei, but because of the nature of the language it is not always possible to tell whether they are separate items, or form a compound, or whether t'u is an adjective indicating a kind of nu. The first passage (3), referring to men working at Li Mountain who were freed in 209 B.C. to fight one of the rebelling armies, makes a clear distinction between two types of people: convicts, and born male slaves ^A'^M.-f'. In proposing the scheme Chang Han had only suggested freeing of t'u. The second passage, dated 197 B.C. and referring to a plan to free men willing to fight Kao-tsu's Empress and Heir-apparent, is indecisive (13). It simply says, "the various bureaus' t'u nu" ^li"^^. Here "convict" and "male slave" could be either two classes, or one special class of "convict-slaves." The third passage ^6), dated 119-113 B.C. and telling of the con- fiscation of property from merchants and others who infringed certain drastic emergency laws, points to an adjectival usage. It seems to distinguish between confiscated slaves il^AifZW, that is, slaves confiscated along with other property, on the one hand, and criminals seized and enslaved ^iN,W, on the other. However, this is only an assumption. The second reference may equally well be read "convicts and male and female slaves," as it is trans- lated in the document. There is an interesting undated passage in the Han chiu i (95), in which it is said that the government took twelve hundred t'u nu from bureaus within the capital city and subordinated them to (or put them under the control of) a single inspectorate. The passage itself could be translated either as "convicts and male slaves," or as "convict-slaves." The interesting point is that in the introductory section to the "Table on the Bureaucracy" in the Ch'ien Han shu, under the heading Ssu-li chiao-wei, a similar statement leaves out nu and mentions only the t'u,^ saying: "Carrying credentials, [the Colonel over the Retainers] was escorted by twelve hundred t'u from the bureaus of the capital city. He arrested [those who practiced] wu-ku, and judicially investigated major [cases of] licentiousness and treachery." Here t'u may be standing for t'u-nu; on the other hand there may be a scribal error, or a misunderstanding, or the meaning "foot soldier" may be indicated. This seems most likely in view of the fact that 1 CHS, 19A, 6a {Han shu pu-chu, 22a). The passage is translated in 95, foot- note 2. I 82 SLAVERY IN THE FORMER HAN DYNASTY the next statement is: "Later his soldiers (^) were abohshed. . . ." Finally, there is a usage dated A.D. 19 which reports that Wang Mang made a great enlistment of imprisoned t'u and people's male slaves IS^A^X to fight as shock troops against the Hsiung-nu. As it makes a distinction between t'u and private slaves, it does not exactly fit the present discussion, but it is interesting because another report of the same event in the same source says Wang Mang enlisted prisoners sentenced to death, and male slaves of officials and plebeians i^EHH^K^X. The conjunction of the two passages shows that the imprisoned t'u in the first item were criminals sentenced to death, and not government slaves (130). In summary, two of the passages show a clear distinction, one is indecisive, and two can be argued to mean "convict-slaves." Now, granting that there is a term "convict-slave," does this mean that fu alone means slave? Obviously it does not. There are many meanings for t'u. Referring to citations in two Chinese dictionaries compiled on modern principles, the Chung-hua ta tzu- tien isolates twenty-four meanings, and the Tz'u yuan gives ten. Not all of these are Han usages, and only a few are nouns such as foot-soldier, commoner giving menial service in government bureaus, and sentenced criminals. But in both dictionaries under the defini- tion meaning a convict the reference to nu comes from the "Treatise on Jurisprudence" of the T'ang shu, or History of the T'ang dynasty (618-906). There are many pre-T'ang historical and legal texts, items from dictionaries, and commentaries, which in one or another way tell of government enslavement of the families or descendants of criminals, and perhaps even of criminals themselves (the distinction is not easy to draw in view of the Chinese family system), but while these employ nu or nu-pei as the term for such people, none earlier than this T'ang shu passage (except Li Chi) equates them with t'u.^ Liang Ch'i-ch'ao, the most scholarly of those who include t'u in their discussions of slavery, speaks of two types of enslavement for crime. He is not so much interested in the Former Han practice as in that throughout Chinese history, but the documents he quotes are the Chou li and Han chiu i.- The first of these dates only shortly 1 1 mean by this that I have not been able to find such references in various Chinese encyclopaedias, legal works, or the Shuo wen chieh tzu. ku lin, with its extensive quotations of Shuo wen commentaries. I may not have looked far enough, but considering that assembling of early references on all subjects is a passion with Chinese scholars, methodically driven to the point of a vice, the search has certainly passed the point of "diminishing returns." 2 Liang Ch'i-ch'ao, op. cit., pp. 532-533. ENSLAVEMENT 83 before Han (probably edited, or "rediscovered" about the beginning of our era), while the second is an early Latter Han work. His first class of government slaves consists of lightly punished people, those sentenced to servitude for one to five years. Though his citations do not refer to these people as t'u, Liang says that they were called fu. Here it is not necessary to go into the types of work such criminals did, or the names of the various kinds of sentences.^ If, however, t'u were people serving criminal sentences from one to five years in length, then they were certainly not slaves any more than convicts in a southern chain-gang are slaves, no matter how slavish their treatment may be. It must be admitted that a confusion between t'u and certain government nu or nu-pei is understandable, and there may indeed have been points of identity. Wang Mang's slaves were certainly given the treatment of convicts, for that is what they were, though they are called nu-pei. Two documents dating very early in the Former Han period describe free people disguising themselves as private slaves, and therefore having their heads shaved, putting on iron collars, and dressing in russet clothes. These were the identify- ing costume of t'u and people sentenced for limited periods.- It is probably significant that the only reported instances during the Former Han period of private or government slaves so treated (except for the Wang Mang case), come at the beginning of the dynasty. At this time private slavery was apparently not very extensive, and perhaps reflected the Ch'in and late Chou treatment of the enslaved families or descendants of executed criminals, who seem to have been the most common government slaves in the pre-Han age. Tattooing the face was another way of distinguishing criminals. It was next to the least of five punishments involving bodily mutila- tion, Ch'ing Pu, a convict working on Li Mountain at the end of 1 See HFHD, vol. I, p. 177, footnote 1, for a translation of the Han chin i pas- sage. Ch'eng Shu-te (op. cit., pp. 51-55) assembles the Han material. He does not equate criminals of this sort with t'u, just as he does not equate t'u and nu, yet the passages he quotes when dealing with punishments of one and three year servitude use t'u contextually. Dubs does equate t'u with slaves and also ap- parently with people sentenced to servitude from one to five years. In his note on the t'u that Kao-tsu was escorting to Li Mountain, he says: "Enslavement or convict labor was a common punishment; a criminal could be sentenced to enslave- ment for a number of years." HFHD, vol. I, p. 34, footnote 1. (Italics mine.) 2 8 and 12. See Ch'eng Shu-te, op. cit., pp. 51-52, for shackling and shaving the heads of criminals; also HFHD, vol. I, p. 177, footnote 1, and Edouard Chavannes, Les documents chinois. . . , p. 63. Document 26 describes the meeting of the slave, Wei Ch'ing, with a convict in an iron collar. 84 SLAVERY IN THE FORMER HAN DYNASTY the Ch'in period, had been tattooed, and took his name from that fact. The practice was theoretically abolished by an edict of Em- peror Wen in 167 B.C., but probably continued to be done at later times. ^ Evidence that enslaved relatives or descendants of executed criminals were tattooed appears in a citation of the Han law code, quoted by the Grand Judge, Chung Yu, toward the end of the Latter Han period. According to the Han code, said Chung Yu, "the wives and children of criminals [PA] are confiscated as male and female slaves, and tattooed on the face." Amplifying this citation he explained that "the punishment of tattooing the face as practiced by Han was preserved from ancient statutes. The genuine male and female slaves of today had ancestors who originally committed crimes. Even though a hundred generations have gone by, still they have tattooed faces [as a sign of] submission to the government." - There seem to have been differences as well as similarities between fu and government nu or nu-pei. There is no evidence that t'u were sold, or given away, which is a useful pragmatic test of slave- ownership. An affirmative conclusion cannot be drawn from absence of evidence. Therefore, while it cannot be asseverated that t'u were not sold, it cannot be proved that they were. There is plenty of documentary proof, on the other hand, that slaves, 1 For example, about a.d. 150 Chu Mu offered to have his face tattooed in expatiation of a misdeed. Cf. HHS, 73, 5b. This could hardly have happened if the custom had been strictly abolished for more than 300 years. 2 San kuo chih, Wei chih, 12, 4a. The statement occurs in the biography of Mao Chieh (ibid., 3b-4b), an important official under T'ai-tsu (i.e., Wu-ti or Ts'ao Ts'ao) shortly before a.d. 220. Someone reported that Mao Chieh had interviewed some "rebels with tattooed faces, whose wives and children had been seized by the government as slaves." (The Chinese construction of the sentence makes it appear that the rebels were the ones whose faces were tattooed, but Chung Yu's statement quoted above, as well as what follows, indicates that it was the enslaved wives and children of former rebels who are meant.) Mao Chieh was thrown into prison because he was reported to have said that this sort of treat- ment was the cause of a current drought. In the trial Chung Yu quoted the Book of history and the Chou li to prove that such enslaving was an ancient practice, and then cited and explained the Han law code, as quoted above. He asked how many tattoo-faced people Mao Chieh had interviewed, and wanted to know how the ignorant "tattoo-faced male and female slaves" had been able to obtain an interview with Mao Chieh to present their grievances. Cited by a Grand Judge in an official trial, this quotation of the Han code, and his explanation, constitute good evidence of the treatment of this type of govern- ment slave. Chung Yu's citation is quoted as evidence of Han conditions by such Chinese students of law and slavery as Ch'eng Shu-te (op. cit., p. 81), Shen Chia-pen (op. cit., 5b, 6a-b), Wang Shih-Chieh (op. cit., p. 307), Liang Ch'i-ch'ao (op. cit., p. 532), Ma Fei-pai (op. cit., p. 389), and several others. Since slaves were tattooed for purposes of identification, the face was the natural place for the marking. The kind of tattoo mark applied in Han times ENSLAVEMENT 86 including government ones, were sold and given away. Many special pardons of t'u were recorded, but very few manumissions of government slaves during the Former Han period. T'u were often recruited to fight in China's frontier wars, but government slaves are not reported to have been. There were a number of revolts of t'u, but none reported for government slaves, which strongly suggests a fundamental difference in treatment. Because there seems to be no contextual evidence that t'u were slaves; because many of those called t'u were merely criminals sentenced to labor from one to five years ;^ because the term has a number of other important meanings; and because t'u in the sense of criminal cannot be equated with any other Former Han term surely or exclusively meaning "slave" — for these reasons it is unjusti- fiable to translate the term as "slave." To treat t'u as slaves distorts the crucial problems of slave numbers, functions, and economic position. In this work the term is translated "convict." Enslavement Because of Economic Distress Famine and slavery in China are cause and effect, and the sale of women and children because of economic distress is a constant factor during all Chinese history when slavery was an established institution. Numerous instances of the sale of children from the beginning of Han times through the Ming period appear in the dismal record of famines spread out year by year in the pages of Chinese encyclopaedias, and many Occidental writers attest to the practice during the last dynasty. As late as 1920-21, women and children, and particularly young girls, were sold in large numbers in a north- China famine which cost 500,000 lives. Yet at that time slavery was already legally abolished in China. Sales during famines doubtless still occur. Extensive warfare almost always causes famine. People flee their homes, trade and communications collapse, crops and grain- is not known, but a description of the practice in Chin times (a.d. 265-419) perhaps throws a little light on this minor point. An official order stated that male or female slaves who ran away were to be tattooed on both eyes with a copper- green substance like ink; if they ran away a second time they were to be tattooed on both cheeks; and if they ran away a third time they were to be tattooed under each eye with a horizontal mark one and a half (Chinese) inches long and half an inch wide (Ch'eng Shu-te, op. cit., p. 354, quoting the Yu yang tsa tsu and the T'ai p'ing yil Ian). 1 Instead of being executed, seriously mutilated, dismissed from office or title, or fined. These were the principal types of punishments; cf. Ch'eng Shu-te, op. cit., pp. 42-63. 86 SLAVERY IN THE FORMER HAN DYNASTY stores are seized or destroyed to weaken the enemy, able-bodied men forced into the army cannot attend to planting and harvesting, and brigands stalk and pillage in the wake of the troops. At the beginning of the Han period, after four years of rebellion and civil war, a great famine swept the Ch'in stronghold of Shensi, normally a rich and fertile region. Grain cost five thousand cash the picul, people ate human flesh, and over half the population died. In 205 B.C. Kao-tsu permitted people to sell their children and migrate to Szechwan for food (7). Probably this order merely sanctioned a widespread practice. Famine does not wait for imperial decrees! Three years later, when the country was scarcely pacified, the Emperor decreed: "Those common people who because of famine have sold themselves to be people's slaves are all to be freed and become commoners." (9.) This edict shows that people had sold themselves as well as their children. The cycle of warfare, famine, and slavery was repeated at the close of the Former Han period. The successful rebellion against Wang Mang degenerated into years of civil warfare. Famine devastated the land, and two edicts of Emperor Kuang-wu evidence the sale of victims. Three years after the murder of Wang Mang, the new Emperor proclaimed that plebeians' wives married and children sold could freely return to their parents {1SJ^). In a.d. 31 a more specific edict ordered that those officials and plebeians who had encountered famine and turmoil, thus becoming slaves and lesser wives, should all be allowed to return home if they wished. To put teeth into the edict the Emperor ordered that any master who dared to restrain them from leaving should be tried according to the "law for selling people." (136.) Records of famine and widespread economic distress caused by droughts, untimely frosts, floods, and wars appear at least twenty times in the "Imperial Annals" of the Ch'ien Han shu during a period of 212 years. Each of these disasters was severe enough to demand government action, or at least imperial solicitude. Four notices report cannibalism. There is no indication in any but the first of these terse and stark records that people sold themselves or their children, but it is likely that every famine produced a new batch of slaves. Several leading statesmen referred to the practice in memorials on general economic conditions. Thus, in 178 B.C. Chia Yi told Emperor Wen that even after several decades of peace public and private stores of grain were ENSLAVEMENT 87 lamentably small. "When there is a lack of timely rains," he said, "the people cast wolfish glances; when the harvest is bad and is not harvested, pleadings to be allowed to sell titles and children have been heard." (A^.) Ch'ao Ts'o also addressed to Emperor Wen a long memorial on agriculture, in which he described the hardships of farmers, burdened with year-long labors, corvee duties, charitable obligations, excessive taxes, and ill-timed legislation. ^ When droughts and floods occur, "those who have [grain] sell at half price [sic!], and those who have none get respite by borrowing, to be repaid double; and therefore there are those who sell their lands and houses, and sell their children and grandchildren in order to repay their creditors." (18.) Chia Chtian-chih attempted to dissuade Emperor Yiian from launching a military expedition against the people of Hai-nan Island in 46 B.C. A year before, a famine in Shantung was so severe that people there resorted to cannibalism. Chia Chtian-chih pictured the unhappy condition of the people in eastern China, who for years had suffered and been forced to leave their homes. Although in human nature no family relationship is closer than that between children and their parents, he said, and nothing is more joyous than the relationship of husbands and wives, still the people are driven to the extremity of marrying off their wives and selling their children. Neither laws nor the sense of righteousness could stop them. He argued that to send off a great military expedition at such a time was no way to cope with famine and to preserve the people (AlS). During Wang Mang's reign famine and warfare against the Hsiung- nu forced the people to flee into the central commanderies, where they were sold as slaves. To prevent this depopulation of the frontiers, Wang Mang ordered public execution for anyone, official or plebeian, who dared to traffic in frontier plebeians (126). An interesting variant of child-selling was "pawning children." Liu An, better known as Huai-nan Tzu, referred to this custom in his letter to young Emperor Wu. He criticized a proposed campaign against the Kingdom of Yiieh because economic conditions were unfavorable. Two years before, a flood of the Yellow River had forced people to practice cannibalism. "For several years," he said, "the harvests have successively not been abundant, and people have had to depend upon selling their honorary ranks and pawning 1 The whole memorial is translated by Herbert A. Giles (Gems of Chinese literature, rev. ed., Shanghai, 1922, pp. 70-73) and Georges Margoulifes (Le Kou- wen chinois, Paris, 1926, pp. 68-73). The essential parts of the famous documents are translated by Duyvendak (op. cit., pp. 54-55). 88 SLAVERY IN THE FORMER HAN DYNASTY their children in order to continue to clothe and feed themselves." (33.) Ju Shun, a third century commentator, explains the practice of pawning children. In Huai-nan people sold children to do slave work. If, after three years, the parents could not redeem them, the children then became male and female slaves, A T'ang dynasty custom is quoted in support of Ju Shun's observation. It was the custom at one place in Kiangsi province to use boys and girls as collateral for borrowing money. If the loan was not repaid in the specified time, or if the interest due equaled the principal, the children were seized and enslaved (33, and footnote 5). Although there is no other evidence in the Han dynasty of this custom, or of debtor slavery specifically, it is plausible as a modified and mitigated form of the well-documented practice of selling children outright. A poor family that could not afford to rear another child, or especially another girl, might allow the newborn child to die. In- fanticide may be used as a measure of economic distress. A destitute family might prefer to leave its baby girl on the doorstep of a wealthy family that could rear her as a slave. Formal histories do not give much precise information about infanticide, yet there are references to it in Han times. In one region so many poor people abandoned their children that Chia Piao announced that parents who let their children die would be punished for murder.^ Legality of Sale into Slavery Two curious facts stand out in the documents regarding enslave- ment because of economic distress. The first is that the sale of free people was apparently extra-legal, justified only in special cases. Kao-tsu gave special permission for people to sell their children. Chia Yi speaks of people pleading to be allowed to sell their chil- dren. Chia Chiian-chih reports that laws could not prevent people from marrying off their wives and selling their children. The second fact is that in the early years of both dynasties people who had been sold into slavery because of famine were ordered to be freed, as though the enslavement were invalid. Emperor Kuang-wu even ordered that slave owners who refused to comply would be tried according to the "law for selling people." These two facts raise an important question. Was the sale of free women and children, or voluntary self-sale, recognized by Han law as a valid transaction? Since Emperor Kuang-wu invoked a "law for selling people" almost at the beginning of his reign, it must 1 HHS, 97, lib; cf. also Ch'eng Shu-te, op. cit., p. 134, for other early cases. ENSLAVEMENT 89 be assumed that it was an already existing law dating from the Former Han period. What was this law? The Chin shu or History of the Chin dynasty (a.d. 265-419), compiled in the seventh century from eighteen earlier works, has a reference to this "law for selling people" in its "Hsing fa chih." It quotes the preface to the Hsin lii or New code, by Ch'en Ch'iin, who died in a.d. 236. He and others prepared this New code for the Wei dynasty of the Three Kingdoms period some time before 229, and the mention in Ch'en's preface of the "law for selling people" probably refers to the Han law; that is, the law invoked by Emperor Kuang-wu in A.D. 31. The preface said merely: "The law on robbery had items on kidnaping, terrorizing, and selling and buying people by persuasion." The heading "Law on Robbery" was one of the six divisions taken over by the Han dynasty from the Ch'in, and incorporated by Hsiao Ho into his "Law in nine sections," the basic Han code. The details of this law on robbery are mostly lost. But Wang Hsien-ch'ien quotes the eighteenth century commentator Hui Tung as writing that it said: "Those who kidnap people or kidnap and sell people or sell people by persuasion or buy people by persua- sion, and make them slaves, shall die." ^ The law of the Northern Wei dynasty (a.d. 386-534) had virtually the same provision in its section on robbery: "Those who kidnap people or kidnap and sell people or sell people by persuasion, and make them male and female slaves, shall die." This law was cited in the trial of a very interesting case in A.D. 502. A man sold his seven-year-old daughter to be the slave of a fellow townsman. The latter purposely bought her for resale, and did resell her in another region without revealing her background. She was thus in danger of losing her identity as a "good" person. Charged with buying by persuasion, he should, according to the law, have been strangled. The same case affords the earliest extant citation of a law against selling members of the family. It was invoked against the girl's » Chin shu, ch. 31 ("Hsing fa chih"), 5a. I translate the word "item" in the plural, because in a.d. 37 Emperor Kuang-wu ordered that owners who dared to restrain plebeians kidnaped into slavery should be tried according to the "law on kidnaping people." Obviously the law on selling people was a separate item from the law on kidnaping people, and both were part of the law on robbery. Cf. Shen Chia-pen, op. cit., "Han lu chih i," ch. 2, 9b, ff. On the Hsin lii of Ch'en Ch'iin, cf. Ch'eng Shu-te, op. cit., p. 229 (preface to his ch. 2, on the Wei dynasty law). On the "Law in nine sections" of Hsiao Ho, cf. Jean Escarra, Le droit chinois, Peking, 1936, pp. 23 and 94. Wang Hsien-ch'ien's commentary is quoted in 136, footnote 2. Liang Ch'i-ch'ao (op. cit., p. 550) also quotes Hui Tung's statement, and locates it as a commentary to the Jih chih lu by the famous scholar Ku Yen-wu. He also points out that this law on robbery was part of "Law in nine sections" by Hsiao Ho. 90 SLAVERY IN THE FORMER HAN DYNASTY father, and read: "Those who sell their children shall be punished for one year. [Those who sell] relatives of the same surname, who are their superiors or elders within the five grades of mourning, shall die. Those who sell their near relatives, or their concubines, or their sons' wives, shall be banished. ^ These citations reflect obscurely the Han attitude toward sale of free people. To judge by later codes, the "law for selling people" — invoked by Emperor Kuang-wu in A.D. 31 against anyone who refused to free slaves who were victims of famine, turmoil, and kidnaping — was part of a law against sale into slavery by force or guile. This was, in turn, part of the larger law on robbery. It is not clear whether the Han code actually had any law against selling members of one's family. But it is evident that such transactions were precarious if brought on by famine and other untoward condi- tions. The slaves might be ordered to be freed even though permis- sion for sale had formerly been granted. Both Liang Ch'i-ch'ao and Wang Shih-chieh, generalizing from the continuity of Chinese law, assert that sale of free people has never been legally recognized in China, though it was regularly practiced. - In Han times people compelled by starvation to sell themselves or their children had to find some private buyer. Purchase of famine victims is not reported among the various measures adopted by the government to relieve or forestall widespread economic distress, and there is no evidence that the government ever bought people who were free up to the time of purchase, or enslaved people for debt. Illegal Forced Enslavement Stringent laws during the Han period did not prevent kidnaping. This indicates that there was a market for slaves which made kidnaping profitable and worth the risk of the death penalty. Shortly before the beginning of the Han period Luan Pu was kidnaped by someone and sold as a male slave in Yen to do a deed of revenge for his master (5). Luan Pu was not a child, but a homeless wanderer who had hired himself out several years before to serve a wineshop keeper. Perhaps it is an indication of the decadence of law and order in the closing years of the Ch'in rule that a young 1 Cf. Ch'eng Shu-te, op. cit. (ch. 5, "Hou Wei lii k'ao"), p. 415. "Liang Ch'i-ch'ao, op. cit., p. 550; Wang Shih-chieh, op. cit., pp. 311-313. I may add, what is only an opinion, that it was probably the chief source of private slaves throughout the period here discussed, as it seems to have been in later times. ENSLAVEMENT 91 man could be kidnaped in Ch'in, transported to Yen two hundred miles away, and sold there without interference. Since it is not known what kinds of legal papers were necessary for slave transactions it is only a guess that he was sold with the connivance of the civil authorities. About the year 190 B.C. Tou Kuang-kuo, the four-year-old son of a poor famil}^ was kidnaped and sold by some one, and his family could not find him. Resold about ten times, and traveling several hundred miles from his home, he was finally bought a number of years later by a man living near Lo-yang. Tou Kuang-kuo's later importance explains why a history devoted primarily to matters of state concern reports the kidnaping of a poor boy. His older sister became the consort of Emperor Wen. About the same time Tou Kuang-kuo escaped death almost miraculously. Casting his horo- scope, he learned that he would soon become a marquis. Therefore he went with his master to the capital and heard that the new Empress was a lady named Tou, from his native town. How he got in touch with her and proved his identity is a moving story. Eventually he did become a marquis as his horoscope had predicted (llf.). These are the only specific cases on record of individual Chinese who were kidnaped and sold into slavery. However, strong robber bands are frequently reported from various parts of the empire and often had to be suppressed by military force. These bands doubtless kidnaped many people for sale. Wang ]Mang, in his edict of A.D. 9, denounced kidnaping of women and children for sale as if it had been a constant occurrence throughout the Han period (122). Three edicts of Emperor Kuang-wu in 31, 36, and 37 specifically ordered that people who had been kidnaped and made slaves during the troubled period of founding his dynasty were to be freed {136, and footnote 3). IMany other cases of kidnaping are indeed reported but it is not clear that the victims were sold. People could be seized for revenge, hostage, ransom, forced labor, sexual gratification, or other reasons. Why did Fan Ping and his accomplices kidnap officials and plebeians in a brief rebellion in 14 B.C.? They may have been seized for hostages of safe passage.^ Why did Hsiung-nu bands raid the Chinese frontiers to kidnap plebeians and officials? The Hsiung-nu kept private slaves (5^), but it is unlikely that they wanted or could absorb the thousands of Chinese reported captured on these recurrent raids. Perhaps they held some of the victims for ransom, but they 1 CHS, 10, 6a. 92 SLAVERY IN THE FORMER HAN DYNASTY may have sold many to Chinese slave traders operating on other parts of the frontier, or to various Central Asian kingdoms farther west. Ransom was the primary motive of Chinese officials who, during the reign of Wang Mang, falsely put seals on people's necks, making them youth-servants. They removed the seals only when they received bribes (A21). Kidnaping was not practiced by disreputable robber-bands or "barbarians" alone. Some of the "best people," relying upon their lofty rank for protection, were guilty. Usually the motive was sexual desire and not sale. Thus, in 130 B.C. the Marquis of Chii-ni kidnaped a man's wife. When this was discovered he was executed and his marquisate abolished. In 74 B.C. the King of Ch'ang-i kidnaped girls on his way to the capital. Chang Fang, the favorite of Emperor Ch'eng, tried to take the daughter of a commoner by force. ^ Han documents lack information on many questions naturally arising about kidnaping for slavery. Were there large-scale kidnap- ing rings? If so, by what methods did they operate, what were the profits, and how did they dispose of their victims? Did they work through corrupt officials? What did the government do, beyond passing laws and executing sentences, to prevent kidnaping? Did it scrutinize titles to ownership of slaves, and listen to slaves' com- plaints? To what extent did kidnaping contribute to the total number of new slaves? These questions must go unsolved because Chinese history concerns itself only with "important" matters. Importation of Foreign Slaves Foreign slaves were very popular with the cosmopolitan upper classes of the T'ang period. ^ Korean girls were in demand as personal maids for wealthy gentlemen, and piracy along the Korean coast 1 Cf. CHS, 16, 6b, and 40, 9a, for the Marquis of Chu-ni; documents 67 and 100 for the King of Ch'ang-i and Chang Fang. Several similar cases in Latter Han times indicate that kidnaping of women, under the cloak of imperial protection, was a by-product of nearly unlimited power enjoyed by distaff relatives of the imperial house, by some nobles, and by high officials. Around a.d. 90 Tou Hsien, a brother-in-law of Emperor Shang, and a leading general, kidnaped women and girls (HHS, 53, 8a) as did the powerful eunuch Hou Lan in 170 (HHS, 108, 6b). In A.D. 150, Liang Chi, brother of the Dowager Empress Liang, and one of the most powerful men of his day, not only kidnaped women and girls, but also seized several thousand "good" people and made them male and female slaves, calling them "self-sold people." (HHS, 64, 6a-b.) ^ The subject is not new. It has been discussed in several works on slavery or on the T'ang period, of which the following are a few more readily accessible: Liang Ch'i-ch'ao, op. cit., pp. 540-542; Chang Hsing-lang, "The importation of negro slaves into China under the T'ang dynasty (a.d. 618-907)," Bulletin of ENSLAVEMENT 93 and trade in the seaports of Shantung helped supply this fashionable need. Turkic, Tibetan, and Uighur slaves were captured in raids or sent as tribute by frontier governments. Dancing girls imported from regions even farther west performed novel Iranian dances; they were probably slaves. Dark-skinned K'un-lun slaves, certainly negroid, were very popular; references to them go back to the fourth and fifth centuries. In T'ang times some K'un-lun slaves may have been African negroes imported by Arab traders. The greatest source of new slaves during this period, however, was south China, where much of the population consisted of aboriginal tribes. Edict after edict vainly attempted to prohibit kidnaping or importing of tribes- people from the regions now comprising FXikien, Kwangtung, Kwangsi, Hunan, Kweichow, Yunnan, and Szechwan. In Former Han times foreign slaves were popular also. Southern Szechwan and Yunnan supplied many non-Chinese youths for the Chinese market. Merchants of Shu and Pa, the regions of modern Chengtu and Chungking, were middlemen in a lucrative trade in slaves taken from P'o and T'ien, and resold in Ch'ang-an (10). P'o, at the juncture of the Yangtze and Min rivers, was a particularly important center for it was the gateway to Yunnan and about equally distant from the capitals of Pa and Shu. T'ien was the region of modern Kunming (Yunnanfu). Whether these slaves were aborigines or members of the gifted T'ai race is uncertain. The men of T'ien and P'o may have been slave traders themselves, capturing people from the surrounding hill tribes for sale to Chinese merchants. Both regions were natural centers for such operations. On the other hand, the term "youth," used for these slaves, usually signifies young boys and girls in the luxury class, and it seems more likely that they would be civilized T'ai children. The passages which describe this trade indicate that it already functioned early in the Han period. Ssu-ma Hsiang-ju, acquainted at first hand with conditions in southwestern China, referred to enslavement of children there in a conversation with Emperor Wu (A?). A remark by Fu Ch'ien of the second century of our era indicates that the trade was continuous. He says: "In the former capital" — that is, the Ch'ang-an of his own day — "there are female slaves from P'o." {10, footnote 2.) theCatholie University of Peking, vol. 7, 1930, pp. 37-59; L. C. Goodrich, "Negroes in China," ibid., vol. 8, 1931, pp. 137-139; Kuwabara Jitzuzo, "On P'u Shou- keng," MRDTB, No. 2, 1928, pp. 61-63; Ishida Mikinosuke, "Etudes sino- iraniennes, I: A propos du Hou-siuan-wou," ibid.. No. 6, 1932, pp. 61-76; Stefan Balazs, "Beitrage zur Wirtschaftsgeschichte der T'ang-Zeit, II, Der Sklaverei," MSOS, vol. 35, 1932, pp. 2-14. 94 SLAVERY IN THE FORMER HAN DYNASTY Yiieh was an ancient coastal state in present Chekiang and northern Fukien, having a cultural element markedly different from that of north China. It is interesting, therefore, to read of a slave woman from Yiieh ordered to "descend to the spirits and invoke curses on Emperor [Wu]." (38.) She may have been a sorceress, accomplished in the black arts of the south-China coast. This ref- erence does not prove that Yiieh slaves were common, but another document shows that there was a market for them. In 113 B.C. Lii Chia, Chancellor under the last three kings of Yiieh, issued a statement accusing the Dowager Queen, who was Chinese, of schem- ing to surrender the kingdom to China. He asserted that she planned to take a large number of her attendants to Ch'ang-an, and there sell them as slaves (50). This accusation would have been meaningless unless the people he was seeking to influence knew there was a ready market for Yiieh slaves at the Han capital. Along the northern frontier the Chinese intensive farming and walled-city type of culture gradually faded off into extensive agricul- ture and pastoral husbandry. The frontier was not a fixed boundary line between two nations and two antagonistic cultures, despite the Great Wall which symbolized the Chinese desire to create such a line. As determined by climate and terrain, it was rather a broad zone of change from river valley agriculture to steppe nomadism. The frontier was inhabited by people of mixed stock who employed applicable techniques of production, political organization, domestic life, and warfare derived from both cultures, and freely modified to suit local conditions.^ From the north the Hsiung-nu constantly harassed the Chinese colonists and indigenes of the frontier, plundering their goods and livestock and kidnaping plebeians. As early as the time of Emperor Wen, Ch'ao Ts'o suggested methods to strengthen the Chinese colonists against these forays. He proposed that the government encourage the colonists to resist the Hsiung-nu and to "recover" what they plundered. If, when the enemy invaded one area, the people of a neighboring region could intercept them and recapture live stock and people kidnaped, then the original owner should give half the plunder, or half its value, to the rescuer. The government would reimburse the owner. Ch'ao Ts'o thought this system would make the Chinese resist the Huns and mutually rescue each other, not for the glory of the Emperor, but for personal gain (19). Ch'ao 1 A most instructive work on the northern frontier is that of Owen Lattimore, Inner Asian frontiers of China, New York, 1940. ENSLAVEMENT 95 Ts'o phrased his proposal as a purely defensive measure for newly established colonies, but the real frontiersmen certainly needed no official encouragement to retaliate or initiate raids against the people of the "outer frontier," who were merely more nomad and less "Chinese" than themselves. Such raids were part of the tech- nique of frontier life. In 33 B.C. Hou Ying gave direct evidence of this practice in a ten- point speech against turning over guardianship of the Wall to the then friendly southern Hsiung-nu. He pointed out that a similar policy had been tried with the Ch'iang people inhabiting eastern Tibet, but that peace had been disturbed because greedy Chinese officials and commoners had invaded their territory and stolen their livestock, produce, women and children. Hou Ying argued that the same thing would happen if China ceased to man the Wall and the Hsiung- nu moved in. What he apparently feared was that the zone of "kidnaping" by Chinese, and "rebellion" by the Hsiung-nu, would be moved in closer to the central core of China. Another of his arguments is probably indirect testimony that the Chinese kidnaped and enslaved their northern neighbors. He reported that the male and female slaves of the frontier people were "melancholy and bitter." Many wanted to escape to the Hsiung-nu lands, where they considered life to be happy. Some did indeed escape and the others were restrained only by the oppressive vigilance of the guard- watchers on the Wall (96).^ Now, slaves who wanted to escape to the happy life of the Hsiung-nu were very likely marginal people or Hsiung-nu themselves. Real Chinese slaves from interior prov- inces would not onlj'- be in general ill-adapted to frontier life, but would be much more likely to long for their own homelands, as did most Chinese colonists. Two incidents in the first century of the Latter Han reflect a trade in foreign slaves which was probably in operation during the period covered by this study. In a.d. 49 the leader of the Wu-huan people west of the Liao (western Manchuria and eastern Inner Mongolia) came with 900 followers to the Chinese court to present "tribute," which consisted of male and female slaves, cattle and horses, bows, and the furs of tiger, leopard, and sable. Because of 1 The statement that watchers of the Wall prevented slaves from escaping to the Hsiung-nu is a minor corroboration of Lattimore's thesis that the Great Wall had a double function: it was just as important to keep the intra-mural frontier peoples under Chinese political control as to protect China from the nomads outside. Cf. Lattimore, op. cit., pp. 480-482. A treaty of peace with the Hsiung-nu in a.d. 5 stipulated as its first item that Chinese absconding to the Hsiung-nu should not be given asylum (96, footnote 4). 96 SLAVERY IN THE FORMER HAN DYNASTY the size of the party the tribute was probably large. In return they received rich imperial "gifts." ^ Exchange of "tribute" and "gifts" between rulers of states in Mongolia and Central Asia and the Chinese government was actually trade.- There is every reason to believe that the first historical notice of slaves in this exchange came long after, not before, they were first used in trade at the frontiers. Shortly before a.d. 90 Li Hsiin was appointed internuncio, or director of guests, and sent to be Assistant Protector-General of the Western Regions. On several occasions hostage princes, am- bassadors, and foreign merchants tried to present to him male and female slaves, Ferghana horses, gold, silver, incense, and rugs, but he refused to accept them.^ Considering Li Hslin's position, these gifts, part of the stock in trade of visiting foreigners, must have been bribes. The slaves probably came from Central Asia, since most of the goods listed were unmistakably products of the "Western Regions." The historian reported Li Hslin's refusal as an illustration of his particularly virtuous character. It may be concluded, there- fore, that officials in similar positions normally accepted such gifts,"* and that slaves were part of the Central Asian trade in Former Han times just as they were later, in T'ang times. These are the principal ways by which free Chinese and foreigners became slaves. Major crimes by family members led exclusively to government slavery; economic distress and kidnaping produced private bondsmen; while foreigners, enslaved by methods generally unknown, went to both categories of owner. During the 230-year period covered by this study, and partic- ularly during the reign of Emperor Wu, China fought a series of extensive and costly wars on its several frontiers. In these foreign wars thousands of the enemy were captured. The custom of enslav- ing prisoners of war has had a very wide distribution in various historical periods, though it was by no means universal. Warfare 1 HHS, 120, 2a. 2 Cf. Frederick J. Teggart, Rome and China, Berkeley, Calif., 1939, pp. 214-216, especially footnote 59; Lattimore, op. cit., p. 175. 3 HHS, 81, la. On other men who held the position around the same time, and the regions under control, cf. Edouard Chavannes, "Trois generaux chinois de la dynastie des Han orientaux," TP, 2nd ser., vol. 7, 1906, pp. 210-269. * As noted by Ma Fei-pai (op. cit., p. 388). Lao Kan (op. cit., p. 9), quoting the Tung kuan chi, emphasizes that the slaves were Hu, "barbarians." ENSLAVEMENT 97 was probably the main source of Roman slaves at the other side of the Eurasian continent at precisely the period here being examined. Therefore the question arises: Did the Chinese make slaves of enemy soldiers and civilians captured in battle? Without close examination of the historical evidence it cannot be lightly assumed that they did. IV. WERE PRISONERS OF WAR ENSLAVED? To this absorbing and perplexing problem there is no decisive answer. Modern Chinese scholars have reached three divergent conclusions: (1) Prisoners captured in war by the Chinese were en- slaved; (2) some prisoners of war may have been enslaved, but this was not characteristic, and prisoners were not an important source of slaves; (3) such captives were not enslaved at all.^ Obviously we are on very uncertain ground. These contradictory conclusions are possible precisely because historical sources are neither explicit nor decisive on the matter of disposition of prisoners. 1 (1) Lao Kan ("The system of slavery during the two Han dynasties," p. 9) states that prisoners of war taken from non-Chinese tribes were all enslaved; but he makes the important observation that barbarian peoples who voluntarily surrendered were not enslaved. Wu Po-lun ("An investigation of slavery in the Western Han," p. 281) believes that enslavement of prisoners was universal in Western Han times, and that one of the principal reasons for the wars with the Hsiung-nu was to capture slaves. (2) Wu Ching-ch'ao ("The slavery system of the Western Han," p. 265) says that prisoners of war were not an important source of slaves in Western Han times. He points out that it is exceedingly difficult to say whether or not Hsiung- nu captured on the field of battle were all enslaved, but considering the large numbers it is hard to believe that they all were. Wang Shih-chieh ("The Chinese slavery system," pp. 309-310 [trans. Toni Pippon, "Beitrag zum Chinesischen Sklavensystem," p. 104]) states that there were several cases during Chinese history when large numbers of prisoners were made slaves, but that this situation usually occurred in times of internal war and dynastic change when Chinese soldiers took soldiers and civilians of rival regions as slaves. Moreover, the practice was more common when non-Chinese ruled the country. He does not discuss Chinese enslavement of non-Chinese enemy. Liang Ch'i-ch'ao ("System of slavery in China," p. 542) does not mention captured enemy as slaves until after T'ang times, when they became common. He considers the practice more typical of northern barbarians who ruled China during much of the time after T'ang. Ma Ch'eng-feng ("Ts'ung hsi Chou tao Sui ch'u chih i-ch'ien-ch'i-pai yli nien ti ching- chi chuan-i [Economic transition during the seventeen hundred years from Western Chou to the beginning of Sui]," Shih Huo, vol. 2, No. 9, Oct. 1, 1935, pp. 400-410, [see p . 403] ) emphasizes that prisoners of war were not the principal source of slaves. (3) T'ao Hsi-sheng (An economic history of Western Han, p. 56) asserts that captives were not made slaves, and that in fact many thousands of them were supported by the government with food and clothing and received rewards. He cites as evidence the case of the folld'wers of the King of Hun-hsieh (cf. I^0, foot- note 1), thus failing to distinguish between captives taken on the field of battle, and enemy who voluntarily surrendered (cf. my discussion, pp. 102-103 below). Yet he represents a strong school of thought on Chinese economic history. Ap- parently T'ao Hsi-sheng held the opposite opinion in his article entitled "A new estimate of the development and process of the forms of Chinese society (Chung- kuo she-hui hsing-shih fa-chan kuo-ch'eng ti hsin ku-ting)," in which he stressed the fundamental position of a slavery economy in Han times. I have not seen this article, but Ma Ch'eng-feng (An economic history of China, vol. 2, p. 237) quotes him as saying that the military conquests of the two Han periods— especially the brilliant conquests of Emperor Wu — were a movement for slave-hunting and commercial expansion. Ma refutes this view with numerous arguments and citations, thus placing himself in the camp which believes that prisoners of war were not enslaved (pp. 237-244) or were not an important source of slaves (p. 322). 98 ENSLAVEMENT OF PRISONERS OF WAR 99 In regard to most campaigns they record the numbers of enemy killed and captured, but none tells what was eventually done with the captives.^ This is even more puzzling because a Chinese ethno- graphical account of the Hsiung-nu, written during the second century B.C., tells explicitly that enemy captured by the Hsiung-nu in battle became the private slaves of their individual captors (5|). Records of Captives How accurate are the records of enemy killed and captured, and was there falsification? Why were records kept? On the whole, reports are quite specific. When the numbers are small they are stated exactly,- but when large they are estimated in round numbers. Frequently if a campaign is described in several different chapters, one account will give a precise figure while the others will be general. Thus in reporting the first campaign led by Ho Ch'u-ping in 121 B.C., two accounts say that he killed or captured over 8,000 Huns, but the laudatory speech by Emperor Wu, probably quoted directly, tells of his killing or capturing 8,960.^* Keeping records of enemy losses as accurately as possible was important because rewards of honorary rank in the official hierarchy > I have searched in vain through the memoirs on those foreign nations with whom the Chinese fought, the chronicles of the Han emperors, and the biographies of the great Han generals for any indication of what was done with enemy soldiers and civilians after they were captured. Chinese writers on Han slavery consulted have likewise produced no positive historical references. ^ Thus CHS, 41, la, which tells that Fan K'uai at various times "beheaded 15," "beheaded 23," "beheaded 40 and captured 16," etc. 3 CHS, 55, 3b, as compared with 6, 6a and 94A, 8b. It is necessary here to discuss the terminology involved in these reports. The most important term is hi }^, which means "to capture an enemy" or "an enemy captured alive in warfare." This definition of la is confirmed by the commentators Chin Shao (ca. A.D. 275) and Yen Shih-ku (a.d. 581-645), and is proved contextually, for example, in the biographies of Fan K'uai, Hsia-hou Ying, and Chin Hsi (CHS, 41, passim, and SC, 95 and 98, passim). Lists of their successes state how many enemy heads were cut off, prisoners captured, enemy surrendered, cities taken, armies defeated, etc., in such a way that lu can only mean prisoners. Furthermore, several instances of lu giving valuable military information show that they were soldiers taken alive (CHS, 55, 5b, twice; 94B, 10b). The term has other meanings. The most common is a general designation for the enemy. The Hsiung-nu were often simply called lu. This is a potential source of confusion with lu meaning "a captured enemy." A third meaning is "male slave." The Tz'u yiian explains that "the ancients used captured lu as household male slaves. Therefore they also called slave-servants 'lu.' " Locus classicus for this meaning is a passage in the SC biography of Li Ssu, in a memorial addressed to the second Ch'in Emperor, which quotes Han Fei-tzu as saying: "It is the affectionate mother who has the prodigal son, while the severe household is without fierce slaves (h<)." (SC, 87, 6b. DerkBodde, trans., China's first unifier, 100 SLAVERY IN THE FORMER HAN DYNASTY were given to officers in proportion to their successes. Very successful generals and other commanding officers were given marquisates which carried incomes expressed in terms of households in the fief. In the heat of battle, records of enemy slain or captured by individual soldiers must have been hard to keep accurately, yet men of the ranks sometimes received money rewards after the return of a highly successful expedition. After the campaigns of 124 and 123 B.C., for example, the soldiers who had cut off heads or captured enemy, received grants which were reported to total more than 200,000 catties of gold. Although this may be an exaggeration, there is no reason to doubt the fact that large money rewards helped to bankrupt the government.^ p. 40; cf. J. J. L. Duyvendak, The book of Lord Shang, p. 291, footnote 1. In Han Fei-tzu the passage appears in ch. 19, [par. 50], p. 26.) Lu is also often used in the generaUzed sense of "despicable person"; cf. the analogous term "caitiff," derived from Latin captivus. The term most commonly used for enemy slain in battle is shoii "^, "head." In statements of successes of military expeditions or of individual commanders a regular formula was used to report the number slain: "$ff^' • • • [number] . . . 15, i.e. heads cut off . . . [so many] . . . degrees [of honorary rank]." This equa- tion of degrees of honorary rank with numbers of enemy slain apparently began with the practice in Ch'in State of granting one degree for each enemy slain in battle. Cf. Duyvendak, op. cit., pp. 63 and 295-303 (cf. p. 147 for a discussion of par. 19); Bodde, op. cit., p. 8. The term chi ^, in the combination shou . . . chi, appears by Han times to have had the value merely of a numerical classifier. In texts referring to the early part of the Han period, enemy soldiers slain and those captured in battle are normally listed separately (cf. CHS, 41, passim). However, reports during and after the reign of Emperor Wu usually give totals of soldiers lost by the enemy. It is impossible to distinguish the number in each category. The formula $/ft^'ti/^ or $ff^^^^» followed by the total number, is often abbreviated to # (or '^, or ^, or ^) ^f^. Furthermore, beginning with the reign of Emperor Wu the word chi "degrees [of honorary rank]," is frequently used in the passages which lump enemy slain with enemy captured. In CHS, 55, 2a, Yen Shih-ku explains the use of chi for both slain and captured. "Originally," he says, "for one head of a slain enemy [a person] was conferred one degree of honorary rank. Therefore one head was called one degree. Accord- ingly they also named one [enemy] taken alive as one degree." For the expression shou lu, Chavannes (MH, vol. Ill, p. 553, footnote 2) adopts the translation "esclaves soumis." This gives shou the special meaning of fu ^R, "to submit." But there appears to be no evidence for this sense before Latter Han times, and it seems incorrect for the previous period to give a special meaning to shou in combination with lu. Dubs discusses this problem (HFHD, vol. II, footnote 7.8 to chapter VI). The act of voluntary surrender was called hsiang 1^. Loot was called lu ^ ; but this term sometimes means captives also. A Latter Han bas-relief of a battle scene shows enemy heads hanging in a rack and prisoners with hands tied behind their backs. Cf. Edouard Chavannes, Mission archeologique dans la Chine septentrionale, 2 vols, plates, 2 vols, text, Paris, 1909, 1913, 1915: pi. XXVI, No. 47, lower left, and pi. XXVIII, No. 50, left center. iCHS, 24b, 3b; SC, 30, 2b (MH, vol. Ill, pp. 552-553). Two hundred thousand catties (chin) of "yellow metal" come to 1,568,000 troy ounces, accord- ENSLAVEMENT OF PRISONERS OF WAR 101 Since success in killing and capturing enemy brought valuable rewards, records were sometimes falsified. But this was dangerous, for discovery of deceit brought severe punishment. The fact that reports were closely scrutinized and falsification was dangerous adds a certain degree of credibility to the figures finally given by the historian. The following trials illustrate the procedure against falsifiers. Wei Shang, Administrator of a commandery in present Suiyuan province, fought a whole day with the highest merit for killing and capturing Hsiung-nu; but because the reports by his staff members did not exactly agree he was tried and not rewarded. Considering this a great injustice, Feng T'ang protested to Emperor Wen that the laws were too strict, rewards too light, and punishments too severe. He said that because of a miscalculation of only six, the Emperor had turned Wei Shang over to civil officials who had stripped him of his honorary rank. Emperor Wen saw the injustice and immediately ordered Feng T'ang to take credentials pardoning Wei Shang and reappointing him Administrator of Yun-chung. In 117 B.C. Kuo Pu-chih, the Marquis of I-kuan, was tried for increasing the count of heads he had taken in a fight with the Hsiung-nu. He should have been executed but was allowed to pay redemption and was merely dismissed from his marquisate. In 71 B.C. Ch'e Shun, the Marquis of Fu-min, was tried for falsely increasing the amount of loot and number of captives taken in a battle with the Hsiung-nu. He committed suicide. ^ Yang P'u tried a most interesting variant on this dangerous trick. He had won considerable merit as a general against the southern Yiieh in 112 B.C., and was made Marquis of Chiang-liang. When the Eastern Yueh "rebelled," Emperor Wu planned to use him again. Because Yang P'u was excessively boastful about his earlier accomplishments, the Emperor sent him a letter of reproof, listing his various faults, the first of which was that in the defeat of P'an-yii (Canton) he had seized people who had surrendered and made them prisoners of war, and had dug up corpses and used them as slain enemy! The significant point in this case is the ing to a personal communication from Dr. Homer H. Dubs. Today, at a price of $35.00 per ounce, this "yellow metal," if real gold, would be worth $54,880,000 in United States currency. 1 For Feng T'ang, see CHS, 50, 3a; SC, 102, 3b. For Kuo Pu-chih, CHS, 17, 8a; SC, 20, 9b. Note the commentary by Yen Shih-ku. For Ch'e Shun, CHS, 18, 8b; 66, 3b; and 94A, 14a. 102 SLAVERY IN THE FORMER HAN DYNASTY clear distinction between enemy captured in warfare and enemy who voluntarily surrendered.^ Differentiation Between Surrendered and Captured Enemy This distinction is important, and it is just here that some Chinese writers confuse the whole issue by failing to differentiate between the two. The distinction was much more sharply drawn than that between enemy slain and captured in battle. The disposi- tion of captured enemy is unknown but several cases show what happened to those who surrendered voluntarily. During the overthrow of the Ch'in dynasty, when a city or a general voluntarily surrendered, the normal procedure was to leave the civilian population alone, but to annex most of the surrendered army. Sometimes defeated soldiers were enlisted in the victorious army." Hsiang Yii executed a reported 200,000 Ch'in soldiers who had been surrendered by their general, perhaps against their will (6). This slaughter, in violation of the accepted code, is one of the blackest crimes charged against him. Foreigners who surrendered voluntarily were not supposed to be mistreated. For example, Li Kuang, consulting a soothsayer, attributed his failure to be promoted to the fact that he once induced 800 rebellious Ch'iang to surrender, and then killed them. The interpreter of psychic emanations agreed, saying that nothing was more calamitous than killing those who had already suiTendered.^ There are several examples before the reign of Emperor Wu of enemy being rewarded for surrendering to China. About 166-164 B.C., Han T'ui-tang and Han Ying, grandsons of a Chinese king who had joined the Hsiung-nu, surrendered with their followers and were given marquisates. When Hsiu-lu and four other Hsiung-nu kings surrendered, about 147 B.C., Emperor Ching made them marquises to encourage other alien kings to submit also. Three more became marquises by surrendering during his reign.'* 1 CHS, 90, 5a. m^^i^MmMn^^jn^. Another irregularity occurred in the same campaign, for in 104 B.C. Sun Hsiang, who inherited the Marquisate of Lin-ts'ai, was tried for having appropriated other people's prisoners (lu) and loot in the battle of P'an-yiieh (CHS, 17, 13b). 2 Cf. HFHD, vol. I, pp. 44, 49, 53-54, 75. 3 CHS, 54, 3a; SC, 109, 3a. * For Han T'ui-tang and Han Ying, see CHS, 33, 4b; 16, 53a, and 54a. For the eight Hsiung-nu kings made marquises, CHS, 17, 3a-4a. ENSLAVEMENT OF PRISONERS OF WAR 103 Between 140 and 119 B.C., ten kings, four chancellors, two majors, a Tang-hu, and the Hsiung-nu Heir-apparent, deposed on his father's death by his uncle, voluntarily surrendered and were given mar- quisates.^ Such surrenders were natural during a period of Chinese military ascendancy. It is highly significant that there is no corre- spondence between the names of these Hsiung-nu who were given marquisates and the names of Hsiung-nu leaders captured between 129, when active war commenced, and 119, the year of China's greatest success. None of the Hsiung-nu leaders named as being captured is listed as receiving a marquisate, and all Hsiung-nu who became marquises are specifically stated to have surrendered. Furthermore, accounts of campaigns in those years often mention captured Hsiung-nu officials by title only, yet even these titles are never reflected in the list of marquises appointed in coiresponding or succeeding years. Obviously, then, those Hsiung-nu officials who resisted the Chinese and were taken in battle were treated very differently from those who surrendered voluntarily. The wars against kingdoms in present southwestern and south- eastern China and in Korea were of short duration, and the Chinese took relatively few prisoners. But the wars against the Hsiung-nu were of a quite different order. Intermittent punitive expeditions and minor warfare characterize the period before Emperor Wu, the "Martial Emperor." Much of his reign was devoted to successive campaigns against the Hsiung-nu, who lost vast territories in the northwest. After 90 B.C. the armed peace was broken only occa- sionally until the time of Wang Mang, when the Hsiung-nu succeeded in recapturing much of their lost territory. Typical Reports of Wars and Captures Historical accounts of the Hsiung-nu wars follow a general pattern. A digest of the seven campaigns between 129 and 119 B.C. indicates the nature of the available information about treatment of prisoners of war — Hsiung-nu, Koreans, Cantonese, or any other people. These campaigns- are interesting because of the extraor- » CHS, 17, 4b-10a; SC, 20, lb-14b (MH, vol. Ill, pp. 161-169). 2 Information on them is found in CHS, 6, the "Annals of Emperor Wu," 3b-7a, under respective years; in the biographies of Wei Ch'ing and Ho-Ch'ii- ping, CHS, 55, lb-3a (SC, HI, lb-3a), and CHS, 55, 3a-6a (SC, 111, 3a-6b), respectively; and in the "Memoir on the Hsiung-nu," CHS, 94A, 7b-9b (SC, 110, 8b-10b). Farther facts appear scattered through biographies of lesser com- manders, statesmen, and the "Treatise on Economics." In preparing this description the material in all these sources was translated and collated, but it is impossible to give a continuous and intelligible narrative 104 SLAVERY IN THE FORMER HAN DYNASTY dinary totals that are reported of persons killed and captured. Furthermore, unique information on slavery at this particular time appears to bear upon the question of enslavement of prisoners. Campaign of 129 B.C. — In the fall or winter of 130 the Hsiung-nu invaded Shang-ku. The following spring Emperor Wu dispatched four armies of 10,000 cavalry each, from four bases on the center of the northern frontier: Shang-ku, Tai, Yen-men, and Ylin-chung. Only Wei Ch'ing, commander of the eastern army, distinguished himself by going to Lung-ch'eng, the place of the annual Hsiung-nu spring assembly, and killing or capturing some 700 of the enemy. The two central forces under Kung-sun Ao and Li Kuang were badly beaten, Kung-sun Ao losing 7,000 soldiers, most of his army. The western force under Kung-sun Ho achieved nothing significant. How Wei Ch'ing was able to reach Lung-ch'eng in west-central Mongolia, when the other three forces made no headway from points much nearer, is not explained. Perhaps he made a rapid dash behind the Hsiung-nu armies. He could not have encountered much resistance for he killed or captured only 700 of the enemy. Campaign of 128 B.C. — In the winter of 129 B.C., and again in the autumn of the next year the Hsiung-nu invaded the north- eastern frontiers at Yu-yang, Liao-hsi, and Yen-men, killing or kidnaping more than 3,000 people. Wei Ch'ing set out from Yen- men with a force of 30,000 cavalry, and Li Hsi started from Tai, and at the same time discuss textual differences and debatable terms. The follow- ing paragraphs, then, are a synthesis. Figures are quoted as given. A copy of Herrmann's Historical and commercial atlas of China (pp. 22-23; p. 17, II; and p. 24) is useful; see also Map, this volume. Place Names in Order of Appearance Shang-ku Southern Chahar, just northwest of Peking. Tai Southern Chahar, northeast of present Yii hsien. Yen-men Northern Shensi, west of Tatung. Yiin-chung Southern Suiyuan, southeast of Paotow. Lung-ch'eng In west-central Mongolia, but not precisely located. Yii-yang Northeast of Peking on the road to Kupeikow. Liao-hsi Northeastern Hopei near Shanhaikuan. Kao-ch'iieh Near northwest bend of the upper Yellow River. Lung-hsi Kansu, near Lanchow. Shuo-fang Southern Suiyuan, on the west side of the Ordos along the Yellow River. Ting-hsiang Northern Shansi, near Yuyii hsien west of Tatung. Shang Eastern Shensi, southeast of Suite hsien. Yu-pei-p'ing Northeastern Hopei and southern Jehol. Pei-ti Eastern Kansu, near Huan hsien. Chii-yen Northern Ningsia, southwest of Gashun-gol. Ch'i-lien Mountains Or Nan-shan, the range skirting the south of western Kansu. Hsiu-ch'u Kansu, around present Wuwei or Lanchow. Hun-hsieh Kansu, around present Kiuchuan or Suchow. Chih-yen Mountain In Outer Mongolia, but unidentified. ENSLAVEMENT OF PRISONERS OF WAR 105 with an army of unreported size. Wei Ch'ing killed or captured several thousand, but considering the force involved this was an unimpressive campaign. Campaign of 127 B.C. — In the spring, Wei Ch'ing and Li Hsi again set out, this time from Ylin-chung, farther west. Following the upper course of the Yellow River, first to the west as far as Kao-ch'iieh and then south to Lung-hsi in Kansu, they defeated the Hun King of Lo-fan, killed or captured several thousand, and took more than a million head of cattle. Farther west they defeated two other kings and captured 3,017 people. Because of these successes China took the land south of the river — that is, within the great bend of the Yellow River: southwestern Suiyuan, western Shensi, and eastern Kansu — and established Shuo-fang Com- mandery. Using the river for a defense, China restored the Ch'in dynasty wall built by Meng T'ien, and transported a hundred thousand people from the interior to populate the Shuo-fang territory, which became a new base. Emperor Wu made a laudatory speech detailing Wei Ch'ing's victories, and enfeoffed him as the Marquis of Ch'ang-p'ing with the income from an additional 3,800 households. Campaign of 12 U B.C. — The Worthy King of the West, in com- mand of the western Hsiung-nu dominions, contested this loss of his territory by invading Shuo-fang and killing or kidnaping many of the colonists. Other Hsiung-nu raiding parties invaded the central northern frontier at Tai, Yen-men, and Ting-hsiang, and even penetrated to Shang Commandery. At this time the frontiers were obviously very loosely defined and held. In the spring of 124 China organized more than 100,000 men, led by six generals under Wei Ch'ing, and an attempt was made to demolish Hsiung-nu opposition in the new northwest area. Wei Ch'ing led 30,000 cavalry from Kao-ch'iieh, while Su Chien, Li Chii, Kung-sun Ho, and Li Ts'ai went jointly from Shuo-fang. Far to the east another army under Li Hsi and Chang Tz'u-kung went out from Yu-pei-p'ing with no important results. The armies of Wei Ch'ing advanced rapidly from the wall six or seven hundred li, took the Worthy King of the W^est completely by surprise, and surrounded his encampment in the night. Although the Hsiung-nu king escaped v/ith a few followers, the Chinese cap- tured some ten petty kings, 15,000 men and women, and several hundred thousand to a million herd animals. Wei Ch'ing then led his troops back to the wall and was greeted by an imperial emissary who conferred on him the title of General-in-Chief. After a review 106 SLAVERY IN THE FORMER HAN DYNASTY of the troops Wei Ch'ing returned to the capital where the Emperor lauded him, giving him an additional fief of 8,700 households and making his three sons marquises. Various senior officers whose troops had distinguished themselves also received rewards. This campaign firmly established Chinese control of the region within the great northward bend of the Yellow River, which, in turn, allowed the Chinese to push out through the narrow corridor of Kansu toward the strategic oases in the desert to the west. Campaign of 123 B.C.— Two successive expeditions against the Hsiung-nu center were successful but not conclusive. Early in the spring Wei Ch'ing and six generals — Kung-sun Ao, Kung-sun Ho, Chao Hsin, Su Chien, Li Kuang, and Li Chii — left Ting-hsiang with more than 100,000 cavalry. The Chinese killed or captured only some 3,000 of the enemy, and the army returned to rest at Ting- hsiang, Yun-chung, and Yen-men. In the second campaign two months later the main army killed or captured some 15,000. Ho Ch'ii-ping first distinguished himself as a major in charge of 800 light cavalry. He advanced several hundred li beyond the main force, killing or capturing 2,028 of the enemy. At the same time, however, a small force of some 3,000 cavalry under Su Chien and Chao Hsin stumbled on a much larger force of the Hsiung-nu, and without support was completely defeated. Su Chien escaped but Chao Hsin surrendered with about 800 of his cavalry and joined the Hsiung-nu Shan-yii. Chao Hsin was originally a petty king of the Huns who had surrendered to China some years before and been granted a marquisate. After he went over to the Hsiung-nu he became the principal adviser of their ruler, who gave him a daughter in marriage. The combined total of enemy killed and captured reached 19,000, but the Shan-yii had maintained his own army intact, so the result was not decisive. Campaign of 121 B.C. — In the spring Ho Ch'ii-ping led a select force of 10,000 cavalry from Lung-hsi some thousand li out through the narrow corridor of modern Kansu province. He killed or cap- tured more than 8,000, and seized the sacrificial gold idoP of the King of Hsiu-ch'u. Early in the summer Ho Ch'ii-ping and Kung-sun Ao went out from Pei-ti and Lung-hsi, leading an army of several myriad cavalry. 1 For recent discussions of this event, see H. H. Dubs ("The 'golden man' of Former Han times," TP, vol. 33, 1937, pp. 1-14) and J. R. Ware ("Once more the 'golden man,' " TP, vol. 34, 1938, pp. 174-178). ENSLAVEMENT OF PRISONERS OF WAR 107 Kung-sun Ao lost the road and did not meet Ho Ch'u-ping at the appointed place. Nevertheless Ho Ch'ii-ping pushed out through Chii-yen, then turned south and delivered a crushing blow near the Ch'i-lien Mountains. In his document of praise, Emperor Wu listed 30,200 enemy killed or captured, 2,500 surrendered, and 59 enemy nobility and 63 high military officers captured. Ho Ch'ii-ping lost only 30 per cent of his forces, which was apparently considered low. He was given 5,400 additional households for his marquisate, and two of his subordinates were enfeoffed. Kung-sun Ao was tried for failing to keep his appointment, but was allowed to pay redemption and was made a commoner. Because of Chinese successes in this campaign the Hsiung-nu kings of Hsiu-ch'u and Hun-hsieh, in western Kansu, decided to submit voluntarily to China. By their surrender the empire acquired areas of great strategic importance. An expedition on the extreme eastern flank was a dismal failure. A force of 10,000 cavalry under Chang Ch'ien and another of 4,000 under Li Kuang set out by different roads from Yu-pei-p'ing to attack the Worthy King of the East. Li Kuang's force was sur- rounded and nearly cut to pieces before Chang Ch'ien arrived to rescue the remnants. Chang Ch'ien was tried for delay on the way, but instead of being executed he was allowed to pay redemption and became a commoner. Campaign of 119 B.C. — This last campaign was the greatest triumph of all, for it shattered the armies of both the Shan-yii and the Worthy King of the East. Neither had been really defeated before; the previous great successes had been in the west. Late in the spring an enormous Chinese army of 100,000 cavalry, several hundred thousand infantry and supply columns, and 140,000 horses privately provided, assembled at Ting-hsiang under the joint com- mand of Wei Ch'ing and Ho Ch'ii-ping. Prisoners taken in a preliminary reconnaissance revealed that the Shan-yii was farther eastward. It was therefore decided that Ho Ch'ii-ping should take part of the army out from Tai, farther east, making a double thrust by armies operating independently. Chao Hsin devised the strategy of the Shan-yii. The Hsiung-nu would withdraw northward and attack the Chinese armies while they were still exhausted from crossing the desert. Wei Ch'ing split his army into four or five columns, to meet at a predetermined place. Two columns failed to arrive at the rendezvous. Nevertheless, as soon as Wei Ch'ing had crossed the 108 SLAVERY IN THE FORMER HAN DYNASTY desert with the main force and sighted the Hsiung-nu he prepared for battle. He ordered his baggage train to form a solid ring for an encampment and kept most of his army hidden. Then he led forward a deceptively small body of 5,000 cavalry. Suddenly a great windstorm blew up and the air was filled with sand, so that the armies were invisible to each other. Wei Ch'ing brought up his right and left wings and encircled the Hsiung-nu, who were actually weaker than the Chinese. In the confusion of battle the Shan-yu escaped northwestward with a hundred cavalry. A little later the Chinese, learning of his escape from an enemy prisoner, sent a body of light cavalry in pursuit, but they failed to capture the Hsiung-nu ruler. The battle continued in great confusion through the night, with both sides suffering heavy losses. By morn- ing the Chinese had won a decisive victory, killing or capturing more than 10,000 Hsiung-nu and putting the rest to flight. Wei Ch'ing then advanced with his army to Chao Hsin's base at Chih-yen Mountain in Outer Mongolia, where he captured the Hsiung-nu grain stores. After resting and provisioning his troops, he burned the "city" and enemy stores. While recrossing the desert he met the two generals that had failed to keep their appoint- ment. Li Kuang committed suicide, while the other general was allowed to pay redemption and became a commoner. The total of enemy killed and captured was 19,000. Ho Ch'u-ping's army of picked fighters was even more successful, though the details of the campaign are not reported. Riding out from Tai and Yu-pei-p'ing some 2,000 li, it defeated the army of the Worthy King of the East, killed or captured more than 70,000 persons, and seized three petty kings and 83 high officials and officers. Chinese casualties were 20 or 30 per cent. The combined totals for the campaign of 119 B.C. are given as 80,000 or 90,000 Hsiung-nu killed and captured. The Chinese losses were "counted by the myriad." They also lost more than 100,000 horses. Thus, although they won a crushing victory from which the Hsiung-nu did not recover in a century, they won it at great cost in men and material. Summary. — The biographies of Wei Ch'ing and Ho Ch'ii-ping conclude with summaries of their military achievements: Wei Ch'ing is credited with killing or capturing more than 50,000 in seven campaigns, while Ho Ch'ii-ping is credited with more than 110,000 in six campaigns, during the last four of which he was a general in charge of his own forces. enslavement of prisoners of war 109 Presumptive Evidence for Enslavement of Captives Were the Hsiung-nu captured in these seven campaigns enslaved? In the historical sources we can find presumptive evidence that some captives were enslaved. Then other questions arise. How many or what proportion of the captives became slaves? Were they an important part of the slave population? What was done with those who were not enslaved? Certain facts associated with the surrender of the King of Hun- hsieh and his people lead to the belief that some Hsiung-nu captives were made slaves. In 121 B.C. Ho Ch'u-ping overran most of present Kansu province, the western stronghold of the Hsiung-nu. The Shan-yil, enraged at his heavy losses, summoned the kings of that area to his court, planning to execute them. Instead of going, they decided to surrender to China, and sent an ambassador to announce their decision. Fearing a plot, Emperor Wu sent Ho Ch'ti-ping to meet them. At the last minute the King of Hsiu-ch'u lost heart, and the King of Hun-hsieh murdered him and merged the two hordes. Just before crossing the river some of the King's officers wanted to withdraw, fearing the treatment they would receive. Ho Ch'li- ping immediately galloped up and slew those who wanted to flee. He must have been attended by a strong force, for it is said that he killed 8,000! Those who surrendered numbered more than 40,000, though they claimed to be 100,000. ^ The Chinese government issued 30,000 carts to transport these people with their belongings to the capital. Various local govern- ments along the route were taxed heavily to feed them. The King of Hun-hsieh received a marquisate having an income from 10,000 households, while three of his petty kings and his chief Tang-hu received marquisates of 500 to 700 households each. The govern- ment gave rich rewards to the surrendered people, supplied them with food and clothes, and established them as members of Dependent States in eastern Kansu, Shensi, and southern Suiyuan, at An-ting, T'ien-shui, Shang Commandery, Hsi-ho, and Wu-ylian. The Bureau for Dependent States had to be expanded to handle the in- creased administrative problems. Presumably the people continued their pastoral economy, for the new territory, though farther east, was similar to that which they had abandoned. They also had to help defend their new homeland. For example, in 111 B.C. other 1 For the account of the surrender see CHS, 6, 6a-b; 55, 4b-5a (SC, 111, 4b); and 94A, 8b (SC, 110, 10a). For marquisates and rewards see also CHS, 17, 8a-b (SC, 20, lOa-Ub); 24B, 3b. For the establishments of the Dependent States, see CHS, 6, 6a-b; for their location see Map, this volume; for the Bureau of 110 SLAVERY IN THE FORMER HAN DYNASTY Hsiung-nu invaded one of the places where these Dependent States had been established, and cavalry from T'ien-shui and An-ting was mustered to fight them. This voluntary surrender opened to Chinese political and mili- tary control most of that long salient of western Kansu which was the strategic highway to the oases kingdoms of Chinese Turkestan. Without control of this region any further push westward would have been impossible. Moreover, this surrender represented the first major success in the Emperor's policy of splitting up the Hsiung- nu empire. Now, when the King of Hun-hsieh surrendered he brought along the Queen and the two sons of the King of Hsiu-ch'u as prisoners. The Chinese government seized them and made them slaves {Jf.1). Furthermore, one of Emperor Wu's officials proposed that the whole horde be enslaved. This official, Chi Yen, was indignant because 500 Chinese merchants were sentenced to death for trading with these Hsiung-nu at Ch'ang-an. A law which prohibited ex- porting or selling weapons and iron to the Hsiung-nu at the frontier passes had been invoked against these merchants, though they had traded at the capital. Chi Yen, who was prefect of the western section of the capital, and a noted champion of the people, considered this a great injustice and protested to Emperor Wu (^0): Now, the Hsiung-nu, attacking and blocking the highways and barriers, broke the [treaty of] peace and friendship. China raised troops to punish them, and the [Chinese] killed and wounded were innumerable, while the expenses were counted by the ten billions [of cash]. Your servant foolishly thinks that when Your Majesty obtains northern barbarians they should all be made slaves and be granted to the families of those who died in the army; [also let (whatever)] is captured be given to them, in order to relieve the empire ['s distress] and satisfy the hearts of the people. Now, supposing You cannot do that. The King of Hun-hsieh comes leading a horde of several myriad [to surrender. You] empty the treasury Dependent States see CHS, 19A, 5b (MH, vol. II, p. 523). On the use of cavalry from T'ien-shui and An-ting see CHS, 6, 9a. A little later, Chao Po-nu led a large force of "cavalry of the Dependent States" to fight the Hsiung-nu; in 104 B.C. Li Kuang-li used 6,000 "cavalry of the Dependent States" against the city of Erh- shih (CHS, 61, 3b and 4a, respectively). Under later emperors there was consider- able use of Ylieh, Hu, and Ch'iang cavalry, but these may have been mercenaries. At the funeral of Ho Ch'ii-ping in 117 B.C. soldiers from the Dependent States, dressed in black armor, were lined up along the way from Ch'ang-an to the cemetery at Mao-ling (CHS, 55, 6a). For further information on "Dependent States" during Former Han times see P. A. Boodberg, "Two notes on the history of the Chinese frontier," HJAS, vol. 1, 1936, pp. 283-307 (see pp. 287-288). ENSLAVEMENT OF PRISONERS OF WAR HI and reward them, and send forth good people to wait on them as though serving spoiled children. How can the ignorant people understand that because they traded [with the Hsiung-nu] in Ch'ang-an the civil officials arrest them for exporting goods as though at a frontier pass? If Your Majesty cannot derive benefit out of the Hsiung-nu for relief of the empire, still more [how can You] execute more than 500 ignorant people because of a trifling law? Your servant ventures to suggest that Your Majesty should not take [such a step]. What does the essential proposition in the first paragraph signify? Does it indicate, as some Chinese writers beheve, that captured Hsiung-nu were not enslaved? One argument is that since Emperor Wu refused to enslave the followers of the King of Hun-hsieh he did not enslave Hsiung-nu captives. This argument is invalid because these were not captured Hsiung-nu fighters, but a special group who had surrendered. Another argument seems to be that since Chi Yen suggested that the Emperor ought to enslave all Hsiung-nu he obtained, therefore it was not customary to enslave any Hsiung-nu. This is equally fallacious. Chi Yen brought forward some unusual suggestion which Em- peror Wu rejected. To interpret his suggestion correctly it is necessary to determine which part of it was novel. Since the arrival of the Hun-hsieh horde in Ch'ang-an brought on the trial of Chinese merchants the novelty probably concerned these surrendered people. Chi Yen suggested that, since the Hsiung-nu had broken the peace treaties and caused China enormous expense in men and treasure, all Hsiung-nu that the Emperor obtained should be enslaved. Cer- tainly he did not create the idea of enslaving Hsiung-nu. Prisoners of war had been enslaved during Shang and Chou times, and three Hsiung-nu were made slaves on this very occasion.^ The novelty of his suggestion must have been, therefore, that these surrendered people be treated like captured Hsiung-nu, and be enslaved. In view of the general situation, in which Hsiung-nu power was weakened and China had gained important areas in Kansu because 1 Whether shortly before or after cannot be precisely determined. It should be pointed out that the Queen and two sons of the King of Hsiu-ch'u may have been enslaved as the family of a "rebel" rather than as captives of war. The distinction is probably fictitious, since from the Chinese point of view Hsiung-nu "rebelled" and had to be "chastised" when they broke the peace and fought. For enslavement of captives in Shang times, cf. H. G. Creel, "Soldier and scholar in ancient China," Pacific Affairs, vol. 7, 1935, pp. 336-343 (see pp. 341-342 and references cited), and The birth of China, London and New York, 1936, pp. 130, 214, 282, and index (New York ed.) under "Slaves." For Chou, especially Ch'un-ch'iu and Chan-kuo periods, consult references cited by Ch'en Hsien- hsuan, "Ch'un-ch'iu ti nu-li [Slavery of the Ch'un-ch'iu period]," Shih Huo, vol. 2, No. 5, Aug. 1, 1935, pp. 234-236; Chang Yin-lin, "Chou tai ti feng-chien she-hui (Feudalism in the Chou dynasty)," CHHP, vol. 10, 1935, pp. 803-836 112 SLAVERY IN THE FORMER HAN DYNASTY of their surrender, it is unlikely that Chi Yen thought Emperor Wu really would enslave them. To do so would probably have destroyed at one stroke the possibility of splitting off other parts of the Hsiung- nu empire in the future. People who surrendered had to be rewarded, and the Hun-hsieh folk were treated so liberally as to strain the treasury. Thus, Chi Yen made the suggestion merely as an effective contrast to the proposed execution of 500 ignorant Chinese merchants who had unwittingly broken a frontier law. His proposition also had a second part: Enslaved Hsiung-nu should be bestowed upon families of those who had died in the Chinese army. This should be done in order to relieve the empire's distress and satisfy the hearts of the people. Here too was a novel suggestion. It must have been an extension of current practice. It would be incredible that Chi Yen should come to court and create the whole idea of enslaving all Hsiung-nu — captives and surrendered alike — and of giving them as recompense to Chinese families. Em.peror Wu's rejection of the suggestion does not prove that captured Hsiung-nu were not enslaved. He rejected the preposterous idea of enslaving these surrendered Hsiung-nu, or the idea of giving enslaved Hsiung-nu to the families of slain Chinese soldiers. What- ever elements may have been novel in Chi Yen's proposal it seems evident that the presupposed idea of enslaving Hsiung-nu captives was thoroughly familiar and probably in practice. Two other documents from this period strengthen the conclusion deduced from Chi Yen's proposal to Emperor Wu. In 127 B.C., because of the costs of the first three campaigns against the Hsiung- nu, the government treasuries and arsenals were empty. As a remedy the government appealed to the people to contribute male and female slaves in exchange for tax exemption or increase in honorary rank (36). Since these slaves were to improve the condition of the treasury and arsenals they must have been put to work in the arsenals or at money-making projects, or sold for revenue.^ The latter is the more probable, since female slaves were also requested. (see pp. 806-807); Liang Ch'i-ch'ao, op. cit., p. 528; Duyvendak, op. cit., p. 298. I have not emphasized historical precedent here because the whole picture of slavery before 200 B.C. is much more fragmentary and confused than that of Han times when documentary material, scant as it is, becomes relatively abundant. Classical references do not prove by context which terms refer unequivocally to slaves, and which refer to serfs, convicts, and free or semi-free menials of various ranks. It does not clarify a doubtful point in Han slavery to cite the practice of a period which is even more obscure. Nor do we know whether the custom of enslaving prisoners — if actually practiced in Chan-kuo times — carried through the Ch'in period into Han. 1 Wu Ching-ch'ao (op. cit., p. 264) believes they were sold. ENSLAVEMENT OF PRISONERS OF WAR 113 Since the government was actively seeking slaves it would seem that it must have enslaved those prisoners— captured in the very wars that had depleted the treasury — who showed most adaptability to ser- vitude, and either sold them for revenue or put them to gainful work. During the period immediately after the last campaign the government appropriated male and female slaves "by the thousand [even to] a myriad" from wealthy Chinese of the merchant class. These slaves were put to work on government enterprises {Jf6). Can it be assumed that a government, needing thousands of slaves for work, would pass by the prisoners it captured in war? Aside from these four documents dating from the period of the great campaigns, there is another broad, etymological indication that some prisoners were enslaved. At the end of the Ch'in period Chang Han surrendered his Ch'in army to Hsiang Yli and the rebel- ling nobles, whose officers and soldiers took advantage of the victory to treat the Ch'in soldiers like slave captives {nu lu). The latter feared that if the nobles were unable to conquer Ch'in they would make captives {lu) of them and take them east {6). What did they fear about being taken east as captives? Having been treated like slave captives, was it not that they would be really enslaved? Here the term lu is important. Used in combination with nu, "male slave," it has its derivative meaning of "slave" as explained (p. 99, footnote 3) in connection with the saying: "It is the affectionate mother who has the prodigal son, while the severe household is without fierce slaves [lu]." The combination nu lu appears with similar meaning in two other translated documents. The biography of Tiao Chien, who made his fortune from the work of slaves, opens with the statement: "In Ch'i it was customary to look down upon male slave captives [nu lu], but Tiao Chien alone appreciated and valued them. Rascally and crafty male slaves are what men suffer from. But Tiao Chien alone gathered and employed them . . . ." (i7.) Ssu-ma Hsiang-ju predicted conditions in the southwest — noted as a source for child slaves— if China did not take over the area, saying: ". . . and children and orphans will be made slave captives [nu lu], bound up and weeping . . . ." {A7.) In both these passages lu has by context the sense of "slave." In another instance, Liu Ch'ii, who had murdered two concubines and their female slaves, said in reference to his fear that they would haunt him: "Those captives [lu] will appear again to terrify me . . . ." {6U.) The fact that the word for "captive" is also used as a term for "slave" is only suggestive. It does not by itself prove that captives 114 SLAVERY IN THE FORMER HAN DYNASTY were enslaved. There is, however, the additional testimony of Chin Shao, who lived during the third century of our era. He says the terms tsang and huo mean "those defeated enemy who have been captured and made slaves." (11, footnote 5.) Finally, why did the Chinese take prisoners of war at all? It must have been difficult to bring them back to China, and if they were of no value it would have been much easier to kill them on the spot. No people are more eminently practical or shrewd than the Chinese. The government and officers certainly profited from herds of cattle and other loot captured in warfare.^ Considering the enormious expense entailed in these wars the conclusion is inescapable that some captives were forced to labor as slaves on government works, or were sold. Proportion Enslaved, and Disposition of Others If this be admitted, the next questions concern the numbers enslaved and their proportion to the total slave population. About two hundred Hsiung-nu nobles or officials were taken in the seven campaigns, and Wei Ch'ing captured 3,017 people in 127 B.C., and more than 15,000 men and women in 124. Aside from these specific figures it is quite impossible to know the proportion of the 160,000 killed or captured that were taken alive. The problem thus turns upon the question of how many could have been absorbed. Elsewhere, it is estimated on the basis of a memorial by Kung Yti (89) that the government owned about 100,000 slaves in 44 B.C., when it was both rich and extravagant. This figure is probably close to the maximum for the Former Han period. That the govern- ment owned many fewer slaves during the time of Emperor Wu may be deduced from the great confiscation of private slaves during the years 119-113, when it appropriated several thousand "[even to] a myriad." (-4^.)- This small number nearly swamped the govern- 1 There may have been a regular system of division of loot between officers and the government. Ch'en T'ang was severely censured, or slandered, for his greed in a campaign of 36 B.C. in which he allowed each person to keep what loot he could get, and himself brought in through the wall more loot than was allowed by military law (CHS, 70, 5a and 5b). In 104 B.C. Sun Hsiang was tried for having appropriated other people's prisoners and loot in the battle of Canton (CHS, 17, 13b). It is told that in preparation for the campaign of 119 B.C., 140,000 private baggage and following horses were gathered, aside from the regular cavalry horses and provision wagons supplied by the government (CHS, 94A, 9a). This looks very much as though the generals expected to bring back a great deal of loot. - Possibly this should be read "by the thousand and the myriad," an in- definite figure representing a large number but not yet "several myriad." The point is discussed in Jt.6, footnote 4. This uncertainty is to be borne in mind whenever this document is cited below. ENSLAVEMENT OF PRISONERS OF WAR 115 ment. The slaves were sent off to parks and ranches or given to various bureaus. Many government bureaus had to be enlarged and new ones established. Still, a multitude of slaves (and convicts) was sent away to work on the grain transport. If roughly ten thousand or even several myriad private Chinese slaves, the most useful and tractable type, were absorbed by the government only with difficulty in seven years it seems unlikely that as many Hsiung- nu could have been enslaved in an equivalent period. Yet during the previous seven years, 125-119, the total number listed as "killed and captured" was roughly 138,000, in addition to the 15,000 men and women definitely stated to have been taken alive in 124. Most adult Hsiung-nu males would have been unsuited for jobs in government bureaus or skilled manufacturing; they could have been used profitably only in ranching and gang labor. Yet multi- tudes of criminals and the corvee system regularly provided as much forced labor as the government could employ. As for ranches, the Han chiu i reports (for some unspecified date, it is true) that 30,000 government male and female slaves tended 300,000 horses on thirty- six imperial ranches scattered on the north and west frontiers {93). But could the government have kept many captured Hsiung-nu warriors as actual slaves on ranches near the frontier? They could have escaped almost at will if given the freedom of movement necessary for tending herds of horses. From the purely practical point of view it is doubtful whether the government could have used many Hsiung-nu warriors as slaves under the prevalent system. If the government was unable to absorb any large number did it sell many into private slavery? Here again there is no evidence, and in speculating it is necessary to anticipate later discussion, which shows that most private slaves were domestics performing service functions that native Chinese could do best. Hsiung-nu warriors, reared in a totally un-Chinese background and speaking a foreign tongue, were fitted for few normal slave occupations. The possi- bility of their escaping must have made them a risky investment, especially in view of the ill-defined and poorly guarded frontier. Probably only the more attractive Hsiung-nu women and children could have been absorbed naturally into the system of private slavery as it is reconstructed from extant documents. The very fact that numbers of killed and captured are not separately reported even in direct quotations suggests that the distinction was not historically important, and that it did not make an important difference to the state, economically, whether enemy were slain or captured. 116 SLAVERY IN THE FORMER HAN DYNASTY Furthermore, there is no observable change in the slavery system during or shortly after the wars. One of the outstanding charac- teristics of Chinese slavery from the beginning of Han times was the absence of a marked racial character of slaves. Though non- Chinese slaves were fashionable, they apparently comprised only a small element in the slave population. Those from south China were always the most important "foreign" type, and enemy captured in the southwest or southeast are much more likely to have been enslaved than Hsiung-nu. But wars in those regions were brief, comparatively few prisoners were taken, and the texts are equally reticent about disposition of enemy captives. Whatever demand there was could probably have been filled easily by slave raids, or by sale to Chinese traders of captives taken in intertribal wars. Thus, while large numbers of Hsiung-nu were captured during the Han period, and some were enslaved, it appears altogether unlikely that they constituted an important part of the servile population of China. What became of prisoners who were not enslaved? The histories do not tell. Prisoners may have been exchanged for Chinese captured by the Hsiung-nu, but this is not reported. If so, the Chinese had the advantage, for on the whole they were more successful in the wars. Perhaps prisoners were ransomed. Neither alternative seems likely in view of the Chinese policy of keeping the Hsiung-nu weak and divided. Captured Hsiung-nu probably could have been sold by the Chinese to Central Asian kingdoms farther west, but there is no report of this. However, the Chinese did sometimes give Hsiung-nu prisoners to their Central Asian allies after a joint campaign. Another alternative would have been for the government to have established them along the inner frontier in regions distant from the place of capture, perhaps among colonies of surrendered Hsiung-nu or their allies. Given grazing land of their own to defend, they could probably have been induced to submit. Several facts about the frontier make this final hypothesis seem reasonable. It was not a national boundary line peopled on either side by mutually hostile nations, but rather a broad zone of mingled Chinese and Hsiung-nu culture, where intensive agriculture gradually faded off and changed into extensive agriculture and herding, then into pure nomadism. Political control fluctuated with the fortunes of war and the relative strength of the imperial Chinese state and of the confederations in Mongolia. Many of the best Chinese generals ENSLAVEMENT OF PRISONERS OF WAR 117 came from the frontier; some were even "Hsiung-nu" — which means that they had once recognized the suzerainty of the Shan-yil, and then had come over to China. The phenomenon of quick and easy "surrender" to China or "flight" to the Hsiung-nu merely manifests the fluid condition of the frontier. Another evidence is the ease with which Hsiung-nu raiders penetrated deep into territory that the Han empire claimed as its own. The frontier was most strongly Chinese when China won cam- paign after campaign, slaughtered and captured Hsiung-nu, split them up into warring groups, and drove them back. When China was on the defensive, or internally weak, the pattern reversed; the frontier turned in upon the empire. Thus it was exactly those periods when China captured far more enemy than it could absorb into slavery that it was probably safest to establish many of them, as we have postulated, along the inner frontier as a buffer and an asset. V. ACQUISITION, HEREDITARY SLAVERY, AND MANUMISSION Free people were enslaved either to the government or to private individuals. Initial enslavement, depending on the process, generally led exclusively to one or the other type of ownership, but there was a real fluidity in transfer of slaves, in both directions, between the government and private masters. Individual slaves did not necessarily belong permanently in one category or the other, and the process of acquiring slaves was not identical with the mechanism of enslavement. Corresponding to the constant recruiting of new slaves was the opposite process of liberation. Enslavement, transfer, and manu- mission are all part of a fluidity of status which appears to charac- terize Chinese slavery, and indeed Chinese society, at this epoch. In political, economic, and social matters China was in a period of youth, of rapid change and growth. Slavery was a part of this flux, although we cannot determine the proportion of the slave population, or the number of individuals, involved. Acquisition of Slaves The government acquired privately owned slaves by confis- cation, by gift, and probably by purchase. Favored individuals received gifts of slaves but most people got them by purchase. Laws threatening confiscation of slaves indicate that the process was recognized, even though histories do not report final details.^ On the other hand, confiscation of property is frequently reported, but only occasionally do we hear of the kinds of property.- When the government made a total confiscation, slaves must often have been included, as when on Emperor Wu's order Wang Wen-shu confiscated all the property owned by a thousand families of "tyran- nical gentry" in a single commandery US). Likewise, when impor- tant families became too powerful, or when they were successfully accused of menacing the imperial house, the whole clan might be stamped out. In such cases the property was probably taken over, 1 Emperor Wu's counselors advised him to confiscate the fields and slaves of merchants daring to own agricultural land U5). When Emperor Ai considered restrictions on the number of fields and slaves that could be owned by people of various ranks it was proposed that after a specified time excess lands and slaves should be confiscated (109). 2 Cf., for example, CHS, 45, 6a-b; 51, lib; 90, 6b. 118 SLAVE TRADE AND MANUMISSION 119 even though the histories fail to tell. More than usual detail is accorded to Tung Hsien, the darling on whom Emperor Ai loaded imperial treasures. After the Emperor died Tung Hsien was forced to commit suicide, and the imperial government confiscated and sold his belongings. The texts do not itemize the property taken, but we know that Tung Hsien was the owner of many slaves {109, 117, 118,119) soil is fair to assume that the government acquired his slaves even if it sold them immediately for revenue. The same thing probably occurred when Ho Kuang's wealthy slave-owning family was -stamped out; and so, too, with unnumbered families, great and small, throughout the dynasty. One confiscation was pushed through on a large scale. As part of the policy of suppressing merchants, and as a scheme to replenish the imperial coffers, investigators all over the empire tried people who had accumulated fortunes in "secondary" occupa- tions. Between about 119 and 113 B.C. they confiscated personal property by the hundred million cash, slaves by the thousand up to a myriad, and large areas of farm land (4-6) ^ Some years earlier Emperor Wu tried a different plan to augment the treasury by allowing plebeians to give male and female slaves in exchange for life exemption from taxes and corvee duty. Those already enjoying exemption could contribute slaves for an increase in rank (36). This is very vague. How many slaves did a person have to give? Did anyone accept this offer, and was the condition of the treasury improved? Information to answer these questions may already have been lacking when Han histories were written, for routine records of fact, invoices, treasury reports, minor census records and the like — the intimate, day-by-day records of govern- ment in operation — were of a transitory nature, and were probably easily scattered or destroyed when they became out of date. Ap- parently the government found this plan difficult to administer, for a new one, introduced only four years later, allowed people to buy titles of "military merit" at specified prices for various grades. "Military merit" entitled the owner to reduction of two degrees of punishment if he committed a crime. As early as 167 B.C. Ch'ao Ts'o had advocated to Emperor Wen the plan of granting honorary ranks in exchange for slaves who would be sent to colonize the frontier. He also suggested that criminals be allowed to atone by giving slaves. "The Emperor followed his counsel and enlisted 1 Ma Ch'eng-feng (An economic history of China, vol. 2, p. 248) considers this a unique case based upon a special emergency law designed simply to authorize confiscation of what the government needed for prosecution of its wars. 120 SLAVERY IN THE FORMER HAN DYNASTY people to remove to beneath the Barrier." (19.) Does this mean that the plan was followed in detail? Did the populace and criminals respond? Again the records are silent. In the opposite direction, slaves went to private individuals as customary grants, rewards for outstanding services, or marks of imperial favor. Probably princesses received a regular allotment when they left the palace. This is deduced from a document telling of a pseudo-princess, treated at court like a real one, who was given a "princess' fields, houses, and male and female slaves." (88.) Emperor Wu gave his long-lost half-sister a million cash, three hundred male and female slaves, a hundred ch'ing (465 acres) of government farm land, and a first-class mansion (31). Because she was not a daughter of Emperor Ching she was not made a princess, but instead became Baroness Hsiu-ch'eng. Chao Fei-yen, who later became Empress, was once given to the Princess of Yang-a (101). Furthermore, princesses are often mentioned as slave owners {26, 27, 39, 61, 109). Probably they all owned scores of slaves, for they were daughters of emperors, maintained elaborate households where they entertained their royal or imperial brothers, and normally married marquises of other surnames than Liu.^ The government occasionally gave slaves to noblemen in reward for services to the state, though grants of money and land were more common. Ho Kuang received 170 male and female slaves, together with rich gifts of gold, cash, silk, horses, and a first-grade mansion {70). Shih Tan was given slaves "numbered by the hundred" because two generations previously his family had pro- tected and reared the imperial infant who became Emperor Hsiian {97). The slaves given to the pseudo-princess, mentioned above, should actually be counted among rewards for services. Sent as a bride to the Wu-sun ruler, she was the spearhead of Chinese diplomacy in that far northwestern country for nearly four decades. Finally at the age of seventy she begged to return to spend her last years in China and be buried in her native soil. When she arrived at the capital the Emperor duly rewarded her. Other imperial grants signalized special favor, and allowed select people of no private means to live in proper luxury. For a brief period a charlatan magician named Luan Ta convinced supersti- 1 There is no evidence to show that their brothers, too, received a regular allotment of government slaves when they were made kings, for no Former Han record tells of a king granted government slaves. The Latter Han cases all describe special gifts in no way connected with enthronement. Cf. HHS, 72, 3a; 72, 5b; 80, 4a; 85, 2a. SLAVE TRADE AND MANUMISSION 121 tious Emperor Wu of his supernatural abilities. The Emperor gave him the title "General of Magicians of the Earth," made him a marquis, presented him with the highest-class mansion for a marquis, a thousand "youths," a princely chariot, extra carriages and horses, and furniture to fill his house. He was even married to one of the Emperor's eldest daughters. Then, convinced that Luan Ta was an imposter, the Emperor had him executed the next year! {J^9.) Other imperial darlings like Chang Fang and Tung Hsien received lavish gifts, and in all likelihood slaves, for they were noted slave owners. When the child Emperor Chao was placed upon the throne his previously unimportant maternal grandaunt received two million cash, male and female slaves, and a mansion and houses to give her "abundance" {58). Wang Mang presented slaves and other gifts of enormous value to the inconspicuous family from which he chose his second Empress (133). This distribution presumably went on all the time, especially to lowly families of the girls who became empresses or imperial favorites. Slaves were indispensable around the court. Buying and Selling Slaves Extant documents on sales of slaves are at once fragmentary and suggestive; they reveal faint outlines and hint at more, but leave the total picture vague. We should like to know who sold slaves and by what methods; about professional dealers and syn- dicates; about sources of supply, and markets for distribution. We want to know the relative importance, in terms of frequency and volume, of dealer sales and owner-to-owner sales, and of wholesale and retail selling; and whether the government regulated and authenticated transactions, and levied sales taxes. We should also like to know, in relation to general Han economy, something about prices, fluctuations in the market, profits, and speculation, and above all, how important slave-trading was in the nation's commerce. Sales were made both by direct transactions between one owner and another and by dealers in slave markets. Wang Pao provides the classic account of direct sale in his semi-humorous account of the purchase of Pien-liao (83). A resident of Chengtu in Szechwan, Wang Pao went in 59 B.C. on business to the Chien River where he called on the widow Yang Hui, who owned a male slave belonging formerly to her husband. Wang Pao requested that the slave be sent out to buy some wine, but Pien-liao refused to go, claiming that the agree- ment or contract with his late master specified that he had only to guard the house. Furious, Wang Pao decided on the spot to buy 122 SLAVERY IN THE FORMER HAN DYNASTY the slave, and the widow agreed. The sale contract, a unique docu- ment, gives the exact date, names and residences of the buyer and seller, name and description of the slave, and the agreed price of 15,000 cash. These items would form the essential core of any- such transaction — the bill of sale and evidence of ownership.^ There is no term for slave dealers, as distinguished from other sorts of merchants, in Han literature. Information about volume of transactions is too scanty to judge whether sales of low-priced slaves supported a group of specialists. Volume, the factor most important for profits in that field, would be a product of urbaniza- tion and the geographical concentration of slave-owning classes, as well as abundant sources of supply. Probably the demand for slaves in the capital district, with its congregation of noble and wealthy slave-owning families, produced such constant turnover that slaves did not eat up dealers' profits between time of purchase and sale. A few other thickly populated areas, particularly the western Shantung and eastern Honan region, might have supported a special- ized trade also. Elsewhere, and early in the Han period, most slave sales were probably handled by agents who simply brought buyer and seller together, or by merchants and other business men as an occasionally profitable sideline. The account of Chi Pu's sale illustrates the point. He was seeking to escape the wrath of Emperor Kao because he had been one of Hsiang Yti's generals. A man named Chou, who lived at P'u- yang on the Shantung-Hopei border, disguised Chi Pu as a slave, and sent him with a score or more Chou family youths to be sold by Chu Chia, who lived at Lu about 100 miles east. Careful reading of Chu Chia's biography, and of this particular incident, reveals 1 The only similar record (cited by Lao Kan, "The system of slavery during the two Han dynasties," p. 2) is the brief item excavated at Chii-yen, which gives the price of two young male slaves as 30,000 cash, and of one adult female slave as 20,000. Cheng Hsiian of the second century tells in his commentary to a passage in the Chou li concerning government supervision of markets that two kinds of contracts were used. Long ones covered sales of people, horses, and cattle, while short ones were used in sales of utensils. Cf. Edouard Biot, trans., Le Tcheou-li, 2 vols., Paris, 1851, 1. 1, p. 52, footnote 3 ; also p. 318, footnote 1. This is good Latter Han evidence on slave contracts, and probably of government supervision. The T'ang lii su i contains several items showing that in T'ang times sales contracts for slaves had to be authenticated by the government. Prices of slaves are interesting only in relation to prices of other commodities, concerning which little is known. Attempts to figure commodity prices for the Western Han period are unfortunately almost worthless, for they are generally based upon two or three references for any single commodity, either undated, or miscellaneously scattered over a 200-year period. Cf., for example, Ch'ii Tui- chih, "Hsi Han wu chia k'ao (The prices of commodities during the Western Han dynasty)," YCHP, vol. 5, 1929, pp. 877-881. SLAVE TRADE AND MANUMISSION 123 that he was not a professional slave trader. He was most famous as a protector of political refugees, which is the reason why Chou was confident that he would harbor Chi Pu until the Emperor could be induced to pardon the refugee. But the stratagem demanded that every detail of the sale appear perfectly natural, for the Emperor had threatened to execute anyone who aided Chi Pu. Hence we must conclude that there was nothing unusual or suspicious about sending a group of slaves a hundred miles for Chu Chia to sell. By profession, Chu Chia was primarily a large-scale farmer who employed innumerable hired (or indentured) workmen. He recog- nized and bought Chi Pu, hid him in a shed in the fields, and had him work with other farm-hands so as to appear to be one of them. Being on friendly terms with Kao-tsu's influential associate, Marquis Hsia-hou Ying, he visited the marquis for several days at Lo-yang, and persuaded him to intercede and win Chi Pu's pardon. From these assembled facts it is clear that a well-established landlord- business man might also traffic in slaves (8).^ This was at the very beginning of the period. Another transaction eighty years later may indicate a more developed sales system. About 120 B.C. Ho Ch'ii-ping was passing through P'ing-yang in western Shansi to take command of his army. He had never met Ho Chung-ju, his father, so the latter was brought to the hostel. Then the general liberally bought his father fields, houses, and male and female slaves, and hurried on to the wars {1^2). This quick transaction, involving land, houses, and slaves, may have been handled by agents who knew of real estate for sale and could speedily assemble a group of slaves. Apparently some merchants did specialize in slaves for the luxury trade. They bought children, especially young girls, trained them as entertainers, and then resold them. Wang Weng-hsu, the mother of Emperor Hsiian, was such a girl. The son of a marquis took her from her parents under an agreement to rear her in exchange for her services, trained her as a singer and dancer, and then sold her to a merchant who was in the business of handling entertainer slaves. A member of the suite of Emperor Wu's Heir-apparent was sent out as an agent to find trained girls for his patron's palace, and bought Wang Weng-hsli with four others from the merchant {55). Well-trained girls commanded very fancy prices, as is suggested by the boastful remark of the King of Chi-pei around 160 B.C., that 1 See his biography in SC, 124, 2a, and CHS, 92, lb-2a. The incident appears also in SC, 100, la. 124 SLAVERY IN THE FORMER HAN DYNASTY he had paid 4,700,000 cash for four girls especially clever at doing tricks (21). He said that he bought the girls at "the common people's place," and even though a doctor pronounced one of them incurably ill, he did not sell her at "the place for nobles." The first of these "places" (so) must have been a market where commoners sold goods, and the second, a market for the nobility. Chia Yi described such market places a decade earlier, saying: "Nowadays people who sell youths dress them up in embroidered clothes and silken shoes with the edges all embellished, and put them into pens." (16.) Slaves dressed like this must have been for the luxury trade, and the end of the passage shows that the "youths" were singers and enter- tainers. Wang Mang briefly described slave markets in his fulmina- tion against the evils of the Ch'in dynasty perpetuated by the House of Han. Accordingly, his statement that male and female slaves were sold in the same kind of pens as were cattle and horses (122) referred equally to conditions in his own day. Instances of kidnaping, purchase of famine victims, slave-raiding and importation of foreign slaves indicate a constant demand perhaps exceeding the normal supply. Slave-raiding and transportation of foreign slaves to the capital suggest some sort of professional organization. The youths acquired by people of Chengtu and Chungking from southern Szechwan and Yunnan (10) went hundreds of miles overland to reach Ch'ang-an.^ Wang Mang's order that "those officials or plebeians who dare traffic in frontier people shall be publicly executed" (126) suggests organized trade in slaves from the frontier. People who wanted to escape the turmoil and famine there perhaps applied to dealers who arranged for their transporta- tion to, and sale in the interior commanderies. This system may have been operated by a syndicate in co-operation with officials. Since Wang Mang wanted to prevent depopulation of the hard-won frontier, he directed his law against those who made it possible for people to leave rather than against the frontiersmen themselves. New slaves transported overland were probably convoyed in parties like groups of convicts, who traveled on foot, with shackles, to work at the frontier or on government construction. Both convicts (t'u) and male slaves- sometimes had their heads shaved 1 For the routes, cf. Albert Herrmann (Historical and commercial atlas of China, p. 20) and L. H. D. Buxton (China, the land and the people, Oxford, 1929, map, p. 141, and discussion, p. 144). - For slaves, see 8, 12, 131, and 132. For convicts, see CHS, 23, 6a and 10a; 90, 2a; and 97A, 2a; also HFHD, vol. I, p. 117, footnote 1. SLAVE TRADE AND MANUMISSION 125 and wore coarse clothes and iron collars as marks of identity. There is no mention of special slave guards, but a hint is provided by the story of Liu Chi before he became Emperor Kao. As a petty official he had to escort a group of convicts from northern Anhwei to Li Mountain near the Ch'in capital some 600 miles distant. He was personally responsible for each convict, and when a few escaped he untied the rest and absconded.^ Later in the period slave traders must have developed a much better system than this primi- tive method of escorting convicts. Since slaves represented a financial investment it may be assumed that those transported against their will were also closely guarded. When females and docile young slaves for the luxury trade were sent from one place to another they went in carts (or probably by boat where possible), for it was important that they arrive in good condition. Mr. Chou sent his household youths by cart the hundred odd miles to Lu (8), and the merchant who bought the female entertainer slave, Wang Weng-hsu, sent her by horse to her new home (55). There is no information to show whether the regular land routes were equipped with special lodgings for slaves in transit comparable to the system of guest houses and lodges established along principal highways for the use of the Emperor or members of the nobility, or the post stations which were used by girls being sent to the imperial seraglio.- These fragments of information about contracts, prices, markets, dealers, and trade routes fit together to form only a vague picture of the slave trade. A marked difference must have existed between selling methods at the capital and in the provinces. Furthermore, we must assume a development from the beginning of the Han period, only shortly removed from the feudalistic Warring States period, to the time of Wang Mang after two centuries of economic growth. By the end of the period slaves were regular commodities, freely sold, and commanding standard prices. When Emperor Ai attempted to limit the number of slaves owned by people of various classes in society, stipulating that after three years all slaves over the specified limits would be confiscated, the mere proposal threw prices into a slump (109). A few years later when Wang Mang attempted to abolish buying and selling of fields and private slaves, 1 HFHD, vol. I, p. 34. 2 When the girl who later became Empress Tou was being sent to the palace she started her journey from such a "transfer house." Cf. lit, and footnote 13 for commentary. 126 SLAVERY IN THE FORMER HAN DYNASTY opposition was so great that he had to repeal the law in three years. The antagonism aroused during that short time contributed to his downfall ten years later {121, 12 J^, A22). The slave trade was not a monopoly of private dealers, for the government itself sold slaves. Near the end of the dynasty the Grand Empress Dowager nee Fu sent an internuncio to various government bureaus to purchase female slaves, buying them at a low price. When he bought eight girls from Mu-chiang Lung's bureau, Mu- chiang Lung formally protested to Emperor Ai, stating that the Grand Empress Dowager had paid too little and asking the Emperor to readjust the price {116). The notable point about Mu-chiang Lung's protest is that his only complaint concerned price. Ap- parently sale was so normal that prices were well-established and understood. The appeal by Emperor Wu, more than a century earlier, that plebeians contribute private slaves to help meet a grave deficiency in the treasury and arsenals suggests, as noted before, that even then the government sold slaves for revenue {36). Secret sale of government slaves and other valuables — a clear case of graft- — was cause for executing the governor of northern Annam in 54 B.C. {86). Hereditary Slavery In the Chinese family system certain types of property were owned by that part of the family group which clung together as a single economic unit. Control of real estate, particularly farm land, was vested in the family head, who held it in trust as a patrimony and was expected to pass it on intact to his eldest son or successor. Sale of land was decided upon by the adult members of the family. If brothers or nephews of the family head established separate economic households they could demand a share of the land. Ownership of slaves was apparently of the same order. Many accounts speak of slaves belonging to a family or household. That they were part of the family estate in T'ang times is indicated by the law on freeing private slaves, which specified that the head of the family must give the slave a document, co-signed by the eldest son and others, and turn over a similar document to the local gov- ernment for validation and filing;^ that is, the son or others who would normally inherit the slave had to agree to his manumission » From the Tang lii su i, ch. 12. See also Liang Ch'i-ch'ao, "System of slavery in China," p. 549, and Wang Shih-chieh, "The Chinese slavery system," p. 324 (translated by Toni Pippon: "Beitrag zum Chinesischen Sklavensystem," p. 127). SLAVE TRADE AND MANUMISSION 127 in writing and file their agreement with the government to prevent later argument. Several cases in the Hou Han shu illustrate the principle of family ownership and the procedure of inheritance at that time. The incidents were recorded to illustrate the character or misfortunes of the individuals in whose biographies they appear, and not to reveal anything about slavery. When Hslieh Pao's parents died some time before a.d. 121, his younger brother's son asked for a division of the property. This indicates that Hslieh Pao had become head of an economic family which included the son of his deceased younger brother. Unable to prevent his nephew from leaving, he divided the male and female slaves into two groups, taking for himself the old ones, to whom he said: "You have worked with me for a long time; you cannot be sent av/ay." He likewise took the poorer fields and sheds and the worn-out household furnishings. ^ The curious procedure which Hsli Wu used to help his younger brothers become officials is a peculiarly Chinese illustration of character. Hsli Wu was the grandfather of an official who served Emperor Ho (a.d. 89-105); therefore the present event probably occurred early in the Latter Han period. He was already a prefect, but his two younger brothers were not yet noted. He therefore proposed that they take their shares of the patrimony and live separately. In the division Hsli Wu himself took the fertile fields, large houses, and sturdy male and female slaves. The people of the countryside all admired the humility and self-denial of the younger brothers and despised Hsli Wu for his greed. Because of this both younger brothers were selected and advanced for office. Only then did Hsii Wu call the family together and in tears reveal the whole scheme. Under his management the property had increased three times in value. He offered it all to his brothers. "Therefore the whole commandery applauded; far and near proclaimed him; and he advanced to the position of Treasurer of the Ch'ang-lo [Palace]." - Chou Tang, the scion of a very wealthy family, was orphaned as a child and reared by a kinsman who mistreated him and refused to give him his family inheritance when he came of age. He there- fore sued in the district court, forcing his kinsman to turn over the estate. Then he distributed his property among his fellow clansmen, 1 HHS, 69, lb. The Feng su Vung i, 4, 8b, tells the same story, but calls Hslieh Pao by his Tzu, "Meng Ch'ang." This version is presumably earlier than that in HHS, and has the same purpose — that of illustrating virtuous character. 2 HHS, 106, 5b. 128 SLAVERY IN THE FORMER HAN DYNASTY freed all his male and female slaves, and went to Ch'ang-an as an itinerant scholar.^ This occurred shortly before Wang Mang usurped the throne, and is evidence that courts upheld the right of a son to inherit family slaves, who were part of the estate to be held in trust during his minority. Going back earlier in the Former Han period, there is the case of the widow Yang Hui, who had her late husband's slave Pien-liao (who haunts this work!); and the widow of Ho Kuang, who owned her husband's household slaves (83 and 72). Presumably Ho Kuang's widow would have passed them on with the estate to his son and grandnephews had not the whole family been stamped out a few years later. In 115 B.C. Hsia-hou P'o was tried and deposed for having had sexual relations with his father's personal female slave (4.8). But his father had been dead eighteen years at the time of the trial. It is clear that Hsia-hou P'o had inherited the bondswoman, but since she had been his father's mistress his relations with her were incestuous. Here Chinese customary morality intervened to limit his rights of possession by inheritance. The converse of slave inheritance is hereditary slave status. In Han times the child of two slave parents was apparently a slave automatically. Grand Judge Chung Yu clearly indicated the hereditary status of government slaves late in the Latter Han period, in his statement on the practice of tattooing enslaved families of criminals: "The genuine male and female slaves of today had ancestors who originally committed crimes. Even though a hundred generations have gone by, still they have tattooed faces [as a sign of] submission to the government." ^ Already in Ch'in times the government owned hereditary slaves, for in 209 B.C. the last Ch'in ruler ordered the Treasurer of the Privy Purse to free the convicts and "born male slaves" working at Li Mountain (3). Yen Shih-ku used the seventh century expression chia sheng nu, "house-born slave," to explain the Han term nu ch'an tzu, translated "born male slave." It is therefore particularly interesting to find his expression used contextually in a T'ang will dated a.d. 865, and discovered by Sir Aurel Stein at Tun-huang. By the terms of the will a nun bequeathed her only property, a "house-born" female slave, to her niece {3, footnote 4). 1 HHS, 113, 2b. 2 San kuo chih, Wei chih, 12, 4a. See p. 84 and footnote 2 for the context of his statement. SLAVE TRADE AND MANUMISSION 129 Ts'ao Hsiao, a government slave woman, had a daughter named Ts'ao Kung, who also belonged to the government. Ts'ao Kung was probably born into slavery although the record does not say so. The alternative — that the girl and her mother became government slaves together — is most unlikely because the mother and the girl's best friend testified in the careful judicial investigation about the murder of Ts'ao Kung's son. This child was the son of Emperor Ch'eng, and might have been heir to the throne. Had Ts'ao Kung once been free the fact would have been reported and probably emphasized, in so important a case {107). Wei Ch'ing complained to a fortune teller that he was "born as another's male slave." {26). He was the son of a slave woman and a free man to whom she was not married. The position of her children gives the clearest internal evidence of hereditary slave status, but is discussed in the chapter on status as part of the subject of marriages between slaves and free people. Manumission of Government Slaves We do not know what proportion of the slave population at any moment had been born free, nor what part of any generation of slaves won their freedom. Yet references to manumission are casual enough to make the act appear not unusual. This is not to say that individual slaves had mathematically favorable chances of becoming free men. On the basis of historical records we can only describe the methods of liberation without much detail about actual process, and attempt to deduce some of the motives for the cases that are preserved. From these points of view the distinction between government and private slaves is important. Most manumissions of government slaves affected groups of people in specified categories. Indeed, there are only two records of the liberation of individuals. In the first instance Emperor Wu personally freed Chin Jih-ti, the son of the Hsiung-nu King of Hsiu-ch'u. Noticing the boy because of his dignified bearing and because the horses in his charge were fat and well groomed, the Emperor questioned him and was impressed by the straightforward account of his background and enslavement in 121 B.C. Emperor Wu immediately freed him and granted him a ceremonial bath, robe and cap {J+l). The ceremony was not part of the process of manumission; it was the process by which a commoner became a Gentleman. Thus Chin Jih-ti rose from enslavement to honorary rank almost in one step. 130 SLAVERY IN THE FORMER HAN DYNASTY A woman named Tse, who worked in the palace, was the other government slave known to have been individually freed. About twenty years before, she had assisted in caring for a child (later Emperor Hsiian) whose father and grandfather were killed in the scandalous witchcraft affair in 91 B.C. (p. 44). This child was secretly reared and finally placed on the throne as Emperor Hsiian. In 63 B.C. the slave woman persuaded her commoner husband to sub- mit a memorandum on her case to Emperor Hsiian, who was ignorant about many of the facts of his childhood. After her plea had been investigated and substantiated, she was freed and made a commoner by imperial edict, received a gift of 100,000 cash, and had an imperial audience (80). Her claim to historical mention lies solely in the fact that she told the Emperor how Ping Chi had saved his life, and this revelation finally brought Ping Chi his due reward. In discussing government slaves, Wei Hung states that they could purchase their freedom for a thousand myriad cash (92). This figure seems very high, and "thousand" may be a scribal error for "ten." The interesting point is that government slaves had sources of revenue with which to free themselves. The only direct testimony to this is the complaint, lodged by the Worthies in the great debate of 81 B.C., that government male slaves accumulated great fortunes (60). Even without such a statement it would be possible to deduce from the duties of government slaves, discussed below, that they had plenty of opportunities for grafts and tips. Whether or not they received stipends, however, is unknown. Five documents report manumissions of government slaves in groups. Four were specific groups and one was general. How and why did the government free its slaves? The cases are as follows: First. — In 209 B.C., when Ch'in was imperiled, the Second Emperor ordered the Treasurer of the Privy Purse to free the convicts and male slaves working at Li Mountain and send them out to fight the invaders (3). Second. — In 160 B.C. Emperor Wen proclaimed a general amnesty and also freed government male and female slaves, who there- upon were to become commoners (2Jf). Third. — In 140 B.C. Emperor Wu pardoned the enslaved families of leaders of the Rebellion of the Seven States (30). Fourth. — In 7 B.C. Emperor Ai sent out to be married from his predecessor's seraglio all the Palace Women who were thirty years of age or younger. He freed and made commoners all government slaves fifty years of age or older (110). SLAVE TRADE AND MANUMISSION 131 Fifth. — In A.D. 31 Emperor Kuang-wu freed and made commoners all those who, during Wang Mang's time, had been seized and made slaves not in accordance with the old laws (135). The first amnesty is reported both in the Ch'in Imperial Annals and in a biography. The four from Han times appear only in the Annals, where they are treated as imperial decisions, with no descrip- tion about the discussion leading up to each, and no suggestion of motive. Nowhere is it stated by what process the slaves were freed ; indeed, only the first case, because it is recorded in a biography, contains evidence that the order was carried out. Dry statements of fact are like precipices beyond which there is nothing but mist. What happened after edicts of liberation? Was there any ceremony of manumission? Were names struck from the ledgers of bureau and palace property? Did the slaves get credentials of freedom? Could they work as free men in their old slave jobs? Were they really free to go if they were useful cogs in bureau routine or valued personal retainers? Did they receive cash grants to start life anew? No single or general motive covers all these manumissions. In the last one Emperor Kuang-wu tried to re-establish justice by freeing people made government slaves under the usurper's laws. In the first case the government offered a bargain: the slaves and convicts would be freed if they fought to defend the dynasty. They were a tough lot, men who had spent years at hard labor, probably having the esprit de corps of convicts; and since they helped Chang Han defeat an army reported as numbering a thousand chariots and a hundred thousand foot, they must have been stout fighters and numerous. Many probably had to stay in the army which Chang Han treacherously surrendered two years later and which Hsiang Yu slaughtered on that grim night before he invaded Ch'in (6). The same reason for manumission, a bargain to be earned by special service, appears several more times during the Former Han dynasty. When Han Hsin planned a revolt against Kao-tsu, who was conducting a campaign against Ch'en Hsi, he plotted to forge an imperial edict pardoning all the government convicts and male slaves, who would then form an army to attack the Empress and the Heir-apparent (13). The plot was foiled, but the idea of manumission in return for fighting is clear. Convict levies were used in wars against the southern and western regions, and government slaves may have served with them to win their freedom. Wang Mang "enlisted" levies of criminals sentenced to death and private male 132 SLAVERY IN THE FORMER HAN DYNASTY slaves to fight as shock troops against the Hsiung-nu {130). Whether contributed in exchange for honorary rank or requisitioned, those who survived the assaults against the Hsiung-nu surely won or took their freedom. Likewise, any private slaves contributed to the government as colonists for the frontiers, following the plan proposed by Ch'ao Ts'o (19), must have been freed in compensation. Two liberations of government slaves in the Latter Han period help to establish the motive for the third manumission, in 140 B.C., in which Emperor Wu pardoned the families of rebels. In A.D. 106 all people belonging to the imperial house or noble clans who had been enslaved for crimes since A.D. 55 (and their descendants) were ordered pardoned and sent away.^ In 110 Emperor An freed all those who had been enslaved since A.D. 76.- The decrees of 140 B.C. and A.D. 106 are parallel in two respects: they occurred during the first year of a new ruler, and they freed members of the imperial clan who had been in servitude for many years. Together, the three cases of manumission form a type, falling into the class of imperial amnesties. During the period before a.d. 9 there were eighty-four general amnesties, or about one every thirty months. Thirty-three others pardoned special classes or groups of criminals.^ Most general amnesties celebrated imperial enthronement, establishment of an empress or heir-apparent, or the adoption of a new reign period. Others signalized the observance of auspicious omens or the occur- rence of Heaven-sent disasters such as earthquakes, fires, and eclipses. Amnesties were considered acts of imperial kindness designed to win popular favor by remitting taxes or corvee duty, halting investi- gations of crimes, and freeing prisoners. The manumission of enslaved families of rebels fits perfectly into this pattern. Did other govern- ment slaves win their freedom in this way, or was Emperor Wu's pardon unique during Former Han times? It is at least unique in having been reported. 1 HHS, 4, 9a. The act was done on behalf of the baby Emperor Shang by the Empress Dowager nee Teng. Cf. HHS, lOA, 9a, and Nancy Lee Swann, "Biography of the Empress Teng," JAOS, vol. 51, 1931, pp. 138-159 (see p. 146). The same edict ordered that the names and duties of all male and female slaves who had the surname Liu, or were old and sick, and who belonged to govern- ment bureaus, treasuries, commandery (offices) and households of kings and marquises of states should be sent up for careful investigation. Presumably slaves with the imperial surname who were really related, and old and sick slaves, were to be freed. 2 HHS, 5, 4a. Cf. p. 79, footnote 2, above. 3 All are conveniently assembled in the Hsi Han hui yao, ch. 63. SLAVE TRADE AND MANUMISSION 133 Emperor Ai's manumission of Palace Women and old slaves (the fourth case) introduces a new principle, age. Whereas Emperor Ai sent the younger Palace Women out to be married, ^ it was the older government male and female slaves, those over fifty, that he eman- cipated. A number of later cases'- in which old slaves were freed suggests that Emperor Ai acted in line with a principle that old slaves deserved to be granted their freedom. Did slaves over fifty consider this manumission a benevolent act, as the historian apparently did? If they were permitted to live in their familiar slave quarters and receive regular food, or if they received stipends and had families with which to live, they might have enjoyed freedom in their declining years. Considering the traditional Chinese attitude of respect and care for the aged, and government sensitivity to popular opinion, it is hard to believe that old slaves were simply turned out to beg. An analogous case in the Three Kingdoms period fortifies this view. In a.d. 239, Emperor Fei ordered all government male and female slaves over sixty to be freed and become "good people." Seven years later he had apparently been informed that aged government slaves were being sold on the market. He ordered his officials to inspect the markets to observe sales of discharged government male and female slaves over seventy, or infirm and crippled ones. "Forsooth," he said, "the government, considering their strength to be exhausted, resells them so they have no place to turn. Let all of them be freed and become 'good people,' and if there are those who cannot support 1 Wei Hung states that kung-jen (Palace Women) were especially selected young slave girls, who were sent out to be married when they reached the age of thirty-five (91). But it is not certain that "kung-jen" was used only for slaves. Emperor Ch'eng's Palace Women were sent out if they were under thirty. Ap- parently the older ones, less likely to be married, had earned the right to stay in security. There are three analogous cases in the Former Han history. Emperor Wen provided in his testamentary decree that his Ladies and women of lower rank down to Junior Maids should be sent back to their families (HFHD, vol. I, p. 271). Emperor Ching likewise provided that his Palace Women should be sent home (op. cit., p. 332). There is no other report of this practice down to 7 B.C., but when Emperor P'ing died early in a.d. 6, Wang Mang ordered that women who had accompanied his Empress into the palace should be sent home and be allowed to marry (CHS, 12, 4b, and notes by Wang Hsien-Ch'ien in HSPC, 12, 10a). There were kung-jen of Emperor Chao still in the palace when Liu Ho came in as Emperor, for he had sexual relations with some of them (69). 2 Emperor Shang's edict cited above probably meant that old and sick slaves would be freed. In A.D. 518 Emperor Wu of Liang freed all (government?) male slaves more than sixty years old and female slaves more than fifty years old. Cf. Ch'eng Shu-te, Chiu ch'ao lii k'ao, ch. 4 ("Liang lij k'ao"), p. 382. A decree of A.D. 566 freed all government slaves over sixty-five from Chiang-ling, and another of a.d. 657 freed all government slaves who were over sixty or invalids. Cf. Liang Ch'i-ch'ao, op. cit., p. 549. 134 SLAVERY IN THE FORMER HAN DYNASTY themselves, let the commanderies and prefectures relieve and support them." 1 Emperor Ai did not state in his edict the kind of care that should be given to those who could not support themselves. While the spirit of the edict appears to be humanitarian, the possibility can- not be excluded that it was an economy measure disguised as a benevolent act.^ A large proportion of government slaves were non-productive, and a great expense to the treasury. Slaves over fifty would be nearly worthless in heavy labor and not very useful in service capacities. It is worth noting, therefore, that about a generation earlier Kung Yii seriously proposed freeing all government slaves in order to cut down expenses. He argued that it would be cheaper to feed them from the government granaries than to keep them as slaves (89). Kung Yii knew about state finance from the inside, having just been treasurer of one of the palaces. This raises a question about Emperor Wen's reason (in the second case) for freeing government slaves near the end of his reign. The act is recorded just after a general amnesty and may have been part of it. Emperor Wen, the most economical and socially minded of all the Han rulers, several times cut government expenses and reduced taxes and labor duties specifically to lighten the burdens on the populace. It is therefore possible that economy was one of his motives in liberating government slaves. This was the only decreed manumission of public or private slaves during both the Former and Latter Han periods not limited to some specified group. Does it follow that all the government slaves were actually freed? Realistically speaking, this is unlikely. Perhaps Emperor Wen meant to authorize a general weeding out of slaves with the dual motives of benevolence and economy. Manumission of Private Slaves Masters could naturally free their own slaves, but we do not know the legal procedure necessary to make the act binding. In T'ang times, as shown above, a slave freed by a master was given a document co-signed by the master's oldest son and other presump- 1 San kuo chih, Wei chih, 4, la and 2b, under years a.d. 239 and a.d. 246. It is a real puzzle as to who might be in the market for aged or sick slaves, unless it might be their free relatives or humanitarian folk. There is no evidence that the Buddhists performed this good work as early as this date. 2 Ma Fei-pai ("Source material on the economic history of Ch'in and Han," pt. 6, "The slavery system," p. 399) holds this view. SLAVE TRADE AND MANUMISSION 135 tive heirs, and a similar document was validated and filed in the local government office. Whether the act was equally formalized in Han times is not clear, for original manumission records, so valuable in the study of Greek and Roman slavery, are still lacking for China of the Han period. The King of Ch'ang-i freed a number of his male slaves and made them Gentlemen and officials {69). Three slave women of the marquises of Ch'eng-tu and Ping-a and of the Noblewoman Hsli were freed and became commoners. They may have been freed by wills, for their masters died in 12, 17 or 16, and 8 B.C., respec- tively. It is more likely that the one who had belonged to Noble- woman Hsli was freed in some other way, perhaps when her mistress was deposed as Empress in 17 B.C. The three ex-slave women were later simply commanded to enter the palace as private slaves of the imperial concubine nee Chao. Later she presumably freed the three women again, for she presented each with ten slave women as a bribe to keep quiet about her evil practices.^ The manner in which Luan Pu won his freedom is not stated. Purchased specifically to do a deed of revenge for his master, he later rose to be a commandant and then a general (5). Probably he was freed as a reward for having successfully performed the deed.- There is likewise no specific statement about the manumission of Tou Kuang-kuo and the several children of Dame Wei, who suddenly rose to prominence as relatives of empresses. Affinal relationship made their freedom automatic.^ The most interesting document on manumission is one recording a slave woman's purchase of her own freedom. Her owner, the Marquis of P'u, kidnaped and re-enslaved her. For this he was brought to trial and deposed (102). This case proves that freedom by purchase was legally recognized even at the expense of a member of the nobility. It also raises the question of peculium, discussed later (pp. 219-220). Paralleling the series of edicts which freed groups of government slaves is another series which freed specified types of private slaves. 1 Document 107; see especially footnotes 20 and 21. 2 Precedent for this assumption is found in a case reported in the Tso chuan, Duke Hsiang, 23rd year (James Legge, The Chinese classics, 2nd ed., 5 vols., Oxford, 1893-95, vol. 5, pt. 2, pp. 497, 501), in which the slave Fei Pao agreed to kill a man in exchange for his freedom. To validate his manumission the "red book," obviously a record of enslavement, was to be burned. ^H; 26, 27, and 29. See discussion of the Wei family, pp. 160-161, below. 136 SLAVERY IN THE FORMER HAN DYNASTY One of these edicts dates from early in the reign of Emperor Kao and five from that of Emperor Kuang-wu, early in the Latter Han. First. — In 202 B.C. Kao-tsu ordered that people who had sold themselves because of famine were all to be freed and become commoners (9). Second.— In a.d. 26 Emperor Kuang-wu ordered that ple- beians' wives and children who had been married off or sold should, if they wished, be allowed to return to their parents (ISJ,). Third. — In 31 the Emperor ordered that officials or commoners who had become slaves or lesser wives through famine and turmoil, or kidnaping by the eastern robbers, should, if they wished, be allowed to go free (136). Fourth. — In 36 an edict freed and made commoners of people in Kansu and Szechwan who had been kidnaped into slavery and had reported to the judiciary but had not been requited. Fifth.— In 37 an edict freed and made commoners of people in Yunnan who had been kidnaped into slavery after a.d. 32. Sixth. — In 39 an edict freed and made commoners of male and female slaves in Yunnan and Kansu who had pled their cases to the local government after A.D. 32. The sellers were not compelled to return the purchase price. ^ These government orders to free private slaves all occurred during the first few years of the Former and Latter Han periods, and were attempts to correct injustices or irregularities that had occurred during the upheaval of founding a dynasty. More than two centuries elapsed between the first and second edicts, and there is no recorded edict that freed private slaves after either dynasty was soundly established. In other words, the government apparently did not interfere with private ownership of slaves properly acquired during normal times. Several close parallels mark the first two manumissions. Both were promulgated almost immediately after the founder of the dynasty had achieved preliminary control, but before he had quelled opposition and established the machinery of civil administration. Both freed people who had sunk into slavery because of economic distress brought on by the chaos of war. Neither indicated how the edict was to be enforced. 1 For the last three edicts see 136, footnote 3. SLAVE TRADE AND MANUMISSION 137 Consider the manumission of famine victims in 202 B.C. The same edict that freed the slaves directed officials to take up their duties, ordered refugees to return home, and granted honorary re- wards and exemption from taxation to people who had distinguished themselves as Kao-tsu's followers. It was merely a proclamation for the world to return from war to peace. Could this manumission have been enforced? Kao-tsu had killed Hiang Yii only a few months before his edict of 202 B.C., and had just established his first capital at Lo-yang. Although he had ordered the demobilization of troops, the country was still filled with private armies. Under these circumstances he lacked the machinery and power to enforce the manumission. It is doubtful that he even meant to enforce it. What would have happened? Wealthy and powerful people would have been robbed of their property. Kao-tsu could not have had the stomach to antagonize them and to threaten his precarious supremacy. Who, moreover, were to go free? A group of helpless folk, sunk into slavery to save themselves from famine. How could they profit his dynasty, and how would his order profit them? For many of them slavery in a well-to-do household must have been preferable to freedom in a war-torn world. If the decree had come ten years later, when the empire was firmly established, it might have been put into effect regardless of opposition. In view of the date, this manumission looks like a humanitarian gesture, placed on record for effect, a plea for public approval neither intended nor capable of being enforced.^ As pointed out. Emperor Kuang-wu's first edict freeing private slaves was closely parallel to Kao-tsu's. His later edicts were more realistic. Limited to particular areas recently conquered, they 1 Chinese students of slavery seem not to have considered this manumission in its historical setting. Since it was an imperial edict it is simply assumed to have been carried out in full. Ma Fei-pai (op. cit., pp. 399-400) even deduces shrewd economic motives for it. He argues that Kao-tsu and Ivuang-wu, also, represented the landlords in opposition to the newly rising commercial class. Before Emperor Ching, govern- ment policy aided agriculture at the expense of commercialism. Since private slaves were one of the most profitable instruments of commercial capitalism and were linked to it much more closely than to landlordism, freeing of private slaves was an attack on commercial capitalism. Such is Ma's thesis. Government economic policies early in the Han period seem to have en- couraged agriculture and demeaned commerce, but there is little indication of a struggle between two hostile economic groups. Furthermore, there is no shred of evidence that slavery was linked more closely to the commercial class than to the landlord class during any part of the Han period. Indeed, for the whole of Han history, there is far more evidence that large landowners were also large slave owners than that merchants were; and while people deriving their wealth 138 SLAVERY IN THE FORMER HAN DYNASTY ordered civil authorities to use specified laws against recalcitrant owners. He may have meant to enforce these edicts, but the normally close connection between officials and the slave-owning class probably thwarted his intention in many cases. A fact about the eleven recorded edicts ordering manumission is worth noting: Only one contains evidence that slaves were actually freed. This does not mean that the edicts were disregarded; it is simply a caution against assuming immediate liberation of all slaves in the groups specified. We do not know how faithfully and thoroughly orders were carried out, or how many slaves were affected. Imperial orders to free government slaves could be carried out through regular administrative channels within the closed system of the imperial government; but orders to the populace to free its slaves at personal economic loss would be vastly more difficult to enforce. On the other hand it cannot be assumed that the eleven edicts represent the only government manumissions of groups during two and one-half centuries. It is only the manumission of larger groups or of important people like Chin Jih-ti or Tse (important for the narrative) that were worth recording. Probably many minor manu- missions, involving individuals or small groups, were overlooked. Every reference to manumission specifying the status of the freed slaves uses the term "commoner" or "common people" ^ except when the slave rose immediately to some honorary rank. This is significant. The terms refer to the vast mass of the Chinese people. Kings, marquises, and the highest ministers of state tumbled to the very same rank when they committed crimes requiring their expulsion from the ruling class. There was an unrestricted transition from slave to commoner status, and a quick transition from plebeian to noble rank in cases of talent or fortunate marriage. On the other hand, there was an equally easy descent from noble to plebeian rank through crime, and from commoner to slave status through economic distress or "rebellion." T'ang law reflects a certain rigidity of status. Government slaves had to pass through two higher stages, from either or both sources did own slaves as a result of that wealth, there is little to prove that slaves were important as producers of wealth. Evidence that slaves contributed to wealth of merchants is scanty for the whole Han period and only one or two items even close to 202 B.C. can be cited. Finally, there is no way of knowing who had bought famine victims or which group might have been harder hit by having its slaves freed. If the manumission was not widely enforced, as I believe, then the latter uncertainty is, of course, beside the point. 1 Documents 9, 2U, 80, 89, 92, 102, 107, 110, 135, 138, and HHS, IB, 3b, 4b, and 5b; 4, 9a; 5, 4a. SLAVE TRADE AND MANUMISSION 139 with separate pardons, before they became liang-jen, "good people." Later still, freed slaves and their descendants were prohibited from taking civil service examinations, the normal road to office.^ In Former Han times, however, slavery was not caste-bound although slaves possessed a definite ascribed status. The next problem is to determine the nature of that status. 1 Cf. Wang Shih-chieh, op. cit., p. 325 (Pippon, op. cit., pp. 128-129), and his whole discussion of legal slave status. VI. STATUS OF SLAVES "Status" is the position or standing of individuals in relation to other individuals, groups, or the state. It is manifested by prerogatives or limitations, rights or duties, enjoyed by or imposed upon individuals in correspondence to their membership in recognized social groups. It is usually characterized by accepted modes of reciprocal behavior between persons of different status. Differences may be formally defined or they may be merely implied by systems of classification according to sex, age, kinship, rank, caste, occupation, membership in special associations, and so forth. The status of an individual may be ascribed to him automatically because of his membership in these groups, or he may achieve certain types by special accomplishments and skills. Servile status is ascribed. When status is determined by social custom and maintained by moral, religious, or social sanctions in various degrees of rigidity it is "customary" status. But certain types of status are codified. Legal status represents the standing of a person before the law, or his relation toward others and the state in matters such as citizen- ship, property rights, marriage, infancy, and majority. If slaves were a well-defined class in Han society we should expect to find them endowed with a particular status in reference to members of other groups or classes; and conversely, if slaves had status different from that of people in all other classes, it would indicate a formal distinction between slaves and other people in the social structure. Historical literature reflects social custom only vaguely. Yet the thoughts of historians and philosophers, and the attitudes of free people toward slaves do reveal something of customary slave status in China during the period here studied. Fortunately, well- codified T'ang and post-T'ang legal materials illuminate the frag- mentary evidences of legal slave status in the Former Han period. Distinct in their degree of formalization, and reconstructed from differing kinds of sources, these two types of status need to be examined separately. Customary Attitude Toward Slaves Han statesmen and writers considered slave status base, but fail to explain in what respects it was demeaning. Apparently they accepted it as a matter of course, and not a subject for speculation. Thus, the historian Pan Ku listed men of humble origin who rose 140 STATUS OF SLAVES 141 to greatness: Pu Shih, the shepherd; Sang Hung-yang, the lowly tradesman; Wei Ch'ing, the slave; and Chin Jih-ti, the surrendered captive {26, footnote 12). Ssu-ma Ch'ien defended Chi Pu's voluntary enslavement to avoid capture as if it were the depth of disgrace, by pointing out that it was at least more noble than suicide, the recourse of slave women, concubines and other mean people (56). He, or some earlier writer from whom he copied, employed slave status as a simile for baseness and ill-treatment in the statement that the daughter of the rich man of Wai-huang looked upon her husband as a hired laborer or male slave {A2), and similarly in the remark that Wang Wen-shu skilfully served those in power, but looked upon them as though they were slaves as soon as they lost their power {A9). Pao Hsiian, a Grandee Censor at the end of the Former Han period, protested that it was against Heaven's intention for the slaves of Tung Hsien to be liberally rewarded, officially employed, and consequently become rich (118). Any personal relation between rulers and slaves was considered vastly demeaning. The contemporaries of Liu Ho, King of Ch'ang-i till 74 B.C., severely censured him for having slaves as friends. This impropriety was one excuse for his impeachment after he became Emperor, and his adviser, Kung Sui, was spared execution primarily because he had protested against it (67, 69). Wang Yin, Liu Hsiang, and Ku Yung all criticized Emperor Ch'eng for taking incognito journeys in the intimate company of male slaves, and for keeping private slaves in his palace (10^). Some moralists considered it highly improper for slaves, or even common free folk, to wear fine clothes, especially any bit of costume emblematic of noble or official status. This attitude had a philo- sophic background, strongly Confucian but inherent in other schools of thought as well. Ranks, functions, duties, and prerogatives of various members of a family group, of classes in society, and of strata in the political organization had to be distinctly and properly differentiated. Likewise the symbols of rank — both physical and abstract — had to correspond. When such distinctions were upheld, when all realms of social and political organization were in proper harmonious relationship, nature too would be harmonious. But defiance of proper differentiation could violently upset natural phenomena and cause physical disasters. According to this school, if slaves were allowed to wear costumes reserved for or symbolic of the highest ranks, the ethical impropriety created an actual physical menace. 142 SLAVERY IN THE FORMER HAN DYNASTY Chia Yi criticized the practice of dressing slaves in beautiful clothes reserved in antiquity for emperors and empresses, and the custom of using stuffs, once reserved for the service of an emperor and his empress, to decorate the walls of rich people's homes. In a memorial to Emperor Wen, he asked: "How can there be no danger that the world will be unsubmissive [to the Emperor] if now the walls of the rooms of commoners may be done with an emperor's clothes; if singers and entertainers [and other] mean people may have the ornaments of an empress?" (16.) The ceremonial cap was the most distinctive emblem of the upper class. When Liu Ho conferred this cap upon a male slave he horrified his contemporaries and shocked later commentators. Kung Sui, his righteous adviser, saw in such acts a veritable peril to the kingdom's gods of the soil and grain (67, footnote 5). Liu Hsiang considered it as approaching "heterodoxy in clothing." Ching Fang, a specialist on disasters and unnatural phenomena, is quoted as having said: "His conduct was not compliant [with natural law]; it brought disaster on the man. A male slave was capped; and the world was [thrown into] anarchy." An unknown writer whose words are now part of the chapter on "Unnatural Phenomena" in the Ch'ien Han shu, said of the act: "A ceremonial cap is honorable clothing; a male slave is a mean person. [The way Liu] Ho enjoyed conferring uncustomary ceremonial cappings with- out reason, was a symbol of [the way he] regarded honors. Cere- monially capping a male slave is equivalent to making the height of honor fall down to the extreme of meanness." (67. y This belief that servile status was demeaning is not supported by any theoretic discussion on the nature of slavery, nor is it mixed with any antagonism toward aliens, for there is no evidence of a marked non-Chinese element in the slave population. It was accepted as self-evident, and the point of view reflected what "every- one" thought at the time. A few Han documents picture this general attitude. The Shih chi and Ch'ien Han shu both tell that in the State of Ch'i "it was customary to look down upon male slave captives." (17.) Early in the Han period Wei Pao refused to rejoin Emperor Kao, saying: "Now the King of Han insults people. He curses 1 This quotation sounds like a commentary. It may have been copied by the authors into the Ch'ien Han shu from some such writer as Liu Hsiang or Ching Fang; but it might be a later commentary which has slipped into the body of the text. The attitude expressed appears, however, to be typical of the common attitude in Former Han times. STATUS OF SLAVES 143 and scolds the nobles and ministers just as [one curses] male slaves, only!" (AS.) Wei Ch'ing's two proud but penniless retainers, Jen An and T'ien Jen, refused to eat at the same mat with the male cavalry slaves of the Princess of P'ing-yang. They cut the mat in two with their knives and sat apart (39). The fact that Wei Ch'ing himself had been a male slave and cavalry man in the same house- hold only a few years earlier lends an ironic touch to the situation. We even have a report of Wei Ch'ing's own attitude toward his slave status when a physiognomist told him that he would become a marquis. To this prediction he is said to have replied, skeptically: "Born as another's male slave, it is sufficient not to be beaten and cursed. How could I get appointed a marquis!" (26.) Thus slaves were apparently considered to be on the lowest level of the social scale. But slave status is only one of the innu- merable kinds of status. An individual does not have a single and permanently fixed customary status, but many, deriving from such factors as sex, age, kinship, occupation, political position, and nationality. In some fields status may change gradually, as with age, or quickly, with accomplishment. The standing of Tou Kuang- kuo before and after he established his identity as the brother of an empress is an example of rapid change. Furthermore, status does not operate abstractly, but concretely, in the relations of individuals each having a combination of statuses. Within the slave group itself there were marked differences in status. Among government slaves, for example, there must have been differences between enslaved families of criminals — wearing felon's dress, their heads shaved, and perhaps even permanently disfigured by tattoo marks around the eyes — and private slaves given to the government or seized by it. The work done by a government slave must itself have determined his status within the slave group, but it is not easy to distinguish the personal qualities, or the circumstances in a slave's acquisition that determined the type of occupation to which he was assigned. Some toiled at hard labor transporting grain, or in the government mint. Others, more skilled and clever, manufactured implements or prepared the palace banquets. Some among them even rose to be "senior slaves," in charge of their lesser fellows. They must have had a customary status above that of the toilers; and as govern- ment employees, knowing the ropes, familiar with bureau affairs, they doubtless enjoyed a commanding position among the common folk who were actually free. 144 SLAVERY IN THE FORMER HAN DYNASTY The elite among the government slaves were probably those who worked in imperial palaces. Most palace slaves were women; many worked in the Concubines' Quarter as servants of palace ladies, though some were in other parts of the palaces, too. Some, like Ts'ao Kung, unwittingly became involved in palace intrigue. The attractive daughter of a government slave woman, Ts'ao Kung was a student clerk attached to the Empress. Although she had a homosexual attachment with another slave girl named Tao Fang, she was "favored" by Emperor Ch'eng and conceived. She boasted to Tao Fang and her mother about her great good fortune. Her prestige was growing rapidly. The child she bore was a boy, the presumptive heir to the throne. Six slave women were placed at her disposal. Had the boy been made Heir-apparent, Ts'ao Kung would probably have become Empress; yet she was legally a slave and theoretically that was still her status. Because of the wild jealousy of the favorite imperial concubine, the Emperor commanded Ts'ao Kung to take poison. Her son was slain, and the six slave women who attended her were compelled to hang themselves. Cheng Ch'i, another slave woman who had nursed the babe for eleven days, lived to tell her side of the story to the investigating com- mission, as did Ts'ao Kung's mother and the friend, Tao Fang (107). This document reveals some aspects of the life of palace slave women. The violent and most dramatic scenes may not be typical of the life of such women, but intrigue was always current in the palace. The slaves saw much and knew the gossip of the court. When freed, or when mixing with other slaves and commoners, they had much to hint at or tell. They enjoyed a customary status loftier than that of most government bondsmen, and far above that theoreti- cally ascribed them as slaves. The humble palace slave woman Tse— whose surname is not even reported, but who had an audience with the Emperor and received a gift that to her was fabulous — must have enjoyed far greater customary status amongst her neighbors because of her former position in the palace than she possessed, after emancipation, from her rank as a commoner (80). The toiler slaves, the servile bureaucrats, and the palace slaves may all have been equal before the law. Certainly they were different from commoners, and were considered beneath them. But within this common status there must have been marked degrees of difference. If a single principle defines that difference it is func- tion. Information about government slavery in Former Han times is too scanty to allow an arrangement of various slave occupations STATUS OF SLAVES 145 in a series of increasing importance. Yet tlie principle of relationship between the function of individuals and their customary status is applicable also to private slaves. Some of Tiao Chien's slaves associated with generals, adminis- trators of commanderies, and chancellors of kingdoms. Tiao Chien trusted them so completely that they preferred their status as his slaves to freedom and noble rank (17). Likewise, the male slaves Feng Tzu-tou and Wang Tzu-fang, Ho Kuang's confidential ad- visers, who were regarded by officialdom as more important than the Chancellor (66, and footnote 3), had a status on the basis of function beyond that of his ordinary slaves. Feng Tzu-tou even had his name listed along with several nobles and high officials in an imperial edict about the conspirators of the Ho family clique,^ a dubious honor made possible primarily because of his importance in the business affairs of the family (72, and footnote 3). So, too, the status of other private slaves probably varied; a servant girl must have been lower in status — both among her fellows and in the social scale — than a female chamberlain slave, who had personal charge of her master's wardrobe and bed. Male slaves who did manual work could hardly have enjoyed the prestige of their fellows who were cavalry escorts and personal guards for their masters. The actual customary status of private slaves was affected also by the status of their masters. The best documented example of this modification of status concerns the slaves of Ho Kuang. His two confidential slaves ranked above their fellows because of their special position, but their status with officialdom arose from the fact that they were the key for access to Ho Kuang. Yet all his slaves enjoyed special prestige. In Ho Kuang's native seat, the male slaves and retainers of the family would go armed into town, fighting and brawling, and no official dared to stop them except Yin Weng- kuei, who, as the historian admiringly reports, enforced the laws impartially (65). Shortly after Ho Kuang's death one of his grand- nephews had the audacity to send a slave in his stead to attend the imperial court. Certainly it was no ordinary slave who could appear before the Emperor at formal assembly, but a top slave of the most powerful family in the realm. Even after their master was dead the Ho family slaves were exceedingly arrogant. Once a party of them disputed the right of way with the slaves of Wei Hsiang, who was Grandee Secretary. Here was a clash between slaves over » CHS, 8, 5a. 146 SLAVERY IN THE FORMER HAN DYNASTY their relative status. To avenge the insult to their prestige, the Ho slaves invaded Wei Hsiang's residency and made him knock his head upon the ground and apologize to them! (72.) This humil- iation was partly responsible for Wei Hsiang's prosecution and final destruction of the Ho clan when he became Chancellor. Simi- larly, the slaves of Tung Hsien, Emperor Ai's favorite, received rich gifts from various officials (117), and "looked on wine as though it were soup, and meat as though it were beans." {118.) Wang Shih-chieh, in a very interesting article on Chinese slavery, ^ devotes a section to a discussion of the legal position of slaves, under six headings: limitation of rights of marriage; limitation on right of examination and official position; punishment of slaves for crimes; crimes against slaves; limitation of rights in law suit; and manu- mission. Drawing on laws, edicts, and cases from all Chinese history, but especially from T'ang times on, he is able to formulate several important generalizations concerning slave status. For the Han period, however, his information is scanty because of the paucity of extant legal material. From T'ang times down to the end of the nineteenth century the legal status of slaves was definitely inferior to that of the free. Slaves who committed crimes of violence, sex, or abuse against their masters or against other free people were much more severely punished than free people who committed the same crimes against each other. On the other hand, masters or other free people who committed these crimes against slaves were lightly punished. Male slaves could not legally marry free women; nor could slaves accuse their masters in court (except in crimes of high treason), even for redress against personal wrong. - Slaves in Criminal Law Because the law code of the Han dynasty has been lost since the sixth century we must go on the one hand to edicts or recorded cases, and on the other to later codes based upon it in order to formulate an opinion regarding the legal status of slaves prior to the 1 Wang Shih-chieh, "The Chinese slavery system" (translated by Toni Pippon, "Beitrag zum Chinesischen Sklavensystem"). On legal aspects see especially pp. 315-325 (Pippon, pp. 113-129). - The legal documentation on these points may be found in Wang Shih-chieh, op. cit., pp. 319-322, 316-318, and 322, respectively. See also Shen Chia-pen, Li tai hsing fa k'ao, "Fen k'ao," ch. 15, pp. lOa-llb, and 29b, and Liang Ch'i- ch'ao, "System of slavery in China," pp. 547-549. STATUS OF SLAVES 147 first century of our era. Yet it is clear that slaves had an inferior position in criminal law as compared with free people not their masters. This generalization is based upon two edicts by the first Emperor of the Latter Han dynasty, referring to legal matters during the preceding period. In A.D. 35 Emperor Kuang-wu revoked the law of public execu- tion for male and female slaves who shot and wounded people (138). No such law is elsewhere recorded for the Former Han period, but an imperial edict revoking such a law is evidence of its earlier existence. The edict does not specify whether "people" means any free person or the master. The phrasing is also vague on the point of accidental or purposeful wounding. Apparently the circumstances in such cases were of no concern; a slave who shot and wounded a free man intentionally, or possibly even accidentally, could legally be executed in public. Was this inequitably severe? Ch'en Lung, a Commandant of Justice in A.D. 94, reported to the throne that there were 610 listed crimes involving the death penalty. 1 In Han times the death sentence was applied in three ways, which differed in degree of severity. The worst penalty was to behead the criminal and exhibit his head in a public place. The second was to cut the criminal in two at the waist. The least severe was public execution (lit., "casting on the market place"), since this did not involve dismemberment of the body.- The severity of each penalty could be increased by extending the sentence to various relatives of the criminal. In practice, and in theory also, the degree of the punishment was conditioned by the rank and social status of the criminal; punishment was not decreed objectively as a pre- established sentence for a specific crime. '^ When a slave shot and wounded a free man the proper sentence was public execution. This was excessively severe treatment judged either by comparison with crimes of free people which resulted in their being publicly executed, or by comparison with the punish- ments legally applied to free people who merely wounded other free people. 1 HHS, 76, 4b. - In 148 B.C. Emperor Ching abolished the punishment of "quartering" and substituted public execution. See HFHD, vol. I, p. 319, especially the important footnote 6.4. Beheading and cutting in two at the waist continued to be practiced, as is shown by cases during the rest of the dynasty. ' Cf., for example, cases incidentally reported in documents 44, 52, 6^, 98, 103, 113. Cf. also HFHD, vol. I, pp. 176-177. 148 SLAVERY IN THE FORMER HAN DYNASTY Analyzing a considerable number of cases in which criminals were actually publicly executed, and examining known laws con- cerning this punishment, we find that this third form of execution was applied in general to four classes of very serious crimes. These classes were: (1) crimes against the Emperor or imperial prerogatives; (2) murder; (3) violation of fundamental morality; and (4) serious corruption or crime by high officials. ^ This shows not only that the law applied to slaves was discriminatingly severe, but also points to a formalized differentiation in Han criminal law between crimes committed by slaves and by free people.- 1 This classification certainly does not accurately represent the Chinese con- ceptual background in which ethics, morality, and law are mingled in a way which defies classification in western terms. The difficulty is immediately apparent in the number of cases which might fall in two or more of these categories. ^ In compiling the following list, items were taken from the first six chapters of Ch'eng Shu-te, Chiu ch'ao Hi k'ao ("Han lu k'ao"). In general only items of the Former Han period referring specifically to public execution, and quoted from the SC or CHS, were taken. The pages referred to are in the one-volume edition of 1935. Class 1 Page 75: Father, mother, wife, children, brothers and sisters should be publicly exe- cuted in cases of treason and inhumanity (also class 3). 77: Forging imperial edicts and disregarding imperial orders. 127: Divulging conversations held in the imperial palace. 145: Entering the imperial audience hall without a pass. 176, 184, 185: Criticizing the imperial family. 102, 180: Counterfeiting, or making false gold. Class 2 131: Murder (2 cases). 133: Having a person murdered by another (5 cases). 135: Wounding a person who then died within a short time. Class 3 115: Inhumanity — an ofladal bearing a grudge against and slandering a member of the government. 117: Unfilial conduct— a son bringing public accusation against his father (cf. also pp. 114-118 j)assim). Class If. 119: Wilful negligence — a Commandant of Justice failing to punish a capital crime. 120: The same — a Prefect failing to prosecute a conspiracy of rebellion (also class 1). 123: Concealing criminals — a marquis concealing a gang of robbers; a Comman- dant of Justice concealing conspirators of rebellion (also class 1). 127, 157: False accusation of innocent people — by a Commandant of Justice, by a marquis. 160: Officials mutually recommending each other for advancement. 163: Being disrespectful about ceremonial purification. 2 Wu Ching-ch'ao ("The slavery system of the Western Han," p. 267) says that this punishment for slaves was one degree more severe than if the guilty person were free. He does not document his statement, but his opinion may be based on an analogous law of T'ang times which specified that male or female slaves who beat "good people" were to receive a punishment two degrees more severe than would be applied to free culprits {Tang lil su i, ch. 22. Cf. Shen STATUS OF SLAVES 149 In determining the normal punishment for a free person who merely wounded another free person, we arrive at the same conclu- sion. A basic principle of Han law was the famous declaration of Emperor Kao after he entered the Ch'in capital in 207 B.C.: "He who kills a person will die; he who wounds a person or robs [will be punished] according to the offense." ^ There are few recorded cases of free people punished for wounding others because such matters were not generally of serious state concern. Two cases involving noblemen and officials happen to be recorded, and are valuable for comparison. In one, a marquis was tried for wounding a man under circumstances which are not reported, and was dismissed.- The other is the case in which Hsiieh K'uang hired a man to attack and horribly disfigure an enemy, for which he was only banished to Tun-huang. During the official debate concerning the crime, which was complicated by a number of interesting factors, the Commandant of Justice enunciated an important legal principle and quoted a significant law. First, "The universal principle of the past and present, unaltered during the Three Dynasties, is that 'murderers die, and wounders [are punished with] mutilation.' " The law he cited was: "[One who] wounds another with a bladed weapon while fighting will be left whole [i.e., unmutilated] and will [be sentenced to] work on [frontier] fortifications and patrol; one who purposely attacks [and wounds another will receive] a punishment increased by one degree; one who plotted with [the actual attacker will receive] the same punishment." ^ Chia-pen, op. cit., p. 29b; Wang Shih-chieh, op. cit., p. 320). In T'ang times there was a special class, called pu ch'ii, whose status was between that of slaves and free people. Punishments for pu ch'il were one degree more severe than for free people, and one degree less severe than for slaves. 1 Cf. CHS, lA, 7a, and HFHD, vol. I, p. 58. Wang Shih-chieh (op. cit., p. 320) emphasizes the fundamental difference between this principle and the law regarding slaves who wounded people. 2 Ch'eng Shu-te, op. cit., p. 76. Another case is there reported of a marquis who was tried for premeditated attack and murder, and was dismissed. ' The word translated as "left whole" refers to those who escaped mutilation after Emperor Wen abolished that form of punishment (cf. HFHD, vol. I, p. 255, and Ch'eng Shu-te, op. cit., pp. 44-45, 178). The increase of punishment by one degree would here probably have involved tattooing on the face as well as frontier service, or increase from four to five years' servitude (ibid., pp. 51-52). This complex case occurred in 7 B.C., and is reported in detail in CHS, 83, 3b-4b. An official named Shen Hsien was slandering a fellow official, Hsueh Hsuan, for lack of loyalty and filial piety. Hsueh K'uang, the son of Hsiieh Hsiian, feared that Shen Hsien would make an official accusation against his father, and therefore plotted to hire a retainer, Yang Ming, to disfigure Shen Hsien so that he could not appear at court. Learning that Shen Hsien was about to be made Colonel over the Retainers, Hsueh K'uang ordered Yang Ming to go ahead 150 SLAVERY IN THE FORMER HAN DYNASTY We thus have a dual contrast. On the one hand, free people were publicly executed only for very serious crimes; yet slaves could be publicly executed merely for wounding free people. On the other hand, if free people wounded other free people they were punished in some equivalent or proportionate degree; but slaves who wounded free people could be punished far beyond the degree of the actual crime itself. In cases involving slaves the punishment was obviously not balanced to equal the criminal act alone, but was determined on the basis of status as well. There is no further documentation in Former Han records about slaves who wounded free people. But the discussion cannot end there. Emperor Kuang-wu revoked a particular law which harshly discriminated against slaves who committed a particular crime. While no other similar specific laws still exist it is almost beyond the realm of possibility that the very elaborate Han code contained an isolated law to deal exclusively with slaves who shot and wounded with the deed. Yang Ming and some others attacked Shen Hsien on the great road in front of the palace gate, cut off his nose and lips, and hacked his body in eight places. In the discussion among the officials as to proper punishment, two opinions developed. The first maintained that Hsiieh Hsiian and his son were both officials; that Hsiieh Hsiian's lack of filial piety was common knowledge and should have been officially known; that the attack on an official by the agent of another official had been made in front of the palace, and in public, so that it had a very bad effect on public morals. Thus the crime was ta pu ching ("great disrespect for the Emperor"), and therefore both Hsiieh K'uang and Yang Ming ought to be publicly executed. The second opinion, advanced by the Commandant of Justice, starts with the quotation of the Code about wounding in fighting, etc. (translated above), and maintained that Shen Hsien was not justified in constantly speaking of Hsiieh Hsiian's bad conduct; that it was a private fight and not different from fights between ordinary civilians simply because it occurred outside the palace gate. The Commandant then cited the unalterable ancient principle that "murderers die, and wounders [are punished by] mutilation." This fundamental principle, a quid pro quo, may be contrasted with the law of public execution as punishment for slaves who wounded free people! Continuing, he contended that to treat Hsiieh K'uang and Yang Ming as being guilty of ta pu ching because they were officials was a violation of the principle of the Ch'un ch'iii that there was no difference between public and private matters; that, considering the original impulse of the crime, Hsiieh K'uang was violently angered by observing a person slandering his father, and was guilty of no other great crime. He believed that Yang Ming and Hsiieh K'uang ought to be sentenced respectively according to the law pertaining to the premeditated attack and wounding of a person, and to plotting with the attacker. But, since both had honorary titles, the sentence ought to be reduced to working on frontier fortifications and patrol, without mutilation. (Yen Shih-ku explains that if a person had honorary rank he received a reduced sentence and was not mutilated. Thus, these men, by getting a reduction, would be given the punishment due for wounding another in an ttnpremeditated fight.) Emperor Ai then put the question for discussion among the high ministers. The Chancellor, K'ung Kuang, and the Grandee Secretary, Shih Tan, agreed with the first decision, while all the other ministers agreed with the second. _ In the end, Hsiieh K'uang had his sentence reduced one degree, and was banished to Tun-huang. What happened to Yang Ming is, characteristically, not told. STATUS OF SLAVES 151 free people; that it had no other laws of this class; and, finally, that this one law happened to be preserved among the relatively few Han laws known to this day. In other words, this law was probably typical rather than unique. It must have been one of a class of laws which discriminated against slaves by condemning them to more severe punishments than those applied to free people guilty of the same or similar crimes. There was a group of laws discriminating against slaves in the T'ang code, and T'ang law was based upon the Han law indirectly through intervening codes, and probably even directly. Moreover, the principle of inequality was carried on in all the later great codes based on T'ang. These codes, in fact, did not even alter many of the details.^ The general continuity of Chinese law is a fact not to be minimized when a principle rather than a particular point is in question. If slaves were in general punished more severely for crimes against free people than were free people themselves, then were free people punished less severely for crimes against slaves than for similar crimes against other free people? This reverse side of the question is suggested by another edict of Emperor Kuang-wu. In A.D. 35 he proclaimed: "In the nature of heaven and earth, man is most important. He who kills a male or female slave will not receive a reduction in punishment." (137.) This imperial edict lays down the principle that, under law, the murder of a slave is no less serious than the murder of a free person ; or, more specifically, that the murderer of a slave could not there- after be sentenced to a lighter penalty than the murderer of a free man. Does this not mean that before A.D. 35 the murderer of a slave properly received a reduced penalty? Modern Chinese writers on slavery conclude that it does.- Analogy with the T'ang code suggests the same conclusion. According to T'ang law, "those 'good 1 Wang Shih-chieh, op. cit., p. 320. For example, the T'ang law (p. 148, foot- note 2) which specified two extra degrees of punishment for slaves who beat and wounded "good people," is only part of the law. Thus, if a slave beat "a good person" and broke a limb or a bone in his body, or blinded one eye, he would be strangled; if the slave beat the "good person" to death he would be beheaded (T'ang lii su i, ch. 22). Again, a male slave who had peaceable sexual relations with a free woman would be punished with two and a half years' servitude; if he raped her he would be banished; if he injured her during rape he would be strangled (ibid., ch. 24). Crimes by slaves against their masters or their masters' relatives were even more severely punished. - Wu Ching-ch'ao (op. cit., p. 267) uses the edict as direct evidence concerning the legal position of slaves in Former Han times, as does Liang Ch'i-ch'ao (op. cit., p. 546). The belief is implied also by Ma Fei-pai ("Source material on the eco- nomic history of Ch'in and Han," pt. 6, "The slavery system," p. 395). 152 SLAVERY IN THE FORMER HAN DYNASTY people' who beat and wound other people's pu ch'u [shall receive sentences] reduced by one degree [from the sentence for beating and wounding] ordinary people. [Those who beat and wound] male or female slaves [shall receive sentences] further reduced by one degree. Those who purposely kill pu ch'u [shall be] strangled; [those who purposely kill] male or female slaves [shall be] banished 3,000 li." ^ Emperor Kuang-wu's edict did not differentiate between people who killed slaves belonging to the government or to other people, and masters who killed their own slaves. Legal Rights of Masters over Slaves Apparently masters had the right to kill their slaves, but under certain conditions only. Even then they were held legally account- able to the government. Just before the beginning of the dynasty one of the rebels against Ch'in used the following stratagem to gain access to the Prefect of Ti in order to murder him. T'ien Tan "deceivingly bound up his male slave, and escorted by [a group of] young bloods, went to the court [of the Prefect of Ti, as though] wishing an interview to [announce his intention to] kill the male slave." He was successful in this stratagem, and having been admitted to court, he killed the Prefect U). This passage must have seemed intelligible to the authors of the Shih chi when they composed or copied it toward the end of the second century B.C. Today, however, its implications are revealed only through the commentators. Fu Ch'ien, writing toward the end of the Latter Han period, explains that formerly anyone killing a male or female slave had to announce it to the government; and that T'ien Tan, wishing to kill the Prefect, bound up his male slave in order to get an interview. Yen Shih-ku adds that he fraudulently bound up his male slave to create the appearance of killing the slave (J^, footnote 4). In later histories, this law or usage becomes clearer, and the right of masters to kill their slaves is more strictly defined. In the History of the Chin dynasty, A.D. 265-419,^ a statement in the 1 T'ang lii su i, ch. 22 (cf. Wang Shih-chieh, op. cit., p. 321; Liang Ch'i-ch'ao, op. cit., p. 547). Wang Shih-chieh says that the Sung law was the same as T'ang, while those of the Yiian, Ming and Ch'ing periods were modified only in detail. The principle was apparently the same. He also cites equally discriminatory T'ang and later laws concerning sexual relations between "good people" and public or private slaves. 2 Chin shu, ch. 30 (Hsing fa chih), 7b, quoted by Ch'eng Shu-te (op. cit. [Chin lii k'ao], p. 275), Liang Ch'i-ch'ao (op. cit., p. 547), and Shen Chia-pen (op. cit., p. 29a). STATUS OF SLAVES 153 section on law says that if a male or female slave resists his master, the master may pay a visit (to an official) and then kill the slave. Early in the T'ang period the law read : "When various male or female slaves commit crimes, those of their masters who kill them without requesting [permission of] the government officials shall be basti- nadoed one hundred strokes." ^ The element which these later laws had in common with the situation in 209 B.C. is the master's right to kill his slaves after receiving permission from the government. In the earliest case it is not stated on what grounds permission would be granted. Tung Chung-shu, one of the foremost Han social philosophers, implied that in practice masters often killed slaves on their own authority, and advocated that Emperor Wu abolish slavery in order to eliminate the terror of autocratic execution (35). Since masters were supposed to obtain permission from the government, were they punished if they neglected to do so? Known cases of masters (or presumed masters) who auto- cratically executed slaves indicate at the outset that such acts were treated as criminal. In most of these cases the master committed other serious crimes as well, so that it is not possible to assess the relative gravity of the slave murder and of the other crimes in terms of the punishment. For example, when the Marquis of Chao was tried for murdering sixteen men, some of whom were slaves {52), the brief report does not differentiate the crimes or tell how many were slaves and how many free. Three out of sixteen people foully murdered by the King of Kuang-ch'uan at the instigation of his Queen were slaves (6^). But in the trial the sixteen murders were treated together. Likewise, Liu Li, the King of Liang, ordered a slave to murder two officials, and then murdered the slave to silence 1 T'ang lii su i, ch. 22. The law continues: "Those [masters] who kill innocent [slaves] shall [be punished by] one year of servitude; those who accidentally kill [their slaves) shall not be tried." Wang Shih-chieh (op. cit., p. 321) states that the Yiian, Ming and Ch'ing codes copied the T'ang law either entirely or in general. In contrast to these light punishments for masters who killed their slaves, is the reverse situation, in which slaves harmed their masters. The T'ang law does not even specify a punishment for slaves who purposely killed their masters, but "all those pu ch'il or male and female slaves who plot to kill their masters shall be decapitated. Those who plot to kill the masters' relatives or mothers' parents shall be strangled; those who [plot to kill and] actually wound [such relatives] shall all be decapitated." Furthermore, "All those pu ch'ii or male and female slaves who accidentally kill their masters shall be strangled. Those who beat their masters' relatives or mothers' parents shall be strangled; those who [thus] actually wound [the masters' relatives] shall all be beheaded. Those who curse [their masters or masters' relatives(?)] shall be [punished by] two years servitude." (T'ang lii su i, ch. 17 and 22.) See also Liang Ch'i-ch'ao (ibid.), Wang Shih-chieh (op. cit., p. 319) and R. Deloustal ("La justice dans I'ancien Annam," BEFEO, vol. 11, 1911, p. 318, footnote 2). 154 SLAVERY IN THE FORMER HAN DYNASTY him (106). He was officially accused of three murders, as well as of other crimes. After the death of Liu Yiian, King Miu of Chao, the Grand Herald memorialized the throne about his former criminal acts, accusing him of killing slaves and of ordering others to be killed after his death. The wording is important. "[Liu] Yuan formerly illicitly killed male and female slaves with a sword." Does this not mean that he killed when he did not have the right to kill? Later, when the King was mortally ill he made a will commanding that his entertainer slaves be buried with him. After his death sixteen of them were compelled to commit suicide. Only the fact that his son had killed an internuncio complicates this case, preventing it from being a clear example of punishment for the illegal killing of slaves. The earlier illegal killing of slaves, the son's killing of the inter- nuncio, and the enforced suicide of sixteen entertainer slaves were equally recorded in the memorial as grounds for abohshing the kingdom (SJf). This lumping together of murders of free people and slaves owned by the murderer is significant. It indicates that the wanton killing of a slave was itself a serious crime. The facts in each of the above cases are on record primarily because the murderer was an important nobleman and a member of the imperial clan, so that the punishment vitally affected an hereditary line and a political unit. Yet in each case the fact that some of the victims were slaves was of enough additional importance in the total picture to be reported specifically. Thus, even in cases where nobles were guilty of ruth- lessly murdering free people, the official records of the cases include the fact that they also murdered their own slaves. The wife of the Marquis of Chiang-ling was so uncontrollably jealous of other women in the household that she murdered forty or more female serving slaves. She was publicly executed for these and other crimes, but the case was so extreme that it does not serve as a good test {82). A decisive one is the matter of Chancellor Wei Hsiang. Two separate reports of the suspicious death of Chancellor Wei Hsiang's female slave chamberlain prove unequivocally that masters could not freely kill innocent slaves. The Ch'ien Han shu relates that Chao Kuang-han, Administrator of the Capital District, feared that Chancellor Wei Hsiang was about to accuse him of past crimes. Through a spy he learned that the Chancellor's female slave chamber- lain had died violently some time before, and suspected that the STATUS OF SLAVES 155 Chancellor's wife had killed her out of jealousy. Armed with this information, he tried to frighten the Chancellor into dropping his charges, but when this failed he sent up a document reporting the crime. The case was referred back to him, as Administrator of the Capital District, to prosecute. Seizing this opportunity, he invaded the chancellery, cross-questioned the wife, took witnesses, and charged the wife with murdering the slave. Chancellor Wei Hsiang was thus driven to submit a document personally affirming his wife's innocence and requesting an imperial commission to investi- gate the slave's death. The Commandant of Justice determined that the Chancellor himself had scolded and beaten the slave because of her faults, and that she had gone to an outside mansion and died, i.e., by suicide. The Chancellor and his wife were thus absolved; but Chao Kuang-han was tried and executed for slandering a high official as well as for his other earlier crimes (76). The account in the present Shih chi, appended by Ch'u Shao- sun, who lived at the very time, is less complete but adds certain valuable details. Regarding Chao Kuang-han's attempt to dissuade the Chancellor from accusing him, it states that he "again sent someone to coerce and frighten Chancellor Wei concerning the matter of his wife's illicit killing of a serving female slave; while [at the same time] he secretly and individually presented a memorial requesting a thorough judicial investigation into it." It also says that "in fact [the female slave] had not been killed with a weapon." The Chancellor's counter-accusation charged Chao Kuang-han with the iniquity of making "a false accusation of his wife's illicit killing of a female slave." (?7.) It is not known what punishment the Chancellor and his wife would have received, or should legally have been sentenced to, if Chao Kuang-han had been able to prove that the wife had murdered the female slave. But it is quite clear that Chao Kuang-han, who knew the law, pressed his suspicion in an official accusation in order to prevent the Chancellor from accusing him of serious crimes. It is equally clear that the Chancellor was thereby placed in such a dangerous situation that he had to request an impartial judicial investigation to establish his wife's innocence. One could hardly demand from the available historical sources more conclusive evi- dence that in Former Han times masters could not legally or securely kill slaves who were innocent of any crime. This is a good and necessary corrective for some modern Chinese writers who assert that masters had absolute and unlimited rights over their slaves. 156 SLAVERY IN THE FORMER HAN DYNASTY Yet Emperor Kuang-wu's edict — "he who kills a male or female slave will not receive a reduction in punishment" — makes it clear that when masters did kill their slaves illegally during the previous part of the dynasty they were punished less severely than they would have been for killing free people. On the other hand, masters seem to have had the right to kill slaves guilty of certain crimes, not now known, provided they first reported to the proper government authorities. The second and third of these conclusions indicate the inferior legal status of slaves, while the first shows that they had some legal protection. Slaves in the Courts of Law The rights of slaves to make accusations or act as witnesses in law courts also illuminate this question of slave status. Under T'ang law a slave could not accuse his master except in cases of rebellion and high treason. A slave who reported other crimes was strangled whether the accusation was true or false, while the master was treated as if he had voluntarily confessed, and was thus excused from punishment or received a reduction in it. Further- more, slaves who accused their masters' relatives were punished by banishment even if the charge was true. The codes from Sung through Ming followed the T'ang code with little change. ^ The principle underlying these laws seems to be part of the larger and fundamental Chinese principle that within a family group — as well as within certain other more extended groups — persons of inferior rank or age could not accuse their superiors or elders. Such accusations profoundly violated the fundamental ethical and moral values which constituted the framework of Chinese society. Now this moral principle was already well entrenched in Former Han times. There are clear records showing that sons who accused their fathers could be punished by execution. For example, the Heir-apparent of the King of Heng-shan submitted a document to court accusing his father of plotting rebellion. Although this was true, and the King was therefore executed, the Heir-apparent was also executed on the grounds of unfilially (pu hsiao) accusing his father. In another instance, an eclipse of the sun was blamed upon the immorality of Chancellor Wang Shang. One of the charges was that his son, Wang Chiin, had actually prepared a document 1 T'ang lu su i, ch. 24. See also Wang Shih-chieh (op. cit., p. 322) and Deloustal (BEFEO, vol. 8, 1908, pp. 191-192). STATUS OF SLAVES 157 accusing his father [of murder and incest(?)]. Wang Chun's wife seized the document to show it to her father, Shih Tan, who was so disgusted by this family schism that he made her leave Wang Shang's house. This family disharmony was listed as one of the reasons why the Chancellor was unfit for his post. He was dismissed. ^ There is no direct evidence that in Han times prohibitions which forbade people of lesser rank in a family to accuse their superiors were extended to slaves accusing their masters. Yet slaves were included within the incest group to the extent that a son might not have relations with a female slave who had been favored by his father.2 There is, furthermore, a censorious and possibly significant statement by the historian concerning Wang Mang's confiscation of illegal profits: "This opened [the way for] officers to inform on their generals, and male and female slaves to inform on their masters." (128.) Though this remark proves nothing concerning laws against accusations by slaves, it does indicate the historian's belief that such conduct was reprehensible. The testimony of slaves was admitted as evidence against their masters in cases already under judicial investigation. This is quite different from an accusation brought by a slave in the first instance. Liu Ch'ii, the unspeakably cruel King of Kuang-ch'uan, was finally brought to justice because a slave who murdered a woman on his orders was apprehended and confessed. An imperial decree ordered that the Queen, concubines, male and female slaves, and other witnesses be put in prison, and their testimony about the sixteen murders formed the basis for the King's banishment (64-). Chao Kuang-han seized the slaves of Chancellor Wei Hsiang as witnesses against their master (76, 77). Six government slave women are listed among the chief witnesses against the imperial concubine nee Chao in the important investigation of the murder of two im- perial sons {107). One document indicates that government slaves could bring accusations against free men. In this case the older slave brother of a government slave woman reported that his sister had been debauched by a Gentleman. It is interesting that the slave's accusa- tion actually reached Chang An-shih, the Superintendent of the Gentlemen of the Palace, even though he suppressed the matter and punished the slave. The historian made an approving comment 1 CHS, 44, 7a; and CHS, 82, 2a, respectively. Cf. documents 37 and 98 for more complete details on the background of both cases. 2 Cf. 37 and 1^8 for cases punished, and 59 and 98 for cases reported with censure. 158 SLAVERY IN THE FORMER HAN DYNASTY that Chang An-shih "always concealed others' misdoings in this manner." {63.) "Mixed Marriages," and Status of Children In the matured slavery system of T'ang times intermarriage between slaves and free people was legally restricted. Free women, especially, could not marry male slaves. If they did, the girl and the slave were separated, and the master, his slave, and the girl's family all were punished. If male or female slaves passed themselves off, or were passed off, as "good people" and became the husbands or wives of "good people," they were punished by two years of penal servitude, and each member of the marriage was returned to his former status.^ There is no record of such legal restrictions on marriage between slaves and free people earlier than T'ang times. Even the class of laws on marriages, under which these restrictions fall in the T'ang law, was not instituted before the Northern Wei period. ^ It seems unlikely, in fact, that such alliances were forbidden in Han law. Early in the Han period, Ch'ao Ts'o memorialized Emperor Wen about the problems of frontier defense, and suggested that male and female criminals, private male and female slaves contributed to the government, and civilian volunteers be sent to establish per- manent agricultural garrisons. He also proposed that the govern- ment should buy mates for those migrants who were single (19). The plan to set up colonies received imperial approval but it is not told whether all details suggested by Ch'ao Ts'o were adopted. All we know, therefore, is Ch'ao Ts'o's proposal that unmarried male and female criminals, slaves, and plebeians be provided with mates who had been purchased and therefore presumably were slaves. After establishing colonies all of them would probably have been freed. The palace slave woman named Tse had a free husband. She worked in the Palace of the Imperial Concubines in 63 B.C., but was in contact with her commoner husband, for she had him submit 1 T'ang lii su i, ch. 14. Cf. Liang Ch'i-ch'ao, op. cit., p. 548, and Wang Shih- chieh, op. cit., pp. 316-317. Liang Ch'i-ch'ao points out that if a male slave tried to elevate his status by marriage he was punished, but if a free woman lowered her status by marrying a slave she (or her family) was much more severely punished. 2 Cf. Ch'eng Shu-te, op. cit. ("Hou Wei lu k'ao"), p. 413, and ("Pei Ch'i lii k'ao"), p. 468. Evidence of earlier legal restrictions may exist, although Shen Chia-pen (op. cit., p. 15a) starts his citations on this subject with the Chin History (i.e., after a.d. 1115). STATUS OF SLAVES 159 for her a memorandum claiming special merit because, some twenty years before, she had nursed the infant who eventually became Emperor Hsiian, Yen Shih-ku's comment about her matrimonial status actually befogs the situation: "It means that at the time when she had not yet become a palace slave woman she had had a former husband, who at this time was among the populace." Does he mean that formerly she had been a free person married to a free man, and then became a palace slave, while her husband remained free? Or was she previously a government or private slave married to a free man, and then became a palace slave? Or was she previously a government or private slave married to a slave who later became free? All that is certain on the basis of the historical text is that she was a palace slave woman with a free husband (80). Wang Lin-ch'ing killed the lover, or husband, of one of his female slaves about 7 B.C. Yen Shih-ku says that the man was an outsider who had had illicit relations with the slave, but the additional com- mentary, also quoted by Wang Hsien-ch'ien, says that he was the husband to whom the serving girl was married. The term hsil (a son-in-law), which is used to designate the man, makes the second interpretation seem more likely, especially as it is used in the Fang yen for the act in which commoners mated female slaves {11, and footnote 4). Wang Lin-ch'ing should have been prosecuted for this. murder, but the official who knew all about the case did not think the time was appropriate, especially in view of the fact that Wang Lin-ch'ing had just been dismissed from court (112). There is no recorded instance of a free woman married to a male slave. This was the type of alliance most rigorously forbidden in T'ang times. Yet a social stigma against such marriages is perhaps reflected in the great reluctance of the Elder-Princess of Yang-hsin to marry Wei Ch'ing, the most eligible of the marquises, because he had once been a slave in her household. His later position, rather than the earlier status, decided the issue, and the Princess arranged the marriage through the Empress nee Wei and Emperor Wu, as intermediaries {26, and footnote 12). Further evidence of a social stigma against mixed marriages resides in the Fang yen definition of the terms tsang and huo, always used demeaningly: "In the north- ern outskirts of Shantung, and in the northern countryside of Hopei all plebeian males who mate with female slaves are called tsang; [plebeian] women who become wives of male slaves are called huo." A variant meaning of these terms, given by the third century writer, Wei Shao, carries the same connotation: "When a good man 160 SLAVERY IN THE FORMER HAN DYNASTY takes a female slave as wife and she bears a child, [the child] is called a huo; when a male slave takes a good woman as wife and she bears a child, [the child] is called a tsang." ^ Thus there is nothing to indicate a legal restriction against marriages between slaves and free people, and there is some reason to believe that in practice they occurred. What was the status of the children of such marriages? No cases reveal the status of children born of a free mother and a slave father. Actual marriages of that sort must have been rare, at least among the classes of society that appear in history.- On this point no conclusion can be squeezed from the available texts. Children of slave women by free men not their masters were probably slaves. The chief evidence for this concerns the children of "old lady" Wei, who was a slave in the household of the Marquis of P'ing-yang. He was the husband of Emperor Wu's older sister, the Elder-Princess of Yang-hsin. "Old lady" Wei had six children, probably in the following order: A boy named Wei Chang-chiin, three girls named Chiin-ju, Shao-erh, and Tzu-fu, and then two boys named Ch'ing and Pu-kuang. She may have been "married" to a man named Wei, who fathered the first boy and Tzu-fu. Whether he was father of the two other girls is uncertain. Ch'ing was the son of a prefectural clerk named Cheng Chi, who had relations with Dame Wei while serving in the marquis' household. Pu-kuang's father is not specified. Ch'ing and Pu-kuang adopted the surname Wei when their older sister became important. All the children grew up in the household of the marquis, and three of them were almost certainly slaves, as the following biographical items reveal. When Wei Ch'ing was still young he went to live with his father, who employed him as a sheep herder. He considered himself a slave, and his father's other children by the real wife "treated him as a male slave and did not count him as a brother." When he gi'ew up he was a cavalry man in the household of the marquis as an attendant for the Princess (26). Cavalry escorts of this sort were frequently slaves (see p. 179). The girl Shao-erh was a "serving one" in the household of the same marquis. The term applied to her was often used as a functional 1 Document 11, and footnote 5. See also discussion of terminology, pp. 68-69, above. 2 Three instances of sexual relations between free women and male slaves are recorded but these were not marriages, and only one resulted in offspring (37, 72, 85). This last was a case of general promiscuity so that the paternity of the child could not be determined and the child was done away with. STATUS OF SLAVES 161 designation of female slaves. Furthermore, she had a child by secret relations with Ho Chung-ju, who, like Wei Ch'ing's father, was a prefectural official serving a term in the marquis' household. This child was Ho Ch'u-ping. He, too, grew up in the marquis' house- hold and did not see his father till twenty years later {29). Wei Tzu-fu was a chorus singer for the Princess of P'ing-yang (27). Comely slave children were regularly trained as entertainers in the great households. Because of her great good fortune we know these details about the family, and its subsequent rise to glory. Shortly after Emperor Wu ascended the throne he visited his older sister. There he saw Wei Tzu-fu singing and dancing in the chorus and was delighted with her. He presented his sister with a thousand catties of gold, and she accordingly memorialized that W^ei Tzu-fu should be sent to the Emperor's palace. She bore Emperor Wu three daughters, and finally, in 128, his first son. Thus she became Empress. The family fortunes began rising as soon as she conceived her first child. Wei Chang-chiin was called to the palace but died before his sister became Empress. Chiin-ju married Emperor Wu's boyhood retainer. Chief of Stud Kung-sun Ho, who later became Chancellor. Shao-erh apparently married Ch'en Chang (a great-grandson of Ch'en P'ing), with whom she had formerly had relations. Wei Ch'ing was made a Grand Palace Grandee, and then so distinguished himself as a general that he won a marquisate and later became General-in-Chief. Nothing is said about the youngest boy, Pu-kuang, but Ho Ch'u-ping, the son of Shao-erh by a secret intimacy, arose from the marquis' household to be an even greater general than his uncle. Ho Ch'ii- ping's younger half-brother by his father's real wife also came to court. This was Ho Kuang, the maker of emperors, who dominated China for two decades after the death of Emperor Wu. The interesting details of this family history should not obscure the point of primary interest — that the children of this slave woman retained their mother's status, or at least grew up in her master's household, until an unexpected event entirely changed their position in society. Unfortunately, it is a unique case in early records; there is no other specific evidence from Han times to support the belief that children by slave women and men not their masters were slaves. There is yet a third combination. A master had sexual rights over his slave woman provided she had not been used by his father. The child of a slave woman and her master might be free, especially if it were a boy. This situation is closely related to concubinage. 162 SLAVERY IN THE FORMER HAN DYNASTY The terms "female slave" and "concubine" sometimes appear to- gether as though there were no great distinction between the two (26, footnote 5; 56; 60, footnote 4; 72; 106), and there are references to selling children and marrying off wives (AlS, ISIi), and to women becoming slaves or "lesser wives" {136). Favored female slaves were within the same incest group, with reference to sons, as the principal wife and concubines. Formal distinctions between the principal wife and all other women, and between various ranks among concubines were probably established by marriage contracts between the man's family and the respective families of the women. Yet the terminology applied to a woman may conceal as much as it reveals, because it indicates her status only at a particular time. In the imperial household, about which information is fullest, the mother of the boy selected as Heir- apparent generally became Empress — if she was lucky enough not to be murdered by a rival — even though she had been only a concubine. 1 Even the son of a slave woman might become Emperor. This happened in the case of Emperor Hstian {55) and might have happened to Emperor Wu's first Heir-apparent {27) and to the son of Emperor Ch'eng by the palace slave Ts'ao Kung {107). Wang Li tried to promote a boy that he claimed was the son of the Emperor Ai by the slave Yang Chi {A18). Superficially the issue in this last case hinged on a question of fact — whether the Emperor was indeed the father — though the matter was probably decided accord- ing to the relative power among aspirants for the regency. When a woman became a favorite her status changed, and this change might be made official by granting her an established rank in the imperial household. Wei Tzu-fu became a /w-jen after she had conceived her first child, and later became Empress. The onetime slave Chao Fei-yen became a chieh-yu, then Empress, even though she had no children. Wang Weng-hsu was scorned by the wives and concubines of the imperial grandson Shih; they called her a "householder," properly considering her a slave. But because she bore a son she became a fu-jen, and when the boy became Emperor she was posthumously called "Empress." Both Ts'ao Kung and Yang Hui could have become empresses, though perhaps fictitiously, had their sons been put on the throne. 1 Cf., for example, document 31, and HFHD, vol. I, pp. 229-300, as well as Dubs's introductions to other reigns, in this and succeeding volumes. For similar situations of inheritance and intrigue, cf. documents 37 and 6^ and the revealing article by Wu Ching-ch'ao, "Liang Han to-ch'i ti chia-t'ing (The polygamous family of the Han dynasty)," CLHP, vol. 1, 1931, pp. 47-57. STATUS OF SLAVES 163 The important point in regard to status of children is this: The originally humble position of the mother was no bar to the child's advancement if there was some good reason for his selection. Similarly, in private households — patterned after, or more prob- ably the model for, the imperial system — when the principal wife bore no son the male child of a slave girl might become heir and his mother at least a concubine. The case of Chu Po, who had no son, is a good example. Wang Mang bought a supposedly fertile slave girl and gave her to Chu Po, hoping that she would bear him a son (108). Had this happened there can be no doubt that the child would have had the status of his father. In spite of the attempt, Chu Po had the calamitous Chinese misfortune to die without a son, and had only a daughter by his real wife. Yiian Shao, a leading general in the wars which finally divided China into the Three Kingdoms, was the eldest son of Yiian Feng, but not by the principal wife. His mother's rank is not stated specifi- cally in his two official biographies, but one commentary says that he was the "commoner son" of Feng; another that he was the son of a concubine. The most specific statement comes in a speech of his enemy Kung-sun Tsan, who, recalling the principle of the Ch'un- ch'iu — that a son takes his rank from that of his mother — said that Shao's mother was a chamberlain slave so that his station was really base and lowly. The mother's original status is confirmed by an angry remark of Yiian Shu, Shao's younger half-brother by their father's principal wife. The two men were rival generals, but Shao was far more successful in acquiring allies. Furious at this situation, Shu said: "Why does the crowd not follow me, but instead follows my family's male slave?" He wrote to Kung-sun Tsan that "Shao is not a son of the Yiian family." ^ Putting the facts together it is safe to deduce that Yiian Shao's mother was, in fact, a female slave chamberlain. When she bore Shao, the first male child, she became a concubine, and the boy was recognized as a member of the family. Then the principal wife belatedly bore Shu, who was the natural heir. Shao's position was less secure. Long after, when the half-brothers were rivals, Shu tried to claim that Shao was really his family slave and had no 1 Biographies of Yiian Shao: HHS, 104A, and San kuo chih, Wei chih, 6, 6b- 13a; ref. la and 6b, respectively. For statements of Kung-sun Tsan and Yiian Shu: HHS, 103, 4a, and 105, 3b. There is reason to believe that Ho Kuang's second wife, Hsien, the mother of Ho Yii, was originally a slave, though this is only the assertion of a second century commentator. Certainly there was nothing inherently improbable in his statement {72, footnote 3). 164 SLAVERY IN THE FORMER HAN DYNASTY right to the family surname. While this happened some time before A.D. 192, long after the end of the Former Han, it indicates for its own period how the problem was handled when a female slave bore a son to her master. The casual intimacies between masters and their slave women must have resulted in innumerable children who never achieved free status, but naturally we know nothing of them. Regarding slave status there are other legal questions which remain unanswered. Was there a conflict between the legal concepts of slaves as property and as human beings? The only suggestions of this are imperial decrees which cited the Confucian precept and principle that "in the nature of heaven and earth, man is most im- portant." {122, 137.) It was this principle which Emperor Kuang- wu adduced as the basis of his law that people who killed slaves would not receive reduction of punishment. It is not clear to what legal protection or redress slaves were entitled when people other than their masters committed crimes of violence or lust against them, or whether, in such cases, their masters could sue for damages. To summarize the legal aspects of slave status during the Former Han period: slaves appear to have had a definitely inferior position in criminal law. They were punished for crimes against free people more severely than free people who committed similar crimes against their fellows. A free person received lighter punishment for killing a slave than for killing a free person. But he could expect to be punished. Masters who killed their innocent slaves were punished, though they could apparently kill slaves guilty of certain crimes provided they received permission from the government first. Probably slaves could not bring legal suit against their masters, but their testimony was accepted in court if a case was already under investigation. There is no evidence that laws forbade slaves to marry free people. Probably children of slave women by men not their masters were normally slaves; but sons by the master might attain full family status. VII. SLAVE OWNERS AND NUMBERS OF SLAVES A knowledge of slave owners is more important for an under- standing of the function of slavery in Chinese economics and social structure than an acquaintance with slaves themselves. When we know what sorts of people owned slaves we can begin to ask why the slaves were owned, and to what distinctive use they were put. Glancing through the translated documents one might gain the impression that all sorts of people owned slaves, and accordingly that the slave population of the Former Han period was very large. How should we classify the following slave owners: The scholar- poet Wang Pao (83); the squire of the Chancellor of Ch'i, who had a male slave attendant (22); leaders of the robber gangs of Ch'ang-an, who used "youth" escorts to give the appearance of being solid citizens {81 ) ; the redresser of wrongs, Ylian She, whose male slave involved him in a nasty mess with the local official (125); a former wet-nurse of Emperor Wu {A6) ; a rich widow (A20) ; the charcoal-burner who owned Tou Kuang-kuo (H); several knights fighting against the army of Wu in the Rebellion of the Seven States (25) ; and three slave women each given ten slaves by the Brilliant Companion nee Chao (107)"! The only common characteristic of these owners is that they were important enough in their own right, or indirectly, to be mentioned with their slaves in literature and history. Yet this characteristic is exceedingly important. It is the nub of the whole question of classifying slave owners. Ownership per se was of no special interest to the Han historians except to illustrate the wealth of an individual, his imperial favor, or some other notable matter in which slaves had to be mentioned to make the facts intelli- gible. Items about slaves have a certain random and incidental quality, but it is no accident that only certain sorts of owners were granted space in the enduring chronicles. They were people who were for some other reason important historically. This situation colors the whole picture of slave ownership, but does not distort it entirely. The ability to own slaves depended largely upon wealth, and there is a real though rough correspondence between wealth and historical importance. Political or military ability and noble status, which made men historically significant, also usually led to riches. Wealth, on the other hand, often opened the way to political or social dominance, and also was occasionally an independent criterion of importance. Yet historical importance and wealth could 166 166 SLAVERY IN THE FORMER HAN DYNASTY be entirely distinct. Many of the scholars and writers accorded significant space in the histories of the Han dynasty were poor men. Thus the gross correlation between historical importance, wealth, and large-scale ownership helps to correct historical bias, while importance dissociated from wealth offers a means to check that bias by presenting a diversity of "important people" among whom to look for types of owners. Furthermore, we are not entirely de- pendent upon analyses of individual owners, for edicts and memorials often indicate the types of people that owned slaves without reference to individuals. Types of Slave Owners Edicts of Emperors Ch'eng and Ai emphasized members of the nobility and high officials as extensive slave owners (105, 110). More than twice as many Former Han documents mentioning specific owners refer to noblemen as to untitled officials and plebeians. Thus the nobility constituted an important slaveholding class. Probably every nobleman of any consequence owned slaves. It is curious that some emperors, empresses, and imperial con- cubines owned private slaves, even though they had an almost un- limited number of government slaves at their disposal. Emperor Ch'eng was severely censured by his high ministers for keeping private male slaves in the palace and taking them as companions on his incognito journeys {10I^). The Empress nee Shang-kuan sent her private male and female slaves to guard the graves of her father and grandfather who had been executed for treason {62). Since she was only eight years old it may be assumed that the slaves were sent on her behalf by some minister. Presumably these slaves had belonged to the Shang-kuan family; some of them may have accom- panied the child Empress into the palace as her personal servants and guards. The imperial concubine nee Chao was given three women as her private slaves {107), and the Empress Dowager nee Fu, who must have been surrounded by palace slaves, nevertheless bought a number from various government bureaus {116). Because kings had royal governments which were replicas of the imperial government early in the Han period, they controlled their government slaves as well as household ones. Yet it appears that private slaves are referred to in the frequent textual association of kings and slaves unless government slaves are specified.^ The King of Ch'i-pei certainly owned the four entertainers he bought in the ' As in A8 and 69. King's household slaves are specified in 12 and probably 75. Other references: 21, 37, 38, U, 6U, 67, 68, 8j^, 85, 106, 109, 110, and 113. SLAVE OWNERS AND NUMBERS OF SLAVES 167 common people's market {21), and Emperor Ai was referring to pri- vately owned ones when he limited kings to two hundred adults {110). Princesses, mentioned in the previous chapter as owners, were given slaves when they set up independent households. For them Emperor Ai thought a hundred male and female slaves should suffice. An impressive number of documents mention marquises as slave owners.^ They were the most numerous group in the nobility, usually wealthy from the fixed incomes of their estates, special government grants, and investments in private land. Many main- tained expensive households and needed slaves as servants and for social display. The nobility form a natural category easy to isolate as a class in society and clearly identified by contemporary writers as a slave- owning class. Other groups are much more difficult to classify. We have to decide which system will be significant from the view- point of economics and social structure. Three groupings, on the basis of occupation or sources of income, appear valid: officials, merchants or manufacturers, and owners of large amounts of land. Noblemen cannot be excluded from any of these groups, nor are the groups entirely distinct from one another. The occupations and sources of income of owners were diversified. Both officials and merchants were frequently landlords; nobles were generally land- lords, often officials, and sometimes merchants. Yet many masters seem to have belonged primarily to one of the three groups, and the groups were recognized in Han times as a way of classifying the people of substance. An imperial decree of 13 B.C. chid lesser officials for emulating the nobility and great ministers in slave-owning and other extrava- gances {105). At the end of the Former Han, out of 130,285 officials probably only a small proportion received enough salary to maintain slaves.^ The vast majority were petty clerks and minor 1 Or presumed owners. The documents are not always specific. Marquises, sons of kings: 51, 52, 115, All; marquises by affinal connection: 59, 61, 62, 82, 98, 99, 100, 108, 109, 111, IH, 119, 120; marquises for military merit (other than in group two): 26, ^8, 1^9, 71, 76, 87, 102, 103. Distinctions between the second and third group are not as clear-cut as appears from their separate listing in CHS, 18, and 16-17. I attempted here, perhaps unwisely, to make distinctions that did not always follow the CHS system. ^ The figure comes from CHS, 19A, 8a. Wu Po-lun ("An investigation of slavery in the Western Han," p. 279) makes the wild estimate that officials averaged a hundred slaves each, thus alone counting some 13,000,000 slaves! Wu Ching- ch'ao ("The slavery system of the Western Han," p. 270) compares salaries of lesser officials with normal living costs to show that most officials lived very unpreten- tiously. See also Ma Ch'eng-feng (An economic history of China, vol. 2, p. 245) for further refutation. 168 SLAVERY IN THE FORMER HAN DYNASTY bureaucrats who could do little more than support their families on their modest incomes from salary and petty graft. But well-to-do officials, especially those seeking to climb by social means, probably maintained as many slaves as they could afford. Lu Chia was given a hundred, together with considerable sums of money, precisely so he could mingle in official society at the court (15). A few of the great officials known to be slave owners are Shang-kuan Chieh and Shang-kuan An {59, 61), Ho Kuang (65), Chang An-shih (71), Wei Hsiang (76), Shih Tan (97), Wang Shang (98), Wang Ch'ung (120), and the various relatives of Empress Dowager nee Wang, who one after another controlled the empire (99). These men formed the pattern of society, and though they were noblemen also, they and their kind must have set the pace for official life. Yiian Ang (23), Shih Fen (A5), and Wang Tsun (A15) were middling important officials probably representative of the lesser slave-owning offi- cialdom. Emperor Ai considered thirty slaves adequate for any official not a nobleman, and probably few were able to maintain that number on their salaries alone. The first requirement of the new dynasty was to revive agricul- ture and place it on a productive, that is to say, tax-paying basis. Attempts were made to prevent business men from investing their profits in farm land, thereby dispossessing independent small farmers who were the backbone of the tax-paying peasantry. Passages dealing with this policy show substantial merchants and manu- facturers to have been a slave-owning class. A recommendation presented to Emperor Wu by his high ministers in 119 B.C. suggested that any merchant who owned private fields should have his land and his slaves confiscated (^5). Between 119 and 113 B.C. the government acquired thousands of slaves, together with fields, houses, and money, by confiscating the accumulated fortunes of well-to-do business men and merchants (^6). Chang Shou-chieh, an important eighth century commentator, states in this connection that merchants were taxed double for their slaves and other property (127, footnote 1). Besides these general indications we have specific cases of merchants or manufacturers who were owners: the Cho family of Szechwan, wealthy from iron-smelting and trade, who had hundreds of "youths" (2); Tiao Chien of Shantung, who used slaves in salt-refining, fishing, and as his traveling and resident agents (17); and, finally, Marquis Chang An-shih, a manufacturer who, during the reign of Emperor Hsiian, became fabulously wealthy from the sale of products turned out by his seven hundred household "youths" (71). SLAVE OWNERS AND NUMBERS OF SLAVES 169 Land was the indestructible foundation of any "permanent" fortune, and the most common investment. Probably all wealthy people put their excess capital into land ; even merchants, stringently restricted by early laws, could not be effectively prevented from acquiring it. Therefore, the classification of landlords is practically synonymous with the classification of wealth. It embraces nobility, officialdom, and business. It also covers a class of owners who were almost exclusively landowners, never acquiring noble or official status and indulging in little trade beyond disposing of the harvests derived from their tenants. The subject of landlordism and slavery is discussed at length in a later section on slaves in agricul- ture (pp. 195 ff.); suffice it here to mention Cho Wen-chiin, who was given a hundred "youths" and a hundred myriad cash by her father and immediately bought fields and houses (28) ; and the edict of Wang Mang, forbidding people to buy or sell lands and slaves (122). Among the groups thus segregated it is difficult to determine which was numerically largest or which owned the most slaves. Yet in spite of the historical bias it is logical to believe that the nobility were the distinctive owner class, possessing more slaves on an average than any other group. The nobility had among all classes the freest access to wealth: through fixed incomes from estates, rents from private farm land, special grants from the treasury, official salary, graft, appropriation under the cloak of imperial favor, and, to a certain extent, the profits of trade. Not all noblemen enjoyed all these sources of wealth, but no other group had access to more than a few of them. Moreover, because of their position in society the nobility organized their mode of living most closely on the pattern of the imperial household, in which slaves played a prominent part. No other group had such a conspicuous need for numerous slaves employed for luxurious living, entertainment, and display. Their mansions, particularly those of the kings, princesses, and marquises through affinal connection, were run like little imperial palaces, and fixed the pattern to be aped by officialdom as well as by powerful local gentry who derived their wealth from land and trade. Numbers of Slaves Individually Owned Chinese students of Han slavery frequently present lists of owners recorded as having many slaves, and use the lists to create the impression that many people owned large groups of slaves, or that the slave population was indeed considerable. Among the 170 SLAVERY IN THE FORMER HAN DYNASTY translated documents the list of owners of many slaves, arranged chronologically, runs as follows: (1) After 228 B.C.: The Cho family owned 800 (variant, 1,000) "youths." (2.) (2) 181-180: Lu Chia was given 100 male and female slaves (15). (3) 144: Cho Wang-sun owned 800 "youths" and guests (or "youth-guests"; SC gives "household 'youths' "), and gave his daughter 100 "youths." (28.) (4) 144: Cheng Ch'eng also owned several hundred ("youths" and guests; as in No. 3) (28). (5) Ca. 140: Emperor Wu gave his half-sister 300 male and female slaves (31). (6) 113: Emperor Wu gave Luan Ta 1,000 "youths." (I^9.) (7) Ca. 87-68: Imperial grants totaling 170 male and female slaves were given to Ho Kuang (70). (8) 74-62: Chang An-shih owned 700 household "youths." (71.) (9) 67: Liu Ho, ex-King of Ch'ang-i, owned 183 male and female slaves (75). (10) 33-15: Imperial grants of "youths" and male slaves to Shih Tan were num- bered by the hundred (97), (11) Before 25: Wang Shang's whole clan owned private male slaves numbered by the thousand (98). (12) Ca. 23: The brothers of the Empress Dowager nee Wang owned "youths" and male slaves numbered by the thousand or hundred (99).^ A few points about this list are worth emphasizing. Only two of the twelve items employ specific figures; all the rest are round numbers in terms of hundreds or thousands. Thus, most of them are symbolic; they are rough estimates and probably not reports of known numbers. Seven use the term "youth" for either the whole figure or part of it. As shown in the discussion of terminology (p. 67) , this term does not mean slaves exclusively; it often means children. While in the above passages the term probably does refer to slaves, we may note in passing that Chang An-shih used his 700 "youths" in household manufacturing, where free child-labor would be just as useful as, and probably less expensive than, slave labor. Also some of Cho Wang-sun's total includes guests, which he had "num- 1 Shortly before Han times, Chang Liang inherited 300 household "youths." (Al.) Two other pre-Han numbers may be relegated to the obscurity of a foot- note as merited by their extreme unreliability: Lii Pu-wei had a myriad household "youths"; Liao Tu had several thousand household "youths." (SC, 85, 2b; and 3a.) The list can also be projected into the Latter Han period with the following references: Before a.d. 59, the family of Tou Jung had male and female slaves numbered by the thousand (HHS, 53, 5a); in 64 the King of Tung-p'ing received an imperial grant of 500 Palace Women and male and female slaves (72, 5b); in 83 the King of Chi-nan had 1,400 male and female slaves (72, 4a); about 92 the King of Ch'ing-ho was given an imperial grant of 300 male and female slaves (85, 2a) ; before 93 the King of Liang-chieh had been given imperial grants of 200 male and female slaves, and other male and female slaves and "green-heads" from government offices (78, 4a); about 150 Liang Chi took several thousand "good people" and made them male and female slaves (64, 6b); undated, Che Kuo had 800 household "youths" (112A, 6b). These statements, while perhaps indicating a trend toward increase in private slave numbers, need to be judged by the same criteria as are applied below to the Former Han documents. SLAVE OWNERS AND NUMBERS OF SLAVES 171 bered by the hundred." ^ At any rate there is a correspondence between large, vague figures and use of the term "youth"; the eleventh item alone mentions a really large number of slaves (nu) only. Finally, in every case but one these figures are given expressly to portray the wealth, power, or imperial favor enjoyed by the owners. Consider the reasons for mentioning slaves in the items not already familiar from the previous discussion. Cho Wang-sun and Cheng Ch'eng were two of the wealthiest men in Szechwan. Emperor Wu was very liberal to his long-lost half-sister, and extraor- dinarily so to the magician Luan Ta, whom he loaded with riches. This latter account is suspect because it may be meant to illus- trate the extent of Emperor Wu's gullibility and superstition. Shih Tan's wealth measures the extent of imperial gratitude for the part his family played in rearing the orphan who became Emperor Hsiian. The statement concerning male slaves owned by Wang Shang's clan appeared in a memorial by the spokesman of the rival clan, designed to show that Wang Shang was a potential menace to the throne. Exaggeration of numbers would seem to be the very essence of such a memorial. Further, it referred to the holdings of the Wang clan, at the apex of its power. The report of slaves owned by Wang Mang's uncles figures in the picture of their enormous wealth and luxury wherein exaggeration might be expected even if it were not written during the dynasty which overthrew Wang Mang, the "usurper." If these figures were recorded as measures of wealth then they cannot be taken as a common measure of slaves individually owned. They represent the unusual, in some instances the phenomenal and amazing, cases. Therefore, the two documents giving specific num- bers are worth considering carefully. The first casually mentions 183 slaves owned by Liu Ho, the former King of Ch'ang-i (75). It is part of an exact and detailed eyewitness description of the living conditions of the former King, made by Chang Ch'ang, an experienced and trusted administrator, at the specific command of Emperor Hsiian. He accompanied his description with an invoice of the male and female slaves and other property of the former King, who, together with his sisters, had been given the entire household wealth of his kingdom. ' Unless we take the Shih chi reading of "household youths," or assume that "youth-guest" is a compound term for a type of slave, which may be correct. 172 SLAVERY IN THE FORMER HAN DYNASTY Slaves were part of that wealth, and the number 183 could be cheeked against the invoice. It is true that Liu Ho was somewhat hard-pressed financially at the time of the investigation; nevertheless, this document must be used as a primary point of reference regarding numbers of slaves owned by single individuals. The second document is equally important as a point reference for reports of large imperial gifts of slaves. During his entire life- time Ho Kuang received 170 male and female slaves by imperial grant. The figure is revealing for two reasons: Ho Kuang was one of the most important political figures during the entire Han period ; and it is possible to compare the gifts of slaves with gifts of other sorts which he received. He was indirectly related to, and directly sponsored by, the family of the Empress nee Wei. Emperor Wu publicly honored him as his most loyal and trustworthy minister, and on his deathbed appointed Ho Kuang one of three regents for young Emperor Chao. His granddaughter became the new Empress, and he became virtual ruler. When Emperor Chao died in 74 B.C. without heirs. Ho Kuang determined the selection of his successor, Liu Ho, and then after twenty-seven days deposed him. He next selected a great-grandson of Emperor Wu to become Emperor Hsiian, and thus continued the dominant figure at court until he died in 68 B.C. Probably few men received richer rewards than he. Among the totals mentioned are fiefs, with the income from an aggregate of 20,000 households, 7,000 catties (about 55,000 troy ounces) of gold, 60,000,000 cash, 30,000 pieces of silk, 170 male and female slaves, 2,000 horses, and a first-grade mansion (70). Taken in conjunction with these other gifts, whether exaggerated or not, the 170 slaves must have been considered a lavish and unusual -imperial gift at that time. This document compels suspicion about statements of even larger imperial gifts of slaves, especially since each of them uses vague round numbers. Two other documents, one near the beginning and the other toward the close of the Former Han period, support the belief that from one hundred to two hundred slaves constituted a large number for any private owner. The first (No. 2 in the list above) tells of 100 male and female slaves, 50 outfits of carriages and horses, and 5,000,000 cash largess that were sent to Lu Chia for his expenses in helping to engineer the destruction of the house of Lii. Because of this wealth Lu Chia was able to mingle at the Han court on an equal footing with dukes and ministers, and he achieved a bad reputation for extravagance (15). SLAVE OWNERS AND NUMBERS OF SLAVES 173 The documents regarding Liu Ho and Ho Kuang, coming near the middle of the period, probably indicate an advance in numbers privately owned as the dynasty progressed. The trend is further suggested by the following occurrence in 7 B.C. When Emperor Ai came to the throne his former tutor, Shih Tan, memorialized concerning the dangerous contrasts between the wealth of the upper classes and the poverty of the masses. His suggestion that wealth be restricted was handed down for discussion, and the Chancellor and Grandee Secretary, both marquises, pro- posed specific limits on the extent of land and number of slaves which various grades of the nobility, officials, and common people might own. For slaves more than ten and less than sixty years of age these limits were 200 for kings, 100 for marquises and princesses, and 30 for kuan-nei marquises, officials, and commoners {109, 110). Since this was a reform we must assume that some individuals owned more slaves than allowed in the various classes. This is confirmed by the statement that the price of fields, houses, and slaves immediately depreciated. On the other hand, the Chancellor and Grandee Secretary were both noblemen, and they apparently made their proposal seriously, believing that the limits were reason- able and that the law could be enforced. As it turned out, the law was too drastic. It was "inconvenient" for members of the Ting and Fu families, who were affinally related to Emperor Ai, and for his favorite, Tung Hsien. The Emperor ordered it to be deferred "temporarily." But note this: The principal objectors were in the groups limited to 30 ch'ing (340 acres) of land, and to either 100 or to 30 slaves. The total impression created by this document is that while some members of the nobility may have had upwards of 200 slaves the statesmen of the time believed that 200 for kings and 100 for regular marquises were adequate and feasible numbers. They miscalculated the opposition of Emperor Ai's affinal relatives, who were marquises, and of the imperial catamite, Tung Hsien, and various relatives not ennobled, who would have been limited to thirty slaves. Since this was the end of a cycle, a period of extreme luxury, the event strengthens the deduction that throughout the Han period private ownership of 100 or 200 slaves was unusual. Probably a few score made a respectable showing for most people, especially since many slaves were used almost exclusively in non- productive capacities. The number "forty or more" is twice men- tioned for slaves of one or the other sex, owned by men who were 174 SLAVERY IN THE FORMER HAN DYNASTY marquises through imperial affinal connection, in a way that might indicate that the number represented approximately the total number of males or females owned. Some time about 64-58 B.C. the pathologically jealous wife of the Marquis of Chiang-ling strangled to death more than forty serving female slaves {82). Likewise, shortly before 15 B.C. the imperial favorite, Chang Fang, who was exceedingly wealthy, sent his senior male slave, Chiin, and forty or more others to invade a government bureau and square a personal grudge {100). It is probable that for this job he mustered all or most of the male slaves available. Total Number of Slaves In attempting to reckon slave numbers during the Former Han period we work in such a vacuum that estimates by Chinese students have ranged as wildly as from 600,000 to between twenty and thirty millions! ^ And this disparity figures in a total population given as 59,594,978 at A.D. 2.- Is there any way honestly to estimate the maximum slave population during Former Han times? Only a very vague way. The starting point for any such attempt is document 89, where Kung Yii, in a memorial dated 44 B.C., speaks of the ten mjrriad and more government slaves. The possible accuracy of this state- ment is discussed extensively in a footnote to that document where it was concluded that Kung Yii was a man of the highest reputation for integrity, and with full access to such facts as were known. If he erred it was probably on the side of exaggeration. The figure is in reasonable accord with other independent figures of special groups in the population. Before 44 B.C. government slaves were probably less numerous. Kung Yii was about eighty years old, and had held several govern- ment offices earlier in his career before he went into retirement. iWu Ching-ch'ao, op. cit., pp. 269-270; and Wu Po-lun, op. cit., pp. 278- 279, respectively. ^ CHS, 28B, 9a. The figure derives from imperial censuses made from time to time for tax purposes, as discussed in 89, footnote 3. Wu Ching-ch'ao (op. cit., p. 270, footnote 6) believes that the actual population was larger, on the assumption of deception to avoid taxation, a regular Chinese practice. On Han population, see also Ma Fei-pai ("Source material on the economic history of Ch'in and Han," pt. 5, "Jen-k'ou chi t'u-ti [Population and land]," Shih Huo, vol. 3, No. 3, Jan. 1, 1936, pp. 102-132), and Lao Kan ("Liang Han hu-chi yii ti-li chih kuan-hsi [Popula- tion and geograpiay in the two Han dynasties]," Academia Sinica, Bulletin of the Institute of History and Philology, vol. 5, pt. 2, Dec, 1935, pp. 179-214 and 215-240). All Chinese population figures are highly suspect; they must be used as rough approximations, indicating trends in growth and decline, and not as accurately determined numbers, even though solemnly recorded down to the last unit. SLAVE OWNERS AND NUMBERS OF SLAVES 175 Possibly he noted, during his second period of office, from 48 to 44 B.C., that the government slave population had grown during his lifetime, for his memorial emphasized the magnitude of the figures. Independently of this deduction, it appears evident that the govern- ment owned many fewer slaves during the reign of Emperor Wu, for when it confiscated from a thousand to a myriad private slaves between 119 and 113 B.C. it was simply embarrassed by the superfluity (^^), which indicates a rather low "absorption point" for new slaves. Wei Hung, living in the first years of the Latter Han dynasty, mentions 30,000 slaves tending horses on 36 govern- ment ranches during the previous epoch (93). His figure, for one of the most extensive occupations of government slaves, is not startlingly large. Unfortunately it is undated and may apply only to the closing years of the period. It is not known whether the number of government slaves in- creased in the half century between 44 B.C. and a.d. 2, or if it did, by how much. Emperor Ai ordered all government slaves over fifty years of age to be freed, but it is not certain how thoroughly the order was carried out. Probably it did not reduce the totals greatly if we assume for China then, as in recent times, a low percentage in the higher age levels of the population. Even a 50 per cent growth in government slave numbers would only bring the total to 150,000 or so, which is suggested as a liberal estimate. Apparently Chinese students have found no more accurate way of determining the private slave population than to multiply Kung Yii's figure for government slaves by some arbitrarily selected multiple. No one has approached the subject by estimating the number of slaves owned by the nobility, who were the most important group of private owners. Chapters 14 to 18 of the Ch'ien Han shu give fairly complete lists of the nobility under various classifications, emphasizing original ennoblement but carrying each line down until it ran out or was dropped from the record. The longer a noble line lasted, the scantier becomes the information about it, especially in regard to the date when it died out. Therefore it is not easy to determine how many noblemen there were at any given period; often it is necessary to guess whether a noble line still existed by counting the number of generations after the last specific date entered. Giving the benefit of the doubt to all uncertain cases, we arrive at a figure not only probably too large, but also including many people who were noblemen in title only, neither prominent nor wealthy. 176 SLAVERY IN THE FORMER HAN DYNASTY On the basis of a careful count there were 17 kings and 206 full marquises (including 77 doubtful cases) in 7 B.C., when Emperor Ai made his famous restrictions on slave-owning. ^ The tables do not list princesses for they did not establish independent lines. Only the sisters of Emperor Ch'eng, and the sisters and aunts of Emperor Ai need be counted. They could hardly have totaled more than ten, for Emperors Ylian and Ch'eng and the father of Emperor Ai were not prolific. Using Emperor Ai's proposed law of 200 adult slaves as a maximum for kings and princesses, and 100 for full marquises (certainly generous for an average), we arrive at surprisingly low figures: 3,400 for kings, ±2,000 for princesses, and 20,600 for full marquises — a total of not more than 26,000. Suppose that the restrictions reduced by an average of one hundred the number of slaves owned by each nobleman. This allows 300 slaves for kings and princesses, and 200 for marquises, and gives a total of only 49,300. Perhaps the fifteen listed kuan-nei marquises should be counted as full marquises even though the Emperor restricted them to 30 slaves each, since these particular marquises were probably wealthy. This would add another 3,000 or 4,500, giving top figures of 29,000 or 53,800 slaves owned by noblemen. By comparison the same calculations were made for the year 44 B.C., when Kung Yii mentioned 100,000 or more government slaves. There are listed for that time 18 kings and 169 marquises, including doubtful cases. Adding a possible 15 princesses, and using the alternative sets of multiples, we arrive at 23,500 or 43,700. Maximum figures were used through every step of these calcula- tions. Kung Yli's estimate of 100,000 or more government slaves was probably a maximum and possibly an exaggeration. All doubtful cases were counted in, listed kuan-nei marquises were treated as full marquises, and the average number of slaves assigned to each group of nobles was probably high even for the richest of them. But there were at least some basic records as points of departure. Any estimate of the remaining slave population is a pure guess — simply a matter of selecting some multiple to apply to the estimated number of slaves owned by nobles. Taking the top figures and multiplying by ten, officials, rich people, and those of moderate means all together owned 437,000 1 The count excludes descendants of all those honorary noblemen who were granted titles in 62 B.C. as collateral descendants of early Han noble lines that had long since run out. They were mostly petty officials or gentry, given titles for sentimental reasons, but neither active nor wealthy. SLAVE OWNERS AND NUMBERS OF SLAVES 177 slaves in 44, and 538,000 in 7 B.C. It would have taken some 22,000 owners in 44, or 27,000 owners in 7 B.C., each able to support a score of slaves, to total those figures. This seems very high when we recall that "itinerant traders and resident merchants, and [people of] middling [wealth] and up were generally ruined" in the great con- fiscation of land, money, and slaves throughout the empire in 119- 113 B.C. Yet they produced only a myriad, or at most a few myriads of slaves! (-4^.) If recorded figures, estimates, and guesses are combined the grand total of government slaves and those of noble and non-noble owners would have been roughly 580,000 in 44 B.C. and 741,000 in 7 B.C. The Chinese population figure of roughly 60,000,000 in a.d. 2 is not exact; the population was probably not less than 50,000,000 and not more than 70,000,000. In the same sense, estimates of the slave population are not accurate; they indicate only that the number was probably less than a million and probably more than 300,000. These calculations, neither scientific nor soundly historic, can be assailed at every step; and yet they portray in a shadowy way the proportion of the total population that was enslaved. It may have been under 1 per cent. This bears upon the heated discussion among Chinese economic historians whether China during the Han period was a "slavery society" or "slave economy society" (nu-li she-hui or nu-li ching-chi she-hui). While their discussion is some- what sterile in the sense that there seems to be no agreement upon the meaning of the terms, and no accepted criteria by which known facts should be judged, it has had the stimulating effect of focusing attention away from pure description to an analysis of the function of slavery in Chinese economy and of slaves in the social structure. These are the subjects of primary interest in the succeeding chapters. VIII. SERVICE FUNCTIONS OF PRIVATE SLAVES The most important thing about slavery is the way in which slaves were used — their function. When we know the function of slaves it is easier to understand the significance of the slavery system itself and to acquire a clearer picture of the social and economic organization of which slavery was a part. In the succeeding discussion two major headings are useful: services and production. The first term covers occupations con- tributing to the comfort, pleasure, or prestige of the private owner, and services for the government. "Production" refers to the pro- duction and processing of goods for sale or use by the master or the government, to merchandising, and to other sorts of activity that increased the owner's wealth. These categories are purely schematic. Owners employed slaves in both ways, and individual slaves worked in both capacities. The first category, services, bears more closely upon society, while the second, production, has more to do with economics. This division clarifies the extent to which slaves participated in the process of production, on the one hand, and the extent to which they were wealth-consuming luxuries, on the other. These two main categories help to explain why the government and private individuals owned slaves at all. Duties of Domestic Slaves One of the principal occupations of household slaves was general servant's work, which is essentially the same throughout the world. Formal histories, recording events of great importance to the state, or presenting biographies of men of affairs, tell little about the work of menials. Casual references mention specific duties, but the mass of routine work — cleaning, washing, cooking, serving, repairing, errand-running and the like — is referred to only in the semi-humorous but very enlightening slave contract by Wang Pao.^ The amount and diversity of work listed in this contract is amazing. On top of multifarious productive duties to be mentioned below, the slave Pien-liao was expected to cook, serve, wash dishes, clean house, 1 The semi-humorous nature of the document does not invalidate it. Even if it were imaginative it would still necessarily be based upon experience and fact, as is pointed out by Ma Fei-pai ("Source material on the economic history of Ch'in and Han, " pt. 6, "The slavery system," p. 393). Whether the account is autobiographical or purely fictitious, Wang Pao, the author, pictures himself as thinking of all the work which might need to be done and of trying to cover every eventuality. 178 Field Museum of Nalural History Anthropology, Vol. 31. Plate I MORTUARY FIGURINE OF A SERVANT OR SLAVE OF THE HAN PERIOD Pien-liao, the slave of Wang Pao, may have worn a similar hat and coat and carried such a broom and dustpan. Collected by D. C. Graham in a cave-tomb near Kiating, Szechwan, and now in the United States National Museum SERVICE FUNCTIONS OF PRIVATE SLAVES 179 tend the barnyard, wash clothes, help in making wine, run errands, and guard the house (83). In other documents we read of a male slave attending his master at the imperial court (22); of another being dispatched to buy meat from a butcher (125); of male slaves caring for horses in the stable at night (7J^); and of ten courtiers becoming slaves to the ex-King of Chao, Chang Ao, so they could follow him to prison (12), undoubtedly to wait upon him. In con- trast with such slaves-of-all-work, domestic slaves in the grander households were highly specialized. Male slaves regularly acted as bodyguards and outriding escorts; they constituted a special class, mounted and armed, who rode with their masters on trips and excursions — in fact, whenever they left the house. Wei Ch'ing as a slave in the household of the Marquis of P'ing-yang was a horseman and attendant for the Princess (26), and we read later of other cavalry slaves in the same household (39). When Emperor Ch'eng went on incognito journeys he was attended by a small group of gentlemen, male slaves, and guests. All were dressed in plain clothes and carried swords, and rode either in small chariots or on horseback (10J^). Wang Lin-ch'ing, a relative of an empress, traveled in a carriage escorted by armed and mounted slaves (112); even the leaders of thieving gangs at Ch'ang-an went about escorted by "youth horsemen," which created the impression that they were fine gentlemen (81). On at least one occasion, which may be representative of a common practice, cavalry slaves went into battle with their masters. During the Rebellion of the Seven States, in 154 B.C., Kuan Fu wanted to avenge the death of his father by a private sortie into the Wu encampment. He called for volunteers among the hardy men of the army who loved him and were willing to follow him. When it came time for the foray, however, the venture was so risky that only two friends dared to go with him. Yet they were unhesitatingly followed into battle by their attendant male slaves. The ten or more horsemen galloped into the Wu army to the very foot of the general's standard, killing and wounding several tens of the enemy (25). Many Latter Han bas-reliefs, sometimes presumably depicting scenes from the lives of the dead, show processions of chariots and cavalry escorts. Some of these outriders may well be slaves occupy- ing positions like those of the cavalry slaves in the previous period.^ 1 Cf. Edouard Chavannes, Mission archeologique dans la Chine septentrionale. Plates, pt. 1, pis. XXIV, XXV, XLIV, L, LII, LV, LVI, LIX, and passim. 180 SLAVERY IN THE FORMER HAN DYNASTY Female slaves, especially young girls, performed intimate services for their masters and mistresses and their guests. Wei Tzu-fu, belonging to the Princess of P'ing-yang, helped young Emperor Wu change clothes when he was a house guest (27). A "serving female slave" (shih-pei) of the rich widow Tso A-chiin helped a drunken guest to bed (A20), and another such girl died of a disease just as she was following the King of Chi-pei to carry his sword when he went out to the toilet (21). Ho Kuang's widow enjoyed being trundled through her elaborate mansion by "serving female slaves" dressed in many-colored silks (72). Wei Shao-erh, the older sister of Wei Ch'ing, and mother of Ho Ch'ii-ping, was a "serving one" in the household of the Marquis of P'ing-yang (29). When Yiian Ang was Chancellor of Wu State he had a "serving child" (23); and Wang Mang once secretly bought a "serving female slave," and gave her to General Chu Tzu-ylian (108). The term fu or fu-pei, translated "chamberlain," or "female slave chamberlain," is explained by Yen Shih-ku as always referring to those who attended to their master's clothes and bed, and he further states that another character /w, meaning "a close favorite," may be used interchangeably (120, footnote 2). Wang Wu's "chamberlain" must have been considerably younger than he, for his son Wang Shang had sexual relations with her (98). When the female slave chamberlain of Wei Hsiang died suddenly. Administrator of the Capital District Chao Kuang-han suspected the Chancellor's wife of killing her out of jealousy (76). Wang Ch'ung was poisoned by his female slave chamberlain and died (120).^ People of fashion in Han times employed musicians, dancers, singers, acrobats, and jugglers to entertain at their banquets. Graphic illustrations of feasts and entertainments also preserved 1 The term fu-pei appears several times in the HHS in ways which also indicate a close relationship between master and female slave chamberlain. In a.d. 95 the Chancellor of the Kingdom of Lo-ch'eng reported two matters against his King, Liu Tang. In violation of an old prohibition that Palace Women sent out to be married could not go to any of the kingdoms (i.e. into a royal palace), Liu Tang commanded a former musician of the imperial Concubines' Quarter to enter his palace, and had relations with her although she was already married to a commoner. He had her husband killed to silence him, and also killed three inner "serving" (slave?) girls to prevent them from talking. Also he took as lesser wife the fu-pei of the late King of Chung-shan (HHS, 80, 2b). Some time after A.D. 162 Fen K'un was memorialized against by a military inspector for going out in uniform accompanied by two fu-pei. No law was found applicable and the charge was dropped (HHS, 68, 4a). The estrangement between Lii Pu and Tung Cho, two of the leading figures at the end of the Latter Han, started because Lii Pu had relations with Tung Cho's fu-pei (HHS, 105, 5b). Yiian Shao was said to have been the son of his father's fu-pei, which cast doubt on his right to the family surname (HHS, 103, 3b-4a, biography of Kung-sun Tsan; see also pp. 163-164, above). SERVICE FUNCTIONS OF PRIVATE SLAVES 181 in bas-reliefs of the Latter Han period probably reflect closely the general conditions only a century or two earlier.^ On these reliefs, dancers, both male and female, some with elongated sleeves, posture in their dances, while drummers beat time on enormous drums. Musicians, pictured individually or in groups, play various wind, string, and percussion instruments. Jugglers nimbly keep three, seven, and even nine balls flying in the air at one time. Acrobats walk on their hands and do cartwheels; in one scene a man supports a p3Tamid of four balancing children, while another precariously keeps his balance treading on a ball. It is in just such occupations that young slaves are regularly mentioned in Former Han documents, and perhaps some of the entertainers pictured in the Latter Han bas-reliefs are actually talented slaves.- The long and careful training necessary for these arts must have begun when they were quite young, and surely enhanced their value greatly. About 175 B.C., Chia Yi protested in a memorial to Emperor Wen about the lavish customs of the day, and he referred in passing to the sale and treatment of young slaves that were probably of the entertainer type. These youths were dressed in "embroidered clothes and silken shoes with the edges all embellished." (16.) The King of Chi-pei explained his lavish payment for four girls by calling atten- tion to their skill at doing tricks (21). Wang Weng-hsii was just such an entertainer. Her training under Liu Chung-ch'ing started when she was only eight or nine years old. Later a merchant from Han- tan came looking for singers and dancers, and Liu Chung-ch'ing sold her in spite of her mother's claim (the father may have been cheating!) that the parents had never been paid a cent for her. The merchant kept her with a group of five entertainers and later probably sold the quintet to the repre- sentative of Emperor Wu's ill-starred Heir-apparent, who came from Ch'ang-an looking for singers and dancers for his patron's palace (55). Before he sold them the merchant presumably hired the girls 1 See Chavannes, op. cit., plates, pt. 1, text, t. 1: pi. XXVIII, No. 49, text, p. 86; pi. XLIX, No. 104, pp. 184-185; pi. LIX, No. 122, p. 201; pi. LXXVIII, No. 149, pp. 226-227; pi. LXXXV, No. 158, p. 232; pi. LXXXVI, No. 160, p. 232; pi. LXXXVII, No. 163, p. 233; pi. XCVII, No. 182, p. 247 (sic); and t. 1, pi. DXLI, No. 1270, p. 277. Cf. also Chao Pang-yen, "Han hua so chien yu-hsi [Sports as seen in Han drawings)," Academia Sinica, Studies presented to Ts'ai Yuan-p'ei on his sixty-fifth birthday, 2 vols., Peking, 1933-35, pt. 1, pp. 325-338. 2 Probably some of the mortuary figurines of musicians and dancers, dating from Han through T'ang times, represent slaves, who were used as entertainers in those epochs. 182 SLAVERY IN THE FORMER HAN DYNASTY out for feasts and parties. Emperor Hsiian's great-grandmother was that chorus singer Wei Tzu-fu, who infatuated young Emperor Wu so that he had no eyes for any of the girls of good family who had been assembled and especially groomed for his inspection (27). Em- peror Ch'eng's second Empress, whose sobriquet was "Flying Swallow," was once a slave dancing-girl in the household of the Princess of Yang-a, where the Emperor first saw and fell in love with her (101). His passion for her was so overwhelming that he deposed the Empress nee Hsii, and made "Flying Swallow" his consort. The "serving one" of the Queen of Heng-shan was a fine dancer (37); the infamous King of Kuang-ch'uan made his singing girls and entertainers play about naked among the guests at banquets {6J^) . When the King of Chao was ill he made a will commanding that those of his male and female slaves who could make music should follow him in death, and sixteen were compelled to commit suicide (84.). In contrast to this dark incident there is a pleasant passage in a letter from Yang Ytin to a friend describing the pleasures of rural retirement (87) : "My home was originally in Ch'in, so I can play Ch'in music; my wife is a girl from Chao, quite good on the drum and lute; and there are several male and female slave singers. After the wine has begun to warm my ears, I gaze up to heaven and beat my pottery drum and sing wu, wu." Girls trained as entertainers were obviously chosen for their natural grace and beauty, and probably received instruction in seductive arts. Numerous references to sexual relations between slave girls and their masters, or to illicit relations with other men, indicate another important "service" function of private female slaves. The term yil, used in combination with shih or pei, appears to indicate a sexual relationship with the master which should be reserved to him alone. For example, the scheming Queen of Heng- shan tried to create a schism between the King and one of his sons (the child of another woman) by inducing the boy to have relations with her own "serving one," already favored by the King. Later the boy was tried and executed for having had relations with the King's "personal female slave." Obviously his relations were incestuous (37, and footnote 4). In another case a son who had relations with his late father's "personal female slave" was tried, committed suicide, and his marquisate was abolished (48). Shang- kuan An had incestuous relations with his stepmother and various of his father's ladies and "serving personal ones." (59.) Again, a king allowed his favorite male slaves to have relations with his SERVICE FUNCTIONS OF PRIVATE SLAVES 183 palace ladies and "personal female slaves." For this and other abnormal acts his kingdom was reduced by four prefectures (85). A number of other cases indicate that masters had sexual rights to their slave women, though these rights were often violated by- other men.i The relationship between concubinage and slavery stands out clearly. Concubinage is often a criterion of a domestic type of slavery. As shown in the discussion of status, the crucial factors in the imperial household were winning the Emperor's "favors," bearing children, and, above all, bearing a son. In the palace, change in rank was complicated by the bitter struggles for power between various consort families. Such struggles undoubtedly occurred also in private families, but much less was at stake. By analogy with cases in the imperial household where considerable details are known, it would seem that slave women not infrequently became concubines "legally," or were concubines in fact. Private slaves sometimes served as guards for the tombs of their dead masters. This service was naturally confined to rich families with extensive burial grounds. Only two documents report it, yet rather casually. After Ho Kuang's death, his widow elaborately expanded and improved the buildings and grounds of the cemetery and secluded there "Sweet Ladies," slave women, and concubines to guard it (72). When Shang-kuan Chieh and his son An were executed, the child Empress nee Shang-kuan sent her private male and female slaves to guard the graves of her father and grandfather (62). Clearly, the slaves were domiciled in the cemeteries, which were isolated regions well outside the city, and they were probably self-sustaining groups that stayed at their posts without constraint. Their duties included guarding and tending the grounds, grave mounds, and temples, and performing periodic ceremonies to the dead; probably these duties were hereditary as long as the family continued important or its memory was strong enough to protect the tomb.2 1 23, 26, 29, 98, 108, 112. A16 reports the case of a marquis who was deposed for daring to have adulterous relations with a female slave in the presence of his wife, the Princess of Yang-i, and for drunkenly cursing her. It must have been his effrontery to the daughter of an Emperor which caused him to be deposed, rather than the fact that he had relations with another woman. * Judging by analogy with documents 55, 58, and 62, see also HFHD, vol. I, p. 140. Speculating further, but without specific proof, it seems likely that such grave-watchers eventually became either free people or hereditary families of nominal slaves or serfs bound only in a special legal or customary way to the memory of those masters whom their ancestors had once served. So long as the descendants continued to care for the graves, their right to use land about the graves would probably have been validated by law or customary usage. 184 SLAVERY IN THE FORMER HAN DYNASTY At least one master used his trusted male slave to manage his financial interests. Ho Kuang loved and favored his "supervising male slave," Feng Tzu-tou, and always consulted with him on business {72). Yen Shih-ku explains the term as meaning a slave who supervises and is familiar with the household business. Feng Tzu-tou's responsibilities must have been great, for, as already shown, Ho Kuang received annually the tax income from 20,000 households, and during his lifetime acquired enormous gifts {70), many of which he probably sold. Feng Tzu-tou was no ordinary slave. Not only did Ho Kuang always consult with him on business, but also he was recognized by the officialdom as being the key for access to Ho Kuang in political matters. "All the officials and lesser people only served Feng Tzu-tou and Wang Tzu-fang, and regarded the Chancellor as less important." {66.) Although legally a slave, Feng Tzu-tou was virtually free. After Ho Kuang's death he had relations with the widow, and is described as often breaking the law. In the annals of Emperor Hsiian he is referred to as a com- moner from Ch'ang-an and listed as one of the important rebels against the throne in &Q B.C. {72, and footnote 3). One of the few slaves known by his full name, he even had a courtesy name, which might be translated "Adonis." ^ Treatment and Position of Domestic Slaves Several observations arise from a study of documents that mention domestic slaves. They were in a certain sense members of the family, well clothed, fed, and housed, and many of them showed a strong personal loyalty towards their masters. Slaves that carried 1 Two passages in the Hou Han shu mention supervising slaves and contain important supplementary information. The first is dated about a.d. 150, and concerns a slave of Liang Chi, the powerful affinal relative of the imperial house who established two emperors on the throne and from whose family came seven marquises, three empresses, six imperial concubines, two generals-in-chief, three princesses, and fifty-seven various officials. Liang Chi's "beloved" supervising male slave, Ch'ing Kung, rose in office to the position of Superintendent of Public Granaries. He also had relations with his master's wife. His power was so great that censors and commandery officials all came to visit and consult with him (HHS, 64, 5b). The second instance is that of the supervising male slave of the eunuch Chang Jang, an influential official and marquis during the reign of Emperor Ling (a.d. 168-188). This supervising slave was entrusted with and controlled his master^s household affairs, and was intermediary for bribes in goods and money. His power and authority were well known. Meng T'o, an aspirant for office, once succeeded in getting to the master by becoming a close friend of the slave and getting him in debt. Thus, when Meng T'o wanted an interview with the eunuch Chang Jang, the supervising slave welcomed him ostentatiously and escorted him into the gate past a crowd of other aspirants who were waiting outside. The other aspirants immediately assumed that Meng T'o was a close friend of the eunuch SERVICE FUNCTIONS OF PRIVATE SLAVES 185 weapons and rode horseback must already have achieved, in fact, a semi-free status. When generations of slaves served in the same household, their relations to the family must have been very close. Young Liu Ho, King of Ch'ang-i, was intimate and companionable with his male slaves and servants (67). This is represented as an objectionable quirk in his character, but later, when he had been deprived of his kingdom, and lived, half-demented, in the seclusion of his palace, his slaves took care of him (75). The slaves of Tung Hsien were treated like members of his family when it came to gifts; officials and Emperor Ai both gave generous amounts of money and valuable presents to them all {117, 119). Ho Kuang's slaves identified themselves closely with their master's importance. When he was at the height of his power they were arrogant and lawless {65), and after his death, when the power of the family was challenged by the Emperor through Grandee Secretary Wei Hsiang, they under- took to humiliate Wei Hsiang — with disastrous results {72). The gossiping of a group of slaves about the rebellion being plotted by the Ho family was overheard by a man trying to sleep in the stable, and was revealed by him (7^). Even the grooms knew of the plot before it was fully matured ! The fact that the famous redresser of wrongs, Ylian She, was held directly responsible for the conduct of his slave also indicates a legal conception of slaves as family members. His slave quarreled with the butcher, chopped him very badly, and then ran away. When the slave did not show up by the next day the officials were about to execute Yiian She on the spot. His friends intervened in the nick of time, and arranged a face-saving device for the magistrate in which Yiian She took on the outer appearance of a slave or criminal, apologized humbly, and was absolved {125). These are but a few specific cases. More general indications of standing in the family are the fact that certain female slaves were counted in the incest group, and the use of the term k'o, "guest" or "retainer," in combination with "male slave" or "youth." ^ and therefore gave him bribes. He divided the bribes with the eunuch, who was very pleased and got him appointed Censor of Liang Chou (HHS, 108, 9b-10a). In both these passages the supervising slaves were similar to Feng Tzu-tou in that they occupied positions of confidence with their influential masters, and were accorded great respect by officials or aspirants. One rose to high office, while the other was in charge of his master's business afifairs and acted as inter- mediary in matters involving bribes. 1 For passages using "slave" and "guest" together see 28, 37, 61, 65, lOU, 118. It is not clear whether nu-k'o and Vung-k'o should be translated as compounds ("male slave-guest," "youth-guest") or as separate terms. The institution of "guests" or "retainers" was well established in Chan-kuo times. "Guests" 186 SLAVERY IN THE FORMER HAN DYNASTY "Guests" were not part of the blood family but were closely asso- ciated economically and socially. On the other hand, slaves had their own surnames, which differed from those of their masters. Here Chinese familism showed its great strength, over-riding economic considerations. Details about the living conditions of domestic slaves are hap- hazard, but aside from several specific cases of ill-treatment, the accounts indicate that they were physically well cared for. This would be especially true of slaves in noble or wealthy households, who probably fared better than most plebeians. Tung Hsien's slaves ate so well that they looked upon wine as though it were soup and meat as though it were beans {118); and two of Wei Ch'ing's squires, guests with him in the home of a princess, were invited to eat with cavalry slaves of the family {39). The squires thought this was a disgrace, but apparently the hostess considered her slaves' food quite good enough for poor men of the Gentleman class. It was fashionable to dress slaves in fancy silks {16, 72, 105), and a memorial praising Wang Mang's frugality and modesty signalized the fact that his slaves wore coarse clothes, as though such costume were unusual in his day {A19). We read twice about the famous court physician Shun-yii Yi examining slaves for their masters {21, 22). On the other hand, some private slaves were apparently branded, though there may have been a Former Han law against it {138). The personal loyalty of slaves to their masters was marked. The male slaves who galloped with Kuan Fu and his two companions into hopeless battle were all killed after fighting valiantly and taking a heavy toll of the enemy {25). When Emperor Kao uncovered a plot on his life he ordered the King of Chao and his rebellious ministers to be arrested and brought to the capital, and warned that anyone in Chao daring to follow the King to Ch'ang-an would be punished by death and the extermination of his three sets of relatives. T'ien Shu, Meng Shu, and some eight others shaved their heads, put on russet clothes and iron collars, and thus disguised as slaves could follow the King to prison {12). Several versions of this story stress the great loyalty of the King's retainers. The interesting thing is that it was considered natural for slaves to follow their masters to prison. attached themselves as advisers, and usually as loyal followers, to important political and military men. Translating the terms separately, in accordance with their separate historical backgrounds, should not obscure the fact that during Han times "attendant slaves" and "guests" had a functional similarity. On "guests," cf. T'ao Hsi-sheng, "Hsi Han ti k'o ['Guests' in the Western Han period]," Shih Huo, vol. 5, No. 1, Jan. 1, 1937, pp. 1-6. W 2 ^ > J f Z •" -S ^ S t B :S 2 - E E ^ Z 3 SERVICE FUNCTIONS OF PRIVATE SLAVES 187 The loyalty of female slaves to their mistresses sometimes made them dangerous to rival ladies. The King of Kuang-ch'uan had two concubines nee Wang, both of whom he promised to make his Queen. Another woman, nee Yang-ch'eng, won his affection, and the two disappointed candidates planned to kill her. When the King discovered this he assembled all his ladies, and in their presence he and his favorite slowly hacked the Whangs to death. Lady Yang-ch'eng then pointed out to the King that the slave women of the two concubines might let the matter leak out. The Wangs' three attendant female slaves were therefore strangled (ff^). Probably the three "attendant" female slaves had come from the Wang family into the King's palace as personal servants for their mistresses. For that reason the King feared that they alone of all the women would reveal the murder.^ Later Yang-ch'eng Chao-hsin, wanting to monopolize the King's affection, persuaded him to use her elder slave woman as gate guard over the Concubines' Quarter, to prevent the other ladies from "licentious promiscuity." Appar- ently she trusted her elder slave women to protect her interests! Loyalty of slaves to their masters was not, however, an absolute quality; in two cases masters killed slaves to prevent them from letting the facts of a crime leak out {106, 129), and, as mentioned before, the imperial concubine nee Chao tried to bribe three of her private female slaves not to reveal her mountainous sins. She had good reason to suspect their loyalty. Formerly household slaves of the rival Wang and Hsii families, and freed by them, the three women had simply been commanded to enter the palace and become her private slaves {107). All these occupations of private slaves — general servants' work, armed escort, entertaining, personal attendance and sexual relations, tomb-watching, and managerial work — were "non-productive" in the sense that they added nothing to the master's wealth. They were luxury occupations, promoting the owner's comfort, pleasure, or prestige. Without these aids to gracious living, noble and wealthy people probably told one another they "simply could not endure." Slaves as Instruments of Power There was yet another "service" function of private slaves, and it has a sinister aspect. This was the use of slaves as instruments of ^ Every upper-class woman may have taken with her a few family slaves as personal servants when she was married. Quite probably the "private" male and female slaves of the child Empress nee Shang-kuan were simply her family slaves who went into the palace to wait on her and protect her from rivals {62). 188 SLAVERY IN THE FORMER HAN DYNASTY power. Because this subject has been generally neglected by- previous writers it merits careful analysis. It demonstrates more clearly than any of the previously mentioned service occupations the relationship between private slavery and the societal structure of the Former Han period. Nobles and landed gentry were the actual rulers of their local areas whether political control was vested in them or not. They forced their will upon the common people through political connec- tion and economic power; they appropriated land by illegal or shifty means, laid tribute, and extorted extra corvee labor for their personal ends. Social philosophers inveighed against these practices over and over again. Part of the phenomenon was the use of armed male slaves — an inevitable development, since the principal duty of a special class of male slaves was to bear arms and ride escort with their masters. Trained as fighters, they were personally loyal to one man, obeying his commands alone. Cases in which masters used their slave retainers to enforce their will on others and to obstruct justice are so common and so flagrant as to indicate that private slavery had a special function in maintaining and extending the power of the slave-owning class. Slaves were the principal or accessory agents in several cases of terrorism. The most notorious concerned a cousin of Emperor Wu, Liu P'eng-li, King of Chi-tung. In the years before 116 B.C. the King made a practice of going out at dusk with several tens of his male slaves and lawless young bloods, plundering and assassinating people "all for sport." More than a hundred cases of people killed were actually divulged. No one dared travel at night in the kingdom. Finally the sons of people killed sent a statement to the imperial court reporting the facts, and the officials begged the Emperor to have the King executed. Emperor Wu "could not bear" to do that; instead he deposed Liu P'eng-li, made him a commoner, and banished him to northwestern Hupei (^4^). Seven other judicial cases reveal that slaves murdered people at their masters' command. In 110 B.C. Liu Lung, the Marquis of Fu, was tried for causing a male slave to kill a man. The marquis- ate was abolished and the marquis was thrown into prison, where he wasted and died (51). When Liu Ch'ii, the infamous King of Kuang-ch'uan, was a boy he had a teacher who criticized and disciplined him. When Liu Ch'ii grew older he drove the teacher out, but some time after- SERVICE FUNCTIONS OF PRIVATE SLAVES 189 wards the latter was appointed an official of the kingdom, and several times got restrictions placed upon the King's household. Liu Ch'ii ordered a male slave to kill the official and his son. Later two of his concubines, the sisters nee T'ao, were killed. Completely destroying the corpse of one girl, Liu Ch'ii delivered to the mother a substitute corpse with the body of the sister. The mother was not deceived, however. She "howled and lamented," demanding the real corpse of her daughter. Queen Yang-ch'eng Chao-hsin ordered a male slave to murder her. The slave was caught, and his confession started an investigation which revealed sixteen murders by the King and his jealous Queen. Though the judicial commission recom- mended public execution. Emperor Wu demanded some other punishment. The King was deposed and banished with his wife to northern Hupei, where he was promised a small estate. On the way he committed suicide and only then was the Queen publicly executed. The kingdom was abolished {6Ii). Chi Tang, Baron of Chou-tzu-nan, was tried in 67 B.C. for causing a male slave to kill his household manager. He was publicly executed and the barony abolished {73). Among the accusations against Chancellor Wang Shang, Marquis of Lo-ch'ang, was the suspicion that he had had a male slave murder his sister's lover. Because Wang Shang was a cousin of Emperor Hslian his crimes were pardoned but he was removed from office and noble position (25 B.C.). Three days later he "fell sick, spat blood, and died." {98.) Hsiao Huo, a seventh generation descendant of Hsiao Ho, was tried in 16 B.C. because he had ordered a male slave to kill a man. The marquisate was abolished, but his sentence was reduced from execution to "guarding the frontier." {103.) A male slave owned by the King of Liang assassinated two high officials on his master's order because they had enraged the King on matters of public business. Then the King killed the slave to silence him. For these and other crimes the judicial officials begged for his execution, but Emperor Ch'eng only reduced his kingdom by five prefectures {106). The Marquis of Wu-an was deposed in 3 B.C. for having caused a male slave to kill a man {115). Eight cases of slaves employed to murder during a period of more than two centuries may appear insignificant. On the contrary, they probably exemplify the importance of slaves as instruments of power. This does not mean that the original archivists or the authors 190 SLAVERY IN THE FORMER HAN DYNASTY of the Shih chi and Ch'ien Han shu recorded the cases to illustrate that phenomenon. From their point of view the fact that slaves were involved was purely incidental; nothing is said of what happened to the actual murderers, the slaves. The important fact, naturally, was that noblemen were tried, with the result that five political units were abolished, a kingdom reduced, a marquis deposed, and a chancellor dismissed. Personal punishments included execution, reduction of sentence from execution to guarding the frontier, banishment, and imprisonment resulting in death. Each case concerned (1) a capital crime, which (2) involved a nobleman, who (3) was actually brought to justice. They were all matters of imperial and historical concern. It is, therefore, legitimate to make three corresponding assumptions: (1) Less serious illegal uses of slaves happened frequently, but were not recorded because the consequences were not important matters of state. (2) Similar capital cases occurred, involving slave-owning commoners who were tried, but the cases were not of enough historical impor- tance to be reported. (3) Other cases of murder by slaves at the behest of their noble masters probably occurred but are not recorded because they were not revealed, were suppressed, were not the principal or most serious charges against the nobleman, or were simply forgotten. It cannot be assumed that every such murder was revealed, tried, punished, recorded, and finally included in the histories. In all probability the eight recorded cases were a mere fraction of the total. If this reasoning is correct, it is justifiable to assume that they typify a general phenomenon. A few more historical tidbits fall into the above category. Luan Pu was purchased for the specific purpose of doing a deed of revenge for his master (5), but only because he later became famous do we know even that much about his enslavement. The King of Huai- nan used a government slave (or slaves) to forge imperial seals and all sorts of official insignia in preparation for rebellion (A8). Private slave women were twice ordered to invoke magic curses against emperors in the hope of causing death. In each instance a king and his queen conspired in the witchcraft, and it happens that both kings committed suicide and both queens were executed (38, 113). Another queen was executed for having ordered a female slave to kill her predecessor by magic poison {37). The last four items concern kings and queens who were actually brought to justice for using slaves in the crimes of rebellion and murder; they fit the formula established for the eight cases just cited. SERVICE FUNCTIONS OF PRIVATE SLAVES 191 Two cases where masters used their slaves to enforce their will against common people happen to be recorded. When Liu Ho, the King of Ch'ang-i, was rushing to the capital to be made Emperor in 74 B.C. he had his senior male slave kidnap girls and load them into a screened carriage to take along with him. Since he was quite close to the capital an official sent out to meet him was able to have the slave arrested and turned over to the guard for punishment in spite of Liu Ho's objections. There is no indication that the girls were set free, but the incident was one of the grounds for deposing Liu Ho after a brief period as Emperor (68, and foot- note 3). Chang Fang was Emperor Ch'eng's cousin and catamite; his wife was the Empress' niece. The Emperor called him his son, and gave him a mansion, chariots, clothes, and jewelry; when he was married the Empress presented him with girls, probably slaves, from her private bureau. He accompanied the Emperor on his famous incognito journeys, and was a cock-fighting and horse-racing "fan," generally arrogant, self-indulgent, prodigal, and debauched. Finally, the Chancellor and Grandee Secretary memorialized the Emperor, revealing some of the crimes of Chang Fang and insisting that he be sent away from the capital. One complaint was that when he learned that a certain commoner was planning to give his daughter to the imperial seraglio, he sent the Inspector of Musicians in the Bureau of Music to get her (for himself) by force. Unsuccessful in this, Chang Fang sent his male slaves to the commoner's home. They broke in and wounded three people (100). This happened in the capital and was perpetrated by the Emperor's personal favorite. Slaves were used not only to enforce the master's will against other people, but also to oppose the government and obstruct justice in cases where proper judicial procedure endangered the owner or his interests. The acts of Princess Kai, the older sister of Emperor Chao, are a clear example. Her lover, a commoner named Ting "Wai-jen," hated the former Prefect of the Capital and sent a "guest" (k'o) to slay him. Because the retainer hid in the Princess' house the officials did not dare arrest him. Hu Chien, Prefect of the city where she lived, boldly surrounded the house with his officers and soldiers, placing it under guard. When the Princess learned about it she assembled her lover and her friend General Shang- kuan An, father of the Empress, and the three galloped up, accom- panied by a large number of male slaves and guests who shot at the officials and chased them off. There was some fighting, for one 192 SLAVERY IN THE FORMER HAN DYNASTY of the Princess's slaves was wounded. This enraged her. She sent her chief gate guard to impeach Hu Chien's subordinate, the Police Chief of the district, for wounding a slave of her household. Hu Chien protected his subordinate by reporting back that he was not guilty. Furious at this, the Princess submitted a letter to court, accusing Hu Chien of having invaded and insulted a Grand Princess, and of shielding an officer who he knew had outrageously wounded her slave. Ho Kuang was able to table her memorial, but later fell ill. Thus co-regent Shang-kuan Chieh, the father of Shang-kuan An, had a chance to take over the case. He sent an official to arrest Hu Chien, who committed suicide. This was recognized by every- one as a great injustice (61). A similar case concerned the same Chang Fang who tried to kidnap the commoner's daughter. Officials ordered to his house to arrest a known thief were locked out and shot at with military crossbows by Chang Fang's slaves under his orders. Again, because he was angered at the Police Chief of the Bureau of Music about some governmental affair, he sent his senior male slave and a gang of forty or more, all armed with military crossbows, to invade the bureau in broad daylight. They shot up the offices, wrecked the furnishings, tied up the children of the chief official, and would not stop until the police chief and his assistants, dressed in the costume of convicts, apologized to Chang Fang personally (100). If some of the flower of the nobility used their slaves to terrorize officials in the capital one may imagine how they lorded it over admin- istrators in the provinces. After Wang Lin-ch'ing had been dismissed from court on the death of his sponsor, one of the Empresses nee Wang, he returned to his native seat, where he and his ruffian associ- ates continued to flout the law. He was advised officially to move on to the next prefecture, but as a last defiant gesture sent back a slave to beat on the assembly drum standing before the prefectural offices. Ho Ping, who was Prefect of the district, set out with officers to arrest Wang Lin-ch'ing, and toward sundown was hot on the trail. In these straits, Wang Lin-ch'ing made one of his cavalry slaves exchange clothes with him and take his place in the carriage, while he galloped off in disguise and escaped. When Ho Ping over- took the party he realized he had lost his man; but he pretended that the slave was the master and had him beheaded on the spot. The slave's head was publicly exhibited with a label proclaiming it to be Wang Lin-ch'ing. Thus the master lost his official identity and the populace believed him dead. When the facts were learned SERVICE FUNCTIONS OF PRIVATE SLAVES 193 at court another Dowager Empress, grieving for a relative of the late Empress Dowager nee Wang, reported to Emperor Ai. Far from being angry, the Emperor upheld Ho Ping and advanced him in position {112, and footnotes). The Ho family slaves and guests living at P'ing-yang, the family's native seat, would carry weapons and go into town fighting and brawling, and the officials could not prevent them. It is specially noted to the credit of Yin Weng-kuei that when he became an official in the town no one — not even the Ho family slaves and guests — dared to break the laws {65). Both these incidents were recorded to illustrate the character and administration of particularly fearless and aggressive officials. This is both a backhand commentary on the average local official, and an unconscious indication of the way powerful individuals probably used their slaves to dominate local areas. Dangers inherent in the lawless use of slaves did not escape the notice of a few high officials in the government, but there is no evidence that they were concerned with the matter as a general phenomenon. 1 Specific crimes were specifically punished, the master being charged with the crime committed by the slave under order. Two memorials show, however, that certain owners of numerous male slaves were considered a menace to the government. Chan- cellor Wang Shang, leader of the clan of Emperor Hsuan's mother, was accused of controlling the palace and embittering the whole empire by his ruthless acts. Supported by a powerful clan whose combined capital was reckoned by the hundred million cash and whose private male slaves were said to be numbered by the thousand, Wang Shang was accused of being a potential threat to the throne. The memorialist considered him much more dangerous than Chii Meng, who, as a lone wolf with many independent braves for fol- lowers, held the balance of power during the Rebellion of the Seven States {98). The other memorial criticized the political power and private conduct of Wang Ken, half-brother of Emperor Yiian's consort, and leader of the other Wang family from which Wang Mang sprang. Immediately after pointing out how he had infringed upon imperial prerogatives it disclosed that he made his male slave attendants wear armor, carry bows and crossbows, and deploy like infantry {111). 1 Unless we assume that one reason for repeated attempts to limit numbers of slaves individually owned was to prevent accumulation of too much terroristic power. There is no textual evidence for this assumption, and none of the restric- tions specifies a limit to the number of male slaves a family could own. 194 SLAVERY IN THE FORMER HAN DYNASTY The fear that private slaves of these powerful consort families might be employed against the imperial house was not imaginary. Throughout the Former Han period consort families frequently acquired so much power that they might have menaced the House of Liu. Between the never-forgotten attempted revolution by the Lii family near the beginning of the dynasty and the successful one by Wang Mang at the end, the clan of nearly every Empress became so dominant that it had to be quelled. If open revolt had broken out, instead of mere political intrigue, private male slaves, loyal and armed, could have been used as shock troops in the fighting. IX. PRODUCTIVE EMPLOYMENT OF PRIVATE SLAVES There is far more evidence that private slaves were used in service capacities, including maintenance of power, than as producers of wealth. How much "productive" work, therefore, did slaves do? What kind of work was it? How important was slave labor in the total process of production during Han times? In a country with an agrarian economy such as China had during the Former Han period, the really basic problem in regard to produc- tive employment of slaves is their use in agriculture. Most slave owners, and probably all large slave holders, also owned agricultural land. Indeed, land was primary; slaves were secondary. There is no reason to doubt that masters who directed their tenants or managed their own farms used some of their male slaves to work during those crucial periods of planting, cultivating, and especially harvesting when all hands were mustered to help. Some masters, likewise, may have kept small groups of unfree cultivators for the regular farm work, and used hired labor at crucial periods. The fundamental question is not, however, whether slaves worked on farms, but whether slave labor was an important form of agricultural labor during the Former Han period. If there is any indication that it may have been important, then the problem is to compare slave labor with other types of agricultural labor existing during the epoch. If slave labor was an important part of agricultural labor, then slavery was correspondingly important to Chinese economy in general. The subject is at the very center of the controversy whether Chinese society was a "slavery society" and whether the economy of Han times was a "slavery economy." It is therefore essential to subject to the sharpest scrutiny those texts which suggest a relationship between slaves and farming. Evidence and Presumption of Slaves in Agriculture There is an initial enigma: only two documents speak definitely of slaves working on farms, and in each only a single slave is mentioned. Chu Chia, the buyer and protector of Chi Pu, hid him in a house in the fields, and instructed his son to be lenient with the slave in the field work and to share his own food with him (8).^ 1 Chu Chia's instructions to his son are not in the CHS, but come from the SC. There is a possibiUty that the SC text is secondary, which would lessen its importance in regard to agriculture. 195 196 SLAVERY IN THE FORMER HAN DYNASTY Chu Chia's son may have been manager of the farm. Since Chi Pu was in disguise as a slave it must have been quite natural for him to be working as a farm hand. Obviously there were other workers in the same fields, but not necessarily slaves, for Chu Chia had innumerable hired (or indentured) laborers, but was not noted as a slave owner. ^ In 59 B.C., Wang Pao wrote into his purchase contract the specifications for agricultural work to be done by his slave. Pien- liao was to hoe the garden, and plant ginger, melons, eggplant, and onions. At the time of the vernal equinox he was to repair the ditches, dikes, and boundary walls of the fields, fertilize the land and prepare it for planting; in the fourth month to transplant; in the ninth to harvest; and in the tenth to gather in the beans. He also had to prune mulberry trees (used for production of silk), and tend an extensive orchard of peach, plum, pear and persimmon trees (83). This interesting document gives insight into work on a farm managed by the owner where the products were for sale and home consumption. But it was certainly not a plantation, for Pien-liao appears to have been the only male slave owned by Wang Pao. There are two groups of documents which, by mentioning fields and slaves in close proximity, suggest a functional connection; i.e. that some of the slaves worked in the fields. The first group reports wealth of certain individuals, or gifts to them, in terms of fields and slaves. The case of Ho Ch'ii-ping, who liberally bought fields, houses, and slaves for his father, merits attention. It seems obvious that the slaves were producers of wealth, and not mere consumers, because the father was a humble man who certainly could not support luxury slaves (42). Likewise, the daughter of Cho Wang- sun was given a large sum of money and a hundred "youths" by her father. She immediately invested in farm land and houses and became a rich person (28). Is it justifiable to assume that the slaves in these and certain similar cases- were agricultural slaves? It may be; and yet the assumption is an arbitrary one. There are similar documents which report wealth in terms of slaves and other goods without mentioning farm land, and furthermore these other types of wealth, such as mansions, carriages, horses, furniture, silks, gold, 1 See pp. 122-123 and footnote. 2 Emperor Wu's gift of 1,000,000 cash, 300 slaves, 100 ch'ing of fields, and a mansion to his half-sister (31), and the reward of a princess' fields, houses, and slaves granted to the Chinese lady who had been sent as bride to a Wu-sun king, when she returned to China (88). PRODUCTIVE USE OF PRIVATE SLAVES 197 and money, have no functional connection with land.^ Thus it may be an error to pick out farm land and slaves and assume that they belong together. They may merely have been two of the standard indices of wealth. The second group of documents includes ministerial protests against extensive owning of fields and slaves, and imperial edicts or laws restricting such ownership. This group merits close study since the documents appear more definitely to imply a functional connection.^ Tung Chung-shu, the outstanding social reformer of the Han period before Wang Mang, was deeply concerned with economic problems besetting the empire during the early years of Emperor Wu's reign. Especially was he the champion of the common people, the great peasant group whose toil he considered the real basis of national wealth. In a speech to Emperor Wu he complained about the way in which nobles and high officials, enjoying big incomes, took advantage of their wealth and position to compete for profits with the common people who could not possibly match them, "Therefore," said Tung Chung-shu, "[the influential people] multiply their male and female slaves, increase their [herds of] cattle and sheep, enlarge their fields and houses, broaden their fixed property, and accumulate goods. Busily engaged in these pursuits without end, they thereby oppress and trample on the common people. The com- mon people are daily pared down and monthly squeezed, gradually becoming greatly impoverished," He urged the Emperor to prevent households that received official income from competing with the people (3^). Does this imply that slaves were an instrument of this competi- tion? Were they used in the fields to produce crops sold in compe- > As in the cases of Ch'en P'ing (15), Luan Ta U9), the grandaunt of Emperor Chao (58), Ho Kuang (70), Shih Tan (97), Wang Mang's uncles (99), and the family of his second Empress (133). "^ Documents of the kind to be discussed have been quoted by Chinese writers as proof or strong presumption that the slaves were used in the fields, but they have been cited without careful analysis, so far as I have read. Tai Chen-hui ("Liang Han nu-li chih-tu [The slavery system of the two Han dynasties)," Shih Huo, vol. 1, No. 7, March 1, 1935, pp. 286-291 [hereafter cited by translated title], see p. 287) uses the association as definite proof. Wu Po-lun ("An investiga- tion of slavery in the Western Han," p. 282) says that the association suggests that slaves were used in agriculture, though he believes that the system of share- crop farming made slaves unimportant. Ma Ch'eng-feng (An economic history of China, vol. 2, pp. 269-271) quotes the passages I am citing and similar HHS ones as proof, though he concludes that in the nation's agriculture, slave labor was not important. Presumably Liang Ch'i-ch'ao ("System of slavery in China," p. 552) had the same sources in mind in his discussion of large-scale landowning and use of slaves in agriculture. 198 SLAVERY IN THE FORMER HAN DYNASTY tition with those of the small farmer? This possibility cannot be ignored, but the development of ideas in the memorial implies that the successful business enterprises of people enjoying government income and personal power as officials or noblemen resulted in the accumulation of slaves, herds, farms, real estate, and other kinds of wealth. In another memorial, closely similar, Tung Chung-shu also mentioned both fields and slaves, but treated them as distinct problems. He urged the Emperor to limit the amount of private farm land any individual could own, "in order to assist those in want." After suggesting two other reforms, the second of which had no connection with land, he urged the abolition of slavery, in order to eliminate the fear of autocratic execution. Then he recommended reduction in taxation and corvee duties (35). There is certainly no suggestion here that slaves were employed in agriculture or that their employment created an economic problem. In 119 B.C. other ministers urged Emperor Wu to forbid resident merchants to own fields. Their recommendation is phrased as though it were an administrative order: "[If anyone] dares to violate this order, confiscate his fields and youths." U5.) Why should mer- chants illegally owning fields have their fields and "youths" confis- cated? Does this mean that the "youths" (probably slaves) worked in the fields? The Ch'ien Han shu variant of the passage uses the term "goods" in place of the Shih chi "youths," thus creating doubt as to which reading is correct. It gives the impression that merchants would have their illegally owned fields confiscated, and as punish- ment lose their other wealth as well, and makes no hint of a functional connection between fields and slaves. Incidentally, during the next seven years when the government confiscated large amounts of farm land and thousands of slaves it did not use them on this land, as it presumably would have done had they been agricultural slaves (^6). The next protest against extensive ownership of fields and slaves, a protest of the same textual order as earlier and later ones, dates a century later. Emperor Ch'eng was censured by his high officials for privately owning both. But the land and slaves were not con- nected. The fields were scattered among those of the people while the private male slaves lived with the Emperor in the Northern Palace to be his companions on his incognito journeys {10J^). The famous attempt by ministers of Emperor Ai to set limits to ownership of land and slaves, reported both in the "Treatise on Economics" and in the "Imperial Annals," is exceedingly suggestive. When Emperor Ai was given the throne in 7 B.C., his former tutor, PRODUCTIVE USE OF PRIVATE SLAVES 199 Shih Tan, delivered a lecture on economic conditions. He recalled the ancient ching-t'ien system of land division, and the fact that early in the Han period Emperor Wen had emphasized agriculture and sericulture, and set an example of thrift. Therefore, said Shih Tan, the people had an abundance, and the accumulation of land, offices, and other sources of income did not occur. Wherefore Emperor Wen "did not set limits on people's fields and male and female slaves." Shih Tan emphasized that now, in contrast to conditions in that day, the overbearing officials and rich plebeians had riches counted by the hundred millions while the poor and weak were increasingly distressed. He recommended that general restric- tions be ordered (109). The Emperor then discussed the problem in an edict i^ Now the vassal kings, the ranking marquises, the princesses, the officials [receiving a salary of] two thousand piculs, and the overbearing, rich people herd many male and female slaves, and fields and residences without limit. They compete with the common people for profit, and the people have lost their occupations, and are heavily distressed and in want. Let limitations and regulations be discussed [110]. Shih Tan's statement that it was unnecessary for Emperor Wen to set limits on fields and slaves, and Emperor Ai's complaint that nobles, high officials, and rich plebeians accumulated slaves and fields and competed with the commoners for profit, would both seem to be prima-facie evidence that fields and slaves were considered in 7 B.C. to be instruments of competition, and that accumulation of them by the powerful minority created an unfavorable economic situation. Chancellor K'ung Kuang and Grandee Secretary Ho Wu proposed the following restrictions: Vassal kings and marquises should be allowed to own fields in their own states; it is not clear whether they could possess unlimited amounts or were restricted to thirty ch'ing (about 340 acres). Marquises who lived at the capital, and princesses likewise, should be limited to thirty ch'ing in prefectures and border marches outside their states. Kuan-nei marquises, who alone had incomes from estates near the capital, and officials and plebeians should all be limited to thirty ch'ing. No merchant could own private fields or become an official. As to male and female slaves between the ages of ten and sixty, vassal kings were to be ' That is to say, the next step in the development of the law was an edict by the Emperor. We do not know who wrote the edict for young Emperor Ai, or exactly whose ideas he was expressing, though Shih Tan, K'ung Kuang, and Ho Wu were the moving spirits in the reform. 200 SLAVERY IN THE FORMER HAN DYNASTY allowed two hundred, marquises and princesses one hundred, and kuan-nei marquises, officials, and plebeians thirty. All private fields and slaves exceeding the areas or numbers allowed in each category- were to be confiscated after a three-year period (109, 110). In considering this law it is assumed to have been seriously proposed by officials who looked upon it as a necessary corrective for social and economic problems of the day. K'ung Kuang, Ho Wu, and other high officials charged with writing the terms of the ordinance, were thoroughly conversant with conditions of land- and slave-owning, and settled upon limits that they believed were equitable. Even though opposition from powerful affinal relatives of the Emperor and from his favorite Tung Hsien prevented the law from going into effect, the high ministers must have believed it could be put into operation within the three years allowed for liquidating excess holdings. It must be observed that the numbers of slaves allowed did not depend upon the amount of land owned, but upon the social position of the master. Kings, having the most elaborate palaces, holding court in imitation of the imperial pattern, and by 7 B.C. possessing the forms but not the power of rule, were considered to need no more than two hundred adult slaves of both sexes. Marquises, whether they lived at their estates or in the capital, needed a maxi- mum of one hundred, as did princesses. Thirty slaves sufficed for lesser ranks of nobility, non-noble officials, and all plebeians.^ 1 An even stronger case develops if, as Dubs states ("Wang Mang and his economic reforms," TP, vol. 35, 1940, p. 246), ". . . kings and marquises should be allowed unrestricted amounts of private land within their kingdoms and marquisates; outside those areas they should be restricted to 3,000 mou (140 acres)." According to this interpretation, instead of everyone being limited to 340 acres of private land, except merchants, who could have none, we find different amounts of land determined by the location of the fields, and by the dwelling place, rank, or occupation of the owner; while slave numbers depended upon the owner's status only. Rank or class Land Slaves Kings Unlimited in own state? 200 Marquises Unlimited in own state? 100 Marquises living at the capital 30 ch'ing 100 (plus unlimited in own state?) Princesses 30 ch'ing 100 Kuan-nei marquises 30 chHng 30 Non-noble officials 30 ch'ing 30 Plebeians 30 ch'ing 30 Merchants None 30 I place this interpretation, which I consider correct, in a footnote because the general argument does not depend upon which particular interpretation of the passage is accepted. In the other reading also, amounts of farm land and numbers of slaves are unrelated. PRODUCTIVE USE OF PRIVATE SLAVES 201 Restrictions on slave-owning seem to have been based, not upon economic considerations, but rather upon the precedent of general sumptuary laws already well established early in the Han period. Slave-owning was the subject of purely sumptuary legislation only six years before and ten years after 7 B.C. In 13 B.C. Emperor Ch'eng delivered an edict censuring the nobility and high officials for their lavish style of living: Some among them are extravagant and neglectful; busy with expanding their mansions and houses, setting out gardens and pools, accumulating numer- ous male and female slaves, dressing them in fancy silks, setting out bells and drums, providing