Beruflich Dokumente
Kultur Dokumente
A SOCIO-CULTURAL APPROACH
BY
ROBIN D. S. YATES
(McGill University)
Abstract
This essay analyzes the nature of slavery in early China from a comparative socio-
cultural perspective, using the sociological approach of Orlando Patterson (Slavery and
Social Death, 1982). Marxist and other theoretical positions are rejected in favor of
viewing slaves not as the object of property, but rather seeing that slaves could not be
the subject of property. In other words, slaves are “dominated non-persons.” The focus
of the inquiry is on interpreting the diverse materials relating to slaves in the Qin legal
documents discovered at Shuihudi, Hunan Province, in 1975. Brief consideration is
given to other statuses, such as the convict status of lichen and liqie (male and female
bondservants), and whether they should be considered slaves or not. The conclusion
emphasizes the importance of analyzing early Chinese slavery within its culturally rich
context of ritual and cosmological conceptions and practices.
Introduction1
Over the last two decades, slavery has become the topic of numerous
studies to such an extent that one might say that slavery has become
1
An earlier draft of this essay was presented at the Conference on Ancient China
and Social Science Generalizations, Airlie House, Virginia, 21–27 June, 1986. I would
like to thank Perry Anderson and Lin Ganquan for their valuable comments on that
draft. A Chinese translation of an earlier version of this article, “Gudai Zhongguo
nulizhi bijiao lishi yanjiu 古代中國奴隸制比較歷史研究 ,” was published in Zhongguoshi
yanjiu 中國史研究 111 (1986), 21–34. This essay is dedicated to my late teacher and
mentor Professor Kwang-chih Chang, who throughout my career, even in his illness,
unfailingly gave me wise advice and generous support. I am also very thankful for the
valuable suggestions and corrections by Lothar von Falkenhausen, Robert E. Murow-
chick, and David Cohen. David Cohen also solved many software problems at the final
stages of the preparation of this essay. To him I am much indebted. I would also like to
express my gratitude to the Killam Foundation and the Canada Council for the Arts,
the Social Sciences and Humanities Research Council of Canada, and Fonds pour la
© Brill, Leiden 2002 JEAA 3, 1–2
284 ROBIN D . S . YATES
an academic industry in its own right (cf. Karras 1991; Miller 1985;
Parish 1989). The reasons for this interest in slavery stem from, on the
one hand, the work of the so-called cliometricians, led by Fogel and
Engerman (1974), who developed new historical and statistical tech-
niques to reevaluate the nature and economics of the slave system from
the sixteenth to the nineteenth centuries in the area that became the
southern United States, and on the other, from studies of ancient slavery
by historians of Classical Greece and Rome (Finley 1968, 1980, 1983).
In addition, Western Marxist scholars attempted to develop the scattered
and rudimentary remarks of Marx and Engels on the theory of pre-
capitalist modes of production into a rigorously scientific formulation.
They have tried to apply Marx’s methodology in general theoretical
terms as well as to the analysis of particular social formations in which
slavery has been found to exist. Perry Anderson (1974) has been among
the most prominent scholars in England to devote attention to the issues.
In France, the group of anthropologists around Claude Meillassoux
(1975, 1978) have examined much ethnographic material from Africa.
In Western studies of modern China, James Watson (1980a, b, c) has
written detailed ethnographic papers and developed more general, com-
parative theories on the relations between Asian and African systems of
slavery. Simultaneaously, Orlando Patterson (1979, 1982) developed his
own, highly original theories on the topic of slavery in general.
The question of slavery in ancient China, however, was passed over
by most Western experts on slavery and even by specialists on China,
despite this growing interest in the topic of slavery among historians in
general. In this essay, I would like to analyze some of the new evidence
relating to slavery in ancient China that has only recently come to light,
in the hopes that the discussions of the question of slavery will lead to a
better understanding of the economic and social basis of the traditional
Chinese state. In addition, I hope to bring the ancient Chinese evidence
to the attention of those involved in the international debate about the
nature of slavery and its role in the development of world history.
Since the volume of studies on slavery is enormous, in this essay I
intend to use material relating to the slave systems of ancient Greece and
Rome to help illuminate relevant issues in ancient China because of their
comparable antiquity and their importance for laying the foundations
of Western and Eastern civilizations respectively. By doing this, I do not
intend to cast any aspersions on the value of drawing comparisons with
other slave systems in other social formations, such as those in Africa and
America: I merely wish to limit the article to manageable proportions.
In China from the Neolithic to the late Western Zhou 西周 and early
Eastern Zhou 東周 periods, there was no strict separation in status be-
tween the lowest members of the society and the highest. All members
were to a certain extent unfree, and it was only in the middle of the
Spring and Autumn period that a market economy really developed. I
am not claiming that there were no slaves at all in early times. Rather,
288 ROBIN D . S . YATES
I wish to argue that there was a radical change in the nature of slavery
in the middle of the Eastern Zhou period, comparable to that discussed
by Anderson in fifth and fourth century bc Greece, when slaves for the
first time came to be bought and sold. I will explain my interpretation
of the nature of slavery in ancient China and the role it played in the
social formation in the following pages. But before embarking upon this
analysis, I would like to point out that I shall draw much of my infor-
mation from the Qin legal documents, which were only discovered in
1975 at Shuihudi 睡虎地 , in Yunmeng 雲夢 , Hubei 湖北 .2 These docu-
ments provide essential information not available to previous scholars
of ancient Chinese slavery. In the West, the most notable are Martin
Wilbur (1943), Wang Yi-t’ung (1953), Edwin Pulleyblank (1958), and
Ch’ü T’ung-tsu (1972).
In 9 ad, the usurper Wang Mang 王莽 proclaimed his inaugural edict
in which he condemned the state of Qin for lacking the Way (Dao 道 ).
Among the greatest of Qin’s faults, according to Wang, was that they
established slave markets that were the same as those for oxen and horses.
And throughout the Han 漢 , it was believed that many hereditary slaves
were the descendants of criminals whose families had been enslaved
under the laws of mutual responsibility (lianzuo 連坐 ) promulgated by
the infamous legalist statesman Shang Yang 商鞅 or Lord Shang 商君 in
Qin in the fourth century bc (Wilbur 1943: 72–79; Tomiya Itaru 1985;
Yates 1987).3 The new laws paved the way for Qin’s successful conquest
of its rivals and its founding of the Chinese imperial order in 221 bc,
and this form of legal enslavement for high treason continued down to
the end of the Qing 清 dynasty. Just what role slavery played in Qin’s
unification effort, therefore, is of vital importance for understanding
the internal dynamics of the Chinese historical process.
A final reason for considering the question of slavery is that the nature
of slavery goes to the heart of the nature of social status and the social
structure not only of the early period in Chinese history, but also in the
later traditional period as well. It even has implications for social status
in the twentieth century. Slavery was only formally abolished in China
itself in the last years of the Qing dynasty (Philip C.C. Huang 2001: 17),4
2
My references in this essay are to Shuihudi Qinmu zhujian Zhengli Xiaozu (1978)
(hereafter abbreviated SHD, but see also Shuihudi Qinmu zhujian Zhengli Xiaozu
(1990). The bibliography on the Shuihudi texts is enormous. I bring the attention
of the reader only to the studies by Xu Fuchang (1993), Li Jing (1985), Wu Fuzhu
(1994), Yu Zongfa (1992), Xing Yitian (1987); Hori Tsuyoshi (1988), and Ôba Osamu
(1991).
3
The Shuo wen 說文 dictionary is one of the texts in which the opinion on the origin
of slaves is to be found (Xu Zhongshu 1981: vol.1, 652).
4
However, the enslavement of ordinary “good” commoners was illegal under the
Qing and punished severely, because the Qing were anxious to maintain the proper
hierarchy in society (Philip C.C. Huang 1996: 94–95).
SLAVERY IN EARLY CHINA 289
1. Slaves as property
Most scholars who have studied the question have emphasized the
fact that slaves were property. Watson, drawing on the work of Finley,
the anthropological work of Meillassoux (1975; 1978) in Africa, and
the sociological writings of H.J. Nieboer,6 has noted that this aspect is
primary, while the fact that slaves are marginal, liminal people, or out-
siders to society, is secondary (Watson 1980b: 8–9). In addition, Watson
emphasizes the fact that slaves’ labor is extracted through coercion. Let
us first take up the question of slaves as property.
Unlike in Rome and generally in Western civil law, property was not
easily alienable in China. There was no concept of absolute rights in
rem (in the thing itself) (Patterson 1982: 20). This was a special charac-
teristic of Roman private law, which, by a legal fiction, gave a person
absolute rights and control over an object.7 In Rome, the two main
5
I place the term “freedom” in quotation marks deliberately. As Finley (1983:
119–120) points out, “[f]reedom is no less complex a concept than ‘servitude’ or
‘bondage’; it is a concept which had no meaning and no existence for most of human
history; it had to be invented finally, and that invention was possible only under very
special conditions.” “Freedom,” he continues, was scarcely possible to say in nineteenth
century Chinese.
6
Nieboer (1900: 7) defines slavery as “the fact that one man is the property or pos-
session of another man, and forced to work for him.”
7
Ownership of a thing (res) could be passed in two different formal acts, either man
290 ROBIN D . S . YATES
sources of wealth were land and slaves, and it is quite possible that this
legal fiction of absolute rights over things derived from the institution
of slavery, and resulted in the definition of slaves as property. Indeed,
the Romans made a clear distinction between possession and owner-
ship, and much of Roman property law was based on the concept of
dominium, even though this term is never defined in the extant sources
of Roman law. Nevertheless, F. Schulz has concluded that “a Roman
principle must have existed to the effect that the right of ownership was
to be as unrestricted as possible and the greatest possible latitude given
to individual action and initiative” (Schulz 1936, quoted in Harrison
1968: 202, note 2).
The situation was different in Athens, however, where legal thinking
was not as highly developed. There was, in fact, “no general term to
describe the law of property; they had no abstract word for ownership”
(Harrison 1968: 201), and the sources are too fragmentary to permit
more than a sketchy interpretation of the data. Although the right and
power to alienate property had emerged by the fourth century bc, where
it had been previously extremely limited, still this power was restricted
to cases in which the owner lacked a legitimate heir, and full rights
of ownership were restricted in several other ways, including those in
which there existed joint ownership and where there were easements.
Under the law of succession, property seems to have been divided into
two parts, that which a man inherited (patrÓa) and that which he had
created or accumulated himself ( £epiktntá) (Harrison 1968: 125). The
former was hard to dispose of at will, while the individual had more
freedom of decision with respect to the latter.
cipatio or cessio in iuris (“cession in court”). Mere transfer of an object, such as a slave,
did not pass the right of ownership (Crook 1967: 141). Things were divided into two
categories, res mancipi such as land and houses on Italian soil, slaves, certain animals
such as beasts of burden and of draft (e.g., oxen) and one kind of easement, the rustic
praedial servitude, and res nec mancipi, which included everything else. The latter could
be transferred by simple delivery (traditio), for example a sheep (Jolowicz and Nicholas
1978: 137). Dominium could not be passed unless the person had the rights of dominium
himself. In fact, dominium could be achieved by simple usucapio (“acquiring dominium by
use”) which required a relatively short period to elapse before the object fell into the
user’s hands, two years for real estate and one year for movable property. As Jolowicz
and Nicholas state (1978: 140), “[t]he Roman law of classical times is dominated by
what is commonly called the absolute conception of ownership which it has evolved
and by the action through which this right is asserted, the vindicatio. Ownership, in
the developed law, may be defined as the unrestricted right of control over a physical
thing, and whosoever has the right can claim the thing he owns wherever it is and no
matter who possesses it.” In Germanic and English systems of law, there existed only
a relative right to possession, and as Patterson (1982: 20) points out, Anglo-American
law rejects the Roman civil view “[b]ecause, first, in both sociological and economic
terms . . . there can be no relation between a person and a thing. Relations only exist
between persons. Second, relations between persons with respect to some object are
always relative, never absolute.”
SLAVERY IN EARLY CHINA 291
Ancient and traditional Chinese law seems to have lacked the notion
of absolute rights over things, and therefore it is not adequate to define
ancient Chinese slaves purely in terms of their position as property (cai
財 ) as almost all scholars have done to date.8 They were indeed property,
but were much more than property.
Hilary Beattie, in her study of Tongcheng 桐城 county in Anhui 安
徽 in the Ming and Qing dynasties, points out that “[p]rivate landhold-
ings were of course always subject to fragmentation on the owner’s
death, though it should be born in mind that property at this time
and subsequently was never freely alienable by the individual owner;
his immediate family and other relatives had rights of objection and
pre-emption which did act as a brake on the rapid dispersion of estates
among outsiders” (Beattie 1979: 8).
In fact, not only did immediate family members and relatives have
rights over such landed property, but so also did neighbors, and this
was true into the twentieth century, at least in parts of China.9 It was
therefore the traditional Chinese family or lineage that held the landed
property jointly and, of course, in essence, this meant the male members
of the lineage.10
Two theories have been advanced by Japanese scholars on this mat-
ter: one holds that family property was the sole, personal, possession
8
Niida Noboru (1944: 905–914) emphasizes this aspect, but it should be born in mind
that the evidence of this definition of slaves as property (caiwu 財物 ) only derives from
Tang and later sources. Whether the newly discovered Han 漢 laws from Jiangling 江
陵 throw some light on the early definition of slaves remains to be seen. Wilbur (1943:
63) defines a slave as “a person who is owned as actual property by 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.”
9
Rev. J. Hutson (1919: 553), for example in the section “Purchase of Property,” notes
that Yo chieh hsia pai (yuejie xiabei 約界下碑 ) is to go round the boundaries of the property
along with the middlemen and neighboring landlords, who have to be invited to a feast;
their presence is equivalent to their recognition of their new neighbour.”
10
James L. Watson (1982: 594) defines “lineage” as a “corporate group which celebrates
ritual unity and is based on demonstrated descent from a common ancestor.” In the conclusion
to his article (Watson 1982: 617–18), he suggests the possibility that Chinese lineages
as anthropologists know them today may not have developed until the end of the Tang
唐 and the beginning of the Song 宋 dynasties. Before that time, the aristocratic elites
were interested in forming alliances and were particularly concerned with matters of
pedigree. They possessed lines not lineages. From the Song on, however, “the balance
shifted from alliance to descent.” While this may be true for the later period, it would
seem that lineages in Watson’s definition of the term did exist in Shang 商 and Zhou
China. Elites traced their descent from a common ancestor, even though in some cases
of the major lineages descent was claimed from a mythological ancestor, and most
lineages held towns or cities that were the ritual centers for their ancestral cults. There
have been many studies of this phenomenon, including Du Zhengsheng (1979) and
Kwang-chih Chang (1976). For recent scholarship on the changing rights of women to
property from Song dynasty times on, see Bernhardt (1999), Birge (2002), Ebrey (1993),
and McKnight and Liu (1999).
292 ROBIN D . S . YATES
of the living common ancestor, and the other is that it was held by
ascendants and descendants in a direct line. Inferiors and juniors may
use the property with the permission of the elders, such that it is not
the private property of the elders and superiors, but is subject to the
management of the latter.11
I would suggest that in the period of the late Warring States, Qin
and early Han, Chinese law resembled that of traditional Africa, in
which Miers and Kopytoff have argued that an individual can be seen
both as a member of the kin group and as part of the group’s corpo-
rate property and constitutes lineage wealth. And, they add, “[n]either
criterion of property nor that of salability can be useful in separating
slavery from simple kinship in African societies,” for kin members
are treated in ways westerners would consider property” (Miers and
Kopytoff 1977: 12; Watson 1980b: 4–5). Supporting evidence for my
view comes from section 5 of the Qin legal documents, no. 4 “Sealing
and Guarding” (fengshou 封守 ), where a list of a commoner’s property
is given as including wife, adult unmarried female daughter, non-adult
male child, two slaves, one male adult and one female non-adult, one
male dog, clothes, house, and ten mulberry trees (McLeod and Yates
1981: 137–39; Hulsewé 1985a: E3, 181–85). Again, in section 4, the
questions and answers on points of law, we find a definition of the
distinction between a so-called public domain denunciation (gongshi gao
公室告 ) and a non-public domain denunciation (fei gongshi gao 非公室
告 ).12 The former refers to a situation in which a person with murder-
ous intent kills, wounds, or robs another person, and the latter is one in
which a son or child robs his father or mother, and they on their own
responsibility (shan 擅 ) kill, mutilate, or shave their children and male
and female slaves (nüqie 女妾 ) (SHD 4.90, 195–96; cf. Huang Zhongye
1991: 143–45).13 In addition, in 4.91, if a child denounces his parents
for a crime or if slaves denounce their master, this too was a non-public
domain denunciation, and the authorities were not to take up the case.
And if a master (zhu 主 ) were on his own responsibility to kill, mutilate,
or shave his children and slaves (chenqie 臣妾 ) this too was a non-public
domain denunciation, and the authorities were not to prosecute. If the
children persisted in their accusations, they themselves incurred guilt
and if another third party took their side, again the authorities were to
ignore him (SHD 4.91, 196; Hulsewé 1985a: D 87, 148–149).
11
Schurmann (1956) observes that “some form of joint family property has existed
in China for at least two millennia,” and that junior members could not be accused of
stealing, only of appropriating common property.
12
Hulsewé (1985a: D 86, 148) translates gongshi as “official.”
13
The first transcriptions of the Shuihudi legal materials found in SHD were presented
unnumbered in Yunmeng Qinmu Zhujian Zhengli Xiaozu 1976a–c.
SLAVERY IN EARLY CHINA 293
Here we can clearly see that the parents (fumu 父母 ), and it is im-
portant to note that the mother is included, it is not just the father, are
alternatively referred to as masters (zhu 主 ), and that they have certain
rights over their children and slaves. I say “certain” intentionally because
of the presence of the term shan “on their own responsibility” in the
definitions. Generally speaking, the parents were expected to request
permission from the authorities before carrying out the punishment.
This gave the children and slaves some legal protection from arbitrary
savagery, yet since the parents were not subject to prosecution even if
they carried out the maximum penalty, the children and slaves cannot
have derived much comfort from the legal provision.
We can also see that the wife was in some ways in an anomalous
position vis-à-vis her status as property: on the one hand, she is treated
as part of her husband’s property, and, on the other, she has property
rights over her children. We find further support for this view in the
almanac texts (rishu 日書 ), where we read on the shouri 受日 (“Receiving
day”), the tenth in the twelve day Qin week, that one could bring into
one’s house people (renmin 人民 ) (i.e., slaves) horses, oxen, husked and
unhusked grain, take in wives and other things (tawu 它物 ) (Shuihudi
Qinmu Zhujian Zhengli Xiaozu 1978: plate 117, slip 752; Rao Zongyi
and Zeng Xiantong 1993: 408; Liu Lexian 1994: 33; cf. Kalinowski
1986; Loewe 1994: 221–26; Poo 1998: 69–101).
The data available at present make it very difficult to determine with
any precision Warring States conceptions of property. Did the ancient
Chinese distinguish between ownership and possession, as the Romans
did, or were their ideas more vague, like those of the Athenians? Cer-
tainly there is no theoretical explication of any such concepts in the
legal documents found at Shuihudi. Furthermore, all the references to
land in the documents seem to imply that the state was the sole owner:
only one reference to the “people’s fields” (mintian 民田 ) suggests, but
does not prove, private ownership of land (SHD 4.138).14 This is not
to say that there was no private ownership of land in the Qin. As is
well known, Han scholars frequently claimed that the scourge of the
accumulation of landed property in fewer and fewer private wealthy
hands derived from Qin land tenure practices and specifically from
Lord Shang’s destruction of the “well-field” (jingtian 井田 ) system (Gao
Min 1979; Xiong Tieji and Wang Ruiming 1981; An Zuozhang 1981;
14
Hulsewé (1985a: D 135, 164) translates mintian as “fields of commoners.” I prefer
to use “commoner” in a more technical legal sense of those with the commoner status
of “member of the rank and file,” who probably could be awarded or attain up to the
eighth rank in the ranking system. Min 民 here might refer to members of the aristo-
cratic lineages of daifu 大夫 rank as well as commoners or to members of those lineages
that did not belong to the ruling clan of Qin, if Gassman’s (2000) interpretation of the
difference between ren 人 and min is correct.
294 ROBIN D . S . YATES
15
Lin Jianming argues that the early Qin state never had a well-field system of land
apportionment, because before its enfeoffment with the old rich farming lands of the
Western Zhou, it was a nomadic pastoral people. When it was enfeoffed, the well-field
system was already breaking down. The earliest land system was, therefore, that of the
so-called “changing-field” (yuantian 轅 /爰田 ), in which the land was divided into three
grades according to quality, superior, middle, and lower, and each family was given a
fixed area of each, and the allotments and residences were changed every three years.
Another description states that each family was given 100 mu 畝 of land which was not
to be changed, 200 mu of land was to be changed once (every three years), and 300
mu of land to be changed twice (every three years). Thus Lin suggests that this system
developed into one in which there was a single allotment which was returned to the
state when the individual reached the retirement age of 60 years. In another way, too,
the “changing-field” system was different from the “well-field” system. In the latter, all
that the peasants (slaves, in Lin’s view) were required to perform for the state was labor
in the fields, but in the former, they were required also to perform military service (fu
賦 ). In this regard, the system of “changing-fields” was the same as the later tuntian 屯
田 “colony” system. Yet in neither system, be it the “well-field” or the early and later
“changing-field” system, did the workers have rights or powers (quan 權 ) over the land,
and it all belonged to the slave-owners, and not to the workers who were slaves (nuli 奴
隸 ). Because of the military needs of the early Qin state, it did not institute the decaying
“well-field” system but the “changing-field” allotment with its military implications.
When Lord Shang reformed the laws in Qin in the fourth century bc, he reintroduced
the “changing-field” allotment system but without the practice of changing every three
years, and thereby he merely redrew the so-called qianmo 阡陌 field boundaries of the
“well-field” system to create his new system.
While this analysis has many valuable points to it, I am not in agreement with all of
Lin’s and the other scholars’ arguments concerning the early Qin land system. For a
recent analysis of bronze inscriptions concerning transfer of title over land and related
disputes in the mid-Western Zhou, see Lau 1999; cf. Shaughnessy 1993).
It is worth noting, however, that a stone stela dating from the Eastern Han dynasty
was discovered in 1973 near Yanshi 偃師 county, Henan 河南 . This land-sale contract
provides solid evidence of one form of joint ownership of land: twenty-five individu-
als, who are named, formed a special group called the fulao dan 父老僤 of Shiting 侍
廷 village, where dan is the name of an association at the village or sub-village level, to
buy 82 mu of land for 61,500 cash (Du Zhengsheng 1990; cf. Yu Weichao 1988; Huang
Shibin 1982; Ning Ke 1982). For other views on the Qin land system, see Li Jiemin
(1981), Zhang Chuanxi (1985: 77–96), and Hulsewé (1985b: 215–218).
16
Zhao had an acre (mu) of 240 paces (bu 步 ), according to the new Sunzi 孫子 (Yates
1988: 218).
17
The Shi ji 史記 (Takigawa Kametarô 1979: j. 15, 42) records that the Qin began
to tax the land in the form of grain (chu zuhe 初租禾 ) in the seventh year of Duke Jian
簡公 (408 bc), and instituted the fu 賦 military service tax in 348 bc under Lord Shang.
(cf. Xiong Tieji and Wang Ruiming 1981: 71; Huang Jinyan 1981: 61–73).
SLAVERY IN EARLY CHINA 295
18
Hulsewé (1988: 173), observes that only one rule in the Qin law refers to “the
misappropriation of land,” but at present it is not possible to determine whether land
could actually be stolen or not.
296 ROBIN D . S . YATES
27
Keightley 1969: Appendix 1 “Slavery in the Shang Dynasty,” 357–79), effectively
dispenses with the arguments advanced by Guo Moruo (1973) and others that certain
groups of forced laborers in the Shang can be identified with slaves, even though there
are some discrepancies between the definitions of the terms employed by him and the
ones that I am accepting in the present paper. Keightley analyzes the terms nu 奴 , pu 僕 ,
chen 臣 , qie 妾 , and zai 宰 as they appear in the oracle bone inscriptions, because these
terms come to be applied to servile and semi-servile statuses in later times. He shows
that Guo has misread the two inscriptions which purport to mention the word nu and
that the element in the ancient graph , which is the modern xin 辛 , may represent
a decorative tail: thus pu probably refers to a figure of relatively high status dressed
in ritual garb. Pu therefore has nothing to do with slaves. With regard to zai 宰 , Guo
argued that they were of the same status as chen 臣 which shows that, since the latter
were slaves, so were the former. In fact, Guo has once again misread the relevant graph
, which may in fact be the ancient form of kou 寇 and probably refers to high status
army officers. As for qie 妾 , Shima Kunio (1958: 337, 482) has shown that they were
“shamaness participants in the ritual of ancestor worship.” But even if they were sac-
rificial victims, this does not prove that they were slaves in the strict sense of the term.
Finally, with regard to chen 臣 , Guo argued that the archaic graph represents an eye in
the vertical position, indicating a prisoner of war with bowed head, and therefore chen
were slaves. Nevertheless, the inscriptional evidence suggests that the xiaochen 小臣 were
rather “members of the royal administration who performed the king’s orders by leading
the peasantry in various tasks” (Keightley, 1969: 369). Keightley concludes that there
is no evidence that Shang state labor employed slavery, or that people were owned like
property, nor any concerning the tasks prisoners of war were forced to perform, even
though perhaps some of them became human sacrificial victims. Finally, there is no
evidence concerning the percentage of the population that was free as opposed to slave.
The conclusion is clear: Shang was based not on slave labor but on forced labor.
28
See Li Hengmei (1999) for a recent discussion of slaves in Spring and Autumn
times, including their origin, status, and occupations.
302 ROBIN D . S . YATES
29
Watson (1980c: 223) notes that in the later traditional period nearly every peasant
household was affected by slavery, and most of the slaves were sold because of extreme
poverty.
30
Hulsewé does not venture an opinion as to whether the pledge is to be understood
as a person or as an object.
304 ROBIN D . S . YATES
Changwei and Song Min (1982) have argued, are private slaves in con-
trast to lichenqie bondservants, whom they interpret as state slaves.
People in this type of status were considered to be property of their
owners and not human in the social sense, yet at the same time, they
were definitely held responsible for their crimes and punished by the
government, and were not considered to be suffering from diminished
capacity in the legal sense. Crimes mentioned include beating their mas-
ters (zhu 主 ) (SHD 4.63; Hulsewé (1985a: 169), raping them (presumably
this applied primarily to mistresses) (SHD 4.63; Hulsewé 1985a: 169),32
secretly planning to murder their masters (SHD 4.64, 184; Hulsewé
1985a: D 60, 140), planning to steal their master’s ox, selling it, taking
the proceeds, and fleeing the state (SHD 4.4, 152–153; Hulsewé 1985a:
D 4,122), unauthorizedly killing their child (SHD 4.61, 183; Hulsewé
1985a: D 59,140), beating their child so that it dies of its wounds (SHD
4.62, 163; Hulsewé 1985a: D 62, 140–142),33 robbery (SHD 4.18,
159–160; Hulsewé 1985a: D 18, 125–126; SHD 4.122: 211; Hulsewé
1985a: D 119, 159–160),34 and finally being a scold (han 悍 ) and refus-
ing to work in the master’s fields.35 It is important to note that the state
claimed the right to punish such delinquent slaves, even though their
masters could mete out the punishment themselves. This was one more
way in which the Qin state tried to dominate all members of society
and gain exclusive control over their lives.
In addition to being punished, slaves could be sold to the government
at the fair market price. In one case, the slave has his foot cut off, and
is made a chengdan forced laborer (SHD 5.14, 259–260; Hulsewé 1985a:
E 15, 193–194; McLeod and Yates 1981: 5.14, 146–147). However, he
first has to state that he has never been manumitted, is not sick, and is
not liable for any other crime. The lingshi 令使 a low official in charge
32
The punishments for beating and raping the master/mistress were the same, but
not stated in the documents. The punishments meted out to the slave in Qin times for
the crimes of beating, raping, and planning the murder of the master/mistress were
probably comparable to those in the Qing dynasty. Immediate decapitation was the
punishment for beating the master and fornication with the master’s main wife or his
daughter, with strangulation after the assizes for such fornication with a master’s con-
cubine. (Meijer 1980: 335).
33
The slave is tattooed on the forehead and cheekbones and then returned to the
master; no hard labor is stipulated.
34
In the latter case, it is stated that it is the state that rewards the individual who ar-
rests a male or female slave for stealing 110 cash, rather than the owner who rewards
him.
35
SHD 5.14, 259–260; Hulsewé 1985a: E 15, 193–194; McLeod and Yates 1981:
5.14, 146–147, where a male slave is a scold and refuses to work in the fields. SHD
5.15: 260–61; Hulsewé 1985a: E 16, 194–195; McLeod and Yates 1981: 5.15, 147–148,
where a female slave confesses that she is a scold, but her occupation is not given. Being
a “scold” was one of the seven legitimate causes for divorce in traditional law (Johnson
1997: 167, Article 189, subcommentary).
306 ROBIN D . S . YATES
36
Hulsewé (1978: 199) and (1985a: A 40, 50, note 9) has demonstrated that the sha-
onei treasuries were part of the metropolitan and regional governmental organization,
being attached to all organs that were involved in financial transactions.
37
Su Yiwu 蘇夷吾 , marquis of Pu 蒲侯 , had his marquisate abolished for this reason
in 18 bc (Hongjia 鴻嘉 year 3) (Wang Xianqian n.d.: j. 17: 25b; Wilbur, 1943: no. 102,
419).
SLAVERY IN EARLY CHINA 307
and is tattooed. Another case in the slips refers to slaves being mutilated,
being made chengdan convicts and then being returned to their masters
(SHD 4.4, 152–153; Hulsewé 1985a: D 4, 122). This surely indicates
that the forced labor punishment was for a fixed period of time, and
therefore convicts cannot have served life sentences, and thus were not
structurally equivalent to slaves, as some scholars have maintained. It
should also be noted that the Qin permitted their convicts to return
to their homes to perform essential agricultural duties at appropriate
times of the year. No mention of slaves is made in this connection, and
therefore it is unclear whether their masters could petition to have their
slaves returned to them on these occasions. That this was possible, I
consider to be quite unlikely.
Slaves could also be used by their masters in the same way as oxen
and horses to be held by the government to pay off their fines and debts.
To be the owner of slaves was thus additionally beneficial. Such slaves
had to wear red clothes, be shackled, and be supervised, if they were
held to work with the chengdan and chong forced laborers (SHD 1 Sikong
12, 84–87; Hulsewé 1985a: A 68, 67–71), just like other chengdan convicts
who were originally of free status (SHD 1 Sikong 12, 89–90; Hulsewé
1985a: A 70, 72–73), and like guixin collectors of firewood and baican
sorters of grain convicts and a status called qunxiali 群下吏 , who had
not suffered the punishment of shaving the whiskers (SHD 1 Sikong 12,
84–87; Hulsewé 1985a: A 68, 67–71). On the other hand, individuals
of the lowest aristocratic rank gongshi 公士 and commoners (shiwu 士伍 ),
who were being held to redeem mutilating punishments and the death
penalty and who were held for chengdan and chong labor, did not have to
suffer the indignity of wearing red clothes and being shackled.38
This difference between slaves and regular “free” convicts redeeming
their crimes or fines is of utmost importance, in my opinion, to see the
difference between slave and convict status. The Qin obviously drew
a careful distinction between redeemable crimes and non-redeemable
crimes and between regular convicts and slaves. Slaves were treated in
the same manner as individuals serving non-redeemable heavy labor
sentences, but at the same time the term “non-redeemable” must be
understood not as meaning that the convict never ended his term, that
his labor was for life, but rather that the redeemable crimes could be
paid off in cash all at once if the criminal had the financial resources.
If he lacked the means, he was held until he paid off the crime. The
rate was fixed at 8 cash per day, 6 cash if the convict received food
38
Hulsewé (1985a: 69, note 10) observes that the phrase “gongshi or less 公士以下 ”
probably refers to gongshi and commoners (shiwu), but this would be rather remark-
able.
308 ROBIN D . S . YATES
from the government (SHD I Sikong 12, 84–87; Hulsewé 1985a: A 68,
67–71). In other words, in a 365 day year, the convict would pay off
2920 cash or 2190 cash, which would work out as 97.3 piculs (shi 石 )
or 73 piculs of grain, since the price of grain was fixed at 30 cash per
picul. Unfortunately, it is nowhere stated what the cash equivalent of
the various hard labor service punishments were, but if, as I believe,
the labor terms ran from one to five years for the watchmen through
to the chengdan the minimum might be 2920 or 2190 to a maximum of
14,600 to 10,950, a very large sum for the average commoner to pay
all at once.
We might note, however, that the government of the Qin state must
have benefited immensely from this practice of cash payment, for if it
did not acquire the forced labor of the criminals, then it received lump
sums of cash, which it could then put to good use to increase its military
and public works budgets and enable it to field large armies for years at
a time and pay for the construction of huge public works, such as the
Great Wall, the First Emperor’s tomb and the many palaces in the Wei
river 渭河 valley.
A further point with regard to slave status is that the wives of slaves
were apparently not graced with the term that meant legitimate wife,
qi 妻 , they are always referred to as qie 妾 “concubine.” This is of great
significance, for it is a practice common to other slave systems through-
out the world (Patterson 1982: 186–190). Slaves are slaves because they
are cut off from rights and obligations of family and their unions are
not recognized as legitimate and binding on the rest of society. Their
“families” could be separated by their owners entirely at their owners’
discretion, although, in general, slave owners kept such families together
to encourage better performance of duties.39 In the Qin, evidently slave
unions were recognized, because they were recognized as parents who
could kill their children. Yet in the customs of the time only legitimate
wives could perform duties in the sacrificial rites to their husbands’
ancestors, concubines could not. It is for this reason that I say that
slave unions were recognized, but were not considered legitimate: con-
cubinage was, however, condoned for the free population, and so slave
unions were assimilated to this secondary form of marriage.
As mentioned above, another important item quoted in section 4
of the legal documents states that if a husband commits a crime and
his wife denounces him first, before the authorities learn of the crime,
she is not subject to the law of arrest (shou 收 ) and her male and female
39
Harrison (1968: 177) remarks that in the case of Athens “although relatively
permanent unions were in fact formed between slaves it is very doubtful whether these
had any juristic validity” (cf. Fogel and Engerman 1974: 49 and 142–144).
SLAVERY IN EARLY CHINA 309
40
I adopt Hulsewé’s explanation of the term toushu 投書 because it makes better
sense in the context in contrast to that of the SHD editors who interpret the term as
“to lodge an anonymous complaint.”
41
Keightley (1969:210–216) argues that these chenqie were not slaves in his definition
of the term, even though they were originally captured by the Western Zhou aristocratic
rulers. The reasons are that chen were never transferred as individuals but in groups
or families (jia 家 ) and that geographical names were frequently attached to them,
indicating “that we are dealing with groups of people whose previous social organiza-
tion or regional location took precedence over their status as ch’en. In other words, the
ch’en’s primary ties were to their family units and to the land” (213). The term chenqie
certainly appears in the Zuozhuan and I would argue that there is a strong possibility
that, with the development of the beginnings of a commodity market system in the late
Spring and Autumn and early Warring States periods, these chenqie began to be sold as
individuals. Huang Ranwei (1978) doubts the interpretation of the terms as indicators
310 ROBIN D . S . YATES
44
Hulsewé translates nilu as “innkeepers.” The precise meaning of the terms jiamen
and nilu however, has been the subject of some debate. The SHD editors (293, note
4) interpret the former as jiamen 假門 to be read as jiamen 賈門 , with the meaning of
“traders” or “merchants” (shangjia/gu 商賈 ). Wu Rongzeng (1995a, 1995b), however,
disagrees and makes a convincing argument in favor of reading jia as jian 監 , with the
full term jianmen 監門 meaning a low-status “door- or gate-keeper.” Yang Heding (1983)
rejects his conclusions and attempts to prove that the first interpretation is correct. Un-
fortunately, his efforts to show the importance of gates (men 門 ) for markets have little
to do with the meaning of this status or functional/occupational term, and I remain
unconvinced at present. Indeed, it is not impossible that jia should be taken as it stands,
and the term jiamen should be understood as low-status bondservants or slaves who had
been loaned (jia 假 ) out by the government to private individuals: Li Jiahao points out
that in Lu dialect the two words jia and jian were interchangeable (Wu Rongzeng 1995b:
171). We find in the Statutes on Granaries, for example 1 Cang 17 (SHD, 48–49) slave
girls not yet of serviceable age (weishi 未使 ) loaned out to the common people (baixing
百姓 ), who are then responsible for clothing and feeding them. Males in Wei could
perhaps have also been loaned out. Nevertheless, in this interpretation, there remains
the problem of the meaning of men.
312 ROBIN D . S . YATES
Shang’s laws stipulated enslavement for the idle (Takigawa 1979: j. 68:
8–9; Duyvendak 1963: 15), and the Guanzi 管子 suggested that they be
sent to the frontier to perform guard duty, a form of punishment (Guanzi
1934: ce 3 pian 78 “Kuidu,” 90). In another Wei state statute preserved
with the household statute in the Qin tomb, the authorities proclaim
that such indigents should be pressed into the army where the generals
should not show them any mercy (SHD, 294–295; Hulsewé 1985a: F
2, 209–210). The king even implies that they should be killed outright,
and, if not outright, they should be used to fill in the moat as human
waste when the need arose during the siege of a city, when an assault
on the walls was considered necessary.45
Bondservants
Let us now turn very briefly to the question of the bondservants (lichenqie),
which, as I have already mentioned, some scholars interpret as convicts
and others as government slaves. I believe that Liu Hainian (1985, 1986)
is correct that some bondservants were convicts whose sentences were
of limited duration and others were permanent government slaves. This
situation gives further evidence that the number of grades in the orders
below the “free” commoners (shiwu 士伍 ) was considerable.
First of all, quite clearly some of them were convicts and probably
served three-year hard labor terms. Others, however, inherited their
status from their father in the following way. Unlike slaves, a male
bondservant could have a wife not of bondservant status, and the
union of the couple was considered legitimate, for the wife was called
qi 妻 . A child born of this union while the father was still a bondser-
vant, however, inherited his father’s status. If on the father’s death, the
wife tried to conceal the circumstances of the child’s birth, she herself
became a bondservant (SHD 4.155, 225–226; Hulsewé 1985a: D 154,
169–170).
It is interesting to note that, generally speaking, marriage to a male
slave resulted in the free woman accepting his status, and this was true
in ancient and modern times both in the East and the West (Patterson
1982: 131). Indeed, in later Chinese custom, the principle of deterior
condicio that the rule in the Qin exemplifies, continued in full force, such
that the child inherited “the status of the parent with the lower status.”
In the Yuan, the rule was changed and the wife took on the status of
the husband (Patterson 1982: 141–144).
45
I follow the rendition of the SHD editors: Hulsewé’s translation does not seem
appropriate here.
SLAVERY IN EARLY CHINA 313
46
The editors place slip 153 at the end of this rule. In Hulsewé’s translation, “If
someone for the redemption of banishment wishes to pay money, this is eight cash per
day.” It has been suggested to me that if lichen were slaves and the status was considered
the equivalent of 5 years at the frontier, to manumit such a slave would be the equiva-
lent of 14,600 cash. This is possible, but it may not be correct to place this slip in this
location, and so I reserve judgment on this matter.
SLAVERY IN EARLY CHINA 315
47
Qiu Xigui accomplished the difficult task of deciphering slips 875–890.
48
The level to which commoners could aspire in the Qin bureaucracy is unknown.
Cho-yun Hsu (1965) gives many examples of low status individuals rising to the high-
est ranks in other states, but I am not sure if this was always true in Qin. Commoners
there may have only risen to the equivalent of the eighth rank in the rank hierarchy.
Under Lord Shang’s laws, an individual who was awarded rank for cutting off heads
in battle could, if he so chose, return the rank and receive an official position instead,
a policy which Hanfeizi deplored (Wang Xianshen 1956: ce 4 pian 43 “Shuoyi,” 22;
Liao 1959: vol.2, 215).
316 ROBIN D . S . YATES
the means of production, such as oxen and tools, even if at present the
extent to which they were able to own the land itself is open to question
(Hulsewé 1985b: 215–18). As we have seen, these commoners could
own slaves themselves, but there was clearly a wide range in financial
circumstances between the richest and the poorest of them, and their
wealth certainly was not at all equal to that of the aristocrats.
Conclusion
What other conclusions can be drawn from the data presented above?
First of all, it must be recognized that there was a range of low statuses
below that of the common people. In non-technical legal language, the
common people were the pingmin 平民 “level people” and were known
also by the term shuren 庶人 or shumin 庶民 . Manumission or freedom
from convict status always used this latter term. However, the legal sta-
tus for a male commoner was shiwu a “member of the rank and file.”
This status is to be compared and contrasted to lower orders, and to
the so-called aristocratic ranks (jue), which commoners could acquire by
cutting off heads in battle, inherit from their fathers, or by performing
other meritorious deeds for the government. These we have mentioned
above, but there is also the possibility that they could purchase them and
win them by producing exceptional quantities of agricultural surplus.
Probably commoners, as in the Han dynasty, could not expect to be
promoted above the eighth rank in the aristocratic hierarchy, a hierar-
chy that stretched up to the king and through him to the cosmos as a
whole. But the evidence for the Qin on the break between commoners
and aristocrats is extremely limited.
Below the commoners, and cut off from the ranking system were the
gate-guards, pawns, migrants, and other low statuses,49 who were denied
rights to land, houses, and registration in households. Below them were
the convicts, and then slaves. Within the convict status, a distinction was
drawn between those who had to wear special red clothing and shackles,
on the lowest level, and the lesser hard labor service convicts, those who
served up to three years. Although all convicts served for only fixed time
periods, yet at the same time, after completion of their sentences, they
still did not regain full commoner status or rights. In the first place, if
they had been mutilated, they were still considered polluting and could
49
One such status was the “man-marmot” (renhe 人貉 ), whose children were forced
to go and take care of their master. If they did not, they were confiscated by the gov-
ernment. If, however, they provided him with food, rather than taking care of him,
they were not be confiscated, but given back to their master (SHD 4.175, 235; Hulsewé
1985a: D 174, 177). As Hulsewé observes, this item is extremely obscure.
SLAVERY IN EARLY CHINA 317
earth.50 It would appear that this was already true in Warring States
times.
From this conceptual ordering of the world, with its emphasis on
the preservation of family ties, of the rites, we can understand how
crime was considered to be disrupting to the entire hierarchy that held
the cosmos together. We can also see why convicts and even more so
slaves were considered to be total outsiders, polluting and dangerous
(cf. Douglas 1966). And finally, we can note how difficult it was for the
Chinese to accept on equal terms peoples on their periphery, such as
the nomads on the northern border and the native tribes of the south
and southwest, as genuine human beings. They did not accept the rites
and the agricultural way of life which defined the social and economic
life of the truly Chinese, and thus by their very nature were of the same
category as slaves. It was thus no accident that the ancient Chinese felt
few qualms in enslaving them when the opportunity or the necessity
arose. Finally, we need to note a remarkable implication of the concep-
tion of slaves and convicts as marginal, liminal people. Because they
were cut off from normal human relationships and were thought of as
socially dead, they themselves could touch the really dead, for both they
and the corpses were dead, and were polluting to the rest of human
society (cf. Yates 1997).51
In this essay, I have not been able to confirm the existence of a slave
society or that the slave mode of production was ever the dominant mode
in ancient China. Nevertheless, I believe I am correct in asserting that
the social institution of slavery, together with its associated penal system,
50
Wallace Johnson (1997: 155) translates the Tang code and subcommentary on this
matter. The subcommentary reads as follows. “The wife is an equal as was said of the
states of Ch’in (Qin) and Chin (Jin). Concubines can be bought and sold and so are
very different from the wife. Slaves are inferior and basically not of the same class as
commoners.
If a person makes the wife a concubine or makes a slave the wife, this violates and
sets aside an agreement that was decided upon, injures the true way of husband and
wife, and insults the principles of human relationships. If the distinction between upper
and lower classes is upset and ritual and principles are scattered, those who commit
such crimes are punished by two years of penal servitude.” For a comprehensive study
of concubines, female slaves and courtesans in the Tang, see Liao Meiyun (1995).
51
It might be valuable to apply the cultural analysis suggested in this article to the
study of eunuchs in Chinese history. The anomalous position of slaves in, or rather
outside, society in the conceptual scheme of things might explain why eunuchs, the
ultimate slaves, were so hated and yet so necessary. At various points in time, they medi-
ated between the sacred Yang (quintessentially male) emperor and the general mass of
the bureaucrats, courtiers, and common folk, who were Yin (quintessentially female in
relation to the emperor) because they were degendered, neither male nor female, and
were “pure” enough to be able to come into contact with the sacred emperor. During
most of Chinese history, however, they stayed within the imperial palace ministering
to the emperor’s private needs and to those of his empress and his harem. In this case
too, they were permitted to do so because they were degendered males.
SLAVERY IN EARLY CHINA 319
Appendix
One of the early Han legal cases found at Zhangjiashan 張家山 docu-
ments how a female slave Mei 美 attempted to reject her re-enslave-
ment by her former master Dian 點 and subsequent sale to the Dafu
Lu 大夫祿 (Jiangling Zhangjiashan Hanjian Zhengli Xiaozu 1993: 22;
Li Xueqin 1993; Peng Hao 1993; Yates 1996). The text of the official
report reads as follows.
The eleventh year, eighth month, bingxu day, jiashen being the first day of
the month, Ao 驁 , Assistant of Jiangling, ventures to submit to higher
authority the (following) case:
In the third month, jisi day, the Dafu Lu made a statement, saying:
“In the sixth year, second month, I bought the slave Mei at the member
of the rank and file (shiwu) Dian’s place. The price was 16,000 cash. But
then in the third month, dingsi day, she absconded. I sought and captured
Mei. Mei said, ‘I should not be considered a slave.’”
• Mei said, “Formerly I was Dian’s slave. In the time of Chu, I went
off and absconded. I surrendered and became a Han. My name was
not written down or enumerated. Dian captured me and registered and
enumerated me again as his slave. He sold me to Lu’s place. I (originally)
had pledged myself; I should not suffer enslavement (again). Accordingly,
I went off and absconded. The rest is as Lu (states).”
• Dian said, “Mei formerly was Dian’s slave. In the time of Chu, she
absconded. In the sixth year, in the second month, I captured Mei. Mei
did not yet have a name and was not yet enumerated. Then I registered
and enumerated her, and sold her to Lu’s place. The rest is as Lu and Mei
(state).”
• I (i.e., Assistant Ao) investigated Mei, (saying), “You, Mei, were former-
ly Dian’s slave. Although in the time of Chu, you went off and absconded,
and surrendered and became a Han, you did not write down or enumerate
(yourself). Dian captured (you) and registered and enumerated you. You
(Mei) thereby again became a slave. That he sold you was appropriate. But
you went off and absconded. What is your explanation?”
• Mei said, “In the time of Chu, I absconded. Dian then became a Han.
He again enslaved me, and sold me. I had pledged myself. I ought not to
have been considered a slave again. So then I went off and absconded.
There is no other explanation.”
• I (Assistant Ao) questioned Mei. She is forty years old. The rest is as
in the deposition.
• I (Assistant Ao) judged the case. Mei formerly was Dian’s slave. In
320 ROBIN D . S . YATES
the time of Chu, she absconded and surrendered to be a Han, but she
did not write down her name or enumerate (herself). Dian captured (her)
and registered and enumerated (her). He made her a slave again, and sold
her to Lu’s place. But Mei went off and absconded. She is forty years
old. She was captured. All (the above-mentioned facts) have been care-
fully investigated.
• I suspect Mei is guilty of a crime, but this may have been discussed in
another county. I venture to submit (the case) to higher authority; when I
referred (it) to higher authority the response was that the post [investiga-
tion] was to be similar to the Granary Laws(?).
• The official (li) matching (dang 當 ) is: Tattoo Mei’s face and . . . ,
and give her to Lu. An alternative states, “She matches (deserves) to be a
commoner (shuren).”
This document is very interesting on a number of points. It provides
information, for example, on the price of slaves in the early Han, and
evidence that Han officials closely followed Qin legal forms and proce-
dures.52 It is the latest in the collection from Jiangling: the date of As-
sistant Ao’s submission corresponds to 196 bc, according to Li Xueqin
(1993: 30) and Peng Hao’s (1993: 32) calculations. The slave Mei, being
forty years old at the time of the case, must have been born well before
the Qin unification, in 236 bc, presumably in the state of Chu, though
the “time of Chu” probably refers to the time of the recreation of the
Chu kingdom under Xiang Yu 項羽 after the collapse of Qin imperial
authority.
The crux of the issue, as Li Xueqin points out, is that Mei failed to
register herself as a free commoner after she had absconded and sub-
mitted to Han authority. She hoped that since her master, an ordinary
commoner, Dian, had also surrendered to the Han, and thus both of
them had changed their recognition of what was the legal authority,
the pledge of her person as debt-slave to Dian under Qin rule had
been invalidated and she should no longer be considered a slave. Even
though it is recorded in 202 bc that Han Gaozu seems to have manu-
mitted slaves who had fallen into bondage as a result of poverty in the
previous period,53 here it appears that one legal opinion was that she
should be punished with tattooing of the face and returned to bondage
in Lu’s house, thus upholding the validity of Dian’s re-enslavement and
subsequent sale of her.
I think that the Han official is suspicious about the circumstances
52
The contention in my article Yates (1987) that shiwu indicated commoner status
and was not the name of a punishment is fully born out in the present case, for the
shiwu Dian owns the slave Mei.
53
Han shu j. 1B, p. 4b; Dubs 1938 vol. 1: 104, reads, “As to those people who because
of famine or hunger have themselves sold their persons to be slaves or slave-girls, let
them all be freed and become common people.”
SLAVERY IN EARLY CHINA 321
References Cited
Liu Weimin 劉偉民 (1975). Zhongguo gudai nubi zhidu shi 中國古代奴婢
制度史 (History of the ancient Chinese slave system). Hong Kong:
Longmen Shudian.
Liu Xiang 劉向 n.d. Shuo yuan 說苑 (Garden of sayings). Sibu congkan
ed. Shanghai: Shangwu Yinshuguan.
Liu Yin 劉寅 (editor) (1975). Liutao zhijie 六韜直解 (Direct explanations
of the Liu tao). Kambun taikei ed. Tokyo: Fuzambô.
Loewe, Michael (1974). “The Grand Inquest — 81 b.c.” In Crisis and
Conflict in Han China 104 B.C. to A.D. 91, pp. 91–112. London: George
Allen and Unwin.
Loewe, Michael (1994). “The almanacs (jih-shu) from Shui-hu-ti: a pre-
liminary survey.” In Divination, Mythology and Monarchy in Han China,
pp. 214–35. Cambridge: Cambridge University Press.
McKnight, Brian E. (1981). The Quality of Mercy: Amnesties and Traditional
Chinese Justice. Honolulu: University of Hawaii Press.
McKnight, Brian E. and James T.C. Liu (trans.) (1999). The Enlightened
Judgments Ch’ing-ming Chi: The Sung Dynasty Collection. Albany: State
University of New York Press.
McLeod, Katrina C.D. (1978). “Law and the symbolism of pollution:
some observations on the Ch’in laws from Yun-meng.” Unpublished
paper, Annual Meeting of the Association for Asian Studies, Chi-
cago.
McLeod, Katrina C.D. and Robin D.S. Yates (1981). “Forms of Ch’in
law: an annotated translation of the Feng-chen shih.” Harvard Journal
of Asiatic Studies 41.1: 111–163.
Meijer, Marinus J. (1980). “Slavery at the end of the Ch’ing Dynasty.” In
Jerome Alan Cohen, R. Randle Edwards, and Fu-mei Chang Ch’en
(eds.), Essays on China’s Legal Tradition, pp. 327–358. Princeton, N.J.:
Princeton University Press.
Meillassoux, Claude (editor) (1975). L’Esclavage en Afrique Precoloniale.
Paris: Maspero.
Meillassoux, Claude (1978). “Correspondence (On slavery).” Economy
and Society 7: 321–331.
Miers, Suzanne and Igor Kopytoff (editors) (1977). Slavery in Africa:
Historical and Anthropological Perspectives. Madison: University of Wis-
consin Press.
Miller, Joseph C. (1985). Slavery: A Worldwide Bibliography 1900–1982.
White Plains, N.Y.: Kraus International.
Musha Akira 武者章 (1980). “Sei-Shû sakumei kinbun bunrui no ko-
koromi 西周冊命金文分類の試み (An attempt at a comprehensive
classification of the Western Zhou bronze inscriptions: the case of
the investiture of feudal vassals (ceming jinwen).” In Matsumaru Michio
松丸道雄 (editor), Sei-Shû seidôki to sono kokka 西周青銅器とその国家
SLAVERY IN EARLY CHINA 327
Address:
Departments of History and East Asian Studies
McGill University
3434 McTavish Street
Montreal QC H3A 1X9
Canada
robin.yates@mcgill.ca