. 6
( 15)


noting). The Chou states™ relations with Tai, including the king™s ensnare-
ment, follow the pattern of relations between Chou and non-Chou preva-
lent in the late Spring and Autumn period and remind us also of the Chin
conquest of the Hsien-yü city of Ku, in 520 b.c. As the conquest of the
various Jung and Ti peoples that had once surrounded the Chou states was
coming to an end, the state of Chin pushed its borders into an area occu-
pied by a kind of people unfamiliar to them, the Hu.
Although “Hu” may have once been used as an ethnonym, in pre-Han
sources it was just a generic term for nomads,7 which, by Han times, had
become synonymous with Hsiung-nu.8 The term “Mo,” mentioned in the
preceding passage together with “Hu,” is an old term that occurs in the
Shih ching and in the Chou li in reference to foreign neighbors.9 Generally
speaking, there is no textual basis indicating that the Hu or Mo constituted
a more or less homogeneous ethnic or linguistic group. Rather, “Hu” was
used as a blanket term that included mounted bowmen who practiced pas-
toral nomadism as their main economic activity. Therefore, Hu signi¬es
an “anthropological type” different from the Jung and Ti, but does not
imply linguistic or ethnic similarities among the different hu.10 According
to an anonymous letter sent to the king of Yen, and recorded in the
Chan-kuo ts™e:

Shih chi chu-yi 110, 2315“16, n. 3 (Burton Watson, trans. Records of the Grand
Historian by Sima Qian [New York and Hong Kong: Columbia University Press
and The Chinese University of Hong Kong, 1993], 2: 132).
Jaroslav Prusek, Chinese Statelets and the Northern Barbarian in the Period

1400“300 B.C. (Dordrecht: Reidel, 1971), pp. 224“25.
This is evident also from the names of border garrisons and companies engaged
in ¬ghting the Hsiung-nu in the Western Regions, such as Ling Hu (Those who
oppress the Hu) and Yen Hu (Those who detest the Hu). See E. Chavannes, Les
documents chinois decouverts par Aurel Stein dans les sables du Turkestan Ori-
ental (Oxford, 1913), p. x. Cf. Lin Kan, Hsiung-nu shih-liao hui-pien (Peking,
1988), pp. 50“51; 152, no. 9; 154, no. 12; 166, no. 47; 167, no. 49.
Chou Li 10, 1a (trans. Biot, Tcheou-li, 2: 264). See also A. W. Rickett, Guanzi:
Political, Economic, and Philosophical Essays from Early China, 2 vols.
(Princeton: Princeton University Press, 1985“98), 1: 388. On this term see also
E. Pulleyblank, “The Chinese and Their Neighbors in Prehistoric and Early
Historic Times,” in The Origins of Chinese Civilization, ed. David N. Keightley
(Berkeley: University of California Press, 1983), pp. 442“46.
At the same time, the differentiation between Hu, on the one hand, and Jung and
Ti, on the other, must not be taken in absolute terms; in some cases the term jung

The Hu and the Yüeh cannot understand one another™s language and cannot
communicate their ideas, but when mountainous waves arise about the boat
they share, they rescue one another as though they were from a single state.
The allies of Shan-tung are now as if they were crossing a river on the same
boat, but when Ch™in troops reach them they will not rescue one another as
though they shared a single state. Their wisdom indeed cannot match that of
the than Hu and Yüeh people.11
Obviously the meaning of this passage cannot be that Hu and Yüeh were
in mutual contact and helped each other against a common threat. Instead,
the name “Yüeh” is used as a generic term for the non-Chinese peoples of
the south,12 while “Hu” describes non-Chinese peoples of the north. We
can thus reasonably say that, by the end of the fourth century b.c., the term
“Hu” applied to various northern ethnic groups (tribes, groups of tribes,
and even states) speaking different languages and generally found living
scattered across a wide territory. Their fragmentation, however, could be
turned, when the need arose, into a superior form of political organization
(a “state”). This explains why hu appears often preceded by a quali¬er that
we may take as a marker for a speci¬c ethnic group, as with the Lin Hu
and the Tung Hu.13 Whether or not it had originally been an ethnonym,
such a designation had been lost by the late Warring States period.
According to Lattimore, the “new names” for the northern peoples that
appear in the Chinese sources re¬‚ect a change in the economic specializa-
tion of non-Chinese peoples, owing to their forcible expulsion into a more
arid ecological zone. Lattimore argues that the Hu and Hsiung-nu were the
Jung and Ti of old, who turned to a nomadic lifestyle once they had been
pushed into the steppe by expanding Chinese polities.14 This in¬‚uential
argument cannot be sustained, however, in the light of the archaeologi-
cal evidence presented earlier, which shows that the transition from semi-
sedentary agro-pastoralism to pastoral nomadism took place in the
Northern Zone of China over a long period.
In fact, we know little about the relations between northern Chinese and
nomads in the period preceding the famous dispute over “Hu” clothing “

continued, somewhat anachronistically, to appear in the sources in reference to
horse-riding peoples, such as the Lin Hu and the Lou Fan; see Shih chi chu-yi
110, 2315.
Chan Kuo Ts™e, 30 (Yen 2), p. 1110 (Crump, Chan-Kuo Ts™e, p. 516).
On the Yüeh people and state, see William Meacham, “Origins and Development
of the Yüeh Coastal Neolithic: A Microcosm of Culture Change on the Mainland
of East Asia,” in The Origins of Chinese Civilization, ed. David N. Keightley
(Berkeley: University of California Press, 1983), pp. 147“75; Pulleyblank, “The
Chinese and Their Neighbors,” pp. 438“41.
Shih chi chu-yi 43, 1322 and 1323, n. 20 (Chavannes, M©m. hist., 5: 44, 81).
Owen Lattimore, Inner Asian Frontiers of China (Boston: Beacon Press, 1962
[1940]), pp. 448“49.


that is, a type of attire suitable for riding horses “ leading to the adoption
of cavalry by the state of Chao, a momentous event that signaled the
growing importance of the northern frontier to the process of state build-
ing among the northern states. References to incursions, submissions, wars,
and court visits are lacking, but we do ¬nd scattered allusions to trade,
suggesting that nomads and sedentary peoples enjoyed peaceful, though
distant, relations, and seeming to contradict the established wisdom that an
endemic and irreconcilable hostility existed between them.15 Compared to
the Ti and Jung of the seventh and sixth centuries b.c., the northern nomads
from about 450 to 330 b.c. were a tame neighbor, far from the dangerous
lot that they would later become.


One area in which relations between nomads and sedentary peoples seem
to have ¬‚ourished is trade, a consideration supported not only by archae-
ological data but also by text references. In the Chan-kuo Ts™e16 there are
marginal, yet clear indications that the Chou states and the Hu traded with
each other, with the Hu exporting horses and furs to the Chou states.17
In another example, the Mu T™ien-tzu chuan (Biography of the Son of
Heaven Mu, a text usually dated around the fourth or third century b.c.),
though largely ¬ctional, mentions information that must have originated
in actual practices and customs.18 In the course of his legendary travels,

This point was convincingly argued by Lattimore, who maintained that “this
frontier was the voluntarily demarcated limit of the convenient expansion of
the Chinese empire; in other words it was not necessitated by the aggression of
the nomads against China.” See Owen Lattimore, “Herders, Farmers, Urban
Culture,” in Pastoral Production and Society: Proceedings of the International
Meeting on Nomadic Pastoralism, Paris 1“3 Dec. 1976, ed. L™Equipe ©cologie et
anthropologie des soci©t©s pastorales (Cambridge: Cambridge University Press,
1979), p. 481.
The general view of this text is that it contains much anecdotal and ¬ctional mate-
rial, and therefore should be taken with a grain of salt when used as a historical
source. Because the stories are set against a historical background that is gener-
ally regarded as pertaining to the period of the Warring States, historians have
found it to be of considerable use in reconstructing a given political or social
context. I concur with this view. Much has been written on the reliability of the
Chan-kuo ts™e as a historical source; for a recent appraisal see Michael Loewe,
ed., Early Chinese Texts (Berkeley: Society for the Study of Early China and the
Institute of East Asian Studies, 1993), pp. 1“11.
Chan Kuo Ts™e, 5 (Ch™in 3), 178 and 18 (Chao 1), 608 (Crump, Chan-Kuo Ts™e,
pp. 55, 324).
Loewe, Early Chinese Texts, pp. 342“46.

Emperor Mu exchanged gifts with the foreign chiefs that he encountered,
possibly a common practice in relations between China and northern pas-
toralists during the fourth century b.c. The largest “gift” received by Son-
of-Heaven Mu was in the form of cattle and sheep, numbering in the
thousands, which pointed to the existence (whether he really visited them
or not) of specialized pastoral peoples. Even more valuable as gifts, and
always mentioned at the head of any list, were horses, numbering several
hundreds. Both the Chinese and foreigners alike valued horses, which had
long been used in China for military and ceremonial purposes, but it was
the north that had a surplus for sale to China. A third type of gift, which
appears often but not always, was cereal, such as millet, indicating that at
least some northern peoples relied on agriculture as well as pastoralism;
other “donations” included wine, dogs, and goats.
The gifts presented by Emperor Mu in return were primarily precious
artifacts, such as deer made of silver or gold, a silver bird, necklaces of gold
or precious-stone beads, pearls, gold bullion, belts adorned with precious
shells, and, sometimes, ¬ne horses in a team (four of the same color), prob-
ably meant to be hitched to a royal carriage and used for display.19 A gift
of women was used to seal an alliance, a practice that points to the role of
bride giving as an instrument of diplomacy.20
Can these gifts ¬nd their material counterparts in the recumbent bronze
deer and golden animal-style plaques so common in the late Warring State
nomadic burials? For the time being, we can register only the possible con-
vergence of archaeological ¬ndings and the travels of Emperor Mu, noting
that archaeology seems to corroborate the trade mentioned in the Mu T™ien-
tzu chuan. As romanticized as the story it presents is, this text points to a
growing Chinese awareness of the economy and ecology of the northern
peoples and re¬‚ects a certain amount of knowledge that must have been
current in late Warring States China.
Other scattered references also make it possible for us to discern a trend
in diplomatic and economic exchanges in which pastoral products were
exchanged for high-value items such as silk. In chapter 129 of the Shih chi,
we ¬nd that the land “north of Lung-men and Chieh-shih is rich in horses,
cattle, sheep, felt, furs, tendons and horns.”21 The commercial value of this
land was not lost on the merchant Wu-chih Lo, who became rich at the
time of the First Emperor of Ch™in. He traded silk for domestic animals

This type of gift may indicate more than a passing courtesy. The envoy of the
Hsiung-nu Hsi-hu-ch™en brought two teams of four horses to the Han emperor
as a present. It is possible then that this tradition of diplomatic relations already
existed in the fourth century b.c. See Shih chi 110, 2896.
See R©mi Mathieu, Le Mu Tianzi zhuan: traduction annot©e: ©tude critique (Paris:
Collège de France, Institut des hautes ©tudes chinoises, 1978), pp. 34“36, 158“61.
Shih chi 129, 3254 (B. Watson, Records, 2: 434).


with the king of a Jung tribe, thereby increasing enormously the size of his
own herds of horses and cattle, whose number “could only be estimated by
the valleyful.”22 Who this king was, we do not know, but the story con-
¬rms the existence of high-volume trade in the third century b.c.
Horses were also imported at that time by Chinese states from Tai, a
region referred to as Ti territory whose inhabitants had urban centers and
were organized into “states.” The export of horses by the Jung and the Ti
peoples, to which the sources refer, may either re¬‚ect the domestic pro-
duction of pastoralists or agro-pastoralists or point to a phenomenon of
indirect trade, whereby these tribes were procuring horses indirectly from
nomads farther north and west and selling them to China.
Finally, in addition to other objects of Chinese provenance, such as the
golden plaques inscribed with characters indicating their weight, hoards of
coins have been discovered in the northern territories.23 Coins and gold are
de¬nitely indicative of a complex network of circulation of goods, although
its mechanisms are not yet well understood. Yet it requires no great stretch
of the imagination to assume that precious-metal artifacts and possibly silk
were traded for horses and furs, and that this barter trade was supplemented
by monetary trade.
The question of trade, however, must be considered not only in terms
of exchange but also in terms of access to markets and communication
routes. For instance, Ssu-ma Ch™ien seems to imply that the fourth-century
b.c. expansion of Ch™in into the land of the Jung and Ti was meant not
only to facilitate trade with the foreigners but also to open up direct and
safe routes with the successor states of Chin. This expansion prevented
marginal peoples, such as the Jung and Ti of Tai, from playing the role of
middlemen in commercial communication among northern states and
between northern states and nomads.24 The later Discourses on Salt and
Iron (Yen T™ieh Lun), which possibly took place between 86 and 81 b.c.,
also refers to the prosperous northern trade, which included the importa-
tion of furs and animals (not just horses, but also other domestic animals
such as donkeys and camels) from the Ch™iang and the Hu in exchange
for a few golden trinkets and some cheap silk.25 This argument was used,
in the ¬rst century b.c., to justify the Han Wu-ti expansionist program,
but there is no reason to assume that such a trade had developed only

Shih chi 129, 3260 (B. Watson, Records, 2: 440).
Hoards of coins minted in Yan have been discovered near Ch™ih-feng. See Hsiang
Ch™un-sung, “Nei meng-ku Ch™ih-feng ti-ch™ü fa-hsien te Chan-kuo ch™ien-pi,”
K™ao-ku 1984.2: 138“44 (cit., chap. 2, n. 71).
Shih chi 129, 3261 (B. Watson, Records, 2: 441).
Esson M. Gale, Discourses on Salt and Iron (Leiden: Brill, 1931; rpt. Taipei:
Ch™eng-wen, 1967), p. 14.

If we hypothesize that its trade with the north was based on China™s need
for horses, we must ask why exactly the horses were needed. The most
obvious answer is that horses became a requirement when Chinese states
began to create cavalry units. The adoption of cavalry had economic and
military consequences, therefore, that are widely held to have deeply altered
the relationship between the north and China.


The central piece of textual evidence concerning the adoption of cavalry
in China is a well-known debate held in 307 b.c. at the court of King
Wu-ling of the state of Chao (325“299 b.c.).26 The debate was over the
adoption of cavalry and mounted archers, inspired by the superior riding
expertise of the nomads, which was staunchly supported by a far-sighted
king against the conservative approach of his advisers. The debate presents
a broad picture of relations between Chao and its neighbors. Moreover, it
shows that promotion of this military reform was meant to strengthen the
state against all of its foes rather than serve as a special measure to ¬ght
off raiding nomads.27
The king™s main aim was to turn his own Chinese people into mounted
warriors to be deployed on Chao™s borders with both Chinese and nomadic
states. But it is doubtful that much of his new military machine actually
was used to “contain” the various hu, for at this time the greatest threat
to the existence of Chao as an independent kingdom came from the other
Central States; indeed, within a few decades, Chao was invaded and con-
quered by Ch™in.
The position of King Wu-ling in the debate was by no means exclusively,
or even primarily, a defensive one. In fact, his rhetoric was de¬nitely focused
on the development of offensive military forces. Of the two stated goals in
Chao foreign policy, namely, the completion of the territorial conquests ini-
tiated by King Wu-ling™s predecessor and the establishment of defenses
against Chung-shan, the ¬rst is openly expansionist. The conquest of the
cities of Lin and Kuo-lang had been obtained about thirty years earlier,
when Chao defeated the Lin Hu nomads,28 and the king believed that the
To call it a debate may be misleading. Rather, it was a spirited defence of this
military reform by the king himself, as he was building consensus by request-
ing the approval of his ministers and members of the aristocracy. Shih chi 46,
1806 ff. (Chavannes, M©m. hist. 5: 70“1); Chan Kuo Ts™e 19 (Yen 2), 653“67
(Crump, Chan-Kuo Ts™e, pp. 296“308).
Shih chi 43, 1806 ff.
At this time, in 332 b.c., a “long wall” was erected, which extended westward
from the county of Yu, in the Hsün-hua commandery, to the county of Lan, in
the T™ai-yüan prefecture. Cf. Chavannes, M©m. hist., 5: 64, n. 1.

time had come to consolidate those territorial gains and expand the state
even further. With Machiavellian determination, the king decided to
promote the change, even though people might ridicule him, not because
he wished to entrench himself against raiding nomads, but because the con-
quest of the hu lands and of the state of Chung-shan was integral to his
political project.
Even so, a secondary foreign policy concern of the state of Chao was
indeed defensive, and stated in the following passage from the Chan-kuo
At present to the east of our state the borders lie along the Yellow, the Pao
and the Lo Rivers: we share these boundaries with the states of Ch™i and
Chung-shan, but command not a single boat or oarman upon them. From
Ch™ang-shan to Tai and Shang-tang we border Yen and the Tung Hu in the
east and Lou-fan, Ch™in and Han in the west. Along this line we do not have
a single mounted archer. Therefore I have gathered boats, and sought the
people who live on the rivers in order to guard our boundaries on the three
rivers. I changed our garments to those of the mounted archers to guard our
borders against Yen, the Three Hu, the Lou Fan, Ch™in and Han.29
However, the main foes do not seem to have been exclusively the nomads
but also Chung-shan and the other Chinese states that surrounded Chao:
Yen in the north; the Hu in the east; and the Lin Hu, the Lou Fan, Ch™in,
and Han in the west. It is clear that Chao™s military problems were caused
not just by the nomads but also by the high level of con¬‚ict among all
the states and tribes. Thus, the adoption of cavalry could be presented
as a solution applicable to all of Chao™s military needs, offensive and
Still another argument presented by the king is more an ideological than
a strategic one. As a ¬lial and loyal follower of his ancestors, King Wu-ling
said that he too, like the ancient rulers Chien and Hsiang, who had sought
glory by ¬ghting against the Hu and the Ti, wanted to conquer those
peoples™ lands. By teaching his people to ride and to shoot arrows from
horses, he might well attain such a goal.
The decision to introduce cavalry was controversial, and although the
king claimed to be ahead of his age, he was equally sure that he would be
“exposed to the hatred of the vulgar people.” The criticism he feared came
¬rst from Kung-tzu Ch™eng, his own uncle, who, summoned to court, pre-
dictably accused the king, by diverging from Central States™ customs, of
abandoning the very traditions that had gained the admiration of the Man
and the Yi. “Man” and “Yi” are doubtless terms used here for the “tamed”
foreigners who have been incorporated into the Chinese states, in contrast

Chan Kuo Ts™e, 19 (Yen 2), p. 657 (Crump, Chan-Kuo Ts™e, p. 299). Cf. also Shih
chi 43, 1806.

to the terms “ especially “Hu,” but also “Jung” and “Ti” “ that continued
to be used loosely as ideological references to the “northern foreigners”
who carried the stigma of being hostile. The gist of Ch™eng™s argument,
which carried with it a veiled threat, was that by trying to conquer the for-
eigners the state might loose the support of its own people and allies.
The king™s response was pragmatic and in line with the tradition of
thought that justi¬ed the use of foreigners as “resources” according to the
needs of the moment. He articulated three main points: ¬rst, and most clas-
sically, the king claimed that the way of the sage lay in being able to adapt
to circumstances as they change. Second, he expressed cosmopolitan con-
cerns, arguing that it was not right to condemn what was different in other
peoples or to be automatically suspicious of what one did not understand
about other countries. Third, the king stated that just as he needed boats
on the rivers that separated Chao from Ch™i and Chung-shan, so he also
needed cavalry on the mountainous borders with Yen, the Eastern Hu,
Ch™in, and Hann: it was just a matter of having the right sort of tools. He
made clear his intention to imitate the enterprise of his ancestors, who had
been openly bent on expansion, stating that “my ancestor Hsiang shared
control of Tai with the barbarians for he intended to strike the Hu”30 and
implying that the ¬nal conquest of Chung-shan was also a concern. In the
end, King Wu-ling convinced Kung-tzu Ch™eng with this argument but was
still faced with criticism from other quarters.
These other arguments can be encapsulated in the notion that the ways
of the ancients should not be changed, as revealed in the phrase “proper
clothing is an injunction of propriety.”31 The opposition to the adoption of
Hu garments on the basis of an analogy between garments and behavior
was voiced by yet another advisor:
Hu clothing is not thought well of in the world, and wearing it would not
be something which would instruct people and make the proprieties com-
plete. If the garment is outlandish, the intentions become disordered; when
custom is ¬‚outed, the people become rebellious. So it is that one who rules a
country does not clothe himself in strange garments. The Central States have
no af¬nity for the activities of the Man and Yi peoples, so this action of
yours is not something which teaches the people and makes the proprieties
In his rebuttal, the king availed himself of well-known rhetorical points,
observing that the ancient kings did not all have the same rules, that the
emperors did not imitate one another, and that in the past laws were created
according to the needs of the time and rites were established according to
the circumstances. The king went on to dismiss any analogy between gar-
Chan Kuo Ts™e, 657 (Crump, Chan-Kuo Ts™e, p. 299).
Chan Kuo Ts™e, 660 (Crump, Chan-Kuo Ts™e, p. 300).
Chan Kuo Ts™e, 663 (Crump, Chan-Kuo Ts™e, p. 302).


ments, proper behavior, and historical change; he concluded by hinting
pointedly at the plain ignorance of the literatus concerning military matters,
quoting the saying: “choose a charioteer for his penmanship and he will
never understand your team of horses.” King Wu-ling clearly saw himself
as being in many ways exceptional in his grasp of his age™s special needs,
and he was not shy in responding to them: “a talent for following the ways
of yesterday is not suf¬cient to improve the world of today.”33
Out of this debate, several points can be made. The adoption of cavalry,
as it is narrated in the Chan-kuo ts™e, does not support the conventional
idea that cavalry was meant to repel attacks by nomads. If we take King
Wu-ling™s debating points as a valid representation of his motives, this
reform had a broad agenda oriented primarily toward strengthening the
state. Reasons for cavalry surely entailed the protection of the borders
against all enemies, but the ultimate goal was offensive, intended speci¬-
cally to subdue the nomads and the state of Chung-shan. Made necessary
by deep changes in the circumstances of war and relations with foreign
peoples, the adoption of cavalry was presented as something revolutionary;
but in fact the king of Chao was following an older tradition in the rela-
tions between Chou and northern non-Chou states, that of tapping the
foreigners as resources so as to increase one™s own chances of survival and
victory. Of course, King Wu-ling thereby had exposed himself to the accu-
sation of becoming a “barbarian” himself. This accusation carried serious
consequences, and it can be argued that the principal reason for the debate
was the king™s willingness to show that the rationale for change was not at
odds with proper behavior or with a tradition that supported the strategic
use of the potential for war represented by foreign peoples.
Foreign peoples must have been key to the realization of the reform not
only as a source of innovative military tactics and a model to emulate but
also in more practical terms. Once past the political stage, the reform
needed to be implemented on the ground. How did King Wu-ling accom-
plish that? The adoption of cavalry probably relied heavily on local recruits,
people who had lived on the frontier for a long time and were acquainted
with the nomads, had traded with them, and had the warlike disposition
that would make them well suited to military tasks. The Shih chi provides
signi¬cant textual evidence in support of this hypothesis:
Yang and P™ing-yang have customarily traded with the area of Ch™in and the
Ti in the west, and with Chung and Tai in the north. Chung and Tai are sit-
uated north of [the old city of] Shih. They border the land of the nomads
(hu) and are subject to frequent raids. The inhabitants are proud and stub-
born, high-spirited and fond of feats of daring and evil, and do not engage
in agriculture or trade. Because the region is so close to the territory of the
northern foreigners (pei yi), armies have been frequently sent there, and when

Chan Kuo Ts™e, 663 (Crump, Chan-Kuo Ts™e, p. 303).

supplies were transported to them from the central states, there was a surplus
[for the local people to live on]. The temperament of the local people is like
that of wild sheep, they are nimble and clever, but not uniform (in their behav-
ior). From the time when the state of Chin was still undivided they were
already a source of trouble because of their violent temperament. Yet, king
Wu-ling pro¬ted by disciplining them, and their folksongs and ways are still
the same as the customs they developed when they were under the rule of
Chao. Therefore [the merchants of] Yang and P™ing-yang carry out their busi-
ness quickly in their areas and obtain whatever they want.34
This passage implies that the peoples of Chung and Tai became a buffer
against the Hu. It also suggests that they were commercially “colonized”
by the merchants of Yang and P™ing-yang, who carried out a pro¬table trade
between Chung and Tai and the western regions of Ch™in and other non-
Chou peoples. But it was in Wu-ling™s time that they were incorporated
within “China,” henceforth constituting an economic reservoir for pastoral
products and a buffer against hostile nomads. Together with other north-
ern pastoral zones, these areas continued to be inhabited by “Jung and Ti,”
whose herds, according to Ssu-ma Ch™ien, were “one of the riches of the
empire.”35 The commercialization of the frontier, the increased strategic
interest in the north, and the need for horses are the preconditions leading
to the appearance of extensive forti¬cations in northern China toward the
end of the Warring States period, an event that radically changed the
concept of the frontier and opened an entirely new phase in the relation-
ship between China and its northern neighbors.

Wall Building

Wall building did not originate in the north, but was a military concept
and a technology imported from within the Central Plain, although its
function may have changed once it was transferred to another context.
Ssu-ma Ch™ien attributes the construction of the earliest walls to King
Hsüan of Ch™i, who surveyed the land and then built a wall on hilltops
and mountain peaks; in the east this wall reached the sea, and in the west
it reached Chi-chou; it was a thousand li long and was meant to protect
Ch™i from the state of Ch™u.36 The Chu-shu chi-nien (Annals of the Bamboo
Books) mentions an attack against Ch™i during the reign of King

Shih chi 129, 3263 (B. Watson, Records, 2: 442).
B. Watson, Records, 2: 441.
For an introduction to the early walls, see Ou Yen, “Wo kuo ch™ang-ch™eng te
k™ao-ku fa-hsien yü yen-chiu,” in Ch™ang-ch™eng kuo-chi hsüeh-shu yen-t™ao-hui,
pp. 250“63.


Wei-lieh in which Chou forces penetrated a long wall (ch™ang yüan).37 The
walls built by Ch™i seem to have been essentially defensive, and to have
marked state boundaries.38 Walls were built not only by the Chinese states.
The state of Chung-shan and even some Jung peoples built walls against
their enemies. In the case of the Yi-ch™ü Jung, the enemy was the state
of Ch™in.
These walls effectively inaugurated a new type of defensive system, pos-
sibly as a consequence of the creation of infantry armies whose maneuver-
ability required more extensive use of, and control over, natural features.
The “long” walls appear also to have been strategic forti¬cations aimed at
asserting a state™s political and military control over a given area. These
areas could be strategically important not just defensively but also offen-
sively, inasmuch as control over a mountain pass or a river ford could either
block an advancing hostile army or secure passage to one™s own troops.
Moreover, like roads, walls provide the logistic infrastructure to facilitate
communication and transportation, vital elements for armies employed in
the occupation or invasion of a foreign territory. Thus military walls could
be an integral part of an expansionist “offensive” project, and we can assess
their function properly only in their historical context.
Another factor that may have favored the building of long walls among
Chinese states was the reduced space for territorial expansion. Whereas
during the Spring and Autumn period the stronger states could expand rel-
atively easily at the expense of non-Chou communities and weaker Chou
states, by the fourth century b.c. the competition had become limited to a
few powerful states, thus increasing the pressure to improve both defensive
and offensive capabilities.
In sharp contrast to the all-embracing cultural and political metaphor
that the “Great Wall of Ten Thousand Li” has become in modern times,
we know little about the genesis and earliest development of northern-
border walled forti¬cations in China. Arthur Waldron has helped to
demythologize the Great Wall as an eternal ¬xture of Chinese civilization,
the product of imperial uni¬cation under the First Emperor of Ch™in, who
joined the pre-existing northern walls, creating a barrier that supposedly
would keep out the nomads. Prior to 221 b.c., the warring states of
China had erected “long walls” (ch™ang ch™eng) both in the Central Plain
and in the north; yet why these walls were erected toward the end of the
Warring States period and how they functioned are still, in my view, open

Chu-shu ch-nien, 2.26b (James Legge, The Chinese Classics, vol. 3: The Shoo
King, 2nd ed. [Oxford: Clarendon, 1985; rtp. Hong Kong: Hong Kong Univer-
sity Press, 1960], “Prolegomena,” p. 169).
Legge, The Shoo King, “Prolegomena,” p. 173.


In the standard interpretation of the genesis of the walls, “long walls”
started to be built between Chinese states sometime during the Warring
States period. The states of Ch™i and Ch™u were particularly precocious, fol-
lowed by the state of Wei. In the north walls were built by the states of
Yen, Chao, and Ch™in between the end of the fourth and the ¬rst half of
the third century b.c., as a defensive measure against nomadic peoples who
were encroaching upon their territory. This measure proved effective, to the
point that the northern states could keep the nomads at bay while concen-
trating on the struggle among themselves. Upon the uni¬cation of the
empire, Ch™in Shih Huang-ti demolished the internal walls and uni¬ed
the external ones, thus creating a continuous barrier between China and
the northern nomadic peoples.
Yet archaeological data on the actual location of the walls and the cul-
tural environment of the areas in which they were erected raise substantial
doubts about the various binary constructions typically used to explain the
historical function of the walls. The walls have been chie¬‚y interpreted as
the product of tensions between the warring states of China and a politi-
cally amorphous north, between agriculture and pastoralism, or, to put not
too ¬ne a point on it, between civilized peoples and barbarians. The study
of the early walls inevitably suffers from assumptions that arose in the
Chinese historical consciousness long after these walls were built, especially
during the Ming dynasty, when the wall actually was a military bastion
intended to ward off further Mongol invasions.
The signi¬cance of “long walls” as a factor in the history of the north-
ern frontier has to be assessed based on the evidential texture provided by
the intersecting threads of the history of wall building and of the actual cul-
tural and political contexts in which the walls were built. In this section, I
intend to examine the evidence in support of the thesis that the early walls
were meant to contain invading nomads, considering it in the light not only
of the walls™ own history but also of the broader historical context. The
result, as we will see, leads to conclusions quite different from the conven-
tional wisdom.

The Written Evidence

References to wall-building activity by the northern states are found in
Chinese historical sources, in particular the Shih chi. As we know, this para-
mount monument of Chinese historical literature was written by Ssu-ma
T™an and his son Ssu-ma Ch™ien about two hundred years after the ¬rst
walls were built on the northern frontier. At this time the uni¬ed nomadic
empire of the Hsiung-nu had created, beginning in 209 b.c., a true politi-
cal and military crisis that the Han dynasty was able to overcome only with
the greatest dif¬culty. Without going into the details of Ssu-ma Ch™ien™s
Location of the N

"Long Walls" built by TUNG HU

200 mi
Ch™in, Chao, and Yen

o R iv e
0 200 km
Extant ruins

Fragmentary ruins or
presumed extension 11
L 13

1. Ti-tao (today: Lin-t™ao) WA L L

2. Lin-t™ao (today: Min-hsien)

3. Ching-pien CHAO W
10 ALL
4. T™o-k™e-t™o 9

5. Hsing-he 5
8 R.
6. Chi-ning T a-h e Yün-chung

7. Chuo-tzu 4
8. Pao-t™ou

9. Huhhot

10. Wu-la-t™e-ch™ien-ch™i

11. Ch™ih-feng

Gulf of

12. Chien-p™ing
R iv


Po - Ha i
13. Fu-hsin
lo w

14. Karachin Banner

Y el





Fen Ri




g River

Lo R i

Ye l l o w

W ei R

Map 3

depiction of the Han“Hsiung-nu confrontation,39 it should be noted that
the historian inscribed such a confrontation in a historical model accord-
ing to which “Chinese” (Hua, Hsia, Chung-yuan, Chung-kuo, etc.) and
“nomads” (Jung, Ti, Hu, Hsiung-nu, etc.) constituted antithetic poles at
odds since the dawn of Chinese history. This polarity between a uni¬ed
north and a uni¬ed south was then projected back into the past.
Yet Ssu-ma Ch™ien™s account of the late Warring States cannot be
regarded as pure ideology. He recorded names and events whose multi-
plicity and unfolding are evidence of the political and ethnic complexity of
both the north, where different groups appear to have been living, and the
south, where, of course, Chinese states were still vying for power. Hence,
although it is vital to remember that Ssu-ma Ch™ien™s narrative of the north-
ern frontier is not itself neutral, one cannot use this argument to dismiss
tout court his narrative of Warring States history, which was based, we
presume, on an extensive knowledge of sources available at the time. With
this critical caveat in mind, I should add that historians have never truly
questioned the reliability of the Shih chi when it comes to the Great Wall.
The question I am especially concerned with is whether the Shih chi actu-
ally supports an interpretation according to which the walls were estab-
lished as a military defense.
The state of Ch™in began to build walls on its northwestern border under
King Chao-hsiang (306“251 b.c.). The pretext for Ch™in™s expansion is
attributed to a “scandalous” series of events. Apparently, the king of the
Yi-ch™ü Jung had illicit intercourse with the Queen Dowager Hsüan, who
bore him two sons. The queen dowager later deceived and killed the king
of the Yi-ch™ü Jung; then she assembled the army and attacked and
destroyed the Yi-ch™ü. After conquering the Jung, Ch™in also expanded to
the north into the territory within the Yellow River™s Great Bend (today™s
Ordos region). In this way, Ch™in acquired the northern commanderies of
Lung-hsi, Pei-ti, and Shang. At this point, the sources say, Ch™in “built a
long wall (ch™ang ch™eng) to guard against the Hu [i.e., nomads].” Ch™in™s
defense line ran from northeast to southwest, extending from the eastern
part of the Yellow River™s “loop” (Shang commandery) to the southern part
of Kansu (Lung-hsi commandery).
The state of Yen expanded mainly in the northeast and occupied
both the maritime region north of the Liao-tung Gulf and the Liao-tung
Peninsula, including, to the west, a large portion of what is today Hopei
province. During the reign of King Chao (311“279 b.c.), General Ch™in
K™ai, who had served as a hostage among the nomads, launched a
surprise attack against the Eastern Hu, who had placed their trust in him,
defeated them, and forced them to retreat “a thousand li.” Yen then “built
extended forti¬cations (that is, the Great Wall, ch™ang ch™eng) from Tsao-

This topic is treated at length in Chapter 8.

yang to Hsiang-p™ing, and established the commanderies of Shang-ku, Yü-
yang, Yü-pei-p™ing, Liao-hsi, and Liao-tung in order to resist (chü) the
The state of Chao built the northernmost line of forti¬cations under King
Wu-ling, and possibly even earlier. Let us review the events as they appear
in the Shih chi. King Wu-ling “in the north attacked the Lin Hu and the
Lou-fan; built long walls, and made a barrier [stretching] from Tai along
the foot of the Yin Mountains to Kao-ch™üeh. He then established the three
commanderies of Yen-men, Yün-chung, and Tai.”41 After the conquest of
Chung-shan in 295 b.c., Chao continued its drive north, advancing into
today™s Inner Mongolia and building a series of forti¬cations to the north
of the Great Bend of the Yellow River in the Ho-t™ao region, where it encir-
cles the Ordos Steppe in a wide loop, thus creating the most advanced
Chinese forti¬ed front deep into nomadic territory. The Shih chi also
informs us that, as a consequence of the northern and western expansion
of the states of Ch™in, Chao, and Yen, the “central states” came to border
directly on the territory inhabited by the Hsiung-nu.42
This is the core textual evidence used by scholars to argue that the
northern walls had a defensive purpose and had been erected as protec-
tion against nomadic attacks. However, none of these statements actually
says that walls were constructed as a response to nomadic attacks on
Chinese people. What they say is that the walls were built “to repel” or “to
contain” the nomads after a substantial drive into foreign lands. The build-
ing of forti¬cations proceeded in tandem with the acquisition of new terri-
tory and the establishment of new administrative units. In other words, the
states of Ch™in, Chao, and Yen needed to protect themselves from the
nomads only after they had taken large portions of territory from other
peoples and had chased the nomads away from their homelands. Surely at
some point the forti¬cations did acquire a “defensive” function, but the
context suggests strongly that this defensive role was subordinated to a
grander strategy, one that was militarily offensive and territorially expan-
sionist, pursued by all three Chinese states. To examine the context more
closely, we need to zoom in on the wall itself, on its construction and its

Technical Features of the Northern Walls

Chinese scholarship has been engaged for centuries in what we may regard
as an antiquarian interest in the Great Wall. Issues of toponomastics, con-
centrating on the solution of textual problems concerning the exact loca-
tion of certain portions of the Great Wall, continue to claim a central place

40 41 42
Han shu, 3748. Shih chi 110, 2885. Shih chi 110, 2885“2886.

in modern scholarship, although these discussions have recently been based
also on evidence derived from archaeological ¬eldwork.43
A survey of the early walls built by Ch™in, Yen, and Chao indicates that,
together with their common features, mostly related to engineering, the
walls also show differences related to topography, strategic choices, and
political relations between the states and the peoples who lived around and
beyond the walls. Naturally, the varying intensity and success of the archae-
ological surveys in the interested regions also accounts for discrepancies.
(The wall of the state of Ch™in, for instance, has been more extensively
researched than the walls of Chao and Yan.)

common characteristics. As already mentioned, the northern walls
were not the ¬rst walls built by Chinese states. The earliest walls had been
built by the states of Wei and Ch™u and are documented in the written
records and in archaeology.44 Hence, the speci¬c technology concerning the
building of these walls was shared, in the fourth century b.c., by all states,
and engineers were available to design the walls and supervise their
The most important common feature of these “walls” was that they con-
stituted an integrated system of forti¬cations that included not only man-
ufactured structures but also natural barriers. These lines of forti¬cations
made extensive use of the natural features of the surrounding topography.
In mountainous terrain, along precipices and ravines or narrow gullies, the
arti¬cial structures may have been limited to a few lookout posts or to stone
walls blocking a mountain pass. Across a river ¬‚oodplain, rolling grass-
lands, or low hills, the walls were invariably built with mixed stones and
earth pressed together. The “walls” also comprised ramparts and ditches,
small and large forts, beacon towers, lookout platforms, watchtowers, and
other structures. Typically, the walls were made of stamped earth and
stones, dug out from the outer side and piled up to form the wall itself,
usually on sloping terrain, so that the ground on the inner side would be
considerably higher than that on the outer side. Moreover, along these walls

For an extensive study of the archaeological context of the Great Wall in Inner
Mongolia during the Warring States, Ch™in, and Han periods, see Kai Shan-lin,
Lu ssu-hsien, “Nei Meng-ku ching-nei Chan-kuo Ch™in, Han ch™ang-ch™eng
yi-chi,” in Chung-kuo k™ao-ku hsüeh-hui ti-yi-tz™u nien-hui lun-wen-chi 1979
(Peking: Wen-wu, 1980), pp. 212“28. For the “Great Wall” region in Ning-hsia
during the same period, see “Ning-hsia ching-nei Chan-kuo Ch™in, Han ch™ang-
ch™eng yi-chi,” in Chung-kuo ch™ang ch™eng yi-chi tiao-ch™a pao-kao chi,
pp. 45“51.
For a selection of references to the Great Wall in written sources, see AA. VV.,
“Ch™ang ch™eng wen-hsien tzu-liao chi-lüeh,” in Chung-kuo ch™ang ch™eng yi-chi
tiao-ch™a pao-kao chi (Peking: Wen-wu, 1981), pp. 119“37.

the archaeologists have found, at regular or sometimes irregular intervals,
mounds of stamped earth that they assume are to be the remains of ele-
vated platforms or towers. On higher ground, such as hilltops or even
mountain peaks, small stone structures have been found, in the shape of
platforms, which are thought to have served as lookout posts.
On the inner side of the walls, at varying distances and intervals, we ¬nd
a number of constructions, in the shape of square or rectangular enclosures,
often made of stone. Such enclosures are taken to have been forts garrisoned
by soldiers, and the largest among them are assumed to have been the local
command stations, where high-ranking of¬cers resided with their troops.
Roads internal to the walls served to connect the various garrisons with
strategically important areas. Beacon towers, also situated on the inner side
of the walls, were used to communicate between the various stations.
Although much of what we say about the organization of the troops along
the walls is speculative, the number of structures and their spatial exten-
sion nevertheless suggest that the ef¬cient use of the walls required a
massive military presence, as well as a ¬nely networked system of couriers,
postal stations, and checkpoints.

the ch™in wall. Investigations of the early Ch™in wall started with Ku
Chieh-kang,45 who visited in the 1930s the southwestern location of the
walls, in the Min-hsien and Lin-t™ao districts of Kansu. Archaeological
investigations were carried out in the 1970s and 1980s and important
contributions to the knowledge of the Ch™in wall were published by Shih
Nien-hai and others. This wall is the best known among the early north-
ern walls, and its location has been de¬ned with accuracy, notwithstanding
some lingering controversy concerning the southwestern terminus. Its total
length is estimated at 1,775 kilometers, and it actually consists of several
separate “walls.” Starting in southern Kansu, the line of the wall cuts
through southern Ning-hsia, enters eastern Kansu, and proceeds into
northern Shansi, ending its run on the bank of the Yellow River in Inner
Mongolia (Zhunggar Banner). An important branch line extends eastward
for 225 kilometers from south of Ching-pien, reaching almost to the
Yellow River.
Three points about the walls™ construction should be emphasized. First,
40 percent of the total line of the wall was built on sloping terrain: the
higher ground was kept on the inside, a moat was dug on the external (and
lower) side, and soil and stones were piled up into a wall. Second,

On the Ch™in wall, see Shih Nien-hai, “Huang-he chung-yu Chan-kuo chi Ch™in
shih chu ch™ang-ch™eng yi-chi te t™an-su,” in Chung-kuo ch™ang ch™eng yi-chi tiao-
ch™ pao-kao chi, pp. 52“67; id., O-erh-to-ssu kao-yüan tung-pu Chan-kuo shih-
ch™i Ch™in ch™ang ch™eng yi-yi t™an-suo chi,” in Chung-kuo ch™ang-ch™eng yi-chi
tiao-ch™a pao-kao chi, pp. 68“75.

forti¬cations have been found near ravines, crevices, and rivers encased in
steep gullies. We do not ¬nd actual walls made of stamped earth, but ele-
vated platforms and occasionally a fort blocking the entrance to a valley.
This type of system accounts for 20 percent of the total length. Third, walls
made of stamped earth are seen across plateaus and ¬‚at or mildly rolling
plains. For construction material the builders used what was available at a
location “ earth, stone, or a combination of both.
On top of the walls, for its entire length, we ¬nd three to four “mounds”
(elevated platforms) per kilometer, amounting to a total of approximately
6,300 mounds. Possibly, these were lookout towers and served also as lodg-
ings for soldiers. Throughout the line of the walls, on the inner side, we
encounter the ruins of forts, round watchtowers, and beacon towers. Some
of these towers are especially high. Citadels and forts are distributed at a
distance of three to ¬ve kilometers from each other, and their internal area
may vary from 3,500 to 10,000 square meters. They are generally walled,
though forts built on steep ravines and gullies do not have walls, as the
natural topography provided suf¬cient protection. Beacon towers are
located at a distance of one to two kilometers from the wall, on the inner
side, and two to three kilometers from each other. In some areas watch-
towers (or beacon towers) are built on high watershed crests at a distance
of three to ¬ve kilometers. The dating of these constructions has been based
on the style of their terracotta tiles; for example, the “cloud” decoration is
typical of the Warring States period. Along the desert and hills of northern
Shansi, researchers have found several roads.
The study of the toponomastics of the walls, based on the historical texts,
and in particular on the Shih chi, has given rise to a series of controversies.
The longest dispute involved the western terminus of the wall, as opinions
were divided between the Min-hsien and the Lin-t™ao districts in Kansu. The
majority opinion now is that, regardless of contrary evidence in the Shih
chi, the walls actually started in Lin-t™ao (then called Ti-tao).46 Shih
Nien-hai, after a detailed examination of the course of the Ch™in wall, re-
marked on its strategic nature. He noticed that the meandering and tortu-
ous line traced by the wall was due to the need to remain on higher ground.
Therefore, its twists and turns follow closely the relief map of the region.
In some cases the wall extends along rivers, some of which were regarded
as military lines of communication. According to Shih, the Ch™in wall ended

A contrary opinion is expressed by Shih Nien-hai, who follows the tradition in
believing that the old Ch™in wall started in Min-hsien because this was the ancient
location of the Lin-t™ao mentioned by Ssu-ma Ch™ien. He explains the ruins of
the wall found in modern Lin-t™ao as belonging to the prefectural capital
Lung-hsi, established early on to administer this territory. See Shih Nien-hai,
“Huang-he chung-yu Chan-kuo chi Ch™in shih chu ch™ang-ch™eng yi-chi te t™an-
su,” p. 60.


in the neighborhood of today™s Shih-erh-lien-ch™eng, in the T™o-k™e-t™o dis-
trict (Inner Mongolia), because this was a ford of the Yellow River of strate-
gic importance, the point at which the Ta-hei River joins the Yellow River.
Following the Ta-hei one could have access to the foothills of the Yin-shan
mountains and from there could reach every part of the Mongolian grass-
land. King Wu-ling of Chao had established the Yün-chung commandery
exactly on the river Ta-hei. Hence, it was important for Ch™in to control
the ford of the Yellow River to guard against possible Chao incursions.
Although Shih remained faithful to the theory that the wall™s purpose was
to defend Ch™in against the nomads, he introduced here the notion that
some of the tracts of northern walls were actually meant either to facilitate
access to the steppes or to protect Ch™in from other Chinese states.

the chao wall. Investigation of the Chao wall began in the 1960s.
Relying on the historical records just mentioned, Chinese archaeologists
trace the beginning of the construction of the Chao wall to the reign of
King Wu-ling, although they cannot provide an exact date. As for its
location, the consensus is that the eastern terminus of the Chao wall is in
Hsing-ho, in Inner Mongolia; then, proceeding slightly to the northwest,
the wall enters the districts of Chi-ning and Chuo-tzu, then continues west
passing north of Huhhot, and ¬nally reaches Pao-t™ou. The line of the wall
follows the foothills of the Ta-ch™ing and Wu-la Mountains, in the Yen-shan
mountain chain. Several long stretches of the walls, especially near the
Ta-ch™ing Mountains, can still be seen very clearly for several scores of
A secondary, northern, line, not as well researched, begins north of
Huhhot and ends north of the northwestern corner of the Yellow River, at
the point where the river turns east. This terminus is supposed to have been
the location of the fortress of Kao-ch™üeh, which appears in the sources.
Along this southern tract, between Huhhot and Pao-t™ou, the wall is made
of rammed earth, while some parts are made of stone. The average height
is 3.5 to 4 meters. On its southern side one can see ruins of small citadels
and beacon towers. Researchers are still debating the eastern and western
ends of this line of forti¬cations, the western terminus being especially prob-
lematic, but it is possible that some ruins discovered in Wu-la-t™e-ch™ien-
ch™i, west of Pao-t™ou, mark the western end.

the yen wall. The Yen wall (again consisting of several separate lines of
“walls”) has been researched from the mid-1970s, but its entire extension
has not been clearly de¬ned yet, especially in its western part.47 These walls,

References to the Yen walls include: Liu Chih-yi, “Chan-kuo Yen pei ch™ang
ch™eng tiao-ch™a” [An Examination of the Great Wall North of Yan during the
Warring States], in Nei Men-ku wen-hua kaogu 1994.1: 51“53, 68; Hsiang


usually attributed to the late fourth century b.c., stretch from northern
Hupei across the Jao Uda League in Inner Mongolia, and seemingly end in
Liao-tung. The central portion, running approximately from the Karachin
East Wing Banner in the west to the Fu-hsin district in the east, has been
investigated more closely. Here we ¬nd three roughly parallel lines of for-
ti¬cations: one running north of Ch™ih-feng, which is attributed to the Ch™in
period; one to the south of this, running south of Ch™ih-feng, which is
regarded as the northern Yen wall; and a third line further south, which is
dated to the Han period. The eastern and western termini have not been
conclusively ascertained yet, but a line of walls has been found as far as
Liao-tung, crossing into north Korea, and tentatively attributed to the state
of Yen. In the west, isolated ruins of the walls are present in various loca-
tions, and it is possible that the Yen western terminus may have been close
to the Chao walls.
The construction of the Yen “walls” does not differ from that of other
Warring States walls, utilizing mainly stamped earth and stone. The stone
walls have a width of about four meters. At the base they are mostly built
on hills and high mountain peaks. The earthen ones are slightly wider at
the bottom (approximately ¬ve meters), are made of black soil, and have
an external ditch. The main branch of the wall, from the Karachin Banner
to Fu-hsin, has a length of 200 kilometers. Along this line researchers have
discovered sixteen forts of various sizes. The largest ones appear on both
banks of the Lao-ha River. One of them is surrounded by earthen walls of
a length of 320 meters from east to west and 260 meters from north to
south. It is assumed that these installations were used to garrison a large
number of troops. The smaller forts have walls of 30 to 40 meters in length
on each side and are built on high places. These were probably smaller
checkpoint stations or posts for patrolling troops. Two round stone plat-
forms have also been found on mountain tops. The distribution of these
auxiliary forti¬cations is not as regular as it is for the Ch™in wall; more-
over, no beacon towers have been found. Based on tile decorations and the
presence of knife coins, pottery, and iron objects, the dating of these walls
attributes them to the late warring states period.
A most interesting feature of these walls concerns the discovery, along
the line of the walls, on high terrain, of several citadels and round habita-
tions built in stone from where archaeologists have recovered artifacts
attributed to the Upper Hsia-chia-tien culture. The archaeologists believe

Ch™un-sung, “Chao-wu-ta-meng Yen Ch™in ch™ang ch™eng yi-chih tiao-ch™a pao-
kao,” in Chung-kuo ch™ang-ch™eng yi-chi tiao-ch™a pao-kao chi, pp. 6“20; Chao
Hua-ch™eng, “Chung-kuo tsao-ch™i ch™ang-ch™eng te k™ao-ku tiao-ch™a yü
yen-chiu,” in Ch™ang-ch™eng kuo-chi hsüeh-shu yen-t™ao-hui [Proceedings of the
International Academic Symposium on the Great Wall], ed. Chung-kuo ch™ang-
ch™eng hsüeh-hui (Chi-lin-shih: Chi-lin Jen-min, 1994), pp. 240“41.


that the original dwellers may have been Tung Hu, that is, a non-Chinese
nomadic group that the written sources place in the northeast and against
whom the state of Yen fought, “pushing them 1,000 li away.” In other
words, these dwellings may have belonged to people who ¬‚ed the area after
Yen attacked them. At the same time, however, the investigators also
con¬rm that both “outside” and “inside” this line of forti¬cations the only
cultural remains are “non-Chinese.”48 A more likely hypothesis is that these
remains belonged to local people who may have been subject to Yen. From
the location of their settlements it cannot be excluded that these people also
performed a military service for Yen, having been either recruited or con-
scripted into the Yen army.
Their study of the walls™ location has led Chinese archaeologists to
believe that the walls did not mark an ecological boundary,49 that is, the
walls were not built to separate steppe and sown, nomad and farmer. In
my opinion, they seem most consistently to have been built (far from the
political and economic centers of each state) with a tactical goal in mind.
This goal was to establish a strong military presence that allowed the state
to control the movement of people, be they nomads, moving across plains,
hills, or mountain passes; peddling merchants; transhumant populations; or
hostile armies. Naturally, the walls also served a defensive purpose, but that
purpose must be seen in the light of what they were actually defending.
From the location of the walls, many miles from farming settlements that
could be related to either Ch™in, Chao, or Yen, and often built in a terri-
tory that could not support agriculture, we can safely exclude that the pro-
tection of Central States™ agriculture was a strategic goal. It is likely then
that the walls were erected to defend the surrounding non-agricultural ter-
ritory and to establish lines of communication and facilitate the movements
of troops as they patrolled this territory, having occupied it by forcing the
local population to submit and driving away recalcitrant nomadic groups.
This military presence may also have facilitated colonization of these areas
by immigrants from the core regions of the states. This point comes into
sharper relief as we examine the “cultural” make-up of the regions fenced
in by the walls. Thus we must study the extent of the cultural presence of
either Chinese (i.e., Central Plain) or non-Chinese (i.e., Northern Zone) cul-
tures, which necessarily relies on the archaeological context, to ascertain
the precise function of the walls.

See in particular Hsiang Ch™un-sung, “Chao-wu-ta-meng Yen Ch™in ch™ang ch™eng
yi-chih tiao-ch™a pao-kao,” in Chung-kuo ch™ang-ch™eng yi-chi tiao-ch™a pao-kao
chi, pp. 6“20.
Tang Xiaofeng, “A Report on the Investigation on the Great Wall of the Qin-Han
period in the Northwest Sector of Inner Mongolia.” Trans. in Chinese Archaeo-
logical Abstracts, vol. 3: Eastern Zhou to Han, ed. Albert E. Dien et al. (Los
Angeles: Institute of Archaeology, 1985), pp. 959“65.


The Archaeological Context

The northern walls were built in the middle of large stretches of grassland.
From the cultural remains recovered from these areas, with the exception
of some Chinese coins and other objects left in the forts by military per-
sonnel, it is obvious that the whole area of the forti¬cations was inhabited
exclusively by non-Chinese, mostly pastoral, people.50 Here I will present
some examples of the material culture of the walls™ regions, based on the
Northern Zone sites discussed earlier (see Chapter 2, especially map 2, for
the location of archaeological sites), that I believe to be close indicators of
the broader context.
Archaeological excavations in the proximity of the section of the Yen
wall near Ch™ih-feng reveal the presence, in the burial ground of T™ieh-
chiang-kou (Ao-han Banner), of typological elements that place it within
the general spectrum of the Northern Zone cultural area, such as the Upper
Hsia-chia-tien and the so-called Ordos bronze cultures. A large number of
bronze objects, such as knives with ringed handles, horse- and bird-motif
ornaments, bell ornaments, buttons, earrings, and belt hooks, place this
area in a cultural context that is fully outside the Central Plain sphere. The
basis for dating the burial place of T™ieh-chiang-kou to the Warring States
period, and to no later than 299 b.c. (date of the construction of the wall
according to the written sources), rests solely on the assumption that no
non-Chinese people could have remained there after the walls had been
As for the Ch™in wall, its region includes some of the most important
early nomadic sites of the Northern Zone cultural complex, sites that are
dated to the latter part of the Warring States. The westernmost line of walls
built by Ch™in is a north-to-south stretch between the districts of Lin-t™ao
and Min-hsien (Kansu). Proceeding in a northeasterly direction, the Ch™in
walls enter Ning-hsia, and then reach just to the north of Ku-yüan. If we
consider that the territory of the Yi-ch™ü, in Shensi, was located to the east
of Ku-yüan, we must assume that the wall was built after Ch™in™s subjuga-
tion of this people. Archaeologically, this area is extremely rich in sites that
show the continuous habitation of people whose culture and even physical

For an investigation of the “Great Wall” territory in Inner Mongolia and its arche-
ological cultures at an early time, see T™ien Kuang-chin, “Nei Meng-ku ch™ang-
ch™eng ti-tai chu k™ao-ku-hsüeh wen-hua yü lin-ching t™ung-ch™i wen-hua
hsiang-hu ying-hsiang kui-lü te yen-chiu,” Nei Meng-ku wen-wu k™ao-ku
1993.1“2, pp. 16“22. A work of similar scope is Han Chia-ku, “Lun ch™ien
ch™ang-ch™eng wen-hua tai chi ch™i hsing-ch™eng [On the Pre-Wall Cultural Zone
and Its Formation], in Ch™ang-ch™eng kuo-chi hsüeh-shu yen-t™ao-hui, pp. 60“72,


characteristics “ according to Chinese archaeologists “ were different from
those of the people of the Central Plain.
The features of the Ku-yüan culture re¬‚ect, from the late Spring and
Autumn period down to the late Warring States period, a fairly homoge-
neous cultural complex characterized by a bronze culture similar to that of
the Ordos region. Among the culture™s bronze objects, we ¬nd the charac-
teristic animal-style ornamental plaques, horse and chariot ¬ttings, and
northern-style weapons.51 Ku-yüan sites show clear connections with sites
in Ch™ing-yang county (Kansu), which also show cultural af¬nity with the
Northern Zone complex. For instance, at the site of Hung-yen, a grave and
a horse-sacri¬ce pit were found containing bronze artifacts that have clear
Ordos matches at sites such as T™ao-hung-pa-la, Hu-lu-ssu-tai, Hsi-ch™a-
kou, and Yü-lung-t™ai.52 To the west of Ku-yüan, in Ch™in-an county
(Kansu), a site was discovered, also dated to the Warring States, that is
unquestionably Northern Zone, and also connected with the Ordos sites of
Tao-hung-pa-la, Hsi-kou-p™an, and Fan-chia-yao-tzu. Here we even ¬nd a
fu, a sacri¬cial bronze cauldron of the type normally associated with north-
ern nomads, including the Hsiung-nu. The sites in the distribution area of
Ku-yüan, Ch™ing-yang, and Ch™in-an counties are all “internal” to the Ch™in
wall, and prove that the region to the northwest of Ch™in was home to pas-
toral-martial cultures, possibly semi-nomadic or even fully nomadic, but
certainly non-Chou. The Chinese presence in the area at this early time can
be detected only at sites connected with the wall forti¬cations themselves,
showing the presence of Chinese troops in an otherwise alien cultural
The site of Su-chi-kou, in the Zhunggar Banner, is located between the
Ch™in wall and the Chao wall in the north. This Hsiung-nu burial site of
the late Warring States period is particularly astonishing for the accumula-
tion of animal-style pole tops that have been found there. These pole tops
include six with small deer statues (four standing, two kneeling), two with
crane heads, one with a sheep head, one with a lion cub, two with kneel-
ing horses, one with a wolf head, and one that is tubular. Other objects
recovered include two disk ornaments with knobs and four bells.53
Although this site was ¬rst assigned to the Han period, it is now believed
to be a late Warring States site of the fourth“third century b.c. The style of
its bronzes has been compared to that of the Hsiung-nu bronzes from the

Lo Feng and Han Kung-le, “Ning-hsia Ku-yüan chin-nien fa-hsien te pei-fang-hsi
ch™ing-t™ung-ch™i,” K™ao-ku 1990.5: 403“18.
Liu Te-chen, and Hsü Chin-chen, “Kan-su Ch™ing-yang Ch™un-ch™iu Chan-kuo
mu-tsang te ch™ing-li,” K™ao-ku 1988.5: 413“24.
Kai Shan-lin, “Nei Meng-ku tzu-chih-ch™ü Chun-ke-erh-ch™i Su-chi-kou ch™u-t™u
yi-pi t™ung-ch™i,” Wen-wu 1965.2: 44“45.


site at Yü-lung-t™ai, also in the Zhunggar Banner. This latter site was also
initially dated to the Former Han dynasty and later to the Warring States
period.54 Of extreme interest is that the Ch™in Great Wall makes a large
curve around the territory of the Zhunggar Banner and, therefore, fully
incorporates these sites, which are among the best-known Hsiung-nu sites
of this period.
Another Warring States Hsiung-nu site is Hu-lu-ssu-t™ai. Its location and
period place it just to the north of the westernmost extension of Chao™s
northern line of forti¬cation.55 Several features in its bronze inventory,
including knives, axes, arrowheads, and ornamental plaques, link the met-
alwork of this site with that found at Yü-lung-t™ai and Su-chi-kou, showing
that throughout the whole area of the Ch™in and Chao walls in the north-
ern part of the Ordos region and beyond the Yellow River there was a rel-
atively homogenous non-Chinese culture. Archaeologists attribute this
culture to the Hsiung-nu, but it is more correctly de¬ned simply as “early
nomadic.” Certainly the walls had not been built “in between” Chinese and
nomads, but instead ran through an alien land inhabited by alien groups.
Some of these groups were incorporated within the perimeter of the walls,
some remained outside.56

The Frontier after the Walls Were Built

In the years following the building of the walls relations between Chinese
states and northern nomads were increasingly fraught with con¬‚ict. This
friction is epitomized in the story of General Li Mu of Chao, recorded in
the Shih chi. Because this is such an important document, often used to
support the notion of aggressive nomadic behavior, it is worth quoting
Li Mu was a valiant general in the northern frontier of Chao. He often stayed
in the T™ai and Yen-men [prefectures] to protect them from the Hsiung-nu.
He had the power to appoint of¬cials, and all the taxes from the towns were

T™ien Kuang-chin and Kuo Su-hsin, “Yü-lung-t™ai Chan-kuo mu,” in T™ien Kuang-
chin and Kuo Su-hsin, O-erh-to-ssu ch™ing-tung ch™i (Peking: Wen-wu, 1986), pp.
T™a La, Liang Chin-ming, “Hu-lu-ssu-t™ai Hsiung-nu mu,” Wen-wu 1980.7:
For a discussion of the issue of “ethnic relations” in the Great Wall region, see
Li Feng-shan, “Lun ch™ang-ch™eng tai tsai Chung-kuo min-tsu kuan-hsi fa-chan
chung te ti-wei” [On the Position of the Great Wall Zone in the Development of
the Chinese Ethnic Relations], in Ch™ang-ch™eng kuo-chi hsüeh-shu yen-t™ao-hui,
pp. 73“85, 366“67. The position taken by this author emphasizes economic
exchanges and a gradual process, since an early time, of cultural integration.


. 6
( 15)