. 10
( 15)


Although it would be far-fetched to suggest that the sole motive for
opening and defending the route to Central Asia was to gain horses needed
for the wars against the Hsiung-nu, there is no doubt that Li Kuang-li™s
campaign of 104“101 b.c. was chie¬‚y intended to acquire horses of a supe-
rior breed and that high-quality stallions were included in the tributary rela-
tions between China and Central Asia.58 In fact, horses had probably been
imported from Central Asia for some time before the arrival of Han armies
in the Tarim Basin, although when this trade began is uncertain.59 Nonethe-
less, imported Central Asian horses could not possibly satisfy the need for
military mounts, which was instead probably ful¬lled chie¬‚y through the
capture of animals in Hsiung-nu territory and through the incorporation
into the Han army of nomadic peoples.60
An ef¬cient cavalry force does not depend solely on horses. It is also
essential to be able to train special troops in shooting from horseback and
in all other cavalry skills. The creation of a cavalry force, one initially
intended especially as a frontier defense force, had already started in 178
b.c.,61 and in the period between 129 and 119 b.c. the mounted soldier was
in fact “the key factor” in Han military campaigns.62 From 130 b.c. to a.d.
23 the Han army comprised two types of soldiers: the ping, “combat sol-
diers for both temporary and permanent service,” who were volunteers and
salaried, and the tsu, that is, conscripted men used chie¬‚y for labor and
guard duty.63 Because cavalry forces were already in existence before Han

Yü, Trade and Expansion in Han China, pp. 119“20; Loewe, “Introduction,” in
China in Central Asia, pp. 43“44.
Yetts argued that already in the ¬rst century b.c. there were two types of
horses in China; see W. Perceval Yetts, “The Horse: A Factor in Early Chinese
History,” in Eurasia Septentrionalis Antiqua, “Minns Volume” 9 (1934): 231“
The “heavenly horses” imported from Ferghana seem to have been used only for
rituals and in parade functions only, whereas the small sturdy Przhevalsky horse
“ possibly the jung ma of the Chinese sources “ continued to be used mainly in
warfare. The account of the war against Ta-yüan, at the end of which only ten
“high quality” horses (shan ma) and three thousand “medium quality” horses
(chung ma) were selected by the Han general to be brought back, shows that
horses from Central Asia were imported in relatively small quantities, owing pos-
sibly to the dif¬culty of taking large herds over such a long distance. Shih chi
chu-yi 123, 4: 2602.
Swann, Food and Money in Ancient China, pp. 168“72.
Chang Chun-shu, “Military Aspects of Han Wu-ti™s Northern and Northwestern
Campaigns,” p. 167; Michael Loewe, “The Campaigns of Han Wu-ti,” in Chinese
Ways in Warfare, ed. Frank A. Kierman and John K. Fairbank (Cambridge Mass.:
Harvard University Press, 1974), p. 101.
Swann, Food and Money in Ancient China, pp. 50“53, 207, n. 326. Chang
Chung-shu (“Military Aspects,” p. 169) maintains that “Emperor Wu reformed

Wu-ti, the appearance of the ping does not coincide, as some have surmised,
with the introduction of a new type of soldier specially trained as a mounted
archer. Rather, the ping were specially trained combat forces used as both
infantry and cavalry and were accompanied by auxiliary troops comprised
of mounted nomads who were also used as frontier guards.64
Because adequate pasture and people skilled in horse breeding were
scarce, the autarchical breeding of horses was dif¬cult to sustain, and during
the reign of Wu-ti home-bred horses, because they were either few or of
poor quality, had to be increasingly replaced with horses imported or
captured from the nomads. The pool of horses available at the time of
the ¬rst campaigns against the Hsiung-nu nevertheless provided enough
mounts to enable the Han to achieve those initial victories that gave them
access to the steppe and to the resources to be found there. By the time of
Han Wu-ti, cavalry had grown “into an independent arm, and ¬nally
became the most important one in the wars against the roving tribes of
Central Asia.”65
Besides cavalry, Wu-ti could also count on progress in military tech-
nology made during the earlier period. The main innovations in defensive
technology seem to have been the Han adoption of ironclad armor, which
replaced the copper-hide armor previously in use, and metal helmets. The
wearing of body armor by horsemen was a practice that was probably bor-
rowed by King Wu-ling from the nomads, who wore a type of cuirass made
of leather obtained from their domestic animals.66 Leather armor was also
used by Chinese soldiers; ironclad armor “ ¬rst introduced during the
Former Han “ seems to be an evolution from the leather type imported from
the nomads,67 and was also used in the ¬ghting against the Hsiung-nu
during the reign of Han Wu-ti.68 The main offensive innovation was repre-
sented by different types of crossbows, which were more or less powerful

the old draft system by establishing a permanent army through the use of mer-
cenaries.” We can take Chang™s statement as a supposition or educated guess, but
not as based on factual evidence.
Loewe, “The Campaigns of Han Wu-ti,” pp. 90“96.
Berthold Laufer, Chinese Clay Figures, pp. 229“30.
Laufer, Chinese Clay Figures, p. 224.
Édouard Chavannes, Les documents chinois decouverts par Aurel Stein dans
les sables du Turkestan Oriental (Oxford: Impr. de l™Universit©, 1913), pp. xv“
A Western Han suit of iron armor was found at Erh-shih-chia-tzu, in Inner Mon-
golia. The armor is made of plates stitched together and divided into several
section for the chest, shoulder, and collar. See Anon., “Hu-he-hao-t™e Erh-shih-
chia-tzu ku-ch™eng ch™u-t™u de Hsi Han t™ieh chia,” K™ao-ku 1975.4: 249“58.
Transl. “The Western Han Iron Armors Unearthed from the Remains of an
Ancient City at Ershijiazi in Huhehot,” in Chinese Archaeological Abstracts, vol.
3: Eastern Zhou to Han, pp. 1349“58.

according to their “pull.”69 In the Han documents excavated by Stein in
Central Asia, and brilliantly studied by Chavannes, there are numerous
references to the crossbow.70 Multiple-shooting, or “repetition,” crossbows
and precision triggers were also developed during the ¬rst half of the Former
Han dynasty, making the crossbow in general a weapon superior to the
nomadic bow. Indeed, these innovations in the basic crossbow design seem
to have accounted for the dif¬culty the nomads met when ¬ghting an orga-
nized Chinese military detachment, even when, as in the case of Li Ling™s
army, the Chinese soldiers were infantry troops much less numerous then
the nomadic forces.71
In sum, the adoption of cavalry from the northern peoples, together with
indigenous technological innovations, made the Chinese army probably on
the whole superior to the Hsiung-nu on the battle¬eld. However, one fun-
damental weakness remained: logistical support and supplies.72 Han states-
men recognized that the need to feed the soldiers in the ¬eld and the cold
northern weather had always constituted limitations to the mobility of
Chinese armies, making it impossible to sustain an expedition for more than
one hundred days.73 Given these limitations, troops stationed far away for
long periods of time had to be self-supporting. Food supplies were provided
by state-sponsored military farms run by t™ien-tsu units (translated as “agri-
cultural conscripts,” “colonists,” or “pioneers”).74
The system of garrisoned forts and beacon ¬res, coupled with the con-
struction of military roads on China™s northern frontier, allowed for a rapid
reaction to nomadic inroads and invasions, but this was merely for defen-
sive purposes. Once an expedition was launched, the border regions could
supply foodstuffs and logistical support only within a limited range; long-
term expeditions in pursuit of retreating nomadic tribes needed to rely on
what they could pillage from the enemy. In the course of the campaign slow-
moving livestock and people could be, and were, captured, but the bulk of
the nomadic soldiers, more mobile than the Chinese, could easily retreat
into the steppe, thereby avoiding direct military confrontations. If an expe-
dition lasted longer than expected, the Chinese soldiers were ill-prepared
to face the rigors of the nomadic lands and found it impossible to survive
in the steppe without supplies and logistical support. In fact, climatic factors

Loewe, Records of the Han Administration, 1: 99; Joseph Needham et al., Science
and Civilization in China, vol. 5: Chemistry and Chemical Technology, Part VI:
Military Technology: Missiles and Sieges (Cambridge: Cambridge University
Press, 1994), p. 142.
On the technical and mechanical aspects of the Han crossbow see Joseph
Needham et al., Science and Civilization in China 5, Part 6: 120“44.
Loewe, “The Campaigns of Han Wu-ti,” pp. 119“22.
Loewe, “The Campaigns of Han Wu-ti,” p. 96. 73 Han shu 94B, 3824.
Loewe, Records of the Han Administration, p. 56.


and food shortages were responsible for a larger number of Chinese casu-
alties than were the nomadic armies.

The Han Campaigns

Following the tradition of stationary border defenses developed in the late
Warring States period, the Han approached this problem investing heavily
in the construction a forti¬ed line of communication with warehouses that
were intended to supply garrison forces.75 The supply system itself included
three types of storage facilities: granaries (ts™ang), intended for the man-
agement of grain, which were kept under strict military surveillance; store-
houses (k™u) where money and weapons were kept; and temporary stations
for the storage of money, grains, and weapons of soldiers garrisoned on the
frontier.76 As already mentioned, this support structure was linked to an
overall strategy of turning the border areas, as they were conquered, into
“commanderies” and other administrative divisions integrated within the
military and civil bureaucracy of the Han. As these areas were brought
under the control of the central government, they were either settled by
Chinese people or garrisoned by Han troops and foreign troops recruited
under the Han insignia. Based on this strategy and on the new-found mil-
itary advantages that underpinned it, the struggle against the Hsiung-nu
and the Han (re)conquest of the north conducted by Wu-ti can be divided
into three phases.

¬rst phase (133“119 b.c.). Wu-ti™s early campaigns were aimed at recov-
ering all lands to the south of the Great Bend of the Yellow River and at
consolidating Han power in those border areas that were more vulnerable
to Hsiung-nu attacks. This ¬rst phase of the Han offensive is characterized
by a series of attacks and counterattacks that did not achieve any lasting
results and failed to establish the military supremacy of either side. They
did show, however, that the Han had become able to mount extensive cam-
The study of these installations is based on the bamboo slips excavated especially
from Chü-yen and Tun-huang. See Loewe, Records of the Han Administration.
Hsü Le-yao, “Han chien so chien ch™ang-ch™eng te hou-ch™in kung-chi hsi-t™ung,”
in Ch™ang-ch™eng kuo-chi hsüeh-shu yen-t™ao-hui lun-wen chi, ed. Chung-kuo
ch™ang-ch™eng hsüeh-hui (Chi-lin-shih: Chi-lin Jen-min, 1994), pp. 116“22,
375“76. The wealth of information provided by the Chü-yen and the Tun-huang
slips for the investigations of Han garrison system in the northwest has been
recently augmented with the discovery of over 22,500 slips at Hsüan-ch™üan, near
Tun-huang; see Wu Jeng-hsiang, “Ssu“chou chih lu shang yu yi zhong-ta k™ao-ku
fa-hsien: Tun-huang hsien Hsüan-ch™üan chih,” in Ch™ang-ch™eng kuo-chi hsüeh-
shu yen-t™ao-hui lun-wen chi, pp. 283“85, 487“88.


paigns and to react to Hsiung-nu incursions. The Han could both strike
deep into the steppe and attack from multiple directions. After the unfor-
tunate episode of the defeat at Ma-yi, Han military efforts to expand into
Hsiung-nu territory began in earnest in 129 b.c., with a large offensive led
by four commanders.
Wei Ch™ing, general of Chariots and Cavalry (chü-ch™i chiang-chün),77
set out from Shang-ku;78 Kung-sun Ho, general of Light Chariots (ch™ing-
chü chiang-chün),79 departed from Yün-chung; Grand Palace Grandee (ta-
chung ta-fu) Kung-sun Ao, appointed cavalry general, was sent from the
Tai commandery; and the commandant of the Guards, Li Kuang, was made
general of the Imperial Cavalry (hsiao-ch™i chiang-chün) and set out from
Yen-men. Each of the four commanders™ armies comprised ten thousand
cavalry. In this expedition, as in later ones, what is surprising is the strik-
ing distance of the Chinese armies. Wei-ch™ing must have penetrated fairly
deeply into Hsiung-nu territory, since he reached Lung-ch™eng, the sacred
site of the Hsiung-nu sacri¬ces to Heaven where the ch™an-yü held large
political gatherings. Here he killed and captured several hundred men.
However, Kung-sun Ao lost seven thousand cavalry, and Li Kuang was
taken prisoner, though he then managed to escape and return home. Both
ransomed their lives (to avoid being executed) and were degraded to
the rank of commoners. Kung-sun Ho did not achieve any merit, either;
so, as Ssu-ma Ch™ien recorded, out of four generals, only one attained
This expedition was followed by a Hsiung-nu counteroffensive, in 128
b.c., when the Hsiung-nu crossed the border in force, killing the governor
of Liao-hsi81 and invading Yen-men, where they killed or captured several
thousand people. Wei Ch™ing attacked them again, and Han An-kuo was
made general of skilled soldiers (ts™ai-kuan chiang-chün) and stationed to
garrison Yü-yang,82 a position he later quit claiming that the Hsiung-nu had

Cf. Hans Bielenstein, The Bureaucracy of Han Times (Cambridge: Cambridge
University Press, 1980), pp. 122“23.
Commandery located in the northern part of present-day Hopei. Included the ter-
itory to the west of today™s Yen-ch™ing county (Peking municipality) and to the
east of the city of Ch™ang-p™ing county; cf. Chung-kuo li-shih ti-t™u chi. The His-
torical Atlas of China, ed. T™an Ch™i-hsiang [Tan Qixiang] et al. (Peking: Ti-t™u
ch™u-pan-she, 1982), 2: 27“28, 3“2. Its administrative center was in Chü-yang.
He was formerly a grand coachman (t™ai-p™u).
Shih chi 111, 2923; Shih-chi chu-yi 109, 2350.
Commandery located in the region to the west of the lower course of today™s
Ta-ling River, in Liao-ning (Chung-kuo li-shih ti-t™u chi, 2: 27“28, 3“5).
Commandery located in the region to the east of Peking. The administrative center
was Yü-yang, situated to the southwest of today™s Mi-yün county (Chung-kuo li-
shih ti-t™u chi, 2: 27“28, 3“3).


¬‚ed far away. A month later, the Hsiung-nu attacked again in force Shang-
ku and Yü-yang.83 The most signi¬cant effort aimed at achieving the “paci-
¬cation of the north” was carried out by Wei Ch™ing, who in 127 b.c.
crossed the Western River84 and reached Kao-ch™üeh,85 killing and captur-
ing twenty-three hundred enemies. After this, he went west to pacify
the region south of the Yellow River, where he inspected the “old frontier”
at Yü-hsi,86 that is, the position occupied by the Ch™in armies after Meng
T™ien™s offensive. Next he crossed the Tzu-ling Mountains from east to
west87 and built a bridge over the Northern River.88 Then he attacked and
defeated the Hsiung-nu local lord P™u-ni at the locality of Fu-li, killed
P™u-ni™s picked soldiers, and took three thousand and seventy-¬ve scouts
prisoner.89 Even such a sweeping military action did not intimidate the
Hsiung-nu, who in the following year, 126 b.c., crossed the borders again
and killed the grand administrator of the Tai commandery, Kung Yu. At
that time the Hsiung-nu also invaded Yen-men and abducted over a thou-
sand people.90
As the Han armies were seeking to gain the upper hand in military
matters, they also proceeded to consolidate the Han defenses along the
northern border, by building a series of forti¬cation works, by transferring
in new settlers, and by establishing new administrative units in the border
regions. The ¬rst change in the organization of the north was made when
the Shuo-fang commandery was established in 126 b.c.91 This strategy
caused the Hsiung-nu to react angrily to what they regarded as a regular
invasion of their territory. As the Shih chi reports, “the Hsiung-nu Wise

Shih chi 108, 2864; Shih-chi chu-yi 108, 2293.
The “Western River” indicates the section of the Yellow River that ¬‚ows south
to north through today™s provinces of Ning-hsia and Inner Mongolia.
This was an important pass across the Yin Mountain Chain (Chung-kuo li-shih
ti-t™u chi, 2: 17“18, 1“5).
Also called Yü-lin Barrier (sai), it was located on the northeastern corner of the
Great Bend of the Yellow River, which was possibly in the northeastern corner
of Shensi province (not marked in Chung-kuo li-shih ti-t™u chi).
Tzu-ling: name of a mountain; its location is unknown. It was possibly situated
to the west of present-day Heng-shan county, in Shensi (not marked in Chung-
kuo li-shih ti-t™u chi).
This is the Wu-chia River, in present-day Inner Mongolia, which once ¬‚owed into
the Yellow River.
Fu-li: name of a fort located to the northwest of present-day Wu-yüan county, in
Inner Mongolia (not marked in Chung-kuo li-shih ti-t™u chi).
Shih chi 111, 2924; Shih-chi chu-yi 111, 2350“1.
Shih chi 112, 2950, and 116, 2995; Shih-chi chu-yi 112, 2436. Commandery
established in 127 b.c. It was located across the northern portion of the Great
Bend of the Yellow River, in today™s Inner Mongolia. Its administrative center was
at Shuo-fang (Chung-kuo li-shih ti-t™u chi, 2: 17“18, 1“7/3“7).

King of the Right was angry that the Han had seized the territory south of
the Yellow River and had built forti¬cations in the Shuo-fang comman-
dery.”92 Hence, in 125 b.c. they invaded the commanderies of Tai, Ting-
hsiang,93 and Shang, killing and carrying away several thousand Han
people.94 As the military confrontation between the two empires escalated,
long-range expeditions of enormous proportions were coupled with the
establishment of additional outlying administrative units and permanent
forces to guard the border areas. By the spring of 123 b.c., Wei Ch™ing, at
the head of six other generals and an army of over one hundred thousand
cavalry, proceeded several hundred li north of Ting-hsiang to attack the
Hsiung-nu.95 The following year ten thousand Hsiung-nu cavalry invaded
Shang-ku,96 but in 121 b.c. General Huo Ch™ü-ping, at the head of ten thou-
sand cavalry, setting off from Lung-hsi97 marched for over one thousand li,
crossed the Yen-chih Mountains,98 and attacked the Hsiung-nu. In the
summer of that year Huo Ch™ü-ping, together with the Ho-ch™i marquis
Kung-sun Ao, again led a force of tens of thousands at a distance of
two thousand li north of Lung-hsi and Pei-ti. They passed Chü-yen99 and
gave battle in the Ch™i-lien Mountains,100 where they killed or captured
over thirty thousand enemies, including seventy “small kings” and lesser
These military confrontations show that the Han had become pro¬cient
in planning and enacting long-range military campaigns and were capable
of making surprise attacks on areas located deep within nomad territory.
This strategy paid off, as discord began to brew within the ranks of the
Hsiung-nu, and some prominent Hsiung-nu leaders started to defect to the

Shih chi 110, 2907; Shih-chi chu-yi 110, 2328.
Commandery located north of the Yellow River, in the region of today™s Ho-lin-
ko-erh county, in Inner Mongolia; its capital was called Ch™eng-yüeh (Chung-
kuo li-shih ti-t™u chi, 2: 17“18, 2“9).
Shih chi 110, 2924; Shih-chi chu-yi 110, 2350.
Shih chi 110, 2907; Shih-chi chu-yi 110, 2328.
Shih chi 110, 2908, Shih-chi chu-yi 110, 2328.
Province located to the south of the upper course of the Wei River, in southern
Kansu (Chung-kuo li-shih ti-t™u chi, 2: 33“34, 5“7/5“8) (see also above).
Mountain range located in present-day Kansu, west of Yung-ch™ang county and
southeast of Shan-tan county. It was known for its excellent pastures (Chung-
kuo li-shih ti-t™u chi, 2: 33“34, 3“6).
County to the southeast of today™s E-chi-na Banner (Edzin-Gol), in Inner Mon-
golia. It lay on the main route of communication between the territory to the
west of the Yellow River and the land north of the desert (Chung-kuo li-shih ti-
t™u chi, 2: 33“34, 2“6).
Mountain located in Kansu, to the south of Chiu-ch™üan city (Chung-kuo li-shih
ti-t™u chi, 2: 33“34, 3“5).
Shih chi 110, 2908; Shih-chi chu-yi 110, 2328“9.


Han, among them the Hun-yeh king. Subsequently, the areas of Lung-hsi,
Pei-ti, and Ho-hsi suffered fewer Hsiung-nu raids, and the Han transferred
the poor people of Kuan-tung102 to settle in the area of Hsin-ch™in-chung103
(south of the Yellow River), which had been seized from the Hsiung-nu and
which was now being populated with Chinese people. The settlement policy
also allowed for a reduction in the number of troops stationed to the north
of Pei-ti.104 The Hsiung-nu continued to attack, so in 119 b.c. the Han
mounted a major expedition to hit deep into Hsiung-nu territory, and the
entire Han army was mobilized. During this expedition Han troops reached
the T™ien-yen Mountains105 and the forti¬ed town of Chao-hsin106 before
turning back. Huo Ch™ü-ping, advancing over two thousand li from Tai,
clashed with the wise king of the left. The Han soldiers killed or captured
over seventy thousand enemies, and the wise king of the left and his gen-
erals all ¬‚ed. Huo Ch™ü-ping performed a feng sacri¬ce at Mount Lang-chü-
hsü107 and a shan sacri¬ce at Mount Ku-yen,108 descended to the Han-hai,
and then turned back. After this, the Hsiung-nu withdrew to the northern
steppe in today™s Mongolia, and their royal court was no longer located
south of the Gobi.109
During this phase the whole military and civil administrative structure
of the border regions was reorganized and gradually came to be structured
on three levels: (1) the commanderies inside the defensive line (pien-sai); (2)
an intermediate area outside the defensive line populated mainly by non-
Chinese peoples but still under the formal control of the Han bureaucracy;
and (3) the territories that were outside the limits of Han control but that
had accepted some form of subordination to the Han.110 To the ¬rst level
belonged the two commanderies of Shuo-fang and Wu-yüan, set up in 127

Region to the east either of the Han-ku Pass (Chung-kuo li-shih ti-t™u chi, 2:
15“16, 4“8) or of T™ung-kuan, a district in Shensi province (not marked in
Chung-kuo li-shih ti-t™u chi).
This indicates the region south of the Great Bend of the Yellow River, in Inner
Mongolia (Chung-kuo li-shih ti-t™u chi, 2: 17“18, 1“6/3“6).
Shih chi 110, 2909“10; Shih-chi chu-yi 110, 2329.
Range located in Outer Mongolia, to the southeast of today™s Khangai
Hsiung-nu city built by Chao Hsin, located to the west of the T™ien-yen
Mountain located to the east of present-day Ulan Bator (Chung-kuo li-shih
ti-t™u chi, 2: 39, 2“4.)
Mountain located to the west of the Lang-chü-hsü Mountains (Chung-kuo
li-shih ti-t™u chi, 2: 39, 2“4).
Shih chi 110, 2910“11; Shih-chi chu-yi 110, 2329“30.
This third level was most probably established towards the end of the reign
of Wu-ti (see below). Cfr. Bielenstein, The Bureaucracy of Han Times,
pp. 109“13.


b.c. in the region north of the Yellow River after the Han conquered those
nomadic lands in the areas of Ordos, Tai, and Yen. The second level of
border administration began to be established before 119 b.c. and com-
prised areas that lay beyond the effective reach of the Chinese military
colonies but were still regarded as subject to Chinese administration. These
territories were de¬ned as Dependent States (shu-kuo), and their chief
of¬cer was the director of Dependent States (tien-shu-kuo).111 The Depen-
dent States were inhabited by a largely non-Chinese, nomadic population
that had “surrendered” (or simply switched allegiance from the Hsiung-nu
to the Chinese) and were intended to be a buffer between the Chinese
defense line and the Hsiung-nu and Ch™iang tribesmen.112 From 121 to 28
b.c. their administrative structure included a chief commandant of a Depen-
dent State (shu-kuo tu-wei), with one assistant (ch™eng), one or more cap-
tains (hou) and battalion commanders (ch™ien-jen),113 and a prefect of the
Nine Successive Interpreters (chiu-yi ling).114

second phase (119“104 b.c.). At the end of the period just discussed
most of the objectives that Wu-ti had set out to accomplish ten years earlier
had been achieved. The northern border had been secured, the Hsiung-nu
had been pushed farther to the north, and the Han were no longer paying
a tribute. Yet the military campaigns were proceeding unabated, and some
of the most resounding Han successes were obtained between 119 and 112
b.c. The Han government continued its policy of generally strengthening
and restructuring the state economy, increasing state control over the ¬nan-
cial resources of the country, and thus possibly allowing the military cam-
paigns to continue. The establishment of the salt and iron monopolies in
119 b.c. signaled the beginning of a wide range of economic reforms, and
the timing may indicate that the rationalization and centralization of the
¬nancial administration may have been made more urgent because the war
effort had depleted state resources. As events after 119 b.c. show, the Han
emperor and his advisers were no longer satis¬ed with their initial objec-
tives and were determined to continue to ¬ght with a broader agenda. The
goal of defeating the Hsiung-nu militarily, already achieved, was re-
placed by a plan for the annihilation of Hsiung-nu political power, to be

This post was inherited from the Ch™in administration, but its authority was
increased in 121 b.c.; see Bielenstein, The Bureaucracy of Han Times, p. 11. On
the establishment of the shu-kuo and their relationship to the central govern-
ment, see Loewe, Records of the Han Administration, 1: 61“63.
Bielenstein, The Bureaucracy of Han Times, p. 109.
Bielenstein has the term “millarian”; I follow here C. Hucker, A Dictionary of
Of¬cial Titles in Imperial China (Stanford: Stanford University Press, 1975), no.
Bielenstein, The Bureaucracy of Han Times, p. 84.


T ™I E N S H


Ch WU-
™i -
Lie 2 WEI



AY zi

Map 5
1. Kao-ch™üeh
Main Places Cited 2. Yen-chih Mts.
in Relation to the Wars 3. Chü-yen
4. Yü-men
Between Han and Hsiung-Nu 5. Tun-huang
6. Lang-chü-hsü
7. Mt. Ku-yen
8. Shou-chiang-ch™eng
9. Han-ku Kuan
Baikal N

u rR
0 300 mi

0 300 km

Ke iver
ri R


O -H

O -T


N an

8 IN

I-P er

™I N











I Peking


Gulf of
v er




Lo Yellow
Fen R



ng R .

River 9

Ha iR


East China


accomplished through a campaign leading to their “international” isola-
tion, both political and economic. By 110 b.c., after several other cam-
paigns, the Han had established the new commandery of Chiu-ch™üan,
which had the strategic purpose of cutting the routes of communication
between the Hsiung-nu and the Ch™iang. In the north the Han expanded its
areas under cultivation to Hsien-lei,115 which became the new frontier.116
The Hsiung-nu were forced to retreat far into the northern steppe and
forests, abandoning not only the region south of the Yellow River but also
the areas south of the Gobi.
The administrative expansion also continued in this phase, with a total
of fourteen border commanderies established in the northern and southern
regions of the empire between 112 and 108 b.c., thus consolidating the ter-
ritorial gains. Westward expansion and an increasing number of military
and commercial expeditions to Central Asia proceeded together with the
establishment of a defensive line from east to west. This line consisted of
permanent, self-supporting garrisons that controlled the movement of the
nomadic tribes, signaled a warning in case of danger, engaged in agricul-
tural production, and protected the lines of communication. The defensive
line also provided steady and reliable logistical support for the Chinese
expeditionary forces and included garrisons such as those along the Edsin-
gol that reached out to the lakes of Chü-yen and penetrated deeply into
nomadic territory.117

third phase (104“87 b.c.). At this point, the way was paved for further
Han expansion into the Western Regions. In 104 b.c. the Han dispatched
the Erh-shih general Li Kuang-li west to attack Ferghana, and ordered
General Kung-sun Ao to build the “City for Receiving Surrender” (Shou-
chiang-ch™eng).118 Han diplomats then came into contact with the kingdoms
of the Tarim Basin. Wang Hui, an envoy who allegedly had been insulted
by the people of Lou-lan, was sent to help Chao P™o-nu attack and defeat
Lou-lan.119 Subsequently a line of forts and stations was built from Chiu-
Hsien-lei: name of a locality in the vicinity of today™s T™a-ch™eng county, in
Sinkiang (not marked in Chung-kuo li-shih ti-t™u chi).
Shih chi 110, 2913; Shih-chi chu-yi 110, 2331.
Loewe, Records of the Han Administration, 1: 55“56, 184, map 3.
Shih chi 110, 2915; Shih-chi chu-yi 110, 2333; The “City for Receiving Surren-
der,” which was built to welcome the Hsiung-nu noblemen who surrendered to
the Han, was located in today™s Inner Mongolia, to the north of the Yin Moun-
tains (Chung-kuo li-shih ti-t™u chi, 2: 39, 2“4).
On this see Hulsew© and Loewe, China in Central Asia, pp. 86“90. On the loca-
tion of Lou-lan see Enoki Kazuo, “The Location of the Capital of Lou-lan and
the Date of the Kharosthi Documents,” Memoirs of the Research Department
of the Toyo Bunko 22 (1963): 125“71.


ch™üan to the Yü-men Pass,120 and Superintendent of the Imperial House-
hold Hsü Tzu-wei left the Wu-yüan121 frontier post and journeyed several
thousand li, building fortresses, walled outposts, and a line of watch sta-
tions that stretched to Mount Lu-ch™ü.122 At the same time, the Han sent
the chief commandant of Archers Carrying Heavy Bows, Lu Po-te, to build
forti¬cations in the Chü-yen Marshes.123
Key to the Han expansion was the Chiu-ch™üan commandery, located in
a strategic position in the Kansu Corridor.124 In 102 b.c. the Hsiung-nu wise
king of the right invaded Chiu-ch™üan and Chang-yi and abducted several
thousand people,125 and in 99 b.c. the Han dispatched the Erh-shih general
Li Kuang-li at the head of thirty thousand cavalry out of Chiu-ch™üan to
attack the wise king of the right in the T™ien Shan.126 This campaign ended
in disaster when, after an initial success, Li Kuang-li was surrounded by the
Hsiung-nu; he barely survived, and his army was all but wiped out. In the
same year, Li Ling also set out from Chiu-ch™üan at the head of ¬ve thou-
sand special infantry troops. His valorous campaign also ended in defeat.
He was confronted by an overwhelming number of enemy troops and even-
tually surrendered to the Hsiung-nu.127 Two years later the Han again dis-
patched Li Kuang-li, this time at the head of sixty thousand cavalry and
one hundred thousand infantry. The Hsiung-nu responded by moving all
their families and property to the north of the Hsü-wu River,128 while the
ch™an-yü waited to the south of the river with one hundred thousand
cavalry.129 The last engagement between Li Kuang-li and the Hsiung-nu

Shih chi 123, 3172, Shih-chi chu-yi 123, 2598; Yü-men (Jade Gate) is located to
the northwest of Tun-huang, in Kansu (Chung-kuo li-shih ti-t™u chi, 2: 33“34,
Commandery located in Inner Mongolia, to the north of the Great Bend of the
Yellow River (Chung-kuo li-shih ti-t™u chi, 2: 17“18, 2“7).
This was the name of the northern extension of today™s Lang Mountains, in Inner
Mongolia (not marked in Chung-kuo li-shih ti-t™u chi).
Shih chi 110, 2916 Shih-chi chu-yi 110, 2333.
Commandery located in present-day Kansu province, east of the Shu-le River,
and west of Kao-t™ai county; its administrative center was Lu-fu, today™s Chiu-
ch™üan city (Chung-kuo li-shih ti-t™u chi, 2: 33“34, 3“4).
Shih chi 110, 2916“7, Shih-chi chu-yi 110, 2334; Chang-yi was a commandery
established in 111 b.c. Its administrative center was in Lu-te. It included the ter-
ritory to the west of today™s Yung-shang, and to the east of Kao-t™ai, in Kansu
(Chung-kuo li-shih ti-t™u chi, 2: 33“34, 3“6).
Shih chi 110, 2917. Shih-chi chu-yi 110, 2334.
On Li Ling™s campaign see Loewe, “The Campaigns of Han Wu-ti,” pp. 119“22.
Chung-kuo li-shih ti-t™u chi, 2: 39, 2“4. This is today the Tula River, in the north
of Mongolia.
Shih chi 110, 2918. Shih-chi chu-yi 110, 2335.


occurred in 90 b.c. On this occasion he was captured, having apparently
lost heart because of the execution of his family at home. Eventually he was
put to death by the Hsiung-nu.130 After 99 b.c. the heyday of Han expan-
sion was over, and in subsequent years a different attitude emerged again,
one more favorable to compromise. On their side, the Hsiung-nu were beset
by internal discord, ineffective leadership, natural calamities, and the seces-
sion and sometimes open rebellion of chieftains and former tributaries.131
This state of affairs led to the maturation of the conditions for a peaceful
settlement between Hsüan-ti and Hu-han-yeh Ch™an-yü in the 50s b.c. The
gains obtained by China during the period of Wu-ti, and especially the
presence of the Han in the Western Regions, were consolidated and
The administration of the Western Regions was formalized with the cre-
ation of the four commanderies of Chiu-ch™üan, Chang-yi, Tun-huang, and
Wu-wei. The ¬rst two were established in 104 b.c., the third shortly after
(certainly before 91 b.c.), and the fourth between 81 and 67 b.c.132 These
commanderies were placed under the control of a grand administrator (t™ai-
shou) with both civil and military responsibilities, aided by a commandant
(tu-wei). It was also stipulated that where the need arose more than one tu-
wei might be appointed and assigned to different districts within the same
A third level in the administration of border territories, also mentioned
earlier, was probably instituted during this period: the pao (protectorates)
system, which presumably covered “an area that lay outside the main line
of communications and defences.”133 Under Wu-ti additional posts were
established for conducting relations with foreign peoples, such as the
colonel protecting the Ch™iang (hu Ch™iang hsiao-wei) and the colonel pro-
tecting the Wu-huan (hu Wu-huan hsiao-wei).134 In 60 b.c., after the heyday
of the Han thrust into Central Asia, the post of protector general (hsi-yü
tu-hu) of the Western Regions was created, followed in 48 b.c. by a new
supernumerary military post, the Wu and Chi colonelcy (wu-chi hsiao-
wei),135 stationed near Turfan. Later, other special of¬cers were established,
including the commandant of agriculture (nung tu-wei), who was in charge
of the agricultural production in the border prefectures, and the comman-
Biographical information on Li Kuang-li is contained in Han shu 97A and Shih
chi 49. For his campaign in Central Asia, see Hulsew© and Loewe, China in
Central Asia, pp. 228“36, especially n. 926. On the intrigue that led to his
family™s execution see Loewe, Crisis and Con¬‚ict, pp. 45, 53“54.
De Crespigny, The Northern Frontier, pp. 186“87.
Loewe, Records of the Han Administration, 1: 62.
Loewe, Records of the Han Administration, 1: 64.
Bielenstein, The Bureaucracy of Han Times, p. 110.
On the wu-chi hsiao-wei, see Hulsew© and Loewe, China in Central Asia, p. 79,
n. 63.

dant of passes (kuan tu-wei).136 By the end of Wu-ti™s reign the Han had
consolidated their authority in the northern and western territories, thus
replacing the Hsiung-nu as the pre-eminent “superpower.”

The Question of the Western Regions

Few issues in the history of the Former Han have been as rich a source for
scholarly hypotheses and speculation as the discussion on the motives that
dragged the Han into the Western Regions. As mentioned earlier, the Han
advance into the west needs to be comprehended at a level beyond the
purely tactical one. The Chinese advance can surely be explained fully only
within the context of the Hsiung-nu wars, but it is dif¬cult to see it solely
as a natural extension of the declared Han need to protect the frontier
regions. Although protection of the border regions was certainly one of the
Han™s initial goals, the war escalated into a total war, and the Han™s ulti-
mate goal became the destruction of the Hsiung-nu empire. The Western
Regions were important for the Hsiung-nu as an economic base, and the
Han™s chief objective in conquering them was to deprive the nomads of this
source of strength and support. To place this interpretation in its proper
context, however, we should ¬rst examine some other theories proposed to
explain why the Chinese armies colonized today™s Sinkiang.
The ¬rst Chinese expedition to Central Asia, led by Chang Ch™ien, was
sent in 139/138 b.c., although this dating remains controversial.137 Thus
the diplomatic initiative to ¬nd allies in the war against the Hsiung-nu
preceded direct military operations, which began only in 133 b.c. How-
ever, Chang Ch™ien™s mission was intercepted by the Hsiung-nu, and when
he ¬nally reached the Yüeh-chih, he found them unwilling to oppose
the Hsiung-nu. The initial contacts with Central Asia, therefore, did not
produce the expected results in military cooperation, but the exploration
of the routes is likely to have contributed considerably to the decision to
undertake a military offensive, as it was from this point that the Chinese
started to learn in detail about the political composition of this region, the
wealth it produced, and the hegemonic position of the Hsiung-nu over
the area.138
The famed Silk Road explorer Sir Aurel Stein believed that the Chinese
expansion into Central Asia was motivated by the need to ¬nd new markets
for the silk produced in China.139 Others have surmised that the chief

Loewe, Records of the Han Administration, 1: 61.
Hulsew© and Loewe, China in Central Asia, p. 209, n. 774.
Hulsew© and Loewe, China in Central Asia, p. 211.
Aurel Stein, Serindia. Detailed Report of Explorations in Central Asia and West-
ernmost China (Oxford: Clarendon Press, 1921), 1: 406.

catalyst for Chinese expansion into the Tarim Basin was the struggle with
the Hsiung-nu for control of the trade routes in order to ¬nd a commercial
outlet in the West for China™s silk surpluses.140 These arguments have gained
widespread support and continue to appear in recent publications.141
Lattimore distanced himself from this approach, however, and endorsed
the notion that the Han was in fact dragged into Central Asia because of
a growing demand for Chinese luxury products, suggesting also that the
economic possibilities offered by the oases for a “prosperous, intensive,
irrigated agriculture” created favorable conditions for self-supporting
Chinese military settlements. Because of their economic similarity to China,
and despite the geographic distance, the oases and pasturelands of Central
Asia occupied by the Han had a higher degree of cohesion with the main
body of China than did the areas peripheral to the steppe.142
These views presuppose the existence of a commercial network in
Central Asia in the second century b.c. capable of supporting and manag-
ing a complex trade network. Yet at the beginning of Wu-ti™s reign there is
no evidence of anything in Central Asia comparable to the intercontinen-
tal trade that would develop over the land and sea routes during the period
stretching from the ¬rst century b.c. to the third and fourth centuries a.d.143
Not until Wu-ti™s reign did the Chinese became cognizant of the possibili-
ties for trade between China and Central and western Asia, even though
these opportunities do not seem to have enjoyed universal appeal.144
Han economic involvement in modern Sinkiang was a consequence of
events and factors that were primarily political. The earliest Chinese exports
to Central Asia seem to be evidence of exchanges of gifts and tribute as a
means of political leverage rather than evidence of commercial items within

Frederick J. Teggart, Rome and China. A Study of Correlations in Historical
Events (Berkeley: University of California Press, 1939), pp. 148“233; Peter Bood-
berg, “Turk, Aryan and Chinese in Ancient Asia,” in Selected Works of Peter A.
Boodberg, ed. Alvin P. Cohen (Berkeley: University of California Press, 1979),
pp. 9“12.
Jagchid and Symons, Peace, War and Trade, pp. 30, 65.
Owen Lattimore, Inner Asian Frontiers of China (Boston: Beacon Press, 1962
[1940]), pp. 500“506.
Manfred Raschke, “New Studies in Roman Commerce with the East,” in Auf-
stieg und Niedergang der romischen Welt, II Principat, ed. H. Temporini and W.
Haase (Berlin: W. de Gruyter, 1978), 9.2: 622“37. It seems that the formative
period for a continental route regularly frequented by professional merchants
and substantially different from the haphazard or occasional ¬‚ow of trade that
had existed thus far, should be dated to this period; see Liu Xinru, Ancient India
and Ancient China: Trade and Religious Exchanges, AD 1“600 (Delhi: Oxford
University Press, 1988), pp. 1“22.
John Ferguson, “China and Rome,” in Aufstieg und Niedergang der romischen
Welt. II Principat, 9.2: 581“603.


the framework of a burgeoning international trade.145 Thus in the sequence
of events extending from military expansion to formal commercial relations
with the nomadic and sedentary peoples of Ta-yüan (Ferghana or Sogdi-
ana)146 and Ta-hsia (Bactria), the “¬‚ag” did not “follow the trade.” On the
contrary, the whole conquest of the Western Regions was due to the unfore-
seen outcomes of military and political developments,147 and the explo-
ration of Central Asia was primarily a strategic move within the wider
context of the political and military confrontation between Han and
Hsiung-nu.148 Among the motives for this course of action were the desire
to ¬nd allies in the West so that a “pincher” offensive could be carried out
against the Hsiung-nu from both ends of the Central Asian steppe and the
desire to isolate the Hsiung-nu from potential allies such as the Ch™iang. A
third motive may have been related to Wu-ti™s desire for “exotic” goods
from Central Asia. These motives, however, do not actually explain why
the Western Regions were such a crucial strategic objective for the Han even
after they had repeatedly beaten the Hsiung-nu.

The Western Regions as a Hsiung-nu Power Base

It has been pointed out that Wu-ti may have intended to establish a direct
connection with Central Asia to undermine Hsiung-nu prestige and author-
ity in that region. Many of the region™s peoples recognized themselves as
Hsiung-nu tributaries, and if China could act as a counterweight to Hsiung-
nu power, the latter™s in¬‚uence would be drastically reduced. But were these
dwellers in the relatively sparse oases truly vital to Hsiung-nu power, and,
if so, how? Unless we understand the relationship between the Hsiung-nu
and the oasis people, we cannot properly assess how Han conquest of the
region affected the war between the two empires.
In Hulsew©™s most perceptive answer to this question, “the only reason
for the Chinese expansion in Central Asia was the desire to stop the inva-
sions of the Hsiung-nu, or, as the ancients said, to ˜cut off their right arm™,
Raschke, “New Studies in Roman Commerce with the East,” p. 611.
Hulsew© and Loewe, China in Central Asia, p. 273.
Loewe, “The Campaigns of Han Wu-ti,” pp. 84“85; id., “Introduction,” in
China in Central Asia, p. 40; Yü Ying-shih, “The Hsiung-nu,” pp. 131“33;
Raschke, “New Studies in Roman Commerce with the East,” pp. 615“17;
M. C. Webb, “The Flag Follows Trade: An Essay on the Necessary Interaction
of Military and Commercial Factors in State Formation,” in Ancient Civilization
and Trade, ed. J. A. Sabloff and C. C. Lamberg-Karlovsky (Albuquerque: Uni-
versity of New Mexico Press, 1975), pp. 179“94.
Leslie and Gardiner, “Chinese Knowledge of Central Asia,” pp. 254“308; J. R.
Gardiner-Gardner, “Chang Ch™ien and Central Asian Ethnography,” Papers of
Far Eastern History 33 (1986): 23“79.

i.e., to deprive them of their western bases.”149 The territorial conquests of
the Han were dictated by the military necessity to cut off the Hsiung-nu
from those areas that supplied them with provisions: the Kansu Corridor,
the oases of Central Asia, and southern Manchuria.150 The “provisions”
should be understood, I believe, as agricultural goods and the products of
urban craftsmen.
We have already seen how, before Han Wu-ti™s offensive, the Hsiung-nu
managed to acquire large amounts of goods by relying on the frontier
markets, raids into Chinese territory, and ho-ch™in payments.151 As Han Wu-
ti took the offensive, the Hsiung-nu could no longer count on the “tribute”
products or on exchanges at the frontier markets, which were also dis-
rupted; at the same time, they suffered heavy losses of animals and pas-
tureland. As the Hsiung-nu retreated, moving their political center farther
and farther to the north, they had to rely more heavily on the agricultural
settlements in Central Asia and in southern Siberia.152 The oases of Sinkiang
were close to those centers, had a long history of contacts with nomadic
peoples, and were rich enough to support a sizeable urban population. They
came to represent the main (though probably not the only) source of supply
for agricultural and other products for the Hsiung-nu after the nomads lost
the revenues from China, especially after the campaigns of Wei Ch™ing and
Huo Ch™ü-ping forced the Hsiung-nu to transfer their political and military
center to the northwest. The Chinese military and political presence in
Central Asia therefore became vital to the Han overall strategy of weaken-
ing the nomads and was accomplished mainly through the establishment of
farming colonies managed by the military. These military-agricultural set-
tlements had multiple functions: to prevent the Hsiung-nu from gaining
access to agricultural products, to serve as advanced logistic support for
Chinese expeditionary armies, and to protect the trade that China was start-
ing to organize with the west.
The close economic ties between the Hsiung-nu and the Western Regions
are evident from several passages in the Shih chi. In chapter 110 there is
mention of the Han strategy “to create a split between the Hsiung-nu and
the states to the west which had up to this time supported them.”153 These

A. F. P. Hulsew©, “Quelques consid©rations sur le commerce de la soie au temps
de la dynastie des Han,” in M©langes de Sinologie offerts à Paul Demi©ville, II
(Paris: Bibliothèque de l™Institut des Hautes Etudes Chinoises, 1974), p. 120.
Hulsew©, “Quelques considèrations sur le commerce de la soie,” p. 125.
Yü, Trade and Expansion in Han China, pp. 99“105.
The Hsiung-nu royal residence was transferred north of the Gobi sometime
after 119 b.c. and, in 106/105 b.c., it was moved once more, this time to the
upper reaches of the Orkhon River, in Mongolia. See Daf¬nà, Il Nomadismo
Centrasiatico, p. 61.
Shih-chi chu-yi 110, 2331“2.


“states” included not only nomadic but also sedentary populations.
Arguably, once the cessation of the ho-ch™in policy brought to an end the
payment of tribute from China, the Hsiung-nu aristocracy became increas-
ingly dependent on other states, whose military inferiority was uncontested,
to replace the lost income. In addition, these areas could provide the farm-
land to support Hsiung-nu troops, who could, if the circumstances required
it, turn themselves into farmers, as they did in the following case: “In the
time of Emperor Chao the Hsiung-nu took the further step of sending four
thousand cavalrymen to work the land at Chü-shih, and after his accession
Emperor Hsüan despatched ¬ve commanders with troops to attack the
Hsiung-nu. Those who were working the land at Chü-shih ¬‚ed in alarm
and Chü-shih resumed relations with Han.”154 Although the Hsiung-nu had
other settlements in the steppes that also provided them with farming and
handicraft products “ as demonstrated by the excavations of burial sites,
such as Noin Ula,155 or of the forti¬ed settlement of Ivolga156 “ it is unlikely
that these centers had very large productive bases.

Hulsew© and Loewe, China in Central Asia, p. 185.
This is located in the northern part of Mongolia (about 100 kilometers north of
Ulan Bator); other important burial sites are those of Sudzhinsk and Deretsuj,
both situated in Transbaikalia, in the Buriat Mongol A.S.S.R. On the Noin Ula
kurgans, see S. I. Rudenko, Kul™tura Khunnov i noinulinskie kurgany (Moskva-
Leningrad: Nauka, 1962); A. Salmony, “The Small Finds of Noin Ula,” Parnas-
sus 8 (1936), 2: 15“20; W. P. Yetts, “Discoveries of the Kozlov Expedition,”
Burlington Magazine 48 (1926): 168“85; K. V. Trever, Excavations in Northern
Mongolia, 1924“1925 (Leningrad: Memoirs of the Academy of History of Mate-
rial Culture 3, 1932).
The Ivolga gorodishche was a Hsiung-nu forti¬ed village located near Ulan Ude
(in the Buryat A.S.S.R.). Russian archaeologists began working on it in the 1920s.
The extensive excavations at the Ivolga site yield ample information on the agri-
cultural activities of the Hsiung-nu. This village had primarily an agricultural
and handicraft economy, although the raising of domestic animals, hunting, and
¬shing also played important roles. The village™s speci¬c function, within the
context of a nomadic society, was to supplement the steppe peoples with those
products, such as grains, textiles, and various artifacts, that a pure nomadic
economy could not supply. Villages such as this were “trading centers” in the
steppe, where the wandering nomads could acquire the products they needed and
“overcome the narrow economic basis of nomadic economy.” The population of
the village “was composed of settled Hsiung-nu, of the aboriginal population
conquered by the Hsiung-nu, and of alien craftsmen from the ranks of deserters
and prisoners of war.” From craniological investigation, however, it appears that
the population of the village was racially homogeneous and belonged to the
South-Siberian branch of the Mongoloid race. See A. V. Davydova, “The Ivolga
Gorodishche. A Monument of the Hsiung-nu Culture in the Trans-Baikal
Region,” Acta Orientalia Academiae Scientiarum Hungaricae 20 (1968):
209“45. A complete bibliography on this archaeological complex is available in



During Wu-ti™s reign the Han offensive not only in¬‚icted serious military
losses on the Hsiung-nu, but also disrupted the “tribute system” of the
Hsiung-nu empire, which had thus far been the cornerstone for the politi-
cal cohesion among the various tribes united by the charismatic leader
Modun. A crucial blow to the nomads™ power was the severance of their
tributary relations with the sedentary regions of Central Asia, from which
the Hsiung-nu had been able to acquire the type of tributary and commer-
cial goods no longer available from China. At the same time, large portions
of nomadic grazing land to the north of the Yellow River and in the west
were seized by the Chinese, putting an additional strain on the nomadic
economy. In the long run, those pressures brought about the political
disintegration of the Hsiung-nu as a uni¬ed steppe empire. Although the
forward strategy of Wu-ti paid some dividends in foreign policy, and
the newly conquered territories were incorporated within the structure of
the Han administration, the strategy also almost bankrupted the Han
because of the rapid depletion of the state™s ¬nances.
The relevance of the Hsiung-nu to Han politics and society during the
lifetime of Ssu-ma Ch™ien can hardly be overestimated. Besides the chapter
that is speci¬cally devoted to them, the Shih chi deals extensively with the
Hsiung-nu in the imperial annals, in the biographies of important political
¬gures, and in the economic chapters (e.g., chapters 30 and 129). Because
of the wars™ impact on Chinese society and because of the objective threat
the Hsiung-nu posed, the nomadic “factor” required extensive treatment.
It is not surprising, therefore, that in the Shih chi the Hsiung-nu are main
characters, not “extras”: they were a new phenomenon that needed to be
explained, or at least to be brought within the boundaries of a world vision
that could account for their existence and make them into the object of his-
torical investigation. However, as momentous as they may have been, the
various aspects of the relations between the Han and the Hsiung-nu do not
fully explain, in themselves, the method of investigation adopted by the
Grand Historian. How Ssu-ma Ch™ien created, without any discernible
model or precedent, an ethnic and a political history of the Hsiung-nu, is
the object of the next section of this book.

Davydova™s book Ivolginskii kompleks (gorodishche i mogil™nik) “ pamiatnik
khunnu v Zabaikal™e (Leningrad: Isd-vo Leningradskogo Universiteta, 1985). On
the presence of agriculture among the Minusinsk early nomads, see also M. P.
Zavitukhina, “The Tagar Culture,” and “The Tashtyk Culture,” in AA. VV.,
Frozen Tombs. The Culture and Art of the Ancient Tribes of Siberia, (London,
1978), p. 94. On the economy of the early nomads, see Nicola Di Cosmo,
“Ancient Inner Asian Nomads: Their Economic Basis and Its Signi¬cance in
Chinese History,” The Journal of Asian Studies 53.4 (1994): 1092“1126.
part iv
chapter seven

In Search of Grass and Water
Ethnography and History of the North in the
Historian™s Records


This section is dedicated to an analysis of how the history of the Hsiung-
nu came to be written by Ssu-ma Ch™ien. As discussed in Part III, the
Hsiung-nu had become a phenomenon whose effects on Han life “ military,
economic, and political “ could not be ignored. However, by itself that con-
sideration is surely insuf¬cient to establish how the historian constructed
his narrative of the northern nomads, and how he was able to incorporate
this narrative within the general scope of his opus magnum.
The issue is important, at the very least, in two respects: there is no
obvious precedent that Ssu-ma Ch™ien could have used for inspiration, and
the pattern established by Ssu-ma™s Hsiung-nu narrative became the model
for representations of northern peoples and Inner Asian states in the sub-
sequent Chinese historical literature.1 As we will see, two orientations can
be detected in accounts of the Hsiung-nu, and of Inner Asia in general: one
empirical, descriptive, and data oriented, the other normative, ideological,
and in¬‚uenced by currents of contemporary thought. Both orientations
were consistent not only with the declared goals of the historian but
also with the general thinking of an age, the early Han, inclined to the
construction of universal cosmological paradigms and uni¬ed historical
patterns. Hence the account of the Hsiung-nu appears as a combination
of various contemporary concerns and intellectual pursuits.

See, for instance, Michael R. Drompp, “The Hsiung-nu Topos in the T™ang
Response to the Collapse of the Uighur Steppe Empire,” Central and Inner Asian
Studies 1 (1987): 1“46; and David B. Honey, “History and Historiography on
the Sixteen States: Some T™ang Topoi on the Nomads,” Journal of Asian History
24.2 (1990): 161“217.

I have therefore divided the materials on the Hsiung-nu into two differ-
ent sets of data. The ¬rst, in this chapter, presents information on the
lifestyle, history, and ethnography of the nomads, much of which must have
been acquired through direct investigation. The second refers to the posi-
tion of the Hsiung-nu and Inner Asia in general as metaphysical agents and
new ingredients of cosmological architectures. As is well known, Ssu-ma™s
Shih chi is far from being a consistent historical narrative. In addition to
its internal contradictions, the narrative is often overlaid with concerns that
deviate from any mimetic description of events. These are not only moral
concerns but also, more broadly, cosmological concerns that stem from a
particular philosophical tradition based on construction of a unitarian
vision of the cosmos and of the principles regulating it. The narrative on
Inner Asia is not exempt from these “normative” concerns, which will be
discussed in Chapter 8.
To appreciate the “empirical” strain in Ssu-ma Ch™ien™s construction
of the Hsiung-nu narrative, however, we must ¬rst look at the role of the
historian, at what Ssu-ma was trying to achieve, and at the general inter-
pretations that have been offered concerning the role of early Han histori-
ography. These questions are essential to understand the likely intellectual
and methodological frame that informed at least part of the Hsiung-nu nar-
rative. Hence I will begin this discussion with some brief considerations
concerning the position of the historian and of history writing in early

The Role of the Historian (Shih) in Early China

The traditional interpretation of the Shuo-wen, followed by Kwang-chih
Chang and others, attributes the meaning of “archivist” to the ancient
graph for shih.2 The existence of a class of functionaries who specialized
in writing about historical matters is documented as early as the Shang
dynasty (c. 1600“1045 b.c.), but the different meanings assigned to the
word shih do not fundamentally affect the question of the role of the his-
torian in pre-imperial times.3 Wang Kuo-wei in China and Naito Torajirô
in Japan came to the conclusion that the character shih (“historian”) orig-
inally represented a hand holding a vessel used to contain tallies at archery
contests and that the of¬cial designated by this character in Shang times

Kwang-chih Chang, Art, Myth and Ritual. The Path to Political Authority in


. 10
( 15)