. 9
( 15)


chapter six

From Peace to War
China™s Shift from Appeasement to
Military Engagement


With the accession of Emperor Wu in 140 b.c., a half-century-long tradi-
tion of foreign relations based on the search for diplomatic solutions and
negotiated agreements came to an end. In the phase that followed, the Han
dynasty assumed an outward-looking, expansion-driven, military-oriented
posture. The Han“Hsiung-nu bipolar system of foreign relations came to
an end, formally, with the breakup of the Hsiung-nu empire and the formal
acceptance by Hu-han-yeh ch™an-yü in 51 b.c. of a position of inferiority
to the Han emperor Hsüan-ti (73“49 b.c.). This development was the direct
result of the successful military and political campaigns during Han Wu-ti™s
reign (140“87 b.c.). The shift from the “peace through kinship” strategy
to the military solution, which took place during the lifetime of Ssu-ma
Ch™ien, is one of the momentous events of Han history, and one whose
repercussions were felt at every level of political and social life. This change
in the means through which relations with the Hsiung-nu were conducted
led to territorial expansion, but it also created economic problems and fos-
tered tensions between government policy makers and the court on the one
side, and the literati on the other. The respective positions were represented
in stark contrast in the Discourses on Salt and Iron held in the early ¬rst
century b.c.1
The plain narrative of the confrontation between Han and Hsiung-nu
during the reign of Han Wu-ti is well known and does not need to be

Especially relevant to our discussion is chüan seven (sections 43“48) of the Yen-
t™ieh lun (ed. Ssu-pu pei-yao).


repeated here;2 moreover, parts of it will be addressed in the following
chapter, as we analyze Ssu-ma Ch™ien™s attitudes toward the Hsiung-nu. It
should be noted that in my discussion of the Hsiung-nu I rely most heavily
on the Shih chi and make recourse to the Han shu only occasionally, when
it provides speci¬c information needed for clari¬cation. The narrative
on the Hsiung-nu in the Han shu (chapter 94) down to about 90 b.c. is
essentially parallel to that of the Shih chi, but the Han shu is much more
detailed concerning several aspects of Hsiung-nu“Han relations because of
the reports sent from Central Asia after the establishment of the Protec-
torate General in 59 b.c. (a point made most eloquently in Michael Loewe™s
introduction to China in Central Asia). However, the period after the com-
pletion of the Shih chi (c. 90 b.c.) has not been included in the present work
other than marginally to explain some of the consequences of the events
taking place in Wu-ti™s time, because the main focus of this and the fol-
lowing (Part IV) section lies in the transformation of the frontier in Ssu-ma
Ch™ien™s time and in its representation in the Shih chi. Given this chrono-
logical limit, and given the deep differences in period, historical outlook,
and personal interest taken by Ssu-ma Ch™ien and Pan Ku, taken here to be
the main authors of, respectively, Shih chi and Han shu, the Han shu has
not been included as a main source. Nonetheless, I do not wish to give the
impression that the Han shu is not an important source. It is, unquestion-
ably, the major source for Han history. It has less validity in this study
because it is not the ¬rst account to detail the history of Han“Hsiung-nu
relations; that is, it was Ssu-ma Ch™ien, not Pan Ku, who made the essen-
tial shift leading to creation of a “history” of Inner Asia within Chinese
historiography. For this reason, we cannot ignore the lingering controversy
concerning the authenticity of some chapters of the Shih chi, in particular
the possible derivation of chapter 123 of the Shih chi from chapter 61 of
the Han shu. This issue had been discussed by a number of scholars and,
to my mind, has not yet been settled.3 This is not surprising for a text such

There are several accounts of the relations between the Han and the Hsiung-nu;
the basic events are narrated in a clear and synthetic manner in Ying-shih Yü,
“The Hsiung-nu,” in Cambridge History of Early Inner Asia, ed. Denis Sinor
(Cambridge: Cambridge University Press, 1990), pp. 118“49. The best study of
the Chinese penetration in the Western Regions is still Anthony Hulsew©, with
an introduction by M. A. N. Loewe, China in Central Asia: The Early Stage, 125
B.C.“A.D. 23. An Annotated Translation of Chapters 61 and 96 of the History of
the Former Han Dynasty (Leiden: Brill, 1979).
On this and related issues, see Hulsew© and Loewe, China in Central Asia, pp.
13“33; Anthony Hulsew©, “The Problem of the Authenticity of Shih-chi,” T™oung
Pao 66 (1975): 83“147; D. D. Leslie, and K. H. J. Gardiner, “Chinese Knowl-
edge of Central Asia,” T™oung Pao 68.4“5 (1982): 254“308; J. R. Gardiner-
Gardner, “Chang Ch™ien and Central Asian Ethnography,” Papers of Far Eastern
History 33 (1986): 23“79; Yves Hervouet, “Le valeur relative de textes du

as the Shih chi whose transmission history remains fuzzy at best for over
a thousand years. However, the controversy is relatively unimportant to
this account, as it refers mainly to the biographies of Chang Ch™ien and
Li Ling, who enter marginally in this work, and about whom additional
information is available in other, non-suspect parts of the Shih chi. Most
of the information discussed here is derived from chapter 110 of the Shih
chi, which does not present textual problems of a magnitude to justify
doubts about its authenticity with respect to chapter 94 of the Han shu.
Instead, in this chapter, I will focus on two problems that to this date remain
First, I will try to clarify the reasons for the shift from the ho-ch™in policy
to a strategy of direct military engagement and territorial expansion. Why
did the “accommodation” strategy come to an end? By analyzing the con-
ditions of its implementation, the arguments put forth by supporters, and
some aspects of the later debate, I will try to link competing orientations
in Han strategic thinking with the historical context of Chinese-Inner Asian
The second problem of interest is the extent to which the campaign
against the Hsiung-nu was carried out by Han Wu-ti and his generals. Its
duration, territorial expansion, forces employed, and expenses required are
nothing short of exceptional even considering the intense military activity
that had marked the history of China until then. As the Han armies marched
through the deserts of Kansu and showed their insignia at the gates of the
oasis-cities of the Western Regions, a new world opened to China™s imagi-
nation. By reaching as far as the T™ien-shan and the Tarim Basin, Han Wu-
ti™s expansion dwarfed even the feats of two of the most blatantly
expansionistic pre-Han rulers, Duke Huan of Ch™i and the First Emperor
of Ch™in. The motivation for such an accomplishment cannot be ascribed
solely to megalomania. The Han political and strategic choices and the deci-
sion to ¬ght a protracted war occurred in a climate of changing policies
aimed at consolidating imperial unity and strengthening China™s economy.
These measures were arguably essential in allowing the Han to sustain the
war effort. Internal political events within the court, after the death of the
empress dowager in 135 b.c., and the waning of her in¬‚uence also were

Che-Ki et du Han-chou,” in Melanges de Sinologie off©rts a Monsieur Paul
Demieville, part II (Paris: Bibliotheque de l™Istitute des Hautes Etudes Chinoises,
1974), pp. 55“76; Paolo Daf¬nà, “The Han Shu Hsi Yu Chuan Re-Translated. A
Review Article,” T™oung-p ao 68.4“5 (1982): 309“39; Kazuo Enoki, “On the
Relationship between the Shih-chi, Bk. 123, and the Han-shu, Bks. 61 and 96,”
Memoirs of the Research Department of the Toyo Bunko 41 (1983): 1“31; Edwin
G. Pulleyblank, “Han China in Central Asia,” International History Review 3
(1981): 278“96.


factors. Han strategy in the war with the Hsiung-nu has been ascribed to
these climactic changes which took place during the ¬rst twenty years of
Wu-ti™s role. Concurrently, however, it is essential that we take account of
Inner Asian political and economic realities, which exposed the Hsiung-nu
to weaknesses that the Han were able to exploit, especially in the early
phases of the confrontation, with considerable success. Finally, this chapter
outlines the restructuring of the northern frontier as it began to take shape
in Wu-ti™s period, including the new administrative organization of the
frontier areas.

Why Did the Ho-ch™in Policy Come to an End?

The dramatic shift in foreign policy that brought China from a defensive
posture based on appeasement to an offensive strategy based on a total
military commitment plunged the Han into a war whose ¬nal victory,
attained at enormous human and economic costs, appeared Pyrrhic to many
contemporary observers, including Ssu-ma Ch™ien. Even though Han states-
men failed to recognize for a long time exactly why the Hsiung-nu kept
invading their territory, they could not have failed to notice that the
presumed advantages that the ho-ch™in policy was meant to yield were
simply not being realized. The Hsiung-nu™s continuous demands for pay-
ments and their frequent raids (whether or not their demands had been
met) could be explained only by resorting to the cultural stereotype, born
out of the Spring and Autumn tradition, of the uncouth, greedy, and violent
foreigner. This was a convenient rationalization, not the least because it
prepared the Chinese psychologically for the unavoidable consequence
of the failure of peace: a painful and prolonged war against a powerful
As we have seen in the previous chapter, the ho-ch™in doctrine continued
to hold sway throughout the ¬rst part of the Former Han, when the
metaphors of equality and complementarity between the two sovereigns had
supplied the Han with a workable, if not ideal, basis for negotiation with
the Hsiung-nu. As peace failed to last, however, and the borders continued
to be routinely violated, the bipolar system of foreign relations “ repre-
sented in the political symbolism whereby the two rulers were portrayed as
“Earth” and “Heaven,” or as “brothers,” and their kingdoms as two com-
plementary universes “ was no longer sustainable. Disharmonious relations
became the norm, precipitating a deep crisis of the entire Han approach to
relations with the Hsiung-nu. It is at this point that military solutions were
invoked, and the paradigm of foreign relations subsequently shifted back
to the template provided by the strategy and arguments for expansion that
had their precedents in the Spring and Autumn period.


The Debates over Ho-Ch™in (135“134 B.C.)

Although an acknowledgment of the desirability of a policy shift away from
ho-ch™in and toward the military engagement of the Hsiung-nu was already
present in the positions held by Ch™ao Ts™o and Chia Yi, it is the two-phase
debate held in 135 and 134 b.c. that provides the most cogent explanation
of the issues at stake concerning Han relations with the Hsiung-nu at the
time of Han Wu-ti,4 and the subsequent endorsement by the ministers and
high of¬cials of the Chinese emperor™s aggressive military stance.5 The pro-
tagonists, Wang Hui and Han An-kuo, argued vigorously over whether to
discontinue the ho-ch™in policy and attack the Hsiung-nu, or continue
to “appease” them. This debate is reported in the Han shu as a prologue
to the misguided military episode that took place at Ma-yi, a border town
where Han troops were disastrously defeated in an attempt to ambush the
Hsiung-nu and capture their chief.
This debate, the fullest contemporary account of the factors held respon-
sible for the switch to an offensive strategy, began in 135 b.c., and was
occasioned by the ch™an-yü™s request to renew the ho-ch™in treaty. As the
emperor asked for his of¬cials™ advice, Wang Hui, a man of the north-
eastern region of Yen who had served for years as a border of¬cial and had
a reputation as an expert in nomadic matters, expressed his point of view:
“When the Han conclude a peace [agreement] with the Hsiung-nu, usually
after just a few years the Hsiung-nu violate the treaty. It would be better
to reject their promises and send soldiers to attack them.” The military
option was rejected by Hann An-kuo, an able rhetorician, who supported
appeasement on the following grounds:
Fighting at a distance of a thousand li, the army will not gain any pro¬t.
Now, the Hsiung-nu depend upon the hooves of their own horses and cherish
the feelings of birds and beasts. They move around like a multitude of birds
and are dif¬cult to capture and control. In order to conquer their territory it
is not suf¬cient to expand. In order to capture their masses it is not enough
to set up a barrier. From high antiquity they have not been subjugated. If the
Han have to strive over many thousands of li in order to gain an advantage,
men and horses will be exhausted and the Bandits will completely control
them. The circumstances will necessarily be perilous. I therefore maintain that
this [attacking] would not be better than ho-ch™in.6

The following discussion is based on Han shu 52, 2398“2403.
This exchange to a certain extent foreshadows the issues at stake in the
Debates on Salt and Iron. See Michael Loewe, Records of Han Administration,
vol. 1: Historical Assessment (Cambridge: Cambridge University Press, 1967),
p. 54.
Han shu 52, 2398.


Admittedly, there was no easy solution. On the one hand, according
to Wang Hui, the ho-ch™in policy was not a permanent solution because
previous treaties had not prevented Hsiung-nu incursions. On the other,
according to Han An-kuo, attacking was no solution either: the futility of
war against the northerners, who move like “birds and beasts,” was an old
theme, but it was reinforced here by the vivid awareness, the result of many
bitter lessons, that the Hsiung-nu had a military advantage when they
fought in their own territory. Han An-kuo identi¬ed clearly the source of
the Han disadvantage in the logistic dif¬culties that would be met by
Chinese armies if required to spend a long time in the ¬eld. His “paci¬st”
argument won the day because his sentiment was shared by the other of¬-
cials, and the ho-ch™in treaty was temporarily renewed.
A second round of the debate took place the following year, 134 b.c.,
when Nie Weng-yi, a prominent man in the frontier city of Ma-yi, sent
Wang Hui to court to propose a plan for ending the con¬‚ict. Nie™s plan
consisted of using promises of gifts to induce the ch™an-yü and his army to
approach Ma-yi and when they did so, slaughtering them. Nie thought that
such a stratagem might succeed in the climate of trust between Han and
Hsiung-nu that had set in after the renewal of the peace treaty. Possibly
because of the tradition of foreign relations with the northerners that, as
we have seen in Chapter 4, justi¬ed the use of trickery, the violation of the
treaty on which this strategy was based was not seen as ethically improper.
At least in part, however, this course of action must have seemed feasible
also because of the diminished value of the treaty itself as a viable means
of preserving good relations and because of the Hsiung-nu™s own breaches
of previous agreements.
It is clear from the circumstances in which the debate was held that the
emperor was aggrieved by the constant, seemingly unlimited demands of
the ch™an-yü, who had continued to raid and plunder the border regions
while waiting for the ho-ch™in treaty to be rati¬ed by the Han. Once again,
Han Wu-ti asked for the of¬cials™ opinion, and this time Wang Hui and
Han An-kuo™s positions are reported in the Shih chi at greater length.
Because this is such an important political debate, marking the beginning
of a new era not only in Han history but also in the history of the relations
between China and northern nomads, I will examine the two of¬cials™ posi-
tions in detail.
Wang Hui™s ¬rst point is only marginally different from his earlier
position. In essence, Wang presents an argument based on the following
contrast: during the Warring States period even the state of Tai could
hold off the nomads, and keep them from raiding and pillaging, but Han
China, although politically united and invested heavily in frontier defenses,
was unable to stop the Hsiung-nu. Hence, Wang stated, no alternative
remained but to attack them. Han An-kuo replied by ¬rst reminding
his audience of the humiliating defeat suffered by Han Kao-tsu at

P™ing-ch™eng,7 which led to the inauguration of ho-ch™in, a policy that, by
providing peace, had “bene¬ted ¬ve generations.” An-kuo emphasized the
idea that the empire should have territorial boundaries “ “the wise and sage
man regards the empire as being limited” “ and that Han emperors should
refrain from attempting conquest. He praised Han Wen-ti for not having
conquered a single inch of Hsiung-nu territory and for having eventually
renewed the peace treaty. This anti-expansionist stance was at the heart of
his support for ho-ch™in, and he reiterated it throughout the debate.
In his rebuttal of Han An-kuo™s defense of ho-ch™in, Wang Hui resorted
to an argument reminiscent of King Wu-ling™s debate for the adoption of
nomadic cavalry garments:
Not so! I have heard that the Five Emperors did not follow each other™s
rituals, and the Three Kings did not repeat each other™s music. This is not
because they antagonized each other, but because every one followed what
was appropriate to the epoch. Moreover, Kao-ti personally dressed in a strong
armor, and armed with sharp weapons, hiding in fogs and mists, immersed
in snow and frost had fought continuously for over ten years, and therefore
he could not avenge the outrage of P™ing-ch™eng. Without force there is no
ability, and therefore he put the hearts of the empire at rest. Today, however,
there are frequent alarms along the frontiers, the soldiers are wounded and
killed, and in China funerary processions follow one after the other. This is
what grieves the benevolent man. For this reason I say that it is appropriate
to strike.8
The “circumstances” required proper action, and the reason Liu Pang
had been defeated and forced to accept the peace terms was that the empire
was not in a position to continue to ¬ght after the many years of civil war
at the end of the Ch™in. But the peace Liu Pang had secured was simply not
there anymore. An-kuo replied by arguing that a change of policy would
make sense only if a higher return could be guaranteed: “if the pro¬t is not
tenfold one should not change trade, and if the achievement is not one hun-
dredfold one should not change habits.” But what would China achieve by
changing course? This was anything but clear. In addition, by comparing
the Yi and Ti of old with the far more powerful Hsiung-nu of his day,
An-kuo captured both the essence of the changes that frontier relations
had undergone and the reason why the Han now faced a more dif¬cult
From the rise of the Three Dynasties the Yi and Ti did not share the calen-
dar or the color of garments [with us]. Without might they could not be con-
The P™ing-ch™eng prefecture was part of the Yen-men commandery, situated east
of present Ta-t™ung. See Hans Bielenstein, “The Restoration of the Han Dynasty,
Vol. 3: ˜The People,™ ” in Bulletin of the Museum of Far Eastern Antiquities 39,
part II (1967): 86, n. 3.
Han shu 52, 2400.


trolled, and strength alone could not make them submit, and yet, as people
who cannot be shepherded (pu mu chih min)9 from remote regions and inac-
cessible lands, they were thought as insuf¬cient to trouble the central states.
But now the Hsiung-nu are light and quick, brave and hasty soldiers, they
arrive like a sudden wind and leave like a disappearing lightning. Their occu-
pation is raising animals, they go hunting with bow and arrow; they follow
their animals according to the availability of pasture, and their abode is not
permanent; they are dif¬cult to capture and control. As for the present, since
long they have caused the border regions to abandon tilling and weaving in
order to support the common activities of the nomads. Their strength cannot
be matched in a balanced way. For this reason I say it is not convenient to
If Yi and Ti were so dif¬cult to control, even though they had no unity,
how much more dif¬cult would it be to oppose the military strength of the
Hsiung-nu? An-kuo™s argument displays a keen knowledge of the nomads™
strengths, which he saw as lying essentially in the resources of their
economy and in the alleged incorporation of the northern frontier regions
within the sphere of their economic activities.
Wang Hui replied with a variation on his previous argument: brilliant
men act according to the circumstances, and hence Duke Mu of Ch™in was
able to defeat the Western Jung and expand his territory. Likewise, Meng
T™ien was able to open up thousands of miles of land and to build forti¬-
cations so that “the Hsiung-nu did not dare water their horses in
the Yellow River.” From this Wang Hui concluded that the Hsiung-nu
could be subjugated only by force and could not be cultivated with benev-
olence. China was now much stronger than the Hsiung-nu, and could afford
to ¬ght them, as they were “like an abscess which must be burst open with
strong crossbows and arrows, and absolutely should not be left to fester.”
This argument was supported by the perceptive realization that the Hsiung-
nu did not have such a ¬rm hold over their subject peoples, and, once
the ¬rst blows had been dealt to the Hsiung-nu leadership, formerly van-
quished nomadic tribes, such as the Yüeh-chih, “will be able to rise [against
the Hsiung-nu] and will submit [to the Han].” It was perfectly clear to
Wang Hui that the political basis of the Hsiung-nu was quite unstable, and
that it would be possible for the Han to exploit divisions among the
Han An-kuo began to lose ground as he continued to repeat the by now
trite motif that the nomads could not be defeated in their territory: “the
clash of two strong winds will weaken [them] to the point that they cannot

The commentary in Han shu 52, 2401, renders this as “people who cannot be
shepherded,” which I have followed. At the same time, a lingering doubt remains
as to whether this expression may refer to “people who did not tend herds,” that
is, people who were not pastoral nomads.


raise a hair or a feather; an arrow shot from a strong bow at the end of its
¬‚ight would not be able to pierce the white plain silk of Lu,” that is to say,
even the strongest army would eventually weaken and be exhausted to the
point that it could not accomplish anything. In this way An-kuo sought to
rebuke the “hawkish” military faction ready to mount ostensibly useless
expeditions that would penetrate deeply into the enemy™s territory but fail
to achieve their stated goals.
Piqued, Wang Hui replied that that was not what he was proposing. His
position was quite different:

Now the reason why I propose to attack them is absolutely not to go out and
penetrate deeply, but to go along with the wishes of the ch™an-yü, and induce
him to come close to the border. In the meanwhile, handpicked bold caval-
rymen and brave infantrymen should lie in ambush and ready for action; we
shall also examine how to protect them from any dangers so that they can
guard against them. Once this situation has been arranged, whether we
encamp to the [Hsiungu-nu] right or to their left, whether we are positioned
in front of them or cut off their rear, the ch™an-yü can be captured, and then
all the Hsiung-nu will be certainly taken.

Wang Hui™s plan for a limited military engagement won the emperor™s
approval. In the end, it seems that Wang Hui defeated his adversary because
Han An-kuo had been unable to demostrate that ho-ch™in could actually
guarantee peace. An-kuo™s main argument was a negative one, relying on
the point that deep military engagement was not desirable because it would
not solve the problem and would simply result in a loss of people. This
position was based on the realistic understanding, with which his contem-
poraries seemed to agree, that the Han army could ¬ght the Hsiung-nu in
the nomads™ own territory only for a limited time; moreover, if a war were
fought on the nomads™ teritory, they would simply keep moving farther and
farther away, so the consequences of a prolonged war would be disastrous
for the Han. Recognizing the strength of this argument, Wang Hui pro-
posed a limited engagement on the frontier aimed at capturing the head of
the Hsiung-nu, arguing that this act would trigger certain reactions in the
enemy camp “ such as the surrender of other Hsiung-nu tribesmen and the
rebellion of the Yüeh-chih “ that would eventually destroy the political
unity of the Hsiung-nu.
The Han decided, then, to pursue the military option. Unfortunately for
them, the whole plot was hopelessly botched, and at the ensuing battle at
Ma-yi the ch™an-yü not only did not fall into the trap, but deeply humili-
ated the Han armies, in¬‚icting a resounding defeat on them. And yet, in a
somewhat perverse way, this disastrous outcome must have appeared as a
con¬rmation of Wang Hui™s fundamental argument. Peace treaties with the
Hsiung-nu did not work, and because the nomads continued to create a
serious threat on the borders and limited military action of the type

attempted at Ma-yi was not successful against them, there was no other
solution but to confront them on their own terms, through full-scale mili-
tary engagement.
The debates between Wang Hui and Han An-kuo make it clear that if
the ho-ch™in policy had worked, managing to ensure the peace and pros-
perity of the borders, the proponents of the offensive strategy could not
have prevailed; indeed, at the time majority opinion was by no means in
favor of military action. That the emperor eventually followed Wang Hui™s
advice was because the ho-ch™in policy was no longer particularly effective,
and even An-kuo™s support for it was not particularly enthusiastic: it had
come to be seen as the lesser of two evils rather than as a truly successful
foreign policy. The treaty system that the Han had established with the
Hsiung-nu at the time of Han Kao-tsu had been quite acceptable to the
Han, and, even though some at court may not have liked it, the majority
had nonetheless learned to live with it. Hence possible arguments to explain
its rejection by Han Wu-ti, such as “the appeasement policy was a humil-
iating one” or “the Chinese were tired of paying” are immaterial to the dis-
cussion; the issue that forced the change of policy was not the ho-ch™in
policy per se, but that it was not working.10 Why was it not working?

Limits and Ultimate Failure of the Ho-Ch™in Policy

From a Chinese perspective, appeasement had failed because the Hsiung-
nu did not respect the treaties: the ho-ch™in policy did not guarantee the
inviolability of China™s borders and the tribute payments were an invest-
ment that did not pay off. These “investments,” by the way, were by no
means tri¬‚ing, and had come to represent a considerable burden on the Han
economy, though one that could be sustained, probably, without eroding
too much the factor of growth that the Han had come to enjoy down to
the reign of Wu-ti.11
Several theories have been put forward to explain the nature of the
Han“Hsiung-nu relationship during the Former Han. Generally speaking,
these agree that the ho-ch™in policy was entered by the Han at a time of
military weakness and political vulnerability but disagree as to the relative

One should also note that, although the shift to an aggressive stance occurred
early in Wu-ti™s time, the debate between “accommodation” and warfare contin-
ued even afterwards. For another debate concerning the different tactics to be
employed against the Hsiung-nu see Han shu 94B, 3825.
For a calculation of the tribute paid to the Hsiung-nu see Bielenstein, “The
Restoration of the Han Dynasty,” pp. 91“92; Yü Ying-shih, Trade and Expan-
sion in Han China: A Study in the Structure of Sino-Barbarian Economic Rela-
tions (Berkeley: University of California Press, 1967), pp. 45“49.

importance of “appeasement” in the larger context of the two peoples™ rela-
tions. According to some, both Han and Hsiung-nu accepted the settlement
solution as a tactical move that from the start was conceived not as a per-
manent solution but as a temporary arrangement. Whereas the Han were
forced to accept a tributary relationship for lack of strength, the Hsiung-
nu found it advantageous for the economic gains. But neither of them
renounced to pursue their own separate interests, even though they might
be in violation of the treaties. On the one hand, the Hsiung-nu did not
refrain from raiding the border, and the Han, on the other hand, planned
to reverse the relationship once the “family connection” established with
the Hsiung-nu court had achieved the goal of relegating the ch™an-yü to a
subordinate position.12 Going even further, other historians look at ho-ch™in
as pure expediency within a general context of permanent warfare, and they
criticize the view that the search for peace and reconciliation actually was
pursued in earnest by either of the two nations. In fact, in the early period
ho-ch™in treaties sealed ¬rst the tributary status of the Han vis-à-vis the
Hsiung-nu and later, after the “surrender”of the southern Hsiung-nu in 51
b.c., the Han primacy. Yet in each period the ho-ch™in treaties constituted
only one side of a broader and more complex relationship.13 Others again
emphasize the economic and political advantages reaped by the Han during
the early period of the ho-ch™in policy. In this perspective, the policy
acquires a more holistic meaning, as it is held responsible for the economic
growth and political consolidation of China, as well as for the promotion
of friendly relations and “cultural exchanges.”14 But regardless of the
relative weight they attributed to “appeasement” as the leading strategy
pursued by both Hsiung-nu and Han until the time of Han Wu-ti, all com-
mentators agreed that peaceful relations broke down more or less regularly,
without seeking an explanation beyond the often-repeated claim that the
Hsiung-nu were insatiable.
This consideration leads us to a question that is central to our under-
standing of the shift to an offensive foreign policy: why is it that the Hsiung-
nu “ if we are to believe the sources “ did not abide by the terms of the
agreements? This question has rarely been engaged in actual historical
terms. More often, scholars have sought explanations in the analysis of
broader patterns of economic and political relations between the Hsiung-
nu and China. In Chapter 4 we discussed the theory that places the cause
of the nomads™ violent raids against the Chinese borders with China™s

Shih Wei-ch™ing, “Kuan-yü Hsi Han cheng-fu yü Hsiung-nu ho-ch™in jo-kan
wen-t™i,” Hsia-men ta-hsüeh hsüeh-pao 1985.4: 21“29.
Chang Ch™ang-ming, “Shih-lun Hsi Han te Han Hsiung kuan-hsi chi ch™i kuan-
hsi chi ho-ch™in cheng-ts™e,” Chiang-huai lun-t™an 1983.6: 83“88.
Lo Ta-yün, “Hsi Han ch™u-ch™i tui Hsiung-nu ho-ch™in te shih-chih,” Yün-nan
min-tsu hsüeh-yüan hsüeh-pao 1985.4: 44“49.


unwillingness to open its border markets or to agree on paying a ¬xed
“tribute.”15 The problem with this theory is that it assumes the existence
of a mutually exclusive relationship between “trade” and “raid,” even
though throughout the duration of the ho-ch™in policy Hsiung-nu incur-
sions often occurred soon after China agreed to open border markets and
to increase its payments. Hence, the failure of the ho-ch™in policy to pre-
serve peaceful relations cannot be linked to a hypothetical failure of an
existing exchange mechanism. But the question of why the nomads raided
China, when tribute was being delivered and markets were open, remains
unanswered, and other explanations that posit a Hsiung-nu desire to
expand their land or to extort more money remain a matter of conjecture
that cannot be supported by the sources.16
By shifting focus from the economic plane to the diplomatic and mili-
tary arenas, however, it is possible to offer a different explanation. As we
have seen, in the calculations of the supporters of the ho-ch™in policy, the
use of ho-ch™in was not simply a choice dictated by the need to establish a
modus vivendi with a more powerful state, but was a long-term strategy
aimed at absorbing the next generation of the Hsiung-nu leadership within
the Han court™s political sphere and neutralizing the Hsiung-nu as an inde-
pendent and inimical state. This strategy can even be regarded as an evo-
lution of Spring and Autumn marriage-diplomacy policies, and it indicates
that the Han were ready to wait and to sacri¬ce part of their revenues to
achieve a long-term political objective. However, the cost of this policy
proved too high once it became clear that the Hsiung-nu leadership would
continue to raid the border areas and did not have the slightest intention
of recognizing the Han state™s superiority. Ho-ch™in treaties appeared to be
hollow, and the whole strategy had to be called into question.
A close examination of the actual treaty violations by the Hsiung-nu sug-
gests that in various instances it was not the ch™an-yü himself who violated
a particular agreement, but his subordinate leaders or Chinese commanders
who had defected to the Hsiung-nu.17 Possibly, then, the central point in
this matter could be identi¬ed in the discrepancy in the power that each of
the parties “signing” the treaty had to guarantee that the letter of the treaty

Sechin Jagchid and Van Jay Symons, Peace, War and Trade along the Great Wall
(Bloomington: Indiana University Press, 1989).
None of these theories can accommodate the sequence of events that we see com-
monly, such as the systematic withdrawal of the Hsiung-nu troops after the inva-
sion of areas sometimes very close to the capital; or the conclusion of a ho-ch™in
agreement after the Han had actually been able to repel the Hsiung-nu, as hap-
pened with the treaty of 174 b.c. between Wen-ti and Modun, signed in the wake
of diplomatic parleys following the successful expedition of 177 b.c. against the
Hsiung-nu wise king of the right.
Han shu 94B, 3754.

would be respected by the entire body politic formally under his authority.
If it can be shown that one of the parties was lacking the full authority to
commute the international agreement into a “law” that each member of his
people subsequently had to abide by, it would then be obvious that any
strategy based on the assumption that pacta sunt servanda would be fun-
damentally ¬‚awed. In other words, to understand the causes of treaty vio-
lations one must ¬rst understand the essential premises on which such
treaties were based, namely, the nature of the authority of each ruler and
of the understanding of “sovereignty” by each of the parties.
Regarding China,18 one can arguably identify four different aspects as
constitutive of a notion of sovereignty in the early imperial period that
would allow the Chinese side to keep an international agreement.19 First,
the position of the emperor as supreme lawgiver; second, the unity of reli-
gious sanction and political power; third, the absence of another source of
authority within the political community; and fourth, the link between the
political authority of the sovereign and a more or less clearly delimited polit-
ical community.20

My intent here is to give a necessarily synthetic and an admittedly cursory descrip-
tion of some of the elements constitutive of the notion of sovereign authority in
Han China to highlight areas of difference between it and the political authority
of the ch™an-yü. This is not intended to sum up all the elements that can be
brought to bear to explain the nature of the monarch™s place in Han politics and
society, and even less its philosophical foundations. Rather, it is intended to high-
light the most glaring structural differences between the Han™s and the Hsiung-
nu™s notion of rulership with respect to their relative abilities to observe an
international treaty.
On the general issue of sovereignty in early imperial China, see Michael Loewe,
“The Authority of the Emperors of Ch™in and Han,” in State and Law in East
Asia. Festschrift Karl Bünger, ed. Dieter Eikemer and Herbert Franke (Wiesbaden:
Harrassowitz, 1981), pp. 80“111.
A similar notion of sovereignty can be seen in practice with the emergence of the
Roman Principate, especially from the second century a.d. With the transition
from the Republic to the Principate the territorial boundaries of Rome became
more clearly de¬ned, and the word “empire” itself (imperium) acquired a terri-
torial meaning originally absent (see Richardson, “Imperium Romanum: Empire
and the Language of Power,” Journal of Roman Studies 81 [1991]: 1“9).
Through the branching out of the imperial administrative institutions, and the
granting of the right of citizenship to all free people (including those living in
distant provinces), the society and politics of the empire became somewhat more
cohesive. The emperor became the head, rather than the mere agent, of the body
politic. His role as supreme lawgiver placed him above the law, and his diviniza-
tion ensured that no superior or alternative authority (either political or religious)
could exist above him. In the opinion of F. H. Hinsley, these are all compelling
arguments for the existence of a notion of sovereignty in the Roman Principate,
well before a theory of sovereignty was developed by sixteenth-century European


As for the ¬rst point, in the “legalist” thought that inspired many of the
political ideas of the early imperial period the notion of a divine lawgiver
is simply absent. According to one of the fathers of the legalist school, Shen
Tao, “law does not come down from Heaven nor does it arise from Earth.
It is nothing else but something that comes forth from among men, conso-
nant with their ideas.”21 However, Anthony Hulsew© has shown that in the
early imperial period the emperor himself acquired the institutional role of
supreme lawgiver. His ordinances and decrees had the authority of the law
and the power to overrule customary or traditional usages. The code of
laws became then the ultimate source of authority within the body politic,
and the only institutional power located above it was the emperor.22
Second, the creation of the empire set in motion a process whereby polit-
ical power and “religious” sanction were reunited and led to the establish-
ment of the principle that the religious sanction to rule could not rest with
a source of authority different from the wielder of political and military
power. The doctrine of Heaven™s Mandate, in which Heaven became the
ultimate source of temporal authority, although foreshadowed in some of
the earlier texts, such as the Book of Songs and the Book of Documents,
and especially in the philosophy of Mencius, developed relatively slowly on
the political plane. It acquired greater relevance, and became part of the
ideology of imperial legitimacy, most probably when the process of con-
solidation of the authority of the emperor had been completed, and it
acquired actual political relevance only at the time of Wang Mang™s acces-
sion (9 a.d.).23 The religious meaning implicit in a title such as huang-ti
already excluded, from the very beginning of the imperial period, the idea
that the ruler™s religious and temporal powers could be separated.24 Surely,
by presenting themselves as repositories of knowledge essential to the
correct management of state affairs, ministers and literati could attempt to
appropriate the right to the correct interpretation of the “will” of Heaven.25
The tension between the formal authority and “religious” sanction of the
emperor and the informal power of his advisors could generate dissent and

thinkers; see his Sovereignty (Cambridge: Cambridge University Press, 1986
[1966]), pp. 27“44.
A. Hulsew©, “Law as One of the Foundations of State Power in Early Imperial
China,” in Foundations and Limits of State Power in China, ed. S. Schram
(London: School of Oriental and African Studies, 1987), p. 12.
Hulsew©, “Law as One of the Foundations of State Power in Early Imperial
China,” pp. 11“32.
Michael Loewe, “The Authority of the Emperors of Ch™in and Han,” pp. 80“111.
Derk Bodde, China™s First Uni¬er: Li Ssu (Leiden: Brill, 1938), p. 31.
A classic study on the political in¬‚uence exercised by various of¬cials on politi-
cal affairs is Wolfram Eberhard, “The Political Function of Astronomy and
Astronomers in Han China,” in Chinese Thought and Institutions, ed. John K.
Fairbank (Chicago: University of Chicago Press, 1957), pp. 37“70.

disobedience, and even lay the foundations for a dynastic change, but in
general it did not produce a separation between the religious and temporal
powers of the emperor.
Third, within the Han political system the emperor also represented the
highest source of political authority in the state. He could issue edicts and
decrees, could appoint or dismiss ministers, and was the supreme judicial
authority. Such authority was, per se, unchallengeable even though the
person of the emperor might be subject to criticism as inadequate to the
task. The postulation that the theoretical separation between the person of
the ruler and the institution of kingship was the abstract point of equilib-
rium of a universal socio-political order was expounded in the writings of
political philosophers of the late Warring States period.26 The centrality of
emperorship to the Han political and moral universe, above and beyond
the qualities of the individual ruler, gave the emperor “sovereign” author-
ity, an authority that could not be challenged without challenging the
dogma of single and undivided rulership that constituted one of the ideo-
logical tenets of the state. In the political conceptions of the early Han, as
re¬‚ected, for instance, in the Huai-nan-tzu, the notion of law, derived
largely from a legalist frame, was tempered by the concerns with moral
values and interests of the people usually associated with Taoist and Con-
fucian ideas. These attempted to subordinate the ruler himself to the rule
of law and thereby curb the potential abuses of a tyrannical power.27 These
elaborations, although setting standards that the ruler was asked to observe,
and therefore making the ruler “accountable,” at the same time did not rec-
ognize any other authority that could legitimately promulgate laws. In other
words, whereas the ruler™s actions could be judged by the law, it was only
the ruler who could legally issue laws. By restructuring the state bureau-
cracy, reducing the power of the hereditary aristocracy, and eliminating
semi-independent power centers, the former Han emperors had been, by
and large, successful in securing not only the ideological but also the polit-
ical primacy of the imperial institution. The hierarchical system of the Han
bureaucracy also ensured that policies would be implemented through
a chain of subordination within which each link was responsible to the
higher one.
Finally, although the principle of imperial rulership was cast in univer-
sal terms, its translation from the philosophical to the political arena
implied something like a change of status. Both in practice and in princi-
ple, outside the state, which the emperor headed, the authority of the
emperor during the early Han was effectively limited by the norms and prac-

Benjamin Schwartz, The World of Thought in Ancient China (Cambridge, Mass.:
Belknap Press, 1985), p. 40.
Roger Ames, The Art of Rulership: A Study in Ancient Chinese Political Thought
(Honolulu: University of Hawaii Press, 1983), pp. 136“41.


tices that had to be observed in foreign relations; as we have seen, this was
the basis of the new world order created by Han Kao-tsu and Modun. It
can be argued that as the ho-ch™in treaty was signed, a claim to universal
rulership needed to be suppressed because it was unrealistic and contra-
dicted the policy of negotiated accommodation. The “boundedness” of
political sovereignty and its territorial limits continued to be endorsed even
after the Hsiung-nu had been thoroughly defeated, and there was no longer
a compelling realpolitik reason to accept such limitation. For example, in
60 b.c., during the debate over the ritual that the ch™an-yü Hu-han-yeh was
supposed to follow as he came to court to present “tribute,” Hsiao Wang-
chih remarked that because the ch™an-yü did not follow the Chinese calen-
dar, his nation should be referred to as an independent state, and he should
thus not be treated as a tributary, but should instead be assigned a rank
above that of the feudal kings.28
Turning to the Hsiung-nu camp, the situation appears very different. In
traditional Inner Asian societies, kin bonds, reliance on customary law, and
segmentation of political power among clans and tribes in steppe pastoral
societies prevented the emergence of any absolute, indivisible, and legally
recognized authority that would be the expression of a given political com-
munity in its entirety. The “state” among the Hsiung-nu was embodied in
certain features of governmental centralization that allowed for a uni¬ed
military leadership and for the existence of a center recognized both within
and outside of the political community, and, as mentioned earlier, it could
survive only through a process of uninterrupted, ongoing negotiation
between the ruler and the other tribal leaders. Thus, among the Hsiung-nu
the ¬gure of the ruler never truly represented a pole of absolute authority
in the sense evoked by the notion of sovereignty; in this system the leader
was more often the ¬rst among equals, whose position of primacy rested
ultimately on the consent obtained from other chieftains and members of
the tribal aristocracy. This consent could be coerced, but the ultimate foun-
dation of the charismatic leader was the voluntary consensus obtained from
his closest advisors, military commanders, and family members, without
whom his rise to power would be impossible. These “electors” could not
be kept in a position of absolute subordination.29

Burton Watson, Courtier and Commoner in Ancient China: Selections from the
History of the Former Han (New York, Columbia University Press, 1974), pp.
On the question of sovereignty and sacral kingship in Inner Asia, see Paul Roux,
“L™origine c©leste de la souverainet© dans les inscriptions pal©o-turques de Mon-
golia et de Siberie,” in The Sacral Kingship (Leiden: Brill, 1959), pp. 231“41;
Mori Masao, “The T™u-chüeh Concept of Sovereign,” Acta Asiatica 41 (1981):
47“75; Osman Turan, “The Ideal of World Dominion among the Medieval
Türks,” Studia Islamica 4 (1955): 77“90; Peter Golden, “Imperial Ideology and


Moreover, the highest points of nomadic political integration, military
might, and territorial expansion were accompanied by claims of universal
rulership that, by failing to identify the state with a limited political com-
munity, posed an equally unsurmountable obstacle to the emergence of the
notion of sovereignty.30 As we have seen, Modun rose to power through his
successful attempt to defend the Hsiung-nu people from the attacks of both
Chinese and Inner Asian foes. Once that mission had been accomplished
there was no speci¬c reason why the ch™an-yü should continue to enjoy
paramount political authority and “supra-tribal” power.31
However, together with the charismatic leader, a new stratum of aristo-
crats and military commanders had emerged, people who had been granted
a privileged position by the ch™an-yü and who had a bond of personal
loyalty with him. The authority they enjoyed and their enhanced military
power strengthened their position among their own followers and power
bases and allowed them to receive the lion™s share of the tribute paid by
China and other states. Members of the new aristocracy ¬lled the ranks of
the “supra-tribal” institutions; they became the emperor™s bodyguard, the
household administrators (ta-tang-hu), and the Ku-tu marquises, who acted
as inspectors or police agents for the ch™an-yü™s government.32 Thus, so
many interests had clustered around the newly authoritative institution of
the ch™an-yü that, as long as the internal and international situation
remained favorable, most continued to pro¬t from their initial conditional
support for the charismatic leader and would not withdraw it without a
compelling reason. The aristocracy™s political power “ previously absolute
within their own tribes “ had been curtailed, but the advantages of the new
political structure were by no means tri¬‚ing, and they could not be easily

the Sources of Political Unity Amongst the Pre-Cinggisid Nomads of Western
Eurasia,” Archivum Eurasiae Medii Aevi 2 (1982): 37“76.
Modun claimed for himself the title ch™an-yü, “established by Heaven” (Han shu
3756). This claim to universal rulership became particularly evident during the
Mongol period. A notable example of such a claim can be found in Hulagu™s letter
to King Louis IX of France. On this, see Paul Meyvaert, “An Unknown Letter of
Hulagu, Il-Khan of Persia, to King Louis IX of France,” Viator 11 (1980):
252“53; also E. Voegelin “The Mongol Orders of Submission to European
Powers, 1245“1255,” Byzantion 15 (1940“41): 378“413.
Nomadic societies can well continue to function in a situation of political seg-
mentation along tribal lines. The supra-tribal organization is not needed except
in special and fairly anomalous cases. On this, see Joseph Fletcher, Jr., “Turco-
Mongolian Monarchic Traditions in the Ottoman Empire,” Harvard Ukrainian
Studies 3“4, pt. I (1979“80): 236“51.
Rafe de Crespigny, Northern Frontier: The Policies and Strategy of the Later Han
Empire. (Canberra: Faculty of Asian Studies, Australian National University,
1984), pp. 177“78.


All of this does not mean that the ch™an-yü now enjoyed absolute author-
ity among his people. For all the organizational skills of its ¬rst leader,
Modun, the Hsiung-nu empire remained a highly tribalized state, with the
Hu-yen, Lan, and Hsü-pu clans holding the highest ranks. Government
posts and ranks in that structure were hereditary and remained within the
aristocratic families they were granted to. On the legal plane the ch™an-yü
certainly was not seen as the supreme “lawgiver,” and customary and tra-
ditional usages remained the basis of the law, to which even the ch™an-yü
had to conform. In the political community he remained the primus inter
pares, rather than the sovereign, that is, he was the agent of the interests
of the community of aristocrats by whom he was supported. That support
gave him the authority to issue orders, but the implementation of those
orders rested always on the consent of the other members of the elite. This
consent could be extracted by force or suasion but could not be compelled
through adherence to a ¬rmly established state ideology or controlled
through the machinery of a state bureaucracy. Indeed, there are many exam-
ples of internal divisions among the Hsiung-nu political community and of
the limits to the scope of the authority of the ch™an-yü.33 For instance, suc-
cession struggles were not uncommon, as in the case of the Lu-li king of
the left, Yi-chih-hsien, who at the death of the Chün-ch™en ch™an-yü in 126
b.c. proclaimed himself ruler and attacked and defeated Yü-shan, the heir
originally appointed by the late ch™an-yü.34 Disputes between the ruler and
his subordinated aristocrats could also easily lead to splits in the confeder-
acy. In 121 b.c. the ch™an-yü was angry with the Hun-yeh king and the
Hsiu-t™u king, who had under their jurisdictions the western territories of
the Hsiung-nu empire, for losing tens of thousands of men to the Han
armies, and he wanted to execute them, but this could not be accomplished
by a simple imperial order.35 Although power struggles are by no means
foreign to Chinese early history, one should emphasize that the bases of the
authority of the ch™an-yü, in terms of economic and military resources, as
well as legitimacy as a sovereign, rested primarily on the web of individual
liaisons and personal loyalties that he had been able to create before and
during his reign.
Did these two very different notions of political authority “ the Han insti-
tutional emperorship and the Hsiung-nu charismatic chieftainship “ play a
It is indicative that the Hsiung-nu empire appeared, from a Chinese perspective,
as beset by “regionalism.” See Yü Ying-shih, “Han Foreign Relations,” The Cam-
bridge History of China, vol. 1: The Ch™in and Han Empires, 221 B.C.“A.D. 220,
ed. Michael Loewe and Denis Twitchett (Cambridge: Cambridge University Press,
1986), p. 392.
Han shu 94A, 3767.
Both of them became fearful and plotted to surrender to the Han (Han shu 94A,
3769). In 104 b.c., the great commander of the left (tso ta tu-wei) of the Hsiung-
nu wanted to kill the ch™an-yü (Han shu 94A, 3775).

role in the arena of foreign relations and diplomatic agreements? Not only
did they play a role, but their fundamental asymmetry can be regarded as
the main reason for the failure of the ho-ch™in policy. The crux of the matter
is that when treaties were concluded, their observance could not exceed
the limits of the authority of each contracting party. As the Han and the
Hsiung-nu notions of central authority were asymmetric, so were their
respective abilities to observe the provisions of the treaties. The best evi-
dence of this “asymmetry” is provided by the protracted raids that contin-
ued despite the Chinese delivery of “appeasement” tribute to the ch™an-yü.
Though this evidence is not extensive, it is compelling because it explicitly
refers to the perpetrators of the raids as subordinate leaders and reveals
that the ch™an-yü was unable to force his own people to abide by the con-
ditions of the treaty that he had rati¬ed.
Immediately after the conclusion of a ho-ch™in treaty with Han Kao-tsu,
Chinese generals who had defected to the Hsiung-nu invaded and pillaged
the border areas, followed by Hsiung-nu generals. These two groups are
said to have violated the treaty often, invading and looting the regions of
T™ai, Yen-men, and Yün-chung.36 Clearly the authority of the ch™an-yü was
not suf¬cient to make them respect the agreement. When Wen-ti came to
the throne, in 180 b.c., the ho-ch™in treaty was renewed,37 but only four
years later, in 176 b.c., the wise king of the right violated the treaty on his
own initiative and invaded China. Because he had settled south of the
Yellow River and was encroaching upon Chinese territory, Wen-ti issued
the following edict, in which he explained the reasons for his military action
against the Hsiung-nu leader:
The Han and the Hsiung-nu made a brotherly pact. For not invading and pil-
laging the border region we granted many precious gifts to the Hsiung-nu.
Today the wise king of the right has left his state [kuo] and has led his people
to settle in the region south of the river. This is not in accord with the agree-
ments. He often crosses the border, capturing and killing of¬cers and soldiers.
He has taken away the Man and the Yi who protected the frontier of the
Shang Commandery, and ordered them not to reside in their customary
[land]. Oppressing border of¬cials, crossing the borders and pillaging is
very arrogant and does not accord with the norm. This is not the treaty [that
we made].38
The ch™an-yü™s reply to these accusations shows that he was quite frustrated
by the situation, though he laid part of the blame for the disturbances on
the Han as well. According to him the Han border of¬cials had “provoked”
the wise king of the right by invading his territory and insulting him. Unfor-
tunately, the wise king had thereafter failed to inform the ch™an-yü and had
instead preferred to follow the advice of other Hsiung-nu leaders, such as
Han shu 94A, 3754; Shih chi chu-yi 110, 2320.
37 38
Han shu 94A, 3756; Shih chi chu-yi 110, 2320. Han shu 94A, 3756.


the Hou-yi-lu “marquis” Nan-chih. In the ch™an-yü™s assessment, the source
of the problem was to be found in the hatred that existed between the wise
king of the right and the Han of¬cials, who “violated the treaty between
the two rulers, departing from fraternal relations.”39 The same type of sit-
uation arose in 162 b.c., when Emperor Wen-ti once again reminded the
ch™an-yü that wicked and evil people on his side were “severing the
harmony” between the two rulers.40
The troubles continued during Ching-ti™s reign. The ho-ch™in treaty con-
cluded between Han Ching-ti and the Chün-ch™en ch™an-yü (159“126 b.c.)
when the latter acceded to the Hsiung-nu throne was ineffecual in preserv-
ing peace. Like the previous ones, this treaty provided for the opening of
border markets and stipulated the bestowal of gifts as well as a princely
consort for the ch™an-yü, but it did not prevent the Hsiung-nu from attack-
ing and pillaging the border regions as they pleased. Here we do not know
who actually carried out the raids, but it seems likely that it was not the
ch™an-yü himself, because the ho-ch™in agreement did not break down, and
even the border markets continued to stay open.41
In another example of insubordination to the ch™an-yü by members of
the Hsiung-nu aristocracy, when the Hu-yen-t™i ch™an-yü (r. 85“68 b.c.)
came into of¬ce he told the Han envoys that he desired the continuation of
the ho-ch™in agreement. However, two of the most important leaders of the
Hsiung-nu politico-military establishment, the wise king of the left and the
Lu-li king of the right, opposed him, thus creating a split in the Hsiung-nu
polity, the results of which were told as follows:
Because [the wise king of the left and the Lu-li king of the right] did not
succeed to the throne, they nurtured resentment, and led their people away,
wishing to go south to surrender to the Han. But fearing they would be unable
to proceed on their own, they forced the Lu-t™u king [to join], and wanted to
go west to surrender to the Wu-sun, so they could plot [with them] an attack
on the Hsiung-nu. The Lu-t™u king reported this, and the ch™an-yü sent men
to investigate; the Lu-li king of the right did not confess, but reversed the
charge, accusing the Lu-t™u king. All the people of the country resented this.
Thereafter the two kings went to live in their territory, and did not attend the
assemblies at Lung-ch™eng.42
What is interesting in this episode is that these two disaffected leaders
¬rst attempted to “surrender to the Han,” then tried to ¬nd allies to ¬ght
against “the Hsiung-nu,” and eventually decided to withdraw to their
own territories. The “surrender” by Hsiung-nu leaders is nothing else
but an attempt to bypass the ch™an-yü™s authority and set up separate
ho-ch™in agreements with the Han. That the ho-ch™in ultimately became an
Han shu 94A, 3756; Shih chi chu-yi 110, 2321.
Shih chi chu-yi 110, 2324“5; Han shu 94A, 3762“3.
Han shu 94A, 3765. 42 Han shu 94A, 3782.


inducement used by the Han to break the unity of the Hsiung-nu is made
clear by the history of the relations between the Han and the leader of the
southern Hsiung-nu, Hu-han-yeh, who agreed to split from the confedera-
tion on the condition that he gained access to Chinese goods. This was pos-
sible because of the relative ease with which members of the Hsiung-nu
aristocracy could secede from the larger political union. Going back to one™s
own territory was an option open to the nobility that the ch™an-yü had no
legal power to oppose, and one that reveals the inherent weakness of his
Finally, one of the most explicit mentions of the ch™an-yü™s lack of
absolute sovereignty is contained in the speech by the Han statesman Hou
Ying on the occasion of a peace treaty negotiated with the Hsiung-nu leader
Hu-han-yeh in 51 b.c. Hou Ying, dismissing Hu-han-yeh™s proposal that
the Han demobilize their frontier forces and let him guard the border
instead, said: “although China possesses the teachings of propriety and
morality and has the death penalty, the masses still violate prohibitions.
How then could we expect the ch™an-yü to keep his followers from violat-
ing the treaty?”43 Obviously a perception existed among Chinese statesmen
that the sovereignty of the ch™an-yü was more limited than that of the
Chinese emperor.
Proceeding from the assumption that the volatility of the Hsiung-
nu political leadership constituted a serious, ¬nally insurmountable, obsta-
cle to the implementation of the ho-ch™in, it would be logical to surmise
that the Han appeasement strategy was meant to strengthen central author-
ity, not to undermine it. In fact, the diplomatic correspondence from 198
b.c. to 133 b.c. supports the hypothesis that the ho-ch™in was speci¬-
cally intended to support the Hsiung-nu ruler, by granting him not only
economic resources to be used to strengthen his internal position but also
the authority to control the whole Hsiung-nu people. There are numerous
instances in which the language of the documents speci¬cally excludes
the existence of sources of authority other than the emperor and the ch™an-
yü. The policy of assigning the title of ch™an-yü to several Hsiung-nu
tribal leaders, which eventually became the classic understanding of
“divide and rule,” was broadly enacted by Wang Mang and during the
Later Han but was never part of the Han strategy in the ¬rst part of the
Former Han.44

Han shu 94B, 3804.
On Wang Mang™s policies toward the Hsiung-nu, see De Crespigny, Northern
Frontier, pp. 194“218. The policy proposed by Ch™ao Ts™o was quite dif-
ferent and meant simply the recruitment of foreign people to ¬ght against other
foreign peoples: it referred to a military strategy rather than to a foreign affairs


The early Han emperors down to the reign of Han Wu-ti were hard-
pressed militarily, and needed peace to allow the nation to recover its
strength, but a durable peace could be achieved only by negotiating an
agreement with a single political ruler who was able to guarantee long-term
stability. Since, as we have already mentioned, the ¬nal goal of the ho-ch™in
strategy was to make the ruling clan of the Hsiung-nu into part of the Han
“family,” it would be logical to assume that the Han interests would best
be served by a policy of “unite and rule,” whereby the Hsiung-nu leader
would be made into an absolute ruler and thus could steer the whole
Hsiung-nu people toward a position friendly to China. Lacking that author-
ity, even though the Hsiung-nu ruler may have been brought under the
political in¬‚uence of China, other Hsiung-nu tribal leaders would surely
continue to ¬ght, and the whole ho-ch™in strategy would be useless. The
Han willingness to accede to the ch™an-yü™s demand that tribute be period-
ically increased may well have been a response to the hope that central
authority among the Hsiung-nu would be strengthened. Unfortunately no
amount of tribute could accomplish that, and the security of the frontier
was constantly threatened.
In conclusion, this examination of the Han relations with the Hsiung-nu
suggests that between 198 and 133 b.c. the northern territories of China
were constantly under pressure because there was no absolute authority
within the Hsiung-nu tribal confederation capable of guaranteeing the
respect of treaty obligations. Under these circumstances, China™s attempts
at pacifying the Hsiung-nu through pay-offs were destined to fail. Since the
power of the ch™an-yü was not only limited but also constantly exposed to
internal challenges, peace proved aleatory. No amount of Chinese support
could change the core structure of nomadic society, and because a basic
notion of sovereignty could not possibly develop without being preceded
by radical transformations in the nomadic state™s social and political
texture, the position of the ruler was destined to remain fragile. Only after
the realization of the ineffectiveness of the diplomatic approach, revealed
in the aforementioned position expressed by Wang Hui, was the ho-ch™in
policy discarded and replaced by military means.

War and Expansion

The political debates over frontier issues help us to clarify the rationale for
war but are less useful when we attempt to understand the goals that the
Han sought to achieve. In fact, there is a clear gap between the stated objec-
tives and the results that the more aggressive policy carried out by Wu-ti
eventually yielded. The Han dynasty expanded territorially far beyond
the boundaries held by the Chou states, and by the mid-¬rst century b.c.


the northern hegemony of the Hsiung-nu empire was broken. The conquest
of the north increased the Han administrative network in the northern
border areas and strengthened and rationalized the border defense system
with military garrisons, ¬xed forti¬cations, and settlements. The Han
western expansion as far as the Tarim Basin led to the submission of a
plethora of small states and to the establishment of commercial ties with
Central Asia.
These feats were most successfully carried out between 121 and 112 b.c.,
and gradually consolidated during and after the reign of Emperor Wu. Yet
the motives behind Han Wu-ti™s decision to embark on an unprecedented
program of territorial expansion are not fully clear, nor can they be clari-
¬ed without our ¬rst considering the actual military capabilities of the Han,
the extent to which the Hsiung-nu empire had grown, and the nature of
Hsiung-nu economic and political power. A long view is required to deter-
mine, for instance, whether the Han expansion was planned or was, on the
contrary, the end-product of a piecemeal process. One of the most puzzling
features of Wu-ti™s wars against the Hsiung-nu is that none of the stated
objectives of the Han offensive seems to justify either the duration of the
war or the extent to which the Han armies advanced into Hsiung-nu terri-
tory. Even the most hawkish positions expressed during the debates that
accompanied the shift from a paci¬st to a militarist posture did not advo-
cate anything like the results eventually achieved. There was no military or
political strategy elaborated in this period that even remotely justi¬ed the
decades of military offensives, political expansion, and territorial acquisi-
tion that were in fact to follow.
Arguments in favor of the expansion were expressed post facto in
the Discourses on Salt and Iron by the “realist” faction of the government
of¬cials. These arguments can be summarized in two points: border defense
and trade opportunities. According to the “ministerial” faction, which
supported military intervention, Wu-ti had been driven to expansion by
the need to protect the central kingdom. The conquest of Hsiung-nu terri-
tory and the establishment of commanderies in the north and west were
intended to guarantee peace in the rest of the kingdom. It was, in other
words, an instance of “defensive acquisition” rather than true expansion.
Moreover, the bene¬ts to be derived from trade were enormous, because by
expanding into these lands China could trade goods of which it had a
surplus (and which were therefore next to worthless on the internal market),
for things of value, such as gold, pack animals, furs, and other precious
objects. To this, the opposite faction “ the “literati” “ replied polemically,
with the usual critique, that in pursuing an expansionist policy, the emperor
intended to increase his power and aggrandize the state at the expense
of the welfare of the people, who became bankrupt in the process. As
for the hypothetical commercial bene¬ts, those imported goods were
not nearly worth the Chinese products exchanged for them; once all

expenses “ including transportation and all phases of production “ had been
calculated, the cost of the Chinese goods was far higher than that of the
imported items.45
However, neither of these relatively limited objectives could justify sub-
duing independent polities (even those that were not hostile) and bringing
distant lands, far beyond the Han frontiers, under Chinese rule. To com-
prehend the reasons that led the Han to displace the power of the Hsiung-
nu throughout the north, and to establish their own supremacy in Central
Asia, we must examine two aspects of the context in which this strategy
emerged. The ¬rst is purely military: it involves the development by the Han
of the ability to launch long-ranging military expeditions, and the tactical
solution adopted by Han Wu-ti of consolidating territorial gains by turning
the conquered regions into administrative areas, which were then incorpo-
rated within the empire. These areas supported a line of forti¬cations that
provided logistical support to Chinese armies pushing farther and farther
west, and therefore constituted the vertebral column of the Han expansion
into the Western Regions.
The second aspect is strategic and refers to the discovery made by the
Han that the war against the Hsiung-nu could not have been won without
severing the essential economic and political links between the Hsiung-nu
and other polities, in particular the kingdoms of the Tarim Basin and the
Ch™iang. This strategy, which was referred to as “cutting the right arm” of
the Hsiung-nu, required the direct Han intervention in the Western Regions;
it was the dogged pursuance of this strategy that eventually provoked a
fatal split in the Hsiung-nu leadership and marked the end of the Hsiung-
nu empire.

Han Wu-ti™s Offensive: Political and Technical Aspects

Even though the Han did realize that the appeasement policy was never
going to guarantee peaceful relations, a switch to the offensive required
certain objective military capabilities, without which the constraints on
any major operations would have been so great, and the costs so high, that
it is unlikely the policy shift would have ever produced positive results. The
realization that the ho-ch™in policy had failed provided the rationale
and the determination to pursue an offensive strategy, but such a realiza-
tion by itself could not have shifted the military balance in favor of the
Han overnight, as the Ma-yi disaster plainly demonstrated. What enabled
China to challenge the Hsiung-nu militarily was a combination of factors

E. M. Gale, Discourses on Salt and Iron. A Debate on State Control of Com-
merce and Industry in Ancient China (Leiden: Brill, 1931 [rpt. 1967]), pp. 14,

that had been accruing slowly and that resulted in an overall increase in
the Chinese offensive capabilities. Two handicaps had prevented the early
Han emperors from engaging the Hsiung-nu successfully: the deep political
divisions within the Han camp, and the sheer battle¬eld superiority of
the Hsiung-nu. Both of these factors had changed considerably by the time
of Wu-ti.
A primary preoccupation of Wu-ti™s predecessors had been the con-
solidation of central authority.46 Han Kao-tsu had been unable to concen-
trate his efforts against the Hsiung-nu, as he was forced to devote his
attention primarily to the consolidation of the central power against the
centrifugal tendencies expressed by the subordinate “kings” (chu-hou-
wang).47 The same policy of centralization was pursued by later emperors,
engaged, sometimes strenuously, in curbing the rebellions against the Han
government by the chu-hou-wang, such as the major challenge in 156 b.c.,
and the defections to the Hsiung-nu by those political and military leaders
who did not want to yield personal power and prestige to the central
The Hsiung-nu were central actors in this “internal” Han political strug-
gle, because the harshness of the Chinese military code, admiration for
Hsiung-nu martial prowess, and, above all, the con¬‚ict of interests between
central power and peripheral potentates often inspired changes of alle-
giance. The Hsiung-nu status as an independent power enabled them to act
as a subversive political and military presence and greatly increased the
threat they posed both to the integrity of China and to the survival of the
Han dynasty. The stories of generals such as Lu Wan, Hann Wang Hsin,
Ch™en Hsi, and others who, after the establishment of the Han dynasty,
decided to rebel against Kao-tsu and join the Hsiung-nu, are unambiguous
in this respect. After having rewarded his allies with ¬efs, Han Kao-tsu had
to strengthen the authority of the central government and, unsurprisingly,
came into con¬‚ict with his former supporters™ yearnings for autonomy. For
instance, Ssu-ma Ch™ien reports that when the king of Yen, Lu Wan, joined
Kao Tsu to wipe out the “rebellious” feudatory Ch™en Hsi, who had enlisted
Hsiung-nu support against the emperor, he was advised by a Chinese defec-
tor to the Hsiung-nu to put an end to his war against Ch™en Hsi and to
make peace with the Hsiung-nu, for, as soon as Ch™en Hsi had been crushed,
Kao Tsu would be turning against Lu Wan himself. In his concluding

Michael Loewe, Crisis and Con¬‚ict in Han China, 104 BC to AD 9 (London:
George Allen and Unwin, 1974), p. 59.
On this question, cf. M. Loewe, “The Former Han Dynasty,” in The Cambridge
History of China, pp. 139“44; W. Eberhard, A History of China (Berkeley: Uni-
versity of California Press, 1977 [1960]), pp. 77“86; Ch™u T™ung-shu, Han Social
Structure (Seattle: University of Washington Press, 1972), pp. 76, 165“66.


remarks to chapter 93, Ssu-ma Ch™ien states that “in the empire [these lords]
aroused suspicion by their great strength, while beyond its borders they
sought aid from the foreigners (man-mo), so that with each day they became
further alienated from the emperor and moved deeper into danger. At last,
when their position became impossible and their wisdom failed, they went
over to the Hsiung-nu.”48
Besides offering a sanctuary and a political alternative to those who
came into con¬‚ict with the central authority of the emperor, the Hsiung-nu
were also militarily strong. The Han had suffered numerous defeats, and
their border defenses had been repeatedly broken, leaving no doubt that
any military encounter with the Hsiung-nu would be costly and the outcome
unpredictable. In fact, that argument provided the rationale for the long-
term endorsement of the ho-ch™in policy, so that any change of policy
needed to be preceded by a program for the strengthening of the Han army.
By the time of Han Wu-ti, conditions regarding military technology and
offensive capabilities had improved to such an extent that the Han army
had become far more competitive than it had been at the time of Han
The program that made this possible focused on two parallel projects:
creating an effective cavalry force and improving the effectiveness of the
weapons. As noted in Chapter 4, the concept of a regular cavalry was
imported into China from the northern peoples, whose armament included
possibly long swords, spears, compound bows, and body armor.49 However,
the bulk of the Chinese armies in the major campaigns against the Hsiung-
nu carried out by Meng T™ien and by Han Kao-tsu was still made of infantry
and charioteers.50 An explicit request to the emperor that cavalry forces
be used against the Hsiung-nu can be traced to the aforementioned memo-
rial by Ch™ao Ts™o in 169 b.c.51 Such a request suggests that, even if the
building of a cavalry was underway, it had not yet been suf¬ciently

Shih chi 93, 2649.
Burchard Brentjes, Arms of the Sakas (Varanasi: Rishi Publications, 1996); M.
Rostovtzeff, Iranians and Greeks in Southern Russia (Oxford: Clarendon, 1922),
pp. 203“204; Jaroslav Prusek, Chinese Statelets and the Northern Barbarians in

the Period 1400“300 B.C. (Dordrecht: Reidel, 1971), pp. 116“17.
Chang Chun-shu, “Military Aspects of Han Wu-ti™s Northern and Northwestern
Campaigns,” Harvard Journal of Asiatic Studies 21 (1966): 167.
Han shu 2281; cf. also Loewe, Crisis and Con¬‚ict in Han China, p. 100; Leon
Wieger, Textes historiques: histoire politique de la Chine depuis l™origine, jusqu™en
1912, 2 vols. (Hsien-hsien: Impr. de Hien-hien, 1922“23), 1: 343“44; Berthold
Laufer, Chinese Clay Figures, Part I, Prolegomena on the History of Defensive
Armor (Chicago: Field Museum of Natural History Publication no. 177, 1914),
p. 229.


China™s success depended on its ability to gain access to a suf¬cient
number of mounts to sustain a military cavalry on a continuing basis. Tra-
ditionally, China had imported horses from the region of Tai and from the
nomads, and the border markets opened between Han and nomads con-
tinued to be used by China as a gateway for the importation of horses from
the reign of Kao-tsu to that of Wu-ti. However, reliance on external sources
would make the military dependent on supplies that might become unavail-
able when they were most needed, namely in wartime. Therefore, facing for
the ¬rst time the problem of raising large numbers of military mounts, the
early Han emperors tried to solve it autarchically and began to breed horses
within China. Attempts to that effect were carried out in areas close to the
northern frontier, in regions possibly not yet placed under cultivation and
inhabited by pastoral peoples.52 The caretakers of these horse-breeding sta-
tions were not only foreigners but also possibly surrendered or captured
Hsiung-nu tribesmen.53 By Wu-ti™s time, mass breeding of horses was well
under way: in 140 b.c. thirty-six breeding stations were already in opera-
tion on the frontier, providing maintenance for three hundred thousand
After some major campaigns were launched deep into nomadic territory
between 129 and 119 b.c., a shortage of horses began to be felt. The state,
therefore, tried to transfer the burden of providing military mounts from
the frontier areas to the whole territory of the empire by promoting the
private breeding of horses to be sold to the government.55 In addition to
the horses that the state purchased from private individuals, a new tax law
stipulated that up to three men in any family could be exempted from mil-
itary duties by presenting one horse each for the army. Finally, a portion of
the revenues from the poll tax on children (k™ou-fu) was earmarked specif-
ically for the purchase of horses for the military.56 Already in 146 b.c. the
export of horses under ten years of age was forbidden, and horses remained
a much sought-after commodity.57

Later dynasties (in particular the T™ang and the Ming) relied far more on
horse imports that the early Han. On the T™ang-Uighur horse trade and its
importance in Chinese economy, see Christopher Beckwith, “The Impact of
Horse and Silk Trade on the Economies of T™ang China and of the Uighur
Empire,” Journal of Economic and Social History of the Orient 34.2 (1991):
C. Martin Wilbur, Slavery in China under the Former Han Dynasty 206 B.C.“A.D.
25 (Chicago, Field Museum of Natural History, 1943), pp. 109“15.
Nancy Lee Swann, Food and Money in Ancient China. The Earliest Economic
History of China to A.D. 25. Han Shu 24 with Related Texts, Han Shu 91 and
Shih Chi 129 (Princeton: Princeton University Press, 1950), p. 38.
Swann, Food and Money in Ancient China, pp. 308“309, 302“304.
Swann, Food and Money in Ancient China, p. 374.
Yü, Trade and Expansion in Han China, pp. 119“20.


. 9
( 15)