Military Culture in Imperial China
EDITED BY
Nicola Di Cosmo
HARVARD UNIVERSITY PRESS
Cambridge, Massachusetts, and London, England 2009
6
Narrative Maneuvers
The Representation of Battle in Tang Historical Writing
DAVID A. GRAFF
Military events such as battles and sieges figure prominently in the historical record of medieval China. The state-sponsored dynastic histories of this period mention hundreds of such episodes. From the point of view of the modern military historian, however, the traditional Chinese treatment of armed conflict leaves much to be desired. One eminent scholar of medieval Chinese warfare has complained, "Even a major military event is mostly mentioned in the official sources in a few words only. 'The army of X was defeated near Y,' 'the city of Z was taken (or successfully defended)'—such are the usual entries."1
This is too harsh an assessment. The dynastic histories allow us to reconstruct a great many battles and campaigns in outline, and even relatively brief accounts of battles occasionally offer valuable snippets of information about weapons, tactics, preferred stratagems and battle plans, the psychology of combat, and the nature of leadership. Turning from the pragmatics of warfare to the realm of culture, the battle and campaign narratives found in medieval Chinese histories also tell us a great deal about the political and intellectual elite's attitudes toward military operations and military men. This chapter thus addresses the fourth definition of military culture as an aesthetic or literary tradition. It examines the way in which battle is represented in a single dynastic history, the Old Tang History (Jiu Tangshu). How are the battle accounts in this work constructed? Which elements are emphasized, and what is
Narrative Maneuvers
downplayed? In what ways might these accounts have been shaped by the earlier documentary sources on which they were based, and what does all of this reveal about the thinking of the mosdy anonymous scholar-officials who created them?
The Old Tang History, covering the period from 618 to 907, was completed in 945 under the short-lived Later Jin dynasty. The nominal author, Liu Xu (887-946), was a senior statesman who had supervised the activities of a team of scholar-officials. With the exception of the last few Tang reigns, Liu's work contained very little original writing that postdated the Tang. The compilers closely followed their principal Tang sources, the National History (Guoshi) of Liu Fang (fl. ca. 755), for the period up to and including the An Lushan Rebellion of 755-763, and the Veritable Records (sbilu), for subsequent reigns; their work generally involved "condensation, summarization, and elimination of surplus verbiage and unwanted material" rather than "active composition."2
The Tang sources on which the Old Tang History was based were themselves the work of government officials, mosdy men assigned to the Historiography Office (shi guan), an agency of the Tang government with its offices in a building on the grounds of the imperial palace in Chang'an. Court diarists recorded the words and deeds of the monarch, as well as other momentous events as they were reported at court. On a regular basis, the diaries and many other types of official (and unofficial) documents were deposited at the Historiography Office, where they provided the basis for the Veritable Records, annalistic accounts of the events of each imperial reign that were—ideally— supposed to be compiled shortly after the passing of the emperor with whom they dealt. The Veritable Records in turn provided the basis for several National Histories, infrequent longer-term compilations covering the entire history of the ruling dynasty from its inception to the time of writing.
Those Tang officials who were assigned to the writing and compilation of history were an elite within the imperial civil service. Most had established their credentials by passing the difficult and prestigious jinshi examination, which tested skill at literary composition as well as rote knowledge of the Confucian canon.5 Like the remonstrance officials (responsible for pointing out the misdeeds of the emperor himself) and certain other categories of imperial advisers and document
Narrative Maneuvers
145
writers, the historians were considered to hold "pure" posts that put them on the fast track to the highest offices in the capital—in contrast to the majority of officials, who tended to languish in minor posts and provincial backwaters. Many official historians later rose to become chief ministers, members of the collective premiership that marked the pinnacle of power in Tang officialdom. Rarely, if ever, did such men hold military offices, exercise military command, or lay claim to military experience of any sort.4 As we shall see, this brilliantly literate elite took a peculiar view of battle. Their representations of military action devote little attention to weapons, tactics, and the clash of arms. Heroic deeds are treated with an inconsistency suggestive of profound ambivalence, while the greatest emphasis is placed on the efficacy of cunning plans and elaborate stratagems.
Battles in the Dynastic History
Two rather different types of battle narratives are found in the Old Tang History, reflecting the basic division between the annals and the biographical chapters in Chinese dynastic histories. The battle accounts that appear in the annals of the Tang emperors (the first twenty chapters of the Old Tang History) are mostly short entries. The treatment in the Su-zong Annals of the battle of Xiangji Temple, which led to the recovery of Chang'an from the rebels in 757, is typical: "Ninth month, day renyin. Did battle with the rebel generals An Shouzhong, Li Guiren and others northwest of Xiangji Temple. The rebel army suffered a major defeat, with sixty thousand men slain. The rebel leader Zhang Tongru quit the capital and fled eastward."5
More substantial battle narratives are located in the biographical chapters of the Old Tang History. If the subject of the biography held an important leadership position during the combat in question, the account may be of considerable length. Two examples are the account of the battle of Xiangji Temple in the biography of Li Siye (d. 759) and the account of the battle of the Huan River in the biography of Ma Sui (726-795). Both passages are far more extensive and detailed than the minimal notices of the same events in the annals of the Old Tang History.6 Many other biographies contain much shorter passages that do not attempt to depict an engagement in its entirety but simply tell of a single episode that occurred during the course of the battle. The
146
Narrative Maneuvers
biography of Qiu Xinggong (586-665), for example, tells the story of his rescue of the future emperor Li Shimin (599-649) from behind the enemy line during a battle outside Luoyang in the spring of 621.7 In other cases, what is presented is a sketchy overview of a campaign lasting weeks or months, with the decisive engagement receiving only the briefest of mentions—or no mention at all. An account of the epic 635 campaign against the nomadic Tuyuhun people of the Kokonor region (in the biography of Qibi Heli [d. 677], a Turk general in the Tang service) closes with the following: "He thereupon selected more than a thousand valiant cavalrymen, went directly into the valley of the Tulun River, and surprised and defeated the Tuyuhun headquarters camp, killing several thousand men and taking more than 200,000 camels, horses, cattle, and sheep."8
Numbers are a common feature in battle narratives of all lengths. Sometimes the numerical strength of one or both armies is given, and often there is an indication of the magnitude of the casualties suffered by the defeated army—the number of men killed and perhaps also the number captured and (as above) a figure for the livestock taken from the enemy. With very few exceptions these numbers are large, round, and suggestive of uncertainty: "ten thousand" or "several thousand" or "fifty thousand" or "several tens of thousands." The very largest numbers tend to be reserved for rebel or "barbarian" forces (especially Tibetans) opposing the imperial armies, while the figures given for the government side tend to be more reasonable. The losses suffered by the government armies are rarely mentioned.9
Weapons and tactics are occasionally spoken of in the battle narratives but are by no means a regular and consistent feature of these passages. The weapons mentioned most often are the cavalry lance {qiang or shuo), the crossbow (nu), and the long-hafted saber (changdao) or long-hafted sword (modao).U) These were far from the most common arms in use at the time and were usually in the hands of relatively small, elite units of specialists.11 Battle formations and tactics are rarely mentioned—and when they are, it seems that they attracted the attention of the historian only because there was something unusual about them. A case in point is the description of a battle fought in 641 against a steppe people, the Xueyantuo, in which both the Tang cavalrymen and their nomadic opponents elected to fight on foot.12 This sort of encounter was far from the norm in steppe warfare!
Narrative Maneuvers
147
In contrast to the uneven and haphazard treatment of tactical matters, the battle plans, strategies, and stratagems of army commanders receive considerable attention in the Old Tang History. The outcome of a battle is rarely portrayed as the result of blind chance or uncontrollable circumstances; rather, it is due to the superior foresight of the victor or the stupidity of the vanquished. Perhaps the finest example of the leader who has anticipated everything and whose calculations are flawless is Li Su (773-821), the conqueror of the rebellious province of Huaixi in 817. To the consternation of his conventionally minded subordinates, Li took advantage of the cover provided by a snowstorm to plunge deep behind enemy lines and take Huaixi's capital, Caizhou, completely by surprise.13
At the other extreme are generals who are defeated because they make foolish plans and reject good advice. In the spring of 621, for example, the Hebei warlord Dou Jiande (573-621), who was advancing to break the Tang siege of Luoyang, ignored a subordinate's recommendation that he bypass a strong Tang defensive position on the main route to the beleaguered city. Instead, Dou was lured into making a frontal assault on the Tang army at the Hulao Pass, suffered a devastating defeat, and was taken prisoner on the field of battle.14 More nuanced examples tell of wise generals who were defeated because political pressures compelled them to adopt inferior plans against their better judgment. The general himself may be blameless, but the basic pattern still holds, with knaves or fools being held responsible for the defeat.15
In battle after battle, the superior commander's weapon of choice is the stratagem—and his opponent almost invariably takes the bait and suffers the predictable defeat. By far the most common gambit is the feigned flight, which seldom fails to achieve the desired result by luring the enemy into an ambush or setting him up for a devastating counterattack. More elaborate schemes are also used to good effect. The following example is from the early years of the Tang dynasty:
The Tuyuhun and the Dangxiang came together to plunder the border, and Chai Shao was ordered to chastise them. The caitiffs occupied the heights and looked down from above, shooting into the midst of Shao's army; arrows fell like rain. Shao thereupon sent men to strum the barbarian guitar \pipa], while two women danced facing one another. The caitiffs thought it strange and, ceasing their
148
Narrative Maneuvers
archery, gathered to watch. Shao saw that the caitiffs' formation was not orderly, and surreptitiously sent elite cavalry to attack them from the rear. The caitiffs collapsed completely, and more than five hundred were slain.16
The Old Tang History account of a single battle, Wang Shichong's (d. 621) defeat of Li Mi (582-619) at Mangshan in the autumn of 618, offers not one but several stratagems. Before the battle, Wang produces an auspicious omen to raise the spirits of his troops. He then deploys his men in a hopeless position with a river at their back to galvanize them to fight desperately, sends 200 cavalry around behind Li Mi's position to raid his camp at the height of the battle, and finally launches a surprise frontal attack.17
Stratagems of this sort were, of course, hardly a new development of the Tang period. Many examples can be found as early as the Spring and Autumn period (722-481 bce), and during the Three Kingdoms era during the third century ce, many such schemes were attributed to the great strategist of the Shu-Han kingdom, Zhuge Liang (181-234). The continuing prominence of stratagems in Chinese battle narratives points to a broader concern with psychological factors that pervades these materials. Much attention is devoted to efforts to raise the spirits of one's own troops or undermine the morale of the enemy, and the likely impact on morale can be a major consideration for a general contemplating a particular course of action.18 Again and again, the defeat of an army is presented as the result of a blow to its morale, such as the death, flight, or capture of its commander. As Peter Boodberg observed in 1930, the crux of the traditional Chinese battle narrative is psychological: battles are won and lost because something happens that causes the men of one side or the other to lose their nerve and flee the battlefield.19
Juxtaposed against the herd behavior of the masses in these battle narratives are descriptions of the heroic deeds of individual warriors. Sometimes these men are army commanders; more often they are subordinate leaders or lesser officers. They are involved in single combat against enemy warriors; they kill and capture men with their own hands, lead charges, and suffer horrendous wounds. We are told that a young warrior in the civil wars that attended the collapse of the Sui dynasty once charged the enemy line, killed several men, cut off the head of one,
Narrative Maneuvers
149
tossed it up in the air, and caught it on the point of his lance.20 At the battle of Xiangji Temple in 757, Li Siye appears to hold off the entire rebel army single-handedly by wielding a long-handled sword.21 Often it is a comrade in peril that gives a warrior the opportunity to demonstrate his valor and prowess.
Heroic deeds such as these are often encountered in the Old Tang History, but they are not evenly distributed. In the biographies of the founding generation of Tang military leaders, men active between the time of the Sui collapse and the early years of Gaozong's reign, at least eighteen instances of heroic action in battle are explicitly stated (not counting the many stupendous feats of the youthful Li Shimin recorded in the first chapter of the Taizong Annals).22 All of these episodes occurred in the roughly three decades between 613 and 645. Brave and bloody deeds then become quite rare in the second half of the seventh century.23 The action does not begin to pick up again until the Tianbao period (742-756). There are at least twelve more instances spread over the three generations from 742 to 820, with five of them concentrated in the years of the An Lushan rebellion (755—763).24 These later descriptions of violent combat lack the sort of over-the-top extravagance presented by several of the early Tang accounts; there is, for example, nothing comparable to the anti-Sui rebel leader who plunged into an enemy formation to seize the man who had put an arrow into his forehead.25 Accounts of heroic action in the Old Tang History trail off rapidly in the early ninth century, and there are no further examples in the biographies after the end of Xianzong's reign in 820.26
Due to the uneven coverage of different time periods in the Old Tang History, any attempt at statistical analysis of trends or patterns over time based on evidence from its pages is an inherently risky enterprise. With this caveat in mind, the impression of a decline in the representation of heroic action after the founding generation of the Tang would actually seem to be reinforced by the fact that the book's annals have a greater density of coverage—figured in pages per year— for the period from 756 to 847 than for the reigns of Gaozu and Taizong (618 to 649).27
The Old Tang History cannot be considered a primary source for the study of Tang warfare. Scattered Tang documents have, however, survived in encyclopedias and literary anthologies, preserved in many cases for their literary merit rather than for their historical value. A few of
150
Narrative Maneuvers
these documents are reports of military operations that should have been sources available to the authors of the dynastic history. How is battle represented in these materials, and to what extent do they appear to have influenced the representation of battle in the histories?
Representation of Battle in the Announcements of Victory
The Complete Tang Prose (Quart Tang wen) collection, compiled in 1814, contains ten examples of Tang "announcements of victory" (lubu). The ten lubu cover a period of more than two centuries, with the earliest probably dating from 672 and the latest from 883.28 In addition, a late Tang military encyclopedia, the Taibai yinjing, includes a blank-form lubu in a chapter providing models of various sorts of documents for the aspiring military secretary.29 On the basis of these specimens, it is possible to generalize with some confidence about the representation of battle in the announcements of victory and compare this to what we have seen in the Old Tang History. (See Appendixes A-C at the end of this chapter.)
The literal meaning of lubu is "exposed announcement," and the term had been used at least from Han times (202 bce-220 ce) to refer to a letter or communication sent without a seal and therefore open to inspection. It was not until the Northern Wei dynasty (386-534) that it acquired the narrower meaning of an official communication reporting a military victory.30 The Wei practice was to write victory announcements on sheets of silk and stick them up on poles for all to see.31 As one Northern Wei prince put it, the lubu was meant to be "announced within the four seas, and exposed to the ears and eyes."32 A ceremony for the reading out of lubu at court was instituted in 589 under the Sui dynasty. The capital officials and foreign envoys were assembled outside the Guangyang Gate, each dressed in his court robes and positioned according to his rank. When the director of the Secretariat (neishi ling) finished reading the announcement of victory, the officials responded by kneeling, dancing, and kneeling again.33
Under the Tang, announcements of victory were submitted to the court by field commanders following the decisive engagement of a successful military campaign. In the seventh century, the leaders submitting lubu were the commanders of expeditionary armies (xingjun zongguan); from the early eighth century on, they were generally military governors
Narrative Maneuvers
151
(jiedushi).H Although the lubu went out in the name of the commander, the actual work of drafting the document was normally done by a secretary. The finished announcement of victory was carried to the Ministry of War in the capital by one or more messengers who were usually high-ranking officers of the victorious army.35 The vice-minister of war (bingbu shilang) forwarded it to the Secretariat in the form of a memorial.36 The Secretariat then reported the lubu to the throne. After the Secretariat received the emperor's rescript, arrangements were made for a court ceremony that followed the basic pattern established by the Sui dynasty: the director of the Secretariat read the announcement of victory to the assembled officials and foreign envoys, who responded by dancing and kneeling. At the conclusion of the ceremony, the minister of war came forward to receive the document from the hands of the director.37 The Ministry of War was then responsible for recording the lubu and reporting it to the Historiography Office.38 Generals in the field were ap-parendy well aware that their reports might be used as sources by the official historians; the model announcement of victory in the Taibai yinjing closes with the wish that it might be "used to grace the books of history."39
Judging from the ten extant specimens, the Tang announcement of victory was a long and elaborate document. It usually began with a recitation of the history of relations between the Tang court and its opponent of the moment, couched in ornate, rhetorical language. For example, one of the lubu written in the early 670s by Luo Binwang (d. 684) observes that the emperor "purifies the central states by expounding the cultural teaching; he controls the nine barbarians by exalting military merit." As for the Man opponents of this Tang expedition to the far southwest, "jackals and wolves have their nature, and owls are difficult to tame. Consequently, they dared to disorder our heavenly constancy and lead the nine races to go back on their word."40 The lubu then notes the dispatch of an expeditionary army in response to the most recent insult, outrage, or provocation. There might be a description of the terrain through which the army has passed on its way to engage the enemy, but once again the language is poetic and impressionistic ("there lofty precipices obscure the sun, and crows have no room to turn on the wing"); the text does not provide the sort of practical geographical information that might allow a reader to plot the campaign on a map.41
The account shifts in the direction of less opaque phrasing as the army moves toward contact with the enemy. There is usually mention of the division of the army into several columns, marching by different routes and with different roles in the operational plan. Above all, names are mentioned. We may have the names of the commanders of the several columns and those of many of their subordinates, along with their offices and titles. (The most extreme example supplies the names of no less than seventy-nine officers.)42 At this point there might be a council of war, with different options debated.43 Then, as the Tang army maneuvers into position and the decisive encounter begins, a literary curtain descends over the battlefield. In the earlier of Luo Binwang's two lubu, we are told:
The bandit leaders Yang Qianqing, Nuo-mo-nong, Nuo-lan-si and others exerted their mantis strength to block the chariot wheels, let loose their mosquito swarm to pack the hills and fill the valleys. [The Tang officers] Liu Huiji, Gao Nufu, Sun Ren'gan and others were all famous for their loyalty and diligence, known far and wide for their wisdom and strategy. Having known the weighty favor of an enlightened ruler, they risked their lives in a certain place; picking up the divine sword Tai A, they faced death every minute. The evil party collapsed when they drew their bows; the demonic followers unraveled when they brandished their blades.44
The descriptions of battle in this and all the other lubu are couched in the most flowery, elaborate terms. "Weapons crossed and blades touched, birds scattered and fish were startled."45 Most of the announcements of victory shed no light on what might actually have happened on the battlefield or why one of the two armies eventually collapsed.46
When the literary pyrotechnics come to an end, the enemy is in full flight, the Tang army in hot pursuit. The announcement of victory then proceeds to an enumeration of the spoils, providing a list that might include the names of enemy leaders killed or captured, the number of enemy soldiers killed, the number taken alive, the number of livestock seized, and even an estimate of the total number of weapons, pieces of armor, and other items of equipment gathered from the field of battle.
Narrative Maneuvers
153
One lubu, for example, records the killing of 3,000 Tibetans, the capture of 1,000, and the seizure of 80,000 cattle, horses, sheep, and camels.47 But this was only a preliminary accounting. All ten lubu close by saying that a separate, more complete register enumerating the spoils of victory will soon be en route to the capital, and the Taibai yinjing provides us with a blank-form example of just such a document.48 In contrast to the detailed treatment accorded losses inflicted on the enemy, the lubu (and the supplementary register) show virtually no interest in the casualties suffered by the Tang army itself. Only one of the ten announcements of victory has anything to say about Tang losses, and here the number given seems implausibly small.49
In addition to the sorts of numbers already mentioned, the Tang announcements of victory provide several other categories of specific and apparently factual information. There is usually some indication of the date on which a battle occurred (including the month but not the year), as well as the duration of the battle.50 And near the end of the lubu, there is almost always a figure given for the total number of engagements (zhen—literally "formations" or deployments into battle formation) that occurred during the course of the campaign. There might be thirty or more zhen in the course of a single campaign.51 The scope of the lubu is the victorious campaign, not the individual battle, and some of the longer and more complex examples include ornate accounts of not one but two battles. Especially noteworthy in this regard is the report of an attack on the Tibetans in the 740s, which begins with a successful offensive and initial victory, followed by a change of fortune, a fighting retreat, and a final Tang triumph.52
This campaign orientation, with emphasis on the preliminaries of the battle rather than the combat itself, is one of several similarities between the announcements of victory and the battle narratives in the Old Tang History. The lubu also resemble the dynastic history in their treatment of numbers. Those given for the government force appear more sober and reliable than those given for the enemy, and in the event of a government victory—and all of the lubu are of course reports of victory—the victors' losses are seldom mentioned. A third similarity is that the announcements of victory show little interest in weapons and tactics. Only very rarely is a specific type of weapon mentioned, and then it would seem to be more rhetorical nourish than matter-of-fact description.53
154
Narrative Maneuvers
The differences between the Old Tang History and the lubu are far more numerous. First and most obviously, the language in the history is less ornate, and the elaborate recitation of the historical background and ideological justification for the campaign is entirely lacking. The history devotes much less attention than the lubu to identifying the various columns and contingents of the imperial field army and to naming their commanders. A few names may be given, but never dozens (as is usual in the announcements of victory). In sharp contrast to the dynastic history, the lubu pay almost no attention to the psychology of battle.54 Nor, as noted earlier, do they usually offer rational explanations for the outcome of battles, as the official historian is wont to do. In the few instances where the lubu do mention stratagems, these are no more complex than the feigned flight. There is no mention whatever of the more elaborate (and often implausible) sorts of stratagems found in so many of the battle narratives in the Old Tang History.
In the very few cases where it is possible to compare the lubu description of a battle with an account of the same engagement in the Old Tang History, the former offers much less in the way of substantive information that might enable us to reconstruct the sequence of events on the battlefield. Li Sheng's (727-793) report of the recovery of Chang'an from rebel forces in 784 indicates that the government troops pressed on the city from the north, but it mentions neither the storming of a stockade defended by the rebels nor a movement by rebel cavalry to threaten the rear of the Tang army, both of which figure in the Old Tang History biography of Li Sheng.55
The announcements of victory also differ significantly from the dynastic history in their treatment of heroic action. Although accounts of single combat, individual feats of arms, and horrendous wounds may be very unevenly distributed through the Old Tang History, they are entirely absent from all extant specimens of the Tang announcement of victory. The bravery of Liu Huiji and his two comrades in batde against the Man people in the early 670s, mentioned in the passage from Luo Binwang's lubu quoted above, is presented in vague, abstract language; there is no hint of the specific acts of violence and heroism described in such graphic and concrete terms in the history. Yet this is as close as the announcements of victory come to depicting the heroic action of individuals.
Narrative Maneuvers
155
Tang Scholars and the Representation of Battle
The Tang announcement of victory served a variety of purposes beyond the simple reporting of a military event. Much of the abundant verbiage in its preliminary sections reconfirmed and reinforced the official self-image of the emperor and his court as supremely virtuous—and the foe as utterly incorrigible, thus validating military action that might otherwise cast doubt on the monarch's virtue by highlighting his failure to transform the recalcitrant by means of his moral influence. The lubu is clear and concrete at only two points, when it lists the names of Tang officers and when it details the spoils of battle and the losses inflicted on the enemy. The latter offers tangible evidence of the magnitude of the victory, the power of Tang arms, and the merit of the victorious commander; the former provides honorable mention for those who assisted in the victory, whose names will be read out to the assembled officials and who may be in line for reward and promotion. The emphasis given to these elements in the lubu is readily comprehensible.
It is not so easy to explain the consistent representation of combat in terms of ornate, conventional images that reveal virtually nothing of what might have been happening on the battlefield. The failure to speak of basic tactics may have something to do with their very ordinariness and the fact that they were normally the business of subalterns, not senior commanders. The historical record to which the lubu might eventually contribute was intended to illuminate the lessons of the past for rulers and statesmen, illustrating moral principles and problems of imperial government and providing models of ethical behavior. It was certainly not intended to contribute to the technical training of junior officers.
Another factor worth considering is the background of the men who wrote the announcements of victory. Of the ten extant lubu, seven were drafted by secretaries rather than by the victorious commanders themselves. From what we know of the careers of the five authors of these seven lubu (two wrote more than one), it would seem that they were men of literary talent and training who had never commanded soldiers in battle.56 It is perhaps not surprising that all of these men, confronted with the chaos of battle, chose to fall back on hoary cliches. One lubu speaks of "ten thousand crossbows shooting together," a phrase
156
Narrative Maneuvers
borrowed from the Historical Records (Shi ji, early first century bce) biography of Sun Bin (a strategist of the fourth century bce), and the model announcement of victory in the Taibai yinjing says there was "enough flowing blood to float a pestle"—an expression that had been used in stories about the ancient battle of Muye (ca. 1045 bce) and was already considered a cliche in the sixth century.57 As the modern military historian John Keegan has observed, the grip of precedent on the writing of battle narratives is tenacious: authors have often turned to earlier literary models to help them make sense of events they do not understand.58
Yet the lubu really do not make sense of the battles they describe. Their language does not clarify; rather, it obscures. Perhaps, as a result of their proximity to the engagements they describe, the lubu authors— or at least those who first established the convention for representing battle in these documents—knew just enough about the confusion of combat to despair of being able to offer a clear, logical, and accurate account of what had just transpired on the field of batde and why one side had won and the other had lost.
How, then, did heroic action and cunning stratagem enter the picture? Since these elements are emphasized in the Old Tang History biographies, the obvious place to look is the sources for those biographies. The "account of conduct" (xingzhuang) was "a sort of extended curriculum vitae" for a recently deceased official of high rank, which his relatives or former subordinates were required to submit to the Department of Merit Assessment (under the Ministry of Personnel) within one year of his death.59 After verification of the information contained therein, the account of conduct was forwarded to the Historiography Office, where it normally provided the basis for a biography that would be written for inclusion in the Veritable Record of the reign in which the official had passed away. And the Veritable Records in turn provided the basis for most of the biographies included in the Old Tang History, either direcdy or by way of the National History compiled by Liu Fang around 760.
When we examine these biographical documents, however, we find very little concrete, anecdotal material that can be connected to specific military events (as opposed to the ever present topoi and cliches). The extant xingzhuang of the Tang general Ma Sui briefly mentions his cutting of the grass around his army's position at the battle of the Huan
Narrative Maneuvers
157
River in 782 in order to frustrate the enemy's attempt to defeat him by means of an incendiary attack. Yet Ma's biography in the Old Tang History includes a far more elaborate account of this battle, complete with detailed description of several stratagems not mentioned in the account of conduct. In this case, the xingzhuang is clearly not the major source for the battle narrative found in the dynastic history.60 The xingzhuang of Yan Zhenqing (708-784), a civil official who organized loyalist resistance against the An Lushan rebellion in Hebei, is also disappointing in that it has much to say about operational dispositions but very little about specific battles/'1 A sampling of the much more numerous private funerary writings dealing with the lives of military men, including both stela inscriptions and epitaphs (muzhiming), reinforces the impression gained from accounts of conduct. Out of twelve such texts, several mention battle plans, dispositions, and deployments, but never in much detail. Stratagems are rarely mentioned, and then only in the vaguest terms—catching the enemy unprepared, attacking the enemy's "vacuity," and so forth.62 Only one of the twelve inscriptions includes a concrete instance of heroic action: as a young officer during the An Lushan rebellion, the future Tang general Hun Zhen (d. 799) killed a rebel commander by putting an arrow through his left shoulder.63
Other possible sources of specific, anecdotal information about the exploits of military men are the edicts of enfeoffment and reward handed down by the emperor, which typically make some reference to the deed or accomplishment that merited the specified reward. Enfeoffments, in particular, were supposed to be reported to the Historiography Office.64 Some seventeen examples of such edicts can be found in two chapters of the Collected Edicts of the Tang (Tang da zhao ling ji), a Northern Song compilation.65 These examples, however, do not provide the sort of specific information about stratagems and heroic action that is found in the biographies of the Old Tang History. An edict granted to the late eighth-century general Li Sheng notes that he "personally put on armor and helmet" and led from the front, while the edict rewarding Li Sheng's son Li Su for his capture of Caizhou in 817 mentions that he took advantage of snowy weather to surprise the enemy.66 But this is the most we are told.
Though the Tang edicts seem to offer little reward to the modern military historian, there is some evidence in the Sui History (Sui shu, an
158
[Narrative Maneuvers
early Tang official history dealing with the years 581-618 and completed in 636) to suggest that reward edicts did sometimes contribute to battle accounts in the biographies. Of particular interest is the biography of Daxi Changru (dates unknown), who led a Sui expedition against the Turks in 582. The text describes the heroic fighting retreat of Daxi's badly outnumbered force, notes that Daxi himself was wounded five times, and goes on to quote from the emperor's reward edict—raising the possibility that at least some of the earlier, more concrete information regarding this general's exploit may have been borrowed from unquoted portions of the same edict.67
Official reports other than announcements of victory—and possibly even announcements of victory other than the few that have survived— may also have contributed some of the anecdotal material found in the longer battle accounts in the Old Tang History. The lubu was only one among a variety of documentary forms used for reporting military events.6" It seems to have been the most elaborately styled, designed to please the ear when read aloud to the assembled court officials and foreign dignitaries. This literary quality surely contributed to the preservation of the lubu (and especially the most elaborate of them) when other, more mundane sorts of military reports were lost. The two "victory letters" (jieshu biao) drafted by Dugu Ji (725-777) in the early 760s, to my knowledge the only surviving examples of this Tang genre, offer an interesting comparison. They are considerably shorter, are less extravagantly phrased, and provide a great deal more substantive information in relation to their length. Dugu's account of the defeat of a Zhejiang bandit gang is particularly informative, providing a detailed description of the decoy stratagem by which the bandits were lured into an ambush and even noting the weapons carried by different contingents of government troops.69
In addition, detailed accounts of the heroic deeds of individual combatants were almost certainly provided in the "provisional" memorials, separate from the lubu itself, that generals were allowed to submit on behalf of subordinates who earned extraordinary merit through such actions as capturing the enemy's leader or seizing their flag.70 Several examples of what appear to be reports of individual battlefield merit from the Beiting protectorate-general are among the Silk Road documents now in the collection of the Yurinkan Museum (Kyoto).71
Narrative Maneuvers
159
Individually or in combination, these several types of sources probably provided official historians writing biographies for the Veritable Records with the raw material they needed to create batde narratives that were more complete, coherent, and intellectually satisfying than those offered in the announcements of victory. Where events on the battlefield are generally quite murky in the lubu, with little explanation offered for the outcome of the combat, the addition of stratagems, ingenious battle plans, and other episodes on the field of battle enabled the official historian (much farther removed from the scene of action in time and space and therefore possibly less awed by it) to bring cause and effect together in such a way that the battle became a story rather than a scene.
The Tang official historian would have been pushed in this direction not only by a need for narrative coherence but also by the precedent of earlier historical writing. Ancient works such as the Commentary of Zuo (Zuozhuan), the Historical Records (Shiji) of Sima Qian (ca. 145-86 bce), and the Han History (Han shu) of Ban Gu (32-92 ce) were regarded with great respect and enjoyed wide circulation among the educated elite of Tang times. Those Tang officials who were assigned to the writing and compilation of history, representing the most brilliantly literate elite within the civil service, would not have been unfamiliar with the great ancient models of their craft. An outstanding example of the influence of precedent on the Tang historian is the treatment of the battle of Mangshan (618) found in the Sui History biographies of Wang Shichong and Pei Renji (d. 619), which both present a picture that is remarkably similar to the account of the battle of Jingxing (205 bce) found in Sima Qian's Historical Records.72 Wang's battle plan at Mangshan is essentially the same as that adopted by Han Xin (d. 196 bce) at Jingxing; he positions his main force with its back to a river, while sending a smaller cavalry detachment through the hills to attack the enemy camp from the rear. The options discussed by Wang's opponent Li Mi and his generals in their council of war differ little from those debated by Han Xin's opponents on the eve of the batde of Jingxing and are even couched in some of the same language.73
Not only were Tang official historians influenced by the form of earlier histories when they wrote about battles, but at a deeper level, their emphases are also consistent with ideas about armed conflict dating back to pre-Qin times, before the unification of the empire in 221 bce. Texts such
160
Narrative Maneuvers
as the Commentary qfZuo and the Historical Records offer many examples of battles decided by successful stratagems, such as the ruse of diminishing the number of his campfires by which Sun Bin lured his archrival to destruction at Maling in 341 bce.74 The tendency of the Tang battle narratives to represent the victorious general as complete master of the situation echoes classical military writings that reveal the commander, in the words of Mark Edward Lewis, as one whose "sage-like powers of calculation and assessment allowed him to discover the meaningful pattern or order in the flux of the campaign and then use that pattern for his own purposes by skillful maneuver and decisive action at the proper moment."75 Moreover, the attention given to morale and psychology in the Tang material also resonates with important aspects of the ancient discourse on the art of war.76
These elements, all of which involve the application of intellectual power in order to achieve military success, appear to be a constant running through the Tang battle narratives from the beginning of the dynasty to its end. In contrast, the representation of heroic action in the same texts is, as we have seen, very unevenly distributed, with the greatest activity concentrated in the period before 650. This pattern strongly suggests that the bloody deeds of individual warriors did not have quite the same level of continuing attraction for Tang official historians. One possible explanation for the decline in the attention given to heroic action would be the changing character of the Tang political elite. During the first two Tang reigns, many of the highest court offices—including the throne itself—were held by warriors who had literally fought their way to power, men who were proud of their bloody hands. These individuals (and their kinsmen and former comrades in arms) were in a position to insist that their battlefield accomplishments be included in the historical record. After the death of Taizong (Li Shimin, the second Tang emperor) in 649, with the advent of palace-raised emperors, the actuarial decline of the old warriors, and the increasingly important role played by the examination system in recruiting elite-track officials, such pressures must have declined rapidly. This interpretation is lent some support by a resurgence in the representation of heroic action at the very end of the Tang period, a phenomenon more visible in the pages of the Old History of the Five Dynasties (Jiu Wudai shi, completed in 974) than in the Old Tang History.77 This was a time when fighting men once again occupied the political center, with
Narrative Maneuvers
161
the warrior spirit especially emphasized in the regime established by the Shatuo Turk leader Li Keyong (856-908) in Hedong (today's Shanxi province).
The unevenness of the attention given to the violent deeds of individual warriors is of a piece with other aspects of the representation of battle in Tang historical writing. Together with the relative brevity of the passages dealing with actual combat and the general neglect of weapons and tactics, it bespeaks a real lack of interest in the technical aspects of warfare as it was understood at the lower command levels. In contrast, the historian's attention is drawn again and again to the more abstract, intellectual, and nontechnical aspects of conflict. More space is often given to the council of war, and to basic strategic choices (such as whether or not to offer battle), than to the battle itself, and once the action is under way, the emphasis is on superior cleverness as the deciding factor. A basic feature of all but the most cursory battle accounts in the Tang histories is that the reason for the outcome of a battle is readily comprehensible. That reason is almost always a battle plan reflecting the superior acumen of one of the two commanders, a plan that is often based on understanding and manipulation of the minds and emotions of the enemy army, its commander, or even one's own soldiers. Rarely does heroic action prove decisive in determining the outcome of an engagement, even when heroes are present on the field of battle. If there is a single message conveyed by most of the Tang battle narratives, it is that battles are won by cleverness rather than strength. Behind this message, we may detect a hint of the Confucian generalist's disdain for the specialized skills and training of military men and the implication that the scholar's talents may actually be better fitted to the successful exercise of military command.
Appendix A:
List of Tang "Announcements of Victory"
1. "Yaozhou's Defeat of the Perverse Rebels Nuo-mo-nong and Yang Qianqing." Written by Luo Binwang (biographies in JTS, chap. 190A, p. 5006; XTS, chap. 201, p. 5742); submitted by Liang Jishou, commander of the expeditionary army for the Yaozhou route, probably in 672 (though possibly as late as 673 or 674). Text in WYYH, chap. 647, pp. la^a; QTW, chap. 199, pp. 2a-6a. Campaign: ZZTJ, chap. 202, p. 6368; JTS, chap. 5, p. 96; XTS, chap. 3, p. 70, chap. 222B, p. 6324.
^Narrative maneuvers
2. "Defeat of the Rebel She-meng-jian and Others in Yaozhou." Written by Luo Binwang; submitted by Liang Jishou. Text in WYYH, chap. 647, pp. 4a-6b; QTW, chap. 199, pp. 6a-10a.
3. "Pacification of the Jizhou Rebels and the Khitan." Written by Zhang Yue (biographies in JTS, chap. 97, p. 3049; XTS, chap. 125, p. 4404); submitted in 697 by Wu Yizong, commander in chief of the expeditionary army for the Shenbing route (biographies in 77$, chap. 183, p. 4737; XTS, chap. 206, p. 5842). Text in WYYH, chap. 647, pp. 6b-l la; QTW, chap. 225, pp. la-6b. Campaign: ZZTJ, chap.
206, pp. 6517, 6520-6521.
4. "Defeat of the Khitan." Written by Fan Heng; submitted in 733 by Xue Chuyu, senior administrator of the Youzhou area command (biography in JTS, chap. 93, p. 2985). Text in WYYH, chap. 647, pp. lla-14b; QTW, chap. 352, pp. 10b-16a. Campaign: ZZTJ, chap. 213, pp. 6800-6802 JTS, chap. 199B, p. 5353.
5. "Hexi's Defeat of the Tibetan Rebels." Written by Fan Heng; probably submitted ca. 746 by Wang Zhongsi, military governor of Hexi (biographies in JTS, chap. 103, p. 3199; XTS, chap. 133, p. 4551). Text in WYYH, chap. 648, pp. la-5a; QTW, chap. 352, pp. 16a-20b. Campaign: ZZT7, chap. 215, p. 6871 (?).
6. "The Jiannan Military Governor's Defeat of the Xishan Rebels." Written by Yang Tan (XTS, chap. 7IB, p. 2356). Probably submitted at some point during the 750s. Text: WYYH, chap. 648, pp. 5a-7a; QTW, chap. 377, pp. 16b-18b.
7. "Guizhou Defeats the Xiyuan Rebels." Written and submitted by Yang Tan, prefect of Guizhou (see Yu Xianhao, 2000, Tang cishi kao quan bian, vol. 5, p. 3244). Probably submitted in 759. Text: WYYH, chap. 648, pp. 7a-9b; QTW, chap. 377, pp. 19a-22a. Campaign: XTS, chap. 220B, p. 6329.
8. "Li Sheng's Recovery of the Western Capital." Written by Yu Gongyi (biographies in JTS, chap. 137, p. 3767; XTS, chap. 203, p. 5784); submitted in 784 by Li Sheng, deputy marshal of the Shence forces of the capital region (biographies in JTS, chap. 133, pp. 4668-4669; XTS, chap. 154, p. 4863). Text: WYYH, chap. 648, pp. 9b-14a; QTW, chap. 513, pp. 16a-20b. Campaign: ZZTJ, chap. 231, pp. 743 4-743 5; JTS, chap. 12, p. 342.
9. "Defeat of the Tibetans." Written and submitted in 801 by Wei Gao, military governor of Jiannan West (biographies in JTS, chap. 140, pp. 3822-3826; XTS, chap. 158, p. 4933). Text: QTW, chap. 453, pp. 4b-6a. Campaign: ZZTJ, chap. 236, pp. 7597-7598; JTS, chap. 13, p. 395.
10. "Recovery of the Capital." Written and submitted in 883 by Yang Fuguang, eunuch generalissimo (biographies \n JTS, chap. 184, pp. 4772-4773; XTS, chap.
207, pp. 5875-5877). Text: QTl^chap. 998, pp. 16a-17b. Campaign: ZZTJ, chap. 255, pp. 8293-8295.
Appendix B:
The Announcement of Victory from the Taibai yinjing
To the Secretariat-Chancellery and the Ministry of War under the Department of State Affairs:
The military governor for the (blank) province and (substantive office), your servant (name), says: Your servant has heard that the Yellow Emperor raised the army of
Narrative Maneuvers
163
Zhuolu, and that Yao and Shun fought the hattle of Banquan. Although the Kingly Way was lofty in ancient times, it was still not possible to avoid punitive expeditions in the four directions. The virtue of our state surpasses that of Tang [Yao] and Yu [Shun]; its merit is the model for the Xia people. The deranged Di are like wriggling worms; they are benighted, deluded, and disrespectful. Forming packs like dogs or sheep, they violated our border posts. Your servant now ordered the chief troop commander (substantive office) (name) to lead horse and foot (however many) men to form the vanguard, and on left and right further charged the outguard commander \yuhou\ (substantive office) (name) to lead (however many) crossbowmen as surprise troops [qibing] to set an ambush at (such-and-such) a place, and the outguard commander's subordinate officer [zongguan] (name) to lead (however many) men armed with long-handled sabers to act as a reserve force, and the deputy military governor (substantive office) (name) to lead (however many) Chinese and foreign troops as the main body. In the (blank) month on the (blank) day at the (blank) hour, our roving cavalry encountered the main army of the enemy at the (name) mountain or river. Dust rose to blot out the sky; flags and banners covered the field. Your servant ordered the chief troop commander (substantive office) (name) to lead the main body [?] to face their charge, and the left and right outguard commanders to extend the two wings. When the battle was just reaching its greatest intensity, the concealed troops struck with stealth. The enemy masses were terrified; the crossbows and long-handled sabers of the outguard commander (name) arrived one after the other. Wherever their points and blades were applied, there was enough flowing blood to float a pestle; wherever the crossbow bolts reached, the chariot tracks were chaotic and the banners were scattered. Casting away their armor and discarding their weapons, the enemy fled. Our army pursued the fugitives for fifty li, fighting (however many) engagements [zhen] from the_y/w hour [3:00 to 5:00 am] to the you hour [5:00 to 7:00 pm]. All of the [enemy] killed and captured and the items taken are just as enumerated above. How could this have been brought about by the merit of men? What we trusted in was the Imperial merit! Your servant respectfully dispatches the vanguard commander (substantive office) (name) to make this known by submitting an announcement of victory in the special hope that it will be announced both within the country and abroad, and will be used to grace the books of history. Your servant (name) bows his head respectfully and speaks. (Blank) year, (blank) month, (blank) day. Submitted by the chief secretary (substantive office), your servant (name).
Source: Li Quan, 1988, chap. 7, pp. 611-612.
Appendix C:
Lubu Supplement from the Taibai yinjing
The administrative assistant (substantive office) (name) and the chief administrator of the expeditionary army (name) have been tasked with the matter of the announcement of victory over the rebel (name) memorialized by the (substantive office) military governor for the (blank) province. We have taken the (names) walled towns of the enemy, numbering (however many); we have captured their leaders (names), (however many); we have killed their senior generals to the number of
1UT
(however many); we have taken (however many) heads. We have obtained (however many) of the enemy's horses, (however many) suits of armor, (however many) flags, (however many) bows and crossbows, (however many) arrows, (however many) spears and shields, and (however many) garments. That which was gained has all been spoken of.
Source: Li Quan, 1988, chap. 7, pp. 610-611.