The fall of the Tang (618–907) was a cataclysmic event with wide repercussions in China and beyond. Decades of civil war, epidemics, and famine left Chinese society in deep disarray. New military and regional elites had unseated the ruling aristocracy. Previously semi-autonomous provinces became independent kingdoms; vassal states and protectorates claimed sovereignty. Announcing the dawn of an uncertain new order, a polycentric structure of power emerged in China that prevailed through the half century known as the Five Dynasties (907–65). The drawn-out transition to the Song dynasty (960–1127) engendered a broad national renewal that transformed Chinese medieval society on the threshold of the early modern age.

The empire’s territorial dissolution had begun decades before the final collapse. As drought and harvest failures took their toll on the rural population, dislocation followed famine and plagues. Religious writers described a society undergoing doomsday retribution.Footnote ¹ Rebels pinned their hopes on regime change. At the same time, pressure on the Tang’s borders built relentlessly. The frontiers of the Tang were, like the limites of the Roman empire, deep defensive zones with ethnically mixed populations. The difficulty of controlling these literally tested the limits of empire: unsustainable social and financial costs of near-permanent border warfare, the inexorable rise of provincial autonomy, and growing civil unrest triggered by insecurity and economic distress. Aggravating these, power struggles and government factionalism resulted in discordant policies and undermined concerted action. The “natural” allegiance of provincial leaders and nominal vassal princes could no longer be taken for granted. Wary of the growing might of military governors, emperors perilously obstructed their effectiveness and entrusted the control of military and regional affairs to palace eunuchs.

This book examines the run-up to the Tang’s collapse through the trajectory of Gao Pian 高駢 (821–87). At the height of his career, Gao controlled the empire’s essential economic resources as military governor of Huainan, and the bulk of its armed forces as commander-in-chief of the Joint Expeditionary Armies. In his capacity as Salt, Iron, and Transport commissioner, he held the financial reins of southern China and oversaw the principal arteries of the transportation and distribution networks. Gao’s military career spanned the tail end in the 840s and 850s of centuries of struggle with Tibet in Inner Asia, two wars with the southwestern kingdom of Nanzhao 南詔 in the 860s and 870s, and the decade of internal insurrection led by Huang Chao 黃巢 and his associates between 875 and 885. Narrating the breakup of the Tang through Gao Pian’s perspective offers a close-up view of the empire’s military, fiscal, and administrative unraveling.

Tang and Its Neighbors

The first half of the Tang had been a phase of imperial expansion. Seeking to restore the territorial extent of the Han empire between the second century bce and the third century ce, Tang armies surmounted vigorous resistance and costly setbacks to push China’s boundaries to the Pamir Mountains in modern Tajikistan, into Mongolia, Manchuria, the Korean peninsula, and Vietnam. At the height of this expansion, the insurrection of An Lushan 安祿山 (755–57) upended the Tang’s imperial ambitions. The rebel military governor of several northeastern frontier commands in Hebei came close to overthrowing the dynasty, forcing the Tang to redeploy its overextended armies to protect the capital and subdue the rebellion. Tang sway over the mixed borderland prefectures governed by “Loose Rein” 羈縻 power-sharing arrangements declined. Caught on the back foot, the Tang military no longer projected in-depth power abroad. Neighboring states seized the advantage. The balance of power in Asia shifted.

The Tibetan empire 吐蕃 had first established its capital in Lhasa under the reign of Srong-btsan-sgam-po 松贊乾布 (r. 629–49), then occupied the Tibetan plateau and annexed large swaths of Central Asia, including Nepal and Kashmir. For two centuries, Tibet and Tang China engaged in a strategic rivalry for control of the frontier prefectures of eastern Qinghai and southern Gansu, the oases states and trade routes in the Hexi corridor, and the horse-breeding grasslands of Inner Asia.Footnote ² An offensive launched by the Tang Emperor Xuanzong 玄宗 (r. 712–56), aiming to limit Tibetan territorial expansion with the help of the Uighur Khaganate, a sprawling Turkic empire in Eurasia, was on the cusp of success when the An Lushan rebellion broke out and Xuanzong was driven into exile in 755. The Tibetan ruler Khri-srong-lde-brtsan 墀松德贊 (r. 755–97) swiftly filled the military vacuum left by withdrawing Tang forces in Qinghai and Gansu. Tibet incorporated eighteen prefectures previously administered under the Tang protectorate of Anxi 安西 (Kucha), twelve former Chinese prefectures in Kokonor (Qinghai), and four Tang garrisons in the Tarim basin. In 763, a Tibetan army of 200,000 invaded Chang’an and forced Xuanzong’s grandson, Emperor Daizong (r. 762–79), to temporarily abandon the capital. The Tibetan empire continued to win Tang territories in Gansu during the late eighth century until the rising Uighur empire supplanted the Tang as Tibet’s principal rival in Inner Asia. Tibet engaged in a prolonged and costly war to contain the Uighurs militarily, while the Tang opted for marriage diplomacy. In 821 a mutual nonaggression pact acknowledged Tibet’s occupation of Chinese territories in Longyou 隴右, Hexi 河西, Anxi, and Beiting 北庭 (Jimsar 吉木薩爾).Footnote ³

Just as Tibetan power began to wane in the 840s, however, a new threat arose on the Tang’s southwestern frontier. Wedged between China and Tibet, the kingdom of Nanzhao was a federation of six tribal principalities (zhao) located in the Dali plain of modern Yunnan province. Under the leadership of the Meng 蒙 clan, the ruling family of the southernmost zhao, it expanded through Yunnan and parts of Burma, Laos, Vietnam, Guizhou, and Sichuan. Emperor Xuanzong had endorsed the initial consolidation of the federation as a kingdom. In 738 he conferred on its ruler, Piluoge 皮羅閣 (r. 728–48), the title prince of Nanzhao, hoping to secure an ally against Tibet and maintain overland access to the Tang protectorate of Annan 安南 (North Vietnam). Border conflicts nevertheless flared up between the two neighbors. In 756, Piluoge’s successor, Geluofeng 閣羅鳳 (r. 748–79), took advantage of the An Lushan rebellion to launch a raid deep into the Chengdu plain.Footnote ⁴

After the kingdom’s breach with the Tang, the Taihe Stele of Virtuous Suasion 太和德化碑 inscription erected in the Nanzhao capital near modern Dali in 766 proclaimed that Nanzhao’s diplomatic balancing act was the result of legitimate grievances against the Tang, which compelled Geluofeng to seek the protection of Tibet.Footnote ⁵ During the years 789–94, Wei Gao 韋皐, the energetic military governor in Western Sichuan 西川 (the region traditionally known as Shu 蜀), succeeded in containing the Tibetan empire while applying a judicious dosage of force and diplomacy to persuade Nanzhao to switch its allegiance back to the Tang.Footnote ⁶ When Wei died in 805, after twenty-one years of quasi-autonomous rule in Sichuan, he left behind a stable frontier. For a while, the Nanzhao ruler Quanfengyou 勸豐佑 (r. 824–59) sent the kingdom’s young leaders across the border to be educated in Chengdu.Footnote ⁷

These peaceful relations ended abruptly, however, with a devastating incursion by Nanzhao troops into the Sichuan plain in the winter of 829–30. During their brief occupation of the western suburbs of Chengdu, the Nanzhao tribesmen plundered and laid waste to part of the city. The Chengdu poet Yong Tao 雍陶 (jinshi 834) lamented the despair of thousands of abducted townspeople.Footnote ⁸ To cope with the aftermath of the disaster, the prominent statesman Li Deyu 李德裕 (787–850) was appointed military governor of Western Sichuan. During Li’s tenure, from 830 to 832, the scarred and impoverished province slowly recovered. Of an estimated 9,000 captives 4,000 returned, including randomly seized inhabitants, artisans, performers, and a prized Persian eye physician.Footnote ⁹ Temporarily thwarted by Li Deyu’s defenses, Nanzhao turned its attention to Burma where earlier forays had already spearheaded a breakthrough to the Gulf of Bengal.Footnote ¹⁰ In 832 Nanzhao sacked the Burmese Pyu 驃 kingdom, abducting many of its citizens as well. Thereafter, the kings of Nanzhao styled themselves lords of Pyu 驃信苴.Footnote ¹¹

Simultaneously, Nanzhao eyed the fertile Red River valley of Annan. A tentative raid was launched there in 816. Border hostilities continued to simmer through the 820s. In 846 and 854, the Tang protectorate’s heartland of Jiaozhou 交州 surrounding modern Hanoi came under attack.Footnote ¹² Relations with the Tang took a further turn for the worse after the succession of Quanfengyou’s son, Shilong 世隆 (r. 859–77), whose title combined two taboo characters from the personal names of the eminent Tang emperors Li Shimin 李世民 (Taizong) and Li Longji 李隆基 (Xuanzong). Tang writers altered the offending title to Qiulong 酋龍, meaning “chieftain Dragon.” In addition to flouting the diplomatic etiquette befitting a vassal prince, Shilong demanded the status of a junior relative of the Tang emperor. Shilong would pose the main threat to the Tang’s borders in the coming decades and become Gao Pian’s implacable adversary on several fronts.Footnote ¹³

Northeast Asia, where at different times the Khitan and Jurchen each held sway over parts or all of post-Tang China, was another region of cross-border tension. The Sui dynasty (581–619) had launched a ruinous war against the powerful Koguryŏ 高句麗 kingdom (trad. 37 bce–668 ce) straddling the territory of modern Manchuria and North Korea. To contain Koguryŏ’s expansion, the emperor Taizong renewed that campaign with the help of tribal and Turkish allies. The Tang eventually accomplished the destruction of Koguryŏ twenty-three years later. Silla 新羅 (trad. 57 bce–935 ce) gradually emerged victorious among the Three Kingdoms occupying the Korean peninsula. Expanding its base in the southern part of the peninsula, Silla first annexed its rival Paekche 百濟 in 660, then took advantage of the fall of Koguryŏ in 668 to establish its rule over the whole peninsula. Bohai 渤海 (Kor. Parhae), the Manchurian part of former Koguryŏ, broke away a generation later to form a separate kingdom under the indigenous Mohe 靺鞨 leader Da Zuorong 大祚榮 (Kor. Tae Ch’oyŏng, d. 719).Footnote ¹⁴ The Tang attempted to establish an eastern protectorate Andong 安東 in Koguryŏ’s former capital Pyongyang 平壤, but its seat had to be repeatedly relocated to more accessible prefectures in Liaoning and Hebei before being abandoned altogether in the wake of the An Lushan rebellion. Tang China’s presence in the Korean peninsula and southern Manchuria came to an end in 756.

Bohai and Silla, two autonomous monarchies with flourishing economies, thrived on exchanges with the continental power as nominal vassal states. Their ruling elites were receptive to Chinese culture and obliged the imperial court by periodically sending hostages, students, and tribute missions to Chang’an.Footnote ¹⁵ Chinese literati-officials approved of Bohai and Silla as “Sinicized.” Students from the two kingdoms, such as Ch’oe Ch’iwŏn 崔致遠 (855–949) from Silla (see under “The General’s Scribe” later in this introduction), regularly competed in the imperial civil service examinations at the Tang capital and either served as military or civil officials in China or put their expertise on Chinese affairs to use in their home countries.Footnote ¹⁶

Internal Fault Lines

The An Lushan rebellion not only weakened the empire’s external influence but also deepened latent fault lines within China: a growing power imbalance between the center and the provinces, government factionalism, and socioeconomic fractures. The geographical distance between the Tang’s northwestern political and strategic center of gravity and its southeastern economic powerhouse accentuated the natural divide between North and South China. The court in Chang’an depended on massive grain shipments from the lower Yangzi via the Grand Canal for public revenues, supplying frontier armies, and feeding the capital’s inhabitants numbering, in the words of the scholar-official Han Yu 韓愈 (768–824), “no less than one million.”Footnote ¹⁷ Since the Sui dynasty, this transportation artery linked previously isolated canal and fluvial sections into a navigable system covering Huainan, Henan, and the North China Plain. Well-furnished garrisons guarded this vulnerable lifeline against frequent attacks and obstruction by insurrectionists. The lower Yellow River formed another interior frontier, separating the comparatively loyalist South from the heavily militarized and notoriously autonomous provinces of Hebei. After An Lushan launched his insurrection against the Tang from there, Hebei ceased to remit taxes to the Tang emperor and appointed its own officials.Footnote ¹⁸ Hebei had maintained a close relationship with the former Koguryŏ kingdom and after 668 continued to pursue commercial links with the Khitan, Xi, and Bohai.

In the first half of the eighth century, the Tang court sought to rein in powerful frontier military commands by dividing the empire into fifteen circuits of inspection. These were placed under the surveillance of commissioners reporting to the Censorate, a body tasked with ensuring government efficiency. Over time, however, the commissioners in their turn acquired increasing autonomy and assumed executive powers in the regions under their supervision.Footnote ¹⁹ For the defense of the frontier zones under their responsibility, military governors maintained militia troops, that is, locally recruited, self-sufficient regiments that engaged in farming when they were not campaigning. In parallel with the institution of circuits of inspection, the court carried out a border defense reform replacing the regional militia troops with professional standing armies. In the reign of Xuanzong, these armies were organized into forty regional commands placed under the authority of a military governor, each of whom frequently oversaw the command’s civil administration as well.Footnote ²⁰ Additional commands were established in the interior to control internal insurrection, the system providing the backbone of the late Tang’s provincial organization.Footnote ²¹ To win the allegiance of such regional commands, post-An Lushan emperors were constrained to cede growing autonomy to their governors. While this expedient procured them the military forces needed to defend the frontiers and maintain internal stability in the short term, it inescapably sowed the seeds of the empire’s eventual breakup. By the middle of the eighth century, regional military governors controlled 85 percent of the empire’s armed forces.Footnote ²² A new study by the historian Nicolas Tackett demonstrates that the court succeeded in temporarily reversing this trend during the first half of the ninth century, when the capital elite again dominated provincial recruitments.Footnote ²³ Nevertheless, in 885, in the wake of the Huang Chao rebellion, only fifteen out of a total of fifty regional governors remained loyal to the emperor. By the end of the Tang, a dozen governors, ruling like monarchs over country-sized territories, would dispose of the empire’s military and fiscal resources at will.Footnote ²⁴

Anticipating the risk of such a development, Emperor Dezong (r. 779–805) took additional measures to oversee provincial appointments and strengthen imperial surveillance of the military forces attached to regional commands. To implement his policy, Dezong relied on palace eunuchs, bypassing the outer court bureaucracy. Originally imperial household intendents, eunuchs increasingly served as commanders of the palace armies and as army supervisors of provincial troops, gaining a considerable hold on military affairs.Footnote ²⁵ Eunuchs also intervened in military and provincial appointments. Since the reign of Emperor Xuanzong, individual eunuchs had reached senior official rank in the state bureaucracy. As intermediaries controlling access to the emperor, they wielded outsized informal powers that lent themselves to favoritism and corruption. Organized into adoptive lineages, eunuch dynasties formed alliances with military and other interest groups, competed for the spoils of lobbying and graft, and fueled contention between the political factions of the dominant chief ministers, whose public confrontations in the course of the ninth century came close to paralyzing government policy.Footnote ²⁶

In the first half of the ninth century the Tang’s registered population stood at about fifty million.Footnote ²⁷ Aggravating the burdens of conscription and levies to sustain near-continuous border warfare, a prolonged drought tormented the populations of large swaths of the country during the reign of Emperor Xizong 僖宗 (873–88). When Xizong ascended the throne, the Hanlin academician and future chief minister Lu Xie 盧攜 (d. 881) gave a stark account of conditions in Henan and Shandong. Lu had personally inspected the devastation caused by drought there the previous year. All the way from Henan to the East China Sea, he reported, the normal yield of spring wheat had halved. The autumn harvest and winter vegetable crops yielded even less. Farmers had already consumed the next season’s seeds and were reduced to eating tree leaves. A combination of famine, taxation, and banditry drove farmers off the land, leaving those who stayed behind to face starvation.Footnote ²⁸ Dislocation and malnutrition spread disease and epidemics. When policing and civil administration became all but impracticable, the affected areas rose in insurrection.

The plight of war refugees, deserting soldiers, and roving gangs of displaced farm workers fanned the ambient violence born from border conflicts and ethnic uprisings until it consumed all levels of society. A progressively isolated central government faced provincial military governors increasingly preoccupied with consolidating their own personal bases. Seeking to curtail the power of the military governors, the court also restricted their capacity to defend the empire. For their part, the military governors struggled to simultaneously contain threats to their borders and internal insurgency in their territories. Suspected of defying Tang imperial authority, they found themselves challenged by insubordinate generals and officials under their own orders. The ever-present risk of troop mutiny needed to be factored into every military operation. A standing army, capable of either safeguarding or imperiling a governor and his headquarters, was a two-edged proposition. It will be seen that Gao Pian mostly relied on small, mobile, and well-trained units of proven loyalty and disbanded large troop concentrations as soon as he could.

A Broken Compact

The themes of allegiance and disobedience run like a red thread through the writings of the late Tang and Five Dynasties period. A minister’s unconditional loyalty to the sovereign was an assumed imperative of social and political morality. Yet inevitably, in times of dynastic change or regime-threatening crisis, the fortunes of war or the reversals of politics forced public figures to choose their camp or fend for themselves. As the regular channels of recruitment broke down, defectors and capitulated rebels replenished the ranks of officialdom. Anxious to escape public censure, individuals struggled to reconcile personal dilemmas and to rationalize expedient decisions. Many officials changed allegiance, some repeatedly. Official historiography summarily condemned ministers who had served more than one master as turncoats and military governors suspected of secessionism as traitors.Footnote ²⁹ History branded Feng Dao 馮道 (882–954), a chief minister under each of the successive Five Dynasties, as the arch-turncoat, despite the precautions he had taken in crafting an apologetic autobiography.Footnote ³⁰ Mindful of the judgment of history, officials professed to sooner die than forsake their sovereign or serve another lord. The rhetorical nature of such declarations is evident from their theatrical setting: Typically, when a new regime arrived in power, civil servants of the previous government would proffer their own execution, only to be reinstated with commendations for loyalty.Footnote ³¹

Long after Tang emperors could credibly lay claim to empire-wide supremacy, ministers and generals continued to comply rhetorically with the protocol of sovereign–vassal allegiance. The case of Hebei is instructive. Notwithstanding the defeat of An Lushan, the three provinces constituting Hebei preserved their autonomy. The chief minister, Li Deyu, cautioned its leaders: “Although the Hebei military are powerful, they are not independent. They need the sanction of appointments and entitlements by the court to assuage their troops.”Footnote ³² The rhetoric of loyalty was reciprocally founded on the mystique of suzerainty, the vassal depending on the sovereign’s appointment for his legitimacy, the sovereign on the vassal’s allegiance for maintaining a semblance of empire. Vassal states such as Silla and Bohai observed the conventions of tributary allegiance to the continental power as an essentially voluntary arrangement based on mutual interest.Footnote ³³

As the gulf between rhetoric and reality widened, the credibility of the device frayed. The kingdom of Nanzhao rescinded its vassal status outright. Military governors openly voiced their disdain for the imperial government’s shortcomings. What triggered such candor? In addition to disloyalty on the part of subjects, the Song historians decried government misrule as one of the causes for the collapse of the Tang. The New History of the Tang (Xin Tang shu 新唐書) cites the case of the military governor of Shannan East, Liu Jurong 劉巨容 (826–89), who allowed the rebel army of Huang Chao to escape after repelling their invasion of his territory, and then justified his inaction with the government’s fickleness and ingratitude (see under “The Year Jihai” in Chapter 8). We can’t verify the exactitude of this account, but it allows the author to make a significant point: A military governor’s allegiance was contingent on the emperor doing right by him. As Liu Jurong reportedly saw it, allegiance was not deferential but contractual.Footnote ³⁴ Once the tacit compact of allegiance was broken, a minister was freed from his obligation. Emperors who were responsible for misrule and heedless to remonstrance lost their mandate and could legitimately be deposed. The Five Dynasties scholar Xu Xuan 徐鉉 (917–92) made the reciprocity of the political engagement explicit: The exercise of power was the shared duty of ruler and minister, he wrote; when the ruler failed in mutuality, he forfeited the allegiance of his ministers.Footnote ³⁵ Xu Xuan, and officials such as Liu Jurong, had a case then. The Confucian sociopolitical ethic held that the responsibility for a breakdown of the natural order of hierarchy and obedience ultimately lay with the ruler.Footnote ³⁶ To secure a subject’s deference, the sovereign needed to maintain that order through his own actions.

The Old History of the Tang (Jiu Tang shu 舊唐書) describes the court of Emperor Xizong as “engaged in internal plotting while the empire broke apart.” Lesser individuals rose to the top, it says, while capable civil servants withdrew from public life to nurse their grievances. When Huang Chao revolted, such disheartened individuals vented their frustration by publicly subscribing to the rebels’ censure of the government.Footnote ³⁷ When the Old History was completed in 945, the director of National History and chief minister Liu Xu 劉昫 (887–946) had no way of knowing whether the fragmentation of the Chinese realm would be permanent. The Song vision of history and statecraft, by contrast, was forged by that dynasty’s protracted struggle to achieve and maintain unity. Neo-Confucian scholar-statesmen such as Ouyang Xiu 歐陽修 (1007–72) and Sima Guang 司馬光 (1019–86) accentuated the primacy of imperial rule, the linear transmission of political legitimacy, and the importance of unconditional loyalty on the part of ministers. With hindsight, the system of regional commands and military governorships appeared to be at the root of political instability and ultimately led to the loss of territorial integrity. Ouyang Xiu prefaced his “Chart of Military Governors” 方鎮表 in the New History with an account of the institution’s origin as a border defense system and its subsequent abuse by increasingly autonomous governors. Ouyang warned his contemporaries: “We can only take heed!”Footnote ³⁸

The Historical Record

But readers of official histories should also beware. As recently as the late twentieth century, historians of the Tang depended almost exclusively on the records transmitted by official historiography. While the kind of broadly sourced history writing that Wang Mingsheng 王鳴盛 (1722–98) advocated (see the Preface) was feasible for later periods of Chinese history that produced abundant archival materials, it could not be readily envisaged for the Tang. Denis Twitchett, author of The Writing of Official History Under the Tang, alerted his readers to the limitations of the official record, which was designed to convey a carefully curated picture of the recent past to posterity.Footnote ³⁹ The compilation of official histories was a political act that served the primary purpose of drawing lessons from past errors and achievements to guide contemporary policy choices. The high political stakes of containing power struggles and averting a repetition of civil war and regime failure focused the Song historians’ attention on identifying the culprits of the Tang debacle. Another limitation was that topics such as military and regional affairs, court factionalism, or eunuch networks that we now regard as important for understanding the period received only scant or partisan treatment. Moreover, in the process of distilling a “standard history” 正史 from the mass of court archives and administrative documents handed down by the Tang Bureau of Historiography, the Song official historians discarded their intermediary sources and those that conveyed diverging accounts (see Appendix D). Other records were lost or destroyed or became obliterated during the chaotic final decades of the dynasty. As a result, later historians were left with few early sources and virtually no archival materials. The prospect of unearthing substantial new evidence seemed remote.

In the past few decades this situation began to change. A growing body of excavated tomb inscriptions, the development of manuscript studies, and a new recognition of the historical value of informal writings including private memoirs, local records, religious texts, and narrative literature added substantial contemporary materials to the available documentation. Anecdotal histories, “tales of the abnormal” 志怪, and “transmitted singularities” 傳奇 all enjoyed wide popularity in Tang and Song times. The interest of this type of material as historical records has been underlined, among others, by Robert Campany and Glen Dudbridge for the pre-Tang and Tang periods, and for the Song, with special emphasis on Hong Mai’s 洪邁 (1123–1202) Record of the Listener (Yijian zhi 夷堅志), by Alister David Inglis.Footnote ⁴⁰ Sources such as these offer glimpses of social strata and sectors of activity including women, clergy, artisans, merchants, soldiers, and farmers that official historians tended to neglect. To be sure, informal sources and private histories can have agendas and need to be evaluated with the same precaution as the products of official historiography. Alongside official records, edicts, and correspondences, the present book weaves their testimonies into a narrative of events that reflects to the greatest possible extent contemporary perspectives on the declining years of the Tang.

Among the informal records of the Tang–Song transition, the collection of anecdotes Trifles from Beimeng (Beimeng suoyan 北夢瑣言) by Sun Guangxian 孫光憲 (896–968) stands out. The work has been closely studied by Fang Rui 房銳. A native of Sichuan, Sun served as an executive officer under the Former Shu kingdom (907–25). During the Later Tang interregnum in Shu (926–33), he traveled downstream to the land of Chu 楚 (see also under “The Middle Yangzi” in Chapter 8), where he served successive rulers of the Five Dynasties state of Jingnan 荊南 (907–64) as deputy censor-in-chief and trusted counselor.Footnote ⁴¹ The name “Beimeng” in the title refers to the domain of the ancient kings of Chu. Sun Guangxian lived into the early years of the Northern Song as a highly regarded scholar and writer.

Du Guangting 杜光庭 (850–933), the Daoist hierarch and court divine under Emperor Xizong (r. 873–88), served as a senior official and confidant to the rulers of Shu after the fall of the Tang. An astute observer of contemporary society and a prolific writer, Du left a rich testimony of his personal experience straddling the Tang–Five Dynasties divide. A dedicated priest, Du composed liturgical prayers and devotional narratives for diverse congregations ranging from the imperial family to village farmers. His stories are akin to the popular “tales of the abnormal” and “transmitted singularities.” Like the latter, they have been superficially assimilated with imaginative literature. In fact, Du was fundamentally concerned with reality, comprising both tangible and mythical dimensions. His aim was to ground the Daoist worldview in evidence drawn from religious experience that bridged the visible and invisible worlds. In his Divine Manifestations of Daoism (Daojiao lingyan ji 道敎靈驗記), Du Guangting exemplified Gao Pian as an eminent lay devotee and illustrated the religious rewards that accrued to him.

A class of writings of particular interest for this book emanated from the headquarters of military governors.Footnote ⁴² Poetry was a convivial form of entertainment at the heart of the social and intellectual life of Gao Pian’s yamen. It was also the medium of choice for giving expression to an individual’s inner life.Footnote ⁴³ Personal but not private, Gao’s poems were intended to circulate.Footnote ⁴⁴ Some poems record Gao’s responses to events, others served as instruments of public communication, others again have circumstantial documentary value. A complete translation of the remains of his once copious output is found in the compilation “Poems of Gao Pian (PGP)” at the end of this book.

Three categories of prose documents produced by Gao’s headquarters deserve our attention. First, a non-negligeable amount of Gao Pian’s own writings and recorded pronouncements survives. Gao’s correspondence, reminiscences, and commemorations add an immediacy to the historical record that is absent from official accounts. Gao’s “Response to Yunnan,” a letter to the king of Nanzhao, is a rare illustration of Tang diplomacy in action (see under “A Letter to Shilong” in Chapter 6); his narration of the battle of Hanoi (under “The Siege of Jiaozhi” in Chapter 3) plunges the reader into the thick of warfare on the frontiers of the empire. Moreover, Gao Pian was intellectually invested in issues of governance and given to explicating his methods of leadership in court memorials and instructions to subordinates. Letting the actors’ own voices be heard above the clamor of controversy, end-of-regime recriminations, and the rationalizations of later historians lends authenticity to their portrait and lets us envisage the forces shaping history around them.

Second, the productions by Gao’s literati retainers form a substantial corpus (see also under “Patterns of Patronage” in Chapter 1). A noteworthy example is the Daoist author Pei Xing’s 裴鉶 (ca. 825–ca. 880) narrative collection Transmitted Singularities (Chuanqi 傳奇), a title that would later serve to label a genre of writings categorized as “proto-fiction.” In reality, Pei Xing scrutinized the world with the eyes of an ethnographer. One of his tales can be read in conjunction with Pei’s stele inscription “Path of Heavenly Might,” a unique witness account of Gao’s hydraulic engineering feat in Annan where the author used his literary talent to lend the general’s engineering skills a prodigious aura (see Chapter 4 and Appendix A). Other literati associates left similarly absorbing testimonies of life in military headquarters. Some commissioned writings were undeniably biased. Gao Pian was a consummate communicator who employed skilled writers to beat his drum using the media at their disposal: poetry, memoirs, tracts, stele inscriptions. Like court writers everywhere, Gao’s literati retainers laced their personal communications with flattery. Selected eulogies from Poems Commemorating Virtue by Gao Pian’s chief secretary Ch’oe Ch’iwŏn (see under “Tilling with My Brush” in Chapter 9) feature as chapter epigraphs in this book. Ch’oe originally presented these admiring vignettes to Gao as part of his recruitment dossier, their allusive fabric showing off the young Korean scholar’s mastery of Tang China’s literary culture.Footnote ⁴⁵

Third, the scribes, secretaries, and chroniclers in Gao’s sprawling administration generated a steady flow of court memorials, reports, military correspondence, letters of appointment, and written instructions to officials. These executive missives naturally obeyed court protocol and administrative convention, but their purpose was to implement policy, not to curry favor with the patron or score partisan points. Although rescript writers and secretarial redactors legitimately included such proxy writings in their personal literary collections, they unmistakably articulated the master’s voice.Footnote ⁴⁶ As operational levers of government, they constitute for us quintessentially primary sources that document history in the making. Ch’oe Ch’iwŏn’s collection Tilling with My Brush at Cassia Grove (Guiyuan bigeng ji 桂苑筆耕集) is a voluminous archive of such documents covering the crucial years of the court-in-exile, the emperor’s struggle to return to power, and Gao’s simultaneous establishment as the overlord of Huainan. Cited for short as the Cassia Grove Collection (Guiyuan ji 桂苑集), it occupies a central place in this book.

The General’s Scribe

In the year 866 ce, Ch’oe Ch’iwŏn, the author of this work, embarked in the kingdom of Silla on a vessel bound for Tang China. Quick-witted and studious, the eleven-year-old boy had been selected to pursue the most advanced education to be had in East Asia. His native town was Gyeongju 慶州, the capital of Silla on the southeastern coast of the Korean peninsula and terrestrial terminus of a branch of the silk road. Merchants, diplomats, and pilgrims traveling to the Tang capitals from Gyeongju rounded the Korean peninsula, then crossed the Sea of Bohai to the northern the tip of Shandong, landed at Dengzhou 登州, and continued overland via Qingzhou 青州 to Luoyang or Chang’an.Footnote ⁴⁷

Ch’iwŏn later recorded the solemn words with which his father, Ch’oe Kyŏnil 崔肩逸, a ranking Silla aristocrat, saw the boy off: “If you do not pass the imperial examination within ten years, do not call yourself my son, nor shall I acknowledge having a son. Go, be diligent about it; never falter in exertion!”Footnote ⁴⁸ Adding to the intellectual and emotional challenges of his task, the youth landed in a country at war with its neighbors and roiled by internal conflict. Overcoming solitude and homesickness, Ch’iwŏn determinedly exceeded his father’s expectations in joining the ranks of close to a hundred scholars from Silla who graduated in China in the course of the ninth century.Footnote ⁴⁹ Even though Ch’oe had won the coveted presented scholar or jinshi 進士 degree in record time, however, he struggled for several years to find an inroad into the civil service of the faltering Tang. He finally succeeded in securing employment in the bustling headquarters of one of the most powerful men in China, the field marshal and military governor Gao Pian. In 885, after nearly twenty years abroad, Ch’oe Ch’iwŏn would return to Silla, escaping the civil war engulfing the Tang empire. In his baggage were numerous scrolls of poetry as well as copies of official documents he had redacted on Gao Pian’s behalf at “Cassia Grove,” a poetic name for Huainan’s capital Yangzhou. Still aged only thirty and looking for employment in his homeland, Ch’oe presented the king of Silla with a selection of his writings composed in China. They included a substantial compendium of poetry and the Cassia Grove Collection in twenty scrolls, Ch’oe’s edition of Gao Pian’s essential military and administrative communications from the years of the Tang’s descent into chaos.

Like the Japanese pilgrim Ennin 圓仁 (793–864), who experienced and chronicled the Huichang Proscription of Buddhism in China a generation earlier, Ch’oe Ch’iwŏn had witnessed a momentous event in the history of China.Footnote ⁵⁰ The monk Ennin, caught in the turbulence of the repression, and the scribe Ch’oe, writing from his perch at the commander-in-chief’s headquarters, each observed China with the eyes of an alert and informed alien. Safely transported beyond the Tang’s shores, Ch’oe Ch’iwŏn’s testimonial came through the empire’s collapse intact. The Cassia Grove Collection earned the already acclaimed young poet fame in Silla as a master of Chinese prose. The book was reverently preserved in that land and eventually found its way back to China (see Appendix B).