An Lushan Rebellion

The An Lushan Rebellion (755–763), also called the An–Shi Rebellion, was a military uprising and civil war against the Tang dynasty in China. It began under An Lushan, a powerful northeastern frontier commander, and continued under An Qingxu, Shi Siming, and Shi Chaoyi. The rebels established a rival imperial state, Yan, and fought Tang forces for control of northern China, including the capitals of Luoyang and Chang’an. (en.wikipedia.org)

The rebellion was a major turning point in medieval Chinese history. Although the Tang defeated the rebel state, the conflict weakened central authority, strengthened provincial military commanders, devastated important agricultural regions, and altered the empire’s relations with neighboring powers. Its effects extended into taxation, population movement, and literary expression. The dynasty nevertheless survived until 907: the rebellion transformed Tang rule rather than immediately ending it. Understanding the conflict requires distinguishing its military course from its longer-term consequences and treating claims about its death toll with particular caution. (academic.oup.com)

Frontier Armies and the Tang Political Order

The early Tang dynasty relied heavily on the fubing system, whose territorially organized soldiers combined farming and training with rotational guard duty and campaign service. This was not simply an untrained peasant militia: military obligations extended over many years. Mobilized troops served under appointed expeditionary commanders, then normally returned home. From the late seventh century, prolonged confrontation with Tibet and the revived Eastern Türk empire made this arrangement increasingly inadequate. Armies could neither defeat their opponents decisively nor safely withdraw, and temporary encampments developed into permanent frontier garrisons. An edict of June 12, 737, accelerated the transition by directing governors to recruit long-service soldiers, known as jian’er, from existing troops and migrant households. Inducements included enhanced compensation, exemption from taxes and corvée, and land and housing for accompanying dependents. Recruitment did not eliminate emergency conscription. (cambridge.org)

The resulting regional commands were headed principally by military commissioners, or jiedushi, appointed by the emperor. Their responsibilities extended beyond battlefield leadership to provisioning, military agriculture, and the administration of resources supporting their armies. Soldiers received grain and silk drawn from frontier and interior tax revenues; commissioners could conduct punitive expeditions and supplement their standing forces with temporary levies. Long service produced cohesive formations familiar with local terrain, but also placed much of the practical organization of military life under commanders stationed far from court. These were still imperial commissions, not yet the hereditary provincial regimes associated with parts of the postwar empire. (assets.cambridge.org)

By 742, ten major frontier commands contained approximately 487,000 troops on the recorded establishments. The northeastern concentration comprised Fanyang, headquartered at Youzhou in present-day Beijing, with 91,400 men; Pinglu, centered on Yingzhou in present-day western Liaoning, with 37,500; and Hedong, based at Taiyuan in Shanxi, with 55,000. Shuofang, around the northwestern bend of the Yellow River, connected the northern frontier with the approaches to the capital region. Farther west, Hexi guarded the Gansu corridor, Longyou faced Tibet across the Qinghai–Gansu frontier, and Anxi and Beiting defended the Tarim Basin and the territory north of the Tian Shan. Jiannan covered the Sichuan frontier, while Lingnan formed the southern command. These establishment figures indicate the distribution of military resources, rather than an exact count of troops immediately available for deployment. The imperial heartland retained guards and recruitment resources, but lacked a comparable concentration of experienced field armies: the initial defense of Luoyang in 755 depended heavily on hastily raised troops, while subsequent defense required frontier reinforcements. (scribd.com)

Recruitment crossed both regional and ethnic boundaries. Most frontier soldiers and officers were Chinese men of relatively modest social background, including peasants and convicts, but northern and northwestern formations also incorporated Khitans, Mohe, Tiele, and Türks. Some entered for pay, others as refugees, and still others as captives. Mounted archery, tactical skill, and demonstrated military ability offered routes to promotion. Commanders themselves reflected this diversity: Gao Xianzhi descended from Goguryeo aristocrats, while Geshu Han had a Türgesh father and a Khotanese mother. His service against An Lushan illustrates why ancestry cannot be treated as a straightforward predictor of allegiance. The organization and patronage of particular armies mattered alongside imperial appointments and individual political choices. (assets.cambridge.org)

The northeastern commands stood above the open North China Plain, with routes south through Hebei toward Yellow River crossings and Luoyang, the eastern capital. West of Luoyang, the river corridor and constricted mountain approaches led toward Tong Pass, which guarded entry into the Wei River basin and Chang’an, near present-day Xi’an. Supply routes linked this northern political and military geography to the Huai and Yangtze basins through the Grand Canal network, particularly the Bian Canal. Grain moving north could be transferred toward the capitals through waterways, granaries, and overland bypasses. In 734, chief minister Pei Yaoqing reorganized this transport system around transfer depots, including Heyin near the junction of the canal and Yellow River, and arrangements for bypassing the dangerous Sanmen river passage. Control of crossings and canal junctions therefore affected both military movement and access to southern provisions. (en.wikipedia.org)

An Lushan’s Rise and the Decision to Rebel

An Lushan emerged from the northeastern frontier society in which the Tang dynasty recruited soldiers and intermediaries familiar with several languages and political traditions. The Tang histories describe him as having a Turkic mother and Sogdian paternal ancestry, although their account of his birth contains legendary elements. His family’s connections crossed the boundary between Turkic rule and Tang service; this background was a route into imperial employment, not in itself an explanation of rebellion. By the 730s he was serving under the northeastern commander Zhang Shougui. A failed attack in 736 nearly brought his execution, but Emperor Xuanzong pardoned him, allowing his military career to continue. (iranicaonline.org)

His decisive promotions concentrated three frontier commands in one officer’s hands: Pinglu in 742, Fanyang in 744, and Hedong in 751. Pinglu was headquartered around Yingzhou, near modern Chaoyang in Liaoning; Fanyang around modern Beijing; and Hedong at Taiyuan in Shanxi. These appointments gave him authority across adjoining northeastern and northern military regions rather than merely command of a single field army. His advancement also depended on chief minister Li Linfu, who favored generals without the scholarly credentials through which successful commanders might compete for senior civil office. The concentration of standing armies provided the structural means for rebellion, but had initially been authorized to serve imperial frontier policy. (en.wikipedia.org)

Repeated visits to Chang’an, near modern Xi’an, made An a personal favorite of Xuanzong and his consort Yang Guifei. Court accounts describe his comic performances and his adoption as Yang’s son, including a mock infant-bathing ceremony in 751. Such anecdotes record his exceptional access to the palace, though they should not be treated as transparent evidence of his private intentions. Imperial favor had concrete benefits: Xuanzong granted him the title Prince of Dongping in 750 and provided an elaborate capital residence in 751. His standing depended on the emperor’s protection as well as the consort’s patronage; military failure against the Khitan in 751 did not end his career. (en.wikipedia.org)

After Li Linfu’s death, Yang Guozhong, Yang Guifei’s cousin and the new chief minister, became An’s principal court rival. Yang repeatedly alleged that An intended to rebel. In 754 he argued that a summons to Chang’an would expose An’s disloyalty because the general would refuse to come; An instead appeared and renewed the emperor’s confidence. Nevertheless, Yang opposed his proposed elevation to the chancellorship. In 755 the conflict reached An’s associates in the capital: Yang had Li Chao and others arrested and secretly executed. An’s son An Qingzong reported these actions to his father, while An increasingly resisted invitations and summons to court. The struggle thus threatened both his political connections and the security of his military establishment. (iranicaonline.org)

The immediate circumstances of the uprising were this escalating contest over access, appointments, and personal security, rather than the existence of frontier armies alone. Yang could attack An’s supporters at court but had not dismantled his northeastern power base; An could mobilize seasoned troops but could no longer rely securely on court protection. On 16 December 755 he launched the uprising at Fanyang, claiming to possess a secret imperial command to remove Yang Guozhong. This proclaimed purpose represented armed intervention as service to Xuanzong against a harmful minister, not initially as an openly announced campaign to replace the Tang dynasty. (universalium.en-academic.com)

The Rebel Advance and the Fall of the Capitals, 755–756

On 16 December 755, An Lushan launched his uprising at Fanyang, near modern Beijing, claiming imperial authorization to punish the chief minister Yang Guozhong. His army advanced south across the North China Plain, where many local administrations lacked forces capable of resisting experienced frontier troops. The speed of the offensive prevented the court from assembling an effective defense of the eastern capital. On 8 January 756, the rebels crossed the Yellow River and occupied Chenliu and Xingyang, approaching Luoyang from the east. Xuanzong had dispatched Feng Changqing to organize its defense, but Feng’s hastily assembled army suffered repeated defeats against the advancing rebels. (en.wikipedia.org)

Luoyang fell on 18 January 756, little more than a month after the uprising began. Feng withdrew westward and joined Gao Xianzhi, who commanded a second defensive force in Shan Commandery, around modern Sanmenxia. Rather than risk encirclement, the two commanders retreated to Tong Pass, the fortified gateway to the metropolitan region surrounding Chang’an, modern Xi’an. An Lushan did not immediately pursue them into the western heartland. On 5 February, the lunar New Year, he proclaimed himself emperor of Great Yan at Luoyang. The appointment of Zhang Tongru and the surrendered Tang official Daxi Xun as chief ministers established a rival imperial government, replacing the initial claim that his campaign merely sought to remove Yang Guozhong. (en.wikipedia.org)

Tong Pass checked the rebel advance for several months. Its defenders could exploit a constricted approach between the river and mountainous terrain rather than meet the rebels on the open plain. The court nevertheless removed the commanders who had established this defensive position: after accusations by the eunuch supervisor Bian Lingcheng, Xuanzong ordered Feng Changqing and Gao Xianzhi executed. Their replacement, Geshu Han, was an experienced commander from the northwestern frontier, recalled despite serious illness. Reinforcements drawn from the western commands strengthened the position, but Geshu’s dependence on subordinate officers complicated the army’s command arrangements. (en.wikipedia.org)

Geshu favored holding the pass while loyalist armies operating north of the Yellow River threatened the rebels’ rear. Xuanzong and Yang Guozhong instead pressed for an offensive, encouraged by reports that the opposing force under Cui Qianyou was weak. Repeated imperial orders compelled Geshu to advance eastward. In July 756, Cui defeated the Tang army in the confined terrain beyond the pass; the survivors could not restore an effective defense. Geshu’s subordinate Huoba Guiren seized him and surrendered him to the rebels. The loss of Tong Pass thus opened the road to Chang’an through the destruction of its covering army, rather than through a prolonged siege of the capital itself. (en.wikipedia.org)

On 14 July, Xuanzong abandoned Chang’an with members of the imperial family, senior officials, palace attendants, and an armed escort, heading toward Sichuan. At Mawei Station the following day, the escort mutinied. Soldiers blamed Yang Guozhong for the crisis and killed him; their commander, Chen Xuanli, then demanded the death of Yang Guifei, the emperor’s consort and Yang Guozhong’s relative. Xuanzong consented under military pressure, and she was put to death. The surviving imperial party continued toward Sichuan, while rebel troops subsequently occupied the abandoned western capital. The killing of the minister and the coerced execution of the consort were distinct acts within the same breakdown of authority over the imperial escort. (en.wikipedia.org)

Tang Resistance and the Recovery of the Capitals, 756–757

In August 756, the crown prince Li Heng assumed the throne at Lingwu, in present-day Ningxia, as Emperor Suzong. His position beside the Shuofang frontier army provided a military base outside the occupied capitals. Xuanzong subsequently recognized the succession from Sichuan and dispatched senior ministers with his formal abdication message. Suzong’s initial entourage reportedly numbered only about thirty civil and military officials; rebuilding the Tang government therefore required assembling both an administration and an army. Incoming ministers, including Fang Guan, helped restore central decision-making, while the court’s relocation to Fengxiang, west of Chang’an, in spring 757 brought it closer to the principal counteroffensive. (en.wikipedia.org)

Guo Ziyi and Li Guangbi supplied experienced leadership and troops from the northern commands. Before Suzong’s accession, their campaign in Hebei had recovered territory around Changshan and threatened communications between the rebel capital at Luoyang and the northeastern base at Fanyang. The collapse of Tong Pass forced them to withdraw westward, but Suzong summoned them to his court rather than abandoning their forces. Guo subsequently became central to the campaign against Chang’an; Li returned to Hedong and successfully defended Taiyuan against Shi Siming’s siege in spring 757. Resistance also arose outside these field armies: Yan Gaoqing, the civilian governor of Changshan, organized loyalist forces against rebel communications before his capture and death. Rebel possession of the capitals thus did not amount to uncontested control of the surrounding provinces. (en.wikipedia.org)

In eastern Henan, Zhang Xun mobilized locally recruited troops and defended Yongqiu, near present-day Qi County, during 756. Repeated attacks failed to eliminate this loyalist stronghold, although the loss of neighboring positions eventually compelled him to withdraw. Early in 757 he joined Xu Yuan at Suiyang, near present-day Shangqiu. These defenses obstructed the rebels’ southeastern advance along the Bian Canal, part of the Grand Canal network. The strategic concern was the Jianghuai region—the Huai valley and lower Yangtze basin—whose revenues and grain resources were increasingly important to sustaining loyalist operations. Holding towns along this approach delayed rebel penetration into the southern supply region; it did not mean that the entire transport route remained open. (www11.ihp.sinica.edu.tw)

At Suiyang, Zhang directed combat while Xu managed supplies and other administrative responsibilities. The traditional account gives their initial combined force as approximately 6,800 men. Yin Ziqi’s rebel army subjected the city to successive attacks through much of 757, while the defenders used sorties and counterattacks to interrupt the siege. Food shortages became acute, and attempts to secure adequate outside relief failed. The city fell late in the year, and Zhang and Xu both died in rebel captivity. Although a rebel tactical victory, the prolonged siege absorbed forces and time that could otherwise have supported an advance toward Jianghuai. (en.wikipedia.org)

The Yan leadership meanwhile underwent a violent succession. On 29 January 757, An Lushan was killed in a conspiracy involving his son An Qingxu, the senior adviser Yan Zhuang, and the attendant Li Zhu’er. An Qingxu succeeded as emperor of Yan, with Yan Zhuang exercising substantial influence over the new regime. This change did not immediately dissolve the rebel military organization: Yan forces continued to contest the approaches to Chang’an, and the Tang still needed a major field campaign to recover either capital. (citeseerx.ist.psu.edu)

In the closing months of 757, Suzong’s son Li Chu, Prince of Guangping, held overall command of the recovery campaign, with Guo Ziyi directing field operations. Shuofang troops and contingents recalled from the western commands fought alongside Uyghur cavalry. Near Xiangji Temple, south of Chang’an, Li Siye’s infantry helped stabilize the Tang line under rebel attack, while cavalry action contributed to the defeat of the Yan army. Rebel commanders abandoned Chang’an, allowing loyalist forces to reoccupy it. The victors then advanced east through the Tong Pass corridor, defeated another rebel force near Shan Commandery, and recovered Luoyang. An Qingxu withdrew to Yecheng with surviving troops, leaving the rival Yan state intact beyond the recovered capitals. (en.wikipedia.org)

Renewed War and the Collapse of Yan, 758–763

In late 758, Tang armies converged on An Qingxu’s remaining stronghold at Yecheng, in the Xiangzhou district near present-day Anyang. The operation brought together forces under Guo Ziyi, Li Guangbi, and several other regional commanders, but Emperor Suzong appointed no supreme commander: the court was reluctant to subordinate either of its two leading generals to the other. Instead, the eunuch Yu Chao’en supervised the campaign as an imperial military observer. After defeating An’s forces outside the city, the loyalists constructed siege works and diverted the Zhang River to flood its defenses. Yet the besiegers themselves became vulnerable to shortages. Shi Siming’s relief forces raided foraging parties and disrupted supply convoys. In April 759, a violent windstorm interrupted the confrontation outside Yecheng; both sides scattered, but the Tang armies withdrew toward their separate bases, abandoning supplies and lifting the siege. (zh.wikisource.org)

Shi Siming’s intervention followed a brief and unstable accommodation with the imperial court. After the rebel loss of the capitals in 757, he had offered his submission while retaining control of Fanyang, in the region of modern Beijing. Suzong recognized him as its military governor, but mutual distrust undermined the arrangement during 758. Shi executed the imperial intermediary Wu Cheng’en after discovering a plot against him and subsequently resumed rebellion. An Qingxu sought his assistance at Yecheng, even offering to surrender the Yan throne. Once the siege had collapsed, Shi induced An to visit his camp under the promise of an alliance, then executed him and absorbed his troops and territory. Returning to Fanyang, Shi proclaimed himself emperor of Yan in 759: the rebel state survived, but leadership passed from An Lushan’s family to a rival military household. (zh.wikisource.org)

In autumn 759, Shi advanced south and recaptured Luoyang. Li Guangbi withdrew from the city to the fortified crossings at Heyang, on the Yellow River, where interconnected positions protected movement between the river’s banks. His defense checked Shi’s advance despite the loss of the eastern capital. This containment was not an uninterrupted stalemate, however. Early in 761, the court pressed Li to attack Luoyang despite his objections; disagreement with Pugu Huai’en over deployment contributed to a loyalist defeat at Mount Mang, after which Heyang also fell. Shi then attempted to advance west through Shan Prefecture toward Chang’an. During this campaign in spring 761, officers associated with his eldest son, Shi Chaoyi, seized and killed him. Shi Chaoyi assumed the imperial title, but many senior rebel commanders, veterans of An Lushan’s establishment, did not reliably obey his summons. (chinese-classics.org)

The decisive loyalist offensive came in autumn 762, under Emperor Daizong, who had succeeded Suzong that year. Tang forces operating with Uyghur reinforcements defeated Shi Chaoyi near Luoyang in November and recovered the city. Pugu Huai’en played a leading role in the victory and subsequent pursuit, while Li Guangbi’s operations formed part of the wider pressure on the remaining rebel positions. Shi Chaoyi escaped eastward and then sought refuge among the northern commands. The loss of Luoyang was therefore not immediately the destruction of his army, but it deprived him of a principal base and accelerated the disintegration of his commanders’ allegiance. (dergipark.org.tr)

During the final pursuit, rebel commanders including Xue Song, Zhang Zhongzhi—subsequently named Li Baochen—Tian Chengsi, and Li Huaixian submitted to Tang authority rather than continue supporting Shi Chaoyi. Their defections closed potential refuges and transferred strategic districts without requiring the loyalists to capture every stronghold by assault. Li Huaixian’s forces pursued Shi Chaoyi after denying him sanctuary in the Fanyang region. Early in 763, with escape routes blocked, Shi hanged himself, ending the succession of rulers claiming the Yan imperial throne. Several surrendering commanders retained their troops and territorial commands under imperial appointments; the destruction of the rival state consequently did not entail the dissolution of all its military establishments. (en.wikipedia.org)

Uyghur Intervention and the Changing Inner Asian Balance

Emperor Suzong secured assistance from the Uyghur Khaganate through an alliance reinforced by military rewards and marriage diplomacy. In 757, the khagan Moyanchur sent cavalry under his son, known in Chinese accounts by the title yabghu, to join the Tang counteroffensive. Their mobility proved important near Xiangji Temple, southwest of Chang’an: Uyghur troops helped prevent an attack on the loyalist rear and cut off the rebels’ retreat toward the city. They subsequently assisted in recovering Luoyang. In 758, Suzong married his daughter Princess Ningguo to Moyanchur, making the military partnership a direct relationship between the ruling houses. Another marriage connected Moyanchur’s successor, Bögü, to the daughter of the Tang commander Pugu Huai’en, who became an important intermediary between the two armies. (dergipark.org.tr)

The alliance did not guarantee Uyghur support irrespective of circumstances. After Xuanzong and Suzong died in 762, Shi Chaoyi approached Bögü with the claim that the Tang lacked a ruler and offered access to its wealth. Emperor Daizong’s envoy Liu Qingtan countered this appeal, and negotiations involving Pugu Huai’en redirected the khagan’s expedition toward assistance to the Tang. A court observer reported approximately 4,000 fighting men, accompanied by a much larger body of dependants and horses; this was a cavalry contingent rather than an independent army capable of occupying northern China. Bögü personally participated in the campaign that recovered Luoyang in November 762, while Uyghur troops also joined the subsequent pursuit of the remaining rebels. (depts.washington.edu)

Victory carried costs for the populations of the recovered cities. In 757, the future Daizong persuaded the Uyghurs to postpone plundering Chang’an until Luoyang had been taken; the resulting sack of Luoyang lasted three days. In 762, renewed pillaging included the burning of temple towers where civilians had sought refuge. The dynastic narrative also identifies Tang troops as participants in the wider devastation, rather than attributing all violence to the foreign contingent. These episodes illustrate the court’s limited capacity to restrain allied forces whose assistance it needed. (dergipark.org.tr)

Silk rewards and the horse–silk trade gave the military relationship a continuing economic dimension. Exchanges expanded from the Qianyuan period, 758–760, linking Tang demand for cavalry mounts with Uyghur access to Chinese textiles. Tang histories record a price of forty bolts of silk per horse, but this should not be treated as a uniform rate governing every transaction: surviving accounts combine purchases, unpaid obligations, tribute exchanges, marriage gifts, and military rewards. Tan Mei Ah’s study of the trade discusses evidence for fluctuating rates and cautions against explaining later silk shortages solely through payments to the Uyghurs. Horses remained a necessary military resource even when their purchase imposed substantial obligations on the court. (cuhk.edu.hk)

The transfer of frontier troops to the civil war also opened opportunities for the Tibetan Empire, particularly in Qinghai and Gansu. Reduced garrisons weakened the defence of the corridors connecting the imperial heartland with its western protectorates. Territorial losses nevertheless accumulated over decades rather than following an instantaneous abandonment of Central Asia in 755 or 763. Liangzhou fell in 764, followed by Ganzhou and Suzhou in 766; Tang positions farther west survived longer, and Beiting’s principal prefecture was not lost until 790. Tibetan expansion progressively interrupted the communications and reinforcement routes on which the remaining garrisons depended. (cambridge.org)

The Tibetan occupation of Chang’an occurred in November 763, after Shi Chaoyi’s death had ended the rebellion earlier that year. Daizong fled the capital, which Tibetan forces held briefly—approximately fifteen days—before withdrawing. This was a separate invasion, not another rebel capture of the city. Its immediate background included renewed Tibetan offensive operations in 762–763 and a Tang promise of silk payments that remained unfulfilled; its military opportunity rested on the weakened western defences created by years of internal war. The withdrawal from Chang’an did not reverse Tibet’s continuing gains along the northwestern frontier. (cir.nii.ac.jp)

Provincial Power and the Reconfiguration of Imperial Rule

The settlement of 763 restored imperial allegiance without dismantling the military establishments that had sustained the rebellion. Emperor Daizong accepted the submission of former Yan commanders who still controlled troops and territory; at the recommendation of the loyalist general Pugu Huai’en, they received appointments as military governors and retained their commands. Removing them risked renewed resistance rather than an orderly restoration of civilian administration. The arrangement separated the destruction of the rival Yan dynasty from the reconquest and disarmament of every district it had governed: commanders could abandon Shi Chaoyi while preserving their own military organizations under Tang commissions. (en.wikipedia.org)

Three appointments established the principal enduring autonomous commands in Hebei. Tian Chengsi received Weibo, initially comprising five prefectures and centered on Wei Prefecture, in the region of modern Handan. Li Baochen governed Chengde, centered in the region of modern Shijiazhuang, while Li Huaixian retained Youzhou-Lulong, headquartered in present-day Beijing. These became known as the “three garrisons of Hebei.” Their origins lay in the conversion of former rebel commands into recognized provincial administrations, not in the creation of legally independent kingdoms. Imperial appointment supplied an official place within the Tang dynasty, while possession of the troops supplied the practical means to resist intervention. (en.wikipedia.org)

Autonomy operated through control of recruitment, revenue, and succession. The strongest northeastern governors retained locally collected taxes to support their armies and sought to transmit office within their families rather than accept unrestricted imperial replacement. Provincial soldiers were nevertheless political actors in their own right, not simply a governor’s hereditary property. Li Huaixian’s murder by his officers in 768 illustrates how command could change through military violence. Nor did the court concede hereditary succession as an unquestioned rule: in 781 Emperor Dezong refused to confirm Li Baochen’s son Li Weiyue as governor of Chengde, precipitating renewed armed conflict. Such disputes concerned who could nominate and install an imperial officeholder, even when the court could not enforce its decision locally. (researchgate.net)

This northeastern pattern did not describe the entire empire. Powerful loyalist commands along the Taihang Mountains and across Henan confronted the autonomous provinces; those in Henan also protected the Bian Canal, a crucial artery of the Grand Canal system carrying southern revenues toward the court. Southern provinces generally maintained smaller garrisons and more often had civilian governors. Their remittances helped sustain heavily armed commands north and west of Chang’an against Tibet. Provincial military government therefore encompassed markedly different relationships with the throne: revenue-supplying regions, strategically subsidized frontier forces, and commands that withheld resources and resisted appointments. (researchgate.net)

The rebellion also expanded military provincial government beyond the former rebel territories. Between 755 and 763, the number of military provinces increased from ten frontier commands to more than forty, most now situated inside the empire. Most new commands were loyalist creations intended to mobilize local manpower and supplies. Yet loyalist origins did not guarantee obedience. The court continued to appoint both military professionals and senior civil officials to major commands, while increasingly relying on enlarged palace armies as a counterweight to provincial forces. (cambridge.org)

Eunuch supervision provided another channel of imperial intervention, distinct from the appointment of provincial commanders. During the war, Yu Chao’en exercised de facto operational control over nine Tang contingents defeated near Xiangzhou in 759. Subsequently, palace forces, including the Shence Army, developed under eunuch control and participated in imperial campaigns. These arrangements did not eliminate dependence on provincial armies: they placed court representatives and centrally controlled troops alongside them, preserving instruments of imperial military action while creating additional tensions within the command structure. (cambridge.org)

Fiscal Reconstruction and Economic Geography

The fighting of 755–763 damaged the agricultural and fiscal foundations of the Tang dynasty in the northern Central Plain. Displacement interrupted cultivation, while military occupation and disrupted administration prevented agricultural surpluses from reaching imperial granaries. Damage to production, however, must be distinguished from the court’s loss of access to revenue: northeastern military governments could collect taxes locally without remitting them to Chang’an. Consequently, renewed cultivation did not necessarily restore central receipts. Postwar reconstruction was not simply a return to prewar arrangements; it required new means of collecting income and moving supplies from territories that remained fiscally accessible. The restoration of grain transport under Liu Yan demonstrates that wartime disruption was followed by substantial rebuilding rather than permanent economic paralysis throughout the north. (cambridge.org)

In 758, while the rebellion continued, the financial administrator Diwu Qi instituted a salt monopoly to obtain revenue outside the damaged system of agricultural and household levies. Liu Yan subsequently reorganized its operation, particularly during 766–780. Rather than having officials undertake every stage of distribution, his administration concentrated on purchasing salt from producers and selling it to merchants at a price incorporating the government’s charge. Merchants transported and resold it, passing the fiscal cost into retail prices. This arrangement reduced the state’s distribution expenses and allowed it to collect an indirect tax at controlled production and wholesale points, without administering every locality in which salt was consumed. Coastal production around the lower Yangtze and Huai rivers became especially important. Reported annual receipts rose from approximately 400,000 strings of cash at the beginning of Liu Yan’s administration to six million by 780, exceeding half the government’s cash income; these figures concern monetary receipts, not the entirety of revenues received in grain and textiles. (atlantis-press.com)

The lower Yangtze region thus acquired greater immediate importance to imperial finance, although its emergence as an economic center was a development extending across centuries. Before 755, transport policy had focused heavily on moving grain from Luoyang westward to the capital. After 763, securing shipments from the Yangtze–Huai basin to Heyin on the Yellow River became essential to replacing supplies formerly available from Hebei and Henan. Southern fiscal resources included agricultural taxes as well as salt income: commercial revenues could also finance grain purchases for northern consumption. This dependence made the rivers, canals, warehouses, and transshipment stations linking the southeast to the capitals integral parts of fiscal reconstruction. (www11.ihp.sinica.edu.tw)

Liu Yan’s rebuilding of this transport system used successive shipping stages rather than requiring vessels to complete the entire journey. Grain gathered at Yangzhou passed through the Grand Canal network and the Bian River to Heyin, then traveled along the Yellow and Wei rivers toward Chang’an. Boats and crews served particular waterways, with cargo transferred at intermediate depots: vessels suitable for one river did not necessarily navigate the next. Channel maintenance, adapted vessels, and specialized crews addressed the differing currents and hazards along the route. Historical accounts credit the reorganized system with transporting approximately 1.1 million shi of grain annually, illustrating the scale of the restored supply operation. (ctext.org)

The Two-Tax reform of 780, introduced under Emperor Dezong on the proposal of chief minister Yang Yan, followed the rebellion’s end by seventeen years. It addressed both wartime dislocation and longer-standing changes in landholding, taxable registration, and exemptions that had weakened the older zu-yong-diao obligations of grain, textiles, and labor service or its commutation. Households were to be registered where they actually resided, whether native or migrant, and assessed according to wealth rather than adult-male status alone; land taxation also remained part of the arrangement. Collection occurred in summer and autumn, with deadlines in the sixth and eleventh lunar months. Yang Yan’s proposal sought to consolidate overlapping levies and allocate taxation according to government expenditure. It did not eliminate practical difficulties: Lu Zhi later criticized the valuation of different kinds of property in monetary terms and argued that mobile traders could evade demands more readily than settled agricultural households. (cambridge.org)

Civilian Experience, Migration, and Population Evidence

Civilian exposure to the war varied sharply by locality. Between 755 and 763, rebel and loyalist armies repeatedly traversed Hebei, Henan, Hedong, and Guannei, making northern agricultural districts especially vulnerable to plunder and displacement. Occupied communities faced demands for provisions, while the wider population supported military operations through grain, cloth, money, requisitioned horses, and compulsory transport labor. These burdens extended beyond places where battles occurred: supplying an army could remove both household resources and workers from civilian production. Much of the Yangtze basin escaped the repeated campaigning experienced in the north, but it was not uniformly peaceful. Liu Zhan’s forces captured Suzhou in 760, and government troops retook it early in 761. (cambridge.org)

Besieged towns experienced a more concentrated crisis. At Suiyang in 757, where Zhang Xun and Xu Yuan maintained resistance to the rebels, dwindling grain supplies led to reduced rations and the consumption of substitute foods, including bark and tea leaves. Historical accounts describe the slaughter of horses for food and, eventually, the killing and consumption of civilians by the garrison. David A. Graff’s study distinguishes these reports from the much less secure claims that tens of thousands were eaten; starvation had already endangered the civilian population, and the surviving narratives do not establish a reliable victim count. Forced recruitment also disrupted families outside besieged cities. Du Fu’s “Three Officials” poems, written in 759, portray officials seeking replacements after the loyalist defeat at Yecheng. They provide contemporary representations of coercion and household loss, rather than a statistical record of recruitment. (www11.ihp.sinica.edu.tw)

Wartime migration carried people from northern China toward the Yangtze basin and other refuges, accelerating a movement that long predated the rebellion. Northern warfare had encouraged southward settlement since the centuries following the Han dynasty’s collapse; the upheaval beginning in 755 was another major episode, not the origin of southern development. Refugee settlement contributed to the growth of commercial cities in the lower Yangtze region, already connected to northern capitals through the Grand Canal. Sichuan offered another destination. Du Fu’s itinerary illustrates a circuitous route rather than a direct flight south: after leaving his official position in 759, he travelled through Qinzhou and Tonggu before reaching Chengdu at the end of the year. This documented individual journey cannot establish the numbers or social composition of the wider refugee movement. (cambridge.org)

The population figures preserved for the Tang dynasty measure administrative registration, not a complete modern population census. Households and their recorded members formed a basis for taxation and service obligations; even before the rebellion, registration did not encompass everyone living within the empire. Du You estimated the prewar population at 13–14 million households, substantially above the official household totals. (cambridge.org) The commonly cited comparison gives 8,914,709 households and 52,919,309 people in 755, against approximately 2.9 million households and 16.9 million people in 764. These totals were compiled under radically different conditions. Flight could leave survivors absent from their original registers, while disrupted local administration and reduced reporting from northern provinces diminished the central government’s coverage. (en.wikipedia.org)

The frequently repeated estimate of approximately 36 million deaths converts the difference between those registration totals into fatalities. That operation conflates mortality with displacement, underregistration, and loss of administrative access. Hans Bielenstein challenged such readings in 1947, emphasizing the disruption of population registration and reliance on simplified tax records. (en.wikipedia.org) The rebellion unquestionably caused severe civilian suffering, but the registers do not separate people killed by violence from those who died of hunger, fled elsewhere, or survived beyond effective enumeration. They therefore cannot securely establish either a 36-million death toll or claims derived from it about the proportion of the world’s population killed. (academic.oup.com)

Poetry, Cultural Responses, and Remembrance

Du Fu experienced the rebellion as a separated husband, a captive in occupied Chang’an, and subsequently a minor official at the loyalist court. Captured while attempting to reach Emperor Suzong in 756, he escaped in 757. His “Spring View” (Chun wang), written during the occupation, juxtaposes enduring mountains and rivers with a devastated political community and an overgrown capital. Flowers and birds, conventionally associated with seasonal pleasure, instead evoke tears and alarm; the extraordinary value assigned to a letter from home connects public warfare to private separation. “Moonlit Night” (Yue ye, 756) approaches that separation indirectly, imagining his wife watching the moon in Fuzhou rather than simply describing his own loneliness. Its imagined viewpoint is a literary construction, not eyewitness testimony about her actions. (en.wikipedia.org)

Du Fu’s representations of military service similarly combine particular situations with crafted voices. The “Three Officials” and “Three Partings,” conventionally dated to 759, portray recruitment through encounters with officials and the perspectives of people compelled to leave their families. In “The Official at Shihao,” an elderly man escapes a nighttime levy while his wife confronts the officer; her account of sons serving at Yecheng and an impoverished household exposes the pressure exerted by loyalist mobilization. The poem gives the woman an extended speech but leaves the officer’s demands largely implicit. Such selective narration makes civilian suffering central without turning the work into an administrative record or a verbatim transcript. (hillpublisher.com)

Li Bai and Wang Wei occupied different political positions from Du Fu. Li Bai entered the staff of Prince Yong, Li Lin, in 756–757; the prince’s defeat in a conflict with Suzong’s government led to Li Bai’s imprisonment and sentence of exile to Yelang, later interrupted by an amnesty. This was involvement in an intra-dynastic struggle, not service to An Lushan’s Yan regime. Wang Wei, by contrast, was captured in Chang’an in 756, taken to Luoyang, and given office under the rebels. After the Tang recovery, he faced charges of collaboration but received a pardon; his brother Wang Jin’s intervention and a poem lamenting rebel festivities at Ningbi Pond contributed to his rehabilitation. Their later reputations as, respectively, an inspired wanderer and a contemplative poet should not substitute for these distinct wartime circumstances. (brill.com)

Responses beyond these famous figures reveal the conventions through which Tang poetry represented disorder. Gao Shi’s 759 “Responding to Supernumerary Pei, a Poem in Place of a Letter” recounts military and official experience but ultimately redirects attention toward withdrawal from dangerous public commitments. Qian Qi’s poem about seeking refuge in a Buddhist temple after the eastern capital fell places historical danger more explicitly in its title than in its verses, where bells, still water, and mountain space offer an imperfect refuge from anxiety. Lucas Rambo Bender interprets such works as attempts to situate immediate catastrophe within more enduring frameworks. Allusion, parallelism, seasonal recurrence, and the language of reclusion organized disrupted experience rather than merely recording it. (cpb-us-w2.wpmucdn.com)

Later remembrance concentrated especially on Xuanzong and Yang Guifei. Bai Juyi composed “Song of Everlasting Sorrow” (Chang hen ge) in 806, fifty years after Yang’s death at Mawei. The narrative moves from imperial desire and neglected rule through flight and bereavement to a supernatural search for the dead consort. Chen Hong’s accompanying prose account likewise incorporates an otherworldly encounter and a secret pledge between the lovers. These episodes document ninth-century literary remembrance, not contemporary evidence for private conversations or supernatural events. The poem and prose narrative subsequently supplied material for Bai Pu’s Rain on the Paulownia Tree and Hong Sheng’s The Palace of Eternal Life, allowing the rebellion to be reworked as both a political catastrophe and a tragedy of imperial love. (ctext.org)

Historical Sources and Interpretive Limits

The principal narrative sources were compiled long after the fighting. The Old Book of Tang (Jiu Tang shu), completed in 945 under the Later Jin and conventionally credited to Liu Xu and his collaborators, preserved substantial material from earlier Tang court histories. The New Book of Tang (Xin Tang shu), completed in 1060 under Ouyang Xiu and Song Qi, revised that inheritance within the intellectual and political setting of the Song dynasty. Sima Guang’s Zizhi Tongjian, presented to the throne in 1084, reorganized earlier evidence into a chronological history intended to illuminate government. None is an eyewitness account as a whole, although each incorporates information originating much closer to the rebellion. (en.wikipedia.org)

Their chronological distance does not make these works uniformly unreliable, but their overlapping sources mean that agreement is not necessarily independent corroboration. Tang historians had compiled veritable records of successive reigns and national histories from court documentation; later compilers inherited and edited this material. Political interpretation therefore entered the record before the Song period. Denis Twitchett observed that historians writing in the late 750s already represented Xuanzong’s reign as a movement from exemplary government toward catastrophe. Song revisions added further selection: Ouyang Xiu shortened the imperial annals to clarify moral lessons, while Song Qi reshaped biographies to make individual successes and failures more instructive. Accounts of loyalty, ministerial misconduct, or imperial negligence must consequently be distinguished from the events they organize and explain. (cambridge.org)

Biographies, administrative compilations, and inscriptions offer different checks on this narrative framework. Excavated epitaphs can supply dated deaths, official appointments, family movements, and references to service in rebel-held territory; inscriptions using rebel reign titles also preserve evidence that later loyalist narratives could suppress or recast. Such texts nevertheless commemorate their subjects and are not neutral statements of allegiance. Administrative evidence likewise requires attention to what was counted. Du You’s Tongdian, for example, preserves figures for military expenditure and population that permit comparison with narrative claims, but estimates of nominal establishments, actual soldiers, and registered households are not interchangeable. David A. Graff’s “The Reach of the Military: Tang” demonstrates how changing assumptions about army strength and population substantially alter calculations of military participation. (xuebao.snnu.edu.cn)

Tang poetry provides contemporary perspectives without functioning as a continuous campaign record. Poems can corroborate the experience of separation, disrupted official careers, or movement between places, but their speakers, allusions, and conventional imagery complicate literal readings. Lucas Rambo Bender’s 2019 study, “Other Poetry on the An Lushan Rebellion,” shows that many poets sought to place upheaval within enduring natural, philosophical, or religious patterns rather than simply document it. The prominence of Du Fu in later remembrance should therefore not make his intensely historical mode representative of all wartime verse. (campuspress.yale.edu)

Modern scholarship has approached causation through several scales of analysis. E. G. Pulleyblank’s The Background of the Rebellion of An Lu-shan (1955) remains a major study of the conditions preceding the uprising. Volume 3 of The Cambridge History of China (1979) supplies complementary treatments: Twitchett’s “Hsüan-tsung (reign 712–56)” examines the prewar reign, C. A. Peterson’s “Court and province in mid- and late T’ang” addresses the transformed provincial order, and Michael T. Dalby’s “Court politics in late T’ang times” scrutinizes political evidence and its omissions. Military organization, court rivalry, regional autonomy, and ethnic networks are analytical emphases, not mutually exclusive or conclusively ranked explanations. (cambridge.org)

Ethnic interpretations particularly require distinctions between ancestry, social networks, and political commitment. Research on Sogdian communities identifies substantial support for An Lushan within some networks, but also Sogdians who supported the Tang; shared origins did not prescribe allegiance. Evidence of subsequent concealment of foreign ancestry indicates altered pressures on identity, not an instantaneous disappearance of cultural diversity. Similarly, Peterson distinguishes the rebellion’s catalytic effects from institutional changes already under way and explicitly rejects its inevitability. Treating 755–763 as a major turning point need not turn the remaining history of the Tang dynasty into a predetermined decline or make the war a single, immediate endpoint of Tang cosmopolitanism. (iranicaonline.org)

References

1. THE REACH OF THE MILITARY: TANGcambridge.org
2. Military history of the Tang dynastyen.wikipedia.org
3. Late Tang China and the World, 750–907 CE: Excerptassets.cambridge.org
4. The Army of Tang Chinascribd.com
5. New Book of Tang: Biographies 52: Zhang, Yuan, Pei 新唐書 : 列傳第五十二　張源裴ctext.org
6. The Chinese Grand Canal World Heritage Site: living heritage in the 21st centuryetheses.whiterose.ac.uk
7. AN LU-SHAN — Encyclopaedia Iranicairanicaonline.org
8. An Lushan — Wikipediaen.wikipedia.org
9. Yang Guozhong — Wikipediaen.wikipedia.org
10. An Lu-shan — Encyclopedia.comencyclopedia.com
11. An Lushan — Universaliumuniversalium.en-academic.com
12. An Lushan rebellionen.wikipedia.org
13. Geshu Hanen.wikipedia.org
14. Emperor Xuanzong of Tangen.wikipedia.org
15. Emperor Suzong of Tangen.wikipedia.org
16. Rebellion of An Lushan | History | Research Starters | EBSCOhostebsco.com
17. Pacific World, Third Series, Number 15shin-ibs.edu
18. Li Guangbien.wikipedia.org
19. The Reach of the Military: Tangcambridge.org
20. Siege of Suiyangen.wikipedia.org
21. University of Albertaciteseerx.ist.psu.edu
22. An Qingxuen.wikipedia.org
23. Battle of Xiangji Templeen.wikipedia.org
24. 資治通鑑/卷220zh.wikisource.org
25. 資治通鑑/卷221zh.wikisource.org
26. 資治通鑑/卷222zh.wikisource.org
27. 资治通鉴 卷221 · 资治通鉴chinese-classics.org
28. Relations between the Uighurs and Tang Chinadergipark.org.tr
29. Shi Chaoyien.wikipedia.org
30. The Uighur Empire According to the T’ang Dynastic Histories: Selectionsdepts.washington.edu
31. Exonerating the Horse Trade for the Shortage of Silk: Yuan Zhen’s “Yin Mountain Route”cuhk.edu.hk
32. Introduction: China on the Eve of the Five Dynastiescambridge.org
33. Four Garrisons of Anxien.wikipedia.org
34. The Tibetan Invasion of Ch’ang-an 長安 the Capital City of the T’ang Empire, in A. D. 763cir.nii.ac.jp
35. Between China and Tibet: Adash.harvard.edu
36. Tian Chengsien.wikipedia.org
37. Chengde Provinceen.wikipedia.org
38. Li Huaixianen.wikipedia.org
39. Emperor Dezong of Tangen.wikipedia.org
40. Economic transformation in the Tang-Song transition (755 to 1127)cambridge.org
41. An Optimal Tax That Destroyed the Governmentsllr.j.u-tokyo.ac.jp
42. Tax System Reform and Its Impacts in the Middle and Late Tang Dynasty. Take Two-Tax System and Commercial Tax System as Examplesatlantis-press.com
43. Salt and State: An Annotated Translation of the Songshi Salt Monopoly Treatisedokumen.pub
44. Salt in Chinese historyen.wikipedia.org
45. General Annals of Jiangnan: Volume 77ctext.org
46. Tang Debate on the Twice-a-Year Tax: Two Memorials For and Against, by Yang Yan and Lu Zhiafe.easia.columbia.edu
47. Book Excerpt! How to Read Chinese Poetry in Contextcupblog.org
48. The Expansion of China’s Imperial Urban Civilization to the South (220–755)cambridge.org
49. Chronologycup.cuhk.edu.hk
50. Du Fuen.wikipedia.org
51. Du Fu, Yue ye - Poésie Tangpoesie-tang.unige.ch
52. Du Fu’s Three Officials and Three Partings: Poetic Historical Testimony and Artistic Paradigm of Social Suffering in the Mid-Tang Dynastyhillpublisher.com
53. A Confucian Poem by Li Bai: Translationbrill.com
54. New Book of Tang: Biographies 127: Literature and Art IIctext.org
55. Wang Weien.wikipedia.org
56. Other Poetry on the An Lushan Rebellion: Notes on Time and Transcendence in Tang Versecpb-us-w2.wpmucdn.com
57. Biography of Everlasting Sorrow, by Chen Hongctext.org
58. Chang Hen Ge (poem)en.wikipedia.org
59. Old Book of Tangen.wikipedia.org
60. New Book of Tangen.wikipedia.org
61. The Writing of Official History under the T’angcambridge.org
62. The Cambridge History of Chinacambridge.org
63. Rewriting Tang Literati Lives: New Views of Literary Composition in Biographies of the New Tang Historychina.usc.edu
64. Publications — Lucas Rambo Bendercampuspress.yale.edu
65. The Background of the Rebellion of An Lu-shan — Reviewcambridge.org