Escape From Rome

by Walter Scheidel

  In the 490s, the regime followed up with “Sinicizing” measures designed to close the gap between the conquest class and the Chinese majority population. These interventions targeted the elite: the capital city was moved to the old Han center of Luoyang; Xianbei and Han elites were encouraged to intermarry; and the use of Chinese language and costume at court became mandatory. This embrace of Chinese culture not only acknowledged existing trends but also reflected rulers’ desire for greater unity. Chinese governors were appointed to govern provinces, Chinese served in militias, and in a shift away from the previous model of dispersed occupation, ethnic Tuoba garrisons were concentrated at the center and the northern frontier.51

  Although these reforms were meant to boost internal cohesion, the move of many Tuoba military units to the steppe periphery proved to be both unpopular and destabilizing. Retaining their native language and customs, these forces were increasingly alienated from the central authorities. In the 520s this led to a rebellion that triggered internal wars, and in 534 split the state in two. These dislocations, however, proved conducive to further state formation as endemic and symmetric conflict between the two successor polities raised demand for institutions that bolstered capabilities.52

  The western successor state (under the Western Wei and then the Northern Zhou dynasties) controlled fewer economic resources and a smaller population. Initially struggling to survive eastern attacks, state leaders co-opted Chinese militias to compensate for the relative dearth of Xianbei warriors. Censuses and land allocation programs were revived. The 550s witnessed the creation of the “24 Armies”: composed of Xianbei and even more so of Chinese, they mobilized commoners through direct recruitment or the co-optation of local forces.

  At the same time, under the fubing system, more adult men were granted land in exchange for regular military service obligations. This added farmer-soldiers to an expanding infantry even as horses obtained via the Gansu corridor helped maintain strong cavalry forces. These policies simultaneously mobilized a growing share of the population for military purposes and buttressed the rural smallholder class. Chinese elites became more widely involved in administration, and bureaucratic reforms created ministries. All of this raised state capacity and helped cement centralized authority. The overall result was a shift toward the kind of well-integrated high-mobilization state that had been characteristic of the Warring States and Western Han periods. The payoff was large: army strength grew from perhaps 50,000 in the 550s to well more than 100,000 by the 570s.53

  By contrast, the eastern successor state continued to be run by Xianbei as officials and fighters. Co-optation of Chinese forces was constrained by the influence of magnate families. Society was riven by greater divisions between locals and Xianbei, and political conditions were less stable.54

  This growing imbalance in state capacity enabled the western state to overcome and annex its eastern counterpart in 577. Soon thereafter it confronted southern China, where development had followed a rather different path. The fact that a single state had been maintained there for 270 years following the Jin retreat masks a considerable degree of state deformation. Initial efforts to retake the north between 349 and 369 failed and led to gradual demilitarization of the southern elite. Centralized state power waned for about a century until the takeover of the Liu Song dynasty in 420: low-born military men wrested power from the great landowning families, commerce and remonetization revived urbanism and checked the influence of landlords, and professional troops replaced magnate-led forces. However, although this arrangement contained rural elites, it also gave rise to intensifying civil wars.55

  Before long, powerful families dominated the scene, whether magnates or military clans, and even the latter were unable to establish a more militarized culture comparable to that of northern China. Military colonies did not take root in the south. The court failed to raise large armies as it faced a two-way struggle against domestic rivals and the growing northern threat. Large landlords’ shielding of the rural population from registration and thus taxation and service obligations was a persistent problem: as a result, military assets were always in short supply, prompting recourse to convicts, vagrants, and aboriginals of low combat effectiveness.56

  Census numbers from northern and southern China highlight a growing disparity in state capacity. In 464, the southern Liu Song regime counted a little more than 900,000 households with some 5 million members, in an area that 300 years earlier had housed twice as many, even before migration and population growth had raised densities further. By 589, the final year of the southern state, the Chen regime registered 500,000 households and 2 million people in a significantly smaller territory, which in 140 had contained closer to 10 million people.

  By comparison, the northeastern state of Northern Qi alone counted 3.3 million households with some 20 million residents in 577, at a time when registration quality was not particularly high. An empire-wide census in 609 suggests an actual total in excess of 9 million households, more than one-sixth of which were bound to be located south of the Yangzi. All this unmistakably shows that the southern authorities were unable to capture more than a modest fraction of the households in their territory, most likely fewer than a third, and an even smaller proportion of its residents, quite possibly as few as one in five.57

  This stark imbalance in state capacity goes a long way in explaining why imperial restoration emanated from northern China. The south had not failed for want of trying. However, none of the invasions attempted between the 340s and the 570s accomplished anything of substance. It is true that geography also contributed: without sufficient cavalry, southern armies lacked mobility and shock power, and heavily relied on a river network that did not support northward operations. Even so, political conditions were crucial. In the long run, southern unity was not only not an advantage but a distinct impediment: southern China experienced none of the bloody yet productive rivalries that drove northern state formation in the fourth and especially in the mid-sixth century.

  Before the late sixth century, northern states had shown little interest in attacking the south, for two reasons. One was that endemic competition among similarly belligerent peers kept northern forces busy and spared the south. The other lay in operational difficulties: it was challenging for northern cavalry to operate against fortified southern positions and in riverine areas. The only major attack, in 383, had ended in defeat and triggered regime collapse.58

  Stable unification of northern China changed all that as it removed internal divisions and generated the resources required to take on the south on its home turf. In the late 580s, the northern Sui state deployed a large fleet and is said to have mobilized 518,000 men to invade a south that was defended by much weaker forces. This massive assault swiftly succeeded, and a subsequent magnate-backed rising against registration and taxation was likewise crushed.59

  As a result, by the 590s most of the lands that had belonged to the Han empire were once again ruled from the old Han capitals of Chang’an and Luoyang. Elite autonomy was safely contained, as the Sui regime abolished noble ranks and hereditary rights: the principal status distinction was between officials and commoners, with the former being rotated among positions to keep them on their toes. Mobilization of military and civilian manpower occurred on a scale not seen for half a millennium. Chang’an was rebuilt with outer walls that extended thirty-five kilometers, and vast numbers of workers—supposedly 2 million—were drafted for construction in Luoyang. A whole network of lavish palace complexes followed.60

  The most ambitious infrastructure project was the construction of an interlinking system of canals that for the first time connected northern and southern China. From the Sanggan (later Beijing) area in the north it traversed the Central Plain and linked the Yellow River, the Huai, the Yangzi, and Hangzhou Bay: at least on occasion, more than a million workers toiled on 2,357 kilometers of canals. This grandiose scheme became a reality in the first decade of the seventh century, a time the later Song historian
Sima Guang called the “height of the Sui”: in many ways a reincarnation of the Han empire, their domain encompassed over 190 prefectures and 1,225 counties and registered almost 9 million households.61

  In operations reminiscent of the terminal phase of the Warring States period or the Xiongnu wars of the Martial Emperor of the Western Han, enormous resources were committed to warfare in the northeast. In 612, the regime followed up on previous invasions of Korea by supposedly mobilizing 1,133,800 combat troops and twice as many support personnel. And although these numbers are undoubtedly hyperbolic, if the troops were divided into no fewer than thirty separate armies that could be supplied separately, a grand total in the hundreds of thousands was in fact perfectly feasible.62

  Even for the freshly unified empire, this was a step too far. These megalomaniacal exertions unleashed popular resistance that brought down the Sui regime and ushered in a brief phase of renewed state weakness. Under the early Tang dynasty, compromises with local power-holders allowed only a small fraction of its subjects to be counted and taxed. Yet by the middle of the eighth century, the official census tally had returned to more than 9 million households with 53 million people, equivalent to Han and Sui levels. More than 600,000 fubing farmer-soldiers were on the government’s rolls, and each year several hundred thousand civilians were drafted for auxiliary service. Fiscal administration was once again highly centralized.63

  These capacities dwarfed anything that could have been achieved in Europe at that time, even in the headiest days of Charlemagne. Rulers could only guess at the number of their subjects, found it increasingly hard to draft them, and no longer had any hope of taxing land outside royal estates.

  What accounted for this dramatic divergence? The evidence I have briefly laid out gives pride of place unequivocally to the coercive capacity of rulers to frame conditions in favor of elite cooperation and subordination. European-style aristocratic polycentrism did not appear in China because the principal dynasts always, and especially from the fifth century CE onward, disposed of sufficient military assets to check the local power of magnates and ensure centralized revenue collection that in turn sustained military resources. This encouraged elite families to commit to imperial rule as a source of wealth and status, and enabled central authorities to stem and reverse recurrent trends toward devolution of power.64

  Cavalry forces that were either drawn from the margins of the steppe zone or supplied with horses imported from the great grasslands, recourse to hereditary soldiers who occupied state-owned land, and more generally the ability to mobilize complementary resources to support their operations proved critical both to the maintenance of centralized state authority and to state-building on an imperial scale. As I argue in chapter 8, thanks to its physical proximity to the steppe frontier, northern China was particularly likely to experience these conditions. Conversely, none of them were present in Latin Europe.

  Intermittent abatements of state power did not fundamentally alter outcomes. Thus, after internal disorder in the 750s, the Tang domain came to be effectively divided among governors, generals, and warlords, only some of whom remitted taxes to the center. Registration levels once again plummeted to just a few million households. Eventually, the collapse of what was left of the Tang regime around 900 precipitated temporary fragmentation, which in its most basic outlines resembles that of the much longer Period of Disunion from the fourth through the sixth centuries.65

  In northern China, early successor states (Later Liang and Later Tang) rested on core zones where an intact bureaucracy maintained population registers but struggled to control provincial governors while a steppe regime, Liao, consolidated its position along the northern frontier. Conflict of the Later Jin dynasty with Liao in the 940s visited defeat and devastation on the capital and the Central Plain region. Liao’s subsequent retreat opened up space for a more centralized regime (Later Zhou, following Later Han): prolonged warfare had depleted the governors’ resources and ability to resist the center, which was now able to rebuild a strong unified army and embark on a grand expansionist program that first unified the north and under the subsequent Song dynasty (from 960) took over the various smaller states of the south.

  Those polities, meanwhile, had replicated earlier conditions, such as a shift from military to civilian rule, strong reliance on commerce, interference by powerful landlords who concealed taxable land, poor military capabilities, and rifts among rival elites. Just as they had done for the Liu Song and Chen of the fifth and sixth centuries, these weaknesses led to the loss of lands north of the Yangzi followed by conquest from the north.66

  The Song empire exceeded the mature Han, Sui, and Tang empires in terms of population, military might, and fiscal capacity. In the late eleventh century, the central authorities counted some 20 million households. The state maintained more than 2,000 well-staffed tax stations to collect dues, overseen by a powerful central fiscal agency. The careers of high government officials often began in public finance. Total central tax revenue equaled at least 3,500 tons of silver, four-fifths of it in cash, roughly four or five times what the Roman empire took in at its peak and approaching a tenth of total output, a highly respectable rate by premodern standards. This bonanza not only supported a capital city of 750,000 but also an army in excess of a million soldiers, at least on paper.67

  Even though the Song soon relinquished control over northern China, the tenth century remained the last time China formally fractured into a handful of states: up to the end of the monarchical era, there was usually just one empire, except in the Southern Song period when there were two. By the year 1000, when the vestiges of Roman imperial institutions had faded away in Latin Europe, the Chinese tradition of hegemonic empire had become more firmly entrenched than ever before.

  “Follow the Money”

  In the early sixth century CE, the cleric and future saint Eugippius, in his biography of Saint Severinus of Noricum, reported a much earlier encounter between himself and the German Odoacer, the future ruler of Italy. As a young man, the latter had stopped on his way across the Alps to pay his respects to the holy man, who duly foretold his coming greatness and bid him farewell with the words, “Go to Italy, go, now covered in mean hides; soon you will make rich gifts to many!” And so Odoacer did, seizing power with the support of Italy’s restless Germanic mercenaries who demanded land assignments. His peers who led Goths, Vandals, Franks, and Lombards in carving up the western Roman empire bowed to the same calls for largesse. State resources became first localized and then privatized, tax collection fell into abeyance, and central authority withered. For centuries, “rich gifts” trumped state power. Imperial restoration became a distant dream.68

  Northern China during the Period of Disunion moved in the opposite direction. In both East and West, conquest regimes had started out with dual-status systems that demarcated newcomers from locals. I already noted the guoren–native divide. The Ostrogoths in Italy did not assume Roman citizenship, were barred from holding civilian office, and may even have been subject to separate laws. In Gaul, the Franks were initially the only group exempt from taxation.69

  In both cases, integration gradually blurred these boundaries. Yet, whereas in Western Europe tax immunity and its replacement by localized rent and service obligations spread across the general population, in China everybody became subjected to homogenized claims by the central state. This crucial difference determined whether it was landed lords or state rulers who captured most of the surplus. State capacity differed accordingly, with noble levies, small armies, and rudimentary administrative structures on the one hand and extensive censuses, huge militaries, and ministries full of literate bureaucrats on the other. The former sustained intense polycentrism, the latter hegemonic empire.

  These outcomes merely represent two extremes on a wider spectrum. Developments in the Middle East, where much of the eastern Roman empire and the entire Sasanian empire succumbed to the caliphate, occupied an intermediate position. Just as the Germans in Euro
pe and the Xiongnu and Xianbei in East Asia, the Arabs had long interacted with neighboring empires: they had accepted their patronage, fought proxy wars, and even settled in imperial lands. Outright takeover was simply the next step.70

  Unlike Germans on Roman soil, Arabs often settled in garrison cities such as Basra and Kufa in Iraq, Fustat (Cairo) in Egypt, and Kairouan in Tunisia. Iraq was occupied by tribesmen from the northern and eastern Arabian peninsula who, tied to their garrisons, gave up pastoralism and became wholly dependent on tax revenue. There and elsewhere, Sasanian and Roman practices of taxation were generally maintained: the Iberian peninsula remained an exception. There the lack of an adequate fiscal infrastructure—which had once existed but failed to survive Visigothic rule—necessitated dispersed settlement. Because the occupying forces received mostly cash stipends, in-kind taxes declined and land grants did not play a significant role. These practices stood in marked contrast to those in Western Europe, where land assignments made land and poll taxes redundant, and in fourth-century CE northern China, where requisitions supported the conquerors until regularized tax systems were restored. As a result, the Arab military class depended directly on the state for its sustenance and status.71

  The principal weakness of these arrangements, discussed in chapter 5, was that the state was focused more on individual provinces than on the caliphal center: regional armies controlled their respective tax bases. The Umayyad system was better at preserving administrative and extractive institutions beyond the local level than the Germanic regimes but less so than the regimes of northern China. It likewise occupied an intermediate position in terms of recognizing nonsoldiers’ customary rights to receive revenue: in this regard, it bore greater resemblance to the increasingly hereditary Germanic land assignments than to Chinese practices, even though the fubing model also invited creeping privatization of state resources.