The second round of civil war, in the 680s, witnessed renewed clashes between forces in Syria, Iraq, Iran, and the Arabian peninsula. The Syrian army, which was instrumental in pacification, subsequently supplanted the forces that had previously held Iraq. As the only remaining military entity west of Iran that was capable of supporting major campaigns (with the help of troops from the Caucasus frontier), it was responsible for the final push against Constantinople in 717–718, and in 741 sent reinforcements to the Maghreb to help suppress a Berber uprising. Meanwhile, the army of Khorasan in Iran gathered strength, assuming a balancing position similar to that previously held by the Iraqi units.
The third round of civil war in the 740s that brought down the Umayyad dynasty once again pitted northern Arab against Yemeni networks and several regional armies against one another. At first, the Syrian forces disintegrated in a conflict between contenders supported by north and south Arab elements. After an attempted defection of Egypt and risings in Iraq had been suppressed, Iran became a center of opposition, driven in part by Arab settlers in Merv and Yemeni elements hostile to the northern Arab-dominated caliphate farther west. Escalating war destroyed the Umayyads, who had had to fall back on northern frontier troops, and ushered in the rule of the Iran-based Abbasids, who immediately shifted the imperial capital to Mesopotamia (first Kufa, then alternately Baghdad and Samarra)—and hence farther away from Europe.25
This new arrangement, however, merely displaced tensions eastward, as Iraqi and Iranian interests collided. This eventually led to a fourth round of open warfare. Ostensibly a conflict over the Abbasid succession, it was fought in the 810s between Arab and Iranian forces but reverberated into the 830s, and encouraged the regime to employ Turkic mercenaries for its protection—a very belated acknowledgment that long-standing internal divisions and enmities could not be overcome and a novel strategy was called for.
By then, the Sunni-Shia rift had crystallized into separate political-military systems. The introduction of foreign mercenaries simply generated new tensions between what effectively represented a new conquest group and local elites who resisted taxation, especially when revenue was collected for the benefit of outsiders hired to keep these elites in check. Once the Turkic units enjoyed a monopoly on organized violence, they readily joined in lucrative partisan conflicts. During the Samarra crisis of the 860s, five different caliphs were raised up by competing Turkic factions.
This resulted in the most dramatic meltdown thus far: Iran was ceded to rebels, the authorities in Egypt stopped remitting tax to the center and pushed into Palestine and Syria, and southern Iraq was buffeted by a huge slave revolt. Brief and incomplete recovery was soon followed by renewed fragmentation from the 920s onward, as Egypt and Syria established de facto independence from the caliphate. In 945, the Buyids from Iran seized control over Baghdad and Iraq. At that point, the Abbasid caliphs, even if they formally hung on until 1258, had already lost any tangible political power.
By the late tenth century, the splintering of the formerly Arab empire had reached new heights: the Samanids controlled eastern Iran, two or three Buyid polities coexisted in western Iran and Iraq, two Hamdanid regimes shared Syria, and Ikhshidids and then Fatimids ruled Egypt. Kurdish dynasts and the Qaramita in the Arabian peninsula rounded off this mix. As described below, the Maghreb and Iberia had by then long gone their separate ways. For several centuries, the trend was unidirectional: fragmentation intensified over time.26
Regionalization and kin and tribal ties that persisted and were reproduced in occupied lands all converged in weakening the center’s grip on its far-flung domain. Fiscal decentralization contributed significantly to the resultant process of segmentation. From early on, Arab soldiers received cash salaries, a practice that helped preserve the existing Roman and Sasanian tax systems.27
At the same time, fiscal redistribution was highly decentralized: the regional armies retained most of the taxes collected in their respective territories. From the outset, the system had been shaped by the expectation that the revenue collected from infidels belonged to their occupiers as a collective, and that the latter were entitled to it by right of conquest and not as a grant from the caliph. More strikingly, the male descendants of the men who set up the original garrisons fully expected to continue to receive emoluments as a hereditary right.
The question of whether receipt of public stipends should be contingent on actual military service or expand into a sinecure was hotly contested between the central authorities and the beneficiaries of the more generous interpretation of entitlements. The related question of whether what was left after these disbursements was to remain in the same region or to be remitted to the court of the caliph likewise proved controversial. In practice, only a small portion of all state revenue ever reached the imperial center. The Umayyad rulers in Damascus relied principally on Syrian taxes but received little from Iraq and Egypt, not to mention from more distant provinces: there is no sign that the Maghreb or Iberia ever contributed anything of substance.28
As a consequence, the caliphs were heavily restrained from accessing and channeling the abundant riches that their vast empire generated. Nominal military strength as derived from the tallies of those on the government payroll was highly deceptive: while some 250,000 to 300,000 men were entitled to stipends, only a small fraction of them could actually be mobilized, and even fewer were ready to be deployed outside their own regions. Caliphs’ attempts to seize at least some of this surplus created tensions with regional interest groups. When arrangements were changed around 700 to limit pay to those on active military duty, this reform helped sustain effective forces in the various regions but did not address the more fundamental problem of effective decentralization.29
Instead, fiscal fragmentation facilitated political divisions. As revenues had always been largely retained in the regions where they had been levied, the establishment of new polities did not entail dramatic changes but simply formalized existing fiscal dispersion. The step from remitting a small fraction of total revenue to none at all was a small one, fraught with symbolic meaning but relatively easy to take in practice.30
Under these circumstances, it was not the eventual splintering of the political domain that was remarkable, but rather that unity endured for as long as it did. Military force was essential in maintaining a degree of cohesion: as long as the army of Syria was capable of suppressing dissent farther east (as it did in the 650s and 680s) or the army of Iran was strong enough to do the same in the west (as in the 740s), it was possible to maintain a single caliphate. When in the ninth century tensions became too difficult to manage, mercenaries were brought in to save the regime. These, however, merely reinforced the divisions between rulers and ruled without being invested in unity across different regions. In the end, their backing of rival factions accelerated political fission.
All this highlights the long-term costs of Arab-style “conquest-lite,” which allowed a partially fragmented core to reproduce its divisions on a grander scale in newly acquired peripheries. Ceding control of much of the surplus to regional armies may well have helped ensure formal loyalty by reducing the benefits of outright independence, especially as long as the Syrian army was still in a position to punish defectors. Yet this also prepared the ground for fragmentation once a cash-deprived center had been weakened by internal divisions and attrition. In this unpromising context, an early Abbasid attempt to centralize the fiscal regime did not survive for long.
The empire had from the beginning been a congeries of franchises run by and for the benefit of discrete regional groups of armed rent-seekers. Tribal and geographical rivalries stoked latent centrifugal tendencies. The complex and interregional revenue flows of the mature Roman empire provide an ideal-typical counterpoint to arrangements in the caliphate, as does the former’s much greater durability.
In the most general terms, conditions in the Umayyad caliphate were not favorable to what we might call the “Arab option”—the restoration of la
rge-scale empire in Europe under Arab control. Conquests did not fill state coffers to support further conquests but primarily expanded the body of beneficiaries without boosting mobilization. This alone greatly reduced the likelihood that the resources of empire could be marshaled in sufficient concentration to break Constantinople or make major inroads into Europe.
More specifically, Umayyad advances west of Egypt, while successful in terms of handing Arab warriors new subjects and lands to exploit, proved fairly ephemeral in terms of extending the area ruled by the center. After the occupation of Egypt, it took the Arabs half a century to seize Carthage. They reached the Straits of Gibraltar shortly afterward, recruiting local Berbers along the way. In an invasion conceived perhaps as not more than a plundering expedition of fewer than 10,000 men, they took advantage of internal divisions within the Visigothic kingdom to overrun most of the Iberian peninsula with the exception of its mountainous northern rim (711–716).
Members of the Visigothic elite readily joined the new conquest class. Unlike farther east, no separate garrison cities were set up, and Arabs and Berbers lived interspersed with the local population. For the remainder of the eighth century, administrative structures remained minimal. Inasmuch as taxes were collected at all, this practice was limited to the more densely settled and developed south.31
Having secured this region with fairly modest effort, its new rulers directed their attention to raids across the Pyrenees, focusing on the sliver of coastland west of the Rhone that was still held by the Visigoths. A foray into the duchy of Aquitaine, formally part of the Merovingian kingdom of the Franks but de facto an autonomous polity, was checked by a major defeat at Toulouse in 721 that cost the governor of al-Andalus (the caliphate’s new Iberian province) his life. Highlighting the fractured nature of the Arab front, the Berber leader in charge of annexing the Visigothic coastlands subsequently struck a deal with Aquitaine to secure independence from his Arab overlords back in Spain. He was in turn killed by the new governor of al-Andalus, who proceeded to attack Aquitaine and took Bordeaux. As he moved farther north to plunder Tours, he was defeated and killed by Frankish forces in 732.
The most important outcome of these operations was that Aquitaine was forced back under Frankish control, strengthening the Christian side in the process. A further Arab-Berber invasion in the late 730s brought new raids but no lasting gains; a third governor was soundly beaten in 737, even though he uncharacteristically survived. The Arabs managed to hold on to coastal Provence until the end of the 750s, when the Franks completed their takeover. In the 760s, the latter fully subdued Aquitaine, and at the end of the eighth century Charlemagne’s forces crossed the Pyrenees to establish strongholds on the other side.
I relate these events in some detail to clarify the central trend, which was a shift in the balance of power against the Arabs over the course of the eighth century. Whereas they had overcome the Visigoths with relative ease, their actions against the more formidable Franks were largely limited to raiding for plunder, and repeatedly failed at great cost. Galvanized by these hostilities, the Franks soon took the initiative and made gradual but steady progress. There is little in this story to support Edward Gibbon’s famously hyperbolic (if ironic) claim that the Arab incursions might have altered the path of European history:
A victorious line of march had been prolonged above a thousand miles from the rock of Gibraltar to the banks of the Loire; the repetition of an equal space would have carried the Saracens to the confines of Poland and the Highlands of Scotland; the Rhine is not more impassable than the Nile or Euphrates, and the Arabian fleet might have sailed without a naval combat into the mouth of the Thames. Perhaps the interpretation of the Koran would now be taught in the schools of Oxford, and her pulpits might demonstrate to a circumcised people the sanctity and truth of the revelation of Mahomet.32
Yet even though the view that all the Franks did at the time was to curb pillaging operations has since gained ground, the potential for plausible counterfactuals is nonetheless worth considering. Frankish military capabilities were relatively modest and still in the process of being restored when the Arabs appeared on the scene. Thus, if a few of these battles had gone the other way—always a credible contingency—the Arabs and Berbers might well have been able to penetrate France more deeply and form alliances with local potentates, just as they had done in Spain and also attempted in the case of Aquitaine. Southern France could have become a launchpad for ongoing assaults to the north and also eastward into Italy. The counterfactual of a sustained Arab presence coupled with some degree of Islamization cannot therefore be dismissed out of hand. The key question, however, is whether any of this could somehow have led to the incorporation of large parts of Europe into a unified Arab empire.33
Based on real-life developments in the western periphery of the caliphate, the answer can only be negative. After the Umayyad dynasty had been wiped out after its defeat in civil war in 750, a lone survivor took over al-Andalus in 756, prompting the first enduring secession from the (now Abbasid-ruled) caliphate. The center’s only attempt to restore control failed in 763 when an army of just a few thousand men was dispatched from North Africa. Local opposition had much greater impact: it took the new defector regime a couple of decades to suppress resistance within Spain itself, and even then the effective power of the emir of Cordoba was confined to the south of the peninsula while the other regions only nominally accepted his overlordship.
Around the middle of the ninth century, efforts at greater centralization on the early Abbasid model finally succeeded in imposing taxation more widely. However, this push for state-building soon alienated local elites and foundered in a prolonged period of internal conflict from the 880s to the 920s. Restoration of state power in the tenth and early eleventh centuries was followed by renewed collapse. By the 1030s the emirate had splintered into some thirty mini-kingdoms (taifs).34
Spain subsequently fell under the influence of Maghreb-based polities, first the Almoravids (around 1100) and then the Almohads (around 1200), until Christian states expanded from farther north. Not only was the only Arab-held region in Europe permanently severed from (and in fact hostile to) the caliphate, it experienced internal instability and fragmentation that worsened over time. Because of this, the Iberian peninsula was completely unsuitable as a staging area for further advances into Europe, even as Frankish power declined after Charlemagne: the Western European powers fractured but the westernmost Arab regimes were weaker still.
Moreover, the mid-eighth-century secession of Spain was part of a more generalized fraying of the caliphate’s western periphery that faced Europe. In 740, a massive Berber revolt threatened the center’s hold on the Maghreb. A substantial force dispatched from Syria to crush the rising failed: although the Tunisian and northeast Algerian core areas of Arab Ifriqiya could be saved, the central and western Maghreb were permanently lost to assorted Berber polities. The revolt also spread to the Berber units that made up the bulk of the occupation forces in Spain and had to be suppressed by the remaining Syrian forces, which then came to blows with the local Arab garrison. Armed conflict between the Syrian expeditionary force (dominated by North Arabian elements) and the regional army of Ifriqiya (dominated by men of Yemeni background) was only barely avoided and cooperation was poor. These tensions, together with the clashes in Spain, show the extent to which internal divisions obstructed coordinated military activity on a larger scale, even for defensive purposes.
Shortly afterward, the fall of the Umayyads prompted the defection of what was left of Ifriqiya in the 750s until it was regained for the Abbasids in 762. From 757, the Rustamids set up an independent theocratic imamate first in Tripolitania and then in Algeria: despite early attempts to suppress it, it survived for a century and a half. From 788, the Idrisids, originating in yet another refugee from a lost power struggle with the Abbasids, took over Morocco and lasted for two centuries. Meanwhile, Ifriqiya itself was run by the Aghlabid family throughout the ninth century
, nominally under Abbasid suzerainty but in practice highly autonomous: governors were no longer appointed by the caliph as family members inherited the position, and there is no sign of tax remittances to the east.
Even that lip service came to an end when the Fatimid (rival) caliphate took over in 909. The Fatimids in turn were the first regional force to advance eastward, reversing three centuries of westward expansion of major Middle Eastern powers from the Sasanians to the Umayyads and Abbasids. Their capture of Egypt in 969 meant that Asian powers were shut out of Africa until the Ottomans appeared in the sixteenth century and established indirect and soon diminishing nominal control as far west as Algeria.
In Spain and North Africa, the power of Arab and Berber states waxed and waned throughout the Middle Ages, but hegemonic empire did not return. At no time after the mid-eighth century was the original caliphate in a position to launch attacks against Europe. Without such backing, none of the peripheral successor states had any hope of establishing larger imperial structures in Europe. I already mentioned the fitful weakening of Arab power in Spain. The Aghlabids of Ifriqiya, relying on their own limited resources, needed seventy-five years to conquer Sicily. Subsequent raids against southern Italy were suppressed in the early tenth century. There is no better illustration of the basic fact that none of these Arab-led polities were capable of sustained expansion into Europe.35
Overall, in no credible counterfactual scenario could Arabs (and their local allies) have established a larger empire on European soil by way of North Africa. Whether an Arab takeover of Constantinople—a less far-fetched alternative outcome—might have allowed substantial advances through the Balkans is more difficult to judge, although the Bulgar empire would have posed a significant obstacle. What matters most, however, is not whether any such advances could have occurred, but whether they might have drawn large parts of Europe into a very large caliphate. The insuperable structural divisions that steadily undermined the unity of the Arab conquest society that had emerged in the mid-seventh century all but rules this out. In the most optimistic scenarios, the best that might have been achieved was the creation of a few additional successor emirates alongside those that popped up in actual history—but not hegemonic empire on anything like the Roman scale.36
Escape From Rome Page 15