12
In June 656, ‘Uthman ibn ‘Affan, commander of the Believers (amr al-mu’minn), deputy of God (khalfat Allh - hence the English title ‘caliph’), was murdered in his house in his capital, Medina in western Arabia. The event convulsed the Arab world, and the First Civil War (fitna) followed, until peace was restored in 661. So much is certain (it was recorded shortly afterwards, although very sketchily, by the Armenian chronicler whom we call Sebeos); the rest is, and was, hotly contested. Was ‘Uthman’s successor ‘Ali (656-61) involved in the murder, as many in the ‘Uthmani party thought, hence the civil war? Was the murder carried out by disaffected pro-‘Ali bedouin extremists against ‘Ali’s will, as one of the earliest Arab historians, Sayf ibn ‘Umar (d. c. 796), claimed? Or was the murder the work of disaffected Egyptian soldiers, tired of ‘Uthman’s attempts to direct the Egyptian grain surplus to Medina and to replace the power of the early Arab conquerors of the provinces from Egypt to Iraq by more traditional tribal leaders - including members of ‘Uthman’s own immediate family, the Umayyads - as other early historians, Ibn Ishaq (d. 767) and especially al-Waqidi (d. 823), report? And, above all, was the murder a justified response to ‘Uthman’s illegitimate acts, which meant that he was no longer properly caliph, or was it illegal, and therefore had to be avenged?
‘Ali may have thought the first of these latter two alternatives. Certainly the later Shi‘ite tradition did - indeed, that tradition thought that ‘Uthman, and maybe his two predecessors, were usurpers, and that ‘Ali had been designated the Prophet’s successor by Muhammad himself at his death in 632. The ‘Uthmanis certainly thought the second, not least Mu‘awiya ibn Abi Sufyan, governor of Syria and ‘Uthman’s second cousin, thus also an Umayyad. Mu‘awiya demanded that ‘Uthman’s murder must be punished and led a Syrian army against ‘Ali’s Iraqi army to Siffin on the Euphrates in 657, where the two sides skirmished for some time. ‘Ali in the end agreed to arbitration on the issue, but lost part of his army - and his strategic advantage - as a result; the dissident group who left him, the Kharijites, were outraged at ‘Ali’s concession, for they thought that only God could judge the issue, not humans. One of them assassinated ‘Ali in 661, after which Mu‘awiya took over as sole caliph (661-80).
So who did kill ‘Uthman, and with what justification? The same question could be asked of many similar deaths in the early Middle Ages, as with Childeric II in Francia in 675, or Edward the Martyr in England in 978. The basic answer is that we do not fully know, and in these two latter cases historians are relatively relaxed about the fact that they do not know; it is enough for them to unpick the different interpretations in the sources so as to identify political alignments. But in the Islamic tradition it was, and is, not so easy. Religious disagreement between Muslim communities tends not to be over the nature of God, as inside early Christianity (a single monotheistic Allah gives less space for debate than the incomprehensible complexities of the Trinity), but rather, much more, over political legitimacy. The basic twenty-first-century division between Sunni and Shi‘a Islam goes straight back to 656, even if the two sides did not call themselves that yet. The Kharijites still exist too, in Sahara oases and Zanzibar, and have not yet forgiven ‘Ali for Siffin. It would even now be hard in the Muslim world to discuss neutrally the behaviour of ‘Uthman or his murderers without taking a position between Sunni and Shi‘a/Kharijite interpretations. And this was even more so around 800, when our first detailed accounts began to be written down, or around 900, when they were collected in the great historical compilations of writers like al-Baladhuri (d. 892) or al-Tabari (d. 923). Even a decision not to be sure who was right in 656 had a doctrinal implication from the eighth century onwards (it was associated with the Murji’ites, the ‘suspenders of judgement’). Indeed, in the ninth century this became common ground in much of what was becoming the majority Sunni tradition, for that tradition held that rulers should not be deposed, and that communal unity was more important than sectarian division (by then, Sunnis accepted both ‘Uthman and ‘Ali as legitimate caliphs; it was Mu‘awiya they had more trouble with). But the whole issue continued to matter, intensely, and all our sources are structured by partisan positions of this kind.
Writing early Arab history is in many ways a harder task than writing the history of other peoples or states in the same period. One reason is the religious importance of every event, as just discussed; this might seem less surprising, perhaps, when one is discussing Muhammad, who was a prophet above anything else, and maybe even his immediate successors, but Arab history right up to 750 has at least in part to be seen through salvationist perspectives. A second is the late date of most of our narrative sources. This ought not to matter too much to early medievalists - mid-seventh-century Byzantine history is mostly accessible only through early ninth-century writers, too, without Byzantinists being more than regretful about it - but the religious importance of the period, and the irreconcilable sectarian positions of our sources, have bothered Arabists much more and have resulted in the rejection by an influential strand of recent historians of all possibility of knowing anything reliable about Muslim Arab history before the 690s at the earliest. It is also the case that, after an absence of narrative sources for the Arabs in the seventh and early eighth centuries, in the ninth and tenth our source material explodes in quantity. There may be as much writing surviving in Arabic (mostly from Iraq) from those two centuries as from the whole of Europe in our whole period. The huge size of this source material, plus the radical nature of recent critiques of it, has led historians of the early caliphate into ever more enclosed discussions of the criteria for its authenticity, and there are remarkably few recent analyses of the details of the period before 750 (or even after it) in itself. The sources themselves are opaque to the inexpert, too; they are frequently made up of quite bitty stories (akhbr), which are given truth-content by chains of informants, maybe going back a couple of centuries, but then often counterposed to other stories that say the exact opposite. One can feel oneself flung into an unfamiliar cultural world, which is further emphasized by the different way in which most historians currently write about it.
And yet the early Arab period is crucial for us to confront. The caliphate did not rule any part of Europe before the Arab-Berber invasion of Spain in 711, but it cannot be excluded from a history of the Continent. For a start, it was the Arabs who broke in half the surviving section of the Roman empire in the seventh century, ending for ever its dream of continued Mediterranean hegemony, and forcing it to reinvent itself as the state we call Byzantium, as we saw in the last chapter. Secondly, the caliphate was itself built on Roman foundations (as also Sassanian Persian foundations). Notwithstanding the difficulty and unfamiliarity of our narrative sources for it, it arguably preserved the parameters of imperial Roman society more completely than any other part of the post-Roman world, at least in the period up to 750; this is a paradox which it is essential to explore. Thirdly, the caliphate was simply richer and more powerful than any other post-Roman polity. By now it was the Arabs that dominated the Mediterranean. After 750, under the ‘Abbasids, the centre of the caliphate moved from Mu‘awiya’s Syria to Iraq, and further from Roman traditions; I shall discuss the ‘Abbasids in less detail as a result in Chapter 14. But the ‘Abbasids, even more than the pre-750 Umayyads, far surpassed their neighbours in their wealth and in the sophistication of their intellectual culture, and we must pay attention to that, both in Chapter 14 and in 15, when we look at the east Mediterranean economy as a whole. This chapter will discuss the Arab conquests and the Umayyad caliphate of Mu‘awiya and his successors. Here, we shall focus on the linked problems of the stabilization of the Arab (or Muslim) political system, and of the issue of social and cultural continuity and change, in the first of the many centuries of Arab dominance of the eastern and southern Mediterranean, and, indeed, of further afield.
Muhammad (c. 570-632) was a merchant in Mecca in western Arabia who around 610 began to get verbal revelations from God; he became a prophet and sought followers. The Arabs were polytheists, although there were substantial Christian and Jewish minorities among them. Muhammad was certainly closest to the Jewish tradition, and was, like the Jews, a very strict monotheist, but the most reliable early Muslim source (the Constitution of Medina, dating to the 620s) makes it clear that Muhammad saw the Believers (the commonest early word for his followers) as separate from Jews. Muhammad’s revelations were later collected as the Qur’an; Muslim tradition says that the basic recension of the text dates to ‘Uthman’s reign as caliph (644-56). Some recent western scholars have argued for a much later date, as late as 800 for John Wansbrough, the early eighth century for Patricia Crone, though Fred Donner makes a good defence for the traditional dating on grounds of content and style. However this may be, it is undeniable and important that elements of the Qur’an were already widely circulating in the late seventh century, as can be seen in the Qur’anic verses prominently displayed on the Dome of the Rock in Jerusalem, the first major piece of Islamic architecture, finished in 691-2, and that they clearly name Islam as a distinct and coherent religion, founded by Muhammad. The exact details of that religion in its first decades cannot be fully reconstructed, and will of course anyway have been understood differently by different people, but it was recognized as new and challenging right from the start. The Meccans were sufficiently threatened by it for Muhammad to have to flee to Medina, a more welcoming town, in 622; his ‘emigration’ there (hijra) marks the formal beginning of Islam, and Hijra dating for the years appears on an Egyptian papyrus as early as 643. Medina and Mecca fought for supremacy across the 620s, but Muhammad took his home town over in, probably, 630; it became the religious centre of Islam, although Medina remained the political centre. Muhammad also, remarkably quickly, extended his authority widely in Arabia, even before the fall of Mecca, and especially afterwards. By Muhammad’s death, the fragmented and warring Arab tribes for the most part recognized a single authority for the first time ever, and an authority of a new type, based on religious revelation.
The first caliphs, Abu Bakr (632-4), ‘Umar I (634-44), ‘Uthman and ‘Ali, were chosen from among Muhammad’s own close companions, and also, as all successive caliphs were too, from Muhammad’s own tribe, the Quraysh; the title khalfa may have existed from at least ‘Uthman’s time. They were both Muhammad’s successors as leaders of the Believers, and God’s representatives on earth. From the beginning, it was seen as essential to keep the Muslim community together, and Abu Bakr’s caliphate was taken up with subduing Arabia more fully; much of it, particularly in the east, had seceded after Muhammad’s death (partly under its own prophets - Muhammad was by no means the only one in the period) and had to be reconquered. This was doubtless easier because of the collapse of Sassanian hegemony along the east and south coasts of the peninsula after Heraclius’ victory over Persia in 628, for, of course, these Arabian events were happening just as the last great Rome-Persia war ended with the exhaustion of both sides and the prostration of the loser. ‘Umar sent Arab armies north in the 630s; after the defeat of the Romans on the River Yarmuk in 636, all Syria and Palestine was laid open to Arab conquest; after the defeat of the Persians at Qadisiyya in about 637, so was what later became called Iraq. By 640 the whole of the fertile crescent south of the Turkish mountains was under Arab control; Egypt was added in 639-42. Iraq and Egypt were thereafter always the economic powerhouses of the caliphate; except for the brief period of Sassanian rule in the 620s, they had not been part of the same political system since the death of Alexander nearly a thousand years before. In the 640s the Arabs took Iran as well; by the time the last shah, Yazdagird III, was killed in 651, Arab armies had reached the modern Iran-Afghanistan border. The conquests stopped here for a generation; but in fifteen years the whole of the Sassanian empire and half the eastern Roman empire had been conquered by the Arabs. Only Alexander, and the Mongols, have ever matched them for speed of conquest, and both Alexander’s empire and that of the Mongols soon broke up into their constituent pieces again. The Arabs, however, kept these territories together for three centuries, and their religion and culture have dominated there ever since.
The Arabs were able fighters, and both the Romans and Persians had long used them as mercenaries, the Ghassanid confederation by the Romans, the Lakhmids for the Persians. Given the exhaustion of the empires in the 630s, and the new-found religious unity of the Arabs, Arab victories and conquests are not in themselves extraordinary, and of course after the first great battles were won in 636-7 every Arab with a camel was likely to want to take part in the conquests and in the wealth they brought. What was less to be expected was that the conquests would hold together. There can never have been many Arabs; Yemen is the only substantial part of the Arabian peninsula that can sustain more than a scattered, largely pastoralist, bedouin population, and even the Arabs who had for long settled in Roman-ruled Syria and Palestine seem to have been on the desert fringes there too, and therefore not so very numerous. Arab settlers would have been hugely outnumbered everywhere, and might not have withstood sustained revolts or Roman counterattack if their unity faltered; alternatively, they were at risk of being absorbed into local populations and losing their cultural identity, as Germanic ruling groups were in every Romano-Germanic polity except England and Bavaria. The absence of early revolts was fortunate (they would have been most dangerous in Iran, where the Persian aristocracy was a military one and early Arab settlement was more or less confined to Khurasan in the north-east); as for the surviving Romans/ Byzantines, they were in no military shape to take advantage of Arab civil wars. But the core reason for the survival of Arab rule as not only a political but also a cultural hegemony was not luck. Rather, it was the result of the decision (traditionally, and plausibly, ascribed to ‘Umar I in 640-42) to settle the Arab armies, not as a landowning aristocracy as in the Germanic West, but as paid garrisons in newly founded cities (amsr), Kufa and Basra in Iraq, Mosul on the edge of the Iraq-Syria borderlands known as the Jazira, Fustat (the future Cairo) in Egypt, and others. The tax revenues of the provinces went to these garrisons above all, who thus were well rewarded for their separation from the socio-political life of the conquered population; being on the local dwn, the register of those entitled to army pay, was a coveted privilege, defended against newcomers as much as possible. ‘Umar’s policy succeeded; relatively little Arab landowning is recorded for any of these core provinces before 750 (although it seems to have been greater in Khurasan, where indeed Arab settlers were eventually Persianized, and also in the later conquest territories of Africa and Spain). This set the template for a structural separation between a paid army and the rest of civil - civilian - society which was greater even than in the Roman empire, and which marked most Muslim political systems ever after.
This decision had several consequences. One was fiscal: the tax system of the Roman - and also Sassanian - empire never broke down, as it did in the West, for it always had an essential political purpose, the payment of a ruling army. Another was, as already implied, that the Arabs were preserved as a separate and superior social stratum. They intermarried with local families, but their children maintained an Arabic language, culture, religion, identity. And they were so separate that anyone from the conquered majority who sought political prominence would have to try to join them, both in culture and in religion. This was less true of the seventh century, when the Arabs discouraged conversion to Islam, and anyway maintained the provincial governments of the conquered provinces, both Roman and Persian, intact. It was possible for two generations after the conquests to be powerful in the civil administration without changing one’s culture or identity at all, as with the Mansur family, prominent Greek Christian administrators in the Umayyad capital, Damascus, into the early eighth century, one of whose members was the important Christian theologian John of Damascus (d. c. 750). But around 700 the basic language of administration was changed to Arabic; from then on bureaucrats would have to be Arabic-speaking and, increasingly, Muslim. The process of conversion, at least for local élites, was indeed seen as an Arabization process; one had to become the client (mawl, plural mawl) of the tribe of an Arab sponsor, and, usually, to change one’s name to an Arab one. Such people ‘became’ Arabs, with access to political power, and perpetuated Arabic language and culture as they did so. Peasant conversion (which existed from the start; Muslims paid lower taxes, at least in theory) did not ever bring political privilege, but very slowly the links of Muslim clientage extended outwards to the peasantry too, and Islamization/Arabization permeated the countryside as it did so. This process was not a large-scale one until the ninth century at the earliest, but it was steady from then on, and by 1000 the majority of the population from Egypt to Iraq probably spoke Arabic. Of the conquest lands, only Iran maintained its original Persian language, by now however written in Arabic script and full of Arab loanwords.
This early separation between Arab élites and the conquered majority also meant that Roman society and Persian society persisted, remarkably unchanged, into the late eighth century and often later. Egyptian documents show that the cities of the Nile valley remained governed by their traditional élites until past 700; all that the Arabization of the administration meant initially was that Greek was used less and less; most of the population continued to speak and write Coptic. Nor was this process instantaneous; we have some two hundred administrative letters (mostly about taxation) from the governor of Egypt, Qurra ibn Sharik (709-15), to Basilios, pagarch or city governor of the small middle Nile city of Aphrodito, modern Kom Ishqaw, and these are for the most part still either in Greek or bilingual in Greek and Arabic. From this point on, pagarchs would be Muslims, with Arabic names; any local family which wanted to continue to control its city would by 730 or so have to convert. Villages were less affected, all the same; throughout the eighth century Coptic overwhelmingly dominates in our village archives, and Arabic is not prominent except in governmental texts until the ninth. Mosques do not appear in our documents either; rural religion was essentially Christian throughout this period. It is possible as a result to write Egyptian social history up to 800 almost without reference to the Arabs at all, for they were so much shut away in Fustat. This would be a mistake, but it is a tempting one.
Syria and Palestine, the other major ex-Roman provinces, show a more nuanced picture, but a similar one. There were always more Arabs in the Levant, from well before Muhammad’s time; some of the most powerful Umayyad-period Arab tribes, notably the Kalb, originated from the Syrian desert fringe. Probably as a result of this long-standing tradition, there were no important amsr in the region; the Arab army of Syria settled in the already existing cities of the Roman empire, less separate from the native population than they were elsewhere. And Damascus became, from Mu‘awiya’s reign onwards, the capital of the caliphate, replacing Medina; Syria was thus the core province of Umayyad government. One might have expected an early Arabization of the Levant as a result of all these factors. But there are remarkably few signs of it. Damascus probably slowly became Arabized once the administrative language changed to Arabic (evidence for the capital is unfortunately not good), but Edessa, at least, certainly did not; its rich Christian written tradition shows a strong and prosperous Syriac-speaking urban élite until well into the ninth century. In the countryside, Nessana in the Negev desert, which has preserved a papyrus archive into the 680s, has hardly any Arabic documents, even though a substantial proportion of its population were ethnic Arabs, and even though one text in Greek seems actually to be a page of a dwn register. (On the other hand, Khirbat al-Mird, in the desert west of the Dead Sea, was already Arabic-speaking in the late seventh century, as a smaller papyrus collection shows.) And the extensive urban and rural archaeology of both Syria and Palestine shows notable continuities; indeed, the Arab conquest is hardly visible in it at all. There were certainly new Arab administrative and religious buildings put up across the region in the next century, but cityscapes were slow to change; and churches were still being built in cities and the countryside into the late eighth century in what is now northern Jordan and elsewhere. The economic implications of this we shall look at in Chapter 15, but the cultural templates of late Roman urban and rural life were as yet unchanged, even in the Umayyad heartland. The ambitious monuments of the Umayyads themselves, which we shall come to shortly, were only an overlay onto these essential continuities. Here, as in Egypt, wider cultural change only began after 750, and maybe later still.
The trouble with this cultural separation, between Arabs and local populations, was that the age-old patronage links between central and local power were cut, particularly once the administration went Arabic. Local power-brokers could hope to deal with central government in the seventh century, as it still spoke their language; one of the Nessana papyri from the 680s shows a local notable, Lord (kyrios) Samuel, organizing village representatives to go to protest to the governor in Gaza about the provincial tax burden. (The governor was certainly an Arab, but he too wrote in Greek for the most part.) In the eighth century, such power-brokers had to choose: whether to stay Christian with their clients and lose purchase with the administration, or to become Arabized mawl, and thus part of government, but risk losing their local links. In Egypt, the latter choice was rare still, and the eighth century saw tensions rising. Tax revolts began in Egypt in 726, and continued on and off for over a century, with particularly serious uprisings in 750, the year of the Umayyad fall, and 812-32. Arab taxation was not obviously heavier than Roman taxation had been, but Egyptian civil society was too cut off from the Arab military élite, and violent resistance resulted. Arab political power was too entrenched by 750, however, to be structurally threatened by this; and the Arabization of the countryside, which had begun by 832, meant that stronger patronage chains could emerge again.
The Arab/non-Arab cultural separation was nonetheless incomplete, for one crucial reason: Islam itself had emerged from the world of late Roman (largely Jewish) religiosity, and had little difficulty in relating to many aspects of the religious landscape it found in the conquered provinces. This is clearest in the least formalized aspects of religion, those least tied up in political power and legitimacy; several early Muslim accounts claim that Muhammad was recognized and respected by Christian holy men, for example, most notably the Syrian hermit Bahira, who turns up in some Christian sources too. Muslims also respected both Jewish and Christian holy places, Jerusalem most notably (which they sought to appropriate), but also Mount Sinai, location of both Christian and Muslim pilgrimage. Perhaps the best example of this is the Umayyad interest in the cult-site of St Sergios at Sergiopolis, in Arabic Rusafa, in the east Syrian steppe south of the Euphrates. In the decades around 500 this was the location of some highly ambitious imperial church building for the pilgrimage centre Sergiopolis had become; it was also situated in a Christian Arab area, and the Ghassanids linked themselves in the sixth century to St Sergios in general and to Rusafa in particular. It is therefore significant that Rusafa was also the caliph Hisham’s favourite country residence in the 730s; he built a mosque there right beside, indeed sharing a courtyard with, one of the major churches of the city, and also a set of shops around the precinct (Hisham was a patron of monumental shop complexes elsewhere, too: see Chapter 15). The caliph was clearly reacting to - indeed, respecting - the religious importance of the place, even though that importance was essentially and traditionally Christian. Rusafa was a Muslim political centre for only two decades at the most, but Sarjis, that is, Sergios, turned into a Muslim holy man in at least some parts of the Arab world in centuries to come. In places like Rusafa, both conquerors and conquered could meet as, in religious terms, some kind of equals.
‘Umar I’s reign was marked by war, and, apart from the establishment of the dwn system, it was not a period of wider-scale state formation. When the first wave of conquests stopped around 651, ‘Uthman found that one danger was that the new provinces risked drifting apart under their new Arab military élites. It is not clear whether under ‘Umar the provinces sent any of their tax revenue back to Medina, but all sources agree that ‘Uthman laid claim to at least some of them, particularly from the agriculturally rich provinces of Egypt and Iraq. ‘Uthman’s equally controversial patronage of kinsmen and tribal leaders as governors, instead of the early Muslims, often of no particular tribal status, who dominated the garrison towns, can be interpreted as the caliph trying to ensure chains of loyalty to him that would stabilize the new Arab political system. Both of these policies aimed to centralize power, and it is likely enough that it was indeed these policies that led to his death in 656. But it was his kinsman Mu‘awiya who won the First Civil War, and Mu‘awiya certainly continued them; he appointed his adopted brother Ziyad (d. 673) to govern Iraq and Iran, for example, and inside Syria linked himself closely to the tribal confederacy dominated by the Kalb, which was the main Arab group in the province. (It is less certain how far he managed to divert provincial revenues to Syria, however; his centralizing practices were above all personal.) Mu‘awiya clearly thought dynastically, and ensured that his son (by a Kalbi mother) Yazid I (680-83) would succeed him. This led at his death to a far more serious rerun of 656-61, the Second Civil War of 680-92.
‘Ali’s son al-Husayn was the first to revolt against Yazid, in 680; he was killed at Karbala’ in Iraq in a one-sided conflict that has resonated ever since in Shi‘a martyrology. In Medina, ‘Abd Allah ibn al-Zubayr, son of another First Civil War leader, also rejected Yazid’s authority, and he established himself as caliph there and in Mecca (683-92), with quite a wide authority for some years. Ibn al-Zubayr was not very militarily active himself, but he had substantial support both in Iraq and in parts of Syria. After Yazid’s death, Kufa, too, revolted under the ‘Alid leader Mukhtar, and was effectively independent in 685-7. And in Syria itself the leading Arab tribes fell out, the Kalb being opposed by the Qays, a coalition of newer settlers from northern Arabia, based in northern Syria and the Jazira, supporters of Ibn al-Zubayr. The Kalb put in a new branch of the Umayyad family as caliphs to confront Ibn al-Zubayr, Marwan I (684-5) and his son ‘Abd al-Malik (685-705), the first Marwanids, and Marwan defeated the Qays at the battle of Marj Rahit north of Damascus in 684. Even then, everything risked breaking up, but ‘Abd al-Malik held on, carried on fighting, and re-established unity with the reconquest of Mecca and the death of Ibn al-Zubayr in 692. What was clear, however, was that he needed a new and more stable political settlement, to avoid renewed chaos leading to the end of Arab rule.
With ‘Abd al-Malik our historical information begins to be rather more reliable and diversified, and we can be more confident in our reconstructions. One thing he did was return to conquest. Westwards from Egypt, Arab armies had rather desultorily moved into the southern parts of Byzantine Africa in the 640s and then the 670s (founding the garrison city of Kairouan in 670); in the late 690s, however, they defeated the powerful Berber tribes of the Algerian plateau, and conquered Africa definitively, taking Carthage in 698. The Berbers took to Arab rule very fast. In 711, under ‘Abd al-Malik’s son al-Walid I, a Berber and Arab army invaded Spain, and by the end of the 710s it controlled nearly all the Iberian peninsula and was raiding into Francia. To the east, Bukhara and Samarkand fell in 706-12, and the Arabs occupied central Asia, and also parts of north-west India. The scene was set for the greatest conquest of all, Constantinople, with the siege of 717-18 led by Maslama, son of ‘Abd al-Malik, although this failed; it turned out that the caliphate had reached its greatest extent under al-Walid, and border wars would be the norm thereafter. These new conquests did not have the economic and political importance of those of 636-51, but they kept the main provincial armies busy and rich, which was better than civil war.
‘Abd al-Malik also ruled the provinces as forcefully as he could. Egypt was entrusted to his brother ‘Abd al-Aziz (d. 704), and shortly after that to the Qaysi governor Qurra, whose surviving letters show him to be very effective in his exactions and his local control. We still cannot see Egyptian wealth going to Syria, and these governors were probably as rich as the caliphs themselves, but they were certainly loyal. Iraq, the most troublesome province for the early Umayyads, was in 694 assigned to the hyper-loyal al-Hajjaj ibn Yusuf, another Qaysi, who governed it (and, after 697, Khurasan as well) until his death in 714; al-Hajjaj was a very tough, not to say oppressive, ruler who provoked a civil war with the Kufans in 701 and established a Syrian army in the zone after that; Iraqi armies withered, and Iraqi taxes went to Damascus from then on. In Syria, ‘Abd al-Malik maintained a balance between Kalbi and Qaysi patronage networks, as these Qaysi governors already imply. The two opposing networks gained in force, all the same; the Kalb joined with immigrants from Yemen who had settled in central Syria, and the alliance is generally from now on called Yamani in our sources; the two networks, which came to include virtually all Arabs, were fierce rivals for patronage from the caliphs, particularly the highly lucrative position of governor. A Yamani or a Qaysi governor could be relied on to appoint only members of his own faction to subordinate posts, but the caliphs themselves were for a long time fairly neutral between the two major groupings.
‘Abd al-Malik established a new public prominence both for Arab culture and for Islam. He Arabized the civil administration, as we have seen. That administration gained ever greater coherence, as is visible, for example, in the highly polished state letters of the senior chancery administrator ‘Abd al-Hamid (dating 725-50), which prefigure the belles-lettrist adab style of the ninth to eleventh centuries, as also the highly literary Byzantine practices of the same period, both discussed later. ‘Abd al-Malik furthermore, for the first time, instituted a coinage that reflected caliphal political power. Previously, Arab coins had imitated Byzantine and Persian models, but in 691-2 new standard-weight coins came in, the gold dnr in the ex-Roman lands and in ex-Persian lands the silverdirham, which had Arabic and Islamic inscriptions, and which after 696 abandoned images for purely verbal decoration. The caliph also, already during the Second Civil War, inaugurated expensive prestige buildings, beginning with the Dome of the Rock, on the spot to which Muhammad reputedly miraculously travelled for a night from Mecca, on top of the old Jewish cult centre of the Temple Mount in Jerusalem, finished in 691-2; this was followed under al-Walid by the neighbouring al-Aqsa mosque in Jerusalem (709-15), the Great Mosque in Medina (706-10), and the huge Great Mosque of Damascus (705-16), which largely survives in its original form, decorated with mosaics, as we have seen. These and other projects were by far the largest-scale buildings in Eurasia west of China in this period, and they all explicitly celebrated a triumphal and rich Islam. They show, it must be added, that some money at least was by now getting to Syria from the provinces. The Umayyads were also giving a Muslim education to their children; one sign is the religious austerity of ‘Umar II ibn ‘Abd al-Aziz (717-20), who alone out of the Umayyad caliphs was regarded as a just ruler by later generations. This austerity was not continued by his successors, but by now the stability of the regime was more assured, as the long and relatively peaceful rule of Hisham, the last son of ‘Abd al-Malik (724-43), shows.
The Umayyads had a terrible press after their fall in 750. They were seen as dynastic rather than ruling by consensus (though the ‘Abbasids would be just as dynastic as they); and as luxurious degenerates, enjoying themselves in their palaces and ignoring the needs of government. They certainly built luxurious palaces; some of them survive, in the Jordan valley and on the Syrian/Jordanian desert fringe, as ambitious in their own way as al-Walid’s mosques, and in two cases (the stuccoes of Khirbat al-Mafjar outside Jericho, the frescoes of the Qusayr ‘Amra bath-house east of ‘Amman) they show a profusion of human forms (often naked and female) that do not look very ‘Islamic’. This represents a private decorative tradition that would have a long future in Muslim societies, all the same, rather than indicating that its Umayyad sponsors had not read the Qur’an properly. (Actually, the Qur’an only opposes idol-worship, not all figurative representations of humans; but a caution about public representational art was certainly already accepted by the Umayyad caliphs, as we saw in Chapter 10, for the outsides of these palaces, often heavily carved in high relief, were entirely geometric and non-figurative, just as the mosque of Damascus was.) Several of the Umayyads did indeed have imaginative personal lives, too; but so have rulers throughout history - including, once again, the ‘Abbasids - without this impacting very greatly on their conceptions of rule. These accusations are simply a damnatio memoriae, like the later Byzantine attacks on Constantine V, rather than an accurate critique of Umayyad government.
The critique of the Umayyads which had the strongest resonance was that they were Arab, not Muslim, rulers. It has lasted ever since, too: even Julius Wellhausen, the great late nineteenth-century historian of the Umayyads, called their realm the ‘Arab kingdom’. It is a particularly false claim. For a start, the Umayyad caliphs took their religious responsibilities very seriously, at least from ‘Abd al-Malik onwards (Mu‘awiya is a rather more shadowy figure). ‘Umar II issued highly religious edicts, and was by no means the only caliph to do so. We have one from al-Walid II (743-4), later considered the dynasty’s most notorious playboy, which is adamant about the religious duties entrusted to him by God. These include the enforcement of religious obedience, the pursuance of ‘that which is most righteous for him in particular and for the Muslims in general’, and, overall, ‘the completion of Islam’; with a few phrases changed, this could be Charlemagne at his most moralizing. Similarly, his cousin and supplanter Yazid III (744) justified his uprising against al-Walid in exclusively Muslim terms. These caliphs indeed felt their religious role more strongly than did the ‘Abbasids, after the fervour of the first ‘Abbasid generation at least, for by the end of the eighth century the task of interpreting religious authority had mostly fallen to a new social group, the ‘ulam’ of scholars (see below, Chapter 14).
It has also been proposed that Umayyad Islam was more ‘Arabic’ than later, universalist, Islam would be. Was Muhammad a prophet only for the Arabs, or for everyone? It has been argued that the early Arab caution about conversion implied the former, and that only the ‘Abbasids really opened their religion to all comers. This, too, is largely an overstated reading. The Arabs undoubtedly believed in their own ethnic superiority, and were at best edgy, at worst hostile, to non-Arabs, including converts. Qusayr ‘Amra also includes a famous fresco of six kings, of the Roman empire, Sassanian Persia, Ethiopia, Visigothic Spain, and two unidentified countries, apparently gesturing to an adjoining fresco of ethnic Arab victory. But conversion was nonetheless seen as normal, and plenty of mawl reached high positions under the Umayyads, notably Musa ibn Nusayr (d. 716), one of the conquerors of Spain, and several later governors of Africa. Al-Hajjaj, the emblematic Umayyad devotee, himself appointed a black African, Sa‘id ibn Jubayr (d. 713), to the post of q (judge) of Kufa, even if he had to rescind the appointment because the Kufans protested against a mawl holding the role. There was, of course, a contradiction between Arab exclusive-ness and Muslim inclusiveness, but it was felt by every Arab, from caliph to foot soldier, until conversion became widespread, for different reasons, in the ninth century; it was also not just a matter of Arab vs. non-Arab, but settled Arab vs. bedouin Arab (each claimed to be the better Muslims), and of course tribe vs. tribe. Arab tribalism had by now little of the desert about it, it can be added; the huge majority of Arabs by 700 lived settled lives and were just competing for military and civilian positions. Their desire to secure such positions for themselves and their families and allies, rather than their rivals, led to tribal-and ethnic-exclusivist actions and rhetoric, but this is true of any society, and would not cease in 750.
An example of this mixture of positions in a single person is the poet al-Farazdaq (d. c. 729); he may have had bedouin origins, but he lived most of his long life in Basra. His poetic palette of camels, gazelles, tents and cavalry warfare was more the standard rhetoric of any Arab poet than nostalgia for the desert. So were his attacks on the honour and sexual morality of people (usually poets) from rival tribes, and his complex love poetry. Al-Farazdaq was Arab through and through; he loathed having to go to ‘an odious land, the country of the blond-haired Greeks of ‘Amman’. But when he wrote eulogies to the caliphs (some fifteen survive, for every caliph from ‘Abd al-Malik to Hisham) his imagery turns Muslim: ‘Run to Islam, justice has returned to us, the scourge which desolated Iraq is dead, there are no more poor on the earth, Sulayman [caliph 715-17] is the treasure of the universe.’ This is not in the least surprising, and indeed precisely recalls the mixed values that any early medieval Christian writer had, western or eastern, as with the glorification of Frankish ethnic and military superiority in Gregory of Tours or Einhard, Christian inclusivism notwithstanding, or indeed the ferocious hostility to Goths of their fellow-Christian Synesios at the start of the fifth century. It is not religious and moral inconsistency that made the Arabs different in our period.
Hisham was the first caliph to face the problems of a no longer expanding frontier. Instead, Khazars and Turks themselves invaded from the north, and were beaten back with some difficulty in the 730s (in the case of the Khazars, by Marwan ibn Muhammad, an able general from the Umayyad family, who became governor of the Jazira). In the far west, too, there was a major Berber revolt in 740-43 which cut off Umayyad access to Spain and even Africa. But these only look like signs of Umayyad collapse in retrospect; they were all dealt with before Hisham’s death. More serious was his famous tight-fistedness with money, for this is a sign that the caliphs had not solved the problem of tax money staying in the provinces it had been collected from. Not only Iraqis but also Egyptian Arabs had lost their military role by now, and the late Umayyad army was overwhelmingly Syrian except in the Berber lands of the far west and in Khurasan in the far east, but this did not lead to any further organizational centralization. Yazid III, indeed, promised not to move tax money outside provinces in his 744 rebel manifesto. At Hisham’s death, furthermore, serious problems did appear, for the Syrian army broke up into Yamani and Qaysi factions. Al-Walid II was not necessarily pro-Qaysi, but Yazid III’s revolt certainly had essentially Yamani support; Marwan in the Jazira, who sought to avenge the murdered al-Walid, ruled the Qaysi province par excellence and recruited a Qaysi army. Yazid died suddenly after a few months, and Marwan replaced him as Marwan II (744-50), but the latter had to spend two years reducing Yamani resistance in Syria, the first time the core caliphal province had ever been under attack.
The years 744-6 are seen as the Third Civil War; this time, unlike the seventh-century fitnas, overall Arab rule was too established to be in danger. But Umayyad rule was another matter. There were Shi‘ite and even Kharijite revolts, with Yamani support, in Iraq in 744-8 too; these were easy enough to confront, as Iraq no longer had an army of its own, but their appearance is a sign of a loss of confidence in the ruling dynasty. And events in Khurasan, where the main eastern army was situated, were even more serious. It emerged that Shi‘ite groups had been quietly preaching revolution there for three decades in favour of the Hashimiyya, the branch of Quraysh that was Muhammad’s immediate family. The Hashimiyya included the descendants of ‘Ali, of course; but they also included the descendants of ‘Abbas, the Prophet’s uncle. In 747 one of the sectarians, Abu Muslim, urged open revolt outside Merv in eastern Khurasan, and very quickly this revolt snowballed to include almost the whole of the Khurasani army. Abu Muslim and his associates chose ‘Abbasids, not ‘Alids, as their religious figureheads, and Abu al-‘Abbas was proclaimed caliph as al-Saffah in 749. The Khurasanis moved westward and defeated Marwan in northern Iraq in 750, then took over Syria and Egypt in the same year, where Marwan was killed. The ‘Abbasid caliphate began here; and when al-Saffah died in 754, his brother Abu Ja‘far, al-Mansur (754-75), soon executed Abu Muslim and took full power for himself. The new regime ended (or at least marginalized) the Qays-Yaman feud, largely because it mattered less in a Khurasani army which was substantially non-Arab; although the ‘Abbasids certainly made full use of Yamani support, they made peace with the Qaysis as soon as they could. The fact that they conquered all the provinces and could thus begin from scratch also allowed them to end the fiscal exclusivity of each provincial dwn. They did not base themselves in Khurasan, however, even though it was their main military support. They chose Iraq, which became the new caliphal province. It was central; it was also the archetypal non-Syrian province. Syria, laid waste by Marwan in 744-6 and again by Abu Muslim in 750 - as well as by a severe earthquake, probably in 749 - became a province like any other, and politically suspect as well. Al-Mansur’s new capital of Baghdad, founded in 762, soon surpassed anything Damascus had ever been, and the style of the caliphate decisively changed.
The Umayyads largely fell because the dominant Syrian army split, losing them both military superiority and hegemony, the sense that their rule was inevitable. This allowed the sort of millenarian Shi‘ism that had fuelled Mukhtar in Kufa in the 680s, and also lesser rebels in subsequent decades, to gain more support than ever before, in the heartland of Islam’s second major army, that of Khurasan. (The third army, that of the Berbers, went its own way.) Abu Muslim was himself a mawl, and he had considerable ethnic Persian support in the Khurasani army. As a result, it was then, and has been since, possible to see the Hashimiyya rising as the rejection of particularist Arab rule by a new Muslim community, based on a rate of conversion to Islam that was higher in Khurasan than anywhere else. But the other elements of the rising were entirely Arab, and they drew their support from the opposite source, the resentment of Yamani Arab soldiers, and of Arab settlers in the east who had been subjected to the local rule - and taxation - of Islamized Persian élites. It is at least clear that the breakdown of Umayyad consensus in Khurasan was the result of an interaction, much greater than elsewhere, and highly tense as well, between Arab settlers and the indigenous majority. This might have broken down into local civil war; but the Shi‘ites managed to convert this tension into a salvation-based unity that overturned the political system. The salvationism was an illusion, and religious revolts (all by now ‘Alid) dotted the ‘Abbasid caliphate, as it had that of the Umayyads before them. But the political direction of a caliphate now rooted in Iraq would be quite different all the same.