14
The Arab geographer Ibn Hawqal (d. c. 990) hated Palermo and the Sicilians. Palermo itself, conquered by the Arabs from the Byzantines in 831, was rich and impressive, and Ibn Hawqal spends many pages on its amenities: the large mosque (the ex-cathedral) which could contain 7,000 people; more than 300 other mosques, in an unparalleled density, sometimes actually adjoining each other; the very numerous and varied markets; the specialized papyrus production, the only one existing outside Egypt; the richly irrigated gardens surrounding. But the Palermitans wasted this latter fertility on cultivating onions, which they ate raw; the consequence was that ‘one does not find in this town any intelligent person, or skilful, or really competent in any scientific discipline, or animated by noble or religious feeling’. No one was qualified to be (judge) there; they were all too unreliable. Schoolmasters were very numerous, but all idiots: they did the job in order to avoid military service; nevertheless, the Sicilians as a whole considered them to be brilliant. They pronounced Arabic wrong; they could not hold down a logical argument (Ibn Hawqal provides examples); they had no idea of what Iraqi legal and theological schools really believed, ‘even though their doctrinal position is very well known’. Nor did the Sicilians know Islamic law properly, particularly in the countryside. Ibn Hawqal was so incensed about all this that he actually wrote a whole book about Sicilian idiocy, unfortunately lost; but he tells us quite enough in his huge geographical survey, The Book of the Depiction of the Earth. He ends amazed that the Sicilians could be so poor, at least these days (in the 970s), when their land was so rich. The only thing they made really well was linen.
What the Sicilians had done to make Ibn Hawqal so cross (geographers often criticized the inhabitants of regions, but this is extreme) is not easy to see. But it is fair to say that he knew what he was talking about. He was born in Nisibis in the upper Tigris valley and was brought up in Baghdad; he left the latter city in 943 for thirty years of travel, to North Africa, Spain, Armenia, Fars and Khurasan in what is now Iran, back to Mesopotamia and Syria, Egypt, and finally to Sicily. He may by now simply have been tired and grumpy, but he had traversed the whole Islamic world. He saw it as a whole, and constantly compared its parts; the great city of Fustat in Egypt, for example, had a third of the surface area of Baghdad, whereas Córdoba in Spain had almost half; the nougat of Manbij in northern Syria was the best he knew except for that of Bukhara in central Asia; the merchants of Sijilmasa in the Moroccan desert were so rich that people in Iraq or Khurasan hardly believed Ibn Hawqal when he told them how much they were worth. Ibn Hawqal made these journeys, however, when the Islamic world was divided into between ten and fifteen separate polities. This hardly poses him a problem; rulers appear casually in his account, some good, most bad, some sufficiently threatening that he had to leave quickly, but all of them simply controlling sections of a single Muslim community. Ibn Hawqal’s geography transcended politics; he, and other geographers like him, saw the Islamic world as essentially a whole.
This cultural and religious unity was first established by the military conquests of the Umayyads. It was made permanent, however, in the century and a half of the ‘Abbasid caliphate, which was politically hegemonic as a centralized state between 750 and 861, and still powerful until around 920; the disunity of Ibn Hawqal’s time (and ever after) was hardly a generation old when he set out from Baghdad. In this chapter, we shall look at the ‘Abbasid achievement, in the decades of their most effective political centralization, and in the creation of a dense religious and scientific written culture in Baghdad, which was strong enough to survive tenth-century fragmentation. We shall then follow the history of two of the successor states, those closest to the European focus of this book, the Fatimids of North Africa and Egypt, and, in particular, the Umayyads of Spain. The Spanish Umayyads were autonomous under ‘Abbasid power already in the 750s, but they too looked to Baghdad for a long time. Baghdad, although by no means part of a history of Europe, or even of the former Roman world, had an economic and cultural importance in the last third of our period that outclassed anywhere in the world, and that certainly impacted on Europe: on Spain, on Constantinople, and even on far-off Aachen, where Charlemagne’s court paid attention to that of Harun al-Rashid, even if the reverse was probably not the case.
In Chapter 12, we left al-Mansur, the second ‘Abbasid caliph (754-75), in control of the whole of the Muslim lands from North Africa to what is now Pakistan. This control was not simply the result of the ‘Abbasid ‘revolution’ of 747-50; the political system was not yet stable in 754, and al-Mansur, in order to feel secure, had to defeat rivals from inside his immediate family and also a serious ‘Alid revolt in 762-3, as well as establishing a balance of power between the Khurasani army which had brought the ‘Abbasids victory and the Iraqi and Syrian factions they displaced. This political settlement was a success, however, the product of al-Mansur’s brilliance as an operator, buttressed by his famed religious austerity and financial caution. It was crystallized in the foundation of a new capital at Baghdad in 762, focused on a monumental round city (no longer surviving), which was the political and ceremonial centre of the caliphate: Baghdad was to be the home of the Khurasani army, the abn’ or ‘sons’, and also of the administrative elite, who came from everywhere in the caliphate, but particularly from Iraq, the ‘Abbasid heartland.
Baghdad seems to have expanded enormously fast; 500,000 inhabitants or upwards seems to me a plausible guess for the ninth century. This was made possible by the water-supply of the Tigris, which runs through it (Damascus has much less water, and had never been anything like so large), as well as by the great agricultural resources of the Jazira between Iraq and Syria and (above all) southern Iraq, the ‘black land’ or Sawad, which were further developed through irrigation projects by the early ‘Abbasids to outstrip the productive wealth even of Egypt. But it was also made possible by ‘Abbasid control, mostly by conquest, of every part of the Islamic world except Spain: al-Mansur had a clean slate, and, after his execution of his great Khurasani general Abu Muslim in 755, owed nothing to anyone. In particular, he could begin the reorganization of the fiscal system that the Umayyads had never managed. The Arabs living in the provinces steadily lost their rights to live off provincial taxation, and it began to flow more consistently to the military and political focus that was Baghdad, a secure resource for the city’s population, whether the soldiers and administrators who were paid by it, or else the mass of shopkeepers, merchants and artisans, and public and private servants, who supplied and depended on them.
That process of fiscal centralization could not be established overnight, of course, given the size and complexity of the caliphate. As we shall see, the 780s-790s and the 830s saw further developments in that direction. But it started with al-Mansur, who already had more resources at his disposal than any previous caliph, or than any Roman emperor since, probably, the fourth century. Al-Mansur can as a result also be seen developing an administrative network that might become capable of organizing and distributing these resources. The Umayyads already had secretaries (kuttb) who had a considerable administrative import- ance, but it is under the early ‘Abbasids that we begin to find them more clearly responsible for separate branches of government or dwns and it is in particular under al-Mansur that we see an executive head of the whole central administrative system appear, the wazr or vizir; the first seems to have been Abu Ayyub (d. 771), who ran al-Mansur’s government for around fifteen years (c. 755-70). The powers of the vizir continued to expand across the ‘Abbasid period, although they were never complete; vizirs did not normally control provincial governors, for that was a caliphal responsibility (although they did control provincial tax officials), and there were always autonomous offices inside Baghdad itself, not least the chamberlain (hjib), who ran the caliph’s large household and often had the caliph’s ear, and who could thus be a serious rival to any vizir. But for the first time we see a clear structure of government in the Arab world, one with its own complex internal politics, as we shall see, made all the more cut-throat by the huge amount of money it had to direct.
Al-Mansur had no doubt as to the dynastic nature of his rule, and, thanks to his removal of rivals, a continuous line of caliphs, all descended from him, held office up to 1517. His son al-Mahdi (775-85) and grandson al-Rashid (786-809) continued his political practices, in a period of general peace and prosperity which aided the trend to centralization. ‘Peace’ is perhaps too bland a term; there were always frontier wars with the Byzantines, and provincial rebellion was far from unknown, particularly in Egypt and in eastern Khurasan, and including a peasant revolt in the Jazira, west of Mosul, in the 770s. But none of them threatened the structure of the state, which continued to develop. Al-Rashid, also known by his birth name of Harun (all ‘Abbasid caliphs had both a birth name and a ruling name, though historians otherwise tend to use only the latter), is by far the best-known ‘Abbasid, and perhaps the best-known medieval Muslim ruler in absolute along with Saladin, thanks to his starring role in the Thousand and One Nights, in its present form a mostly late medieval collection of stories. In his lifetime, however, although an active general, he was a relatively retiring figure in internal politics, devoted largely to ceremonial. Between 786 and 803 the state was dominated by his vizir Yahya ibn Khalid ibn Barmak (d. 805), son of one of al-Mansur’s leading officials, and Harun’s old tutor. Yahya ran the government together with his sons Ja‘far (Harun’s closest friend and associate, both in life and in the Thousand and One Nights) and al-Fadl, who distributed most of the offices of state between them and also a succession of provincial governorships; together they are known as the Barmakids. The Barmakids ever after had a high reputation for being skilled and honest administrators, and they seem indeed to have been so; they were the principal architects of the mature ‘Abbasid fiscal system, bypassing provincial governors (except when they themselves held such offices), and directing ever higher proportions of tax revenue to Baghdad. Their memory was also enhanced by their abrupt fall, when in 803, almost out of the blue, al-Rashid had Ja‘far beheaded and his relatives imprisoned, for no obvious reason except, presumably, his growing resentment of the family’s power. Arab writers pondered for centuries the tragedy of the ideal administrator, Yahya, brought down by an almost-as-ideal caliph - especially as it was only a few years before al-Rashid’s own death ushered in a serious civil war.
It was standard ‘Abbasid practice for rulers to seek to control the succession by naming first and then second heirs; this frequently did not work out, as political alignments changed, but it at least helped to ensure that the initial heir would succeed without opposition from his presumed successor. Al-Rashid went one further: he designated one of his sons, al-Amin, as the next caliph (809-13), and another, al-Ma’mun, as his successor, but he also assigned al-Ma’mun an apanage, Khurasan, in which he was to be effectively autonomous during his brother’s reign. This was probably because Khurasan had become a tense province again, with local aristocracies unwilling to accept the right of Baghdad to take their tax (ironically, to pay the ex-Khurasani abn’ army, in the capital and on the Byzantine frontier); that would cease, at least temporarily, once al-Rashid died, and Khurasanis could feel that they had a future caliph who would safeguard their interests. The tensions did not stop with the division of 809, however, and now each side had an ‘Abbasid at its head. Al-Amin at once tried to undermine his brother’s rule, and the Khurasanis persuaded al-Ma’mun to declare independence in 810. Unexpectedly, his general Tahir ibn al-Husayn defeated al-Amin’s large invading abn’ army in 811, and al-Ma’mun, now claiming the caliphate (811-33), sent Tahir against Baghdad.
Tahir besieged the capital for a year, until he managed to break down local resistance in 813; al-Amin was caught and killed. Al-Ma’mun however stayed in Khurasan, making Merv (now in Turkmenistan) his capital; furthermore, he showed in this period a Shi‘ite commitment, above all through his unique decision to make an ‘Alid his heir in 817, ‘Ali ibn Musa, whose ruling name was to be al-Rida, ‘the chosen one’. This secured the loyalty of parts of Khurasan and Iraq, but alienated the rest of the caliphate. Baghdad revolted again, choosing a brother of al-Rashid, Ibrahim, as the caliph al-Mubarak; Egypt, too, which had had its own civil war between supporters of the rival brothers since 812, fell into chaos in 819 with the most serious tax revolt of the Christian population since 750. Al-Ma’mun had to backtrack, and moved to Baghdad, and definitively away from ‘Alid imagery, in 819. Iraq fell into line straight away and Ibrahim fled (he survived this debacle and was reconciled in 825; he died at court in 839). Egypt, however, took much longer to subdue; al-Ma’mun had to lead an army there himself in 832 to subjugate it properly. Only then, just before the caliph’s death, did he have full control over his father’s domains, with the exception of North Africa, an always rather marginal province, which never returned to ‘Abbasid rule.
The civil war of 811-13 thus unleashed trouble. The resentment of the provinces over taxation was perennial; the more the ‘Abbasids ensured taxes were sent to Iraq, the more acute local resistance would be. In the Umayyad period, this resistance could be posed in terms of loyalty to the person of the caliph (it was just that local Arab armies should have the right to keep provincial taxation); but, if that right was no longer recognized, the risk was that the province would throw off caliphal authority altogether, as first with al-Ma’mun himself in Khurasan. This would indeed eventually lead to the break-up of caliphal power. But it is necessary to stress that it did not do so yet. Al-Ma’mun kept the loyalty and cooperation - and the taxation - of Khurasan, largely thanks to the family of his general Tahir, who provided four generations of Tahirid governors there from 821 to 873, but who were simultaneously rulers of the city of Baghdad, which depended on provincial revenue. Egypt, at the other end of the caliphate, was finally quiet after 832. Al-Ma’mun’s army, no longer based on the early ‘Abbasid abn’, was initially a rather uncertain collection of east Iranian aristo cratic levies, who had trouble taking Baghdad against informal gangs of civilians (‘ayyrn) even though the defending regular army disinte- grated; but he, and especially his military-minded brother and successor al-Mu‘tasim (833-42), built up an army of mercenaries, particularly from Turkic central Asia, many of whom were former slaves, whom our sources generically refer to as Turks. This was an effective fighting force, not sufficiently Islamized to have its own political programme, not associated with any particular province of the caliphate, and very loyal, at least to al-Mu‘tasim. They provided the muscle behind the last really big ‘Abbasid attack on the Byzantine empire, which took Amorion in 838, and Turkish leaders were increasingly used as provincial governors. With the provinces quiescent, a model army, and an increasingly elaborate and extensive fiscal and administrative machine, the 830s and 840s under al-Mu‘tasim and his more colourless son al-Wathiq (842-7) represented a new high point for the centralized ‘Abbasid state, one that could have real staying power: or so one might have thought.
Ninth-century Baghdad, huge, wealthy and politically central as it was, became a real cultural focus. The startlingly large number of surviving works in Arabic from the ninth and tenth centuries, mostly (particularly before the 930s or so) written in or near the capital, themselves attest to it. They are only a portion of what was actually written, too, as is shown by the Fihrist or Index of al-Nadim (d. c. 990), which lists over 6,000 book titles, nearly all written in the last 250 years (this far outweighs the 279 Greek books in Photios’ Bibliothk, though Photios had at least read them all), or by an anecdote in the Fihrist itself about the 600 cases of books allegedly possessed by the historian al-Waqidi (d. 823) - an impossible figure for such an early date, but significant as a tenth-century image. Theology, philosophy, law, poetry, administration, history, medicine, science and geography all had their experts in this hyperactive cultural world.
These branches of knowledge increasingly developed their own microcultures, with lawyers above all reading other lawyers, historians reading other historians, poets reading other poets. They were tied together, all the same, by two main networks, one cultural-religious, one literary. The intellectual strata as a whole were seen as a community of scholars, the ‘ulam’ (from ‘ilm, ‘religious knowledge’). The community was defined initially and principally in terms of religious expertise, but came soon to extend out to the more specialized disciplines; its identity is most visible in biographical dictionaries of scholars, which were already being written in the early ninth century. It was this community, led by Qur’anic scholars and jurists, which was increasingly seen, in a religion with no formal priesthood or ecclesiastical hierarchy, as the determin- ators of what Islam was and how it should be understood, and indeed, in the twenty-first century, it still is.
The community did not, of course, always agree. We have already encountered the fault-line between Sunni and Shi‘a, which crystallized as alternative political-religious systems in the ninth century. Each of these systems, however, also had their own sub-systems, rival schools of thought about how religion, political practice and law ought to be conducted. Inside what would be called the Sunni tradition, for instance, there was from early in the eighth century considerable debate about the degree to which Islamic legal practice (shar‘a) should be based on legislation (presumably by caliphs), or else reasoning from basic ethical principles derived from the Qur’an, or else on the increasingly elaborate sets of ‘tradition’ (adth), obiter dicta attributed to Muhammad the Prophet on almost every legal or moral issue imaginable. (These pronouncements in reality gave a religious legitimacy to local custom, although custom on its own was never regarded as a legitimate fount of law.) The ‘traditionists’ essentially won out, but the four main law schools of medieval Sunni Islam, looking respectively to Abu Hanifa (d. 767), Malik (d. 795), al-Shafi‘i (d. 820) - the most intellectually influential - and Ibn Hanbal (d. 855), varied considerably in their commitment to adth, with Hanafis most receptive to legal reasoning and Hanbalis most rigidly attached to literal readings of adth. These schools, and other less long-lasting ones, achieved a mutual toleration all the same, as each constitutive of Sunni ‘ulam’ opinion, and by 900 or so they had developed what has been called the ‘closing of the gate of independent reasoning’: no new law or legal opinion, including by a caliph or other political leader, would, in theory, any longer be acceptable. Islamic law thus became increasingly fixed (even if legal practice did not). This served further to define the ‘ulam’ as a cultural grouping, although other disciplines continued to develop for centuries, much as the doctrinal rules of eastern and western Christendom bounded the developing thought-worlds of Europe throughout the Middle Ages as well.
The other way in which the realms of written culture were linked was through adab, roughly translatable as ‘polite education’, or ‘literary etiquette’. This became the foundation of Arab written culture by around 800, and remained so throughout our period and beyond. It linked learning with stylistic elegance, and required of its practitioners a general knowledge of most of the intellectual disciplines of the period, but particularly language, poetry, stories, administrative practice and adth. The administrative practice is the give-away: adab was above all a qualification for careers in government. It was the exact equivalent of the senatorial literary education of the Roman empire and of the classical and theological training necessary for administrators in Byzantium after 900, except that the knowledge it required was mostly of a much more recent vintage. And indeed the scope of intellectual activity in Baghdad and other centres showed the range of skills that were acceptable in government; intellectuals from the geographer Ibn Khurradadhbih (d. c. 885) to the seriously influential and original philosopher-physician Ibn Sina (Avicenna, d. 1037) held governmental and administrative offices. This range marks one of the particularities of adab. So also, however, does storytelling. Literary culture gave considerable space to narratives; ‘Abbasid histories are composed of thousands of short exemplary accounts, with plenty of direct quotations, supposedly taken from the lips of caliphs and their advisers. Rhetorical skill required remarkably recondite knowledge as part of such storytelling; hence the existence of several encyclopedias of ‘curiosities’, such as that of al-Tha‘alibi (d. 1038), which contains such information as the name of the first Arab to wear dark silks, the first caliph to build a hospital, the vizir with the longest unbroken chain of ancestors who were also vizirs, the most generous female pilgrim, the two caliphs who each killed three political rivals whose names began with the same letter, and the alarming (but untrue) fact that every sixth caliph was ‘inevitably’ deposed. This knowledge, these days restricted to adolescent boys, was in this period a requirement for statecraft, along with knowing how to write a letter properly and memorizing the Qur’an.
The strata of professional administrators, from vizirs and other senior secretaries down to the clerks in provincial tax offices, were complex, and generated their own cultural traditions. There are collections of administrative exemplary stories, just as there are political ones in histories; accounts of how and why individuals got promoted and demoted, and of the clever things they said to heads of dwns and vizirs. Nishwr al-muara, Desultory Conversations, another adab text, by the Basra judge al-Tanukhi (d. 994), shows how dense this specifically administrative historical memory could be, and how it extended, even in the late tenth century, without a break back to the caliphates of the mid-ninth, and even of al-Rashid and al-Mahdi. Among other things, one is struck by how accidental promotions could be in this world, as ordinary officials came to the eye of the powerful. Al-Fadl ibn Marwan (d. c. 845), a kitchen steward to an aristocrat and then a minor clerk in al-Rashid’s time, made enough money to buy land and live in the country during the siege of Baghdad, where he reputedly gave hospitality unknowingly to the future caliph al-Mu‘tasim; thanks to this chance, he rose steadily in the administration, and became vizir at his patron’s accession in 833 - though, conversely, he was soon dismissed (in 836), and had to pay huge sums in fines, because he tried to prevent the caliph from spending public money. The chance of fate was linked to a good deal of administrative competence; al-Fadl was an able administrator who brought in considerable revenues to at least two caliphs. It is also clear that plenty of these revenues stuck to his own fingers, given his wealth in the 830s. Much paperwork was indeed expended to try to cut down peculation, but al-Tanukhi’s stories show that this could easily be subverted, with misleading papers put in the records, until or unless rivals uncovered the fraud.
One gains a picture of a tight but very jealous administrative community, in which a common profession counted as a tie of kinship (as al-Fadl said, quoting a retired clerk whom he met as a youth), but in which promotion often depended on the destruction of others. At least al-Fadl kept his life in 836; plenty of others, including in particular many vizirs, did not. To say that administrative and court politics was cut-throat is indeed an understatement; unlucky ‘Abbasid politicians could die by tortures as inventive as those of the Merovingians, or indeed more so, as ‘Abbasid science was more developed - al-Fadl’s successor as vizir, Ibn al-Zayyat (836-47), died in a torture machine of his own devising. But Ibn al-Zayyat had also supposedly kept his position as vizir at the accession of al-Wathiq in 842, even though the new caliph loathed him and had sworn to kill him, because he was the only senior official who could compose a formal letter to the satisfaction of the ruler. This mixture of ambition, greed, violence and genuine professionalism marked the administrative class as a whole, or at least its upper echelons.
The complex and dangerous world of the administration was mirrored in the other two arenas of caliphal politics, the army and the caliphal household. The civil administration and the army are often seen as rivals in ‘Abbasid historiography, much as in middle Byzantine historiography, and probably as wrongly; as in Byzantium, the same person could do both, as with the Barmakid al-Fadl and the Tahirid ‘Abd Allah ibn Tahir (d. 845), and even the occasional Turkish general, such as Utamish (d. 863), who held the vizirate for a year before his death. Factions in reality crossed both areas of government without difficulty, even when the Turks, disliked and distinct, came to dominate the army. The numerous large palaces of the ‘Abbasids also had their own staff, not least the even more numerous slave mistresses of the caliphs, whose head was either a queen, or, if the caliph did not formally marry - which was the norm after the early ninth century - a queen-mother; the factions crossed into this arena too.
As with the Merovingians, equally dynastically minded and polygamous, political influence for women in the ‘Abbasid period tended to be restricted to the mothers of caliphs or designated future caliphs. The most famous examples of this were Khayzuran (d. 789), the mother of Harun al-Rashid, and Zubayda (d. 831), al-Rashid’s wife and mother of al-Amin. Zubayda even kept some of her influence after al-Ma’mun overthrew al-Amin - she brokered, for example, the reconciliation of the anti-caliph Ibrahim in 825. But it has to be said that ‘Abbasid political practice gave less scope to female protagonism than either the Frankish or the Byzantine tradition. The complicated and ever-developing ceremonial of the ‘Abbasid caliphate, which must have matched that of the tenth-century Byzantines, had rather less space for women as public players; but it is above all the case that succession rules focused on choosing appropriate candidates for caliph meant that child caliphs, for whom mothers could act as regents, were less common than royal minors were in Byzantium or Francia. The first was not until al-Muqtadir (908-32), whose reign was indeed dominated by his formidable mother, a Byzantine ex-slave called Shaghab (‘troublesome’), or, simply, al-Sayyida (‘the lady’). Shaghab (d. 933) is not handled in a consistently hostile way by the sources, despite their general suspicion of female power, magnified by the disasters of her son’s reign; she followed Zubayda in making public displays of charity on a large scale, a recognizable ‘Abbasid gendered female role, thanks to her vast wealth, and this allowed at least some chroniclers to depict her neutrally. Shaghab established a parallel bureaucratic hierarchy of male secretaries and female stewardesses which exercised direct power in these decades. It is important, however, to recognize that such offices were already normal in the female areas of the palaces. Queens, chief mistresses and caliphal mothers had long been wealthy, and needed administrators to run their affairs; if, on rare occasions, such as under Shaghab, these took over caliphal politics too, they had all the qualifications to do so.
Caliphs are portrayed in the sources in conventional ways, al-Mansur as eloquent and ascetic, al-Mahdi as generous and poetry-loving, al-Mu‘tasim as martial, and so on. Al-Ma’mun (who conventionally had a sense of humour and a gift for poetry) is perhaps the one who most established his own identity through his actions. His attraction to Shi‘ism is one such, which did not end when he backed down over his ‘Alid heir in 818-19. So is his patronage of scientists, who engaged in a programme of translations of Greek scientific works, Ptolemy, Galen, Euclid and so on, and the determination (among other things) of an accurate calculation of the circumference of the earth: this came to be carried out from a library and scientific research centre known as the Bayt al-Hikma, ‘House of Wisdom’, founded by the caliph in Baghdad in 830. Al-Ma’mun was also a doctrinal protagonist, sympathetic to a rationalist school of Islam known as Mu‘tazilism. The role of the caliph as a religious authority, which was seen as normal in the Umayyad period, and which was urged on al-Mansur by his Persian secretary and adviser Ibn al-Muqaffa‘ (d. c. 757) at the start of the ‘Abbasid caliphate, was being undermined by the growth of the authority of the ‘ulam’, but al-Ma’mun had a sufficient confidence in his mission to put doctrine into the heart of politics. In 833 he decided that one element of Mu‘tazil- ist thought, the doctrine of the createdness of the Qur’an (that is, that God had created the book within time; it had not pre-existed the world), was sufficiently important that all judges and ‘ulam’ should be forced to subscribe to it, particularly the ‘traditionists’, who were bitterly opposed to it. Almost alone, Ibn Hanbal defied him, and went to prison. The created Qur’an remained a tenet of the next two caliphs as well, and was only abandoned in 847, at the accession of al-Wathiq’s brother al-Mutawakkil (847-61). This period, of the so-called mina or ‘inqui- sition’, is the only one in which a doctrinal issue mattered politically in medieval Islam, as opposed to the permanent debates about the legitimacy of early caliphs. The apparent obscurity of the religious issue at stake is one element that reminds us of the Christological schisms of the later Roman empire. The sense one has of a political regime using such an issue to kick religious extremists into line is also a reminder of the near-contemporary Second Iconoclasm in Byzantium, and indeed al-Ma’mun recalls his younger contemporary Theophilos in his interest in religious-philosophical debate as well. Why al-Ma’mun chose the created Qur’an as the issue to make a stand on is, however, even less clear than the reasons for the Iconoclast controversy. It may be that any issue would have done, to re-establish caliphal religious authority, especially in the face of the ‘traditionists’. But the mina failed; Ibn Hanbal returned; after 849 doctrine was fully in the hands of the ‘ulam’, and caliphs - and, still more, their tenth-century supplanters in Iraq and Iran, who did not have their formal religious role as ‘commanders of the believers’ - became essentially secular powers. They would be patrons of intellectuals, jurists, ‘ulam’, but not intellectuals themselves.
Al-Mu‘tasim’s Turkish army got on particularly badly with the Bagh- dadis, who were after all the heirs of the previous paid army, the abn’, so the caliph built a new capital at Samarra, further up the Tigris, and moved both himself and his army there in 836. The establishment of new capitals was a standard part of early ‘Abbasid political affirmation; Baghdad itself was the key exemplar, and al-Rashid’s period in Raqqa (796-808) and al-Ma’mun’s in Merv (811-18) were others. Samarra was the most serious foundation after Baghdad, and was, as usual with the ‘Abbasids, built on a huge scale: its ruins extend along the Tigris for 40 kilometres. All the same, like Raqqa earlier, it was not intended to rival Baghdad as a population centre, and it remained largely a military and administrative centre during its period as the capital, 836-92. The problem was that the caliph was thus isolated together with his army. Both the Umayyads and the early ‘Abbasids used armies paid out of general taxation, which were separated from their areas of origin, the early Arab settlers in their amr, the Khurasani abn’ in Baghdad. In this respect, the Turks were not unusual, except that they came from beyond the frontiers, and they would have plenty of successors in the more fragmented tenth century too. There was always a tension between the paid military and the rest of tax-paying society in the medieval Arab world as a result of this pattern. Furthermore, because provincial élites converted to Islam, above all in the ninth century, and were matched by Arab settler families acquiring land - in the early eighth century in Khurasan, the late eighth in the boom-town hinterland of Baghdad, the late ninth in Egypt - there therefore came to be Muslim provincial aristocracies who could be very resentful of the political power and the financial weight of the army. This was particularly so in Khurasan, where the pre-Islamic Persian ruling class largely remained, with highly aristocratic and military values, however Islamized by now. Some of this Persian ruling class did indeed join al-Ma’mun’s and al-Mu‘tasim’s army, like al-Afshin of Ushrusana (d. 841), a hardly Muslim prince from central Asia, although he, significantly, perished because he was thought to have plotted against the Turks.
The caliphs could not, however, simply leave military affairs to local aristocracies; they would have instantly lost their tax revenues, and the caliphate would have broken up very fast. Given that, they might as well pay men from outside the caliphate, who had no aristocratic pretensions and were at least good at their job. But there were dangers too. In an anecdote laden with hindsight, the historian al-Tabari has the Tahirid Ishaq ibn Ibrahim tell al-Mu‘tasim: ‘your brother considered the roots and made use of them, and their branches flourished exceedingly; whereas the commander of the believers has utilized only branches, which have not flourished because they lacked roots.’ Which is to say: al-Ma’mun used Tahirids like myself, and other people rooted in the community, and that worked; but you use the Turks, who do not have such roots, and this is a real problem. Al-Mu‘tasim is supposed to have sadly recognized the truth of this. However this may be, the deracination of the Turks ceased to be an advantage when al-Mutawakkil turned against them in the 850s and sought to bring down their leaders, for they had nowhere to go. In the end, they responded by assassinating him in 861. This unleashed a decade of crisis in Samarra, 861-70, in which Turkish factions set up five caliphs in turn and killed three of them; the crisis extended back to Baghdad when one of them, al-Musta‘in (862-6), fled to the old capital and its Tahirid governor, with a section of the Turks, and Baghdad was besieged and captured again in 865-6. Stability only returned in 870 when the ‘Abbasid family developed its own military strongman, Abu Ahmad al-Muwaffaq (d. 891), who had in fact led the siege of Baghdad and was very close to the surviving Turkish leadership; he was put in charge of the army by his brother al-Mu‘tamid, who was caliph by now (870-92), and left the latter in Samarra while he gradually transferred himself to Baghdad. When al-Muwaffaq’s son and heir al-Mu‘tadid became caliph (892- 902), he formally re-established Baghdad as the capital, and the Samarra interlude ended.
The years 861-70 were not so very long, but, like the civil war of the 810s, they opened up fault-lines in the ‘Abbasid polity which were hard to close. The revived ‘Abbasid protagonism of 870-908 (it extended to al-Mu‘tadid’s son al-Muktafi, 902-8) faced widespread difficulties. Iranian rebels, the Saffarids (they did not have aristocratic roots, and they were close to fringe Muslim sects), had defeated the Tahirids in Khurasan between 867 and 873, and marched on Iraq; they were defeated there in 876, but they continued to control much of Iran, paying taxes only intermittently. The Turkish governor of Egypt, Ahmad ibn Tulun (868-84) was not directly opposed to the ‘Abbasids, but he too did not pay much tax to Iraq, and he extended his power into Syria and Palestine, which thus did not pay much either; only after his son Khumarawayh (884-96) succeeded him did an ‘Abbasid army manage to re-establish a greater measure of tax-paying from the Tulunid provinces, and not until 905 did the ‘Abbasids regain direct rule in Egypt. Only in Iraq did the ‘Abbasids exercise fiscal control in the 870s and 880s, and here, around Basra in the south, they faced a huge slave revolt, of the Zanj, African slaves used to maintain the irrigation system: this revolt, lasting from 869 to 883, was the most successful slave uprising in history before the Haitian revolt of 1791, resulting in an independent Shi‘ite state which was only destroyed by four years of war under al-Muwaffaq in 879-83. The ‘Abbasids were seriously short of money until the mid-880s, and even after that had to fight without a break, with their still-Turkish armies, to keep on top of events. They succeeded in their core lands, with the exception of Iran, which increasingly slid away under local dynasties. But they could not afford to relax their pressure. After 908, al-Muqtadir was a very inattentive ruler, and his mother Shaghab did not have control of the army. By the 920s, with infighting inside the bureaucracy, rival generals in Iraq, bedouin raids from the Arabian desert, and Syrian and Egyptian governors who had begun to stop paying taxes again, the gains of recent decades were all lost; in the 930s caliphs began to be deposed once more, and after 936 the caliph lost all power to a military governor, the amr al-umar’, ‘amir of amirs’. In 945 Ahmad ibn Buya (d. 967), from the most successful of the rising dynasties of Iran, the Buyids, took Baghdad, and became amr al-umar’ with the ruling name of Mu’izz al-Dawla, ‘fortifier of the [still nominally ‘Abbasid] state’. Iraq was controlled from western Iran from then on for a century.
The break-up of the ‘Abbasid caliphate, for a hundred years the strongest state in the world (Tang China had run into trouble in and after the 750s), would ideally need as detailed an account and set of explanations as did that of the Roman empire. If I dispose of the sequence of events in a couple of pages, it is only because by now, after the 860s, its history hardly extended beyond Iraq except for brief periods, and is too far from the history of Europe. The tenth century in the Islamic world was, as already observed, even more fragmented, with the Samanids and then the Ghaznavids in eastern Iran, two or three Buyid polities in western Iran and Iraq, two Hamdanid polities in Aleppo and (more briefly) Mosul, a set of Kurdish dynasties in the mountains to their north and east, the Qaramita in the Arabian desert, the Ikhshidids and then the Fatimids in Egypt, and other smaller polities too - as well as those of the Maghreb, which had not been under ‘Abbasid control since the early ninth century or even before, the Aghlabids and then the Fatimids in what is now Tunisia and Sicily, the Idrisids in what is now Morocco, and the Umayyads in Spain. We cannot follow all their histories here. But before we look at two of them, we do need to take stock of the century of ‘Abbasid unity and of its failure.
One simple reason why the ‘Abbasid caliphate broke up was that it was too large. Local societies were too different; communications were always slow; the caliphate was larger than the Roman empire, and did not have a sea, with its relatively easy bulk transport, at its heart. Conquests and reconquests, with new ruling armies and a clean slate, helped periodic reunifications: in 636-51, 747-50, 811-13, as subsequently with the Buyids, and the Seljuk Turks in the 1040s and later, but tensions would always rise again. This was particularly the case in Khurasan and in Iran as a whole, whose pre-Islamic ruling class, with some military protagonism, survived better than elsewhere (and whose pre-Islamic past was still celebrated by Muslims in oral and written literature, unlike anywhere further west except Spain); and which, being mountainous, was much harder to control in depth; significantly, the most successful and long-lasting later Islamic empire, the Ottomans, never held Iran. Trouble for the ‘Abbasids generally began in Iran; Iraq and Egypt were much easier to rule, and Syria was not any sort of power-centre for two centuries after the fall of the Umayyads.
This straightforward geopolitical argument is largely backed up by one basic point about the tenth-century Muslim successor states: they were almost all tax-raising states with a central paid army and bureaucracy, just as the caliphate had been. Only some of the Kurdish states of southern Anatolia and the Iranian mountains, followed by bedouin dynasties in Syria and the Jazira in the eleventh century, had a simpler structure, based on block gifts of tribute to armed transhumant groups. Unlike at the end of the western Roman empire, there was no structural breakdown inside the majority of these smaller polities. Unlike in the Romano-Germanic kingdoms, the new ruling groups were not concerned to make themselves into a landowning aristocracy. Land indeed did not bring political power in most medieval Muslim societies, only state position did that: or so it seemed to medieval political actors. Wealth, too, was most reliably obtained through positions in the state; and old families, whose longevity was ensured by private wealth - inevitably in land, in the Muslim as in the Christian world - were not especially privileged in any Islamic state structure, even in Iran. The political model established by ‘Umar I and two centuries of Umayyad and ‘Abbasid caliphs thus continued to hold. Indeed, it intensified, as the idea of ex-slaves holding military power, with no links to local communities and no family background, first experimented with al-Mu‘tasim’s Turks, became an increasingly common model in later centuries. Independence from the caliphate just meant that taxation stayed in the province concerned and paid a local army: a basic aim of provincial élites from the Umayyad period onwards, and only fully overridden by the strongest ‘Abbasid rulers, with reversions whenever ‘Abbasid control slipped, as in the 810s and 860s. From this standpoint, the break-up of the caliphate could even be seen as unproblematic, as simply consisting of the reversion of politics to its optimum size, the province.
Broadly, I think this interpretation is a fair one. But it does concentrate attention too much on the state; provincial societies get left out of the equation. Local social leaders were hugely diverse, extending from the old families of parts of Iran to the rapidly changing Iraqi élites, who tended simply to be the heirs of the most recent wave of administrators, who had made money from taxation and settled down; all the same, they existed everywhere. They certainly did have land by now, and also sometimes commercial wealth, which they turned into land as well. The great local political centres, almost all urban - major cities like Aleppo, Mosul, Rayy (modern Teheran), Merv and Nishapur in Khurasan - were full of local elite families, of ‘ulam’ and others, who sought the post ofqa, an important focus of local power, and who squabbled over local and provincial position, rather than seeking it from the state; here, land, private wealth and birth did matter (being an ‘Alid was increasingly chic, especially in Iran), just as it did in the West. ‘Abbasid governors always had to come to terms with local power-broking families, or else they would fail: they would be unable to collect tax (a process itself controlled by local figures), or face revolt, or both. So did the smaller-scale rulers of the tenth century. And indeed this in itself shows that there was a relationship between local societies and the ‘state class’. Even the most deracinated army family could put down local roots, at least as rulers, as the Tulunids did in Egypt; and all rulers, bureaucrats and local military men had to negotiate with their subjects, or at least the richest of them. Some sections of the ‘state class’, particularly the civil administration, had origins in local societies, too; they, at least, had tight local obligations.
All the same, a separation between the ‘state class’ and local and provincial societies did exist, and was a problem. By and large, making a career in the local city and making a career in the state were different, not only in the geographically large-scale ‘Abbasid caliphate but in the provincial polities of the tenth century as well. This meant that local societies could view the changing fates of their rulers with a certain equanimity: the latter were largely external figures, whether benevolent or violent, generous or fiscally harsh, cultured or martial, without a structural connection to the strata of the governed. As government became more secular, now that the fate of Islam had devolved to the ‘ulam’, the salvationist imagery of right rule so effectively invoked by Abu Muslim and the early ‘Abbasids was also no longer part of most political programmes. Only the Fatimids tried it in the tenth century, as we shall see in a moment. When a local ruler faced military failure, then, because a blockage in the tax supply made it hard to pay troops, or simply because of defeat in battle, he could be replaced without local society really being involved, as long as the new ruler did not take over too violently. There were certainly some examples of a loyalist protagonism by local élites, as when the citizens of Mosul in 989 expelled the Buyids and temporarily restored their earlier rulers, the Hamdanids, but they were not so very many. On one level, indeed, the very ease with which the ‘Abbasids lost control in the 910s to 940s, to be replaced by regimes which for the most part resembled them, was a real structural failure: however dismal the period was, it ought to have been possible for someone to make more of a stand, a heroic loser committed to an older legitimacy. The ‘Abbasids did not leave stories of that kind, and nor did the Buyids later. The stories that continued to hold attention were still Sassanian - or else of the timeless fantasy Baghdad of Harun al-Rashid and the Thousand and One Nights.
The Fatimids were the most successful, richest and most stable of the tenth-century Muslim states. They outlived their major rivals, the Buyids, by over a century, and indeed ruled over all, first in Kairouan in Ifriqiya, modern Tunisia, and then (after 973) in newly conquered Egypt, for more than two hundred and fifty years, 909-1171. They also represent, as just observed, the only serious attempt at a salvationist revival after the early ‘Abbasids, and are thus a special case in the tenth-century Islamic world. Their salvationism was, however, Shi‘ite, not Sunni. The first Fatimid, ‘Ubayd Allah al-Mahdi, was an Isma‘ili Shi‘a living in Syria, who belonged to one of the sects of Shi‘ism which held that a hidden imm or supreme spiritual leader, descended from the caliph ‘Ali, would return to redeem the world. In around 899 he declared - controversially inside the Isma‘ili movement, which he split in two - that he was himself the imam. He had to flee Syria, and ended up among the Kutama Berbers of modern Algeria, a sensible move, for the Berbers often had ‘Alid sympathies - an earlier ‘Alid exile, Idris ibn ‘Abd Allah (d. 795), had founded the Idrisid kingdom in central Morocco in 789. The Berbers were also good fighters, and were the core of the Fatimid army until well after our period ends. The Kutama adopted al-Mahdi as a charismatic leader, and keenly took to the role he offered them as the equivalent to the Khurasanis in the ‘Abbasid ‘revolution’. Their general, an Iraqi named Abu ‘Abd Allah, the Fatimid version of Abu Muslim, took Ifriqiya from the faltering Aghlabid dynasty in 909, and al-Mahdi proclaimed himself caliph (910-34) outside Kairouan a year later. Like Abu Muslim, Abu ‘Abd Allah was also killed by his patron-protégé inside a year, and al-Mahdi was not troubled by rivals thereafter.
Like both the ‘Abbasids and the Aghlabids, al-Mahdi set up his own capital in 920, at Mahdiyya on the Tunisian coast. He used the same governmental structures as the Aghlabids, although his Isma‘ili messianism set himself, and his Kutama army, apart from his Sunni subjects. That messianism, however, meant that al-Mahdi would not be content with Ifriqiya; from the start, the Fatimids looked eastwards, with raids on Egypt. This strategy was deflected by another salvationist Berber revolt, by Kharijites this time, in 944-7, but it was defeated, and by 960 al-Mahdi’s great-grandson al-Mu‘izz (953-75) ruled all North Africa, unified for the first time since the 730s. This stability allowed a renewed attack on Egypt, which was rudderless after the recent death of Abu’l Misk Kafur, a black ex-slave, a eunuch of fabled ugliness, who had ruled Egypt with skill and vision for twenty-two years (946-68). The Fatimid general Jawhar (d. 976), another ex-slave, a Slav this time, took the country with little violence in 969, and al-Mu‘izz moved there four years later. Jawhar and later generals pursued Fatimid ambitions on into Palestine and Syria, but they ran aground around Damascus, and when the frontier stabilized in the 990s it did so between Damascus and Aleppo. Fatimid expansionism stopped, and a modus vivendi emerged in Syria between the main regional powers, the Fatimids, the Buyids, and, since the 950s, the Byzantines, as we saw in the last chapter. Perhaps surprisingly, by the 990s the caliphs, now situated stably in wealthy Egypt, were prepared to let control over Ifriqiya slip, to a family of hereditary governors; from now on the Fatimids would be an Egyptian and Levantine power, which they remained for nearly two centuries more.
It is easy to see 909-10 as a rerun of 749-50, and at one level one whose religious fervour had greater staying power, for the Fatimids began a long way from the old power-centres of the Islamic world, which they would have to fight for longer to reach - indeed, they never reached Baghdad. As Shi‘ite imams, too, the Fatimid caliphs did not have to pay attention to the ‘ulam’ in any of their domains, for that was by definition Sunni, and anyway an imam drew his authority direct from God. But, even more than in Ifriqiya, Fatimid rule in Egypt was simply a continuation of the - already effective - rule of their predecessors. The Kutama in Egypt and Syria were another paid army, far from home, like the abn’ and the Turks. Al-Mu‘izz and his successors recentralized the fiscal administration of Egypt, as had the early ‘Abbasids, but in Egypt it had never been very decentralized. A strong state aided commercial development, but in any case Egypt had by now outstripped Iraq again as a productive region. In large part, the Fatimids allowed it to develop simply by creating stability; Egypt remained one of the major Islamic powers until the very end of the Middle Ages as a result, with a political protagonism unmatched since Cleopatra. Their administrative capital, al-Qahira, that is, Cairo, was founded in 969 just outside the previous provincial capital Fustat, which remained the commercial focus of Egypt; Fustat-Cairo was for a long time the major economic powerhouse of the whole eastern Mediterranean, surpassing even Baghdad, as we shall see in more detail in the next chapter.
So the Fatimids can be construed simply as normal rulers of the tenth century and onwards, just successful at it, and lucky with the region they ruled. All the same, this did not make the Fatimids exactly the same as their peers elsewhere in the Islamic world. Isma‘ilism, a secretive sect with esoteric and abstract Neoplatonist elements, including a complex letter and number symbolism, continued to mark out the court and the army, isolated among an ocean of Sunnis, Coptic Christians and Jews, and caliphs could continue to have messianic dreams: not least al-Hakim (996-1021), who erected anti-Sunni slogans on Sunni mosques, who demolished the church of the Holy Sepulchre in Jerusalem, and who was, and still is, venerated as divine by the Druzes of Lebanon. Al-Hakim was also a capricious and violent autocrat in a rather more familiar mould, but his religious imagery marks out the originality of the Fatimids nonetheless.
Tariq ibn Ziyad, the Berber leader of a largely Berber army, invaded Visigothic Spain for the Umayyad caliphs of Damascus and defeated and killed King Roderic in 711. The Berbers and Arabs had taken nearly all the peninsula by around 718. Muslim armies raided into Francia for another decade and a half, but without much commitment to conquest; Spain - al-Andalus in Arabic - was already on the very edge of their world, and it is likely that, if it had not fallen so easily, they would have stopped at the Straits of Gibraltar. Be that as it may, the occupation of the peninsula was quick. With the Visigothic army defeated, the Muslims made separate treaties with several local lords, in particular Theodemir in south-east Spain in 713. They did not base themselves in the old Visigothic capital of Toledo, but in Córdoba, in the rich south; Toledo was rather more of a frontier area, with an extensive uncontrolled land further north in the Duero valley between Muslim al-Andalus and the Christian polities of the northern fringe of the peninsula. At Córdoba, a succession of governors ruled, chosen by the caliphs. Al-Andalus looked like a normal, if outlying, province of the caliphate. It was as affected as was North Africa by the great Berber revolt of 740, but Caliph Hisham sent Syrian armies into Spain in 742, who won back the peninsula in 742-3 and settled there, thus increasing the Arab element of the Muslim settlement. The Syrians in Spain replicated the Qays- Yaman faction-fighting of the fertile crescent, however, and for a decade from 745 there was civil war between them. When the Umayyads were overthrown in Syria in 750 and largely wiped out as a family, one of Hisham’s grandsons, ‘Abd al-Rahman ibn Mu‘awiya, fled to the Berber kin of his mother, first in Africa and then, in 755, in Spain. Here he found support, both from Berber lineages and from the Yamani Arab opponents of the Qaysi governor, Yusuf al-Fihri. (The Yamanis in Spain were thus pro-Umayyad, not anti, as they had come to be by 749 in the East.) Inside a year he had defeated Yusuf and had taken Córdoba. ‘Abd al-Rahman I then ruled as amr for more than thirty years, 756-88, wholly independent of his ‘Abbasid enemies in Baghdad. So did his descendants, until 1031.
Spain was not like most of the other caliphal provinces, however. It was far more decentralized, and also, for a century at least, had a rather simpler economy than many, more like the economies of the rest of western Europe, with relatively unskilled and far more localized artisanal production, than like the economically complex and heavily urbanized provinces of the caliphate, Egypt or Syria or Iraq. Even its major cities, which under the Arabs as under the Visigoths were Córdoba, Seville, Mérida, Toledo, Zaragoza and a few others, were for a long time relatively small by comparison with those of the eastern Mediterranean. Spain was also, crucially, one of the only provinces conquered by the Arabs which did not have more than a fragmentary tax system. The standard procedures for Arab occupation, based on a paid military élite in a (perhaps new) garrison city, were thus impractical. The Berbers, newly Islamized (when converted at all) in the 710s, anyway doubtless wanted simply to settle on conquered land, and did so. But even the Syrians, who were sent in in the 740s as a normal paid army, soon settled on the land too - initially as tax-farmers, soon as landowners - and just did military service (for which they were paid by the campaign); they intermarried with the Visigothic aristocracy, and into the tenth century, as we shall see, there were families who were proud of both their Arab and their Gothic ancestry.
The amirs took what tax they could from the start, and were heavily criticized by chroniclers for it from the start (as witnessed by a mid-century Christian source, still in Latin, the Chronicle of 754). All the same, they had none of the fiscal control of governors elsewhere. Unlike anywhere else in the caliphate, they had to face a Muslim landed aristocracy from (nearly) the start as well, who might be able to resist tax-paying more successfully than their still-Christian neighbours. Nor was there much of a paid ‘state class’, either civilian or military, for some time. The existence of the frontier with the Christians in the north also led to a military-political fragmentation, with half of al-Andalus separated off into marches (thugr), based on central-northern centres like Toledo and Zaragoza, or Tudela, power-base of the ex-Visigothic Banu Qasi family, over which the Umayyad amirs, based in the south, had little control for a century and more. Spain is very regionally diverse, with bad communications, and the Muslim conquest had caused its local societies to move sharply in different directions; these contrasts were also further exacerbated by the diversities of Arab and Berber settlement. The Berbers, for example, seem to have settled in tight tribal groups in more marginal areas, but to have become ordinary (and Arabized) landowners when living in or near cities. Given this local diversity, this political fragmentation, and the need for the Umayyad amirs from the start to recognize the relevance of the politics of land, Muslim Spain was indeed as much part of western Europe as it was part of the Arab political environment.
Faced with this reality, the Umayyads were eventually rather successful for a time, but it was a long process and it was far from straightforward. ‘Abd al-Rahman I essentially established the centrality of his own family, which was a task not yet completed in 756 - the Banu Fihri, a powerful family in both Africa and Spain, who had supplied four governors in al-Andalus alone, were still revolting into the 780s. Father- son succession then followed into the 880s without a break, and, although there were certainly succession disputes between sons, and killings of potential rivals, there was actually no protracted disagreement about which Umayyad should rule until after 1000, a remarkable record, and one which both aided stability and was made possible by it. The state was still fairly skeletal until the 820s, however. ‘Abd al-Rahman I did employ a small paid army, but it is unlikely that his tax-base extended far outside the Córdoba-Seville region, linked by the lowlands of the Guadalquivir valley, and attempts by his grandson al-Hakam I (796- 822) to stabilize that taxation led to revolt in 818, not only in marcher centres like Toledo, where uprising was fairly frequent, but among the urban population of Córdoba itself. It was not until ‘Abd al-Rahman II (822-52), a subtler ruler, that an administrative system resembling that of the caliphs of the East took shape, with higher taxation, a bureaucratic class (headed here by the hjib, the chamberlain, not by the vizir - the latter was a lesser office in Spain, and there were usually several of them) and a wider political control. ‘Abd al-Rahman II in 825 built a new city, Murcia, in the previously marginal south-east, and settled it with Arab loyalists; he confronted the rebellious tendencies of Mérida by building a large internal fortress there in 835, and another in Toledo in 837; and he developed a formal court in Córdoba, now fast expanding as a city, whose growth in power, wealth and buying-power meant that it would not henceforth be disadvantageous to the capital for the amir to be strong there.
Al-Andalus under ‘Abd al-Rahman II and his son Muhammad (852- 86), seen from the standpoint of the state, thus came more and more to match the ‘Abbasid heartland. The former patronized poets and scholars from the East, not least the important Iraqi musician and poet Ziryab (d. 857), who was rewarded for coming west by a huge salary. ‘Abd al-Rahman’s reign was also marked by the crystallization of an ‘ulam’ on an entirely eastern model, dominated by the Maliki law school, and soon present in every major city and plenty of minor ones. Al-Andalus, with its Umayyad legitimist tradition, was almost devoid of the disputes about right rule that were so important elsewhere, and even its law was not up for discussion. This in part marks its provinciality by comparison with the East, but the cultural continuum that linked them was unbroken; that would remain true in Ibn Hawqal’s time, as we have already seen. Indeed, Spanish historians, once history-writing began in the peninsula (with ‘Abd al-Malik ibn Habib, d. 853, a wide-ranging intellectual), were capable of writing in detail about eastern events on occasion; Andalusis were consistently informed about what went on in the ‘Abbasid world. The population was also, even if slowly, converting to Islam; a majority of al-Andalus was probably not Muslim until well into the tenth century, and Christians and Jews never ceased to be influential in Andalusi culture, but political leaders and major political centres were in general mostly Muslim now. A sign of this is the strange minority movement known as the ‘martyrs of Córdoba’, Christian extremists led by Eulogius (d. 859) and Alvar, who deliberately provoked their death in the capital by insulting Islam in public in the 850s. There were less than fifty of them, and they were clearly unrepresentative of the still-large Córdoba Christian community, despite the fascination their writings (conveniently in Latin) have had for recent scholars; but the desperation of their stand implies that they saw only extreme measures as adequate against the steady advance of Muslim hegemony.
This process of increasing amiral power on eastern political models was falling apart, however, by Muhammad’s death, and the 880s-920s were a long period of generalized disturbance or fitna. Muhammad already had trouble with Toledo and Mérida; he made peace with the former in 873, and sacked the latter in 868, but then nearby Badajoz, which became an alternative political centre to Mérida in the 870s, turned to revolt too under the former Méridan leader ‘Abd al-Rahman ibn Marwan al-Jilliqi (d. 892). In the 880s ‘Umar ibn Hafsun (d. 917) also revolted from his base at Bobastro in the far south, above Málaga. Under Muhammad’s son ‘Abd Allah (888-912), more and more local lords established effective independence, both in the marches and in the Andalusian heartland of the Guadalquivir valley. ‘Abd Allah was an ineffective and reclusive ruler, but the problem was a wider one. The Muslim landed aristocracy, many of whom (including Ibn al-Jilliqi and Ibn Hafsun) had at least partial Visigothic ancestry, had effective local bases and local loyalties. They could be happy with an expanding state, from which they could benefit, even though the growing fiscal demands of that state were opposed to their immediate interests, but if the state faltered they would look to their localities, rather than to the person of the amir. Beneath the ‘Abbasid-style political system in Córdoba, that is to say, the more western-style local political practice, already discussed, continued to exist. Iran, with its surviving Sassanian aristocratic families, offers the closest parallel, including the survival of pre-Arabic political imagery in local social memory; the Zoroastrian legitimists that can be found in Iran as late as the tenth century have their parallel in ‘Umar ibn Hafsun, who actually converted to Christianity in 898. But Iran also had other regions with strong paid armies and depoliticized local societies, which tended to dominate politically. In Spain, the permanent paid army was still not substantial, and military service was largely controlled, as in other parts of the West, by the very aristocrats whose loyalty was now in doubt. When even Seville in 899 established effective autonomy under a member of one of its local élite families, Ibrahim ibn al-Hajjaj (d. 911), called ‘king’ (malik) in the sources, the state risked breaking up.
‘Abd Allah’s grandson and successor, ‘Abd al-Rahman III (912-61), was the ruler who reversed this trend, and by doing so he inaugurated three generations of strong central power, the strongest known in Spain between the Romans and the thirteenth century. ‘Abd al-Rahman III understood that the only way to cope with this decentralization was to fight, systematically and without a break. In only two years he re-established control over the Guadalquivir valley; thereafter he pushed outwards, expanding his army as he did so, not just in the old amiral heartland but in the marches as well. Bobastro fell in 928, Badajoz in 930, Toledo in 932. ‘Abd al-Rahman for the most part incorporated the lords he uprooted into his army or else into the civilian state class in Córdoba, but they were, crucially, separated from their local power- bases and incorporated into a tax-based political system that was less superficial in its similarity with the East than in the previous century. This was underlined further by a great increase in slave and ex-slave soldiers, who were mostly Saqliba, ‘Slavs’ (though the word extended to include other northern Europeans). From as early as 916 this enlarged army was also sent north against the Christians, which further allowed ‘Abd al-Rahman (who, unusually, often led his own troops) to impose himself in the marches. In the end, he came fully to control all of al-Andalus except the Upper March in the far north-east, whose lords gave him military service and tax but remained autonomous. Even there, the main old ex-Visigothic family, the Banu Qasi, had lost its power by 907, and was replaced as a regional focus by the Tujibis, a family close to the Umayyads, which had been given Zaragoza in 890 in one of Amir ‘Abd Allah’s rare effective interventions. This hegemony was not weakened, except partially in the Upper March, by ‘Abd al-Rahman’s only serious military defeat, against the Christians of León in 939 (see Chapter 20). This overall success, plus the collapse of ‘Abbasid power in the same period and the Fatimid establishment of a rival Shi‘a caliphate in 910, led ‘Abd al-Rahman III to proclaim himself caliph, as al-Nasir, in 929.
The tenth century was the period when the ceremonial of the ruler developed most fully. Córdoba gained a series of new suburbs, and, with its monumental mosque in the centre, greatly enlarged by ‘Abd al-Rahman’s son al-Hakam II (961-76), moved into the league of Constantinople and Cairo as a metropolis. ‘Abd al-Rahman also founded around 940 an impressive new court and administrative centre at Madinat al-Zahra’, just north-west of the city. Here, caliphal ritual is recorded in a number of texts, from the Lifeof John of Gorze, ambassador for Otto I in around 953-6, intransigent in its (and its subject’s) hostility to Islam but unwillingly impressed by the complexity of the court, to the 971-5 section of the history by ‘Isa al-Razi (d. 989), preserved a century later in theMuqtabis of Ibn Hayyan (d. 1076), which provides us with several detailed accounts of particular ceremonial moments at the high points of the Muslim religious year. In the caliph’s main reception hall at Madinat al-Zahra’, all major officials had their allotted positions, in two lines, with the caliph at the end; the majesty of caliphal power was intended to be, and was, made very clear.
The tenth century was also a period of larger-scale economic activity. We shall see in the next chapter that al-Andalus participated in Mediterranean exchange, through the port of Almería, founded (or, rather, walled and expanded) by ‘Abd al-Rahman III in 955. Internally, too, we can see in recent archaeology the development of centralized and professional artisanal production of ceramics and glass, including glazed pottery in east Mediterranean styles, not least a ‘green and manganese’ decorated ware, which appears extensively on Spanish sites of the period, and which seems to have been made largely in Córdoba and other major centres. That latter ware has explicit caliphal associations, as can be seen in the frequent inscription al-mulk (‘power’) along the edges of plates and bowls, especially but not only in Madinat al-Zahra’. But this sort of artisanal activity cannot be in itself ascribed to ‘Abd al-Rahman or his political success. Tenth-century artisanal work built on that of the ninth, which was notably more professional than that of the eighth; it testifies to the steady development of hierarchies of wealth and élite demand in most of the Muslim parts of the peninsula. (Not the Christian parts; but Arab-made artisanal goods, especially carpets, cloth and leather, were nonetheless prized there as luxuries.) One thing this growing economic complexity shows is that the rich aristocracies of the ninth century had by no means gone away; they had simply been absorbed into the caliphal political hierarchy, or else into the local ‘ulam’ hier archies of the cities of al-Andalus - or else both, for Spain was not that large, and the deracinated Slav (and, later, Berber) armies were only part of the ‘state class’. Their identity and assumptions are well expressed by the historian and grammarian Ibn al-Qutiya (d. 977), son of a judge in Seville, who wrote a chatty history full of stories about the huge landed wealth of his ancestors, who supposedly included Sara ‘the Goth’ (al-Qtiya), granddaughter of King Wittiza; Ibn al-Qutiya was nonethe- less as focused on the doings of the Umayyads as any other historian, and clearly bought into the values of the court. All that ‘Abd al-Rahman did here - not a small thing, however - was to create the political foundation for the linkage of the local economies and societies of the ninth century in a single network, covering the whole of the Spanish caliphate.
Al-Hakam continued his father’s political practices; he was well known as a literary patron, too. His military expansion, especially in 972-5, was southwards, into Morocco, which had been largely left to its own devices after the Fatimid move into Egypt. At his death, however, his son al-Hisham II (976-1009, 1010-13) was only fifteen; power was seized by one of al-Hakam’s military leaders in Morocco, Muhammad ibn Abi ‘Amir, who had a loyal detachment of Berbers to help him win a coup against their Slav rivals. Ibn Abi ‘Amir steadily eliminated all other powerful figures in the court, and in 981 assumed supreme power as ruling ajib for a figurehead caliph, even giving himself the ruling title of al-Mansur (in Spanish Almanzor, 981-1002). Al-Mansur greatly developed the Berber component of his army to counterbalance the Slavs. He fought in Morocco, too; but he principally sent his armies to the north, against the Christian kingdoms and principalities, whom he defeated time and again, notably but not only in the devastating sack of Barcelona in 984 and of Santiago de Compostela in the far north-west in 997; his son al-Muzaffar (1002-8) continued this as well. In this military dominance, coupled with a substantial internal stability, and a continuation of the central ceremonial role of Córdoba - where al- Mansur built yet another suburban administrative centre, Madinat al-Zahira - the Umayyad caliphate appeared to reach its height.
As with the ‘Abbasid high point under al-Mu‘tasim and al-Wathiq, however, this hegemony would not last. Indeed, almost as soon as al-Muzaffar died, al-Andalus disintegrated into a twenty-year civil war (1009-31). The detailed reasons for this lie outside our period; they essentially lie with the political ineptness of al-Muzaffar’s successors, and power-struggles between Berber and Slav leaders. But this fitna was far more serious than its predecessor a century earlier; it included a violent sack of Córdoba itself in 1013, and the abandonment of the nomination of caliphs altogether, by now all of them figureheads, in 1031. By that date al-Andalus was divided between thirty or so kingdoms, known as the Taifas (from t’ifa, ‘faction’), and it never recovered ‘Abd al-Rahman’s political unity or al-Mansur’s military protagonism. This collapse was so fast and so complete - far faster than that of the ‘Abbasids, and resulting in independent polities that were in many cases single city territories, far smaller than the successor states in the East - that it needs some comment.
Some of the Taifa kingdoms were ruled by regional army commanders, Slav or Berber, who simply turned their commands into autonomous, and then independent, units as central authority collapsed in the 1010s, as in the East. Some, especially in the north-east, were ruled by long-standing families whose local power had been recognized even by ‘Abd al-Rahman III, the Tujibis in Zaragoza or the Dhi’l-Nunids of the upland Santaver area, who in 1018 occupied Toledo. But some, including perhaps the richest, Seville, were taken over by local landowners who had civic, not state, office: not necessarily from the same families who had dominated around 900, but at least from the same social stratum. We have to conclude that ‘Abd al-Rahman III had not definitively ended the presumption, which had always been stronger in al-Andalus than elsewhere in the Muslim world, that landownership brought potential rights to political authority. And, even more important: notwithstanding the substantial territorial reorganizations of the caliphal period - with governorships both large and small tightly controlled by central government, and many of the local fortifications of the first fitna simply taken over by the state - ‘Abd al-Rahman and his successors had not succeeded fully in undermining that other core Spanish presupposition, that practical politics was local. In both these respects, the Visigothic inheritance of al-Andalus comes out in the Taifa period. The amirs and caliphs succeeded in establishing a tax-based state, such as had not existed in Spain since the Roman empire, and this indeed continued under the Taifas; but they did not manage to move their Andalusi population to the assumptions that prevailed in Egypt or Iraq, even in the fragmented tenth century, that only the control of the state mattered, and that a land-based local politics was marginal. When the state faltered, in the 1010s as in the 880s and, earlier, in the 710s, Spain’s localities at once moved centre stage. When a degree of reunification belatedly came this time, with the Almoravids at the end of the eleventh century, the Christians had taken Toledo and the whole balance of power had shifted.