Crone, Slaves on Horses

March 4, 2017 | Author: Fireliel | Category: N/A
Share Embed Donate


Short Description

Download Crone, Slaves on Horses...

Description

SLAVES ON HORSES THE EVOLUTION OF THE ISLAMIC POLITY

SLAVES ON HORSES THE EVOLUTION OF THE ISLAMIC POLITY

PATRICIA CRONE Lecturer in Islamic History University of Oxford

CAMBRIDGE UNIVERSITY PRESS LONDON

CAMBRIDGE NEW YORK NEW ROCHELLE MELBOURNE SYDNEY

PUBLISHED BYTHE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE The Pitt Building, Trumpington Street, Cambridge, United Kingdom CAMBRIDGE UNIVERSITY PRESS The Edinburgh Building, Cambridge CB2 2RU, UK 40 West 20th Street, New York NY 10011-4211, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia Ruiz de Alarcon 13,28014 Madrid, Spain Dock House, The Waterfront, Cape Town 8001, South Africa http './/www. Cambridge, org ©Cambridge University Press 1980 This book is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 1980 First paperback edition 2003 A catalogue recordfor this book is available from the British Library Library of Congress Cataloguing in Publication data Crone, Patricia, 1945Slaves on horses. Based on part of the author's thesis, University of London, 1974. Bibliography: p. Includes index. 1. Islamic Empire - Politics and government. I. Title. DS38.5.C76 9O9'.O9767O1 79-50234 ISBN 0 521 22961 8 hardback ISBN 0 521 52940 9 paperback

For T. C. and V. C.

I have seen slaves upon horses, and princes walking as slaves upon the earth.

Eccles. 1 0 :j.

CONTENTS

Preface

ix

A note on conventions

x PART I: INTRODUCTION

1 Historiographical introduction

3

2 The nature of the Arab conquest

18

PART II: THE EVOLUTION OF THE CONQUEST SOCIETY

3 4 5 6 7 8

The Sufyanid pattern, 661—84 [41—64] Syria of 684 [64] The Marwanid evolution, 684—744 [64—126] The Marwanid faction Syria of 744 [ 126J Umayyad clientage

29 34 37 42 46 49

PART III: THE FAILURE OF THE ISLAMIC EMPIRE

9 The abortive service aristocracy 10 The emergence of the slave soldiers 11 The emergence of the medieval polity

61 74 82

APPENDICES

I The ashrafoi Syria and Iraq II The subgovernors of Syria, 685 — 744 [65 — 126] III The subgovernors of Iraq and its dependencies, 694—744 [75-126] IV The Yamaniyya and the Qaysiyya V The 'Abbasid servants, 750—813 [132—98]

93 123 129 154 173

VI Mawlfi in the sense of 'kinsman' Notes

201 vii

viii

Contents

Bibliography General index

2

72

Prosopographical index

2

93

PREFACE

This book is a reincarnation of the first part of my thesis ('The Mawali in the Umayyad Period', University of London Ph.D., 1974) in a form so different that theologians might dispute the identity. I should like to thank Professor B. Lewis, who supervised me in 1969—73, Professor M. J. Kister, who helped me during a term in Jerusalem in 1972, Robert Irwin, whose queries inspired two pages of part III, and Dr Martin Hinds, whose criticisms inspired many more. Above all I wish to thank Michael Cook, who read the entire typescript in both its past and its present form, and whose advice I have nearly always followed, if not always with good grace. I also owe a special debt to Magister E. Iversen for suggesting to me, many years ago, the unfamiliar idea of becoming a historian. Needless to say, not even Magister Iversen can be held responsible for the result. P.C.

A note on conventions

Dates in the text are A.D., but hijri&axzs have been added in square brackets where appropriate; in the appendices and notes all dates are hijn unless otherwise specified. The full names of Arabic authors are given in the bibliography, but only the short forms are used elsewhere.

PART I

INTRODUCTION

HISTORIOGRAPHICAL INTRODUCTION

This work presents an explanation of how and why slave soldiers came to be a central feature of the Muslim polity. The conceptual framework in which the explanation is set is that of Hagarism, and to the extent that the crux of the explanation has already appeared there,1 this work maybe regarded as simply an overextended footnote. There is, however, one respect in which the two works differ radically; for where Hagarism rejected the Islamic tradition, the present work is squarely based upon it. This apparent lack of historiographical morality may meet with some disapproval, but it arises from the nature of Islamic historiography itself. Whereas the religious tradition is such that it must be accepted or rejected in toto, the secular tradition can to some extent be taken to pieces, and though a great deal of it has to be discarded, there remains enough for a coherent historical account. Before going on to the subject of this book, it is worth lending substance to this claim. Muslim knowledge of the Muslim past was transmitted orally for about a century and a half.2 Whatever the attitude to the permissibility of writing history,3 little history was actually written until the late Umayyad period,4 and the first historical works proper were only composed in early 'Abbasid Iraq.5 The fact that history was transmitted orally does not, of course, in itself mean that it was transmitted unreliably. Human brains can become memory banks of astonishing capacities, procedures can be devised for the transfer of memory from one bank to another, and professional memorizers easily hold their own against copyists in the business of perfect replication: the Vedas, Panini's grammar and the Avesta were all transmitted for centuries by such men. But rigorous procedures along these lines are only adopted for the transmission of highly authoritative works which need to be immutably preserved, not for works of religious innovators; for where classics need to be preserved, new ideas need above all to be spread, and inasmuch as they engender change, they cannot well be shielded from it. Adherents of a new religion necessarily inhabit a different world from that of the founder himself: were it otherwise, his attempt at a religious paradigm 3

4

Introduction

shift would have failed. Hence they will go over their tradition oblivious of the problems with which the founder struggled, struggling with problems which the founder never envisaged, and in so doing not only elaborate, but also reshape the tradition which they received. And since the world of our grandparents, as not quite that of our parents, easily becomes ancient history of which we know little and understand even less, the founder must resign himself to the fact that it takes only three generations for his life and works to be thoroughly reshaped:6 the only insurance policy he can take out against it is to write his own authoritative works.7 Oral transmission in the formative period of a new religion, in short, does not mean faithful preservation, but rapid transformation of the tradition. Thus against the Hindu Vedas we can set the Buddhist Skandhaka, in which the life of the Buddha was first presented.8 It was a grandchild of the Buddha's generation who created this authoritative work in an effort to outbid the Vedas. Formally it was a biography. Substantively it was an exposition of monastic rules interspersed with entertaining legends, in which remains of the tradition from which the biography was recast could still be found, but which was otherwise devoid of historicity. And thanks to its success it is directly or indirectly the source for the bulk of our knowledge of the Buddha's life today.9 Similarly thanks to its success, the Sira of Ibn Ishaq is practically our only source for the life of Muhammad preserved within the Islamic tradition. The work is late: written not by a grandchild, but a great-grandchild of the Prophet's generation, it gives us the view for which classical Islam had settled.10 And written by a member of the 'ulama\ the scholars who had by then emerged as the classical bearers of the Islamic tradition, the picture which it offers is also one sided: how the Umayyad caliphs remembered their Prophet we shall never know. That it is unhistorical is only what one would expect, but it has an extraordinary capacity to resist internal criticism, a feature unparalleled in either the Skandhakaot the Gospels, but characteristic of the entire Islamic tradition, and most pronounced in the Koran: one can take the picture presented or one can leave it, but one cannot work, with it.l ' This peculiar characteristic arises from a combination of the circumstances and the method of transmission. The circumstances were those of drastic change. Whereas Buddhism and Christianity spread by slow infiltration, the coming of Islam was by contrast an explosion. In the course of a few decades the Arabs exchanged their ancestral paganism for monotheism, the desert for a habitation in the settled Middle East, tribal innocence for state structures, poverty for massive wealth, and undisturbed

Historiographical introduction

5

provinciality for exposure to the world's polemical attention. Rarely have a preacher and his followers lived in such discontinuous environments: what made sense to Muhammad made none to Mu'awiya, let alone to 'Abdal-Malik. Even so, the Arabs might well have retained a more integral recollection of the past had they not proceeded to adopt an atomistic method of transmission. The transmitters memorized, not coherent narratives or the components of one, but isolated sayings, short accounts of people's acts, brief references to historical events and the like. It was a method evolved by the Jewish rabbis for the transmission of the Oral Law, and the Mishnah was handed down with the same rigorous attention to immutability as were the Vedas. But it was also a method which, once the rigour was relaxed, made for even greater mutability than that exemplified in the formation of the Skandhaka. Being short and disparate, the components of the tradition were easily detached from context, forgotten or given a new meaning by the addition of a single word or two. Rabbinical memories of the past not only suffered rapid attrition and deformation, but also tended to be found in a variety of versions set in a variety of contexts in answer to a variety of problems, with the overall effect that the original contours of the tradition were blurred beyond all hopes of recognition.12 For the rabbis the past was constantly disintegrating into amorphous bits even at the most stable of times. For the Arabs the combination of atomistic transmission and rapid change was to mean both fast erosion of old structures and fast appearance of new ones. To this came a further circumstance. Muhammad was no rabbi. Whereas Jesus may have been a teacher whose doctrine may well have been handed down in accordance with the normal methods of rabbinic transmission,13 Muhammad was a militant preacher whose message can only have been transmitted bflmana, not bfl4aj\ that is to say only the general meaning was passed on. For one thing, rabbinic methods of transmission were not current among the bedouin; and for another, the immediate disciples of a man whose biography was for some two hundred years studied under the title of 'Urn al-magha^, the Prophet's campaigns,14 are unlikely to have devoted their lives to the memorization of hadith. In time, of course, Muhammad's words were to be transmitted with the usual attention to immutability, both orally and in writing, and he himself to some extent laid down his sword to assume the role of the authoritative teacher of the Sira.15 But that is not how things began. The Muslim rabbis to whom we owe the Prophet's biography were not the original memory banks of the Prophet's tradition. The Prophet's heirs were the caliphs, to whose unitary leadership the

6

Introduction

embryonic religion owed its initial survival. The 'ulama' appear with the Oral Tradition itself, perhaps in the mid-Umayyad period, perhaps before, and the history of Islam thereafter is to a large extent the history of their victorious emergence. The tradition as we have it is the outcome of a clash between two rival claimants to religious authority at a time when Islam was still in formation. We have, in other words, a situation in which the Arabs were rent by acute internal tension and exposed to scathing external polemics, under the pressure of which current doctrines were constantly running out of plausibility. As the caliphs pushed new doctrines at their subjects and the nascent 'ulama took them up, worked them over and rejected them, the past was broken into splinters, and the bits and pieces combined and recombined in different patterns, forgotten as they lost their relevance or overlaid by the masses of new material which the pressure generated: it is no accident that whereas the logia of Jesus have remained fairly small in number, those of Muhammad can be collected by the volume. For over a century the landscape of the Muslim past was thus exposed to a weathering so violent that its shapes were reduced to dust and rubble and deposited in secondary patterns, mixed with foreign debris and shifting with the wind. Only in the later half of the Umayyad period, when the doctrinal structures of Islam began to acquire viability, did the whirlwind gradually subside. The onset of calmer weathers did not, of course, mark the immediate stabilization of the Islamic tradition. On the one hand, the controversies over the Oral Law continued to generate Prophetic hadiths into the ninth century;16 and on the other hand, the Muslim rabbis now began not just to collect but also to sift and tidy up the tradition, an activity which issued in the compilation of the first historical works in early 'Abbasid Iraq. Nonetheless, it is clear that it was in the course of the first hundred years that the basic damage was done. For the hadiths from the late Umayyad period onwards can to some extent be dated and used for a reconstruction of the evolution of Islamic theology and law.17 And the rabbinic censorship, though far from trivial, eliminated only the remains of a landscape which had already been eroded. That much is clear from Ibn Hisham who, as he tells us, omitted from his recension of Ibn Ishaq's Sira everything without direct bearing on the Koran, things which he felt to be repugnant or which might cause offence, poems not attested elsewhere, as well as matters which a certain transmitter could not accept as trustworthy.18 Despite his reference to delicate topics, Ibn Hisham clearly saw himself as an editor rather than a censor: most of what he omitted had long ceased to be dangerous. We have in fact examples of badly censored works in Muslim eschatological books,19 particularly the

Historiographical introduction

7

Kitab al-fitan of Nu'aym b. Hammad, who happily defines the mahdi as he who guides people to the original Torah and Gospel ;2° and though Jewish and Christian material is conspicuously present in these works, the doctrinal formations of which it is the residue can no longer be restored: the structural damage had been inflicted in the course of oral transmission. But it is above all our one surviving document which conclusively demonstrates this point. The Constitution of Medina is preserved in Ibn Ishaq's Sira, in which it sticks out like a piece of solid rock in an accumulation of rubble,21 and there is another recension in the Kitab al-amiual of the ninth-century Abu 'Ubayd.22 Abu 'Ubayd's version, which is later than Ibn Ishaq's, is a typical product of written transmission: it has copyists' mistakes,23 interpolations,24 several of the by now unintelligible clauses have been omitted,25 and it has also been equipped with an isndd;z6 but otherwise the text is the same. The Constitution, however, also survives in a number of hadiths. The hadiths are all short; they mention two or three of the numerous clauses of the document, but do not spell them out; they characterize the document as a scroll coming from the Prophet, but leave the occasion on which it was written unidentified, and turn on the point that the scroll was in the possession of 'All.27 Whereas written transmission exposed the document to a certain amount of weathering which it withstood extremely well, oral transmission resulted in the disintegration of the text, the loss of the context and a shift of the general meaning: the document which marked the foundation of the Prophet's polity has been reduced to a point about the special knowledge of the Prophet's cousin. The religious tradition of Islam is thus a monument to the destruction rather than the preservation of the past. It is in the Slra of the Prophet that this destruction is most thorough, but it affects the entire account of the religious evolution of Islam until the second half of the Umayyad period; and inasmuch as politics were endowed with religious meaning, it affects political history no less. There is not much to tell between the sira of the shaykhayn, the first two caliphs, and that of the Prophet: both consist of secondary structures stuffed with masses of legal and doctrinal hadiths. The hadiths do at least have the merit of being identifiable as the product of the late Umayyad and early 'Abbasid controversies, and though they constitute a sizeable proportion of our information about the conquests/8 they taper off with the coming of the Umayyads. For if the reign of the first four caliphs was sira, a normative pattern, that of the Umayyads, by contrast, was jatvr, paradigmatic tyranny, and where the fiscal rectitude of the first four caliphs is spelt out in a profusion of detail, thefiscaloppression of the Umayyads is summarily

8

Introduction

dismissed in a number of stereotype accounts which, for all their bias and oversimplification, do in fact contain some historical truth.29 And by the time of the 'Abbasids the lawyers had begun to reach their classical positions on the subject; the fiscal policy of the 'Abbasids was therefore neither sira norfaunr,but simply history, of which the sources do not have all that much to say. The secondary structures, however, do not taper off until the second half of the Umayyad period. They are manifest in the mass of material on the battle of Siffih30 and in the received version of the Tawwabun;51 the accounts of Mukhtar successfully blur what was clearly a dangerous message and defuse it by systematic ridicule,32 while those of Shablb and Mutarrif, the Kharijites in the days of Hajjaj, conversely turn minor rebels into prodigious heroes and pinnacles of piety of riveting interest to the chroniclers.35 It is only with the revolts of the Yemeni generals, Zayd b. 'All, 'Abdallah b. Mu'awiya and Dahhak the Kharijite, that we find highly charged events described in fairly neutral terms,54 and by the time of the 'Abbasids, of course, the Sunni attitude had set: 'Alid rebels continue to receive sympathetic attention,35 but the successors of the prodigious Shablb in the Jazira are dismissed in short notices to the effect that they rebelled and were defeated.36 The legal and doctrinal hadiths are thus only one of the problems which the Islamic tradition presents. Were they the only problem, we should still have a fairly good idea of how Islam began; but the basic trouble is that these hadfths are a layer deposited relatively late and that the layer underneath consists of rubble reorganized in minimal order. No scholar in his most extravagant fantasies would dream of reconstructing the Constitution of Medina from its debris in the hadtth about 'All; and yet scholars are doing precisely that when they reconstruct the origins of Islam from its debris in the Islamic tradition. Islamic historiography, however, does not consist only of a religious tradition, but also of a tribal one; and the question to which we must now turn is the extent to which the tribal recollection of the past survived with its structures intact. What the Arabs did with their tribal tradition can best be set out against the background of Iceland. Icelandic and Islamic history unexpectedly sha're the feature of beginning with a hijra: as the future Icelanders made their exodus from Norwegian monarchy in the name of their ancestral freedom,57 so the future Muslims made theirs from Arab paganism in the name of thek ancestral God. And both hijras led into an isolation, physical in the one case and moral in the other, which enabled the mubdjtrun to retain and elaborate the values in the name of which they had walked out.38

Historiographical introduction

9

Hence, for the Icelanders and the Muslims the heroic past was no mere backdrop to history, but history par excellence, the classical age embodying their abiding values and on which their intellectual efforts were spent. Where the Greeks or the Germans remember their jdhiliyya, barbarian past, only from an epic, and others not at all, the Icelanders and the Muslims, by contrast, became assiduous collectors of antiquities relating to the country they had left,39 the exodus,40 and the society which ensued.41 The character of these works is nonetheless very different. Where the Icelandic jdhiliyya merely escaped from monarchy and survived the coming of Christianity, the Arab jdhiliyya by contrast interacted with an Arab religion and state. Hence, where the Icelandic material is either historical or epic in character,42 the Arab material bears all the marks of having been through religious discussions. The Landnamabok and Islendingabok simply recorded the past on the basis of oral tradition collected while the classical society was still in existence, and the Islendingasogur evoked this past in literary works composed during the agonized centuries when the classical society caved in;45 but where Ari recorded and Sturla evoked, the Arabs argued, and the books of futuh and ansdb are thoroughly rabbinicized. The tradition has been broken up. Coherent narratives, though they do exist, are rare;44 and for all that heroic prowess and lapidary style are common enough, the fragments of which the tradition came to consist are so many residues of religious arguments. At the same time pagan timelessness has been replaced by monotheist history. The heroes are sometimes pious and sometimes impious, but of heroic fatalism there is none;45 and where the sagas are pure family history, the futuh and ansdb are that and a good deal more besides. The tribal tradition was, like politics, endowed with religious meaning, and for that reason it did not escape the ravages of the whirlwind. There is no qualitative difference between the tribal and the strictly religious material in the Sira, the Constitution of Medina being once more the only exception;46 accounts of the conquests, insofar as they do not consist of legal and doctrinal hadiths, are formulaic and schematized;47 tribal and religious history up to the accession of Mu'awiya are largely beyond disentanglement ;48 and the careers of the Umayyad ashrdfdxt as stereotyped as the accounts of Umayyad fiscal policy.49 It is, however, undeniable that the tribal tradition was located off the centre of the whirlwind, and suffered less damage as a result. Where the Stra is marked by secondary constructions, the ayydm are simply legendary ;5° there is occasional material relating to the period between the ridda and the first civil war, above all in Sayf b. 'Umar, which is strikingly alive;51 and there is still more relating to the subsequent period, and above

io

Introduction

all the second civil war, which is manifestly historical.52 The fact that material is alive does not necessarily mean that it is true, but it does mean that it has been through an undisturbed transmission such as the religious tradition did not enjoy: of the Prophet the tribesmen remembered nothing, but of their own history they obviously did remember something. But it is not much, and what is worse, much of it is of very little use. What the tribal tradition preserved was above all personalia: who married, divorced and killed whom, who was the first to say and do such and such, who was the most generous of the Arabs, what so-and-so said on a certain occasion, and so forth, in short the chit-chat and gossip of the Arab tribal sessions. Of such material a ninth-century scholar was to make an entire collection, the Kitdb al-muhabbar, which must rank with the Guinness Book, of Records among the greatest compilations of useless information.53 It was material which was well equipped to withstand the effects of atomization, and it was, of course, precisely the stuff of which the Icelanders made world literature; but it is not the stuff of history. Whether one approaches Islamic historiography from the angle of the religious or the tribal tradition, its overall character thus remains the same: the bulk of it is debris of an obliterated past. The pattern in which the debris began to be arranged in the eighth century A.D. acquired the status of historiographical sunna54 in the ninth, the century in which the classical works of history and hadith were compiled. The tradition did not, of course, entirely cease to change on reduction to writing, but basically the canon had now been closed and endowed with the same kind, if not quite the same degree, of sanctity as that which was attached to the Prophet's words; and both were passed on without substantial modifications, complete with ikhtilaf and ijma\ disagreement and agreement. The works on which the canon was based were compilations pure and simple. Had historical works composed before the subsidence of the temptestuous weathers come down to us, we might very well have had the excitement of seeing early Islamic history through independent minds; but because the tradition has been shattered, all the later historians could do was to collect its remains.55 The works of the first compilers —Abu Mikhnaf, Sayf b. 'Umar, 'Awana, Ibn Ishaq, Ibn al-Kalbl and so forth — are accordingly mere- piles of disparate traditions reflecting no one personality, school, time or place: as the Medinese Ibn Ishaq transmits traditions in favour of Iraq, so the Iraqi Sayf has traditions against it ;56 'Awana, despite his Syrian origins, is no Umayyad zealot;57 and all the compilations are characterized by the inclusion of material in support of conflicting legal and doctrinal persuasions.58 Inasmuch as the classical sources consist largely of extracts or free

Historiographical introduction

11

renditions of these works, they could not easily be very different in character. We have an apparent abundance of rich and diversified sources for the history of the first two centuries. Sunnis and Shi'ites, Iraqis and provincials, Arabs and Persians all contributed over the years to the mountain of universal chronicles, local histories, genealogical works, biographical dictionaries, legal handbooks, collections of poetry, of proverbs and of gossip, heresiographies, polemical tracts and essays which shield the Muslim past from the unholy designs of modern historians.59 But the diversity is depressingly deceptive. Ya'qubl gives us nothing like the Shi'ite experience of Islamic history, merely the same body of tradition as the Sunni Tabarl with curses in appropriate places ;6° similarly local historians such as Azdl have no local experiences and few local sources, but merely pick out from the canon what was of local interest;61 compilers of biographical dictionaries picked out theirprosopa, jurists and historians their hadiths on taxation, and Persian historians simply translated their selections into Persian; Baladhuri's A.nsab is a universal chronicle genealogically arranged, Ibn 'Asakir's Ta'rikh a biographical dictionary topographically based, and so on adinfinitum : where ever one turns, one finds compilers of different dates, origin and doctrinal persuasions presenting the same canon in different arrangements and selections.62 This does of course have its practical advantages. Inasmuch as every compiler will have bits of the canon not found elsewhere, one can go on finding new material even in late sources; and in theory one ought to read the entire corpus of Muslim literature on the period before venturing an opinion on what it was about.63 But in practice, of course, this is not feasible, and one all too soon reaches the point of diminishing returns: in a late local chronicle written in Persian such as the Tdrikh-i Sistan there is admittedly bric-a-brac which is not found elsewhere; but there is little else. The source material thus consisted of an invariable canon formecl between a hundred and fifty and two hundred years after the Prophet's death. It is for that reason that it is so extraordinarily impenetrable. Passing from one source to another and finding them very much the same, one is harassed by an exasperating feeling that one cannot see. And in fact one cannot see. Whoever comes from the Mediterranean world of late antiquity to that of the Arab conquerors must be struck by the apparently total lack of continuity: the Syria to which Heradius bade his moving farewell seems to have vanished, not just from Byzantine rule, but from the face of the earth. Nothing in the Arab accounts of the conquests betrays the fact that the Arabs were moving into the colourful world described by historians

12

Introduction

of late antiquity: in the east the Arabs saw kisras and marqtbdns, in the west qaysars and batrfqs, but of whatever else they saw, they took no notice ;64 and for the better part of the Umayyad period, the only non-Muslim presence to come through in the sources is that of Khurasan. The Syrian pillar saints dispensing grace to local Arab tribesmen, the Coptic peasants, riotous Alexandrines or sophisticated Nestorians at home at the King of Kings' court, all these have been conjured away at a stroke and replaced by faceless 'uluj and nasara: one comes straight from late antiquity to classical Islam.65 Unvaried and impenetrable, the tradition is also marked by an extraordinary unreality. The accounts which the sources push at us never convince, and if one accepts the descriptions of Muhammad's years in Mecca, 'All's fiscal policy in Kufa or the course of the battle of Siffln, it is because the sources offer no alternatives, not because they ring true. In part, of course, this unreality arises from the fact that what the sources would have us believe cannot be true: new religions do not spring fully-fledged from the heads of prophets, old civilizations are not conjured away. But more particularly it reflects the circumstance that the tradition which the sources preserve was dead; for whereas the epic has compelling verisimilitude even when its information is wrong, the Islamic tradition is completely unpersuasive even when its information is correct. Thus Noth dismisses the use of takbirs as battle cries as a mere literary topos,66 and as it happens a Syriac source proves him wrong ;6? but had it not been for the Syriac source, who other than the most ijihiri of historians would have believed it?68 The epic evokes a lived experience, but the Islamic tradition had been through too many upheavals to retain much vividness: true or false, it has all become dust in the eyes of the historians. But above all the tradition is marked by high entropy. Unsurprisingly, it is full of contradictions, confusions, inconsistencies and anomalies, and if these could be ordered a certain meaning might emerge. But the debris is dejectingly resistant to internal criticism, and because it cannot be ordered, nothing much can be proved or disproved. There is nothing, within the Islamic tradition, that one can do with Baladhuri's statement that the qibla in the first Kufan mosque was to the west :69 either it is false or else it is odd, but why it should be there and what it means God only knows. It is similarly odd that 'Umar is known as the Faruq, that there are so many Fatimas, that 'All is sometimes Muhammad's brother,70 and that there is so much pointless information; but all one can do is to note that there are oddities, and in time one gets inured to them. It is a tradition in which information means nothing and leads nowhere; it just happens to be there and lends itself to little but arrangement by majority and minority opinion.

Historiographical introduction

13

The inertia of the source material comes across very strongly in modern scholarship on the first two centuries of Islam. The bulk of it has an alarming tendency to degenerate into mere rearrangements of the same old canon — Muslim chronicles in modern languages and graced with modern titles. Most of the rest consists of reinterpretations in which the order derives less from the sources than from our own ideas of what life ought to be about — modern preoccupations graced with Muslim facts and footnotes.71 This combination of traditional rearrangement and modern preoccupations does little to uncover the landscape that we are all trying to see: things can occasionally be brought to fit, but one all too rarely experiences illumination.72 And for the same reason new interpretations do not generate much in the way of new research. Theories and facts do not mesh, paradigms produce no puzzles and puzzles no paradigms:73 we are forever shifting rubble in our own peculiar field without appreciable effect on the work of our successors or that going on in adjoining areas. Hence what patterns we opt for hardly seems to matter: maybe Muhammad was a Fabian socialist, or maybe he merely wanted sons; maybe the Umayyad feuds were tribal or maybe that was how Umayyad politicians chose to argue. What difference does it make? We know as little as and understand no more than before. The inertia of the source material is similarly reflected in the inordinate time it has taken for a helpful Quellenkritik. to emerge. In 1899 Wellhausen applied to Islamic historiography the principles oi literary criticism which had paid off so handsomely in his study of the Pentateuch; and since in both cases he was up against tribal and religious traditions belatedly committed to writing, one might have expected his 'Prolegomena zur altesten Geschichte des Islams' to have been as revolutionary a work as was his Prolegomena 7&r altesten Geschichte

Israels.14 But it is not altogether surprising that it was not. The Biblical redactors offer us sections of the Israelite tradition at different stages of crystallization, and their testimonies can accordingly be profitably compared and weighed against each other.7 5 But the Muslim tradition was the outcome, not of a slow crystallization, but of an explosion; the first compilers were not redactors, but collectors of debris whose works are strikingly devoid of overall unity; and no particular illuminations ensue from their comparison. The Syrian, Medinese and Iraqi schools in which Wellhausen found his J, E, D and P, do not exist: where Engnell and other iconoclasts have vainly mustered all their energy and ingenuity in their effort to see the Pentateuch as a collection of uncoordinated hadiths,16 Noth has effortlessly and conclusively demonstrated the fallacy of seeing the Muslim compilers as Pentateuchal redactors.77 After Wellhausen the most striking feature of Islamic Quellenkritik

14

Introduction

its absence. It was only in 1967 that Sellheim published his stratigraphy of the Sira, a work notable for its failure to relate itself to either Wellhausen or Schacht, and for its definition of a Grundschicht so broad that the basic problems of the formation of the Prophet's biography were evaded.78 And not until 1968 did Wellhausen's ideas begin to be taken up by Noth.79 Noth himself has adopted a form-critical approach, and the result is both enlightening and wholly negative. Form-criticism is, like literary criticism, a method evolved for the study of the Pentateuch. Biblical form-critics treat Wellhausen's redactions rather as conglomerates in which each individual component has its own individual history, and in pursuing these they take us back in time. But just as the Islamic tradition is not the product of either slow crystallization or a gradual deposition of identifiable layers, so also it is not a conglomerate in which ancient materials have come together in a more recent setting. Hence, where Biblical form-critics take us back in time, Noth by contrast takes us forward. He demonstrates time and time again that the components of the Islamic tradition are secondary constructions, the history of which we are not invited to pursue: they simply have to be discarded. Where Biblical form-criticism takes us to the sources behind the sources, Noth exposes us to a gaping void behind the sources. And the practical outcome of his Quellenkritik is accordingly not the rewriting of Islamic history, but a warning to foolhardy Islamic historians.80 By far the most important contributions, however, have come from the field of Hadlth. Here too there was a notable delay. Already in 1890 Goldziher demonstrated that the bulk of the traditions attributed to the Prophet in fact originate in the doctrinal and legal controversies of the second and third centuries of the hijra*1 and his ideas were taken up by Lammens and Becker.82 But thereafter the implications of Goldziher's theories were quietly forgotten, and not until the 1940s did they receive systematic development at the hands of Schacht.83 With Schacht, however, things did begin to move. His work on Islamic law for the first time related atomistic hadiths to time and place and used them for the reconstruction of an evolution,84 a feat which has generated the first and as yet the only line of cumulative research in early Islamic studies.85 At the same time his work on Islamic historiography demonstrated that second-century hadiths abound in the accounts of the Prophet and the Rashidun,86 and that the earliest historiographical literature took the form of dry lists of names chronologically arranged87 — ta'rikh as opposed to hadlth and akhbar. Among historians the response to Schacht has varied from defensiveness to deafness,88 and there is no denying that the implications of his theories are, like those of Noth, both negative and hard to contest. That the bulk

Historiographical introduction

15

of the Sira and lives of the Rashidun consists of second-century hadlths has not been disputed by any historian,89 and this point may be taken as conceded. But if the surface of the tradition consists of debris from the controversies of the late Umayyad and early 'Abbasid periods, the presumption must be that the layer underneath consists of similar debris from the controversies which preceded them.90 The fact that so much of the Sira has no apparent doctrinal point is not, of course, a proof of its historicity: of the lives of prophets little is remembered or invented unless it has a point. And the pointlessness testifies, not to the extraordinary detachment of seventh-century Arab reporters, but to the extraordinary erosion of seventh-century religious and historical structures.91 The question which Schacht's theories beg is whether the chronological and prosopographical skeleton identified by him as the Grundschicht of the Sira can withstand critical inspection, and it is remarkable, but perhaps not insignificant, that no historians have so far rushed to its defence. It cannot withstand such inspection. The chronology of the Sira is internally weak,92 schematized,93 doctrinally inspired,94 and contradicted by contemporary non- Muslim sources on one crucial point.95 And that the prosopography shares these features needs hardly to be pointed out.96 There is of course no doubt that Muhammad lived in the 620s and 630s A.D., that he fought in wars, and that he had followers some of whose names are likely to have been preserved. But the precise when, what and who, on which our interpretations stand and fall, bear all the marks of having been through the mill of rabbinic arguments and subsequently tidied up. As far as the origins of Islam are concerned, the only way to escape the entropy is thus to step outside. It is our luck that, unlike historians of the Buddha, we can step outside: all the while that Islamic historians have been struggling with their inert tradition, they have had available to them the Greek, Armenian, Hebrew, Aramaic, Syriac and Coptic literatures of non- Muslim neighbours and subjects of the Arab conquerors, to a large extent edited and translated at the end of the last century and the beginning of the present, and left to collect dust in the libraries ever since. It is a striking testimony to the suppression of the non-Islamic Middle East from the Muslim sources that not only have these literatures been ignored for questions other than the chronology of the conquests and the transmission of Greek philosophy and science, but they have also been felt to be quite rightly ignored.97 Of course these sources are hostile, and from a classical Islamic view they have simply got everything wrong; but unless we are willing to entertain the notion of an all-pervading literary conspiracy between the non-Muslim peoples of the Middle East, the

16

Introduction

crucial point remains that they have got things wrong on very much the same points. That might not, it is true, have impressed the medieval Muslims who held the Jews and Christians capable of having maliciously deleted from their scriptures precisely the same passages relating to the coming of Islam; but as the Jews and Christians retorted, given their wide geographical and social distribution, they could scarcely have vented their anti-Muslim feelings with such uniform results.98 It is because there is agreement between the independent and contemporary witnesses of the non-Muslim world that their testimony must be considered; and it can hardly be claimed that they do not help: whichever way one chooses to interpret them, they leave no doubt that Islam was like other religions the product of a religious evolution. Stepping outside is, however, not the only solution as far as the political history of the Arabs after the Rashidun is concerned." Here too the Grundschicht consists of a chronological and prosopographical framework, and that the Arab horror anonymitatis contributed to the proliferation of names here as elsewhere can hardly be open to doubt ;IO° but the lists include the names of governors who can be checked against the evidence of numismatics, papyrology and epigraphy, and against the testimony of nonMuslim sources, and the result of such a check is unshakeable, surprising and impressive agreement.101 Who compiled these lists, when and why is one of the most intriguing problems of Islamic historiography ;102 but what matters in the present context is that the one thing we can pride ourselves on knowing in early Islamic history is who held power and when. It is thus not surprising to find that whereas the non- Muslim sources offer a wholly new picture of the religion that was to become Islam, they generally confirm the familiar outline of the society that was to become the Muslim polity;103 and since they do not usually offer many details, their importance is necessarily reduced. Not that this does much to justify the reluctance of Islamic historians to touch a non-Muslim source. Syriac sources offer a contemporary account of the revolt of Mukhtar,104 descriptions of a proto- mamluk. army under Mansur105 and a slave revolt in Harran;106 and had it occurred to Dennett to glance at a collection of Nestorian responsa edited, translated and indexed in 1914, he would not have had to write his Conversion and Voil-tax in 19 5 o to prove that the Arabs did indeed impose a tax on the unbelievers' heads.107 But the fact remains that for political history the non- Muslim sources offer additional, not alternative, information. The obvious way to tackle early Islamic history is, in other words, prosopographical. To the extent that the pages of the Muslim chronicles are littered with names, prosopography is of course nothing but a fancy

Historiographical introduction

17

word for what every historian of that period finds himself to be doing. But early Islamic history has to be almost exclusively prosopographical. There is, to be sure, a scatter of tribal traditions and stereotypes which can be used, but the vast mass of information is gossip which cannot be used for what it asserts, only for what it conveys, primarily the background and status of the persons gossipped about.108 The gossip provides a context for the men in power, and without such context the lists would be of little use to us. But it does not provide much else.

2 THE NATURE OF THE ARAB CONQUEST

A Ch'i-tan prince of the Liao dynasty in China once caught a servant reading a book. It was an embarrassing moment, for needless to say the book was Chinese; hurriedly, the prince hid the book in his sleeve and cautioned the servant in future to do his reading in secret.109 Similarly, 'Umar I once caught an Arab copying the book of Daniel. It was not, however, an embarrassing but a frightening moment, for 'Umar thrashed the man repeatedly to the accompaniment of the verse 'we have revealed to you an Arabic Koran' until the wretched victim cried out that he repented.110 There could be no question of reading Daniel or other foreign writings on the side. These stories nicely catch the contrast between Central Asian and Arab conquest. The Ch'i-tan episode was merely one of the many undignified moments the Central Asian conquerors had to endure in the course of their invariably vain attempts to resist Sinification. There were few who were not determined morally to stay in their ancestral 'forests of OtukanY11 and yet there were none who escaped at least a measure of cultural assimilation: even the Ch'i-tan, who had so aggressively insisted that they possessed a respectable civilization of their own,112 took a Chinese type of administration with them when they escaped to western Turkestan."3 And no barbarian conquest of China ever resulted in the formation of a new civilization. But the Arabs had conquered the Middle East in the name of a jealous God, a God who dwelt among the tribes and spoke in their language, and morally they did remain in Mecca: where the Ch'i-tan could at best translate the Chinese classics, the Arabs read their own Koran and tribal poetry. And, culturally, the outcome of thenconquest was indeed a new civilization: where the Ch'i-tan adhered to Chinese values even in western Turkestan, the Syrians and the Persians adopted Arab values even in the settled Middle East. What are the features of Central Asia and Arabia in terms of which this extraordinary difference can be explained? We may start with Central Asia, the paradigmatic home of barbarian conquerors. Two points are relevant here. Firstly, in terms of ecology 18

The nature ofthe A rab conquest 114

19

Central Asia is a land of steppe. The steppe is close to ideal pastoral land: if generally too poor to be exploited by agriculture,"5 it is generally also too rich to be wasted on camels.116 The steppe pastoralists can keep a wide variety of domestic animals,"7 and above all they can specialize in horses. Horses permit a high ratio of livestock to man,"8 and moreover they are in the nature of cash cattle."9 It is perhaps for these reasons that the process of sedentarization through excessive wealth and poverty, which elsewhere siphons off the top and the bottom of the social pyramid,120 scarcely seems to have been operative on the steppe: just as vast herds could be accumulated before the point of diminishing returns was reached,121 so a large number of impoverished tribesmen could be kept in business as shepherds.122 Central Asian tribes thus disposed of greater endogenous resources than is common in a pastoral context and possessed a correspondingly high potential for internal organization. Secondly, in terms of geopolitics Central Asia was a huge sea of barbarians set in the midst of interlocking continents. Thanks to its border on the Siberian forest in the north, it was open to barbarian incomers who would upset existing polities and set migrations going.123 And being surrounded by the four civilizations of China, India, Iran and the West, it was the recipient of a steady flow of moral and material resources from the settled states, some arriving in the form of imperial subsidies, others being left behind by the commercial and diplomatic caravans that traversed the steppe, or by the garrisons and missionaries who followed in their wake: as Greek fabrics, Graeco- Roman masks and Chinese lacquers could find a common grave in Pazyryk and Noin Ula, so Manichaeans, Nestorians and Buddhists were all to reach the Orkhon at their appointed times.124 There were thus two external sources of commotion in the barbarian sea, and the very size of the sea was such that the commotion could reach gigantic proportions.125 Hence the potential for internal organization was liable to be encashed. This was particularly so in the area along the Chinese wall. Mongolia though split into two by the Gobi, formed a compact steppe wide open to the forests in the north, but in head-on collision with a compact civilization in the south — a situation very unlike the patchwork of oases, steppe and desert which constituted the Transoxanian border of Iran, where nobody could dream of building a single wall to keep the barbarians out.126 Tribes certainly did enter Mongolia from the north,127 and Chinese resources certainly did pour into it from the south, but the only safety-valve was the narrow Jungarian corridor to the west. Accordingly, very high pressure could be built up in the Mongolian steppe, and for this reason Mongolia was the classic site of Central Asian state formation.128 Generally, Turkish

20

Introduction

and Mongol tribes are highly stratified. The lineages are ranked in the order of seniority on the principle that no man is his brother's equal,129 and divided into two estates, nobles and commoners, the 'white' and the 'black bones' respectively,150 and these may be further subdivided. The nobles collect taxes and services from the commoners and command them in war, and descent groups are or may easily be reorganized as units of internal administration.131 Periodically, the warfare endemic in Mongolia would act on these tribes to produce larger political structures. On the one hand, a chief became indispensable: 'a body fares badly without a head', as the conventional wisdom had it.132 And on the other hand, kinship ties were slowly being ground away by the savagery and length of the struggles.133 As tribes were broken up by dispersal or enslavement,134 social stratification encroached on segmentation,135 free retainers clustered around nobles and chiefs,136 and one of these would eventually subdue his neighbours, distribute them in military and administrative units headed by his vassals, kinsmen or the nobles,137 and commonly mark the foundation of his polity by the promulgation of laws.138 It was thus possible to found a state in the steppe; but the caravan trade notwithstanding, it was scarcely possible to maintain it there. Hence such states had no option but to conquer, and from the Hsiung-nu in the third century B.C. to the Manchus in the seventeenth century A.D. northern China was the seat of a long succession of barbarian states bent on the absorption of their rivals along the wall on the one hand, and the annexation of the fertile lands behind it on the other. At the other end of Central Asia, by contrast, the steppe was sprawling and civilization well tucked away behind the Caucasus and the Danube. Here, then, the pressure was deflated. There was of course no lack of tribes coming in from the east, any more than there was lack of revenues pouring in from the south; but on the one hand, the tribes were free to spread out in the almost endless steppe, and on the other hand, the almost endless steppe lacked a natural centre of expansion: the Crimea was a place of refuge,139 not a place from where to conquer. Typically, therefore, the tribal states of southern Russia were loose structures. A layer of tribal rulers was spread thinly over a local population of pastoralists, peasants and hunters; military organization was usually restricted to a royal bodyguard and an army of nobles; and resources came largely in the form of tribute from the subject population and revenues from whatever trading colony the area might house at the time. The basic structure of the Mongolian states — kings, retainers and an aristocracy in command of the tribes — was usually present, but the tightness was entirely absent, and that holds true of both the Iranian states of the Scythians and Sarmatians,140

The nature of the Arab conquest

21

141

the Germanic ones of the Goths and the Rus, and the Turco-Mongol ones of the Huns, Khazars, Volga Bulgars or Tatars.142 Eventually, of course, civilization began to close in on the steppe: where the Sarmatians could spread into Poland at a slight push from the Alans in the first centuries A.D., the Cumans had to negotiate their entry into Hungary when threatened by the Mongols in the thirteenth, while the Kazakhs scarcely budged when slaughtered by the Kalmuks in the seventeenth and eighteenth. And then the tightly organized states did make their appearance.143 But by then it was too late for world conquest. And until then the barbarians of southern Russia had a simple choice between staying in the steppe at the cost of failure to conquer civilization, and conquest of civilization at the cost of losing their tribal homes.144 In practice they usually stayed in the steppe:145 when the Byzantine empire finally fell to the Turks, it fell to those who came from the east. That is not to say that until then civilization was safe from tribal incursions in the west. Because tribal pressure in Central Asia was concentrated in the east, it was in the west that migrations tended to end up, so that southern Russia became instead a dumping ground for unwanted barbarians. The up-and-coming states along the Chinese walls would send their defeated rivals through the Jungarian corridor, from where the waves of displacement would eventually reach the Russian steppe. Hence where China was typically faced with attempts at concerted conquest by barbarian states such as those of the Hsiung-nu or the Yuan-yuan, Byzantium typically had to endure invasions of barbarian hordes such as those of the Huns or the Avars; while Iran, in all respects in between, suffered a bit of both.146 Europe and Arabia can both be seen as variants on the Central Asian pattern. Up to a point, Northern Europe and Central Asia are directly comparable: along the Rhine, as along the wall, barbarians were in direct confrontation with a compact civilization, and here as there the confrontation engendered state structures among the tribes —kings, comitati, military and administrative divisions,147 and possibly even laws.148 But for one thing, Europe was a land of forests, not of steppe, and thus underdeveloped rather than deprived. And for another, it was peninsular in shape. The sea placed a limit on the barbarians who could come from the north, while the eastern frontier, tiny as it was by Central Asian standards and moreover mountainous in parts, was not at all impossible to defend. Neither the ecological potential nor its dcfensibility can have been very obvious to those who witnessed the days of the migrations. But the Germanic tribes could hardly have overrun the Roman empire without the

22

Introduction

intervention of Central Asia in the shape of the Huns: Germanic state structures were too embryonic and Germanic mobility too limited for a conquest of the Chingizid type ;149 and the Goths who broke the frontier as terrified refugees from the Huns, or the Franks who crossed a deserted Rhine with their cattle and cumbrous ox-carts, were certainly a far cry from the devastating horsemen who swept down on the Chinese.150 Conversely, it is not accidental that the Germanic states which escaped imperial reconquest proved viable, so that in Europe the establishment of barbarian states with a veneer of civilization was cumulative.151 As Charlemagne smote the Avars and baptized the Saxons, so even his feeble successors could withstand the onslaught of the Vikings, whose mighty display of barbarian savagery soon degenerated into mercenary and commercial services to the civilized south, and colonization of the empty north. And though the Mongols could still take the Hunnish road to Hungary, the Cuman refugees did not conquer France. But in Central Asia the barbarian states regularly fell, if not to the Chinese then to other barbarians.152 Central Asia was the region not of stable but of vanished nations, the umam khdliya of the Turks and Mongols who left behind a long tradition of tribal unification and conquest. Evidently, this tradition was primarily about Mongolia and China; but whereas the barbarians of Europe, settled in their stable states, forgot their ambition to replace 'Romania' by 'GothiaY53 those of Central Asia by contrast learnt that there was more to the world than the Orkhon and the Middle Kingdom. On the one hand, it was clear that control of the entire steppe was required for the conquest of the Middle Kingdom;154 and on the other, control of the entire steppe led to awareness of the civilized world beyond it.155 And it was doubtless thanks to this piling up of barbarian experience that Chingiz could conceive the ideas deep in the wilds of Mongolia of conquering the world, so that instead of setting up yet another peripheral state that sent the losers westward, he combined the conquest and the invasion pattern in a single snowballing conquest of both China and the west.156 In Central Asia it was thus barbarian conquest which was cumulative, and it was only in the sixteenth century that the stability of the Mongol polity finally issued in the definitive establishment of civilization there. If Europe was too rich and too well-protected to conform to the Central Asian pattern, Arabia by contrast was too poor and too isolated. Ecologically, the deprivation of the desert is extreme: sheep and goats can be reared only along the edges, but in the interior only camels can subsist.157 And inasmuch as camels demand a considerable investment of labour158

The nature of the Arab conquest

23

without yielding a commensurate return on the market, differentiation of wealth along the lines of the Central Asian pastoralists could not arise.159 Geopolitically, the peninsula was simply a backwater. No tribes pressed south in search of its meagre pastures: what tribal movement there was went in the opposite direction. And what interest the settled states displayed in the peninsula was limited to the fertile strips along the*eastern coast and in the south. l6° The Yemen, for all that it might be Felix, was no China, and by Central Asian standards the traffic it engendered was derisory, so that commercially the Arabs could make it only in the Syrian desert or at sea. Inner Arabia thus remained all but innocent of foreign contamination. There may have been Manichaeans in Mecca just as there may have been Manichaeans in Siberia,161 but against the Manichaeans, Buddhists and Christians who populated the oases of the Tarim basin, writing in some eleven languages and about as many scripts,162 Arabia had only the Jews of Wadil-Qura and the Christians of Najran; and these scarcely even wrote. Hence, where the history of Central Asia is one of endless political upheavals, that of Arabia by contrast is one of tribal immutability: there is not much to tell between the Arabia of the Bible and the Arabia of Musil's Rwala.163 In the sparsely inhabited and uniformly impoverished desert social stratification remained trivial. Tribal nobility, sharafi conferred a prestige as elusive as that of the 'good family' among the bourgeoisie; an acquired status,164 it entailed no formal privileges or bans on intermarriage, and its occupants collected no taxes, transmitted no orders and had no tribal units to command. Similarly, chiefs were invariably peers among equals,165 who did not issue orders as much as formulate a general consensus.166 It is true, of course, that warfare might increase their authority dramatically ; l6? but just as there were few endogenous resources for the chief to work on, so also there was no erosion of kinship ties:168 it is precisely because there were so few resources to fight for that warfare in Arabia never came near the ferocity of the wars between the tribes in Chingiz's Mongolia. Tribes were rarely dispersed: the fate of the Bajila is a marvel, not the norm.169 And still less were they collectively enslaved or executed.170. Where the Central Asian tribes had wars, the Arabs typically had feuds;171 they went to battle for the sake of honour and excitement and occasionally for wells or pastures; but though warfare might trigger the formation of confederacies, it did not lead to states. There is thus no parallel in Arabia to the political tradition of the Mongols or the Turks. The Arabs could scarcely even have afforded the Veblenian waste of human lives, animals and material objects — women, servants, horses, lacquers, textiles — that went into the burial of a Hsiung-nu or a

24

Introduction

Scythian chief.172 And where the Central Asian tribes have a profuse and eclectic vocabulary of political titulature,173 the Arabs made good with shaykh, sayyid and a few other terms. The halifw&s no retainer,174 the chiefly slaves no ordo, military and administrative divisions appear only after the conquests,175 and there never was an Arab Ydsa. It was of course possible for the Arabs to have states in the Syrian desert, where civilization was ready to assist with commercial revenues and imperial subsidies. But by the same token such states were forced into undignified dependence. In times of imperial weakness they might either pander to civilization as commercial statelets as they did in Petra, Hatra and Palmyra, or infiltrate it as settled kings as they did in Emesa and Edessa; but those who like the Nabateans or Zenobia tried both commerce and conquest were a shortlived menace at best, and culturally they were no menace at all.J 7
View more...

Comments

Copyright ©2017 KUPDF Inc.
SUPPORT KUPDF