Muslim Spain and Portugal PDF

Muslim Spain and
Portugal
A Political History of al-Andalus
Downloaded by 95.180.49.217 at 13:07 29 August 2017
HUGH KENNEDY
First published 1996 by Pearson Education Limited
Published 2014 by Routledge
2 Park Square, Milton Park, Abingdon, Oxon 0X14 4RN
711 Third Avenue, New York, NY 10017
Routledge is an imprint of the Taylor & Francis Group, an informa business
Copyright 1996, Taylor & Francis
All rights reserved. No part of this book may be reprinted or reproduced or utilised in any
form or by any electronic, mechanical, or other means, now known or hereafter invented,
including photocopying and recording, or in any information storage or retrieval system,

without permission in writing from the publishers.
Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional
practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described
herein. In using such information or methods they should be mindful of their own safety
and the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or
editors, assume any liability for any injury and/or damage to persons or properly as a
matter of products liability, negligence or otherwise, or from any use or operation of any
methods, products, instructions, or ideas contained in the material herein.
ISBN 978-0-582-49515-9 (pbk)
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging in Publication Data
Kennedy, Hugh (Hugh N.)
Muslim Spain and Portugal: a political history of Al-Andalus /
Hugh Kennedy,
p. cm.
Includes bibliographical references and index.
ISBN 0-582-49515-6 (PPR).ISBN 0-582-29968-3 (CSD)
1. SpainHistory711-1516. 2. MuslimsSpainHistory.
3. PortugalHistoryTo 1385. 4. MuslimsPortugalHistory.
I. Title.
DP102.K46 1996
946'02dc20 96-22764
CIP
Set by 35 in Baskerville 10/12pt
Contents
L i s t of M a p s viii
Acknowledgements
Note o nNames and Dates xi

Glossary xii
Introduction xiv
1. T h e C o n q u e s t a n d the A g e of the A m i r s , 711-56 1

The Iberian background 1
The Muslim conquest, 711-16 3
The settlement of the Muslims 16
The early governors, 714-41 18
The coming of the Syrians, 741-56 23
2. T h e Umayyad Amirate, 756-852 30

'Abd al-Rahmn I and the establishment of the
Umayyads, 755-88 30
The reigns of Hishm I and al-Hakam I, 788-822 38
The reign of 'Abd al-Rahmn II, 822-52 44
The administration of the early Umayyad amirate 49
The lands of the northern frontier, 788-852 54
The Iberian peninsula i n the ninth century 59
3. Muhammad, al-Mundhir and A b d

63
Allh: the S l i d e i n t o A n a r c h y , 8 5 2 - 9 1 2
The accession and early years of the A m i r
Muhammad 63
The growing crisis of the Umayyad amirate, 852-86 67
The reign of ' A b d Allh and the decay of
the amirate, 888-912 73
4. T h e G o l d e n A g e of the Umayyad Caliphate,

912-76 82
V
vi Muslim Spain and Portugal
The policies and strategies of 'Abd al-Rahmn III 82

The years of expansion, 912-29 87
Umayyad intervention i n North Africa 95
The later years of 'Abd al-Rahmn III, 939-61 97
The reign of al-Hakam II al-Mustansir, 961-76 99
The caliphate i n Mediterranean society 106
5. The 'Amirids and the Collapse of the Caliphate

of Cordoba 109
The rise of al-Mansr, 976-81 109

Al-Mansr i n power, 981-1002 115
The rule of al-Muzaffar, 1002-8 122
The collapse of the Cordovan state, 1008-31 124
6. The Taifa Kingdoms 130

The emergence of the Taifa kingdoms 130
Arab patrician Taifas: Seville, Zaragoza and Cordoba 134
The ' o l d ' Berber Taifas: Badajoz, Toledo and
Albarracin 138
The saqliba Taifas of the Levante 140
The Zirids of Granada and other 'new' Berber
Taifas 141
Taifa politics 143
The Christian advance, 1057-86 149
7. The Empire of the Almoravids 154

The origins of the Almoravid movement, c. 1050-86 154
The Almoravid conquest of al-Andalus, 1086-1102 161
The establishment of the Almoravid regime i n
al-Andalus 166
The reign of ' A l b. Ysuf b. Tshf n: the years of
victory, 1106-17 172
The structure of Almoravid government i n
al-Andalus 174
The reign of 'Al b. Ysuf b. Tshf n: the years of
decline, 1118-42 179
8. The Second Taifas 189
9. The Early Almohad Caliphate 196

The origins of the Almohad movement, c. 1100-30 196
Contents vii
'Abd al-Mu'min and the construction of the

Almohad caliphate, 1130-63 200
The caliphate of Ab Ya'qb Ysuf: the early years,
1163-72 216
The Huete campaign, 1172 223
Ab Ya'qb Ysuf: the later years, 1173-84 231
10. The Later A l m o h a d Caliphate 237

The caliphate of Ab Ysuf Ya'qub
al-Mansr, 1184-99 237

The caliphate of al-Nsir, 1199-1213 249
The collapse of A l m o h a d rule i n al-Andalus, 1213-28 256
The twilight of al-Andalus, 1228-48 266
11. T h e N a s r i d s of G r a n a d a 273
Ibn al-Ahmar and the foundation of the Nasrid
kingdom, 1232-73 273
Nasrids, Merinids and Castilians, 1273-1333 280
The golden years of the Nasrid kingdom:
Ysuf I and Muhammad V , 1333-91 288
Internal divisions and external threats, 1391-1464 292
The decline and fall of the Kingdom of Granada,
1464-92 299
F a r e w e l l to a l - A n d a l u s 305
A p p e n d i x 1: G o v e r n o r s of a l - A n d a l u s a n d
Taifa Kings 309
A p p e n d i x 2 : F a m i l y Trees of the R u l i n g
D y n a s t i e s of a l - A n d a l u s 312
Bibliography 316
Maps 325
Index 329
L i s t of Maps
1. Al-Andalus (Spain and Portugal) 326-7
2. Al-'Udwa (Morocco) 328

viii
This book is dedicated to my daughter
Katharine, in memory of happy days in
Ubeda and Granada.
ix
Acknowledgements
It is with genuine and heartfelt pleasure that I wish to acknow-

ledge the help and encouragement given to me i n the writing of
this book by colleagues i n Spain. In this context I would particularly

like to mention Eduardo Manzano Moreno, Maribel Fierro, Manuela
Marin and Mercedes Garcia-Arenal who have welcomed me at the
Concejo Superior de Investigaciones Cientificas i n Madrid, encour-
aged me i n this work and introduced me to the writings of other
Spanish authorities which I might otherwise have neglected. It has
been a privilege to work with such a lively and interesting group of
scholars.
I owe a major debt of gratitude to Andrew M a c L e n n a n of
Longman, without whose encouragement this book would never
have been written and without whose patience it might well have
been cut off before it could be finished. It is easy to forget how
important enthusiastic and knowledgeable history editors, like
Andrew, are i n the development of the subject.
I would also like to thank generations of St Andrews students
who have sat through my courses on Christian and Muslim Spain
and provided both audience and critique for many of the ideas
presented here. My colleagues i n the Department of Mediaeval
History here have made this possible by encouraging the develop-
ment of courses on Islamic History. I am grateful to the School of
History for providing funds for visiting Spain and to Rob Bartlett
for his companionship on a trip through the more remote parts of
Extremadura. As always I must thank Helen and Robert Irwin for
providing such a welcoming base i n L o n d o n . A n d last but not least
my wife Hilary, for unfailing love and encouragement.
x
Note on Names and Dates
I have tried to make this book user-friendly to non-Arabists while

remaining faithful to the evidence. Personal names are transliter
ated according to the Cambridge History of Islam system, now gener

ally used i n English language publications. Place-names are given
in the modern Spanish and Portuguese forms for the Iberian
peninsula, and standard modern transliterations for N o r t h Africa.
Al-Andalus refers to those areas of the Iberian peninsula which
were under Muslim rule at the time being described, so that, for
example, it is much smaller i n 1200 than i n 1000. Andalucia al
ways refers to the modern region of that name. I have used the
English term Morocco to describe that part of the Maghreb which
lies within the boundaries of the modern kingdom but which was
known to the Muslims of al-Andalus as the 'Udwa, the land on
the other side of the Straits. I have also sometimes used the ana
chronistic Tunisia i n place of the contemporary but more obscure
Ifrqya. I have used C o m m o n E r a (AD) dates. This sometimes leads
to imprecisions when one Muslim year includes parts of two C E
ones, and where there is uncertainty I have used both, so an event
dated i n the Arabic sources to 500 will be described as occurring i n
1106/7.
xi
Glossary
akhbr: short historical narratives

alcazaba: fortress, usually i n a city (from Arabic al-qasaba)
balod: local, a term used to describe those Muslims who settled i n

al-Andalus before the coming of the Syrians i n 741
cortes: assemblies i n Christian kingdoms of the Iberian peninsula,
including lords, churchmen and representatives of towns
dhimm: a protected person, a term used for Christians or Jews
living under Muslim rule
dnr: Muslim gold coin
dirham: Muslim silver coin
faqh, pl. fuqah': a man learned i n Muslim law
fat: lit. young man, hence slave soldier (cf. ghulm)
ghz: warrior for Islam, one who participates i n the jihd
ghulm, p l . ghilmn: page, hence slave soldier (cf. fat)
hadra: capital city and surrounding area
hfiz: lit. one who knows the Qur'an by heart, used as an adminis
trative title i n the Almohad caliphate
hjib: door keeper or chamberlain, title of chief minister
hashm: army
huerta: fertile irrigated area surrounding cities like Valencia, Murcia
'iqt': land or revenues given i n payment for military service
jihd: Muslim Holy War (cf. ghz)
jizya: poll-tax paid by non-Muslims i n an Islamic state
j u n d : lit. army, hence one of the districts i n the south of al-Andalus
settled by Syrian troops after 741
jund: soldier recruited from one of the junds
ktib, p l . kuttb: secretary
kharj: land-tax, sometimes used for tax i n general
kra: administrative division of al-Andalus
maghrim: taxes not sanctioned by Muslim law
mawl, pl. marvl: client or freedman, sometimes used of all non-
Arab converts to Islam i n the first century A H
muwallad: Muslim from native Iberian stock
xii
Glossary xiii
qd: Muslim judge

ra's, p l . ru'as: chief
Reconquista: Christian reconquest of Muslim Spain and Portugal
rizq: rations given to soldiers as part of their payment
shib al-madna: administrator of an Andalusi city
s'ifa: summer expedition of Muslims against Christians
sayyid: lord, tide given to all members of the ruling A l m o h a d
dynasty
shaykh: o l d man, hence tribal chief or venerable teacher
shurta: police force

sijil: document, usually confirmation of office or property
siqlab , p l . saqliba: Slav, originally used of slave soldiers of eastern
European origin who served i n the armies of Cordoba, later of
all white slave soldiers and mercenaries
sq: market
tlib: lit. student, used as an administrative title i n the A l m o h a d
caliphate
thughr: areas of al-Andalus bordering o n Christian territory, fron
tier zones
'ulam: men learned i n Muslim sciences and law
'ushr, p l . 'ushr: tithes paid by Muslims
Vega: fertile plain to the west of Granada
wl: governor of province
wazr. honorific title given to senior administrators i n Umayyad
times, vizier
za'm: leader
Introduction
This book is intended to provide an account of the political history

of al-Andalus, the parts of the Iberian peninsula under Muslim rule,
between 711, the date of the first Muslim invasion, and 1492 when
the last independent Muslim power, the Kingdom of Granada, was
destroyed.
By political history I do not simply mean the narratives of rulers
and battles, though these are of course important, but also the
understanding of the structures which lie behind political events
and decisions. The most obvious of these structures were the rul-
ing dynasties, where they came from, who their most powerful sup-
porters were and how they attempted to secure a justification and
legitimacy for the exercise of power. The most important function
of a pre-modern Islamic state was the raising and paying of the
military forces. This determined the composition of the elite, the
system of taxation and revenue raising and ultimately the success or
failure of the regime. The structure of the military is an essential
part of political history. Another concern is the reach and range
of government and the extent to which the rulers i n Cordoba,
Seville or Granada were able to make their authority felt through-
out al-Andalus. This i n turn leads to the examination of local elites
and pressure groups and to the consideration of their origins, nature
and power.
This is not a history of the Reconquista. O f course the Christian
powers to the north always affected the history of al-Andalus, and
from the eleventh century onwards they became a threatening and
dominant presence, but the struggle against the Christians was only
one, and not always the most important, concern of the rulers of
al-Andalus: maintaining their own authority i n the Muslim-held areas
was usually the first priority, and the affairs of North Africa were
often as pressing as those of the Christian frontier. This work at-
tempts to see al-Andalus as a Muslim political society among others
like it. Its rulers and administrators were always keenly aware that
their land was part of a wider Muslim commonwealth and it was to
xiv
Introduction XV
this commonwealth, rather than to their northern neighbours, that

they looked for contacts and political ideas. The ultimate failure
and extinction of alAndalus should not be allowed to overshadow
the whole of its 800year history.
This book is not an intellectual and cultural history of alAndalus.
This is not because these things are unimportant, or that the Muslims
of alAndalus d i d not make a major contribution i n these fields, but
simply because they lie beyond the scope of this study except i n so
far as they affected, or illustrate, political developments. Similarly,
there has recendy been much fascinating work o n such topics as
rural settlement, landscape, irrigation technology and cuisine, but

none of these are treated here.
There is a certain unavoidable inconsistency of texture i n this
work. A t some periods we are comparatively well informed about
political events and the scope and operations of government. A t
other periods our sources are much more limited and we can only
discern a bare oudine. N o r is it true that more recent parts of
the history of alAndalus are better known than the earlier ones: we
are well informed, for example, about the reign of alHakam II
(96176) because of the survival of alRz's court chronicle, but
the period 11841210 is an almost complete blank.
Any broadbrush history of this sort is bound to be heavily de
pendent o n the works of others. For the history of alAndalus we
have two major political histories which are classics and remain the
basis for all future research: E. LviProvenal's celebrated Histoire
de l ' E s p a g n e Musulmane, which covers the centuries when Cordoba
was the capital (7111031), and Ambrosio H u i c i Miranda's much
less well known Historia Politica del Imperio Almohade. T o these two
can be added J . Bosch Vila, Los Almoravides, and Rachel Ari's
L'Espagne Musulmane au temps des Nasrides (12321492).
The last two decades have seen a massive increase i n the scope
and intensity of research, which has meant that i n many ways the
history of alAndalus is better known and understood than the his
tory of any other part of the premodern Muslim world, and meth
odologies for treating some important aspects, prosopography and
archaeological evidence for example, are more developed. It is per
haps invidious to single out individuals, but mention should be made
of some of the main advances. The period up to 1031 has been the
subject of intensive study i n Spain and the works of E. Manzano
Moreno o n political structures, and Pedro Chalmeta and M . Barcelo
o n administrative and fiscal history, are fundamental. Also of major
importance are the five volumes of the Estudios OnomasticoBiograficos
xvi Muslim Spain and Portugal
de Al-Andalus edited by Manuela Marin and others which have added

a whole new dimension to our understanding of the Umayyad re-
gime. The period of the Taifa kings i n the eleventh century has
recently been superbly covered i n Los Reinos de Taifas: Al-Andalus
en el Siglo XI, edited by M.J. Viguera Molins as vol. viii of the
Menendez Pidal, Historia de Espana.
In contrast, the periods of the Almoravids and Almohads have
been studied more by French historians. The work of V. Lagardre
has greatly increased our understanding of the Almoravid move-
ment itself, while P. Guichard's Les Musulmans de Valence is an out-

standing work of regional history. The archaeological evidence i n
its broadest context has been studied i n A. Bazzana, P. Cressier and
P. Guichard, Les Chteaux Ruraux d'al-Andalus. There has been less
recent work o n the Almohads and Nasrids, but important contribu-
tions have been made i n R. Ari, Nasrides, L.P. Harvey, Islamic Spain,
1250-1500, and R. Manzano Rodriguez, Los Benimerines.
Numerous other authors, whose works are cited i n footnotes and
in the bibliography, have increased our understanding of the his-
tory of al-Andalus and I am dependent on and grateful to them all.
If this work succeeds i n providing an overview of the subject and
recent research and introducing it to others, be they Orientalists,
western mediaevalists or interested general readers, then it will have
succeeded i n its purpose.
CHAPTER ONE
The Conquest and the Age of the

Amirs, 711-56
The Iberian background
T h e Iberian p e n i n s u l a , divided into the great provinces of

Tarraconensis, Carthaginensis, Baetica, Lusitania and Gallaecia, had
been one of the richest and most developed areas of the western
Roman Empire, but for three centuries before the coming of the
Muslims it had been dominated by warrior aristocracies of Ger-
manic origin. The most successful of these were the Visigoths who
had first entered the peninsula i n the early fifth century. With the
accession of K i n g Leovigild i n 569 the Visigothic monarchy entered
on a century and a half of stability during which kings, based i n
Toledo, exercised effective power.
The kingdom shared many of the characteristics of the post-
Roman kingdoms i n France and Italy. In general the Visigothic
monarchy was a reasonably strong and effective instrument of
government: it d i d not suffer the internal divisions and progressive
debility of the contemporary Merovingian monarchy i n France and,
right up until the Muslim invasions, the kings maintained their
control over most of the Iberian peninsula. In theory the monarchy
was elective and successive church councils of Toledo i n the mid-
seventh century had laid down the rules: the king was to be elected
by the bishops and nobles. H e was to be a catholic Christian, a
Goth by descent and of free birth. H e was to be elected either i n
Toledo or on the site of the previous king's death, and before his
accession he had to swear to uphold the laws of the realm. In
practice, the choice of monarchs was confined to the most import-
ant lineages and there was a natural tendency for fathers to wish
to pass their crowns to their sons, as Leovigild d i d to Recared i n
586 and Egica d i d to Witiza i n 702. A t the same time, there seems
1
2 Muslim Spain and Portugal
to have been a strong feeling among the nobility against the estab-
lishment of a purely hereditary succession and perhaps that a lin-
eage which had held the crown too long should be replaced.
Under the king, the chief men of the state were the nobles,
mostly of Gothic origin, and the bishops, mostly recruited from the
Hispano-Roman landowning class, although the distinctions between
these two groups must have largely disappeared by the beginning
of the eighth century. Besides providing spiritual leadership, the
bishops were also among the largest landowners and most power-
ful political figures i n the land. The nobles, who sometimes bore
the title of dux (duke) or comes (count), were also owners of large,
often underexploited estates cultivated by semi-free peasants. It was
the nobles too who provided the army: apart from a royal guard,
there was no standing army and the nobles brought their followers
in response to the royal summons and the king might reward them
with gifts of gold or silver. In general this simple military system
seems to have functioned fairly successfully, but it probably meant
that the bulk of the troops owed their first loyalties to their lords,
rather than to the monarchy.
It is impossible to make any precise assessment of the popula-
tion or economy of the peninsula. It has been plausibly suggested 1
that the population had been about six million i n the early Roman
period but had been reduced by plague and war to four million by
the later Visigothic period. Archaeological evidence shows that the
large open cities of the earlier period had shrunk into small forti-
fied settlements. Country estates and their buildings were certainly
more primitive than the great latifundia and villas of the imperial
Roman period.
Economic life was almost entirely localised: there is little evi-
dence of long-distance trade and both the small towns and the
large estates were effectively economically self-sufficient. The circu-
lation of coinage was extremely limited and most transactions were
conducted by barter.
Later Visigothic Spain and Portugal was a fairly stable society
and, apart from a limited Byzantine incursion around Cartagena i n
the south-east, there had been no outside invasion for a couple of
centuries. O n the other hand, we can picture a very empty land-
scape, where setdements were few, far between, poor and primitive.
Agricultural resources were i n many cases neglected or underex-
ploited. There were areas, too, notably i n the northern mountains
1. B. Reilly, T h e M e d i e v a l S p a i n s (Cambridge, 1993), p. 7.

The Conquest and the Age of the Amirs, 711-56 3
where the Basques and the Asturians lived, where the people were
totally independent of any form of royal control and where a primi-
tive mountain society vigorously resisted outside control. Such was
the land the Muslims invaded.
The M u s l i m conquest 711-16
The Muslim invasion of Spain and Portugal was i n many ways the
logical and necessary extension of the conquest of North Africa.
Before the coming of the Muslims the area of the modern states of
Libya, Tunisia, Algeria and Morocco was occupied by two groups,
the Byzantines and the Berbers. The Byzantines held a number of
strongholds on the coast, notably T r i p o l i and Carthage, which they
kept supplied by sea. When they reconquered the land from the
Vandals i n the reign of Justinian, they had established an elaborate
system of defences on the southern frontiers of the setded areas,
but these seem to have been abandoned by the mid-seventh cen-
tury when the Muslims began to attack. The remaining Byzantine
garrisons defended their coastal strongholds stubbornly and they
held out much longer than the garrisons i n Syria and Palestine had
done a generation before, but they could easily be bypassed by over-
land invaders and were only a real threat when allied to the Berber
tribes of the area.
The Berbers were the real power i n the land. They were, and still
are, the indigenous inhabitants of North Africa, with their own, u n -
written, language, quite distinct from either Latin or Arabic. Their
social structure was tribal and they seem to have owed their first
loyalties to their k i n . Apart from this, there were wide variations
of lifestyle. Some Berbers were acculturated to the Byzantine world
and many were Christians. Others seem to have lived a much more
separate existence and some at least were still pagans. There were
Berbers who lived i n the cities, many more who lived as farmers i n
2. T h e literature o n the M u s l i m conquest o f al-Andalus is vast a n d fairly uneven

in quality. T h e classic account i n E . Lvi-Provencal, H E M , i, pp. 1-89, is still a use-
ful starting point. A . D . T a h a , T h e M u s l i m C o n q u e s t a n d S e t t l e m e n t of N o r t h A f r i c a a n d
S p a i n ( L o n d o n a n d New York, 1989), is a meticulous, detailed but somewhat u n -
critical account. R. Collins, T h e A r a b C o n q u e s t of S p a i n , 7 1 0 - 7 9 7 ( O x f o r d , 1989), c o n -
tains important insights into Christian life a n d literature o f the eighth century but
is flawed by a contemptuous a n d u n c o m p r e h e n d i n g attitude to the A r a b i c sources.
A m o r e balanced overview o f the p e r i o d by the doyen o f historians o f al-Andalus is
P. Chalmeta, I n v a s i o n e I s l a m i z a c i o n : l a s u m i s i o n de H i s p a n i a y l a f o r m a c i o n de a l - A n d a l u s
( M a d r i d , 1994), a n d I have relied o n this at many points.
mountain villages, some who kept sheep and goats i n the steppe
lands and yet others who wandered as the Touareg do today i n the
remote and awesome wastes of the Sahara. Berber genealogies are
much less fully recorded than Arabic ones and it is difficult to gauge
whether the scattered references we have reflect a static relationship
or whether some groups were expanding at the expense of others.
According to Arab sources, the Berber tribes were divided into
two groups, called Butr and Barnis, just as Arab tribes were divided
into Qays/Mudar and Yemen. Most of the Berbers who joined the
Muslim conquest and settled i n al-Andalus came from the Butr

group. They seem to have retained their tribal identities and prob
ably their pagan religion. In the sixth century these tribes were mov
ing west from Tripolitania and putting pressure on the Byzantine
settlements. The Barnis, by contrast, were older-established tribes
3
who had entered into closer relations with the Byzantines and had
in many cases converted to Christianity. This suggests that the Mus
lims assumed the leadership and gave extra momentum to an exist
ing movement of populations among the Berbers, and this goes
some way to explaining the success and completeness of their con
quests. In practice, the difference between Butr and Barnis seems
to have had little effect on the politics of al-Andalus, unlike the mur
derous disputes between Qays/Mudar and Yemen among the Arabs,
and divisions among the Berber tribes were based on smaller units
of individual tribes and extended families and their relationship
to Arab groups.
The conquest of North Africa had begun as early as 22/642 4
when the conqueror of Egypt, ' A m r b. al-s, led an expedition

to Barqa i n Cyrenaica. From there he dispatched an army to Zawla,
an oasis settlement to the south, led by 'Uqba b. Nfi'al-Fihr. 'Uqba
came from a branch of Quraysh, the Prophet Muhammad's tribe,
and his father Nfi' had been one of the first Muslims to setde i n
Egypt. H e came from the elite of early Islamic society and he used
his position to make contacts among the Berber people of the area,
alliances that were to make his family the most powerful i n North
Africa and al-Andalus before the coming of the Umayyads i n the
mid-eighth century. When ' A m r returned to Egypt, he left 'Uqba
in charge at Barqa.
The conquest of North Africa was difficult, pardy because of
3. O n this see E . M a n z a n o M o r e n o , L a F r o n t e r a de A l - A n d a l u s en epoca de los Omeyas

( M a d r i d , 1991), pp. 2 3 4 - 5 .
4. F o r the conquest o f N o r t h Africa see T a h a , C o n q u e s t a n d Settlement, pp. 5 5 - 8 3 ,
a n d Chalmeta, I n v a s i o n e I s l a m i z a c i o n , pp. 72-94.
political disputes among the Muslims but more because of the vig
orous resistance put up both by the Berber tribes of the interior
and the garrisons of Byzantine cities like T r i p o l i and Carthage o n
the coast. More than the other Arab commanders, ' U q b a seems to
have understood that the key to subduing N o r t h Africa was to enrol
the support of Berber tribes. In 50/670 he founded the Muslim
setdement of Qayrawn, away from the coast, i n the central plain
of Tunisia. Like earlier Arab garrison cities at Kfa, Basra and Fustt,
this was designed to be a setdement where the Muslims could pre
serve their identity and from which they could dominate the sur
rounding country.
In 681 ' U q b a led a spectacular raid to the west i n which he
reached Tangier and the Adantic coast, although there were no
Muslim setdements beyond modern Tunisia at this stage. This
was his last and greatest achievement, but his memory lingered o n
and his sons continued to play a very important role i n the Muslim
politics of North Africa. There followed a period when the Arabs
were almost driven out and Qayrawn itself fell to the Berber leader
Kusayla. The Muslims d i d not recover the initiative until 74/694
when the Caliph ' A b d al-Malik sent an army of Syrians led by Hassn
b. al-Nu'mn al-Ghassn. H e captured the last Byzantine outpost
at Carthage and defeated the Berber leader, the priestess Khina,
and i n 82/701 established himself firmly i n Qayrawn. H e was able
to do this, not only because of his Syrian troops, but because of his
policy of working with the Berbers. Some tribes, like the Luwta,
seem to have remained allies of the Arabs throughout; many others
came over after the defeat of Khina, including her own sons. They
converted to Islam and were enrolled i n the Muslim diwn, receiv
ing a share of the spoils like the Arabs.
Hassn was dismissed by the governor of Egypt, who super
vised the western provinces i n 704, probably because he was too
successful, and was replaced by Ms b. Nusayr, a man of obscure
origins who had risen i n the financial administration of the Umayyad
empire. H e continued Hassn's policy of recruiting converted
Berbers into the Muslim armies and using this new force to extend
his control further to the west until, i n about 90/708, he took
Tangier and appointed a Berber supporter of his, Triq b. Ziyd,
as governor.
The conquest of North Africa had been achieved by an alliance
of Arabs and Berbers i n the name of Islam. As the conquest pro
ceeded, so the importance of the Berber contribution increased.
By the time the Muslims were conquering the area of modern
Morocco, it is probable that the great majority of the troops i n

their army were Berbers, newly converted to Islam. These Berber
troops received a share of the booty but, apart from Triq b. Ziyd
in Tangier, they do not seem to have occupied positions of politi
cal importance. Many Berbers entered intow l 'agreements with
important Arab groups or individuals. They were then described as
m a w l (pl. mawl) of so and so (Triq b. Ziyd, the Berber gov
ernor of Tangier and probably a man of considerable importance
in his own community, for example, is described as m a w l of Ms
b. Nusayr, the Arab governor of Qayrawn). This relationship can

be described as a sort of clientage, by which members of the con
quered peoples were converted (you could not be a non-Muslim
mawl) and given a position i n the Muslim community i n exchange
for their loyalty and support. These networks were very important
5
in the fluid politics of early Muslim North Africa and al-Andalus

and were often more useful than tribal followings i n building up a
power base: both the family of 'Uqba b. Nfi' and the Umayyads de
pended heavily on their m a w l to support their political ambitions.
The governors ( w l or ' m i l are the two Arabic terms used to
describe this office), by contrast, were dependent for their author
ity on the governor of Egypt, and a change of command i n Fustt
(Old Cairo) almost certainly meant a change i n Qayrawn. This
pattern became even more pronounced i n al-Andalus, where the
position of the governor was constandy threatened by changes of
policies or personnel i n Qayrawn or Fustat. In these circumstances,
it was difficult for a commanding personality to establish himself
for long and the governors were often transient figures who made
little impact on the country.
The conquerors fed off further conquests. It is true that subject
Christian and pagan Berbers are said to have been obliged to pay
jizya or poll-tax, but there is litde indication of any formal tax-
gathering machinery. Most of the soldiers must have served i n the
hope of booty and new lands rather than for a salary and Muslim
dominion i n N o r t h Africa had to expand to survive. If the booty
dried up and no new opportunities appeared, then the groups
and tribes would turn i n on each other and disintegration would
inevitably follow. The conquest of Tangier effectively meant the
end of westward expansion; now only Spain could offer the sort of
opportunities the state needed to be able to survive.
Our understanding of the Muslim conquest of al-Andalus and
5. F o r the role o f m a w l in the Umayyad caliphate, see P. C r o n e , Slaves o n Horses.

T h e E v o l u t i o n of I s l a m i c P o l i t y (Cambridge, 1980), pp. 49-57.
The Conquest and the Age of the Amirs, 71156 7
the establishment of Arab rule is hampered by the nature of the

sources. N o contemporary Arabic accounts of the conquest sur
vive and the earliest major sources which have been passed down
to us are collections of historical anecdotes (akhbr) preserved i n
a number of works dating from the tenth century onwards, not
ably the anonymous Akhbr alMajm'a (Collection of Anecdotes) from
possibly c. 940 and the Ta'rkh Iftith alAndalus
6
(History of the Con
quest of alAndalus) of Ibn alQtiya (d. 977). Both these collections 7
arrange their materials more or less i n chronological order but they

are not annals and are more concerned with vivid and interesting
stories than the careful ordering of events. The Akhbris particularly
important for the preUmayyad period, while Ibn alQtiya gives
vivid and gossipy accounts of the courts of the Umayyad amirs.
In the tenth century these accounts were edited and system
atised using the criteria of annalistic historiography developed i n
the eastern Islamic world by such authorities as alMad'in (d. 839)
and alTabar(d. 923). In alAndalus this editing seems to have
been the work of the Rz family, originally from Rayy i n central
Iran, who had come to alAndalus as merchants i n the late ninth
century. According to his son Is (d. 989), it was A h m a d b. Ms
alRz (d. 955) who took the akhbr which people i n alAndalus
had not previously been very interested i n and ordered them
(daxvwana) according to the rules of historical science. The writ 8
ings of the Rzs, father and son, have largely been lost but they
were used, and often incorporated entirely, with acknowledge
ments, by the great eleventhcentury compiler Ibn Hayyn (d. 1076).
M u c h of Ibn Hayyan's work has i n turn been lost, including the
sections which dealt with the conquests and the early amirs. Some
of his material has, however, been preserved i n shorter works, like
the anonymous Fath alAndalus of c. 1100, and later abbreviated9
recensions i n annalistic compilations like Ibn Idhr's Bayn al

M a g h r i l r of about 1300.
1 0
The fact that the sources as they have reached us were written
down at least two centuries after the events has meant that fierce
6. E d . with Spanish trans., E . L afuente y Alcantara ( M a d r i d , 1867). I have used

the dating suggested by C h a l m e t a ( I n v a s i o n e I s l a m i z a c i o n , p. 50) because o f the
archaic nature o f the text, but other authorities prefer an eleventhcentury date.
T h e text is analysed i n detail i n C . SanchezAlbornoz, E l ' A j b a r M a y m u ' a . Cuestiones
h i s t o r i o g r a f t c a s q u e s u s c i t a (Buenos Aires, 1944).
7. E d . with Spanish trans., J . Ribera (Madrid, 1926).
8. Q u o t e d i n Chalmeta, I n v a s i o n e I s l a m i z a c i o n , p. 45.
9. E d . L . M o l i n a (Madrid, 1994).
10. T h e history o f the conquest a n d the A m i r s is covered i n vol. ii, ed. G.S. C o l i n
a n d E . L viProvenal ( L e i d e n , 1948).
controversy has raged about the relative merits and reliability of

these sources. Opinions have varied between historians like Taha,
on the one hand, who accept the Arabic narratives almost com
pletely, and Collins, who holds that the Arabic tradition is virtually
11
worthless. 12
It is important to attempt to assess the reliability of this material.

Clearly these Arab histories are biased i n the sense that they are i n
favour of Muslim victories and claimed that these were the result of
God's support, but this sort of open partisanship does not present
real problems to the modern historian. There are, however, a var
iety of other ways i n which the material needs to be treated with

caution.
There is material which is clearly legendary or folkloric, like the
story of the locked chamber i n Toledo which K i n g Roderick was
rash enough to open, only to find that the interior was covered by
paintings of Arab warriors, and, probably, the story of Count Julian
and the rape of his daughter by K i n g Roderick. These stories, with
their obvious predictive and entertaining functions, are unlikely
to mislead historians. The use of topoi and conventional phrases,
expressions and characterisation borrowed from eastern Islamic
sources may also give a false impression of detailed accuracy. 13
There may also have been more hard-headed reasons for being
economical with the truth. The nature of the conquest affected
the status of the lands conquered: if they were conquered by
force ('anwatan) they became the property of the conquerors, the
indivisible f a y (immovable booty) of the Muslims, and the pro
'
ceeds from these properties were to be used for the benefit of the
Muslims as their ruler saw fit. If the lands were taken peacefully
(sulhan), on the other hand, they continued to be the absolute
property of the inhabitants and would only pass into Muslim hands
by inheritance, purchase or conversion of the owner, i n which
case they would be the absolute property of their Muslim owners.
There is some evidence of two historiographical traditions within
the accounts of the conquests. The first, reported by the Rzs and
14
11. See C o n q u e s t a n d Settlement, pp. 4-14. 12. A r a b C o n q u e s t , pp. 34-5.

13. F o r these problems i n early Islamic historiography i n general, see A . N o t h ,
T h e E a r l y A r a b i c H i s t o r i c a l T r a d i t i o n (new e d n with L.I. C o n r a d , Princeton, 1994); for
a detailed discussion o f similar problems i n the historiography o f al-Andalus, see B.
M u n z e l , F e i n d e , N a c h b a r n , B u n d n i s p a r t n e r (Munster, 1994).
14. F o r these ideas see M a n z a n o M o r e n o , 'Arabes, berberes e indigenas: al-Andalus
en su p r i m e r periodo de formation', paper presented at the Congress o n
I n c a s t e l l e m e n t o , F r e n c h S c h o o l i n R o m e , 1994, pp. 3-12. In press.
other sources close to the Umayyad court, emphasises the force

ful nature of the conquest, since conquest by force would give the
Umayyads the right to dispose of the lands, whereas other accounts
talk of take-over by agreement and so emphasise the rights of the
owners. This may account for disagreements i n the sources about
the nature of the conquest, and such details as the fall of Seville,
which is said to have surrendered peacefully and then rebelled and
had to be subdued by force, may be explained as attempts to con
flate two contradictory traditions. In the end, however, it must be
admitted that these divisions of opinion could simply be the result
of genuine confusion over events which happened long ago.

The fact that these sources, i n the form i n which they have been
handed down to us, are much later need not undermine their
credibility. The Arabic historical tradition laid great emphasis o n
preserving the wording and forms of old accounts and much of the
work of compilers like Ibn Hayyn was basically editing and repub
lishing older materials, rather than composing a new account. Later
chronicles can contain important nuggets of information which
survive from much earlier times: the most important account of the
nature of the settlement of the Syrian junds i n al-Andalus after 741,
for example, is found i n fragments of al-Rz embedded i n the late
fourteenth-century Ihta of Ibn al-Khatib, composed i n its present
form 650 years after the events it describes. 15
In addition to the Arabic texts there are also Latin sources. O f

these by far the most important is the so-called Chronicle of 754, also
known as the Mozarabic Chronicle. This was composed i n al-Andalus,
probably i n the mid-eighth century. Not surprisingly, its viewpoint
is very different from the one presented by the Muslim chronicles,
and for the Christian author, the invasion is a major disaster rather
than a God-given triumph. There are differences over details, but
there is a considerable measure of agreement about the broad out
lines of events. Given its very early date, the evidence of the Chron
icle of 754 must carry great weight and it is helpful that it tends
to corroborate rather than undermine the oudines of the Muslim
tradition. 16
The sources for the Muslim conquest and establishment are as

patchy as they are for most other areas of western Europe i n the
early eighth century. Accidents of survival may play a large part
in shaping our understanding. There is always a danger i n over-
interpreting fragments of information and giving them more
15. See below, p. 50. 16. See Collins, A r a b C o n q u e s t , pp. 2 6 - 3 4 .

importance than they deserve. Despite all these qualifications, how

ever, we can attempt a plausible reconstruction of events.
Most of the early Arabic accounts of the beginning of the
conquest of al-Andalus tell of a Count Julian of Ceuta, perhaps a
Byzantine official with close contacts with the Visigothic rulers of
Spain across the Straits. H e had sent his daughter, as was the cus
17
tom among the Visigothic nobility, to the royal court to complete

her education. Here she was assaulted by K i n g Roderick and com
plained to her father, upon which the outraged governor turned to
the local Muslim commander as an ally i n revenge. Whatever the
literal truth of this story, it probably reflects a situation i n which

the Visigothic K i n g Roderick was resented by an important section
of society who were prepared to call i n help from outside.
During the seventh century, the Visigothic monarchy was both
elective and hereditary. In 693 Egica had associated his son Witiza
in his rule, and when he died i n 702 Witiza succeeded h i m as king.
In his turn, he attempted to do the same for his own son Akhila
and gave h i m the governorate of Narbonne, but when he died
in 710 there was a coup d'tati n Toledo, probably engineered by
nobles who had no wish to see one family retaining the crown for
too long. Power was seized by Roderick, not apparendy a member
of the ruling family, who was able to defeat the army of Akhila and
his brothers Alamundo and Ardabast. In terms of Visigothic prac
tice, Roderick was probably a legitimate king by election, but the
circumstances did mean that there were a number of influential
people at the court who felt themselves wronged; they might well
have caused scandalous stories about the new king to circulate and
would not have been sad to see h i m humiliated by outside invaders.
Like some other Muslim conquests, 'mr b. al-'s's expedition
to Egypt, for example, the invasion of Spain seems to have been
undertaken on local initiative without the approval of the hierar
chy, represented i n this case by the governor of Iftqya at Qayrawn,
Ms b. Nusayr, and ultimately by the Caliph al-Wald b. ' A b d al-
Malik i n Damascus. Majority opinion holds that Triq b. Ziyd,
governor of Tangier, with a force of perhaps 7,000 to 12,000 men,
mostly Berbers, crossed the Straits of Gibraltar i n A p r i l 711 and
established themselves first on the rock which still bears Triq's
name, Gibraltar, before moving on to occupy Algeciras and the
17. F o r recent assessments o f the J u l i a n story, see T a h a , C o n q u e s t a n d S e t t l e m e n t ,

pp. 8 4 - 8 , a n d the full discussion i n Chalmeta, I n v a s i o n e I s l a m i z a c i o n , pp. 112-20.
M u n z e l , F e i n d e , N a c h b a r n , B u n d n i s p a r t n e r , pp. 37-54, has a detailed discussion of the
early sources.
surrounding countryside. Opinions differ as to whether they en

countered any immediate resistance but are agreed that Roderick
and the bulk of the Visigothic army were campaigning i n the north
of the country and it was not until the summer that he was able to
lead his army south to counter the invasion.
There followed a batde which was to determine the future of
the Iberian peninsula for the next eight centuries. It seems to have
lasted for a number of days around 20 July 711. The exact site of
the battle has been the subject of prolonged debate, but it prob
ably lay to the south-east of Medina Sidonia. The sources give very
large numbers for Roderick's army, between 40,000 and 100,000,

and while these are certainly exaggerated, it is likely that Roderick's
army was significandy larger than Tariq's and may have amounted
to between 24,000 and 30,000 men. Arab sources suggest that there
18
were divisions i n the Visigothic ranks and that the brothers of Witiza
at least hoped that the Muslims would defeat Roderick and then
depart, leaving them to assume the crown they felt was rightly theirs.
Whatever the reasons, the Visigothic army suffered a shattering de
feat, Roderick was missing, presumed dead, and members of the
army scattered throughout Iberia, spreading defeatist sentiment. 19
If the brothers of Witiza had imagined that the Muslims would

abandon the fruits of their victory, they were soon undeceived. In
the aftermath of his victory, Tariq moved with speed and confid
ence, aided by Count Julian. H e dispatched a force under Mughth
al-Rm to Cordoba. Here Mughth encountered serious resist
ance from the governor and a small garrison who defended them
selves i n a church after the Muslims had entered the city through a
hole i n the walls: it was three months before they surrendered and
were executed. Malaga and the district of Elvira (the area around
the later city of Granada, which did not really become important
until the eleventh century) seem to have been taken by small units
without much trouble, but Triq himself moved on quickly to the
Visigothic capital at Toledo. There was no resistance. Most of the
people, apart from the Jews, had abandoned the city and Triq was
able to spend the winter of 711-12 there.
Triq's spectacular success attracted the attention of his su
perior, Ms b. Nusayr, i n Qayrawn and he was understandably
eager to associate himself and his supporters with the triumph and
the booty which went with it. The next year he set out with a large
18. Chalmeta, I n v a s i o n e I s l a m i z a c i o n , p. 134.
19. F o r differing o p i n i o n s o n the site a n d course o f the battle, see T a h a , Conquest
a n d S e t t l e m e n t , pp. 8 8 - 9 0 ; Chalmeta, I n v a s i o n e I s l a m i z a c i o n , pp. 132-44.
force, 18,000 men we are told, which included many Arabs. H e

20 21
landed at Algeciras and adopted a strategy which would enable h i m

and his men to make conquests of their own. H e went first to the
great fortress of Carmona, which had been bypassed by Triq, which
was captured with the help of some of Count Julian's men who
pretended to be fugitives and so gained access to the city. H e then
went on to take Seville, which is said to have resisted for some
months before being taken by force. We are given no more details
about this siege and it may be that the resistance was 'invented'
so that the city could be said to have been taken by storm and its
lands confiscated. Ms subdued the neighbouring towns and went

on north to Merida, where the Arab chroniclers comment on the
splendour of the Roman remains. Here there was serious resist
ance, the garrison made a sortie and siege engines were required
to force it into submission i n July 713. While this was going on,
Ms sent his son *Abd al-Azz to the east; when he came to Orihuela,
then the most important city i n the Murcia (the city of Murcia
itself was another later foundation) district, he was met by the local
commander or dux, Theodemir, with whom he made a treaty
whose lenient terms meant effective local autonomy and freedom
of Christian worship i n exchange for goodwill and a modest tribute
to be paid i n cash, wheat, barley, thickened grape juice, vinegar,
honey and o i l . As a result of this, the Murcia area was known to
22
the Arabs as Tudmr for centuries afterwards.

After the fall of Merida, Ms headed for Toledo whence Triq
came to meet h i m . Inevitably, when the two forces d i d j o i n at
Talavera there were tensions and reproaches, but they patched up
their relationship and wintered together i n Toledo. In the spring
of 714 campaigning began again with expeditions which led to the
nominal subjection of Galicia and the Ebro valley.
Their triumphant career was brought to a halt by a summons
to present themselves before the Caliph i n Damascus. Reluctandy,
the two leaders left Spain i n September 714, never to return, Ms
leaving his son A b d al-Azz as governor.
While most of the peninsula had now been visited by Muslim
armies, the conquest was by no means complete. The establishment
20. A k h b r a l - M a j m ' a , p. 15; other sources say 10,000: see T a h a , C o n q u e s t a n d

S e t t l e m e n t , p. 94.
21. F o r Ms's invasion a n d the conquests that followed, see T a h a , C o n q u e s t a n d
S e t t l e m e n t , pp. 94-102, a n d Chalmeta, I n v a s i o n e I s l a m i z a c i o n , pp. 168-205.
22. T h e full text a n d translation can be f o u n d i n C . Melville a n d A . Ubaydli,
C h r i s t i a n s a n d M o o r s i n S p a i n , v o l iii A r a b i c Sources (Warminster, 1992), pp. 10-13.
of Muslim power seems to have occurred i n two stages. The first

was the take-over of the main cities and the fertile lands of the
south and the Levante, i n some cases with the assistance or at least
agreement of members of the Visigothic nobility like the sons of
Witiza and Theodemir. The second phase involved the conquest of
the north-east, where there is some evidence that Visigothic rule
continued under Akhila until the governorate of al-Samh (718-21)
and the making of peace agreements with the Visigothic lords of
the Ebro valley area and other remote districts like the mountains
north of Malaga. In this way the conquest of al-Andalus resembled,
o n a smaller scale, the Muslim conquest of Iran where the main

cities and lines of communication were first secured and only later
were agreements reached with the inhabitants of oudying areas.
The reasons for the success of the Muslim conquest have been
much debated. From the Muslim side the explanation was simple:
G o d had willed it. For Christian commentators from the author
of the ninth-century Chronicle of Alfonso III onwards, things were
much more problematic and they were faced with the question: how
could G o d have allowed this disaster to afflict a Christian people?
For mediaeval authors the most plausible explanation was that the
Visigoths were immoral and had disobeyed God's commandments,
while more secular modern historians have tended to look for signs
of political decay and weakness i n the Visigothic kingdom.
For the author of the Chronicle of 754 the blame lay with the
ambitions of Roderick, the treachery of Oppa, son of K i n g Egica,
who conspired with the Arabs, and the cowardice of the Archbishop
of Toledo, Sindered, who fled to Rome rather than remaining with
his flock. For the Chronicle
23
of Alfonso III it was the immorality of
K i n g Witiza, who besides having many wives and concubines him-
self, ordered his bishops and deacons to marry, and the treachery
of his sons that led to the Christian dbcle. 24
In fact, the evidence for the decadence of the Visigothic king-

dom is non-existent. In many ways it seems to have been stronger
25
in its final years than ever before: the damaging religious division
between A r i a n Visigoths and their Orthodox subjects had been
laid to rest for more than a century and there is by the end of the
seventh century no real sign of a split between Roman and Visigothic
elements. The Jews certainly suffered severe legal disabilities and i n -
termittent persecution and it is clear that they preferred to remain
23. Caps. 52-4. 24. Caps. 5-7.

25. F o r a g o o d discussion o f the arguments, see Collins, A r a b C o n q u e s t , pp. 6-22.
in their cities and accept Muslim rule than to j o i n their Christian

fellow countrymen in flight, but there is no reliable evidence that
they actively supported or encouraged the invaders. It is also true
that the Basques remained outside effective Visigothic control, and
indeed Roderick was campaigning i n that area at the time of the
first Muslim invasion, but separatism i n the northern mountains
had been a feature of political life for centuries and it was no more
dangerous or more threatening than before. In fact, i n some ways
the very strength of the kingdom made it easier to conquer: if it
had been divided into numerous local lordships and principalities,
they would no doubt have put up stiff resistance. As it was, Visigothic

Spain, like Anglo-Saxon England i n 1066, was centralised to the
extent that the defeat of the royal army left the entire land open
to the invaders.
The Visigothic monarchy suffered a short-term political crisis at
the accession of Roderick, resulting i n the defection of important
elements of the ruling class which may i n turn have contributed
to a major military defeat which left the country defenceless. The
army he led against the invaders was certainly large enough for
the purpose; its military worth is less easy to assess. Clearly the
batde was lost, but this might have been the result of bad luck
or bad decision-making on the day rather than long-term military
weakness.
It is striking that there seems to have been litde attempt to
defend the cities effectively or to raise a second army. Apart from
resistance i n the northern mountains, which was at the beginning
not conducted by the Visigoths, only Cordoba and Merida put up
any effective defence. The smallness of the numbers is noticeable:
Tariq is said to have sent Mughth against Cordoba with only 300
men, which might not have been effective if the governor had been
able to raise more than 400 to defend it. The city seems to have
been decrepit: the Roman bridge was broken and there was a major
hole i n the ramparts. When the Muslims arrived at Orihuela the
governor Theodemir had so few men that he had to dress up women
as soldiers and put them on the ramparts. No-one seems to have
defended Toledo, a superb natural fortress. Perhaps this failure
was pardy a result of lack of population, or at least of arms-bearing
population. When Saladin invaded the Kingdom of Jerusalem i n
1187, the Crusaders mustered a large army to oppose h i m but to do
so they had to empty their cities and fortifications of men. When
they lost the batde of Hattin, their splendid casdes were left virtu
ally undefended and most of them soon fell to the invaders. This
may well have been the same i n the case of Visigothic Spain. It was
not perhaps that numbers of the population i n general were lack
ing, but rather that numbers of military men were inadequate.
There seems to have been a sharp division i n Visigothic society
between military and non-military classes and there is no mention
of popular or civilian resistance. In Cordoba the Muslims were
informed by a shepherd they met that most of the people had fled
and that only the garrison of 400 and the du'af remained. Du'af
is an Arabic word which refers to the non-arms-bearing part of the
population; it was clearly considered out of the question that such
people would participate i n the defence of the city.

Another factor i n the lack of resistance may have been that
the Visigoths d i d not take the invasion sufficiendy seriously. It has
already been noted that Roderick's opponents among the aristoc
racy were hoping that the invaders would defeat h i m and depart,
leaving them i n charge, and it is interesting to see that this attitude
was shared by at least some of the invaders. It seems that Triq had
to persuade Ms to allow the Muslims to settle i n Spain and as late
as 717 the Caliph 'Umar b. 'Abd al-'Azz appointed a governor
whose mission it was to evacuate the Muslims from the Iberian
peninsula. It must have seemed to many i n Spain that the logical
response to the Muslim raiders was to retreat to their mountain
fastnesses with what they could carry and wait for them to take their
booty and go.
O n the whole the Muslims offered generous terms which cer
tainly made surrender a more attractive option, whereas unsuccess
ful resistance could, as the unfortunate defenders of Cordoba found,
lead to death. In Andalucia the sons of Witiza seem to have been
allowed to retain possession of the royal lands; i n the Murcia area
26
the terms amounted to local autonomy; i n Merida the inhabitants

were allowed to keep their possessions but the property of those
who had been killed i n the battle for the city, those who had fled
north and of the churches, was confiscated. In the later phases of
the conquest many Visigothic lords i n the Ebro valley area were
allowed to retain their lands and status and soon converted to
Islam: amongst the best known of these were the Ban Qasi (Casius)
of Tudela and the Ban 'Amrs of Huesca, who formed dynasties
which dominated the area for two centuries after the Muslim con
quest. Except for the action against church lands, this fits i n well
27
with what we know of Muslim terms elsewhere; the local people
26. Manzano 'Arabes, berberes e indigenas', p. 13. 27. Ibid., p p . 18-19.

were allowed to remain i n possession of their lands as long as they

paid a land tax and a poll-tax to the conquerors.
The settlement of the M u s l i m s
In the event Spain proved too attractive to its conquerors and they
did not leave but chose to setde i n the new lands. In the eastern
Islamic world it had been the intention that the Muslims should
setde only i n certain garrison towns, like Kfa and Basra i n Iraq,
Fustt (Old Cairo) i n Egypt and Qayrawn i n Ifrqya, and that they
should live off the taxes of the land. N o effort seems to have been
made to develop such a system i n Spain, and Muslim setdement
seems to have been haphazard and determined by the interests of
the setders rather than any overall scheme. The conquerors setded
down as property owners and do not seem to have received the 'at ',
or pensions, which were such an important feature of early Muslim
society i n the Middle East. This i n turn meant that there was no
need for the elaborate and precocious bureaucracy which had grown
up to service the system, nor any need to compile dwns or lists of
those entided to pensions. The slow development of bureaucracy
in al-Andalus may i n turn have been a reason for the slow develop
ment of a literary culture: not until the time of A b d al-Rahmn II
(822-52) did administration and a native literary culture begin to
appear.
Cordoba became the capital shordy after the initial conquest and
remained crucial to the politics of the period. Control of the capital
was vital for anyone who sought to govern the country and no
serious attempt was made to shift the capital elsewhere. The reasons
for this are not clear at first. It was not the Roman or Byzantine cap
ital and had litde obvious strategic importance. It was, however, at
the hinge of a number of important routes. To the north the roads
ran through the passes of the Sierra Morena to Calatrava, Toledo
and eventually to the Ebro valley. T o the east the upper Guadal
quivir valley gave access to the Levante, while south and west Elvira
(Granada) and Seville were easily accessible. The geography of al-
Andalus always made communications a problem, but Cordoba was
probably the least inconvenient site for the centre of government.
Perhaps more important was the rich agricultural hinterland of
the city and the fact that this was densely setded by the Muslims.
These resources of supplies and men were often able to sustain
the rulers when, i n the reign of the A m i r ' A b d Allh (888-912) for
example, other areas slipped out of their control. Toledo, another

possible capital and the centre of Visigothic power, was clearly lack
ing i n such local resources.
The non-Muslim population probably paid some taxes to Cor
doba, probably a jizya or poll-tax and perhaps a land tax as well;
a late source based on much earlier narratives recounts how a
Christian count, Ardabast, son of the old king Witiza, rather than
a Muslim bureaucrat was i n charge of collecting the tax from the
Christians. 28
Naturally, people from the same tribal and ethnic backgrounds

tended to settle i n the same areas and i n areas they considered

appropriate for their lifestyle. O n the whole the Arabs setded i n the
main cities and the fertile irrigated areas of the Guadalquivir valley,
the Levante around Murcia and Zaragoza and the middle Ebro
valley. In some places there were concentrations of men from the
same tribes i n the same area, L a k h m around Malaga for example,
and the Judhms and Tujbs who setded i n Zaragoza were to
dominate the political life of the city and the surrounding area for
centuries to come. Other tribes were more dispersed.
It would be wrong to think of these Arabs as Bedouin. Apart
from the Qurashis, almost all of them belonged to the Yemeni
group of Arab tribes. Some of these, like L a k h m and Judhm, had
lived i n the steppe lands of Jordan and southern Palestine, while
others, like Khawlan and Ma'fir, came from Yemen proper, a land
of cities, well-built villages and carefully tended farms. In either
case they would have been familiar with urban and agricultural life.
Many of them were second or third generation immigrants to North
Africa, reared i n such urban setdements as Fustt and Qayrawn
and well placed to take advantage of the opportunities offered by
their newly conquered lands. Tribal loyalties do not seem to have
been strong. The setders acquired land i n absolute ownership,
sometimes it would seem by inheritance, marrying the daughters of
the previous Visigothic owners, as A b d al-Azz b. Ms married
Roderick's daughter and Sara, granddaughter of Witiza, married
two Arab husbands i n succession and founded a dynasty which
produced, among others, the tenth-century historian Ibn al-Qutiya. 29
The integration of important elements of the Visigothic aristocracy

into the new Muslim ruling class certainly accounts for some of the
lack of opposition.
28. See the important passage i n Ibn al-Khatb, Al-Ihta fi akhbr Gharnta, ed. M .
I n a n (4 vols, Cairo, 1973-77), i, pp. 100-5.
29. M a n z a n o 'Arabes, berberes e indigenas', p. 15.
The Berbers were widely distributed throughout al-Andalus.

Many setded i n the central Meseta, Extremadura and the whole of
the north and west apart from Zaragoza and its environs. Certain
important cities like Toledo and Merida lay i n areas with a pre-
dominandy Berber population. There was also a significant Berber
presence i n the Valencia area, where they probably practised trans-
humance, wintering flocks on the coastal plains and moving west
into the mountains for the summer, and the Guadalquivir valley.
These lands were i n some ways less rich and inviting than the areas
setded by the Arabs and it has been suggested that the Berbers
were obliged to accept inferior lands despite the major contribution

they had made to the conquests. However, there are reasons for
thinking that this may not have been the case. As we have seen,
there was no overall direction or system i n the allocation of lands
and, given the small numbers of the conquerors and the vast extent
of the conquered lands, they would certainly have been able to take
over richer areas had they wished to. Some Berbers were certainly
farmers, used to irrigated agriculture, but the majority seem to
have been pastoral people and it was natural that they should gravi
tate to the familiar pastoral environments of the Iberian peninsula.
It is quite possible that many Berbers brought flocks with them
with the result that they had no alternative but to look for pasture.
This did not mean that there were not conflicts between Arabs
and Berbers which eventually erupted i n civil war, but there is no
evidence that grievances over land were among the causes of this.
The early governors, 714-41

The years between the departure of Ms b. Nusayr i n 714 and
the installation of Balj b. Bishr al-Qushayras governor i n 741 were
the period when these early setders enjoyed unopposed power. A t
first glance it is a period of great confusion: governors succeeded
each other with bewildering speed and only one (Anbasa b. Suhaym
al-Kalb, 721-26) lasted for more than a year or two. Beneath this
confusion, however, there are several common themes. The settlers
wanted to control the riches of al-Andalus for themselves, and to
increase that wealth by raiding areas i n France, which were still i n
the Dr al-Harb (the House of War, that is, non-Muslim territory)
and from which booty could be obtained. It did not especially matter
to them whether the governors who ruled the country were chosen
by them, as sometimes happened, particularly when the previous
governor had died a violent death, or appointed from outside,
as long as they were responsive to their needs. In contrast to the

period which followed 741, this first quarter of a century was com
paratively free of internal strife and rebellion.
W h e n Ms b. Nusayr departed, he left his son 'Abd al-Azz i n
charge. H e established himself as governor i n Seville and married
Roderick's widow. The story i n the Arabic sources explains that she
persuaded h i m to wear a crown and adopt other royal pretensions,
as the Visigothic kings had, and that the Arab notables saw this as
unlslamic and assassinated him. This certainly reflects their fears
that he was attempting to make the governorship his family's prop
erty. They may also have been concerned that he was encouraging
new setders from North Africa and the Middle East who would
demand a share of the wealth of the country. The struggle of the
early arrivals to maintain their privileged status i n the face of chal
lenges from later immigrants was to be a major source of unrest i n
the period of the governors. After 'Abd al-Azz no governor died
30
a violent death before 741 except at the hands of the infidel, and
there were no rebellions of importance.
Muslim Spain was only loosely attached to the main body of
the caliphate, but, given the vast distances involved, it is impress
ive to see how much influence the Umayyad caliphs i n Damascus
could exert on this, the newest and most distant of their provinces.
Governors were i n general appointed by the governor of Ifrqya
in Qayrawn or sometimes, as i n the case of al-Samh b. Mlik al-
Khawln, by the Caliph i n person, and they were almost always
outsiders to the province with no local power base or following.
This did not always mean that the wishes of the local people were
ignored and the Akhbr al-Majm'a states, perhaps with exaggerated
local patriotism, that if they d i d not like a governor or if he at
tempted to undermine their autonomy and privileged fiscal status,
they would write to the Caliph who would send one who pleased
them. Sometimes, if the governor was killed, the settlers would
31
'elect' a successor, but such elections were not usually accepted

by Qayrawn; when al-Samh b. Mlik was killed i n the attack o n
Toulouse, they chose 'Abd al-Rahmn b. 'Abd Allh al-Ghfiq, but
the governor of Qayrawn sent 'Anbasa b. Suhaym. When 'Anbasa
was himself killed while raiding i n France i n 726 the local people
again 'elected' and again their nomination was rejected, but i n 730
30. T h e assassination of A b d al-Azz is discussed in Chalmeta, I n v a s i o n e I s l a m i z a c i o n ,

pp. 245-54. Ibn ' A b d al-Hakam's account of the event is translated i n Melville a n d
Ubaydli, C h r i s t i a n s a n d M o o r s , pp. 1 4 - 1 7 .
31. Akhbr al-Majm'a, p. 25.
an unpopular outsider was replaced by 'Abd al-Rahmn al-Ghfiq

as a sop to local opinion.
The Caliph i n Damascus was remembered i n the Friday prayers
and his name appeared on the coins. Whether he received any
more tangible benefits is not clear. In Muslim law a fifth of the
wealth seized at the time of conquest belonged to the Caliph as a
fifth had originally been reserved for the Prophet Muhammad, but
it is unlikely that the Caliph ever received any revenue from this
fifth. Any attempts to secure a share of the revenues of al-Andalus
for the Umayyad government were vigorously resisted by the early
settlers.
In 718 the reforming Caliph 'Umar II (717-20) appointed al-
Samh b. Mlik al-Khawln as governor to implement the fiscal
reforms he had been introducing throughout the caliphate. The 32
governor arrived with a body of followers said to have been equal

in number to the original conquerors. His first task, apparendy,
was to send to the Caliph 'a detailed description of al-Andalus, its
rivers and its seas' - a sort of Doomsday Book. Then, i n order to
provide for his newly arrived followers, he intended to separate the
land that had been taken by force ( ' a n w a t a n ) from the land that
had been taken as a result of a treaty (sulhan) and to take a fifth of
the 'anwatan as the government's share and distribute the remain
ing four-fifths among the new arrivals. Predictably, this aroused the
indignation of the 'People of the Conquest' and they sent a depu
tation to the Caliph who responded by confirming the villages i n
the hands of those who had taken them as booty. They would be 33
liable to pay the ' u s h r or tithe on this land. The newcomers were
accommodated by dividing the caliphal fifth among them i n terri
torial concessions ( i q t d ' a t ) . For the first time the land was assessed
for taxation, with the intention that the surplus should be sent
to Damascus, as happened i n other provinces of the caliphate. In
the case of al-Andalus, al-Samh was allowed to use such revenue
as remained after the payment of salaries and the expenses of the
Holy War for the rebuilding of the bridge and city wall i n Cordoba.
In short, there is no unequivocal evidence that the caliphs received
anything beyond gifts at the time of conquest from al-Andalus or
that any financial strings bound the province to the central govern
ment. After the death of the Caliph i n 720 and al-Samh i n 721,
these centralising measures seem to have been allowed to lapse.
32. I have followed the discussion of al-Samh's governorate i n Chalmeta, I n v a s i o i
e I s l a m i z a c i o n , pp. 259-68; see also Manzano, 'Arabes, berberes e indigenas', p. 21
33. Akhbr a l - M a j m ' a , pp. 2 3 - 4 .
The Conquest and the Age of the Amirs. 711-56 21
The governors also issued coins from the earliest times i n order
to facilitate the distribution of booty and then the payment of taxes.
The first of these date from immediately after the first conquests,
711-12. Though they bear Latin inscriptions, such as the Islamic
34
monotheist formula ' i n nomine domini n o n nisi deus solus n o n

deus similis', they are based not on Visigothic coins but o n those
issued by the Muslims i n N o r t h Africa and ultimately on Byzantine
originals. U n d e r the governorate of al-Hurr (716-18) coins with
bilingual Latin/Arabic inscriptions were minted, and i n 720 al-Samh,
as part of his fiscal reforms, minted the first purely Arab coins, with
the sort of clear and elegant inscriptions which characterised

Umayyad coinage throughout the caliphate. This suggests a devel
opment of financial administration, although it is not until the
reign of 'Abd al-Rahmn II (822-52) that we hear of an organised
mint i n Cordoba. G o l d coins were struck i n small numbers i n the
early years, but the use of gold was abandoned after 744-45 and
thereafter silver dirhams were the standard issues.
This political establishment was typical of many areas of the early
Islamic state, the conquerors and their descendants firmly i n con
trol, using the revenues of the province as they saw fit, obliging
governors to work with them or be forced out and bitterly hostile
to any attempts to undermine their position, either by central gov
ernment or by any other group attempting to grab a share for
themselves.
Attacks on France and the few remaining areas of resistance
in the north of the Iberian peninsula were the major events. Many
members of the Visigothic aristocracy had made their peace with
the Muslims, but a few had fled northwards, among them Pelayo. 35
In the north these exiles j o i n e d forces with the local people, always
resistant to any central government. Together they opposed the Mus
lims i n the difficult and inaccessible mountain areas of the north
ern flanks of the Picos de Europa. Here they seem to have elected
Pelayo as king and been victorious i n a small encounter known to
history as the batde of Covadonga and traditionally dated to 717.
The reports of the Arab historians do not mention this trifling set
back, but for later Christian sources the batde of Covadonga marks
34. T h e standard work o n the early coinage o f al-Andalus is A . Balaguer, L a s

E m i s i o n e s t r a n s i c i o n a l e s A r a b e - m u s u l m a n e s de H i s p a n i a (Barcelona, 1976). Early coins
are also discussed a n d illustrated i n Chalmeta, I n v a s i o n e I s l a m i z a c i o n , pp. 242-5.
35. F o r a full discussion of the history a n d historiography o f Pelayo, see Collins,
A r a b C o n q u e s t , pp. 141-51, i n which he suggests that Pelayo may have b e e n a local
l o r d like T h e o d e m i r with his power base i n the north-west.
the first stage i n the long struggle to expel the Muslims from Spain
and has acquired a legendary importance.
At the time the Muslims were much more concerned to invade
France, which offered prospects of booty wholly lacking i n the inhos
pitable valleys of the Cantabrian mountains. Details are sketchy, but
there were at least four expeditions, the first launched at the very
beginning of the conquests, the other three led by the governors
in person. In 721 al-Samh b. Mlik al-Khawln led an expedition
against Toulouse on which he himself was killed, and i n the sum
mer of 725 'Anbasa b. Suhaym al-Kalb led a lightning raid right

up the Rhone valley to Burgundy where the army pillaged Autun.
Finally, the popular and well-respected governor of al-Andalus, 'Abd
al-Rahmn b. 'Abd Allh al-Ghfiq, led an expedition through
western France which was finally and disastrously defeated by Charles
Martel at the batde of Poitiers (actually fought at Moussais la Bataille,
to the north of Poitiers) i n October 732. 36
Historians from Gibbon onwards saw the batde of Poitiers as a

major turning point and an event which marked the end of the
great Muslim conquests which had begun a century before. More
recently, scholars have tended to play down its effect: it was clearly
significant i n establishing the power of Charles Martel and the
Carolingians i n France, but it also had profound consequences
in Muslim Spain. It signalled the end of the ghantma (booty) econ
omy. U p to this point pressure on resources i n al-Andalus could
be relieved by raiding and dissatisfied people had the opportunity
to acquire more wealth. It was a popular activity: Ibn Idhri, writ
ing around 1300, looked back with nostalgia to the simplicities of
these early, vigorous days: In those days the people of al-Andalus
were admirable and excellent, determined i n Holy War and eager
for God's rewards so they threw themselves on the Christians (Rm)
in warfare and siege.' These were not wars of expansion and no
attempt seems to have been made by the Muslims to settle any
area north of Narbonne. The purposes of the raids were to take
booty and to establish the prestige of the governor, whose most
important public functions were the leadership of the Muslims
in prayer i n the mosque and i n battle against the infidel. After 732
these opportunities were no longer available to anything like the
same extent. The Muslims of al-Andalus were obliged to live off
the finite resources of their own adopted country and competition
for revenues and status soon gave rise to savage feuds.
36. F o r the Poitiers campaign see Lvi-Provencal, H E M , i, pp. 61-5; Chalmeta,

I n v a s i o n e I s l a m i z a c i o n , pp. 2 8 4 - 8 .
The c o m i n g of the Syrians, 741-56
This first phase of Muslim rule was brought to an end by a major

upheaval i n North Africa and its side-effects i n al-Andalus. In 740
there was a Berber uprising against Arab rule. The main reasons
for this seem to have been fiscal. During the period of conquest
and expansion, the Arabs had been happy to grant their Berber
followers the tax privileges which went with being full members of
the Muslim community. With the end of expansion, the governor
and financial administrator of Egypt, 'Ubayd Allh b. al-Habhb,

tried to impose the kharj (land tax) on these Berbers and reduce
them to a subordinate status i n order to increase the revenue yield
now required to pay the Syrian army, the backbone of the caliphate.
Further fuel was added to their justifiable resentment by the prac
tice of taking Berber children for the harems of the Umayyad elite.
A m o n g some Berbers, these resentments led to the adoption of
Kharijite beliefs. The Kharijites, a puritanical sect which tried to
preserve the virtue of the earliest days of Islam, as they saw it,
rejected the authority of the Umayyad caliphs and refused to pay
taxes to them.
In a very short period of time, the whole Maghreb had slipped
from the control of the governors. In response to the complete
defeat of the local forces, the Caliph Hishm (724-43) set about
recruiting a new army i n Syria. Syria at this time was divided into
a number of fairly small administrative units called junds which
were used as the basis on which the army was raised. Soldiers were
recruited from the j u n d s of Qinnasrn, Hims, Damascus and Jordan
and on their way to North Africa they were j o i n e d by soldiers from
Egypt who were held to form a separate j u n d . These junds were
divided by more than bureaucratic lines: by the middle Umayyad
period all of them were dominated by one or other of the major
tribal groups, Qays/Mudar or Yemen.
The origins of the division between northern or Qays/Mudar
Arab tribes and southern or Yemen has been the subject of consid
erable controversy. In theory these groups were super-tribes, united
by a common ancestry against their foes, but this genealogical the
ory disguised parties based on regional and political interests. T o
add to the confusion, it would seem that these labels meant dif
ferent things i n different areas. In Syria the division was basically
between those tribes (Yemen) which were established i n the area
before the coming of Islam and the others (Qays/Mudar) which
migrated northwards from Arabia i n the aftermath of the Muslim
conquests. These divisions had been sharpened and made much

more acute at the bloody batde of Marj Rhit i n 684 when the
Yemen supporters of the Umayyads had defeated their Qaysop
ponents, inflicting injuries which were never forgotten. Subsequendy
there had been some reconciliation between the two parties and
Qayss too supported the Umayyads, but the rift remained and in
creasingly undermined the power of the Umayyad caliphate. O f the
junds at this time, Damascus and especially Qinnasrn were Qaysl
dominated, Hims, Jordan and Palestine were Yemeni, as were most
of the Egyptians.
The military expedition against the Berbers was led by a Qays
commander, Kulthm b. 'Iyd, with his nephew Balj b. Bishr al
Qushayras secondincommand. It was not a success: i n Septem
ber/October 741 the Muslim army was defeated on the River Sebou
in northern Morocco and the survivors, about 10,000 strong, now
led by Balj b. Bishr, fled north to Ceuta. Here, completely cut off
from their homeland, ragged and starving, they were i n desperate
straits. In vain they appealed to the governor of alAndalus, Ibn
Qatan, for help, but he refused.
Meanwhile, the Berber uprising i n North Africa had resulted
in a similar upheaval i n alAndalus. There had been rumblings
of discontent i n the peninsula before. According to the Chronicle
of 754, a Berber (the Chronicle distinguishes the Berbers (Mauri)
from Arabs (Saraceni)) called Munuz, hearing that his fellow Ber
bers i n North Africa were being oppressed by the Arabs, raised a
rebellion i n the northern frontiers of alAndalus, possibly trying to
establish an independent lordship i n Cerdaiia. This was probably
in 729. The next year Munuz made an alliance with Eudo, Duke
of Aquitaine, but i n 731 the governor, ' A b d alRahmn alGhfiq,
launched an expedition against h i m and he was finally surrounded
in Cerdaa and was obliged to commit suicide i n order to evade
capture. 37
The rebellion of 741 was much more serious. 'Ubayd Allh b.

alHabhb had appointed 'Uqba b. alHajjj alSall as governor of
alAndalus to implement his strict fiscal policies. The Andalusi
Arabs were fiercely opposed to h i m because the policies threatened
their fiscal status and, perhaps, because they feared it would provoke
the Berbers of alAndalus to a similar rebellion.
In 740 there was a coup d'tat i n which 'Uqba was forced to resign
37. T h e only source which mentions this is the C h r o n i c l e o f 754, cap. 79: for the
chronology see Chalmeta, I n v a s i o n e I s l a m i z a c i o n , pp. 2823.
and was replaced by ' A b d al-Malik b. Qatan al-Fihr, a venerable

figure chosen by the people of al-Andalus. It seems, however, that
this change was too late to avoid an uprising. In the autumn of
741 there was a major revolt i n the north-west and the Arabs were
driven out of all the lands north of the Cordillera Central, so putting
an end to the ephemeral Muslim occupation of this area.
The Berbers marched south towards Cordoba and Ibn Qatan
found himself unable to resist them effectively. In his panic he
looked across the Straits for allies. Balj and *Abd al-Malik began to
do business. Balj's men were so desperate that he was prepared to

agree to almost any conditions which would see them fed, and he
accepted that his men should fight the Berbers and then return to
N o r t h Africa when their work was done. His only real stipulation
was that they should return as a group, rather than be dispersed to
be at the mercy of their enemies. In the spring of 742 the Syrians
were helped to cross into al-Andalus.
The batde-hardened Syrians under Balj joined the Andalusi Arabs
led by Ibn Qatan to defeat the Berbers i n a fierce batde near Toledo.
In the aftermath of victory, the Syrians were reluctant to leave this
fertile and promising land, where, we are told, the Arabs lived 'like
kings', and return to the hardships of North Africa. Relations be
tween Balj and ' A b d al-Malik soon broke down and Balj launched
a coup which left the old governor dead and himself and his Syrian
followers i n control. This could not go unavenged and the old-
established Arabs (now called Baladiyn (people of the country) i n
distinction to the Shmiyn or Syrians) launched a counter-attack,
led by two of 'Abd al-Malik's sons. In August 742 this attempt was
decisively defeated near Cordoba, but Balj was mortally wounded
in the batde. Nonetheless, the Syrians remained i n possession of
Cordoba and chose a new governor from their own ranks who de
feated the opposition, composed of both Arabs and Berbers, which
had regrouped i n the Merida area. Many Arabs of distinguished
lineage were sold cheaply as slaves. 38
The next year (743), a new governor was sent from Qayrawn
by Hanzala b. Safwn, apparentiy i n response to a petition from
Andalusis of all parties who wanted peace. Ab'l-Khattar al-Husm
b. Dirr al-Kalb was a member of the Yemeni aristocracy of Damas
cus (Kalb were one of the leading Yemenite tribes of Syria) and he
attempted to solve the outstanding problems of the province. H e
first secured the release of all the Arab and Berber captives and
38. F o r these events see Chalmeta, I n v a s i o n e I s l a m i z a c i o n , pp. 307-27.

then set about providing for the Syrians who now clearly had to
be accommodated i n al-Andalus.
Ab'1-Khattr, perhaps on the advice of Count Ardabast, settled
the Syrians i n a methodical way. They were already organised
in junds and each j u n d was setded i n a different area, chosen,
allegedly, because it resembled their Syrian homeland. Thus the
j u n d of Damascus was settled at Elvira (Granada), the j u n d of
Jordan i n Rayyu (Malaga and Archidona), the j u n d of Palestine
in Sidonia, the j u n d of Hims i n Seville and Niebla, the j u n d of
Qinnasrn i n Jaen. The j u n d of Egypt, possibly the largest, was

divided between the Algarve i n the west and Tudmr, the lands
originally governed by the treaty with the Visigoth Theodemir which
now seem to have been opened up for Muslim setdement, i n the
east. It is said that they were given a third of the property of the
local people to live off, though again it is not clear whether this
should be considered as revenues or actual lands to cultivate: cer
tainly the image of these violent Syrian warriors suddenly setding
down to plough the fields, prune the vines and dig the irrigation
ditches is faintly improbable. It is most likely that the Syrians
were given a third of the revenues paid by the people of the areas
their j u n d was setded i n . They were not concentrated i n Cordoba
or other garrison cities but dispersed throughout the area of the
j u n d . The limited evidence suggests that they were responsible for
collecting the revenues themselves and that they were obliged to
pay the government muqta't, basically a fixed sum, from the rev
enues they collected. In exchange for this livelihood, the Syrians
were obliged to do military service. 39
The events of 741-43 profoundly changed the political charac

ter of Muslim Spain. It substantially increased the Arab element i n
the population, especially i n those rural areas i n the south which
were to be the heartland of al-Andalus for centuries to come. It also
increased the Syrian element. Most of the early setders were ulti
mately of south Arabian origin, but the new arrivals came from
Syria and the area the Arabs called al-Jazra (the island), that is, the
steppe lands of northern Iraq and Syria between the Tigris and
Euphrates rivers. Many of these had long-standing loyalty to the
Umayyad family. A l o n g with the Arab tribesmen, there also arrived
a significant number of m a w l of the Umayyad family, probably
non-Arab native Syrians or prisoners of war from other regions of
39. See Chalmeta, I n v a s i o n e I s l a m i z a c i o n , pp. 331-5; for a full discussion o f the
settlement and fiscal obligation of t h e j u n d s , see E . Manzano M o r e n o , ' E l asentamiento
y la organizacion de los und-s sirios en al-Andalus', A l - Q a n t a r a xiv (1993), 3 3 0 - 8 .
the Muslim empire. They formed an important group without tribal

affiliations, owing their loyalty only to the ruling dynasty, and their
presence i n al-Andalus was vital to the later success of the Umayyads
in establishing themselves as rulers.
The most immediate consequence was also the most destructive,
the introduction into al-Andalus of the fierce Qays/Mudar versus
Yemen disputes. Previously the overwhelming dominance of Yemen
had meant that there was littie conflict, but the arrival of Balj,
aggressively Qays, and his followers changed that. The defeat and
temporary enslavement of many of the Yemenis added bitterness
from their side to an already inflammatory mixture. It was diffi

cult for any governor to break free from the constraints of the feud,
since not to be generous to members of their own group would
leave them without any reliable support. It was probably this i n
security which led Abl-Khattr, an outsider to al-Andalus without
any local power base of his own, to rely o n and favour the Yemenis.
The Qayss could not afford to tolerate a governor who would only
dispense favours to their hated rivals, and they found a new leader
in al-Sumayl b. Htim al-Kilb, hard-bitten, brutal and fanatically
devoted to the Qayscause.
War could not long be averted. The Qayss were fewer i n num
ber but stronger i n the Cordoba area and possibly more effective
militarily. Al-Sumayl set out to divide his enemies and succeeded i n
winning over the leader of the Yemeni tribes of L a k h m andjudhm.
Both these tribes had long been setded i n Syria and may have felt
more i n common with other Syrians than with the Yemeni tribes
of South Arabia. This coalition rose i n revolt and i n A p r i l 745 de
feated Abl-Khattr, who was taken prisoner but was soon rescued.
There followed a period of confusion when different members of
the coalition tried to take control as governor, until i n January 747
al-Sumayl, who had remained behind the scenes, not claiming power
for himself, produced an outside candidate, Ysuf b. ' A b d al-Rahmn
al-Fihr.
It was a shrewd choice. Ysuf was already an o l d man and al-
Sumayl could expect h i m to be a pliable instrument. A t the same
time, he had positive advantages. H e was a direct descendant of
'Uqba b. Nfi', hero of the early Muslim conquests i n N o r t h Africa.
Fihr, being a branch of Quraysh, were usually considered to belong
to the Qaysgroup, but Quraysh, being the Prophet's tribe, had
always occupied an intermediate position somewhat outside tribal
divisions and could hope to attract loyalty from all parties. In addi
tion, the family had built up many contacts and alliances among
the Berbers and could hope for support from some Berber groups.
These assets recommended this otherwise rather undistinguished
old gendeman to al-Sumayl and were also to mean that the Fihrs
were to be the only serious rivals to the Umayyads for control of the
whole of al-Andalus.
Ysuf began to exercise his power as his patron would have
expected, excluding Yemenis from the fruits of office. In 747 Ysuf 40
and al-Sumayl, aided, we are told, by the tradesmen of Cordoba i n

cluding the butchers with their knives, defeated the Yemeni counter
attack at Secunda, on the south bank of the river opposite the city
of Cordoba. Ysuf began to grow i n confidence. The collapse of the
Umayyad caliphate of Damascus i n 747-50 i n the face of 'Abbasid
attacks from the east meant that he became an independent ruler.
H e felt strong enough to remove al-Sumayl to Zaragoza where he
occupied himself helping victims of the terrible famine of the early
750s. But the old enmities were merely dormant. In 755 al-Sumayl
was besieged i n Zaragoza by Yemeni elements, Yusuf was powerless
to help h i m and he was only saved by an expedition of Qaysvolun
teers from the south. It was into this environment of deadly fac
tional conflict that the emissaries of ' A b d al-Rahmn b. Mu'wiya,
the Umayyad, arrived.
During this period of instability, the northern frontier of al-
Andalus stabilised. In the immediate aftermath of the conquest
41
the Muslims penetrated as far as the north coast, leaving only a few
upland valleys i n the Pyrenees and isolated pockets of resistance on
the northern slopes of the Picos de Europa; the sources even speak
of a Muslim 'governor' i n Gijon. Most of the plains of the Duero
and the mountains of Galicia and Cantabria were occupied, i n so
far as there was a Muslim presence at all, by groups of Berbers, but
we know very litde about this and there is no record of permanent
or significant occupation i n these areas; nor were there any major
campaigns, the Arabs finding France much more rewarding terri
tory. Almost by default, the areas north of the Duero slipped out
of Muslim hands after the Berber rebellion of 741 caused many of
them to come south where they were heavily defeated i n batde. A
long famine which began i n 750 seems to have caused many of the
survivors to leave for North Africa. The Christian King Alfonso I
(739-57) was able to establish some fortified outposts i n the Duero
plains and to raid even further south. The Cordillera Central marked
40. Ysuf's governorate is described i n Chalmeta, I n v a s i o n e Islamizacion, pp. 335-48.
41. F o r the historical geography of the frontier regions see Manzano Moreno,
Frontera.
the most northerly limits of Muslim occupation i n the western half

of the peninsula, Coimbra, Coria, Talavera, Madrid, Guadalajara
and Medinaceli all being frontier setdements. This position was
hardly to change for three centuries. T o the north of these setde-
ments, the land seems to have been almost completely uninhabited
except by wandering shepherds until the southernmost outposts
of Christian setdement were encountered at places like L e o n and
Astorga. 42
The eastern sector of the frontier presented a very different pic-

ture. Here Muslim setdement pressed up to and into the Pyrenees;

Pamplona, Tudela, Huesca, Girona and Narbonne were i n Muslim
hands. Only i n Narbonne, captured by Charles Mattel's son Pepin,
probably i n 759, was their rule challenged. The northern outposts
43
of Muslim setdement lay not on an east-west axis but on a north-

east to south-west line, and this was to remain the position until
well into the twelfth century.
By 757 the Muslim presence i n Andalus was clearly there to stay.
Local resistance had effectively disappeared, a new generation of
Arabs and Berbers born i n al-Andalus was growing up and converts
were beginning to be made among the indigenous people. But the
new conquerors had conspicuously failed to develop a viable polit-
ical system. After the 'Abbasid revolution of 750, al-Andalus was no
longer part of a wider Muslim empire: it remained to be seen what
alternative polity would emerge to fill the gap.
42. See C . Sanchez-Albornoz, D e s p o b l a c i o n y R e p o b l a c i o n del v a l l e d e l D u e r o (Buenos

Aires, 1966).
43. M a n z a n o M o r e n o , F r o n t e r a , pp. 75-7.
CHAPTER TWO
The Umayyad Amirate, 756-852 1

' A b d a l - R a h m a n I and the establishment

of the Umayyads, 755-88
The Umayyad family were members of the Prophet's tribe, Quraysh,

and distant cousins of Muhammad himself. In the early days of 2
Islam, the Umayyad chief Ab Sufyn had been one of the lead
ing opponents of the new religion, but, following the triumph of
Muhammad after 628, Ab Sufyan's son Mu'wiya rapidly became
one of the most important figures i n the early Muslim state. After
the conquest of Syria he was appointed governor and, i n 661, with
the assassination of Muhammad's cousin and son-in-law, 'Al, he be
came Caliph of the entire Muslim world (661-80). Mu'wiya's direct
line died out with his son Yazd i n 684. After a vigorous civil war be
tween the supporters of the Umayyads and their enemies, the caliph
ate was seized by Mu'wiya's second cousin Marwn b. al-Hakam
(684-85) and his son 'Abd al-Malik (685-705), so inaugurating a
second Umayyad dynasty, sometimes known as the Marwanids.
For half a century, the Umayyads ruled the whole of the Islamic
world from Sind and Samarqand i n the east to newly conquered
al-Andalus i n the west, but there was always opposition from those
1. Secondary sources for the early Umayyad amirate are very limited. F o r the
reign o f ' A b d al-Rahmn I we have the final chapter o f P. Chalmeta, I n v a s i o n
e I s l a m i z a c i o n : l a s u m i s i o n de H i s p a n i a y l a f o r m a c i o n de a l - A n d a l u s (Madrid, 1994),
pp. 349-87. Thereafter, the fundamental account remains E . Lvi-Provencal, H E M ,
i, pp. 139-278. E . M a n z a n o M o r e n o , L a F r o n t e r a de A l - A n d a l u s en epoca de los Omeyas
(Madrid, 1991) is invaluable, not just for the study o f the frontier zones but for
wider questions o f the political structure o f the amirate.
2. F o r the role o f the Umayyads i n the eastern Islamic world, H . Kennedy, T h e
P r o p h e t a n d the A g e of the C a l i p h a t e s ( L o n d o n , 1986), pp. 82-123, a n d G.R. Hawting,
T h e F i r s t D y n a s t y of I s l a m ( L o n d o n , 1986).
30
The Umayyad Amirate, 756852 31
who felt that the hereditary monarchy of a family who had been
among Muhammad's leading enemies was unlslamic and those who
resented the domination of the Muslim world by the Syrian mili
tary supporters of the dynasty. From 747 to 750 there was a vast
upheaval, the 'Abbasid revolution, which swept the Umayyads and
their Syrian supporters from power. Most of the members of the
ruling house were rounded up and executed, but a few, mosdy less
prominent, individuals were able to lie low and make their escape.
One such was ' A b d alRahmn b. Mu'wiya, a young grandson of
the great Caliph Hishm (72443). After some hairbreadth es

capes, he fled to North Africa, accompanied only by a few mawl,
among them Badr, later to be his righthand man i n alAndalus. H i s
first intention seems to have been to secure Ifriqlya (Tunisia), but
the governor, 'Abd alRahmn alFihr, was hostile and he was obliged
to seek refuge among his mother's relations, the Nafza Berbers.
Thwarted i n Africa, he sent Badr to make contact with the Umayyad
m a w l among the Syrian junds i n alAndalus. There were said to
have been 500 of them i n the dwn, led by 'Ubayd Allh b. 'Uthmn
and ' A b d Allh b. Khlid of Damascus and Yusuf b. Bukht of
Qinnasrn. A t first they tried to attract the support of the Qays
leader, alSumayl, then under siege i n Zaragoza, but he refused,
fearing that ' A b d alRahmn would like to make himself effective
ruler, so they turned to the opposition Yemenis for support. In the
early autumn of 755, after more than five years o n the r u n , 'Abd al
Rahmn crossed to Almuecar o n the south coast of alAndalus.
A t first he was given refuge i n the nearby homes of his mawl,
Ibn Khlid and Ab 'Uthmn, protected by 300 horsemen. After
the attempt to reach a compromise with Ysuf alFihrand alSumayl,
'Abd alRahmn began to make contact with Yemeni leaders through
out the south. By the next spring (756) he had recruited an army
of about 2,000 Umayyad mawl and Yemeni jundis and marched on
Cordoba. Here his supporters fought and defeated the Qaysarmy
of Ysuf and alSumayl and, i n May 756, he entered the capital.
The proclamation of 'Abd alRahmn b. Mu'wiya as A m i r i n the
mosque of Cordoba o n Friday 14 May 756 was not the e n d of the
Umayyad seizure of power i n alAndalus, but only the end of the be
ginning. The new A m i r was determined to build up a secure power
base i n alAndalus which would enable h i m to survive and pass o n
the tide to his descendants i n a way no previous governor had been
able to do. H e also intended to establish himself as an indepen
dent ruler. After the first year, he no longer had the names of the
'Abbasid caliphs of the east acknowledged i n the Friday prayers i n
Cordoba, and while he retained the fairly modest tide of Amir, he

acknowledged no temporal superior.
'Abd al-Rahmn had a number of advantages over his rivals for
power. H e was a member of the Prophet's tribe of Quraysh and of
the family of the Umayyad caliphs: no-one could deny his high
descent. O n the other hand there were many other Qurashs i n al-
Andalus, notably Ysuf al-Fihrhimself, and the Umayyad caliphs
had been violendy and completely rejected by the eastern Islamic
world, which might have encouraged sympathy among the Syrians
of al-Andalus but not much respect. The Umayyads were very much
outsiders as well: as far as we can tell, no member of the extensive

clan had ever visited al-Andalus before. ' A b d al-Rahmn's father
Mu'awiya had died when he was a boy and had not played an im
portant role i n the later Umayyad caliphate, and he himself was
still a young man of 26 with no real political experience.
In addition to the reputation, being an Umayyad d i d bring 'Abd
al-Rahmn a decisive advantage i n terms of a small but committed
band of supporters and the opportunity to recruit more. The band
of supporters consisted of the Umayyad m a w l in al-Andalus with
whom Badr had originally made contact, notably Ab 'Uthmn, his
son-in-law 'Abd Allh b. Khlid and Yusuf b. Bukht. T o these were
added ' A b d al-Karim and 'Abd al-Malik, grandsons of one of the
leaders of the original conquest of Spain, Mughth al-Rm, himself
said to have been a m a w l of the Umayyad Caliph al-Wald, and
Tammm b. 'Alqama, who was probably a member or m a w l of
Thaqf, a tribe with close ties to the Umayyads. These mawl seem
to have had extensive possessions i n Cordoba and the lands of
Elvira and Rayyu, but they had litde prestige i n the tribal politics
and rivalries which dominated al-Andalus and the arrival of the
Umayyad gave them their only hope of real political power. But this
group was not restricted to al-Andalus: there were Umayyad mawl
in other parts of the Muslim world who would be keen to come and
support the new regime.
In addition to the mawl, many members of the Umayyad family
continued to arrive from the east and they too were committed
supporters of the new Amir. The most important of these was ' A b d
al-Malik b. 'Umar b. Marwn: it was he who advised 'Abd al-Rahmn
to drop the name of the 'Abbasid Caliph al-Mansr from the Friday
prayers and he became governor of Seville and one of the Amir's
most reliable generals. In this way, ' A b d al-Rahmn built up a fol
lowing, based on the mawdU and his own clan and geographically
centred i n Cordoba and the south-east, that transcended the tribal
followings which his opponents could command and which was

entirely dedicated to his success. None of his rivals could achieve
that.
'Abd al-Rahmn b. Mu'awiya was to reign as A m i r for 33 years
after his proclamation at Cordoba, and the length of his rule was
a major factor i n the success of the Umayyads; if he had died or
been killed as so many other governors had been after a few years,
the Umayyad regime would no doubt have followed its ephemeral
predecessors into obscurity.
'Abd al-Rahmn embarked on a policy of ensuring that he was

acknowledged as A m i r throughout the whole peninsula, but he
faced many rivals i n al-Andalus. In the Arabic chronicles, with their
marked Umayyad/Cordovan sympathies, those who resisted are por
trayed as rebels against a legitimate Umayyad authority; i n reality
they were local chiefs or strong men attempting to preserve their
influence against expanding and encroaching Umayyad power. The
previous amirs of Cordoba had had litde coercive power beyond
the strength of their own following: real influence i n the provinces
remained i n the hands of the leaders of the junds or other local
notables. One of 'Abd al-Rahmn's main objectives was to expand
the authority of Cordoba throughout the Muslim-held areas of the
Iberian peninsula.
In these circumstances, it is hardly surprising that many people
resented the attempts of the early Umayyad amirs to rob them
of their autonomy and impose a measure of civil and fiscal con
trol over them. There were other reasons for resentment too. The
arriving members of the Umayyad family needed estates and, as the
Syrian junds were not property owners, lands had to be confiscated
from the Baladis and the Christians. It was at this time that the bulk
of the estates of Count Ardabast were confiscated and, probably, that
the areas covered by the pact with Theodemir were opened up for
Muslim setdement. 3
'Abd al-Rahmn's first problem was to destroy the power of

his predecessors, Yusuf al-Fihri and al-Sumayl, who still remained
in the field despite their recent defeat. At first a compromise was
arranged and Ysuf was able to keep his possessions and reside i n
Cordoba, but it appears that he soon found this intolerable and he
escaped to Merida where he raised a large army of Berbers. How
ever, he was defeated by troops loyal to 'Abd al-Rahmn and i n
759/60 was murdered near Toledo. The Fihris enjoyed extensive
3. See Chalmeta, I n v a s i o n e I s l a m i z a c i o n , pp. 262-3.

support among the Berbers and were the only family who could
rival the Umayyads; they were not to give up easily. It was seven
years before ' A b d al-Rahmn felt strong enough to challenge their
hold on Toledo. In 764 he sent two of his most trusted commanders,
Badr and Tammm b. 'Alqama, against the city where Hishm b.
'Urwa al-Fihri was holding out and he was captured and executed.
There were at least two other attempts by Fihris to establish them
selves i n the peninsula. The first of these came i n 778-79 when
'Abd al-Rahmn b. Habb al-Fihri, known as the Slav from his blond
complexion and blue eyes, landed i n the east of the country and
attracted some support before being killed by one of his Ber

ber supporters. Then, i n the autumn of 785, almost at the end of 4
'Abd al-Rahmn's reign, Yusuf al-Fihri's son Muhammad gathered

an army i n the Toledo area. Defeated by the Umayyad troops, he
fled west towards Coria where he was isolated and killed. 'Abd
al-Rahmn went on to chastise the Nafza Berbers i n the area, pre
sumably because they had supported the rebel. The Fihris felt they
had as good a claim to al-Andalus as the Umayyads and they could
mobilise widespread support among the Berbers, but they seem to
have been no match for the Umayyad mawdli and their allies among
the Syrian junds.
Another threat was posed by agents of the 'Abbasid caliph Ab
Ja'far al-Mansr. Al-Mansr made one serious attempt to regain
control of al-Andalus i n 763. H e sent no troops but a standard and
a diploma of investiture as A m i r of al-Andalus. The local leader was
al-'Al b. al-Mughth al-Yahsub from Beja i n southern Portugal,
who gathered a large number of supporters in the west of al-Andalus.
'Abd al-Rahmn left Cordoba and moved west to the powerful for
tress at Carmona, high on the hill where the casde still dominates
the surrounding plains. Here he fought what was probably the most
desperate encounter of his reign. The chronicler Ibn Idhr takes
up the story: 5
H e fortified himself there with his mawdli, his faithful supporters

[thiqat] and the rest of his men and al-'Al b. al-Mughth began a
close investment. W h e n he had besieged the town for many days
and the siege dragged o n , al-'Al's army began to get resdess. ' A b d
al-Rahmn knew that they were impatient and thinking of bridles
and saddles and he ordered a fire to be built and he ordered his
4. For the dating of al-Rz, quoted by Ibn Idhr, see A l - B a y a n a l - m u g h r i b f i akhbr

a l - A n d a l u s w a ' l - M a g h r i b , ed. E . Lvi-Provencal a n d G.S. C o l i n ( L e i d e n , 1948), i i ,
pp. 5 7 - 8 .
5. Ibid., pp. 51-2.
companions to burn the scabbards of their swords and said to them,

'Come out with me against this crowd, determined never to return!'
They were seven h u n d r e d of them, courageous m e n a n d famous
heroes. So they took their swords i n their hands searching for their
enemy. T h e battle lasted long until G o d d i d H i s marvellous work
and the army and companions of al-'Al quaked a n d fled a n d their
fate became a warning for the people of the world. Al-'Al was among
the first to be killed.
Only by such desperate actions d i d the Umayyad amirate sur

vive. The dead head of the defeated rebel was enbalmed and sent
east with a merchant who left it one night with a note attached i n
the markets of Qayrawan, the nearest 'Abbasid outpost (in Tunisia,
other accounts say it was taken to Medina and left outside the
caliphal tent). When al-Mansr was informed he is said to have
remarked, 'We all belong to G o d . We sent this miserable man to
his death. Praise be to G o d who has put the sea between me and
this devil!' It hardly needs to be said that this imaginative story
appears only i n Andalusi histories and finds no place i n 'Abbasid
6
chronicles. This debacle meant the end of direct 'Abbasid interfer

ence. After this there were exchanges of abusive correspondence,
but the Umayyads of Spain posed no real threat to the Baghdad
regime and the 'Abbasids were soon preoccupied with problems
much nearer home.
The rebellion of al-'Al b. al-Mughth was fuelled by local griev
ances as well as 'Abbasid encouragement. ' A b d al-Rahmn was
determined to assert his power over the Syrian j u n d s . Before the
coming of the Umayyads, the leaders (ra'is, pl. ru'as') of the junds
were either self-appointed or chosen by their followers: they were
effectively independent and could make war or alliances as they saw
fit. 'Abd al-Rahmn set out to replace the old-style leaders with his
own supporters. The clearest example of this comes from the junds
of Egypt and Hims setded i n Beja and Seville. It seems that ' A b d
al-Rahmn had appointed the Umayyad ' A b d al-Malik b. 'Umar b.
Marwn as governor of Seville and the west and that this was re
sented by local leaders who felt that this was an encroachment o n
their power. Al-'Al had attracted widespread support i n the area
and ' A b d al-Rahmn is said to have been worried that the soldiers
from Seville i n his own army would desert to the rebel cause. In 7
766 one Sa'd al-Matarrebelled i n Niebla and took over Seville

6. See for example A l - B a y n ; A k h b r al-Majm'a, ed. a n d Spanish translation E .
Lafuente y Alcantara ( M a d r i d , 1867), p. 103.
7. Al-Banvn i i , p. 52.
before being killed by the Amir. In the same year he had another
leader from the area, Ab'1-Sabbh Yahya al-Yahsub, executed i n
Cordoba. The dead man's followers sought revenge and the people
of Seville joined his cousins i n an attempt to take Cordoba by sur
prise. It was not until 774 that the rebellion was finally defeated by
'Abd al-Malik b. 'Umar. This seems to have completed the defeat of
the j u n d i leaders i n this area and their subjection to effective con
trol from Cordoba. 8
Another ideological challenge came from an 'Alid pretender. In

the eastern Islamic world, the members of the house of 'Al, first
cousin to the Prophet and husband of his daughter Ftima, were
the most powerful focus of opposition to both Umayyad and 'Abbasid
caliphates. Most of their following, however, came from Iraq and
points east and they had litde support among the Syrians who formed
the majority of the Arab setders i n al-Andalus. Heterodox religious
opinions were more widespread among the Berbers, however. In
North Africa at this time, these opinions usually took the form of
Kharijism, but i n 768 and 770 a Berber of the tribe of Miknsa
called Shaqyb. 'Abd al-Whid led a revolt, claiming to be related
to the 'Alids. His rebellion began i n Santaver, i n the hills around
9
Cuenca, but for the next nine years he dominated much of the
sparsely inhabited upland country between Santaver and Coria and
Medellin far to the west. It was a guerrilla war, the Berbers retreat
ing to the mountains on approach of the Amir's army and return
ing to the villages and plains when they had gone. The rebellion
was an irritant, but the soi-disant Fatimid seems to have attracted
no support amongst the Arabs or the town dwellers and 'Abd al-
Rahmn was also able to make an alliance with Hill al-Madyn,
described as head ( r a s ) of the Berbers i n the east of al-Andalus.
In the end, i n 776-77, Shaqya was taken by treachery and killed.
Although he clearly commanded support among the Berbers,
there is litde evidence that this, or indeed any other of the Berber
disturbances of the time, were motivated by Berber resentment
against Arab control. 'Abd al-Rahmn had Berber supporters and
Berber opponents just as he had Arab supporters and Arab oppon
ents; the real distinction was between those who threw i n their lot
with Cordoba and those who were determined to maintain their
own local independence.
8. F o r the chronology o f these rebellions see Lvi-Provencal, H E M , i, pp. 110-12;

for the political analysis see E . M a u z a n o M o r e n o , ' E l asentamiento y la organizacion
de los yund-s sirios en al-Andalus', A l - Q a n t a r a xiv (1993), 338-46.
9. T h e revolt is discussed i n M a n z a n o M o r e n o , F r o n t e r a , pp. 2 3 8 - 4 9 .
There were areas of localised resistance i n the west and north

of the peninsula. Not surprisingly, 'Abd al-Rahmn had difficulty
imposing his authority o n distant Zaragoza. The Arab notables 10
of Zaragoza and the Ebro valley resisted Umayyad attempts to take

the city as they had resisted the attempts of previous Amirs, and
the tradition of local autonomy was well established. In about
774 Sulaymn b. Yaqzn al-A'rb, lord of Barcelona and Girona,
and al-Husayn b. Yahyb. Sa'd b. 'Ubda, descendant of a famous
Madinan companion of the Prophet i n Zaragoza, were the main
powers i n the area. In order to maintain his independent status,
Sulayman sent a mission to Charlemagne's court at Paderborn to

ask his support, probably i n return for the overlordship of Zaragoza.
Charlemagne arrived i n the Ebro valley i n 778 but al-Husayn re
fused to cooperate with the Emperor who, having no siege engines,
had to withdraw, humiliated, from the walls of the city. It was dur
ing his return journey that the Basques attacked his rear guard i n
the Pyrenean pass at Roncesvalles, giving rise to the legends which
are recounted i n the Song of Roland, though the first text of this
great epic does not appear for three centuries after the event and
the historical detail has become wildly confused.
In the aftermath of this debacle, Sulaymn was executed by his
erstwhile ally al-Husayn, but his sons inherited his position i n Bar
celona and Girona. It seems that the A m i r 'Abd al-Rahmn was
alarmed by this turn of events. In 781 he led a military expedition
to demand the submission of al-Husayn b. Yahya and to re-establish
Muslim control i n the Upper Ebro valley. A t first al-Husayn ac
cepted the Amir's authority and was confirmed as governor of the
city, but the next year he threw off this allegiance. 'Abd al-Rahmn
returned and assaulted the city with siege engines (manjantq), and
al-Husayn was captured and executed and severe measures taken
against the townspeople. The whole complex episode shows how
the Umayyad A m i r tried to establish his authority over the local
magnates by a mixture of diplomacy and occasional force but that,
as long as they were content to accept his overlordship, he was
prepared to leave them i n peace. By the end of his reign, Merida
and Toledo were being governed by his sons, Hishm and Sulayman,
but it is not clear that the Umayyads exercised any direct control
10. T h e c o m p l e x events i n Zaragoza are discussed i n C h a l m e t a , I n v a s i o n e

I s l a m i z a c i o n , pp. 367-81, whose account I have largely followed. M a n z a n o M o r e n o ,
F r o n t e r a , discusses the history a n d status o f these local magnate families (pp. 208-23)
a n d ' A b d al-Rahmn's attempts to control the city (pp. 315-18). M.J. V i g u e r a M o l i n s
gives a simplified account o f events ( A r a g o n M u s u l m a n , Zaragoza, 1981, pp. 5 7 - 6 5 ) .
in the Ebro valley. In one area they certainly d i d not: it was i n

785, according to the Chronicle of Moissac, that the inhabitants of
11
Girona threw off the authority of the sons of Sulaymn b. Yaqzn

and handed their city over to the Carolingians. The event produced
no reaction i n Cordoba and is not noted by the Arab chroniclers.
Not all of 'Abd al-Rahmn's energies were devoted to assuring
his military control over al-Andalus. His reign saw the confirmation
of the position of Cordoba as the capital of al-Andalus. The status
of the city was confirmed when 'Abd al-Rahmn built a country
palace at al-Rusfa to the north-east of the town, i n addition to the

Qasr i n the centre of town, and i n 785 he began the construction
of the mosque which still survives, with many later additions, to the
present day.
The reigns of H i s h m I and a l - H a k m 7,

788-822
'Abd al-Rahmn b. Mu'wiya died on 7 October 788. His achieve
ments had been enormous, but he had put them all i n jeopardy by
his hesitation about the succession. While it was clear he intended
that the amirate should remain hereditary within the Umayyad
house, it was not certain which of his sons should succeed h i m .
The principle of primogeniture, which became so important to the
ruling houses of mediaeval Christendom, was never established i n
the Umayyad family: designation by the previous ruler and accept
ance by other members of the family and their supporters were the
determining factors i n deciding who amongst its members should
succeed. 'Abd al-Rahmn had three sons who emerged as possible
candidates: Sulaymn, who had been born i n the east and was now
in his early forties, Hishm, who had been born i n Cordoba i n 757,
and 'Abd Allh, who does not seem to have been a serious candi
date at this stage but was to become an important figure i n years
to come. Neither Sulayman nor Hishm were at their father's death
bed, Sulayman being i n Toledo, Hishm i n Merida. Some say that
Hishm was his father's final choice, others that the dying A m i r
told ' A b d Allah to acknowledge whichever of the two reached
Cordoba first: Hishm, he said, had the advantages of piety and a
good education, Sulayman of age, bravery and the devotion of the
Syrians. 12
11. Q u o t e d i n Levi-Provencal, H E M , ii, p. 128.

12. Ibn Idhr, A l - B a y a n , ii, p. 61.
It only took six days for Hishm to come from Merida, and 'Abd
Allh greeted h i m as ruler and handed over the seal of office,
but his other brother was not prepared to accept this verdict
and gathered his supporters to march south. There was a short,
sharp conflict near Jaen and Sulayman's men were defeated. It
took almost two months for Hishm to reduce Toledo and oblige his
brother to surrender, but i n 789 Sulayman was paid 60,000 dinrs
in cash, possibly half the annual income of the amirate at this time,
and was forced to leave for N o r t h Africa and promise not to return:
Umayyads might be defeated and disgraced but, at this time, they
would not be executed like any common rebel, for that would under
mine the status of the whole ruling house. 13
As far as we can tell, the new amir faced litde internal opposition
after the defeat of his brothers. H e had indirect dealings with the
magnates of the Ebro valley. In 788/9 Sa'id b. al-Husayn al-Ansr,
whose father had held Zaragoza against the Umayyads i n the pre
vious reign, took the city again, proclaiming himself A m i r . H e was
soon defeated, not by Umayyad troops but by Ms b. Fortun b.
Qasi, whose grandfather had been Count of the area i n Visigothic
times and one of the earliest and most distinguished converts
to Islam. In 791/2 Zaragoza was taken over by Matrh, the son
14
of Sulaymn b. Yaqzan, and an expedition was sent from Cordoba

to drive h i m out. The problem was solved, however, when Matrh
was murdered by one 'Amrs b. Yusuf while he was out hunting and
the city was handed over to the Umayyad forces.
'Amrs came from a muwallad (native Muslim) background
and is first recorded as a ghuldm (military page) i n the service of
Sulayman b. Yaqzan and his family. H e now seems to have decided
to throw i n his lot with the Umayyads against his old masters: he
and his family were rewarded by the favour of the ruling dynasty
and during the next century they became one of the most powerful
families i n the Upper M a r c h . The rise of the Ban Qasi and Ban
15
'Amrs at this time, allied with the Umayyads against the leading
Arab families of the area, marks the beginning of the entry of the
muwallads into the political life of al-Andalus.
The Arab chroniclers give Hishm a pious and ascetic personality
13. Levi-Provencal, H E M , ii, pp. 139-41.

14. Ibn Idhr, A l - B a y d n , ii, p. 62: see art. 'Ban Kas' by P. C h a l m e t a i n E n c y c l o
p a e d i a of I s l a m new e d n (Leiden, I 9 6 0 - ) , a n d A . C a n a d a Juste, ' L o s Ban Qasi ( 7 1 4 -
924)', P r i n c i p e de V i a n a clviii-clix (1980), 5-90.
15. T h e origins o f the family are discussed i n M a n z a n o M o r e n o , F r o n t e r a ,
pp. 219-22.
and say that he sought to exert his authority rather by personal

example and by leading the Muslims against the Christians of the
n o r t h . His commitment to good works expressed itself i n the com
16
pletion of the first phase of the mosque i n Cordoba, including the

ablution facilities and the minaret and the repair of the bridge. A n
other manifestation of his piety was the sending of military expedi
tions against the Christians, among them a number of campaigns
in the Asturias and Upper Ebro and a major attack o n Narbonne
and Carcassonne i n 793, when the Count of Toulouse, William
Short-Nose, was killed. Hishm himself d i d not lead these cam

paigns but entrusted command to the two brothers, ' A b d al-Malik
and ' A b d al-Karim b. ' A b d al-Wahid b. Mughth. These campaigns
were important i n asserting the role of the Umayyads as leaders of
all the Muslims of al-Andalus and of bringing Umayyad armies to
parts of the country where they would not normally penetrate. In
this Hishm was establishing a precedent which would be followed
by many of his successors.
'Abd al-Karim was Hishm's chief minister until his death fairly
early i n the reign. The inner circle also included the secretary 's
b. Shuhayd, whose father had served ' A b d al-Rahmn and whose
family was to provide Cordoba with bureaucrats for the next three
centuries as well as one of al-Andalus's greatest poets. Hishm's mod
est piety, attendance at funerals and visitations of the sick earned
him such a reputation that the great Medinan scholar, Malik b.
Anas, whose work so profoundly affected al-Andalus, is said to have
wished that Hishm could make the pilgrimage i n person (which
of course he could not, Mecca and Medina being i n the power of
the 'Abbasid caliphs of Baghdad).
The connection with Medina was to prove very important. As
far as we can tell, Andalusis never visited Damascus or Baghdad at
this time, although there was a good deal of immigration from the
eastern Islamic world, but they could and d i d go o n pilgrimage
to M e d i n a and Mecca i n comparatively large numbers, especially
those who aspired to be faqihs or qddls. It was not surprising, there
fore, that the knowledge of Muslim law that they picked up was the
teaching of the pre-eminent Medinan master, Malik b. Anas. Malik
(d. 795) was the founder of the oldest of the four 'orthodox' schools
of Islamic law (the other three being the Hanbal, the Shfi' and
the Hanafi, all developed i n the ninth century). Mliki law, as ex
pounded i n the Kitdb al-Muwattd of Malik himself and later works,
3
16. See for example the character sketch i n Ibn Idhr, A l - B a y a n , i i , p p . 6 5 - 6 .

notably al-Mudaxvwana al-Kubrd of Sahnun of Qayrawan (d. 854),

was essentially the codified practice of Medina. As such it was prac
tical and much concerned with ordering everyday life and especially
with the facilitating of trade and commerce. A t the same time, the
system left litde room for abstract speculation and virtually none
at all for innovation. The fact that it could be adopted as a fully
worked out system no doubt increased its attractions for an iso
lated Muslim colony like that i n al-Andalus, whose members wanted
practical solutions rather than opportunities for debate. It also gave
great power and status to a small, self-perpetuating group of scholars
who were familiar with the doctrines, notably, i n the first genera
tion, Yahya b. Yahya al-Layth (d. 848), a pupil of Malik himself,
who by his advice effectively controlled the appointment and dis
missal of the qddts of Cordoba. The adoption of the Malikite school
so completely meant that al-Andalus was spared the sort of sectar
ian dissensions which tore Baghdad society apart i n the third/tenth
and early fourth/eleventh centuries. Malikism was hostile alike to
Kharijism, Shi'ism and even to the more innocent forms of Sufism,
and none of them really secured widespread support i n al-Andalus.
O n the debit side, however, Malikism offered a rather formal and
rigorist version of Islam, more concerned with correct performance
than inspiration, and it discouraged speculation and discussion: if
Cordoba was spared the conflicts that convulsed Baghdad, it also
saw none of the intellectual excitement that accompanied them.
Hishm died on 17 A p r i l 796. H e was careful to leave no uncer
tainty about the identity of his chosen successor and his son al-
Hakam, now 26 years old, was duly accepted as A m i r i n Cordoba.
The sources portray h i m as a very different character from his
father. T a l l , thin, haughty and strikingly dark i n complexion,
the new prince was to prove a formidable ruler. H e was fond of
women and wine, too much so some said, and was no mean poet,
but he was remembered more for his cunning, his implacable ruth-
lessness and the awesome speed with which he reacted to news of
disaffection.
And disaffection d i d not take long to become apparent. His
father had been opposed by his two brothers, Sulayman and 'Abd
Allh. Both these were still alive, i n exile i n North Africa, and they
were now determined to regain their lost inheritance. The turbu
lent Sulayman spent the next four years wandering the country,
attempting to build up enough support, largely among the Berbers
of the south, to dislodge his nephew. H e was defeated i n a number
of encounters and was finally surrendered to al-Hakam by the Berber
governor of Merida, Asbagh b. Wansus. H e was executed i n 800,

the first member of the ruling family to suffer this indignity.
'Abd Allah was much more successful i n establishing a posi
tion for himself and his family i n the Umayyad polity. H e tried to
establish himself i n the Upper March, even visiting Charlemagne
in Aachen to solicit his support, without much success. In 800 he
returned and attempted to establish himself at Huesca, but he was
forced on to Valencia. Valencia was not, at this stage, a very import
ant Muslim centre, but 'Abd Allah was able to use the area as a
base for negotiation with his nephew. In 802 an agreement was

finally reached by which he was established i n the city and given
a salary (rizq) of 1,000 dinrs a month, presumably to pay his sup
porters, and an annual bonus (ma'drif) of 1,000 dinrs. This ar 17
rangement led to the Valencia area being effectively an appanage

of this branch of the Umayyad family, and 'Abd Allah was known
thereafter as al-Balans (the Valencian). In fact the arrangement
worked well: this area had never been under the effective control
of Cordoba and 'Abd Allah's rule brought it within the Umayyad
orbit, while his military ambitions, and those of his son 'Ubayd
Allah, were directed against the Christians of the north.
The most famous and dramatic events of al-Hakam's reign
were revolts i n Cordoba itself. The causes of the widespread dis
18
content are by no means clear but are probably connected with

al-Hakam's autocratic determination to stamp his authority on this
and other cities and to oblige the people to pay taxes. H e had also
distanced himself from the local elite by recruiting a private body
guard commanded by a local Christian, al-Rab' b. Theodulfo. A l -
Hakam was basically trying to establish a reliable military and fiscal
foundation for the amirate, but it was not surprising that some pious
Muslims were numbered among the dissidents. Others said it was
the natural resdessness of the Cordovans. The unrest occurred i n
19
two stages. In 805 there was a conspiracy among certain notables

of Cordoba to mount a coup d'etat and put al-Hakam's cousin
Muhammad b. al-Qasim on the throne. Unfortunately for the con
spirators, Muhammad did not share their enthusiasm for the pro
ject, which he revealed to the Amir. H e sent his trusted secretary
to overhear a clandestine meeting and record the names that were
mentioned. This the secretary did, though being careful, according
17. T h e terms are given i n Ibn Idhr, A l - B a y a n , ii, pp. 70-1.

18. These disturbances are discussed in full in Levi-Provengal, H E M , i, pp. 160-73.
19. Ibn Idhr, A l - B a y a n , ii, pp. 85-6.
to Ibn al-Qutiya, to reveal his presence before his own name

20
came up. The conspirators were treated without mercy and 72 of

them were executed. It seems to have been a movement within the
elite, for the names that we have are either Arab or, like Masrur
al-Khdim, members of the palace staff, and it was i n no sense
anti-Umayyad or an attempt at social revolution, but was rather an
attempt by leaders of Muslim society to preserve status and privi
lege which they felt the A m i r was trying to undermine.
Numerous executions followed and Cordoba was surprised and
shocked by the Amir's severity. His action alienated many who had
not been actively involved i n the coup attempt. For thirteen years
discontent rumbled on and the A m i r fortified himself i n the city
and became more and more dependent on his guards. T h e n i n 818
opposition erupted again i n a spectacular manner. There was a
widespread uprising i n the populous suburb, usually referred to
simply as al-Rabad (the suburb), which lay to the south of the city
itself, across the Guadalquivir river. The uprising attracted support
from such respectable figures as the jurist Yahya b. Yahya al-Layth
as well as the populace i n general (called the 'dmma or the sawdd'
in the Arabic sources). The causes of the discontent are disputed.
According to some sources there were complaints about taxation
and especially the raising of 'unQuranic' taxes (maghdrim or wazd'if)
to supplement the 'legitimate' tax base of the Umayyad amirs which
was very limited. Al-Hakam is said to have been the first ruler of
al-Andalus to acquire mamlk (slave) soldiers, and recruiting and
maintaining the sort of full-time guard al-Hakam built up would
certainly have required additional sources of income. The com
bination of resentments among the Cordovans at being excluded
from power and paying more taxes brought together a broad cross-
section of the population to oppose the Amir.
If the cause of the rebellion is unclear, the results are not: the
insurgents attempted to cross the bridge and storm the city proper
but they were beaten off by loyal troops, led by the hdjib (chief
minister), 'Abd al-Karm b. Mughth, while two members of the
ruling family with military experience, 'Ubayd Allh b. 'Abd Allh
al-Balans and Ishaq b. al-Mundhir, led troops out of a side gate to
attack the rebels from the rear. Their defeat was total and the
vengeance of the A m i r was terrible. After an initial slaughter, he
ordered that the suburb south of the river be destroyed and its
20. Ibn al-Qtya, T a ' n k h iftitah a l - A n d a l u s , ed. a n d trans. J . M . N i c h o l s ( u n p u b

lished P h D thesis, C h a p e l H i l l , 1975), pp. 111-15.
inhabitants driven into exile. Only a few, like the jurist Yahya b.
Yahya al-Layth, went into hiding and were later pardoned. The
most important consequence of this diaspora was that many of
the refugees went to Morocco where they setded i n Fes on the
opposite bank of the river from the existing colony of immigrants
from Qayrawan. In this way the twin setdements which still form
the core of the ancient city of Fes, the quarter of the Qarawiym
and the quarter of the Andalusiym, came into existence.
The events i n Cordoba overshadowed the last years of al-Hakam's
reign. H e seems to have become something of a recluse before his

death i n 822, but he did establish the succession firmly. During his
last illness, leading members of the court took the oath of al
legiance (bay'a) to his son 'Abd al-Rahmn i n his house and then to
his other son al-Mughra, who became second i n line to the throne.
The taking of the oath was continued i n the mosque i n Cordoba
where al-Mughra remained for some days, receiving pledges of al
legiance on the minbar (pulpit) of the mosque from the rest of the
population. This public ceremonial was based direcdy on 'Abbasid
models and reveals the increasing self-confidence and aspirations
of the Umayyad amirs.
The reign of ' A b d a l - R a h m a n I I , 8 2 2 - 5 2

About 30 years old, the new ruler was already experienced i n
political and military affairs, having led expeditions to Toledo and
the northern frontiers. H e was tall and slighdy stooping; people
noticed his wide, dark eyes and full, henna-died beard. H e was to
rule al-Andalus for 30 years from 822 to 852, and his reign saw
the coming of age of al-Andalus i n terms of developing the mecha
nisms of a mature Muslim state and a genuine indigenous Muslim
culture. 'Abd al-Rahmn seems from the beginning to have tried to
move away from his father's policy of repression and to be more
responsive to Islamic susceptibilities. Even as his father was dying,
he secured the execution of the commander of the Christian body
guard, al-Rabf, and the demolition of the wine market i n Cordoba,
and during his reign this concern was to be expressed i n a renewed
enthusiasm for the jihad and a programme of mosque building.
The administration became more formal and bureaucratic and
took on the structures it retained until the end of Umayyad rule i n
the early eleventh century. A t its head was the hdjib, a word which
21
21. See Lvi-Provencal, H E M , i, pp. 2 5 6 - 9 .

originally meant door-keeper or chamberlain, a meaning it retained

in the Islamic east. In Cordoba, however, the hjib was effectively
the prime minister, holding his own court or majlis at the palace
gate where messengers or petitioners would report. Below h i m were
the wazrs: i n the east, the waztr of the 'Abbasid caliphate was the
chief administrator and head of the civil servants (kuttb, sing, ktib)
who ran the bureaucracy. In al-Andalus, the waztrs were much more
general purpose officials who might well lead an army or govern a
city, and the term was sometimes used as an honorary title. There
was also a degree of overlap and the hjib could also be a wazr.
U n d e r ' A b d al-Rahmn the wazrs were given salaries of 300 dnrs.

The A m i r also had a personal secretary (ktib) who was often one
of his closest advisers. A d t wn (administrative office) was organised
to arrange the collection of taxes, and the standard Muslim institu
tions of the sikka, to mint coins, and the tirz, to provide the official
textiles, were set up.
For the first part of his reign he continued to make use of his
father's advisers. The veteran *Abd al-Karim b. ' A b d al-Whid b.
Mughth seems to have maintained his position as chief army
commander and hjib as he had been for the previous two reigns,
but after his death o n campaign against the Christians i n 824 his
family seems to have disappeared from the scene. H i s place i n the
administration was taken by s b. Shuhayd, famous for his incor
ruptibility, who became waztr and hjib from 833 onwards and also
commanded the horse at the time of the Viking attack o n Seville
in 844. ' A b d al-Rahmn had gathered a number of trusted intim
ates i n the years when he was heir apparent, among them Sufyn
b. ' A b d Rabbihi, a Berber of obscure origins who became his ktib,
and ' A b d Allh b. Sinn, who rebuilt the walls of Seville after the
Viking attack. F r o m a military point of view, the most important
of these newcomers were the two brothers, ' A b d al-Rahmn and
Muhammad b. Rustam. These were scions of the Rustamid dynasty
of Tahert (in Algeria) and were the first of many North African
princelings brought over with their followers to serve the rulers of
al-Andalus, Muhammad particularly distinguishing himself against
the Vikings. In the second half of his reign, ' A b d al-Rahmn came
to rely increasingly o n the eunuch Nasr, the first, but no means the
last, eunuch to achieve major political influence i n al-Andalus.
The simple household of ' A b d al-Rahmn I was gradually trans
formed into the formal court of 'Abd al-Rahmn II, with its courtiers
sheltering a secluded and remote monarch who rarely appeared to
his subjects. The A m i r surrounded himself with a growing number
of eunuchs and slave girls. This development of the court on east

ern models had a cultural aspect to it as well. In this the Umayyads
were helped by the catastrophe which overwhelmed the 'Abbasid
court between the outbreak of civil war i n 811 and the final entry
of al-Ma'mn to Baghdad i n 819: palaces were looted and burned
and talented poets turned out on the streets. A m o n g these was one
'Al b. Nfi', called Ziryab. H e was an Iraqi who had studied under
the greatest of the early 'Abbasid singers, Ishaq al-Mawsil. Accord
ing to the Andalucian story, he was forced to leave the 'Abbasid
court because of his master's jealousy and, after a spell i n North

Africa, he arrived i n al-Andalus i n 822. Here he set himself up not
only as a musician but as an arbiter of taste i n dress and food,
remaining the uncontested Beau Brummell (the analogy is Lvi-
Provenal's) of Cordovan society until his death i n 857. Whatever
the political vicissitudes of the reign, this period marks the first age
of Andalucian culture, silver if not golden. Apart from al-Hakam II,
' A b d al-Rahmn was the most intellectual of the Umayyad sover
eigns and encouraged scholars and poets, including the eccentric
scientist 'Abbs b. Firns who, among other things, made himself
wings and attempted to fly. N o r must ' A b d al-Rahmn's building
work be forgotten. Apart from military architecture, there was the
surviving extension of the mosque at Cordoba and the mosque at
Seville, fragments of which can still be found i n the court of the
church of San Salvador, and a mosque at Jaen. O f course these
buildings had a political value: just like the jihd, his commitment
to mosque building showed the A m i r as a truly Muslim ruler.
Perhaps the most picturesque indication of Cordoba as the heir
of Baghdad is given i n the story of the necklace of al-Shif'. This
necklace had been made for Zubayda, wife of Hrn al-Rashd, and
during the dispersal of 'Abbasid treasures during the civil wars which
followed that caliph's death i n 809, it was bought for 10,000 dnrs
by ' A b d al-Rahmn's agents and given by h i m to his favourite, al-
Shif'. With the collapse of the caliphate of Cordoba it passed to
the Dh'l-Nnids of Toledo. After E l C i d took Valencia i n 1095, the
necklace was handed over to h i m and given to his wife Jimena. It
later appears among the possessions of D o n Alvaro de Luna, Con
stable of Castile, and may even have been worn by Isabella the
Catholic before it finally disappeared from record. Whether history
or romance, the story of the necklace is symbolic of the changing
fortunes of Baghdad and Cordoba, 'Abbasids and Umayyads, i n this
period.22
22. Ibid., pp. 2 6 4 - 5 .

The A m i r had to cope with troubles i n the area of Tudmr i n the

east of the country. The old treaty arrangements of the time of the
conquest had broken down and that meant that Arab settlers were
competing for lands. They divided as ever into Mudar and Yemen
and fought not the government but each other. There was anarchy
for seven years before the Yemen chief Abl-Shammkh Muhammad
b. Ibrhm al-Judhm finally surrendered to ' A b d al-Rahmn. H e
pacified the area i n two ways: he offered Abl-Shammkh a pos
ition i n his army, where we find h i m guarding Calatrava during one
of the campaigns against Toledo, and he caused his governor to
build a new capital, the city of Murcia, founded i n 831, i n which an

Umayyad garrison could be stationed.
Two other notable events distinguish the reign of 'Abd al-Rahmn.
The first of these was the V i k i n g attack on Seville. In the summer
of 844 (229), 80 Viking ships, having been driven away from Lisbon
by the local governor, sailed up the Guadalquivir river, as they had
sailed up so many other western European estuaries and, basing
themselves on the island now known as the Isla Menor, attacked
the unwalled city of Seville. This they looted and pillaged and, find
ing that the Guadalquivir was not navigable any higher, took to
the land. 'Abd al-Rahmn, forewarned by the governor of Lisbon,
acted swiftly and effectively: troops were summoned from all areas
and even the recalcitrant Ms b. Ms b. Q a s led his followers
from the Upper March. The Umayyad armies under Muhammad b.
Rustam and the eunuch Nasr decisively defeated the invaders i n a
land battle; many were killed and most of the rest returned to their
ships and fled. Some, however, remained and settled i n the lower
Guadalquivir area, where they converted by and by to Islam and
lived reformed and blameless lives, selling cheeses to the Sevillanos.
The episode shows the effectiveness of the Cordovan state when
faced with an unexpected attack by an unknown enemy, and it is
only fair to contrast this swift mobilisation with the feeble and chaotic
response of Anglo-Saxon and Carolingian contemporaries. The long-
term response was equally effective. Seville was walled and a naval
arsenal (dr al-sin'a) established there which certainly frustrated
further raids i n 859 and 966. There were, of course, other reasons
why the Vikings were not the menace i n al-Andalus that they were
in Britain or France. It lay further from their lands of origin and
the lack of navigable rivers meant that they were unable to pen
etrate deeply into the country. The Vikings were defeated because
they were fighting inland; if they had been able to bring their long-
ships right up to Cordoba when they first arrived, the result might
have been very different. As it was, the invasions which d i d so much
Bibliography
Arabic sources
'Abd Allh b. Buluggn alZiri alSanhji (d. after 483/1090), Al-

Tibyn, ed. E. LviProvenal (Cairo, 1963); annotated trans., A.T.
Tibi (L eiden, 1986).
'Abd al-Whid alMarrkush (c. 669/1270), Al-Mu'jib, ed. M.S. al
'Uryan (Cairo, 1949); trans. A . H u i c i Miranda (Tetuan, 1955).
Akhbr al-Majm'a, ed. and Spanish trans., E. Lafuente y Alcantara
(Madrid, 1867).
Cronica Anonima de 'Abd al-Rahmn III al-Nsir, ed. with Spanish
trans., E. Lvi-Provenal and E. Garcia Gomez (Madrid, 1950).
Al-Dabb, Ab Ja'far A h m a d b. Yahy (d. 599/1024), Bughyat al-
multamisfl ta'rkh rijl al-Andalus, ed. F. Codera (Madrid, 1884-
85).
Al-Baydhaq, Documents indits d'Histoire Almohade, ed. E. Lvi-Provenal
(Paris, 1928).
Fath al-Andalus, ed. L. M o l i n a (Madrid, 1994).
Al-Hull al-Mawshya, ed. I.S. Allouche (Rabat, 1936); also ed. A .
Zamama and S. Zakkar (Casablanca, 1979).
Al-Khushn, Ab ' A b d Allh Muhammad b. Hrith (d. 361/972),
Qudt Qurtuba, ed. with Spanish trans., J. Ribera (Madrid, 1914).
Ibn al-Abbr, Ab ' A b d Allh Muhammad b. ' A b d Allb al-Quda'
(d. 658/1260), Al-Takmila, ed. F. Codera (Madrid, 1887-90).
, Al-Hulla al-siyar', ed. H . Mu'nis (2 vols, Cairo, 1963).
Ibn ' A b d al-Hakam, 'Abd al.Rahmn b. 'Abd Allh (d. 257/871),
Futh Misr wa'l-Maghrib wa'l-Andalus, ed. C. Torrey (New Haven,
1922); Spanish trans., E. Vidal Beltran (Valencia, 1966).
Ibn 'Abd al-Mun'im al-Himyar, Ab 'Abd Allh Muhammad b. ' A b d
Allh (9th/15th century), Rawd al-Mi'tr fi akhbr al-aqtr, i n partial
edn with French trans., E. Lvi-Provengal, La Peninsule Ibrique au
Moyen-Age (Leiden, 1938).
Ibn 'Abdn, Muhammad b. ' A b d Allh al-Ishbl (6th/12th cen
tury), Seville Musulmane, trans. E. Levi-Provengal (Paris, 1947).
316
Bibliography 317
Ibn Ab Zar', Ab'lHasan 'Al b. ' A b d Allh alFs (8th/14th cen

tury), Rawd al-Qtrts, ed. C. Tornberg (Uppsala, 184346); Span
ish trans., A . H u i c i Miranda (2 vols, Valencia, 1964).
Ibn alAthr, Ab'lHasan Muhammad b. 'Al (d. 630/1233), Al-
Kmil f-Ta'rkh, ed. C.J. Tornberg (12 vols, L eiden, 185367);
partial French trans., E. Fagnan, Annales du Maghreb et de l'Espagne
(Algiers, 1901).
Ibn Bashkuwal, Abu'lQsim Khalaf b. 'Abd alMalik (d. 578/1183),
Al-Sila f ta'rkh a'immat al-Andalus, ed. F. Codera (Madrid, 1882);
also ed. I. alHusayni (Cairo, 1955).

Ibn Bassm, Ab'lHasan Al' alShantarn (d. 542/1147), Al-Dhakhra
f mahsin ahl al-Jazra, ed. I. 'Abbas (Beirut, 1978).
Ibn alFaradi, Abu'lWad 'Abd Allh b. Muhammad (d. 403/1013),
Ta'rkh 'ulam' al-Andalus, ed. F. Codera (Madrid, 1890); also ed.
I. alHusayni (Cairo, 1954).
Ibn Habb, 'Abd alMalik (d. 238/852), Kitb al-Ta'rkh, ed. J . Aguad
(Madrid, 1991).
Ibn Hawqal, Ab'lQsim Muhammad b. Hawqal (c. 380/990), Srat
al-ard, ed. J . H . Kramers (L eiden, 1939).
Ibn Hayyan, Ab Marwn Hayyn b. Khalaf alQurtub (d. 469/
1076), Al-Muqtabis min anb ahl al-Andalus (Al-Muqtabis II), ed.
M.A. Makki (Cairo, 1971).
, Al-Muqtabis fi ta'rkh rijl al-Andalus (Al-Muqtabis III), ed. M .
Antufia (Paris, 1937).
, Al-Muqtabis V, ed. P. Chalmeta, F. Corriente and M . Sobh
(Madrid, 1979); Spanish trans., M . Viguera and F. Corriente
(Zaragoza, 1981).
, Al-Muqtabis fi akhbr balad al-Andalus (al-Hakam II) (Al-Muqtabis
VII), ed. A.A. alHajji (Beirut, 1965); Spanish trans., E. Garcia
Gomez, Anales Palatinos del Califa de Cordoba al-Hakam II (Madrid,
1967).
Ibn Hazm, Ab Muhammad 'Al b. Sa'd alQurtub (d. 456/1064),
Jamharat ansb al-Arab, ed. A . H a r u n (Cairo, 1962).
, Ras'il Ibn Hazm, ed. I. 'Abbas (Cairo, 1955).
, The Ring of the Dove, trans. A.J. A r b e n y (L ondon, 1953).
Ibn Idhr alMarrkush, Ab'l'Abbs A h m a d b. M u h a m m a d
(c. 712/1312), Al-Baydn al-mughribfi akhbr al-Andalus wa'l-Maghrib,
vols. III ed. E. LviProvenal and G.S. C o l i n (L eiden, 1948);
vol. I l l ed. E. LeviProvenal (Paris, 1930); vol. IV ed. I. Abbas
(Beirut, 1967); Part III (Almohads) ed. A . H u i c i Miranda (Tetuan,
1963), trans, idem (2 vols, Tetuan, 1953).
Ibn alKardabs, Ab Marwn ' A b d alMalik alTawzari (6th/ 12th
318 Muslim Spain and P ortugal
century), Al-Iktif' f akhbr al-khulaf', ed. A . M . al'Abbadi (Mad

rid, 196566); Spanish trans., F. Maillo (Madrid, 1986).
Ibn Khaldn, Ab Zayd ' A b d alRahmn alHadram (d. 808/1406),
Al-Ibr wa diwn al-mubtadd' wa'l-khabar, ed. S. Zakkar (8 vols,
Beirut, 1981); annotated trans, of the Muqaddima (Introduc
tion), F. Rosenthal (3 vols, L o n d o n , 1958).
Ibn alKhatb, Ab ' A b d Allh L isn alDn Muhammad b. ' A b d
Allh (d. 776/1374), A'ml al-a'lmf man buyi'a qabla 'l-ihtilm
min mulk al-lslm, ed. E. L viProvenal (Beirut, 1956).
, Al-Ihta fi akhbr Gharnta, ed. M . 'Inan (4 vols, Cairo,

197377).
, Al-Lamha al-badriya fi ta'rkh al-dawla al-nasriya (Beirut, 1978).
Ibn alQattn, Ab Muhammad Hasan b. 'Al alMarrkush (7th/
13th century), Nuzm al-jumn li-tartb m salafa min akhbr al-zamn,
ed. M.A. Makki (Tetuan, 1964, and Beirut,'1990).
Ibn alQtya, Ab Bakr Muhammad b. 'Umar (d. 367/977), Ta'rkh
iftith al-Andalus, ed. a n d trans. J . M . Nichols (unpublished
P h D thesis, Chapel H i l l , 1975; available on University Microfilms
International); also ed. with Spanish trans., J . Ribera (Madrid,
1868).
Ibn Shib alSalt, Ab Marwan 'Abd alMalik b. Muhammad alBj
(c. 594/1198), Al-Mann bi'1-Imma, ed. A . alHadi alTazi (Beirut,
1964); Spanish trans., A . H u i c i Miranda (Valencia, 1969).
Ibn Sa'd, AblHasan 'Al b. Ms b. Sa'd alMaghrib (d. 685/
1286), Al-Mughrib f hul al-Maghrib, ed. S. Dayf (2 vols, Cairo,
195355).
AlIdrs, Ab 'Abd Allh Muhammad b. Muhammad alHammd
(c. 560/1164), Nuzhat al-mushtq (Naples, 1982).
AlMaqqari, Shihb aldn Ab'lAbbs A h m a d b. Muhammad al
Tilimsn (d. 1041/1631), Nafh al-tb min ghusn al-Andalus al-ratb,
ed. I. 'Abbas (Beirut, 1968); trans., P. de Gayangos, History of the
Mohammaden Dynasties in Spain (2 vols, L o n d o n , 184041).
Nubdhat al-asrfi akhbr Ban Nasr, ed. A. Bustani with Spanish trans.
C. Quiros (L arache, 1940).
Al'Udhri, A h m a d b. 'Umar alDal'i (d. 478/1085), Nuss 'an al-
Andalus min kitabT a r s 'al-akhbr watawz'ed.A.al-thr, ed. A. alAhwani
(Madrid, 1956).
AlWanshars, Ab'l'Abbs A h m a d b. Yahy alTilimsn (d. 914/
1508), Al-Mi'yar al-mughrib wa'l-jam' al-mu'rib 'an fatzu ahl Ifrqya
wa'l-Andalus wa'l-Maghrib (Fez, 189697), ed. M . Hajji (Beirut,
198183).

Muslim Spain and Portugal PDF

Uploaded by

Copyright:

Available Formats

Muslim Spain and Portugal PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Muslim Spain and Portugal PDF

Uploaded by

Copyright:

Available Formats

Muslim Spain and

including photocopying and recording, or in any information storage or retrieval system,

Note o nNames and Dates xi

1. T h e C o n q u e s t a n d the A g e of the A m i r s , 711-56 1

2. T h e Umayyad Amirate, 756-852 30

3. Muhammad, al-Mundhir and A b d

4. T h e G o l d e n A g e of the Umayyad Caliphate,

The policies and strategies of 'Abd al-Rahmn III 82

5. The 'Amirids and the Collapse of the Caliphate

The rise of al-Mansr, 976-81 109

6. The Taifa Kingdoms 130

7. The Empire of the Almoravids 154

8. The Second Taifas 189

9. The Early Almohad Caliphate 196

'Abd al-Mu'min and the construction of the

10. The Later A l m o h a d Caliphate 237

al-Mansr, 1184-99 237

1. Al-Andalus (Spain and Portugal) 326-7

2. Al-'Udwa (Morocco) 328

It is with genuine and heartfelt pleasure that I wish to acknow-

this book by colleagues i n Spain. In this context I would particularly

I have tried to make this book user-friendly to non-Arabists while

ated according to the Cambridge History of Islam system, now gener

akhbr: short historical narratives

balod: local, a term used to describe those Muslims who settled i n

qd: Muslim judge

shurta: police force

This book is intended to provide an account of the political history

this commonwealth, rather than to their northern neighbours, that

rural settlement, landscape, irrigation technology and cuisine, but

de Al-Andalus edited by Manuela Marin and others which have added

ment itself, while P. Guichard's Les Musulmans de Valence is an out-

The Conquest and the Age of the

The Iberian background

T h e Iberian p e n i n s u l a , divided into the great provinces of

1. B. Reilly, T h e M e d i e v a l S p a i n s (Cambridge, 1993), p. 7.

The M u s l i m conquest 711-16

2. T h e literature o n the M u s l i m conquest o f al-Andalus is vast a n d fairly uneven

Muslim conquest and settled i n al-Andalus came from the Butr

when the conqueror of Egypt, ' A m r b. al-s, led an expedition

3. O n this see E . M a n z a n o M o r e n o , L a F r o n t e r a de A l - A n d a l u s en epoca de los Omeyas

Morocco, it is probable that the great majority of the troops i n

b. Nusayr, the Arab governor of Qayrawn). This relationship can

in the fluid politics of early Muslim North Africa and al-Andalus

5. F o r the role o f m a w l in the Umayyad caliphate, see P. C r o n e , Slaves o n Horses.

the establishment of Arab rule is hampered by the nature of the

arrange their materials more or less i n chronological order but they

recensions i n annalistic compilations like Ibn Idhr's Bayn al

6. E d . with Spanish trans., E . L afuente y Alcantara ( M a d r i d , 1867). I have used

controversy has raged about the relative merits and reliability of

It is important to attempt to assess the reliability of this material.

iety of other ways i n which the material needs to be treated with

11. See C o n q u e s t a n d Settlement, pp. 4-14. 12. A r a b C o n q u e s t , pp. 34-5.

other sources close to the Umayyad court, emphasises the force

of genuine confusion over events which happened long ago.

In addition to the Arabic texts there are also Latin sources. O f

The sources for the Muslim conquest and establishment are as

15. See below, p. 50. 16. See Collins, A r a b C o n q u e s t , pp. 2 6 - 3 4 .

importance than they deserve. Despite all these qualifications, how