The Medieval Turks: Collected Essays 9781474485951

Citation preview

The Medieval Turks

Edinburgh Studies in Classical Islamic History and Culture Series Editor: Carole Hillenbrand Titles in the series include: The Body in Arabic Love Poetry: The ‘Udhri Tradition Jokha Alharthi Arabian Drugs in Early Medieval Mediterranean Medicine Zohar Amar and Efraim Lev Towards a History of Libraries in Yemen Hassan Ansari and Sabine Schmidtke The Abbasid Caliphate of Cairo, 1261–1517: Out of the Shadows Mustafa Banister The Medieval Western Maghrib: Cities, Patronage and Power Amira K. Bennison Christian Monastic Life in Early Islam Bradley Bowman Keeping the Peace in Premodern Islam: Diplomacy under the Mamluk Sultanate, 1250–1517 Malika Dekkiche Queens, Concubines and Eunuchs in Medieval Islam Taef El-­Azhari Islamic Political Thought in the Mamluk Period Mohamad El-­Merheb The Kharijites in Early Islamic Historical Tradition: Heroes and Villains Hannah-­Lena Hagemann Islam and the Crusades: Collected Papers Carole Hillenbrand The Medieval Turks: Collected Papers Carole Hillenbrand Classical Islam: Collected Papers Carole Hillenbrand The Books of Burhān al-Dīn: Literacy and Book Ownership in Mamluk Jerusalem Said Aljoumani and Konrad Hirschler Medieval Damascus: Plurality and Diversity in an Arabic Library – The Ashrafīya Library Catalogue Konrad Hirschler A Monument to Medieval Syrian Book Culture: The Library of Ibn ‘Abd al-Hādī Konrad Hirschler The Popularisation of Sufism in Ayyubid and Mamluk Egypt: State and Society, 1173–1325 Nathan Hofer Defining Anthropomorphism: The Challenge of Islamic Traditionalism Livnat Holtzman Making Mongol History: Rashid al-Din and the Jami‘ al-Tawarikh Stefan Kamola Lyrics of Life: Sa‘di on Love, Cosmopolitanism and Care of the Self Fatemeh Keshavarz Art, Allegory and The Rise of Shiism In Iran, 1487–1565 Chad Kia The Administration of Justice in Medieval Egypt: From the 7th to the 12th Century Yaacov Lev The Queen of Sheba’s Gift: A History of the True Balsam of Matarea Marcus Milwright Ruling from a Red Canopy: Political Authority in the Medieval Islamic World, from Anatolia to South Asia Colin P. Mitchell Islam, Christianity and the Realms of the Miraculous: A Comparative Exploration Ian Richard Netton Sacred Place and Sacred Time in the Medieval Islamic Middle East: A Historical Perspective Daniella Talmon-­Heller Conquered Populations in Early Islam: Non-Arabs, Slaves and the Sons of Slave Mothers Elizabeth Urban edinburghuniversitypress.com/series/escihc

The Medieval Turks Collected Papers

Carole Hillenbrand

Edinburgh University Press is one of the leading university presses in the UK. We publish academic books and journals in our selected subject areas across the humanities and social sciences, combining cutting-­edge scholarship with high editorial and production values to produce academic works of lasting importance. For more information visit our website: edinburghuniversitypress.com © Carole Hillenbrand, 2022 Edinburgh University Press Ltd The ­Tun – ­Holyrood Road 12 (2f ) Jackson’s Entry Edinburgh EH8 8PJ Typeset in 11/15 Adobe Garamond by Cheshire Typesetting Ltd, Cuddington, Cheshire, printed and bound in Great Britain A CIP record for this book is available from the British Library ISBN 978 1 4744 8594 4 (hardback) ISBN 978 1 4744 8595 1 (webready PDF) ISBN 978 1 4744 8597 5 (epub) The right of Carole Hillenbrand to be identified as author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988 and the Copyright and Related Rights Regulations 2003 (SI No. 2498). Published with the support of the University of Edinburgh Scholarly Publishing Initiatives Fund.

Contents

List of Figuresvii Acknowledgements ix Preface x   1 The Career of Najm al-­Din İl-­Ghazi

1

  2 The Establishment of Artuqid Power in Diyar Bakr in the Twelfth Century

50

  3 The History of the Jazira, 1100–1250: A Short Introduction

76

 4 Malazgird

89

 5 Marwanids

94

 6 Mayyafariqin

98

  7 Mu‘in al-­Din Parwana: The Servant of Two Masters?

108

 8 Mu‘in al-­Din Sulayman Parwana

117

  9 1092: A Murderous Year

121

10 Ibn al-‘Adim’s Biography of the Seljuq Sultan, Alp Arslan

136

11 The Power Struggle between the Seljuqs and the Isma‘ilis of Alamut, 487–518/1094–1124: The Seljuq Perspective

145

12 Some Reflections on Seljuq Historiography

163

vi  |  the m ed ie va l tur k s 13 Women in the Seljuq Period

182

14 Ravandi, the Seljuq Court at Konya and the Persianisation of Anatolian Cities

202

15 Artuqids

215

16 What’s in a Name? Tughtegin – ‘The Minister of the Antichrist’? 220 17 Aspects of the Court of the Great Seljuqs

234

18 Nizam al-­Mulk: A Maverick Vizier?

254

19 The Life And Times of ‘Amid al-­Mulk al-­Kunduri

270

20 The Nizamiyya Madrasas

288

21 The Life and Times of the Artuqid Ruler Najm al-­Din Alpı (r. 548/1154–572/1176)

302

22 What Is Special about Seljuq History?

321

23 ‘The View from Above’: Muslim Perceptions of the Turks of Syria and the Jazira in the Period 1070–1176

336

24 A Cosmopolitan Frontier State: Relations among the Kurds, Arabs, Byzantines, Armenians, Persians and Turks under the Marwanids of Diyar Bakr, 990–1085

361

Original Sources of the Items in this Volume 376 Index379

Figures

16.1 Turkish warrior, galloping into battle, carrying a tugh227 16.2 Ottoman tughs captured at the Battle of Vienna, 1683 227 17.1 Frontispiece of the Kitab al-Diryaq of Pseudo-­Galen, probably Mosul, c. 1250239 17.2 Al-­Jazari nawba scene, Kitab fi Ma‘rifat al-Hiyal al-Handasiyya, Iraq, 1206 242 17.3 Illustration from Varqa va-Ghulshah by ‘Ayyuqi, probably Konya, c. 1250245 21.1 Copper dirham of Najm al-­Din Alpı, 1156–76 307 21.2 Another copper dirham of Najm al-­Din Alpı 308 22.1 Map of the Seljuq empire 322 22.2 ‘Turkish cavalry’, from Rashid al-­Din’s Jami‘ al-tawarikh, Iran, 1314324 22.3  Demon in Chains, in Siyah Qalam style, Iran or Central Asia, c. 1453325 22.4 ‘Court tents’, from Rashid al-­Din’s Jami‘ al-tawarikh, Iran, 1314326 22.5 ‘Malikshah’, from Rashid al-­Din’s, Jami‘ al-tawarikh, Iran 1314327 22.6 Modern Turkish Manzikert celebratory stamp, 1971 328 22.7 Mausoleum of Seljuq sultan Sanjar 329 23.1 Conversation among Turkish nomads 336 23.2 ‘Cavalry pursuit’, from Rashid al-­Din’s, Jami‘ al-tawarikh, Iran, 1314343

viii  |  the m ed ie va l tur k s 23.3 Parthian shot from a fifteenth-­century Persian manuscript 23.4 Two nomad women in conversation 23.5 Shaman 23.6 Minbar of Nur al-­Din, the Aqsa Mosque, Jerusalem 24.1 Map of Marwanid Kurds 24.2 Tower on the walls of Mayyafariqin

344 345 351 353 362 366

Acknowledgements

I should like to express my great thanks to a number of people at Edinburgh University Press who have helped in various ways in the publication of this book: Nicola Ramsey, Kirsty Woods, Eddie Clark, Caitlin Murphy; and copy-­editor Lel Gillingwater and indexer Samantha Clark. I am grateful to the original publishers of these articles for permitting them to be reprinted in this volume. As always, I am extremely indebted to my husband Robert for his advice and support during the preparation of this book.

ix

Preface

I have always been fascinated by languages and I like to speak them as well as read them. As a teenager I made annual visits to France and indeed spent a year there teaching in a Burgundian convent school before entering university. I also visited Germany and Italy regularly to improve my knowledge of those languages. So my very first visit to the Middle East, which was to Turkey in 1967 and lasted two months, opened fascinating new linguistic horizons. As an undergraduate at Cambridge I had studied for a degree in modern and medieval European languages, specialising in French political thought, Romance Linguistics, Old French, Rumanian and Medieval Latin. So my initial predilection for languages became still more firmly established. During my three undergraduate years I spent valuable time in Germany, Italy, Rumania and, above all, France. Those eight weeks in Turkey proved to be a new and fascinating experience which turned my career aspirations upside down. Turkey’s diversity, the contrast between the celebrated historic city of Istanbul and the rarely visited furthest eastern regions of the country, and a journey by sea along the coast of the Black Sea, proved memorable and exciting. Very soon I felt drawn to finding out more about the multicultural history of eastern Turkey, and in particular the area around Lake Van and the remote cities of Kars, Amid (now known as Diyar Bakr) and Mayyafariqin (now called Silvan) close to the Turkish border with Georgia and Armenia. There we hitch-­hiked x

pr ef a ce  | xi and took lengthy walks in search of little-­known Armenian, Georgian and Kurdish sites and monuments as well as Turkish ones. Knowing no Turkish at this point, we found that local men had learned some German during and after the Second World War; this helped with information about the location of the key places that we wanted to visit. On my return to Britain I finally decided to return to academic life. I went to Oxford to study Middle Eastern Languages, beginning with Arabic. I was thrilled to be the first person at Somerville College ever to study Arabic. My degree course was interrupted after two years by my husband’s need to visit Iran to carry out fieldwork for his DPhil, which was on the subject of medieval Iranian mausolea. We stayed in Iran for a whole year; this gave me a great opportunity to learn Persian and to visit many wonderful historical sites. I explore the impact of that year in more detail in the preface to Volume 3 of this set of my reprinted papers. Once back in Oxford I was fortunate to be taught modern and Ottoman Turkish by three scholars: Professors Victor Ménage and Geoffrey Lewis and Dr Richard Repp. Then in 1972 I moved to Edinburgh, after my husband had been appointed Lecturer in Byzantine and Islamic Art there, and began a PhD under the supervision of another Turkish specialist, John Walsh. He was passionately committed to Ottoman studies and wanted me to undertake research on Ottoman Turkish history. He suggested that I should work on the diaries of Marino Sanuto, since I knew Italian, but I decided instead to go back in historical time to study the nomadic Artuqid Turkish dynasty, which ruled eastern Anatolia and the Jazira in the twelfth century. It quickly became clear that this would involve editing, translating and analysing the sections of two unpublished Arabic manuscripts in the British Library which related the history of the Artuqids in Mayyafariqin, work which had been identified as a desideratum by earlier scholars, among them Amedroz, Minorsky and Cahen. The palaeographical challenges of this task were daunting; the scripts in both manuscripts were rough and ready, and the texts were unpointed. This chronicle, known as Ta’rikh Mayyafariqin wa-Amid, was written by Ibn al-­Azraq al-­Fariqi (died sometime after 1176–7), a scribe and official who worked for several Artuqid rulers. After completing my PhD in 1978, I was appointed as Lecturer in Arabic at the University of Edinburgh. For the next thirty years I taught, among

xii  |  the me d ie va l tur k s much else, the first-­year undergraduate Islamic history course, which covered the period from the rise of Islam to the Ottoman and Safavid empires until 1725. I was particularly keen, after the experience of my PhD, to continue studying the medieval Turks. The history of the early nomadic Turks in Central Asia and further east had already been analysed with great erudition by famous experts such as Wilhelm Barthold, Vladimir Minorsky, Richard Frye, Robert Dankoff and others. But there was still much to do. Once the nomadic Seljuq Turks had become the rulers of Iran, Iraq, the Jazira and Syria from the mid-­eleventh to the end of the twelfth century, their history was destined to be recorded in works written in Arabic and Persian. At this time the Turks themselves had scarcely found a voice to tell their own story; they were nomadic newcomers to the Middle East and unschooled in its tradition of writing detailed historical chronicles. The same absence of contemporary Turkish language sources also marks the earlier history of the pre-­Seljuq Turks in the Islamic world, notably that of the slave soldiers of the ‘Abbasid caliphs in Baghdad and the Ghaznavids in Iran and Afghanistan. The scholarship of the last forty years has done a great deal to fill this gap in modern knowledge of this period of Turkish dominion. I therefore remained determined that the centre of gravity in my research should be the history of the Seljuq Turks from the eleventh to the thirteenth centuries, a period when they remained the salient power in the eastern Muslim world from Central Asia to western Anatolia and the borders of Egypt. So from the beginning of my employment at the University of Edinburgh in 1979, I studied the Arabic and Persian primary sources that dealt with Seljuq history. I also consulted the copious modern secondary scholarship in Turkish written on this subject. Fortunately, in the thirty years of my career at Edinburgh, I was able to publish steadily on the Seljuqs, as well as teaching their history to first-­year undergraduates in my Islamic History course. Unfortunately, only two of the fifty-­nine postgraduates whom I supervised, virtually all of them working for a PhD, chose to study Turkish topics. I have given public lectures or conference papers on Seljuq history, many of them essentially the first versions of the papers reprinted in this volume, in the universities of Istanbul, Mardin, Bir Zeit, Salamanca, New York, St Louis, Yale and Notre Dame, as well as at a good number of British universi-

p re f a ce | xiii ties. I have also regularly published a series of articles and book chapters about the period of the Seljuqs, including such topics as military matters, the sultanate, the Seljuq approach to Islam, the ­vizierate – e­ specially the career of Nizam al-­Mulk – ­and the role of Seljuq women. In 2007, I undertook a more difficult research task, namely exploring the full context of the celebrated and seminal battle that consolidated the Turkish presence in Anatolia: the Seljuq victory over the Byzantine army in the Battle of Manzikert in 1071. This was perhaps the pivotal medieval event in that country before 1453, which saw the fall of Constantinople to the Ottoman Turks. For the account of the actual battle, I drew on a good number of primary Arabic and Persian texts, as well as a few Byzantine, Armenian and Syriac sources which had been translated into French and English. However, the book also examined the impact of this battle in later Turkish history; this required a prolonged analysis of Ottoman Turkish sources as well as other Ottoman material translated into French, German and English. The last part of the book explored how the Turks of modern times have understood the battle and its aftermath, and that required close study of the writings of key twentieth-­century Turkish scholars. For understandable reasons, Manzikert has for centuries remained a crucial part of Turkish identity: a long-­lasting and much-­cherished symbolic memory of past Turkish greatness right up to the present day. Hence the title I chose for it: Turkish Myth and Muslim Symbol: The Battle of Manzikert. I was very happy that this book, published by Edinburgh University Press, was immediately favoured with a Turkish translation. I first encountered the medieval Turks at the outset of my career as a historian of the medieval Muslim world, and I count myself fortunate to have done so. Vivid memories not just of Istanbul but of such key Seljuq centres as Konya, Sivas, Kayseri, Erzerum, Alanya and Antalya, and more generally of eastern Anatolia where they were such a dominant presence and have left such magnificent monuments, have informed my study of the medieval sources for many decades. And the numerous visits I have made to Turkey since 1967 have served to embed in my consciousness a constant awareness of the land of Anatolia, its plains and mountains, its rivers and ­valleys – ­a land of breath-­taking natural beauty. Those memories of the land, its people and its monuments have served as a corrective to the view of the

xiv  |  the m ed ie va l tur k s Seljuqs seen through the distorting lens of the medieval historians who were predominantly Arabs and Persians, and who wrote in Arabic and Persian, not Turkish. Once again, then, personal experience can decisively inform scholarship.

This book is dedicated to a remarkable trio of modern scholars of medieval Turkish history – Peter Golden, Gary Leiser and Andrew Peacock

1 The Career of Najm al-­Din ˙Il-­Ghazi

Introduction

˙I

l-­Ghazi has received scant treatment by Islamic and Western historians alike. Historians of Islam have concentrated inevitably and justifiably on the vital task of establishing a basic chronology from the tangled information provided by the medieval chroniclers.1 Scholars of European history, on the other hand, have treated İl-­Ghazi merely as an incidental figure in the wider sweep of Crusader history which spanned so much more than the career of this one Turkish amir, active at a time of Muslim disunity and Crusader conquest.2 Had İl-­Ghazi lived half a century later, he might conceivably have been accorded more systematic scholarly study, such as the impressive corpus of work that has appeared on Nur al-­Din3 and especially Saladin.4 It is perhaps premature to attempt to analyse the career of İl-­Ghazi in a way which deviates from the standard approach of outlining the various stages of his career in chronological sequence and providing documentary support for the information given. Yet many good studies of the chronological kind have already been made, notably the article by Claude Cahen5 in which from a wide array of sources he gives in broad outline the complex political history of the period covered by the Artuqid section of Ibn al-­Azraq’s Ta’rikh Mayyafariqin wa-Amid. This deals with the period from the establishment of İl-­Ghazi in Diyar Bakr to the reign of the third Artuqid ruler of Mardin and Mayyafariqin, Najm al-­Din Alpı. More recently, the Turkish scholar, Ali Sevim, has written at some length about ‘the political deeds’ of Artuq, İl-­Ghazi’s father, Suqman, his brother, and İl-­Ghazi himself;6 and Osman 1

2  |  the me di e va l tur k s Turan devotes six pages to İl-­Ghazi in his recent history of eastern Turkey.7 He adds nothing new, however. More summary and general still are the articles on the Artuqids, and more especially on İl-­Ghazi, in the Encyclopaedia of Islam8 and the Türk Ansiklopedesi.9 Although of varying standard, these provide a skeletal chronological framework for the career of İl-­Ghazi. The various works listed in this paragraph may differ in points of detail, but their basic concerns remain similar; to establish a chronology which includes the major political events of İl-­Ghazi’s life and not to analyse the implications of the various facts cited. If a new primary source came to light, there would be justification for a fresh analysis of the detailed facts year by year. In the absence of such a bonus, there is a need, as Gibb indicated many years ago, for monographs on the important figures of the First Crusade, such as Tughtegin, Zengi and İl-­Ghazi.10 In this article, therefore, the career and personality of İl-­Ghazi will be considered under certain broad themes, in the hope that new light may be shed on this little-­known but significant ruler. The sources used are those concerned with the First Crusade, which are too well known11 to be described yet again. The work of Ibn al-­Azraq will also be used as a source, but principally for İl-­Ghazi’s activities in Diyar Bakr and above all for the author’s account of İl-­Ghazi’s campaign into Georgia in 515/1121–2. Historians of the Crusades have discussed what they take to be the flamboyantly barbaric personality of İl-­Ghazi and despite the somewhat lacunary information available have pronounced strong judgements upon him. Röhricht, as a creature of his times, condemns İl-­Ghazi as a brute given over to drunkenness and excess,12 whilst the same opinion is expressed with greater restraint by Grousset, who labels him a coarse old trooper.13 Runciman has more level-­headed criticisms of İl-­Ghazi,14 but even he judges him to a great extent, though not entirely, from the viewpoint of Crusader history. Such an attitude is totally justified in a scholar who is writing about the Crusades, but it presents a rather one-­sided view of İl-­Ghazi, leaving undiscussed many vital facets of his career. From the Muslim standpoint his dealings with the Franks were of secondary importance in comparison with his crucial role in the Jazira. It is indeed a commonplace of Islamic history that the advent of the Franks in the last decade of the eleventh century and their territorial expan-

T h e Ca r ee r of Na jm a l -D in il - Gh a z i | 3 sion in northern Syria thereafter exercised no dominant influence in the internecine political struggles of the Islamic heartland.15 Furthermore, İl-­ Ghazi, throughout his career, was orientated eastwards towards the Jazira and Baghdad, even to western I­ran – t­hat is, to the territory of his overlord the ­sultan – a­ nd his ambitions and activities were inextricably linked with developments in the Seljuq empire. It is especially against the background of late Seljuq history, therefore, that İl-­Ghazi’s achievements should be evaluated. In the discussion which follows, therefore, İl-­Ghazi’s career will be assessed primarily in the context of Seljuq decline. His role as a semi-­nomadic chief will also be considered. His relations with the Franks cannot of course be ignored, but the Franks will be accorded subsidiary status as a small but significant political entity operative in northern Syria in the early twelfth century. They will not take over the foreground of the discussion, as is standard practice in works on the Crusades. The following abbreviations are used in this article:





Anon. Syr. Chron. A. S. Tritton, ‘The First and Second Crusades from an Anonymous Syriac Chronicle’, with notes by H. A. R. Gibb, JRAS (1933), 69–101, 273–305. al-‘Azimi C. Cahen, ed., ‘La chronique abrégée d’al‘Azimi’, JA CCXXX (1938), 353–44. Bar Hebraeus The Chronography of Gregory Abû’l Faraj, the Hebrew Physician, commonly known as Bar Hebraeus, tr. E. A. W. Budge (London, 1932). al-­Bundari Zubdat al-nusra wa nukhbat al-‘usra, ed. M. T. Houtsma (Leiden, 1889). al-­Husayni Akhbar al-Dawla al-Seljuqiyya, ed. M. Iqbal (Lahore, 1933). Ibn al-‘Adim, Bughyat Bughyat al-talab fi Ta’rikh Halab, ed. A. Sevim (Ankara, 1976). Ibn al-‘Adim, Zubda Zubdat al-halab min Ta’rikh Halab II, ed. S. Dahan (Damascus, 1954). Ibn al-­Athir, X al-Kamil fi’l-ta’rikh X, ed. C. J. Tornberg (Leiden and Uppsala, 1864).

4  |  the me di e va l tur k s

Ibn al-­Athir, Atab. Ta’rikh al-dawla al-atabakiyya, ed. A. Tulaimat (Cairo, 1963). Ibn al-­Azraq, Ms. A Ta’rikh Mayyafariqin wa-Amid, B.M. Or. 5803. Ibn al-­Azraq, Ms. B Ta’rikh Mayyafariqin wa-Amid, B.M. Or. 6310. Ibn al-­Azraq, ed. ‘Awad Ta’rikh al-Fariqi, ed. B. A. L. ‘Awad (Cairo, 1959). Ibn al-­Furat Ta’rikh al-Duwal wa’l-Muluk, Vienna MS A.F. 811. Ibn Khallikan Wafayat al-a‘yan, tr. Baron W. M. de Slane (Paris, 1843–71). Ibn al-­Qalanisi Dhayl Ta’rikh Dimashq, ed. H. F. Amedroz (Leiden, 1908). Marsh 333 Ibn Shaddad, al-A‘laq al-khatira, Bodleian Ms. March 333. Matthew of Edessa Patmut’iwn, tr. E. Dulaurier (Paris, 1858). Michael the Syrian Chronique de Michel le Syrien, tr. J.-B. Chabot (Paris, 1899–1914). RHC Recueil des Historiens des Croisades, Historiens orientaux (Paris, 1872–1906). Sibt b. al-­Jawzi, ed. Jewett Mir’at al-Zaman (Hyderabad, 1951); facsimile ed. by J. Jewett (Chicago, 1907). Usama, tr. Hitti Memoirs of an Arab-Syrian Gentleman, tr. P. K. Hitti (repr. Beirut, 1964). The Career of İl-­Ghazi – ­First Phase The decline of the Seljuq empire did not manifest itself markedly until the deaths in quick succession of Nizam al-­Mulk and Sultan Malikshah in 485/1092–3, although the seeds of its decay had been sown long before.16 During the early part of his career, that is before 485/1092–3, İl-­Ghazi’s role was most probably identical with that of many prominent leaders of the Turcoman tribes who had helped to bring the Seljuqs to power in the Islamic world. He fought in the service of the Seljuq sultans wherever they might send him and was presumably rewarded. At this stage there would have been

T h e Ca r ee r of Na jm a l -D in il - Gh a z i | 5 little opportunity for him to gain independence or to acquire territory on his own account. Sultan Malikshah and Nizam al-­Mulk seem to have exercised a tight control and commanded loyalty from their military commanders, who were ordered to take cities and subjugate ­enemies – ­but in the name of Malikshah.17 Before 485/1092–3 İl-­Ghazi probably fought with his father, Artuq, on his various military campaigns to Diyar Bakr18 and Jerusalem,19 and while Artuq was alive, İl-­Ghazi was under his controlling influence. The Career of İl-­Ghazi – ­Second Phase The death of Malikshah in 485/1092–3 and that of Artuq shortly afterwards afforded İl-­Ghazi the beginning of a freedom from constraint, both personal and governmental, which was further aided by the death of his elder brother Suqman in 498/1104–5,20 whereupon İl-­Ghazi became head of the Artuqid family. Between the death of Malikshah and İl-­Ghazi’s acquisition of the fortress of Mardin some time around 502/1108–9, İl-­Ghazi played an important part in the bitter and protracted conflict between the successors of Malikshah. He also participated in the equally dangerous struggle for power amongst the Seljuq military commanders themselves, of whom he was one. All of them were eager to emancipate themselves from centralised control and acquire territory of their own. İl-­Ghazi’s brother, Suqman, as well as Aq Sunqur al-­Bursuqi, Chökermish and others all played the same game. They aligned themselves according to political expediency with one or other of the Seljuq contenders. According to the caprice of fortune they could gain wealth and power if they had backed the winning side, but they found death or imprisonment if they lost. In the last decades of the eleventh century, then, İl-­Ghazi sought advancement in the same way as the other Turcoman chiefs in the Seljuq empire by attaching himself to a powerful scion of the Seljuq family. He was active in northern Syria in the struggle between Duqaq and Ridwan21 and he held power briefly in Jerusalem. Artuq, İl-­Ghazi’s father, had been given Jerusalem as an iqta‘ by Tutush in 478/1085–6. When Artuq died, Suqman and İl-­Ghazi succeeded him there. It is not altogether clear which of the two brothers was in charge, or whether they ruled jointly.22 Whatever the arrangement in Jerusalem, it was short-­lived, since the Fatimids, profiting from Seljuq weakness, attacked the city in 489/1095–6 and took it from the Artuqid family.23 At that point

6  |  the me di e va l tur k s both Suqman and İl-­Ghazi were there, with their cousin Sevinch and their nephew Yaquti. The presence of this cluster of Artuqid princes may be interpreted as an attempt by the family to carve out an appanage for themselves in the Jerusalem area. After al-­Afdal had granted them safe conduct out of Jerusalem, the two brothers parted company, Suqman staying in Edessa whilst İl-­Ghazi went to Iraq.24 The Jerusalem interlude is not of great significance in İl-­Ghazi’s career, but it is an interesting foretaste of the geographical distance he was time and again to cover in his search for profit. Of paramount importance, however, to an understanding of İl-­Ghazi’s motivation and later career is the move he made from Jerusalem to the service of the future Sultan Muhammad at a time when the latter was struggling for the supremacy in the western part of the Seljuq empire, namely western Iran and Iraq. Sanjar, Muhammad’s full brother, was already ruler of Khurasan. The move to Iraq by İl-­Ghazi was not necessarily an obvious one, for he could have thrown in his lot with Ridwan who was operating just north of the Jerusalem area. Suqman, it will be noted, decided to remain in Syria, and this decision may have influenced İl-­Ghazi. It is also possible that representatives of the Artuqid family were still at Hulwan, his father’s old iqta‘.25 İl-­Ghazi’s relationship with Sultan Muhammad overshadowed his whole career and henceforth he was never free of the spectre of Muhammad at his shoulder until the latter’s death at the end of 511/1118. The presence was not always a malevolent one. When İl-­Ghazi went to Iraq, some time after 489/1095–6, he probably went to Hulwan, and thereafter gave military service to Muhammad. In 494/1100–1, Sultan Muhammad and Sanjar made their way to Baghdad and were joined at Hulwan by İl-­Ghazi whom Ibn al-­Athir praises, saying that he served the sultan well.26 No doubt as a reward for his services, Sultan Muhammad appointed him shihna of Baghdad in the following year.27 The post of shihna in Baghdad was powerful and prestigious. The shihna was the main representative of the sultan in Baghdad, charged with maintenance of law and order and with the task of restraining any attempt by the caliph to interfere in the sultan’s sphere of influence.28 The evidence given by Ibn al-­Athir reveals that İl-­Ghazi became attached to this office, although the people of Baghdad did not share this enthusiasm. The presence in and

T h e Ca r ee r of Na jm a l -D in il - Gh a z i | 7 around Baghdad of İl-­Ghazi’s Turcoman groups engendered civil unrest and riots. On one occasion, in Rajab 495/April–May 1102, İl-­Ghazi, instead of using diplomacy and his power as their leader to subdue his men, set about looting the city in retaliation for the seizure of one of his followers. Only the efforts of the caliph, al-­Mustazhir, prevented İl-­Ghazi from extending the pillaging elsewhere in the city.29 There is no record that Muhammad exacted retribution for this gross misrule. Perhaps the enjoyment he found in controlling Baghdad and his lack of any territory apart from Hulwan contributed to a disastrous political mistake which İl-­Ghazi made in 497/1103–4 when he changed his allegiance from Muhammad to Barkyaruq. Under the terms of the uneasy peace treaty made between Muhammad, Barkyaruq and Sanjar in that year,30 Barkyaruq was to be overlord of Baghdad. Instead of removing himself from that city to seek service elsewhere with Muhammad, who had been given the territories of Armenia, Azerbaijan, the Jazira, Mosul and ­Syria – ­territories which lay more within İl-­Ghazi’s sphere of operations than south-­western Iran and southern ­Iraq – İ­ l-­Ghazi remained in Baghdad and inserted Barkyaruq’s name in the khutba in Jumada I 497/February 1104.31 It is not hard to guess İl-­ Ghazi’s motivation in changing sides. He explained his reasons to Sadaqa who announced his intention of driving İl-­Ghazi out of Baghdad. He told Sadaqa that his only possession until he became shihna of Baghdad had been Hulwan, and that it was because of Baghdad that he had recognised Barkyaruq as his sultan. Sadaqa apparently understood these arguments and returned to Hilla.32 Nor was İl-­Ghazi’s new-­found loyalty likely to ingratiate him with Sultan Barkyaruq, who, if Ibn al-­Athir’s chronology is to be trusted,33 had already sent his own shihna, Gümüshtegin, to Baghdad in Rabi‘ I 496/December 1102, in an attempt to oust İl-­Ghazi, his brother’s candidate. İl-­Ghazi had allied himself with Sadaqa, the lord of Hilla, who had recently shifted his allegiance from Barkyaruq to Muhammad,34 and with his own brother Suqman, whom he had summoned from Hisn Kayfa, and all together they had succeeded in driving Gümüshtegin out of Baghdad.35 It seems unlikely that their joint action was instigated by Sultan Muhammad although he was no doubt glad of the discomfiture of Gümüshtegin. But İl-­ Ghazi had now rendered himself unacceptable to Sultan Muhammad and was at best reluctantly allowed by Barkyaruq to continue as

8  |  the me di e va l tur k s shihna. İl-­Ghazi’s visit to Isfahan in 498/1104–5 to see Barkyaruq may well have been motivated by a desire to persuade the sultan of his loyalty, as well as to press him to come to Baghdad to strengthen his precarious hold over the city. It was en route to Baghdad that Barkyaruq died.36 Instead of attempting at this point to make amends to Muhammad by mentioning his name in the khutba at Baghdad, a move which might conceivably have reinstated himself in Muhammad’s favour, İl-­Ghazi proceeded with Barkyaruq’s infant son, Malikshah, and his atabeg Ayaz, as far as Baghdad. There they proclaimed Malikshah sultan.37 This move was of course prompted by a desire to acquire more power for themselves since they could rule through the small boy. İl-­Ghazi’s attachment to his post in Baghdad had induced him to make another egregious error of judgement. Soon after, Muhammad marched on Baghdad and assumed control of his brother Barkyaruq’s territories,38 thereby putting an end to İl-­Ghazi’s ambitions in Iraq. Sultan Muhammad’s coolness towards him was understandable; so too was the sultan’s next move, which was to replace İl-­Ghazi39 at Baghdad in 498/1105 by Aq Sunqur al-­Bursuqi.40 With the establishment of Sultan Muhammad as sole Seljuq ruler in the west in that year, and in view of İl-­Ghazi’s record of insubordination and unpredictability, a rift between the two was inevitable. In the years that followed, Sultan Muhammad’s attitude was to harden from probable coolness to anger and hostility, whilst İl-­Ghazi’s personal animosity towards the sultan provided a bitter spur for many of his subsequent acts of insubordination. By a strange coincidence, İl-­Ghazi found himself deprived of position and prestige at the same time as his elder brother Suqman, who had possession of Hisn Kayfa and Mardin, died on his way to answer an appeal from Tughtegin in Damascus.41 İl-­Ghazi’s decision to go to Diyar Bakr and assume the leadership of those members of his family who were already there was to pave the way for his eventual acquisition of his own territory. The Career of İl-­Ghazi – ­Third Phase After the struggle for power between Barkyaruq and Muhammad had ended and the latter had become ruler of western Iran and Iraq, the Seljuq empire gained a sovereign who recognised the need to impose his control over the dissident factions within his territories and to restore unity.42 His removal of

T h e Ca r ee r of Na jm a l -D in il - Gh a z i | 9 İl-­Ghazi from the post of shihna of Baghdad indicated clearly his determination to tolerate no amir of doubtful loyalty in such a key post. After his departure from Baghdad, İl-­Ghazi stayed in Diyar Bakr, far from the new sultan’s capital. His deliberate policy appears to have been to ally himself with the enemies of the sultan, whilst initially, at least, maintaining a semblance of obedience. In 501/1107–8, the caliph attempted to mediate in a quarrel between Sadaqa and the sultan. In his reply, Sadaqa informed the caliph that Chavlı Saqao and İl-­Ghazi had agreed to make common cause with him in any war that he might have with the sultan or any other person. He had only to call and they would come with their troops.43 This is one explicit indication of İl-­Ghazi’s declared willingness to lend support to powerful enemies of the sultan. Once secure in the fortress of Mardin, well known to be virtually impregnable and situated far from Muhammad’s centre of g­ overnment – a­ fortress which he took in 502/1108–­9 – İ­l-­Ghazi gradually became bolder in his hostility to the sultan. Muhammad appointed Mawdud to the post of governor of Mosul, dismissing Chavlı Saqao in 501/1107–8.44 This position was of key importance in consolidating the sultan’s authority over the Jazira and carried with it great prestige. Opinions vary as to the extent to which the successive armies which were sent out from Mosul between 500/1106–7 and 509/1115–16 acted on the direct orders of the sultan. Some scholars have seen these campaigns as expressing the ambitions of successive governors of Mosul; others regard them as inspired at least in part by the sultan, who called on his amirs to participate in a joint enterprise against rebel amirs (İl-­Ghazi and Tughtegin) and the Franks, in order to restore his direct control over Syria and Diyar Bakr.45 Whatever the exact motivation of such campaigns, a series of armies were sent out from Mosul, under Mawdud, Aq Sunqur al-­ Bursuqi and Bursuq b. Bursuq, within the space of five years. Their activities directly affected İl-­Ghazi, already incensed by the appointment of Mawdud as governor of Mosul. Two of these armies were sent via Mardin so that efforts could be made to bring İl-­Ghazi to heel. These efforts were invariably unsuccessful. In the event, the latter two armies also experienced humiliating reverses. At first, İl-­Ghazi’s own actions betrayed indecision. He may well still have cherished hopes of the sultan’s favour, especially if after a period in

10  |  the m ed ie va l tur k s disgrace he had shown his behaviour to be exemplary. This is a plausible explanation for his inconsistency with Chavlı, the dispossessed ruler of Mosul, who wanted to enlist İl-­Ghazi’s help in 502/1108–9. İl-­Ghazi was too weak at that time to resist Chavlı’s threats, especially when the latter actually came to Mardin, and so he unwillingly accompanied him. But he grasped the first possible opportunity to escape back to Mardin.46 A similar inconsistency marked İl-­Ghazi’s conduct in 503/1109–10, when he went to Syria with his Turcoman groups in answer to the sultan’s appeal to take part in a joint campaign with other amirs under the leadership of Mawdud, lord of Mosul.47 Although he participated in the fighting, İl-­Ghazi quarrelled with Suqman al-­Qutbi during the campaign. Suqman was a close territorial rival of his, and his career, in contrast to İl-­Ghazi’s, had advanced smoothly through his calculated loyalty to the sultan. After the quarrel, İl-­Ghazi withdrew to Mardin.48 There he remained, refusing to fight. When the campaign was renewed the following year (504/1110–11) a vast army assembled under the leadership of Mawdud. As İl-­Ghazi did not dare at this early stage to refuse the sultan altogether nor to oppose his army, he did send troops, but under the leadership of his son Ayaz, not himself. He was the only amir summoned to fight who did not appear in person.49 The murder of Mawdud in 507/1113–1450 removed one powerful rival from İl-­Ghazi’s path but in the event its consequences were to fan still further his resentment against Muhammad. İl-­Ghazi had already been ignored once by the sultan when the latter had appointed Mawdud as governor of Mosul. Now, in 508/1114–15, the sultan replaced Mawdud with Aq Sunqur al-­Bursuqi, the man who had taken over İl-­Ghazi’s position as shihna in Baghdad.51 İl-­Ghazi’s short-­lived attempt at conformity with the sultan’s wishes had proved fruitless, and he refused to answer the next call to arms from the sultan. In 508/1114–15 Aq Sunqur al-­Bursuqi went to Mardin and forced the reluctant amir to submit to the sultan’s authority. The most İl-­Ghazi would concede was to hand over a body of troops under his son, Ayaz. Aq Sunqur did not accept this insult lightly. He subsequently arrested Ayaz and plundered the countryside around Mardin.52 İl-­Ghazi enlisted help from his nephew, Da’ud of Hisn Kayfa, fought and defeated Aq Sunqur and freed Ayaz.53 By this defeat of Aq Sunqur al-­Bursuqi, İl-­Ghazi’s transparently half-­ hearted support of the sultan had given way to open defiance.

T he C a re er o f Na j m a l-D in il - Gh a z i | 11 At this juncture, İl-­Ghazi apparently began to fear the wrath of the sultan, even in the security of his fortress at Mardin, especially after he had received threatening letters from Muhammad. Seeking an ally with a similar outlook, he fled to Tughtegin in Damascus and joined forces with him.54 Tughtegin had just cause to believe that he had incurred the sultan’s displeasure after Mawdud had been murdered in the previous year whilst in his company at Damascus. Although the blame for the assassination had been placed as usual on the Isma‘ilis, some of the chroniclers suggest Tughtegin’s own complicity in the deed.55 The alliance formed by İl-­Ghazi and Tughtegin in 508/1114–15 was of long duration and mutually beneficial. Together they made a treaty that same year with Roger of Antioch56 and awaited the arrival of the combined forces of the sultan sent out under a new general, Bursuq. The twin objectives of this campaign were to quell the pride of İl-­Ghazi and to prosecute the jihad against the Franks.57 The order of priorities is significant here. After İl-­Ghazi, Tughtegin and Roger of Antioch had assembled their troops and Bursuq’s army had arrived near Aleppo, no battle actually took place. After eight days, Bursuq retreated, fell into an ambush set by Roger at a place called Danith, and was defeated. The battle took place in Rabi‘ II 509/September 1115.58 This serious defeat of an army sent out under the auspices of the sultan appears to have aroused feelings of guilt and fear in Tughtegin and perhaps in İl-­Ghazi too. Tughtegin in Damascus was closer to Frankish territories and had more to lose from a complete break with the Seljuqs. He therefore broke off his alliance with Roger and made his peace with Sultan Muhammad.59 It seems that Tughtegin was not asked to break off relations with İl-­Ghazi as the price of peace; he may even have spoken up for İl-­Ghazi at Baghdad. İl-­Ghazi did not feel the same pressures as his ally, although he did not fight again on the same side as the Franks. He apparently saw no need to seek pardon from the sultan since he had no cause to fear any more reprisals.60 Nevertheless he did not provoke the sultan further by continued defiance. He waited until the sultan’s death at the end of 511 and then sent his son, Temürtash, to Muhammad’s successor, Mahmud, with whom no doubt he hoped to have more friendly but indirect relations.61 For his part, after Danith Sultan Muhammad made no other serious attempts to quell the ambition of İl-­Ghazi, nor did he send out another large

12  |  the m ed ie va l tur k s army into Syria. İl-­Ghazi’s rebellious stance had prevented direct interference by the Seljuq sultan in his affairs, if somewhat belatedly (the sultan died in 511/1118 and İl-­Ghazi himself in 516/1122), and had permitted him to establish his own small but independent territory in the Jazira. After Muhammad’s death, and the succession problems that ensued, İl-­Ghazi wielded genuine power in his own right. From the viewpoint of the external enemies of Seljuq power in the Near East, especially the Franks, and the Georgians, who had begun harassing the northern borders of Muhammad’s territories,62 the sultan’s struggles to impose order within his empire by disciplining amirs like İl-­Ghazi successfully diverted Muhammad’s attention away from their own activities, wasted his valuable time and energy and prevented him from undertaking serious military action against them. These reflections prompt a somewhat critical assessment of the role of Sultan Muhammad. Muhammad has been praised by Muslim chroniclers who see him as a strong ruler who brought order and unity to a decaying state, torn apart by internal strife.63 Whilst fortune favoured him by the absence of any strong rival claimants to the throne, his efforts to establish his authority over his amirs, especially İl-­Ghazi and Tughtegin, were remarkably unsuccessful. Fear of reprisals by Muhammad may have dogged the steps of these two rebels but it did not generally deter them from continuing to pursue their own independent policies. Once Mawdud had been killed, and two successive armies from Mosul had been defeated, one by İl-­Ghazi himself and one by the Franks at Danith, Sultan Muhammad’s hold on affairs in Syria and Jazira became seriously enfeebled. His death provided the opportunity for even more independent actions by the amirs of the Jazira and Syria.64 Although his turbulent relationship with Muhammad looms large in İl-­Ghazi’s career from the fall of Jerusalem onwards, that career may also be viewed as a continual effort by İl-­Ghazi to establish himself in the area of Diyar Bakr and to extend his sphere of influence in the surrounding area. The process by which he acquired territory must have been arduous and complex. The difficulty of holding onto any city or citadel in a land where the central government barely controlled the contending local warlords was a perennial one in the medieval period. But at this time it was exacerbated by the presence in Diyar Bakr of the nomadic or semi-­nomadic Turcoman groups from whose ranks İl-­Ghazi had come. The two major exigencies of

T he C a re er o f Na j m a l-D in il - Gh a z i | 13 nomadic existence had to be satisfied: one was grazing lands for their flocks, the other was booty. The inherent mobility of nomadic life, and the raids necessary to make a livelihood, explain on the one hand the vast geographical area, extending from Jerusalem to Tiflis, from Aleppo to Isfahan, which was covered by İl-­Ghazi and his Turcomans, and on the other highlight some of the difficulties they must have experienced in becoming entrenched in one specific geographical area. The area which İl-­Ghazi had selected for his territory was Diyar Bakr. He had probably visited it with his father, Artuq, who took part in Ibn Jahir’s campaign to seize Mayyafariqin in the name of Malikshah, stamp out the Marwanids and take their treasures.65 The territory of Diyar Bakr was eminently suitable for İl-­Ghazi’s purposes. His brother had possession of Hisn Kayfa and Mardin and other members of the Artuqid family were there too.66 Diyar Bakr had good grazing lands as well as a chain of citadels which had existed since early Islamic times to man the frontier against Byzantium.67 Moreover, the area was conveniently far from the heartland of Seljuq power and this made it fairly safe from interference on the part of the central government. The interplay between the settled ­areas – ­the cities of Mardin, Mayyafariqin and other smaller Artuqid ­possessions – ­and İl-­Ghazi’s Turcomans must have been violent and disruptive, although the custom of the times demanded that a city should be protected by its citadel, held by some armed force. The sources abound with examples of cities such as Tiflis,68 Aleppo69 and Mayyafariqin desperately seeking military protection from various Turcoman and other leaders, since to be defenceless was even more of a disaster than to be ruled despotically.70 Ibn al-­Azraq is an unreliable source for any assessment of the effect of İl-­Ghazi and his Turcomans on the cities he owned. His testimony is blatantly pro-­Artuqid, since he worked for İl-­Ghazi’s son, Temürtash, who incidentally receives far more eulogies than his father. The changes which Ibn al-­Azraq claims that İl-­Ghazi made are therefore intrinsically suspect, but they are nonetheless revealing. İl-­Ghazi is said to have abolished billeting in the houses of Mayyafariqin, most of which was in ruins, and to have made the countryside safe from robbers who plundered the caravans in the immediate area of the town.71 There is, however, no reason to assume that in

14  |  the m ed ie va l tur k s the years 502–16/1108–22, when İl-­Ghazi’s centre of operations was Diyar Bakr, there was any improvement in the uneasy relationship between his Turcoman troops and the local inhabitants. The detailed description given by Ibn al-­Athir of one turbulent incident at Baghdad during İl-­Ghazi’s time as shihna72 is probably a more accurate picture of the impact of the Turcomans on Mardin, Mayyafariqin and Aleppo than that implied by Ibn al-­Azraq’s bland eulogies. The real beginnings of a fruitful relationship between the Turcomans and the people of the cities ruled by the Artuqid family must have been made in the reign of Temürtash, who held on to Mayyafariqin and Mardin for thirty years. Genuine sedentarisation of course took very much longer than a single generation.73 After acquiring Mardin around 502/1108–9, İl-­Ghazi must have wielded power unofficially in Mayyafariqin too74 and his grip on the city must have tightened after the death of its overlord, Suqman al-­Qutbi, in 504/1110–11. The official date given by Ibn al-­Azraq for his acquisition of the city is 512/1118–19.75 The vulnerability of the hinterland of this city after Suqman al-­Qutbi’s death was soon revealed, as ‘despotic hands reached out’ to snatch whatever territory they could.76 Prominent amongst the claimants for this temporarily ungoverned territory was İl-­Ghazi himself, who by that time had probably established himself as the overlord of the other amirs in that area. These included Toghan Arslan of Arzan and Bitlis, who answered his call to arms on the Tiflis campaign,77 and Amir Ibrahim, the lord of Amid.78 The more recent rulers of Mayyafariqin had been local amirs and it must have seemed natural for the pattern to continue. Thus in 498/1104–5 Qilij Arslan of Malatya had entered Mayyafariqin and received oaths of allegiance from the local amirs already established in the area.79 After Qilij Arslan had drowned in the Khabur in 500/1105–6 Suqman al-­Qutbi had assumed overall control of Diyar Bakr from his centre at Akhlat.80 With his death the most likely choice was İl-­Ghazi. İl-­Ghazi’s centre of power was clearly Mardin. The sources refer to him constantly as lord of Mardin and it was a key city for him to hold in view of his predominant political interests in Diyar Bakr and the Jazira. The area of Diyar Bakr seems to have held considerable appeal to Artuq and his sons and Mardin, in particular, became the principal base for İl-­Ghazi, the place to which he always returned, no matter how far afield he campaigned. Suqman,

T he C a re er o f Na j m a l-D in il - Gh a z i | 15 the brother of İl-­Ghazi, had already in 490/1096–7 specified the future Artuqid connection with the area. Declaring his intention to march first on Diyar Bakr and take it from the rebels who had seized it, he announced: ‘Once strengthened in that country, I will leave my family there.’81 Mardin was the key to the continuance of Artuqid power after his death. If he had chosen Aleppo instead, his family name would probably have sunk without trace. İl-­Ghazi at Aleppo İl-­Ghazi showed an unfortunate lack of judgement in his two major attempts to secure territory outside Diyar Bakr. His first blunder was his involvement in the affairs of the city of Aleppo. After the Crusaders had gained possession of Tripoli in 502/1108–9, a rough equilibrium had been established in northern Syria. This uneasy balance was liable to be disturbed at any moment if Aleppo acquired a new ruler, Muslim or Frank. This city therefore became the focal point for the ultimate hegemony in northern Syria. At first, İl-­Ghazi was only one of many chiefs interested in possessing Aleppo but by a series of complicated events, described in detail in Ibn al-‘Adim,82 he became master of the city in 511/1117–18. According to Ibn al-‘Adim, the townspeople sent for his help only very reluctantly, as a desperate measure.83 Aleppo was a vital strategic point but, as indicated above, it was dangerous to own, especially as the Franks, now masters of Antioch and Edessa, became increasingly aggressive; and its economy was in a lamentable state. Although the other cities owned by İl-­Ghazi – ­Mardin and ­Mayyafariqin – ­had been subjected to constant changes of government and their surrounding countryside had been ravaged,84 the plight of Aleppo seems to have been especially pitiful. It had fallen prey to the depredations of the Turcomans some thirty or forty years earlier and had been continually harassed thereafter.85 According to Ibn al-‘Adim, İl-­Ghazi was not the Aleppans’ first choice of protector, though they desperately needed someone to guarantee the security of the city.86 Nor was his first attempt at gaining possession of the city successful.87 He departed, leaving his son Temürtash behind, according to some sources as a hostage.88 On his second entry into Aleppo, however, İl-­Ghazi gained access to the citadel89 and managed to make himself acceptable as the new ruler of the city.

16  |  the m ed ie va l tur k s Reference has already been made to the numerous vicissitudes experienced by the once prosperous city of Aleppo in the previous two decades. A succession of r­ulers – ­Tutush, Ridwan, Alp Arslan al-­Akhras, Sultan-­Shah (the two latter controlled by the eunuch Lu’lu’) had taken the city ever further from centralised control by the Seljuq government and by gross misrule had brought it to a lamentable state. Thus the Aleppo which greeted İl-­Ghazi and his Turcomans was not destined to exert a permanent appeal to them. Above all, İl-­Ghazi found the treasury empty.90 There was nothing with which to satisfy his Turcoman supporters who had come with him from Mardin. He stayed long enough to put the affairs of the city in order, placed his son Temürtash in charge as his deputy there and hastened back to Mardin.91 In the period between his taking of Aleppo and his death in 516/1122, İl-­Ghazi seems to have visited the city regularly but never to have held it in particular esteem.92 His choice of Mardin rather than Aleppo as the centre of his power was astute. It would have been understandable had he opted for Aleppo, since it was a large and prestigious city, far more significant than Mardin on all counts. But it was much nearer the Franks’ sphere of operations, and the Isma‘ilis were very strong there. These factors would quickly have made his tenure of the city precarious and would have required a greater commitment on his part. As a result his descendants would probably not have entrenched themselves so firmly in their chosen area. İl-­Ghazi, after the revolt of his son Sulayman, whom he had left as his deputy in Aleppo during the Tiflis campaign, seems to have made in 515/1121–2 a final bid for strong rule in Aleppo. He allied himself with Ridwan’s family by marrying his daughter and removed his own son from the role of deputy there, appointing his nephew Sulayman b. ‘Abd al-­Jabbar in his stead. He tried to give himself and the city a respite by concluding a truce with the Franks for a whole year. But it was too late for the Artuqids to keep a firm foothold in Aleppo. Temürtash, who was to suceed his father at Aleppo, lacked the grandiose territorial ambitions of his father and was indifferent to the fate of the city after İl-­Ghazi’s death. This indifference culminated in his probably deliberate move of letting the city slip from his hands. This was in fact a logical extension of the policy of İl-­Ghazi, who considered the city unsuitable to serve as his base. The account which relates how on the way to

T he C a re er o f Na j m a l-D in il - Gh a z i | 17 Tiflis İl-­Ghazi offered Aleppo to Dubays, his son-­in-­law, on condition that he helped him on their return in a plan which İl-­Ghazi had to take Antioch, is an indication of the little value which İl-­Ghazi placed on his possession of the city.93 Becoming involved in the affairs of Aleppo was exhausting and unprofitable. Answering the appeals for help from the people of Tiflis,94 however, was an egregious error. Sultan Tughril, who ruled in Arran,95 and who was therefore the nearest person to whom the people of Tiflis might turn in trouble, had been unsuccessful in curbing the incursions of King David the Restorer into the area. İl-­Ghazi, whose reputation was at its height at the time of the request for assistance from Tiflis, must have been flattered by this appeal. Besides, Georgia was a frontier area, well suited for someone who w ­ anted – ­as İl-­Ghazi had consistently ­done – ­to rule independently of the central power. He had the requisite military strength, but an analysis on the exhausting schedule of activities which preceded his departure for Georgia reveals that he was foolhardy to undertake such a long journey into unfamiliar terrain. He had campaigned early in his career into parts of Armenia,96 but it is very unlikely that he had penetrated Georgia itself. The resounding defeat which David and his son Dimitri inflicted on İl-­Ghazi and his army (the battle is analysed in detail below) was a bitter humiliation from which İl-­Ghazi’s pride did not recover. Most of his men were killed or taken prisoner and he escaped with Dubays and a small band of followers to Mardin.97 Ibn al-­Azraq’s account of the Tiflis campaign is the most detailed one by a Muslim author. Ibn al-­Qalanisi attempts to soften this defeat by the assertion that the Muslim troops were victorius initially but were then routed by the Georgians.98 Other writers have based their accounts on Ibn al-­Azraq.99 In fact, the description of the battle between David and İl-­Ghazi is one of the few sections of the Ta’rikh Mayyafariqin wa-Amid where greater credence may be placed on Ibn al-­Azraq than on other more reliable historians of the time. It was typical of İl-­Ghazi’s uncoordinated, almost frenetic, military activity in the last years of his life that he should have allowed himself to be tempted to undertake the journey to Tiflis. Instead of acquiring the abundant booty and other material gain which no doubt to a great extent motivated the campaign, İl-­Ghazi sustained considerable losses of men and possessions.

18  |  the m ed ie va l tur k s Whether İl-­Ghazi seriously contemplated a government of his own in Tiflis is difficult to say. The Tiflis episode may well have been regarded as a mere military operation, with the lure of booty for his men and the chance of another glorious victory for İl-­Ghazi, whose swollen pride is mentioned by the hostitle Christian sources.100 This same sentiment may well have inspired İl-­Ghazi briefly with a vision of his rule over a principality extending from Aleppo to Tiflis. If so, the dream was soon shattered. İl-­Ghazi’s Military Ability His Relationship with the Turcomans İl-­Ghazi depended for his military strength on his bands of Turcomans who, according to the sources, were based in the area around Mardin.101 İl-­Ghazi would embark on a campaign in the spring and would return with his Turcomans in the autumn to Mardin. The Turcomans were known to be fierce fighters and presented a spectacle which excited and fascinated the onlooker. Ibn al-­Athir describes them thus: Each of them would come with a bag containing flour and mutton.102

Ibn al-­Qalanisi likens them to birds and beasts of prey: He (Tughtegin) found the Turkmens already assembled thither from every quarter and every direction in vast numbers and manifest strength, as lions seeking their prey and gerfalcons hovering over their victims.103

While they were renowned for their undoubted fighting powers, the Turcomans were also known to be unruly and wayward. They were motivated only by lust for plunder.104 Any leader who depended on Turcoman troops therefore had to have the funds to pay them promptly; at the very least he had to be assured of ample booty in the campaign undertaken. İl-­Ghazi’s relationship with his Turcomans moulded his career, to his advantage and detriment alike. The chronicles agree that he possessed an unusual power over them and that he inspired loyalty and devotion from them. Michael the Syrian says that the Turcomans ‘were very docile to İl-­ Ghazi’.105 Already in 499/1105–6 when İl-­Ghazi was shihna of Baghdad, Ibn

T he C a re er o f Na j m a l-D in il - Gh a z i | 19 al-­Athir describes him as strengthened by the large number of Turcomans who had joined his service.106 In Ramadan of that year İl-­Ghazi, accompanied by 10,000 horsemen, went with Malik Ridwan of Aleppo on his campaign to take Nasibin. Through the machinations of Chökermish, Ridwan turned against İl-­Ghazi and placed him in chains. When the news of the capture of İl-­Ghazi reached his Turcomans they were so enraged at the treatment meted out to their leader that they attacked Ridwan’s men, plundering and pillaging livestock and other possessions.107 The behaviour of the Turcoman troops after a victory in battle could be disciplined and restrained under İl-­Ghazi’s command. There was not always a wild rush for booty. After the victory over Roger of Antioch at Balat, for example, the Turcoman chiefs came bearing their spoils to İl-­Ghazi, who had taken over Roger’s tent. He allowed the chiefs to retain what booty they had taken, reserving for himself only a few items, to present as gifts to other rulers.108 If such calculated generosity was typical of his dealings with his Turcomans, their loyalty to him would not be surprising. On the other hand, İl-­Ghazi’s operations were also seriously curtailed by his use of Turcoman troops. Since the Turcomans could be persuaded to participate in a campaign only if there was the prospect of booty, İl-­Ghazi was unable to consolidate his gains in a given area. Once a battle was finished, the Turcomans were unwilling to linger. They wished to return to Mardin straight after the battle with their spoils. According to Ibn al-­Athir, İl-­Ghazi never embarked on long campaigns against the Franks. Every hour mattered, since the Turcomans were anxious to return as quickly as possible. In fact, if the campaign became prolonged, they would disperse as he had no money to give them.109 It is clear from such comments, and from the evidence implicit in the sources, that even on campaign the aims of İl-­Ghazi and his Turcomans were not always very compatible. But in times of peace the Turcomans were apt to be a much greater obstacle. They would never be able to share in the ambition of İl-­Ghazi (or any other like-­minded amir) to establish a permanent principality somewhere. Settling in a given area would mean that the interests of its inhabitants would, economically speaking, coincide broadly with those of their ruler. They could not be plundered like the enemy. In time the Turcomans would therefore have to travel increasingly

20  |  the m ed ie va l tur k s further afield in search of booty, for after a comparatively short period they would systematically have laid waste the border areas. Alternatively, they would have had to settle on the land, a process which might take generations to be accomplished. The Turcoman response to the empty treasury at Aleppo exercised a great influence on İl-­Ghazi’s attitude to that city and made him leave for Mardin, earlier perhaps than he had intended.110 Furthermore, the blame for İl-­Ghazi’s failure to take Antioch after his victory at Balat may be attributed at least in part to the Turcomans. They refused to stay on after the battle and thus rendered İl-­Ghazi incapable of moving on to Antioch, which as Ibn al-­Qalanisi stresses lay defenceless after the death of Roger at Balat.111 Their short-­term tactics thus foiled any possible long-­term strategy which İl-­Ghazi may have planned. Despite İl-­Ghazi’s much-­vaunted power over his Turcoman bands and their crucial role in his one major victory at Balat, there are signs that a rift was developing between him and them. As early as the period before Balat, when Tughtegin and İl-­Ghazi were planning joint operations against the Franks, they discussed the unreliability of the Turcomans and they both went personally to Mardin to attempt to persuade the Turcomans to take part in the forthcoming campaign: It was agreed between them that the amir Najm al-­Din Il-­Ghazi b. Ortuq, for the better executing of his undertaking, should proceed to Mardin in order to assemble the Turkmens from his provinces, and urge them to destroy the factions of infidelity and error. It was considered desirable that the amir Zahir al-­Din [Tughtegin] should go with him to reinforce their purpose and facilitate the realization of their hopes.112

A widening gulf between the Seljuqs of Iran and the Turcomans had long been perceptible and had led to friction and misunderstanding. Such a gulf was inevitable as the Seljuqs assumed the trappings of Islamic rulers in the Persian style, thus divorcing themselves from the Turcomans on whose military support they had depended and to whose ranks they had initially belonged. The same development may well have occurred gradually with İl-­ Ghazi and his Turcomans, as their interests diverged and İl-­Ghazi developed a taste for wider political power.

T he C a re er o f Na j m a l-D in il - Gh a z i | 21 Such a development is impossible to prove. But certain evidence which dates from İl-­Ghazi’s involvement in Aleppan affairs deserves discussion at greater length in this context. As has already been mentioned, the Turcomans found Aleppo wanting in funds and urged İl-­Ghazi to return home to the area around Mardin. Aleppo was not an attractive territory to the Turcomans. It had been much more the arena of war than Diyar Bakr and was still the focus for great military activity. İl-­Ghazi, on the other hand, returned to Aleppo the following year and, as Ibn al-­Qalanisi states, spent some time with Tughtegin persuading the Turcomans to come with them.113 İl-­Ghazi’s political aspirations in Aleppo, where he had been invited to come after many factions had been struggling for supremacy there, were clearly in conflict with the Turcomans’ desire to remain in Diyar Bakr. It is at this point that, according to Sivan’s interpretation, İl-­Ghazi made the first conscious use of jihad propaganda to impose his authority over his Turcoman forces. The particular ­occasion – ­before Balat in 513/1119–20114 – made such an appeal well-­timed. İl-­Ghazi is accorded an important role in Sivan’s analysis of the gradual reawakening of the concept of jihad amongst the Muslims in the early twelfth century.115 After explaining the factors which contributed to the absence of an awareness of jihad amongst the Muslims of northern Syria at the time of Frankish expansion and consolidation in the area in the period 491–512/1097–1118, Sivan analyses the beginnings of a change of attitude in the face of increasing aggression on the part of the Franks who were threatening to take Aleppo around 1118.116 Sivan asserts that in 1118, when İl-­Ghazi came back to Aleppo for the second time after the city had been under siege from the Franks, and when he finally assumed control there, he seems to have been in no way imbued with the zeal of a ‘warrior of the faith’. He had been an ally of the Franks in the past and had been enticed into Aleppo only by the promise of considerable financial rewards. The following year, with increasing pressure on Aleppo from Roger of Antioch, however, İl-­Ghazi found himself in a city in a ferment with the idea of jihad. He now realised full well the use he could make of this ‘galvanising’ idea to boost the morale of his troops in this dangerous enterprise.117 The appearance of the qadi Ibn Khashshab before the assembled Turcoman troops is seen by Sivan as the first instance of the use of an official jihad propaganda in the Muslims’ fight against the Crusaders. İl-­Ghazi made

22  |  the m ed ie va l tur k s his amirs swear an oath ‘to do their duty courageously, to fight heroically, and not to retreat, even if they had to shed all their blood for the holy war’.118 After encountering initial indifference and even contempt,119 Ibn al-­Khashshab brought ‘tears of ecstasy’ to the eyes of the Turcomans, who then went into battle and won a glorious victory over the Franks. Sivan’s evidence is certainly persuasive. It may well be that İl-­Ghazi had perceived the benefits of jihad propaganda and used it to strengthen his flagging hold over his troops.120 The events of the following year are also significant. In that year, 514/1120–1, İl-­Ghazi crossed the Euphrates with his men on a new campaign, having made the unusual prohibition that they were not to pillage during the campaign. When some of his men disobeyed, he punished them by shaving and hamstringing them. At this point, his troops disbanded and deserted him, leaving him with only a small handful of men. It was very fortunate for him that Tughtegin came in time to save him from Frankish attacks.121 The sequel showed, then, that İl-­Ghazi was ill-­advised to decree that his soldiers were not to pillage during the campaign. His motives for taking this unusual step are not clear. Perhaps he thought it would be wasteful to pillage territory that would soon be his. Or perhaps he was determined that his men would concentrate their energies only on fighting. Presumably he was confident that he could control them with the promise of battle plunder, in place of the casual pillaging of fields and livestock which was their normal practice. But what most probably motivated his rash decree was a misplaced confidence in his own abilities as a general, a confidence greatly strengthened by his recent victory at Balat. During the years 513/1119–20 and 514/1120–1, therefore, İl-­Ghazi’s control over the Turcomans seems to have been slackening. His troops were subsequently decimated in the Tiflis campaign and it may well be that if the casualty figures were as high as the chroniclers suggest122 İl-­Ghazi’s forces were irremediably depleted. He engaged in no major offensive after Tiflis in 513/1121–22, nor is there evidence that he was planning one, and he died a year later.123

T he C a re er o f Na j m a l-D in il - Gh a z i | 23 İl-­Ghazi – A ­ Military Strategist? It would be false to attribute to İl-­Ghazi a consistent, well-­planned military strategy, even in the period of his success from 502–16/1108–22. He took as allies known enemies of the Seljuq sultan Muhammad, such as Dubays. He also allied himself with rulers who were near at hand, such as Tughtegin and Roger of Antioch. But he played the same game of making and breaking alliances as the other local chieftains in the struggle for power in Northern Syria. İl-­Ghazi’s military abilities may be highlighted by a detailed analysis of his two major battles, Balat in 513/1119 and Tiflis in 515/1121. These are well-­documented. They also represent victory and defeat and thus present a certain balance. In 513/1119 Aleppo stood in great danger of being taken by the Franks after Roger of Antioch had harassed the area around the city and captured Buza‘a. The people of Aleppo summoned İl-­Ghazi who came from Mardin in Safar 513/June 1119.124 Roger of Antioch, ceding to pressure from the lords of Frankish castles in the area around ­Aleppo – ­their territories were being ravaged by İl-­Ghazi – ­decided to fight alone. He positioned himself in a valley at Balat. Despite a marked numerical superiority125 İl-­Ghazi seems to have been unsure whether to attack or not, since he was awaiting the arrival of Tughtegin with re-­inforcements. Once again the pressing demands of his Turcoman troops forced his hand and he attacked. His victory was inevitable, for Roger had placed himself in an extremely vulnerable position and did not expect İl-­Ghazi to attack. The Turcomans rushed into the valley from three sides at the same time and the Franks were powerless to resist. Roger was killed and the flower of the Norman cavalry was cut down. Ibn al-­Qalanisi writes that the Turcomans swooped down like a flight of falcons wanting to protect their nests126 and even the anonymous Christian writer of the Syriac chronicle is moved to an unusual descriptive detail, saying that the Turcomans showered arrows on the camp like clouds of hail.127 It was a victory calculated to delight the Turcomans. It had not taken long and afforded booty and prisoners. Yet İl-­Ghazi had been lucky that Roger had made a fatal blunder in sallying forth too soon and in positioning himself so unwisely. İl-­Ghazi was fortunate too that he had allowed himself, against his better judgement, to be swept along by the demands of his troops

24  |  the m ed ie va l tur k s for an early engagement with the enemy. Moreover, he had enjoyed a considerable advantage in numbers. Lastly, the battle had suited Turcoman military tactics and had not required the talents of a gifted general. It did require strategic ability, however, to plan the next move in the territorial game and İl-­Ghazi lacked the military acumen to make that move. The accession of Roger of Antioch had united the Franks in northern Syria. With his death, the Franks were in disarray and Antioch lay defenceless. It is clear that the people of Antioch were afraid that İl-­Ghazi was coming to attack. They made what preparations they could,128 but İl-­Ghazi did not follow up the great opportunity that presented itself. Instead his troops, following the time-­honoured practices of nomadic life, dispersed with their booty and according to Western sources perpetrated a full-­scale massacre of Frankish prisoners.129 The Turcomans would not contemplate further battles and thus Antioch slipped from İl-­Ghazi’s hands. But it is not even entirely clear that İl-­Ghazi himself had intended to exploit this rare opportunity. The sources say that he was unable to fight for three weeks after the battle. According to some writers, including Usama,130 the illness was caused by excessive bouts of drinking in celebration of his victory. Such a report would accord ill with the image of a great strategist, but it tallies with earlier accounts of Muslim historians stressing the drunken habits of İl-­Ghazi. At all events, by the time he fought the Franks again, in August, the outcome was indecisive; Baldwin had had time to restore the morale of the Franks and forestall any immediate threat to Antioch.131 İl-­Ghazi has incurred much opprobrium from modern historians for his failure to follow up his victory at Balat by an attack on Antioch. As Runciman expresses it: The great Artuqid campaign fizzled out. It had achieved nothing material for the Muslims, except for a few frontier-­posts and the easing of Frankish pressure on Aleppo.132

This kind of criticism is inappropriate to an analysis of İl-­Ghazi, since it seems to endow him with an overall strategy, a master plan, which extended beyond the immediate campaign with which he was concerned at any time. This was manifestly not the case with him or his contemporaries. They shifted their allegiance to meet the demands of the moment. İl-­Ghazi would have

T he C a re er o f Na j m a l-D in il - Gh a z i | 25 needed exceptional gifts of imagination, decisiveness and leadership to have been able to exploit his sudden victory to the full. Zengi and Nur al-­Din may well have been endowed with such qualities, but they did not fall to the lot of İl-­Ghazi. The irony was that this unexpected triumph of his, achieved by a fortunate concatenation of circumstances, should have been the first major Muslim victory against the Franks, who called it the Ager Sanguinis. With his Turcomans, this rebellious enemy of the sultan had achieved a feat that had been denied to more official campaigns mounted with the blessing of the sultan himself. Some lines of al-‘Azimi express the extent of İl-­Ghazi’s prestige after Balat: Say what you want, your wish will be granted. After the Creator, our reliance is on you.133

It is all the more ironic, therefore, to reflect that after Balat İl-­Ghazi’s career petered out miserably. İl-­Ghazi did not, however, entirely lose sight of a possible attack on Antioch. In 514/1120–1 he made for the city, but it was too late for him to recapture the victorious mood of the previous year. He forbade his troops to pillage and had to withdraw from the neighbourhood of Antioch, and it was at this stage, as mentioned above, that his Turcomans, disgruntled at the lack of plunder and money, left him in the lurch.134 It is conceivable that İl-­Ghazi still had his eye on Antioch during the Tiflis campaign in the following year, since he is reported by one source (as noted above) to have sworn an oath with Dubays that he would give him Aleppo if Dubays would bring money and troops to help him attack Antioch.135 When he returned from Tiflis in 515/1121, he was occupied for some time with affairs at Aleppo, where his son, Sulayman, had risen up against him.136 The Franks could, therefore, continue to consolidate their position in northern Syria unhampered by their principal opponent. One source describes İl-­Ghazi before the Tiflis campaign as ‘uplifted’.137 Certainly the immediate consequence of his victory at Balat was a complacency expressed in an orgy of festivities and this held him back from a sustained effort against Antioch. Soon the news of Balat was resounding throughout the Near East, and it brought him glory. Robes of honour came

26  |  the m ed ie va l tur k s from the caliph and verses were penned in his honour. This new-­found fame and sense of his own achievement, coupled with the recent arrival of ­Dubays – ­who brought great wealth and more t­ roops – ­and with the promise of future plunder in Georgia, must have all contributed to İl-­Ghazi’s decision to answer the appeal from Tiflis. Several accounts of the Tiflis campaign survive.138 Ibn al-­Qalanisi is anxious to soften the humiliation of the Muslim defeat. He relates that the Muslims were initially victorious but that the Georgians subsequently routed them, inflicting great losses.139 Ibn al-­Athir’s short but factual account mistakenly puts the campaign in 514/1120–1 but the protagonists are the same; Tughril, Dubays and İl-­Ghazi. Greater credence than usual may be placed on Ibn al-­Azraq’s account. He himself visited Georgia some thirty years after the battle and probably discussed it with participants from the Georgian side.140 His account reveals that in this instance, at least, a desire to produce an accurate statement of the facts outweighed his usual wish to please the Artuqids. His description of a bitter Muslim defeat tallies with Christian sources, such as the Anonymous Syriac Chronicle141 and the Georgian version.142 According to the version of Ibn al-­Azraq, copied later by Sibt b. al-­ Ghazi and his troops, accompanied by Dubays, marched in Jawzi,143 İl-­ 515/1121–2 from Mardin and advanced almost as far as Tiflis. Here an interesting parallel may be drawn with Balat. Georgia was unfamiliar terrain and İl-­Ghazi and his men camped in a valley below a mountain, tired after their very long march. Sultan Tughril, who was in Arran, had been asked to send troops too and according to Ibn al-­Azraq they had not yet arrived. Nor had reinforcements from İl-­Ghazi’s ‘vassal’, Toghan Arslan al-­Ahdab of Arzan. Just as the Turcomans had done at Balat, the Georgian troops under Dimitri, the son of King David the Restorer, swooped dowm on the Muslim troops from above and decisively defeated them. The mighty defender of Islam, the Star of Religion who only the year before had brought such glory to the Muslim cause, was now lucky to escape with Dubays and a handful of men and return in disgrace to Mardin. His men he could not save. In defeat as in victory, then, İl-­Ghazi showed little of the military panache and resourcefulness of Zengi or Saladin. His victory at Balat had restored Muslim morale for a while but it did not mark the beginning of

T he C a re er o f Na j m a l-D in il - Gh a z i | 27 a counter-­Crusade, and it was quickly balanced by an equally spectacular defeat, significantly enough at Christian hands. İl-­Ghazi lacked the qualities of a general, though he was on the whole a good leader of Turcoman nomads. Sultan Muhammad may have been influenced by his awareness of İl-­Ghazi’s rash nature and military limitations, as well as by their mutual animosity, when he systematically refused to appoint İl-­Ghazi as overall leader of his military campaigns, naming instead Aq Sunqur al-­Bursuqi and then Bursuq b. Bursuq. Subsequent events proved that his judgement was sound. İl-­Ghazi’s Political Ability İl-­Ghazi was fortunate in the period from 502–13/1108–20 in that his potential rivals were gradually eliminated. In 507/1113–14 the murder of Mawdud, the energetic governor of Mosul, probably instigated by Tughtegin, who was tired of Mawdud’s interference in Syrian affairs and jealous of his military successes, removed a skilful military commander from Sultan Muhammad, and one who might conceivably have quelled İl-­Ghazi. In the immediate are of Mardin, Qilij, to whom the amirs of Diyar Bakr had declared their allegiance in 498/1104–5,144 had been defeated in battle and drowned.145 The career of Suqman al-­Qutbi, which, as noted above, had advanced smoothly thanks to a policy of strict adherence to the sultan’s commands, reached its apogee when he took Mayyafariqin, now without an overlord, in 502/1108–9.146 He had already ruled the area around Akhlat and the local princelings thought it prudent to declare suzerainty to him.147 It is significant that Suqman took possession of Mayyafariqin in the same year as İl-­Ghazi gained Mardin, already in the hands of his family. It seems likely that both had their eyes on Mayyafariqin and that by seizing it first Suqman was underlining his dominance in Diyar Bakr. Their peaceful coexistence was virtually impossible. Fortunately for İl-­Ghazi, Suqman died soon afterwards,148 and thereby vacated the overlordship of Diyar Bakr for his rival. Ibn al-­Athir describes the raid which İl-­Ghazi led on Suqman’s men as they returned to Akhlat with their master’s coffin.149 This was not merely an attempt to obtain plunder but was a deliberate move to destroy Suqman’s troops and thereby weaken the power of the principality nearest his own. Although the attack on Suqman al-­Qutbi’s troops was unsuccessful, Suqman’s son, Ibrahim, never offered a serious threat to İl-­Ghazi’s growing power in Diyar Bakr. The

28  |  the m ed ie va l tur k s struggle for power within the Artuqid family itself, and especially between İl-­Ghazi and his brother’s son, Ibrahim b. Suqman, seems to have been resolved with the death of Ibrahim around 502/1108–9.150 For a short time Suqman had ruled Mardin and Hisn Kayfa151 jointly. Ibrahim had wished to do likewise after his father’s death in 498/1104–5, when he inherited Mardin and ruled it from Hisn Kayfa. Thereafter, when İl-­Ghazi came to Diyar Bakr, dismissed from his post as shihna and without obvious territorial possessions, there must have been a struggle for control of Mardin between him and Ibrahim, culminating in Ibrahim’s death. Ibrahim’s brother, Da’ud, then took Hisn Kayfa,152 whilst İl-­Ghazi ruled at Mardin. Thus these two cities, the cornerstones of Artuqid power, began to be ruled separately, a pattern which was to continue for a long time. After the death of Ibrahim, İl-­Ghazi wielded supremacy over the Artuqid family.153 The amir whose power most nearly approached his own was his nephew Da’ud, whose subordinate status may be inferred from the fact that he sent troops to İl-­Ghazi when he was called upon to do so.154 Da’ud seems to have maintained good relations with İl-­Ghazi and not to have interfered with his uncle’s territorial ambitions. Such deference and restraint were not, however, evident once İl-­Ghazi had died, for the two cousins, Da’ud and Temürtash, were destined to clash on several occasions in their power struggle in Diyar Bakr. Another of İl-­Ghazi’s nephews, Balak, is mentioned as his constant companion in arms in the last year of İl-­Ghazi’s life and seems to have possessed greater force of personality and energy than either the wayward, frivolous Sulayman or the more docile Temürtash. Certainly, from the viewpoint of the continuation of jihad against the Franks, Balak was İl-­Ghazi’s heir, not his sons. Balak’s qualities as a warrior were recognised by more than one medieval historian. He died, however, soon after his uncle in 518/1124. Between 502/1108–9 and 511/1117–18, İl-­Ghazi managed to entrench himself at Mardin and extend his sphere of influence over the amirs in the surrounding area. After the death of Sultan Muhammad in 511/1118 and the succession quarrels that ensued, the Jazira was at last freed from the interference of the Seljuq sultans. İl-­Ghazi now wielded genuine power in his own right. Even a hostile source such as the Anonymous Syriac Chronicle cannot deny the extent of İl-­Ghazi’s prestige and power just before his death:

T he C a re er o f Na j m a l-D in il - Gh a z i | 29 Ghazi had become powerful and was uplifted because he ruled his own land, that of his brother Suqman’s sons, and that of his uncle’s son Daud,155 as far as Assyria, Armenia, and the land of the Iberians.156

İl-­Ghazi and Administration In his long article on the Artuqids, which spans the political careers of İl-­ Ghazi, Temürtash and Najm al-­Din Alpı, Cahen discusses briefly the nature of the state which İl-­Ghazi built up in Mayyafariqin and Mardin.157 Although he emphasises that İl-­Ghazi was still a nomadic leader who spent his time in wide-­ranging expeditions, he adds that while other Turkish chiefs were mere pillagers İl-­Ghazi busied himself with administration and religious matters.158 This statement is self-­contradictory and is not supported by the evidence. The events of the last years of İl-­Ghazi’s life and the vast geographical areas he covered left him little time for administration, even if he had possessed the inclination or ability for it. The two cities for which documentary evidence is available, namely Aleppo and Mayyafariqin, seem to have been governed under İl-­Ghazi in the manner customary during the period. Local families, such as the Banu Nubata in Mayyafariqin, would be handed the task of the day-­to-­day administration of the city.159 Whether in the winter months İl-­Ghazi took a direct part in the affairs of Mardin, a city which he seems to have preferred to Mayyafariqin, is impossible to say. But it is unlikely that his interests went further than the initial removal of burdensome taxation to ingratiate himself with the inhabitants, as in Mayyafariqin,160 followed by his use of customary methods161 to raise funds for his men and for his military activities. It is hard to join in Cahen’s eulogy of İl-­Ghazi, when, basing his enthusiasm upon the naively pro-­Artuqid statements of Ibn al-­Azraq, he writes: Par sa force même, il redonne aux chemins une relative sécurité profitable au commerce, et peut alléger les impôts qui écrasaient les populations ruinées.162

The same sweeping praise is lavished on İl-­Ghazi by Sevim who labels him ‘not only a shrewd soldier but also a politician and a good administrator’.163 This is hard to reconcile with the description by Ibn al-‘Adim of events at Aleppo when İl-­Ghazi and his Turcomans arrived to take command of the

30  |  the m ed ie va l tur k s city. Rapacity was then the key motivation and it was not mitigated by the slightest interest in the well-­being of the urban populace.164 Ibn al-­Athir’s account of İl-­Ghazi at Aleppo agrees in spirit with that of Ibn al-‘Adim, whilst adding more details.165 Descriptions of a conquering military chief entering a city and removing burdensome taxation from its inhabitants abound in the sources and are couched in stereotyped phrases.166 There is no justification for attaching particular significance to the actions of İl-­Ghazi when he in turn conquered a city. It may, however, be conceded that by the time İl-­Ghazi had become a significant power in the political struggle in northern Syria around 511/1117–18 he seems to have acquired confidence and some rudimentary political skills. His alliance with Tughtegin in Damascus in 508/1114–15 and his earlier experience as a military leader for Tutush and Ridwan had given him a knowledge of Syrian politics. The support of Tughtegin proved invaluable. Alone, İl-­Ghazi might not have withstood the pressure exerted on him by Sultan Muhammad. But the relationship between İl-­Ghazi and Tughtegin was not always harmonious. Whilst, according to obviously biased sources, they had both been guilty of savagery after Balat, especially at Aleppo where the Frankish prisoners were massacred for the enjoyment of the population,167 the death of the Frankish leader, Robert, at the hands of Tughtegin himself seems to have angered even İl-­Ghazi. İl-­Ghazi had sent Robert to Tughtegin in an attempt to extort a higher ransom from him but Tughtegin rashly beheaded him with his own sword.168 In general, İl-­Ghazi’s attitude to Frankish prisoners seems to have been level-­headed. He saw his most prominent captives as a source of ransom money, especially as he was often short of the funds needed to pay his troops.169 By a fortuitous set of circumstances, İl-­Ghazi did not wait for Tughtegin to arrive with his troops before Balat. Since he attacked alone, the glory of the first major Muslim victory against the infidel fell exclusively to İl-­Ghazi. In the years between his taking Mardin and his establishment in Aleppo, İl-­Ghazi’s, relations with the Franks were charcterised by flexibility and pragmatism. Contacts on the personal level between Frankish and Muslim rulers at this time are well-­documented.170 İl-­Ghazi was impressed on occasion by individual Frankish leaders, one of whom he allowed to live because of ‘his proud bearing’.171 Since the Franks were a significant power group in north-

T he C a re er o f Na j m a l-D in il - Gh a z i | 31 ern Syria, İl-­Ghazi inevitably came into conflict with them on occasions, sometimes emerging triumphant, sometimes not. When Bursuq was sent out at the head of a large army in Dhu’l-­Hijja 508/May 1115 to take reprisals against İl-­Ghazi and Tughtegin, the two rebellious amirs were in abject fear of the sultan and did not feel strong enough to face the forthcoming attack alone. They therefore made an alliance with Roger of Antioch. Thus, when Bursuq’s army arrived, the local princes of northern Syria, Muslim and Frankish alike, joined forces against the outside aggressor. The defence of territory clearly counted for more than the defence of the faith. But it would appear even then that the two rebel amirs were reluctant to actually engage with the army of Bursuq, which fell into an ambush set by Roger alone. After Roger’s victory over Bursuq’s forces at Tall Danith on 22 Rabi‘ II, 509/14 September 1115, Syria was indeed rid of the Seljuq army. Tughtegin understood (possibly more clearly than İl-­Ghazi) the wider implications of this defeat of Bursuq and immediately went to Baghdad to seek a reconciliation with Sultan Muhammad. No doubt he feared the evergrowing power of the Franks in Syria and felt isolated. He might also have experienced a certain guilt at fighting with Christians against fellow Muslims and he can scarcely have been unaware how his behaviour might be construed in Iraq, where the Franks offered no threat. As for İl-­Ghazi, he reaped one crucial benefit from the defeat of Danith: no longer were his projects thwarted by the sultan. But he too never again allied himself with the Franks. Indeed, after the battle at Danith there were some signs of a hardening of İl-­Ghazi’s attitude towards the Franks. Various reasons may be adduced for this change. The accession of Roger of Antioch had united the Franks in northern Syria and their firm establishment in that area had been helped by increasing dissensions amongst the Muslim amirs after the deaths of Mawdud and Ridwan in 507/1113–14. By the year 511/1117–18, Roger of Antioch was harassing the area around Aleppo and was threatening the city itself. Combined with a heightened awareness of the danger of his new foe was İl-­Ghazi’s new-­found freedom from direct Seljuq interference and from opposition from within his own territory of Diyar Bakr. Sultan Muhammad died in 511/1118. Sanjar was still preoccupied with affairs further east. Mahmud, the new sultan, was embroiled with the problem of rival contenders for his throne.172 In his own territory of Diyar Bakr, İl-­Ghazi could at

32  |  the m ed ie va l tur k s last feel secure. The new sultan had asked him to take over Mayyafariqin in 512/1118–19, probably in response to a request from İl-­Ghazi. Indeed, Mahmud was scarcely in a position to refuse such a request from a senior amir whose lands were distant and whose loyalty was suspect.173 İl-­Ghazi also ruled Harran, Ra’s al-‘Ain and Nasibin.174 Armed with a new confidence and the ability to act independently, İl-­Ghazi and Tughtegin planned a joint campaign against the Franks in 512/1118–19. According to Ibn al-­Athir, İl-­Ghazi could muster 20,000 men at Mardin175 before embarking on the undertaking which culminated in Balat (513/1119–20). Although İl-­Ghazi sent a perfunctory message to the new sultan asking for help against the Franks,176 he appears to have received no answer. This request was probably nothing more than a formality in any case. It is difficult to assess the extent to which Sivan’s arguments may be accepted when he suggests that in the short period between İl-­Ghazi’s entry into Aleppo in 512/1118–19 and the battle against Roger of Antioch in 513/1119 İl-­Ghazi changed from someone in no way imbued with the zeal of a mujahid to a leader who was fully aware of the benefits of using jihad propaganda.177 If such a theory could be proved it would show that İl-­Ghazi possessed a certain awareness of political realities. It is perfectly possible that brief contact with pietist circles in Aleppo, where the idea of jihad was re-­awakening, could have given an extra edge to İl-­Ghazi’s attitude towards the Franks. Yet it is well known that he stayed in Aleppo for only very short periods at a time. Certainly the presence before the troops at Balat of Ibn al-­Khashshab, the qadi who had played an important role in establishing İl-­Ghazi at Aleppo and who is shown by Sivan to have made a major contribution to the re-­awakening in Aleppo of an awareness of jihad, is very significant. Further support for Sivan’s theory is provided by the oath of allegiance which İl-­Ghazi made his amirs swear before going into battle against Roger. Ibn al-­Khashshab may well have suggested that the taking of such an oath, backed up by his own eloquence to the Turcoman troops, would be a possible means of holding their loyalty a little longer. There is, however, a major flaw in Sivan’s argument. His theory would be persuasive if it could be demonstrated that İl-­Ghazi continued to use jihad propaganda in his subsequent campaigns. As noted above, after his victory at Balat, the caliph sent him robes of honour and adulatory verses were penned

T he C a re er o f Na j m a l-D in il - Gh a z i | 33 in his honour. The efficacity of the formula had been proved, İl-­Ghazi’s own reputation had been enhanced, and the Turcomans had fought like lions, after being moved to tears by the eloquence of the qadi. Yet it was patently not the case that İl-­Ghazi continued to inspire his troops with the spirit of jihad in his subsequent undertakings against the Franks. Sivan states rather lamely that it is not certain that in the less spectacular battles which ensued İl-­Ghazi continued to use these propaganda themes. The truth of the matter was that İl-­Ghazi in no sense built on his victory over Roger at Balat. He failed signally to follow up either the politico-­ military or the psychological advantages which he had gained. Instead, he basked in a haze of glory and then dissipated his energies in a series of minor military operations.178 He may deliberately have invited Ibn al-­Khashshab to speak to his troops and, like his own Turcomans, he may even have been swayed momentarily by an atmosphere of religious fervour; but his deliberate adoption of a jihad propaganda seems unlikely. It would have been more in character with the opportunistic Realpolitik which İl-­Ghazi and his ilk habitually practised for the adoption of jihad propaganda to have been a ploy. If so, such propaganda would have been lightly adopted for reasons of ­expedience – ­after all, it could do no ­harm – ­and as lightly discarded when it had served its turn. After their victory at Balat, the Turcomans quickly began to disperse with their booty and could not be persuaded to linger for any motive, material or religious. Presumably after so devastating a victory, with its correspondingly rich booty, it would have been obvious that no further major opportunities for plunder would present themselves. The ardour of the Turcomans to engage in battle with Roger of Antioch is not very likely to have been of a religious nature and could more realistically be attributed to their desire to acquire more booty for themselves rather than await the arrival of Tughtegin with his troops who would be as rapacious as they were. In fact, İl-­Ghazi was very reluctant to begin fighting until his ally came but was pushed into action by the urgent demands of his Turcoman amirs.179 In spite of the Tiflis fiasco İl-­ Ghazi’s reputation in northern Syria remained prestigious enough, possibly because of the very remoteness of Tiflis. In his general discussion about Georgian affairs at this time, Osman Turan says that İl-­Ghazi died and could not undertake another campaign

34  |  the m ed ie va l tur k s into Georgia.180 He then mentions a letter which, according to Bar Hebraeus (and apparently only him) the Byzantine emperor wrote to İl-­Ghazi in which he informed İl-­Ghazi that a large Crusader force was coming to the coasts of Syria. He then added: Prepare to fight them. If you need it, I will help you with an army of 30,000 men.181

This anecdote, although it has an apocryphal flavour, is testimony to the prestige which İl-­Ghazi had gained at Balat. It was to him, not Tughtegin, that the Byzantine emperor wrote making this offer of military support. Relations with the Isma‘ilis İl-­Ghazi was aware of the political necessity of establishing some kind of rapport or at any rate modus vivendi with the Isma‘ilis, especially while he was master of Aleppo, a city which had been bolstered by their military strength in the area. They had in fact provided military support for Aleppo for several years and Ridwan had indeed shown them outright favour. According to Ibn al-­Furat,182 İl-­Ghazi seems to have engaged in a wily game of diplomacy from which he emerged the victor. He was approached on one occasion in Mardin by a messenger from the leading Isma‘ili in Aleppo who asked him to hand over the citadel of al-­Sharif to the Isma‘ilis. İl-­Ghazi coolly announced that he had just given orders for the demolition of that fortress and added that he would hand it over if the work of destruction had not already begun. Whilst his secretary occupied the envoy, İl-­Ghazi despatched a message by pigeon ordering the immediate demolition of the citadel of al-­Sharif and ordering that some prominent citizens of Aleppo should occupy the ground where it had previously stood. The messenger from the Isma‘ilis arrived back in Aleppo to find the building no longer standing. İl-­Ghazi later informed the Isma‘ilis that he would gladly have given it to them if they had come to him secretly and not announced their intentions publicly.183 The citadel had been an important strategic point in Aleppo and clearly İl-­Ghazi had at once realised the danger of ceding it to the Isma‘ilis.

T he C a re er o f Na j m a l-D in il - Gh a z i | 35 Relations with the Caliphate İl-­Ghazi’s relationship with the ‘Abbasid caliph was one of indifference or at times hostility. This was a characteristic which was shared by later Seljuq sultans and many of their amirs.184 Dubays b. Sadaqa, for example, who was a thorn in the side of the caliph al-­Mustarshid and had ravaged the countryside around Baghdad, sought refuge first at Qal‘at Ja‘bar with the Arab ruler185 there and then with İl-­Ghazi, who became so absorbed with his guest and accorded him such lavish hospitality that he neglected his administrative duties. Dubays, as a further mark of their mutual bond, married İl-­Ghazi’s daughter, Farkhunda Khatun.186 When al-­Mustarshid wrote to İl-­Ghazi, sending him robes of honour, he requested him at the same time to break off relations with Dubays. Whilst İl-­Ghazi agreed to participate in the jihad, he refused to sever his links with Dubays.187 Religion If actions are a clue to inner convictions it is unlikely that İl-­Ghazi’s commitment to Islam was profound. His drinking orgies must have exceeded the norms of his time for they are described with evident relish by the chroniclers.188 Whilst other military and religious leaders, including apparently the caliph al-­Rashid,189 also enjoyed this activity forbidden by Islam, they are not recorded as indulging to the same extent as İl-­Ghazi. It is not appropriate here to discuss in detail the exact nature of the form of Islam practised by the Turcomans in the early twelfth century. What is probable, however, is that their commitment to Islam was only superficial. İl-­Ghazi, as leader of the Turcomans in Diyar Bakr, was essentially flexible in his approach to the established religious groups with which he had to deal. Towards the Seljuq sultan Muhammad, the champion of Sunni orthodoxy, and towards the ‘Abbasid caliphs, he was generally hostile and uncooperative. This sprang from his desire for territorial independence, not from his championing of any heterodox creed. He actively supported enemies of both the sultan and the caliph. Yet he made gestures of reconciliation towards Sultan Mahmud, and the caliph, al-­Mustarshid, as mentioned above, sent him robes of honour to mark his victory at Balat, according him the honorific title Najm al-­Din. İl-­Ghazi is also said by Ibn Shaddad to have founded a

36  |  the m ed ie va l tur k s Sunni madrasa in Mardin.190 In this he was following a widespread practice of his times. On the other hand, İl-­Ghazi is likely to have been familiar with Shi‘ite Islam during his time in Syria, both when he was in the service of Malik Ridwan whose Shi‘ite sympathies were well known,191 and in his own rule in the latter years of his life at Aleppo. Moreover, the Isma‘ilis were powerful at this time both in northern Syria and in Diyar Bakr itself, as the slaughter of several hundred of their number at Amid in 519/1125192 and the deep influence exerted by one Isma‘ili shaykh over İl-­Ghazi’s son, Temürtash, testify.193 İl-­Ghazi is known to have established contact with the Isma‘ilis in northern Syria whilst he governed Aleppo though it is evident that he distrusted them. Personality İl-­Ghazi was, in short, an energetic military leader and at times an astute politician. He was wayward and independent, uncontrolled on occasions in his personal and political life and erratic in his judgements. Only in the later part of his life did he acquire Mardin and Mayyafariqin, but he clung to them tenaciously in difficult times and thus assured the continuation of his family in that area through the agency of his cautious and peaceable son, Temürtash. Indeed, the Artuqids maintained themselves in their area of choice for three centuries, an achievement unmatched by any other minor Turkish dynasty in Anatolia. İl-­Ghazi’s own career was characterised by tremendous, indefatigable military endeavours over far too wide a geographical area, which extended from Tiflis to Aleppo and from Hulwan to Mardin. He did not possess the political maturity to concentrate his efforts in a smaller area, but the lesson of his life was not lost on his less ambitious son Temürtash, who consolidated his hold on the heartland of Artuqid power and deliberately rejected the unreal aspirations of his father.194 Conclusions Against the background of the disintegration of the Seljuq empire after the death of Malikshah, İl-­Ghazi, as chief of a large group of Turcomans in the Jazira, succeeded in establishing himself in the area of Mardin and Mayyafariqin and paving the way for his successors to remain there.

T he C a re er o f Na j m a l-D in il - Gh a z i | 37 The process of his emancipation from Seljuq control was gradual. After serving the minor Seljuq princes of Syria, he seized Mardin, the key to his future successes. By alliances with the enemies of Sultan Muhammad, above all Tughtegin, İl-­Ghazi managed to ruin the attempts of the sultan to impose his authority on the western reaches of Seljuq territory. İl-­Ghazi’s political approach was pragmatic and flexible, motivated by territorial and personal ambition. The basis of his power was his Turcoman forces but by the end of his long career there are signs that a rift was developing between them. Until the last few years of his life, İl-­Ghazi’s political fate was linked with that of the Jazira and orientated eastwards towards the centre of Seljuq power. He did not look westwards to Byzantium nor were his efforts directed primarily against the Franks. Only after the death of Sultan Muhammad, when he began to enjoy true independence and was freed from the spectre of Muhammad’s retaliation, did İl-­Ghazi attempt positively to curb Frankish power in the area of Aleppo. In the wider sweep of Islamic history, İl-­Ghazi played a significant role in the slow undermining of the central Seljuq authority and in the gradual establishment of independent emirates in the Jazira and northern Syria. His victory at Balat was an early indication of the source of the future Muslim Counter-­Crusade against the Franks. That revanche was destined to come not from the efforts of Seljuq sultans far away in western Iran but from small dynasties established locally at Aleppo and other cities in Syria and at Mosul. Yet it is ironic that before Balat İl-­Ghazi’s actions positively helped the Franks. The attention of Sultan Muhammad was often occupied in trying to curb his rebellious stance, whilst in the meantime the Franks were able to consolidate their position and expand their territories in Northern Syria. The armies from Mosul, even under the command of the energetic Mawdud, directed their efforts towards the subjugation of İl-­Ghazi and Tughtegin instead of attacking the Franks. Yet even if İl-­Ghazi had joined the armies sent from Mosul in campaigns against the infidels, it is unlikely that the Muslims would have achieved the successes of Nur al-­Din and Saladin. The religious climate of the early decades of the twelfth century was only beginning to betray a consciousness that the Franks might be different from other factions fighting for hegemony in Syria. İl-­Ghazi was too inconsistent and

38  |  the m ed ie va l tur k s unpredictable, too bound to his semi-­nomadic heritage, to have achieved the successes of Zengi, whilst he lacked the religious motivation shown by Nur al-­Din in his later years. Notes 1. Cf. W. B. Stevenson, The Crusaders in the East (repr. Beirut, 1968), 97–109; R. Grousset, Histoire des Croisades et du Royaume Franc de Jérusalem I (Paris 1934), 548–83; and R. Röhricht, Geschichte des Königreichs Jerusalem (Innsbruck, 1898), 106–8, 115–17, 134–6, 138–43, 150–5. İl-­Ghazi is also discussed in K. M. Setton and M. W. Baldwin, A History of the Crusades I (Madison, Milwaukee and London, 1969), 114–16, 171–4, 403–5, 412–8, 450–2. 2. Cf. S. Runciman, A History of the Crusades II (Cambridge, 1954), 107–8, 111–12, 121–3, 128–9, 131–4, 147–55, 158–62. 3. Cf. especially N. Elisséeff, Nur ad-Din, un grand prince musulman de Syrie au temps des Croisades I–III (Damascus 1967), and H. A. R. Gibb, ‘The Career of Nur ad-­Din’, in Setton and Baldwin, op. cit., 513–27. 4. Cf. especially S. Lane-­Poole, Saladin and the Fall of the Kingdom of Jerusalem (London and New York, 1898); H. A. R. Gibb, ‘The rise of Saladin, 1169–89’, in Setton and Baldwin, op. cit., 563–89; idem, Saladin (Oxford, 1971); A. S. Ehrenkreutz, Saladin (Albany, New York, 1972). 5. C. Cahen, ‘Le Diyar Bakr au temps des premiers Urtukides’, Journal Asiatique CCXXVII (1935), 227–40. These pages deal specifically with İl-­Ghazi’s political career. Cf. also M. F. Köprülü’s ‘critique’ of the above article in Belleten, Cilt 1, sayi 1 (1937) 283–8. Köprülü raises few major objections to Cahen’s article, contenting himself with criticisms of Cahen’s renderings of Turkish names. 6. A. Sevim, ‘Artuklarin Soyu ve Artuk Bey’in Siyasi Faaliyetleri’, Belleten, Cilt xxvi, sayi 101 (1962), 121–47; ‘Artukog˘lu Sökmen’in Siyasi Faaliyetleri’, ibid., 103, 501–20; ‘Artuk Og˘lu İlgazi’, ibid., 104, 649–91. 7. O. Turan, Dog˘lu Anadolu Türk Devletleri Tarihi (Istanbul, 1973), 145–51. 8. C. Cahen, art: Artukids, Encyclopaedia of Islam, 2nd edn, 667; E. Honigmann. art: Ortokids, EI1; H. Süssheim, art: Ilghazi, EI1. 9. M. F. Köprülü, art: Artuk Og˘ullarï, Islam Ansiklopedisi, 617–25. 10. H. A. R. Gibb, ‘Notes on the Arabic materials for the history of the early Crusades’, Bulletin of the School of Oriental Studies VII (1933–45), 739.

T he C a re er o f Na j m a l-D in il - Gh a z i | 39 11. Cf. the bibliography in Runciman (op. cit., 480–5) and in the individual chapters on the First Crusade in Setton and Baldwin (op. cit., 135–76, 368–403, 410–48). 12. Op. cit., 136. 13. Op. cit., 560. 14. Op. cit., 155. 15. Cf. especially the analysis of the Muslim attitude to the Franks in E. Sivan, L’Islam et la Croisade, Idéologie et Propagande dans les Réactions Musulmanes aux Croisades (Paris, 1968), 24–35. 16. For an analysis of Seljuq history after the death of Malikshah, cf. C. E. Bosworth, ‘The political and dynastic history of the Iranian world’, in J. A. Boyle (ed.), The Cambridge History of Iran V (Cambridge, 1968), especially 102–18, and M. F. Sanaullah, The Decline of the Seljuqid Empire (Calcutta, 1938). 17. For example, the campaign conducted by Ibn Jahir against the Marwanids in Diyar Bakr in 477/1084–5. The purpose of this campaign was to quell the Marwanids and secure Diyar Bakr for the Seljuqs (Ibn al-­Athir, X, 86–8, 93–4; Ibn al-­Azraq, ed. ‘Awad, 208–12; Ibn al-­Athir, Atab., 5). 18. Ibn al-­Azraq, ed. ‘Awad, 210. 19. Tutush gave Jerusalem to Artuq as an iqta‘ in 478/1085–6 (Ibn al-­Athir, Atab., 7; Ibn Khallikan, III, 171–2). 20. Ibn al-­Athir, X, 268; Ibn al-­Qalanisi, 147; Matthew of Edessa, 256. 21. Before he died in 488/1095, Tutush asked his supporters to pledge their allegiance to his son Ridwan. Amongst the amirs who were in Tutush’s service was İl-­Ghazi whom Tutush had left to fight alongside his son Ridwan (Ibn al-­Athir, X, 167). 22. It would appear that at one stage İl-­Ghazi was governor in Jerusalem, and not Suqman, since after Duqaq had detained İl-­Ghazi in Damascus, as a move in his struggle against his brother Ridwan, Suqman went to take over from his brother’s deputies in Jerusalem (Ibn al-‘Adim, Bughyat, 139; idem, Zubda, 124). On the other hand, Ibn al-­Athir says that Jerusalem had been given to Suqman by Tutush (X, 193). 23. It is hard to share the enthusiasm for Artuqid rule expressed by Ibn Khallikan in his biography of the Fatimid caliph, al-­Musta‘li. Describing the capture of Jerusalem by al-­Afdal, he writes that al-­Afdal ‘then appointed a governor to rule it in his name, but this person, being unable to resist the Franks, yielded the city up to them; had it been in the possession of the Ortuk family, it would have been better for the Muslim people’ (I, 160).

40  |  the m ed ie va l tur k s 24. Ibn al-­Athir, X, 193; Ibn al-­Azraq. Ms. A, f. 160b. 25. Ibn Khallikan, III, 171. 26. Ibn al-­Athir, X, 210. 27. Ibid., 225; Ibn Khallikan, III, 171–2. 28. For an analysis of the role of the shihna, see C. L. Klausner, The Seljuk Vezirate, A Study of Civil Administration, 1055–1194 (Cambridge, MA, 1973), 20, 28 and 123. 29. Ibn al-­Athir, X, 231–2. The caliph sent the chief qadi and a professor from the Nizamiyya madrasa to intercede (ibid.). 30. Ibn al-­Qalanisi, 147; Sibt b. al-­Jawzi, ed. Jewett, 8; Ibn al-­Athir, X, 253. 31. Ibn al-­Athir, X, 254. 32. Ibid. 33. Gibb has shown that Ibn al-­Athir frequently changes the dates given by Ibn al-­Qalanisi ‘and always does so wrongly’ (op. cit., 747). 34. Sadaqa had originally sided with Berk-­Yaruq. He changed his allegiance to Muhammad after Barkyaruq’s vizier, Abu’l-­Mahasin al-­Dihistani, demanded a large sum of money from him in 494/1100–1 and threatened to use force against him (Ibn al-­Athir, X, 209). 35. Ibid., 244–6. 36. Ibid., 262; al-­Bundari, 90. 37. Ibn al-­Athir, X, 262. 38. al-­Bundari, 80; Ibn al-­Athir, X, 265. 39. There seems to be no record of a quarrel between Muhammad and İl-­Ghazi although Süssheim justifiably infers that one did take place (EI1, art: İl-­Ghazi). The chronology of İl-­Ghazi’s departure from Baghdad and the death of his brother Suqman is not clear. Perhaps he left Baghdad of his own accord, out of fear of possible reprisals from Muhammad and with the aim of taking over his brother’s lands. Alternatively, Muhammad may have sent him away on his own arrival in Baghdad. 40. Ibn al-­Athir, X, 268; Ibn al-­Adim, Bughyat, 204 (he gives no date.) 41. Ibn al-­Athir, X, 268; Ibn Khallikan, III, 171–2; Ibn al-­Qalanisi, 147. 42. Cf. Sanaullah, op. cit., 114. Whether Muhammad was stronger than his brother, as the sources imply, is discussed later in this article. 43. Ibn al-­Athir, X, 308. 44. Idem, X, 319. 45. Gibb thinks that Grousset was wrong to regard every offensive against the Franks as a ‘counter-­Crusade’ (‘Notes on the Arabic materials for the history

T he C a re er o f Na j m a l-D in il - Gh a z i | 41 of the early Crusades’, BSOS VII (1933–5), 741). Gibb regards as misleading the idea that the expeditions sent out under the governors of Mosul were the result of Seljuq intervention. The one exception to this was the expedition of Bursuq in 509/1115–16 which was directed against the Muslim princes, not the Franks (ibid., 742). 46. Ibn al-­Athir, X, 321. This retreat by İl-­Ghazi was not prompted by loyalty to Muhammad, as Turan implies when he says that İl-­Ghazi, in spite of the fact that Muhammad had dismissed him from the post of shihna of Baghdad, would not help and returned to Mardin (op. cit., 146). Turan also confuses Chökermish (who died in 500/1106–7) and Chavlı. Neither shortly before, nor after, this incident did İl-­Ghazi reveal himself to be loyal to Muhammad. His refusal to co-­operate with Chavlı was probably caused by an awareness of his own lack of strength at that point and perhaps by a feeling that he might still be restored to Muhammad’s favour eventually. He may also have had doubts about the particular suitability of Chavlı as an ally. 47. Ibn al-‘Adim, Zubda, 154–5; Ibn al-­Qalanisi, 169–70. According to Ibn al-­ Qalanisi, İl-­Ghazi and his Turcomans joined up with the troops of Mawdud and Suqman al-­Qutbi at Jazira Bani Numayr. During the campaign İl-­Ghazi and Suqman quarrelled; İl-­Ghazi went off to Mardin whilst Suqman seized Balak, İl-­Ghazi’s nephew, and took him back to his own city (Akhlat) in chains. This campaign is not mentioned by Ibn al-­Athir. 48. Ibn al-­Qalanisi, 170. Cahen’s article on the Artuqids (EI 2) is misleading in its analysis of İl-­Ghazi’s relationship with Sultan Muhammad. Cahen dates the cooling of relations between them to the rupture in mid-­campaign between İl-­Ghazi and Suqman al-­Qutbi. A more likely moment for the beginning of animosity between Muhammad and İl-­Ghazi would be in 498/1104–5 when Muhammad dismissed İl-­Ghazi from the office of shihna of Baghdad after he had shifted his allegiance to his brother Barkyaruq. 49. For an account of this campaign, see Ibn al-­Qalanisi, who gives the date as 504/1110–11 (op. cit., 174–5) and Ibn al-­Athir, who has two versions (X, 340–1 and Atab., 18). In the Kamil he records the episode under the year 505/1111–12. It was on this campaign that Suqman al-­Qutbi died. According to Ibn al-­Athir, İl-­Ghazi made a surprise attack on his men who were taking his body back to Akhlat (ibid.; see also Ibn al-­Furat, I, f. 54b.) 50. Ibn al-­Qalanisi, 187; Ibn al-­Athir, X, 347–8; Michael the Syrian, 216; Matthew of Edessa, 285–6; al-‘Azimi, 382; Anon. Syr. Chron., 85. According to Fink, Mawdud of Mosul was motivated by religious idealism in his offensives against

42  |  the m ed ie va l tur k s the Franks (H. S. Fink, ‘Mawdud of Mosul, precursor of Saladin’, The Muslim World XLIII (1953), 18–27). Sivan refutes his claims, saying that Fink does not give convincing evidence that Mawdud attempted to give his campaign a religious character (op. cit., 56, n. 8). 51. Ibn al-­Athir, X, 350–1; Sibt b. al-­Jawzi, 52. 52. Ibn al-­Athir, X, 351. 53. Ibn al-­Athir, X, 352; Michael the Syrian, 216–17; Matthew of Edessa, 287; Ibn al-­Azraq, ed. ‘Awad, 284. 54. Michael the Syrian, 217; Ibn al-­Athir, X, 352. 55. Ibn al-­Athir, X, 352; Matthew of Edessa, 285. Sibt b. al-­Jawzi mentions the rumour but rejects it as untrue (op. cit., 51). Ibn al-­Qalanisi, as the Damascus chronicler and from a viewpoint of warm enthusiasm for Tughtegin, places the blame squarely on the Isma‘ilis and emphasises Tughtegin’s profound grief at Mawdud’s death (op. cit., 187). 56. Ibn al-­Athir, X, 352; Matthew of Edessa, 292. 57. Ibn al-­Athir, X, 356–7. 58. Matthew of Edessa, 292; Usama, tr. Hitti, 102–6. 59. Ibn al-­Qalanisi, 193; Sibt b. al-­Jawzi, 55–6; Ibn al-­Athir, X, 360. 60. Runciman mentions only Tughtegin’s reconciliation with Sultan Muhammad (op. cit., 133). Stevenson, on the other hand, states that ‘Ilgazi and Tugtakin both effected their reconciliation with the sultan’ (op. cit., 100). There would appear to be no evidence of this rapprochement in the chronicles. Indeed, it would have been more consistent if İl-­Ghazi had made no move towards the sultan. 61. Ibn al-­Athir, X, 418. 62. Ibn al-­Qalanisi, 168; M. F. Brosset, Histoire de la Géorgie (St Petersburg, 1858), I, 358–61; Ibn al-­Azraq, Ms. A, f. 161b; Ibn al-­Athir, X, 398–9. 63. al-­Bundari, 134; al-­Husayni, 82; Ibn Khallikan, Wafayat al-a‘yan (Cairo, 1310), II, 47. The latter two writers praise Muhammad’s effort against the Isma‘ilis. 64. It is difficult to agree with Turan’s view that ‘Sultan Mahmud inherited a strong state from his father, Sultan Muhammad’ (O. Turan, Selcuklular Tarihi ve Türk Islam Medeniyeti (Ankaram 1965), 182). A closer analysis of the period would indicate that at least in the west Sultan Muhammad’s authority had been seriously undermined by İl-­Ghazi and Tughtegin. At best, he left a state strong in the central area of his empire only. For a balanced assessment of Sultan Muhammad, cf. Bosworth, ‘The Iranian world’, 113–14. 65. Ibn al-­Azraq, ed. ‘Awad, 208.

T he C a re er o f Na j m a l-D in il - Gh a z i | 43 66. Ibn al-­Azraq, Ms. A, f. 160b; ‘Awad, op. cit., 269. 67. These are listed by Ibn Shaddad (Marsh 333, ff. 65a–b). 68. According to Ibn al-­Azraq, the people of Tiflis had already sought the protection of the Seljuq malik, Tughril, in Arran; but in 515/1121–2, as his help had proved inadequate, they asked İl-­Ghazi to come (Ms. A. f. 161a). 69. Aleppo had a succession of rulers after the death of Ridwan in 507/1113–14. The initiative for their taking command frequently came from influential citizens, especially Ibn al-­Khashshab. 70. This attitude was shown by the people of Aleppo who finally asked İl-­Ghazi to take over control of their city in 511/1117–18. He was clearly their last resort (Ibn al-‘Adim, Zubda, 185). 71. Ibn al-­Azraq, Ms. A, f. 161a. 72. In Rajab 495/April–May 1102, whilst İl-­Ghazi was away from Baghdad, some of his men wanted to cross the Tigris. A sailor who took a long time to answer their request for a boat was promptly killed. The murderer, who was seized by the populace, was then rescued by İl-­Ghazi’s son, whilst the townspeople threw stones at the Turcomans. Although the hajib al-bab seized those people who were involved in the incident, İl-­Ghazi was not satisfied. He went over to the sailors’ quarter and pillaged it. The story did not end there. Some of İl-­Ghazi’s men were killed; still more were drowned crossing the river. Only the efforts of the caliph, who sent the chief qadi, with a professor from the Nizamiyya madrasa, prevented İl-­Ghazi from taking more retaliatory measures on the western side of Baghdad (Ibn al-­Athir, X, 231–2.) 73. Cf. S. Vryonis, Jr, The Decline of Medieval Hellenism in Asia Minor and the Process of Islamization from the Eleventh through the Fifteenth Century (Berkeley, Los Angeles and London, 1971), 143–288. 74. The fate of Mayyafariqin need not necessarily, of course, have been linked to that of Mardin. Mayyafariqin had been the centre of Marwanid power, whereas under Qilij Arslan b. Sulayman b. Qutlumush the town had been ruled from Malatya. It was subsequently taken over by Suqman al-­Qutbi, who ruled it from Akhlat. It was therefore no novelty for the fortunes of Mayyafariqin to be linked with those of a neighbouring city ruled by a dominant power. 75. Ibn al-­Azraq is very specific on the date of İl-­Ghazi’s acquisition of Mayyafariqin: 14 Jumada II 512 (Ms. A. f. 161a). According to Ibn al-­Athir, however, İl‑Ghazi received Mayyafariqin in 515/1121–2 (X, 418). This may well have been only the formalisation of his de facto rule over the city.

44  |  the m ed ie va l tur k s 76. Ibn al-­Azraq, ed. ‘Awad, 283. 77. Ibn al-­Azraq, Ms. A, f. 161b. 78. Ibn al-­Azraq lists the local princelings of Diyar Bakr (ed. ‘Awad, 272). They include Qizil Arslan of Is‘ird, Shahrukh of Hani and Ibrahim of Amid. 79. Ibn al-‘Azraq, ed. ‘Awad, 272. 80. Ibid., 275. 81. Ibn al-‘Adim, Zubda, 129. 82. Ibn al-‘Adim, Zubda, 180–6. 83. ‘The inhabitants, despairing of finding support from any Muslim prince, decided to send ­deputations . . . ­to İl-­Ghazi’ (Ibn al-‘Adim, ibid., 185). 84. Ibn al-­Azraq, Ms. A, f. 161a. 85. Cf. S. Zakkar, The Emirate of Aleppo 1004–1094 (Beirut 1971), 186–234. 86. Nor was he the only ruler interested in acquiring Aleppo (Ibn al-‘Adim, Zubda, 180). 87. In 511/1117–18, İl-­Ghazi was invited to come from Mardin and rid the city of Aq Sunqur. On this first occasion, İl-­Ghazi was not able to gain access to the main citadel and had to make do with the qal‘at al-sharif (ibid., 179–80). 88. Ibn al-­Athir, X, 373; Ibn al-­Qalanisi, 199; al-‘Azimi, 384. 89. Ibn al-‘Adim, Zubda, 185. Sivan says that the personal intervention of Ibn al-­ Khashshab was needed before power in Aleppo could be transferred to İl-­Ghazi (op. cit., 41, citing Ibn Abi Tayyi’ in Ibn al-­Furat, I. ff. 121b–122b). 90. According to Ibn al-­Athir, the treasury in Aleppo was empty because the eunuchs had shared out its contents amongst themselves (op. cit., X, 372). 91. According to Ibn al-‘Adim, the revenues of Aleppo were not adequate for the needs of İl-­Ghazi and his Turcomans. He became estranged from the people of Aleppo and left for Mardin (Zubda, 180). 92. İl-­Ghazi’s attitude to Aleppo is summed up by Ibn al-‘Adim who says that the city was too weakened to be a suitable place in which İl-­Ghazi might live (ibid., 193). 93. The report that Dubays joined in an attack on Aleppo after İl-­Ghazi’s death reveals his continuing interest in the city (Ibn al-‘Adim, Bughyat, 205). 94. Such is the version of Ibn al-­Azraq (Ms. A, f. 161a). Ibn al-­Qalanisi, on the other hand, relates that İl-­Ghazi embarked on the campaign to Tiflis in answer to a call from Sultan Tughril (op. cit., 204–5). Cf. also Matthew of Edessa, 304. 95. Tughril was the brother of Sultan Mahmud. He had first ruled in the area of

T he C a re er o f Na j m a l-D in il - Gh a z i | 45 Zanjan (al-­Bundari, 134; al-­Husayni, 90). In 514/1120–1 he rebelled against Mahmud and took Ganja (Ibn al-­Athir, X, 383–5). 96. Minorsky, citing Vardan, states: ‘In 1092 famine visited Armenia, and then the country was invaded by Al-­Khazi (Ēl-­Ghazi b. Artuq). The latter, with 7,000 men, reached Ani but the Pahlavid Grigor slew his giant brother and the aggressors were beaten off’ (Studies in Caucasian History (London, 1953), 82). 97. Ibn al-­Azraq, Ms. A, f. 161b. 98. Op. cit., 204–5. 99. Sibt b. al-­Jawzi, 101–2; Abu’l-­Mahasin, RHC, III, 499. 100. Anon. Syr. Chron., 89; Matthew of Edessa, 304. 101. Sibt b. al-­Jawzi, 73; Ibn al-‘Adim, Zubda, 193. 102. Op. cit., X, 400. 103. Ibn al-­Qalanisi, tr. Gibb, 159. 104. Ibn al-­Athir, X, 400. 105. Op. cit., 217. 106. Ibn al-­Athir, X, 280. 107. Ibn al-­Athir, X, 280. 108. Ibn al-‘Adim, Zubda, 190. 109. Ibn al-­Athir, X, 400. 110. Ibn al-‘Adim, Zubda, 180. 111. Ibn al-­Qalanisi, 201. This point is also made by Ibn al-‘Adim (Zubda, 191). 112. Ibn al-­Qalanisi, tr. Gibb, 158–9. 113. Ibn al-­Qalanisi, 200. 114. Sivan, op. cit., 41. 115. Ibid., 32–43. 116. Sivan, op. cit., 39. 117. Ibid., 41. 118. Ibn al-‘Adim, Zubda, 188. 119. One Turcoman soldier is reported to have cried out, ‘Was it to obey this man with a turban that we left our country?’ (Ibn al-‘Adim, Zubda, 189). 120. This whole issue is discussed at greater length on pages xxx infra. 121. Ibn al-‘Adim, Zubda, 195. 122. Ibn al-­Azraq says that İl-­Ghazi escaped with only twenty horsemen (Ms. B, f. 103b). 123. For accounts of the death of İl-­Ghazi, cf. Bar Hebraeus, 250; al-‘Azimi, 389; Ibn al-­Athir, X, 426; Ibn al-‘Adim, Zubda, 206; Ibn al-­Azraq, Ms. A, f. 162b; Sibt b. al-­Jawzi, 103; Ibn al-­Qalanisi, 208.

46  |  the m ed ie va l tur k s 124. Ibn al-‘Adim, Zubda, 187. 125. The sources vary widely as to the quantities of troops on both sides, but it would appear that İl-­Ghazi’s men greatly outnumbered Roger’s. 126. Op. cit., 201. 127. Op. cit., 88. For other descriptions of the battle, cf. Ibn al-‘Adim, Zubda, 188–9 and Michael the Syrian, 204. 128. Runciman, op. cit., 151, citing Walter the Chancellor. 129. Ibid. Ibn al-‘Adim gives no indication that there was a particularly barbarous massacre of Frankish prisoners. He simply states that İl-­Ghazi took ransom money from rich captives, killing the rest (Ibn al-‘Adim, Zubda, 193). 130. Usama, tr. Hitti, 149. 131. Runciman, op. cit., 154. 132. Runciman, op. cit., 155. 133. These lines are quoted in Ibn al-­Athir, X, 390. 134. Ibn al-‘Adim, Zubda, 195. 135. Ibid., 200. 136. For an account of Sulayman’s rebellion, cf. Ibn al-‘Adim, Zubda, 200–2; Ibn al-­Athir, X, 417–18. 137. Anon. Syr. Chron., 89. 138. Ibn al-­Athir, X, 398–9; Ibn al-­Qalanisi, 204–5; Sibt b. al-­Jawzi, 101; Ibn al-­Azraq, Ms. A, f. 161a–b; Matthew of Edessa, 304–5; al-‘Azimi, 388; Anon. Syr. Chron., 89. 139. Ibn al-­Qalanisi, 204–5. 140. Ibn al-­Azraq, Ms. A, f. 161b. 141. Op. cit., 89: ‘He returned in shame, escaped to his own land and shortly after fell sick and died.’ 142. The Georgian Chronicle emphasises the great number of Muslim troops ‘which filled the country like the sands of the sea’ and gives specific details about the enormous booty taken (Brosset, op. cit., I, 366). 143. Op. cit., 101–2. 144. Ibn al-­Azraq, ed. ‘Awad, 272. 145. Sibt b. al-­Jawzi, 17–18. 146. Suqman al-­Qutbi fought on Muhammad’s side in 496/1102–3 (Ibn al-­Athir, X, 247–8) and in 498/1104–5 (ibid., 264). Ibn al-­Azraq relates his taking possession of Mayyafariqin (op. cit., ed. ‘Awad, 274–5). 147. Ibn al-­Azraq, ed. ‘Awad, 274–5.

T he C a re er o f Na j m a l-D in il - Gh a z i | 47 148. For the death of Suqman al-­Qutbi, cf. Ibn al-­Athir, Atab., 18; Sibt b. al-­Jawzi, 36. 149. Ibid. Cf. also Ibn al-­Furat, f. 54b (apud Cahen, ‘Le Diyar Bakr’, 233). 150. Ibn al-­Azraq writes that Ibrahim succeeded Suqman, his father, but he does not give a date for Ibrahim’s death (ed. ‘Awad, 269; Ms. A, f. 160b). 151. Suqman had acquired Hisn Kayfa after he had helped Musa in a campaign against Chökermish in 495/1101–2 (Ibn al-­Athir, X, 235). He gained possession of Mardin sometime between this date and his own death, after ‘Ali, the wali of its citadel, had invited him to take the town (ibid., 270). 152. Ibn al-­Azraq, ed. ‘Awad, 269. 153. The struggle between İl-­Ghazi and his brother’s son, Suqman, epitomises a major problem of succession amongst the ruling Turcoman families; namely whether a brother or the eldest son should inherit territories. After the death of Alp Arslan in 465/1072, Qavurt, his brother, claimed the greater right to the inheritance whereas Malikshah adduced the argument ‘A brother does not inherit where there is a son’ (Bosworth, op. cit., 88). In this unresolved inheritance problem the outcome usually depended on a trial of strength. 154. İl-­Ghazi called on his nephew to help him fight Aq Sunqur al-­Bursuqi in 508/1114–15 (Michael the Syrian, 216–17; Matthew of Edessa, 287; Ibn al-­Athir, X, 352). 155. Sic. 156. Anon. Syr. Chron., 89. 157. Cahen, op. cit., 237. 158. Ibid. 159. Members of this family consistently held high office in Mayyafariqin in the Marwanid period and also under the Artuqids. 160. Ibn al-­Azraq, Ms. A, f. 161a. 161. In 511/1117–18 İl-­Ghazi mulcted Aleppan officials of large sums of money with which to buy a truce with the Franks. This move was intended to give him time to fetch reinforcements from Mardin (Ibn al-­Athir, X, 373). 162. Cahen, op. cit., 237. 163. Sevim, ‘Artuk Og˘lu Ilgazi’, 690. 164. Op. cit., 180. 165. Op. cit., X, 372–3. 166. Cf. the phrases used by Ibn al-‘Adim to describe how Aq Sunqur al-­Bursuqi removed the burden of taxation from the people of Aleppo (Bughyat, 204). 167. Runciman, op. cit., 151.

48  |  the m ed ie va l tur k s 168. Usama, tr. Hitti, 149. 169. ‘İlghazi sent a messenger, reproaching the atabek, saying, “We are in need of even one dinar to pay our Turkoman soldiers. Here was a man who had fixed his own ransom at ten thousand d ­ inars . . . a­ nd thou hast killed him!”’ (ibid., 149–50). 170. Cf. especially the memoirs of Usama. 171. Runciman, op. cit., 150. İl-­Ghazi sent this leader, Reynald Mazoir, his ring as a guarantee of safety (Röhricht, op. cit., 135). 172. Cf. Bosworth, op. cit., 119–121. 173. Ibn al-­Azraq, Ms. A, f. 161a. The date of 515/1121–2 given by Ibn al-­Athir is less trustworthy than that of Ibn al-­Azraq who gives the exact day, month and year of the entry of İl-­Ghazi into the city. According to Ibn al-­Athir, İl-­Ghazi sent Temürtash to Sultan Mahmud in 515/1121–2 to plead for mercy for Dubays. It was at this point that the sultan added Mayyafariqin to the iqta‘s already held by İl-­Ghazi (X, 418). 174. Ibn Shaddad, Marsh 333, ff. 17b, 43b and 39b. 175. Ibn al-­Athir, X, 389. 176. Ibid., X, 383. 177. This was mentioned briefly supra, xxx. 178. After Balat, İl-­Ghazi took only Atharib and Zardana (Ibn al-­Athir, X, 390). 179. This is certainly the interpretation given by Sibt b. al-­Jawzi (op. cit., 79). 180. O. Turan, Selcuklular Zamanïnda Türkiye, 166. 181. Ibid., citing Bar Hebraeus, tr. Budge, 249–50. 182. Quoted by E. Quatremère in ‘Notice historique sur les Ismaëliens’, in G. W. Rzewusky (ed.), Fundgruben des Orients (Vienna, 1818), IV, 346. 183. This incident from Ibn al-­Furat is quoted in full by Quatremère (ibid.). Cf. also Ibn al-‘Adim, Zubda, 199. 184. Cf. the stormy relationship between the caliph al-­Mustarshid and Sultan Mas‘ud which culminated in the caliph’s murder in 529/1135. 185. Ibn al-­Azraq, Ms. A, f. 165a–b; Ibn al-‘Adim, Zubda, 197; Ibn al-­Qalanisi, 202. 186. Sibt b. al-­Jawzi, 89; Ibn al-­Azraq, Ms. A, f. 161b. 187. Ibn al-­Athir, X, 399–400. 188. Michael the Syrian, 217; Ibn al-­Athir, X, 352; Usama, 149; Ibn al-­Qalanisi, 191; Sibt b. al-­Jawzi, 52. 189. Ibn al-­Azraq, Ms. A, f. 167a. 190. Ibn Shaddad, Marsh 333, f. 130b.

T he C a re er o f Na j m a l-D in il - Gh a z i | 49 191. Ridwan had even given the khutba in the name of the Fatimids for a short time in 490/1096–7 (Ibn al-­Athir, X, 184). R. W. Crawford views Ridwan’s attitude to the Isma‘ilis as prompted by an awareness of political realities rather than personal or doctrinal motives (‘Ridwan the maligned’, in J. Kritzeck and R. Bayly Winder (eds), The World of Islam: Studies in Honour of Philip K. Hitti (London, 1959), 135–144). 192. Ibn al-­Athir, X, 441; Sibt b. al-­Jawzi, 69. 193. Ibn al-­Azraq, Ms. A, f. 171a–b. 194. After his extremely detailed article, Sevim can only pass a favourable judgement on İl-­Ghazi in his concluding two pages. He labels him a shrewd soldier, a politician and a good administrator (‘Artuk Og˘lu Ilgazi’, 690). He ignores other less positive interpretations which the evidence provokes.

2 The Establishment of Artuqid Power in Diyar Bakr in the Twelfth Century

I. Introduction

T

he most important events in the life of the Artuqid ruler Najm al-­Din İl-­Ghazi are well known. In particular, his victory over Roger of Antioch in 513/1119 at Balat and his ill-­starred campaign into Georgia in 515/1121 are recorded in the majority of contemporary chronicles. Little attention has, however, been paid to the immediate successors of İl-­Ghazi who established themselves in Diyar Bakr after his death in 516/1122. The purpose of this article is to provide a fuller account than has hitherto been available of the careers of İl-­Ghazi’s two most important sons and to comment on their wider significance in the history of the Jazira in the twelfth century. The major work on the early Artuqids remains a pioneer article by Claude Cahen published in 1935.1 In this work Cahen sketched the major political outlines of the reigns of the first three Artuqid r­ ulers – ­İl-­Ghazi, Temürtash and Najm al-­Din A ­ lpı – ­covering the period 500–70/1106–75. In so doing he established for the first time the basic chronology of this confused period in the history of the Jazira. The broad scope of his article, however, necessarily precluded a more detailed examination of the activities of any one of these significant rulers. The present article is intended to fill at least part of that gap by considering the thirty years’ reign of Temürtash, a period which saw the gradual consolidation of Artuqid rule in this area. The most important source for the history of the Artuqids is the town chronicle of Mayyafariqin, the Ta’rikh Mayyafariqin wa-Amid of Ibn al-­ Azraq al-­Fariqi. The value of this work as a source for the history of the 50

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 51 Jazira, northern Syria and Iraq has long been recognised. Amedroz2 and Minorsky3 were the first to use the text and Amedroz transcribed some relevant sections of the text in the footnotes to his edition of Ibn al-­Qalanisi.4 Minorsky ­analysed and translated the account by Ibn al-­Azraq of İl-­Ghazi’s campaign to Tiflis in 515/11215 and Cahen drew heavily on the Ta’rikh Mayyafariqin for his 1935 article. When Amedroz first wrote about Ibn al-­Azraq’s history, the text was known in only two manuscript versions, both of which are in the British Library. These are catalogued under the numbers Or. 5803 and Or. 6310. No further copies of the text have subsequently come to light. The section of the text concerned with Marwanid history has been edited and published by ‘Awad.6 More recently, Ibn al-­Azraq’s account of the reigns of İl-­Ghazi and Temürtash was edited and translated, with commentary, by the present author as part of an unpublished doctoral thesis.7 II. The Succession Problem The sources8 are generally agreed that at the death of İl-­Ghazi in 516/1122–3, his son Sulayman inherited Mayyafariqin whilst another son, Temürtash, succeeded him at Mardin. Aleppo, which İl-­Ghazi also ruled at the time of his death, fell to his nephew, Badr al-­Dawla Sulayman b. ‘Abd al-­Jabbar.9 Behind this apparently straightforward division of territories, however, there remain several unresolved issues which deserve further discussion and on which the Ta’rikh Mayyafariqin wa-Amid sheds some new light. According to lbn al-­Azraq, Temürtash b. İl-­Ghazi stayed behind in Mardin with his atabeg whilst İl-­Ghazi attempted to make the journey to Mayyafariqin with his wife and Sulayman.10 Temürtash experienced no ­serious difficulties in assuming control of the Artuqid possession of Mardin on the news of his father’s death, since he already held the citadel. Besides, it would appear likely that İl-­Ghazi’s hold over that city was firm and that he felt a definite attachment to it. The sources frequently mention that İl-­Ghazi regularly returned to Mardin for rest or reinforcements11 and he must have  regarded it as his real base. He had held Mardin from around 502/1108–9. The situation at Mayyafariqin was markedly different. Since Ibn Jahir had wrested Mayyafariqin from its Marwanid overlord in 478/1085–612 the

52  |  the m ed ie va l tur k s city had been prey to a long succession of rulers, to the detriment of its economic and social welfare. The date of the beginning of İl-­Ghazi’s rule at Mayyafariqin is given by Ibn al-­Azraq as 14 Jumada II 512.13 Ibn al-­Athir, however, writes that İl-­Ghazi was awarded the city as an iqta‘ by Sultan Mahmud as late as 515/1121–2.14 Whatever the actual date of İl-­Ghazi’s acquisition of Mayyafariqin, and even if lbn al-­Azraq’s much earlier date is accepted, İl-­Ghazi did not have the opportunity to spend enough time in the city to consolidate his position there. An analysis of his activities during the years 512–16/1118–22 reveals that he could not possibly have accorded Mayyafariqin more than a passing interest.15 Ibn al-­Azraq’s history contains a most interesting anecdote in which he describes the subterfuge employed by the widow and son of İl-­Ghazi in order to gain access to the citadel of Mayyafariqin.16 On their arrival by night outside the gate of the city, the dead body of İl-­Ghazi was put on his horse. The wali of the city opened the gate and only when the retinue had reached the very centre of the citadel was the death of İl-­Ghazi made known. Ruses of this kind, which concealed the death of a sovereign or an important leader, were far from rare in this period and were often the work of women who wanted to secure the succession for their sons or some other favourite.17 In this particular instance, İl-­Ghazi’s widow was helped by the fact that she was near Mayyafariqin and that she and Sulayman could travel there with the body by night. Had İl-­Ghazi died further from his seat of power, her ruse might well have failed. The khatun had every reason to doubt that Sulayman would be warmly welcomed as the new ruler of Mayyafariqin and to be aware of the need to secure the citadel before announcing the death of İl-­Ghazi. There is no reason to assume that İl-­Ghazi had behaved towards the inhabitants of Mayyafariqin in a more lenient way than any of that city’s earlier overlords. His treatment of the inhabitants of Aleppo and the villages and countryside of northern Syria shows him to have been a child of his times and casts c­ onsiderable doubt on the veracity of the eulogistic account of him given by Ibn al-­Azraq.18 It cannot therefore be assumed that the people of Mayyafariqin would automatically have welcomed a son of his as their new ruler.

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 53 Quite apart from this consideration, it is not at all certain that İl-­Ghazi had in fact appointed Sulayman as his successor in Mayyafariqin. Indeed, Michael the Syrian expressly states that İl-­Ghazi ordered that his son Temürtash should rule after him; but as the latter was not present at the time, İl-­Ghazi’s son Sulayman, who was with him, accompanied him to Mayyafariqin and when he died en route buried him in that city. He then ruled there.19 Sulayman’s succession at Mayyafariqin may well have been effected in spite of the wishes of his father, whose attitude towards him at this time is unclear. The sources agree that Sulayman, who had been left in Aleppo by İl-­Ghazi to manage affairs there, rebelled against his father in 515/1121–2, the year before İl-­Ghazi died. The precise nature of this revolt is not clear and Ibn al-‘Adim, who describes it in great detail, gives a very confusing account.20 lbn al-­Athir relates that in 515/1121–2 Sulayman, egged on by bad advisers, rebelled against his father at Aleppo. İl-­Ghazi rushed there, seized his son, who was full of excuses, and meted out terrible punishments to his accomplices.21 Ibn al-‘Adim, on the other hand, gives two accounts of the rebellion. The one relates that Sulayman was displeased with the demands made on him by his father and rose up against him. During İl-­Ghazi’s absence on his ill-­starred campaign into Georgia, Sulayman expelled various amirs from Aleppo and imposed his own corrupt, despotic rule on the city.22 The other account given by lbn al-‘Adim describes how İl-­Ghazi wrote to his son asking him to stage an apparent ‘rebellion’ in order to free İl-­Ghazi from the obligation of giving Aleppo to Dubays b. Sadaqa, as he had rashly promised to do while on the Tiflis campaign. Sulayman, who was of feckless temperament, actually did rebel. İl-­Ghazi reached Qal‘at Ja‘bar and there met messengers from Sulayman imploring his pardon.23 The opportunistic timing of the uprising, the speed with which İl-­Ghazi suppressed it, and the details of the torture which he inflicted on Sulayman’s confederates, all support the theory that Sulayman acted of his own accord. On the other hand, Dubays was undoubtedly interested in acquiring Aleppo, as subsequent events proved,24 so there may have been some germ of truth in the Tiflis story. Whether Sulayman’s rebellion was genuine or rigged, he was removed by his father from his position at Aleppo and replaced there by İl-­Ghazi’s

54  |  the m ed ie va l tur k s nephew, Badr al-­Dawla Sulayman b. ‘Abd al-­Jabbar. According to lbn al-­ Athir, İl-­Ghazi yielded to his paternal feelings and spared his son’s life. Sulayman fled soon afterwards to Damascus and took refuge with Tughtegin, the father of the khatun. Tughtegin interceded on Sulayman’s behalf but to no avail.25 Ibn al-­Furat also emphasises that İl-­Ghazi could not forgive Sulayman.26 In his forays against the Franks during the last year of his life, İl-­Ghazi turned increasingly to another nephew, Balak, who was to prove his real successor in northern Syria until his premature death in 518/1124.27 It was to Balak that İl-­Ghazi entrusted his two sons, Sulayman and Temürtash, at his death.28 This statement by Matthew of Edessa and the presence of Sulayman with İl-­Ghazi on his last journey to Mayyafariqin would suggest that some kind of reconciliation had actually taken place between them. Indeed, the fact that İl-­Ghazi was prepared to undertake the journey from Mardin to Mayyafariqin at all might be interpreted as a last effort by a dying man to assert his presence in a city of whose loyalty he was unconvinced and thereby to ensure the succession of his son Sulayman there. Conversely, the presence of Sulayman by İl-­Ghazi’s side might be construed as the action of a son waiting to profit from his father’s impending death and making certain that he was on the spot. This could have been his own idea or that of the khatun, who may well have been the motivating force behind their entry into Mayyafariqin. Ibn al-‘Adim describes Sulayman as feckless and his short-­lived rule at Aleppo as despotic.29 He was very young and easily swayed by the evil counsels of his entourage. Thus the events which immediately preceded the entry of Sulayman into Mayyafariqin and his acquisition of its citadel in 516/1122–3 cast an interesting light on the dangerous ruse which he and the khatun practised. They also explain to a certain extent the necessity of such a ploy. It is ironic that, after these elaborate manoeuvres and the eventually successful outcome of the subterfuge, Sulayman was to enjoy only a short-­lived rule in Mayyafariqin. III. Shams al-­Dawla Sulayman at Mayyafariqin, 516–18/1122–4 Little is known of the short reign of Sulayman at Mayyafariqin. The known facts come from Ibn al-­Furat, Michael the Syrian and Ibn al-­Azraq. Had Sulayman lived longer, he would probably have pursued more aggressive

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 55 policies than his brother, Temürtash, at Mardin. Indeed, there are signs that Sulayman had territorial ambitions, not only in the immediate area of Mayyafariqin but probably as far as Malatya. Sulayman took Haza from his cousin, Da’ud, villages around Mayyafariqin from Qurti, the ruler of Arzan, and – more significantly – he seized Khartabirt from Balak,30 either while Balak was absent on campaign or at his death in Rabi‘ I 518/ May 1124. Sulayman’s marriage to a daughter of Sultan Qilij Arslan of Malatya, which is recorded by Ibn al-­Azraq,31 may well have been prompted by a desire on the part of Sulayman to emulate his more successful cousin, Balak, who had also allied himself to the Seljuqs of Rum and acquired more power thereby.32 The proximity of Sulayman’s brother, Temürtash, who now held the citadel at Mardin, did not present problems initially. Ibn al-­Qalanisi relates that at first the two brothers remained friendly.33 Relations between them were still harmonious as late as Rajah 518/September 1124 when, according to Ibn al-‘Adim, Temürtash (who had become disastrously enmeshed in Aleppan affairs) left to go to Diyar Bakr to fetch help from his brother, Sulayman.34 lbn al-­Qalanisi records that later on a quarrel arose between the two brothers ‘which lasted because of both their faults’.35 It would have been comparatively easy for harmony between them to be maintained when Temürtash was often away from Mardin in the service of Balak36 and it is not surprising that dissension really began when Temürtash returned to Diyar Bakr and no doubt interfered in the territorial ambitions of his brother. lbn al-­Furat in fact implies that Sulayman was preparing to attack Temürtash.37 Whatever the cause of the rift between the two brothers, Temürtash considered it more important to remain at Mardin than to answer numerous urgent summons for him to return to Aleppo. A month or so after the departure of Temürtash from Aleppo to Mardin, Sulayman died most opportunely in Ramadan 518/October–November 112438 at Mayyafariqin. This may not have been a mere coincidence. The predictable struggle for hegemony in Diyar Bakr between the two brothers was best settled by the death of one of them. Any complicity on the part of Temürtash in his brother’s death must, however, remain a matter for speculation, especially since lbn al-­Azraq’s testimony favours Temürtash throughout

56  |  the m ed ie va l tur k s and the author would have suppressed any evidence which might reflect badly on his former master. Sulayman does not emerge from the sources as an especially admirable figure. Aside from lbn al-­Azraq’s biased evidence,39 Ibn al-‘Adim confirms that Sulayman was frivolous and a trouble­maker.40 Temürtash may have lacked the flamboyant vigour of his father but he possessed the tenacity and adaptability necessary to stay in the area of Mardin and Mayyafariqin for thirty years. IV. The Sources for the Reign of Temürtash at Mardin and Mayyafariqin It is not the intention in this article to discuss year by year the detailed events in the period of Temürtash’s reign at Mardin and Mayyafariqin. It seems appropriate, instead, to attempt to impose some order and, if possible, interpretations on the wealth of detailed incidents mentioned in Ibn al-­Azraq and other chronicles. The major sources for the rule of Temürtash at Mardin and Mayyafariqin are Ibn al-­Azraq’s text and the History of the Atabegs of Mosul by Ibn al-­Athir. These two works form an interesting foil to one another; the one is highly biased in favour of Temürtash and the other sets out to eulogise the achievements of Zengi and his descendants. Both these works often treat the same historical events but because of the very different slant placed on them by their authors the narratives vary markedly. As well as these two major sources for the history of Diyar Bakr during the thirty years that Temürtash ruled, some events of the period are also mentioned by Michael the Syrian, Ibn al-­ Furat, al-‘Azimi and lbn al-‘Adim. None of these sources have the particular bias of Ibn al-­Azraq or of Ibn al-­Athir and they can therefore in some measure be used as a control. Ibn al-­Azraq’s testimony on the reign of Temürtash is especially valuable for its occasional references to institutions and to the social and economic life of the time. Such references are all too rare and are often only partially explained but they provide information which is not found elsewhere. Ibn al-­ Athir’s History of the Atabegs of Mosul, on the other hand, sheds a fascinating light on the relationship between Temürtash and Zengi. This work provides detailed information on the figure who dominated most of Temürtash’s reign

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 57 and who even on the pages of a tacitly hostile source such as Ibn al-­Azraq’s history is mentioned almost as often as Temürtash himself. V. The Early Career of Temürtash b. İl-­Ghazi Although İl-­Ghazi had a number of sons, Sulayman and Temürtash are the only two who play a significant role in Ibn al-­Azraq’s history. The reason for this is simple: İl-­Ghazi’s other sons had all predeceased him.41 Sulayman and Temürtash were still apparently quite young when their father died in 516/1122–3.42 During İl-­Ghazi’s lifetime, Temürtash had despite his youth performed two useful tasks for his father. He had been left in Aleppo by İl-­Ghazi in 511/1117–18 while the latter went back to Mardin to collect ­reinforcements.43 On another occasion, in 515/1111–12, Temürtash had been sent by his father to Sultan Mahmud to intercede on behalf of Dubays.44 On the death of İl-­Ghazi, Temürtash acquired possession of Mardin without difficulty. The next two years, 516–18/1122–4, he spent in the service of his energetic cousin, Balak, to whose care his father had entrusted him before he died. Temürtash was present at Balak’s siege of Manbij in 518/1124. Indeed, Balak was preparing to leave Temürtash in charge of the siege at Manbij when he was killed outside the walls of the town.45 Significantly, at his death Balak bequeathed his estates to Temürtash46 who proceeded to Aleppo and took possession of it in Rabi‘ I 518/May 1124.47 Temürtash has received much opprobrium for his short administration at Aleppo. Stevenson calls him ‘incompetent’48 whilst Gibb writes that Aleppo had reached the climax of its misfortunes ‘since it was now reduced to dependence on İl-­Ghazi’s indolent son and successor at Mardin, Timurtash’.49 There is certainly ample evidence for a number of blunders and for a marked lack of judgement on the part of Temürtash at this time. His release of Baldwin was based on the perhaps naive assumption that Baldwin would keep his word. Far from doing so, Baldwin refused to comply with the terms of the agreement made with Temürtash and even allied himself with Dubays although he had specifically promised not to do so. After a humiliating defeat outside Aleppo at the hands of Dubays, Temürtash withdrew to Mardin to try to obtain help from his brother, Sulayman,50 and left Aleppo to its fate.

58  |  the m ed ie va l tur k s Whilst Temürtash’s brief attempt at governing Aleppo was far from glorious, his own contribution to the city’s decline may well have been exaggerated. Aleppo was notoriously difficult to rule and a boy of eighteen was scarcely the ideal person to be in charge of it. His preferred modus operandi in his later life was negotiation rather than ruse or the use of military strength, but at Aleppo in 518/1124 he lacked the experience and advisers to deal with such seasoned practitioners of the political art as Baldwin. Aleppo at this particular juncture was coveted by several powerful amirs or ­groups – ­Dubays, who might justifiably have expected to receive the city as a reward for his loyalty to the Artuqids; the dispossessed Badr al-­Dawla, still smarting from the humiliation of being dismissed from his governorship of the city by Balak; and, as always, the Franks. The various conflicting interests of these factions exerted a continuous pressure on Temürtash, who must soon have felt out of his depth in Aleppo. Once he had arrived back at Mardin, Temürtash embarked on policies which furthered his own modest territorial ambitions in that area and he seems deliberately to have turned his back on Aleppo. This attitude on the part of Temürtash is severely condemned by Ibn al-‘Adim, who as the town­chronicler of Aleppo naturally sees Temürtash’s actions in an adverse light and condemns his preoccupation with the affairs of Diyar Bakr to the exclusion of Aleppo.51 When messengers arrived from Aleppo to Mardin in 518/1124, Temürtash temporised with them, absorbed as he was with the death of his brother Sulayman and his own acquisition of the town of Mayyafariqin. He even went so far as to imprison the Aleppan envoys, who escaped and sought help for their city from another source, Aq Sunqur al-­ Bursuqi from Mosul.52 The involvement of Temürtash in the affairs of Aleppo was a singular failure. With a crass political blunder and a military defeat behind him, he retreated into the area of Mardin which his father, İl-­Ghazi, had always regarded as his home base. Thus ended the unsatisfactory bolstering of Aleppo by the Turcoman forces of Diyar Bakr, a process which had begun in 511/1117–18. Thereafter Temürtash rarely occupied himself with political affairs outside the immediate area of Diyar Bakr. He owed his tenure of Aleppo, in any case, to the bequest of Balak rather than to his own efforts to expand his territory. In this

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 59 respect, he is to be contrasted with his father, whose ambition extended well beyond Aleppo. It may be concluded that Temürtash was temperamentally unsuited for the life of virtually ceaseless campaigning which the older generation of his family had led. Viewed from the vantage-­point of medieval Islamic history as a whole, this decision to withdraw to Diyar Bakr marks the end of the uncoordinated but grandiose Artuqid territorial ambitions which had prompted İl-­Ghazi and his father Artuq before him to cover vast stretches of the Islamic world. As far as the livelihood of the Artuqid family was concerned, however, Temürtash’s decision to retreat to the remote, difficult terrain of Diyar Bakr could be judged as realistic and astute. On the death of his brother, Sulayman, Temürtash took possession of Mayyafariqin53 and ruled the two cities jointly for thirty years. VI. The Reign of Temürtash until the Death of Zengi in 541/1146 Whilst Temürtash ruled at Mardin from 516/1122 and at Mayyafariqin from 518/1124–5, his cousin Da’ud had governed Hisn Kayfa since around 502/1108–9. Da’ud had been prepared to accept the overall leadership of İl-­ Ghazi and had furnished him with troops on occasion. He was not, however, likely to allow Temürtash to dominate him in similar fashion. It is true that Temürtash succeeded in seizing Mayyafariqin before Da’ud, but this was one of the few occasions when he independently outwitted his more bellicose and enterprising cousin. lbn al-­Azraq never explicitly states that Da’ud had the upper hand in his relationship with Temürtash but stresses in unusually emotive terms the savagery with which Da’ud pillaged and plundered. Ibn al-­Athir, on the other hand, who has no reason to take sides between the two Artuqids, clearly implies that it was Da’ud, not Temürtash, who presented the real threat to Zengi’s aspirations in Diyar Bakr and that Da’ud had inherited that much-­prized ability to inspire devotion in the Turcomans which İl-­Ghazi had possessed before him. According to lbn al-­Athir, Zengi would have been kept busy if his only adversary had been Da’ud, who enjoyed such prestige with the Turcomans that all those capable of carrying arms would join him. lbn al-­Athir admires Da’ud’s resilience and his ability to return to fight only a few days after a crushing defeat.54

60  |  the m ed ie va l tur k s In the early years of Temürtash’s reign, Da’ud seized the territories formerly held by Balak, such as Khartabirt,55 and even participated with the ruler of Akhlat in a campaign against the Georgians.56 Any further ambitions Da’ud may have had were soon, however, laid aside in the face of the threat of Zengi who was made governor of Mosul in 520/1126. The incipient rivalry between Da’ud and his cousin Temürtash was forgotten at the prospect of a common foe who was preparing to invade their territories. Once Zengi had established himself at Mosul, his first act of aggression was to lay siege that same year (520/1126) to Nasibin, which belonged to Temürtash. Temürtash appealed to Da’ud for help against Zengi and this was promised him. After a well-­known incident involving Zengi’s interception of a pigeon carrying a message from Temürtash to the governor of Nasibin, Zengi tricked the garrison of the town into surrendering it to him.57 This initial act by Zengi united the Artuqids and together they came to an agreement in 524/1130 with İl-­Aldı, the ruler of Amid, and other Turcoman chiefs to defend their territories against Zengi. In spite of their large numbers, the combined Artuqid forces were defeated by Zengi who took Dara and Sarja.58 After this defeat at the hands of Zengi, the uneasy alliance between the two Artuqid cousins was at an end. There are signs that each of them reverted to their own natural mode of government; Temürtash consolidated his hold on the two cities under his command, whilst Da’ud energetically raised his sights to new conquests. The year after the defeat at Sarja, Da’ud seized several citadels south of Lake Van.59 It would appear that Temürtash was worried by the implications of Da’ud’s actions and that he sought a new protector. It is not clear whether he first approached Zengi or whether Zengi suggested an alliance with him. Even if their future collaboration was at Temürtash’s instigation, Zengi dominated the relationship from the outset and knew how to exploit the rivalry between the two Artuqid cousins to his own advantage. Zengi and Temürtash clashed with Da’ud outside Amid in 528/1134 and Da’ud was defeated. Zengi then went on to take the citadel of al-­Sur which he handed over to Temürtash.60 Although Ibn al-­Azraq is silent at this point, al-‘Azimi clearly states that Temürtash had become Zengi’s vassal.61

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 61 Zengi’s gift of this citadel and other relatively unimportant possessions to Temürtash was a clever move on his part since he thereby weakened both Da’ud and Temürtash. Temürtash, well-­pleased for a brief moment with his new possessions but not dangerously strengthened by them, could not enjoy for long the security afforded him by his new-­found protector who spent only short periods in Diyar Bakr. Once Zengi had gone away, Da’ud, enraged by his loss of territories and Temürtash’s agreement with Zengi, terrified Temürtash to such an extent that in 530/1135–6 he even demolished the rabad, and another suburb of his own city of Mayyafariqin, presumably because he felt unable to defend them against the depredations of his cousin.62 By setting the two Artuqids against one another, Zengi cleverly kept them occupied in Diyar Bakr and successfully diverted Da’ud’s energies away from Zengi’s own sphere of influence. Temürtash had now tried to join forces with both Zengi and Da’ud. In neither case had the alliance brought him significant advantages. In spite of the coalition with Da’ud, Temürtash had lost Nasibin and when he had subsequently thrown in his lot with Zengi, this had resulted in an exacerbation of the hostility between himself and Da’ud. From 530/1135–6 onwards, Temürtash was formally attached to neither of his two rivals. For a decade he played a pragmatic but dangerous game of shifting alliances and no doubt careful diplomacy. The sources are silent on the exact details of his  policies toward Da’ud and Zengi, but the results of his unobtrusive  stance were  ­successful. By the end of 541/1146–7, both Da’ud and Zengi were dead. Temürtash was still alive and still had possession of the two cities of Mardin and Mayyafariqin. His waiting game had proved worth while. Before Zengi’s death in 541/1146, their relationship had seriously deteriorated. One event which sparked off a quarrel between them occurred in 533/1138–9 when the governor of Nasibin, Abu Bakr, fled to Temürtash for protection from Zengi. Zengi, whose brutality and iron discipline63 were legendary, asked for Abu Bakr to be handed over to him.64 When Temürtash refused, Zengi came to Mardin. Temürtash extricated himself from this dangerous situation by conceding Dara to Zengi, as well as giving his daughter in marriage;65 ­but – ­if Ibn al-­Azraq is to be ­believed – ­he did not yield up Abu Bakr.

62  |  the m ed ie va l tur k s The agreement reached in 536/1141–2 between Da’ud and Temürtash probably prompted Zengi’s demand to Temürtash to send his chief official, Habashi, to him.66 Zengi’s motives remain unknown but the incident ended with the assassination of Habashi at Zengi’s camp. Temürtash could not fail to be worried by the sudden expansion of Zengi’s lands immediately after the death of Da’ud in 539/1144–5, when Zengi seized nearly all Da’ud’s territories.67 Ibn al-­Azraq comments at this point that Temürtash and Zengi quarrelled, but he gives no details.68 Even before Da’ud’s death, there are signs that Temürtash sought an alliance with Da’ud and Joscelin, the Frankish ruler of Edessa,69 since Zengi was approaching the height of his power and was not prepared to brook resistance from his former vassal. In 540/1145–6 Zengi came as far as Tall al-­Shaykh and threatened the town of Mayyafariqin70 before withdrawing. Presumably this was a demonstration of strength on his part, intended to intimidate Temürtash. It was exceedingly fortunate for the latter that very soon afterwards, in 541/1146, Zengi himself was assassinated at Qal‘at Ja‘bar. With the removal in the space of two years of his two main rivals who had constantly harassed and restricted him, Temürtash suddenly found himself free to act in a more independent way than hitherto, especially since he himself had now been in power for a long period. As already noted, by the end of Zengi’s career a degree of personal animosity had crept into his relationship with Temürtash. Indeed, relations between them resembled those between a lion and a mouse. Ibn al-­Athir sheds very interesting light on this. Several farmers had left Mosul to go to Mardin. Zengi was annoyed by this and asked Temürtash to send them back. Temürtash refused, saying that he treated farmers well, taking only one-­tenth of their crops, and that if Zengi had done likewise, the farmers would not have left his territory. Zengi’s reply to this gesture of defiance was characteristically chilling:71 Tell your master that if he took (only) a one-­hundredth share, that would be a large amount for him, occupied as he is with his pleasures in the citadel at ­Mardin . . . ­If it had not been for me, he would long since have ceased to drink water safely at Mardin and the Franks would have taken it (the city).72

Temürtash sent back the farmers.

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 63 VII. Temürtash at the Height of his Power, 541–8/1146–53 According to Ibn al-­Azraq, Temürtash reacted with joyful alacrity to the news of Zengi’s death. He left his garden immediately and seized a number of neighbouring territories, such as Hani and Siwan.73 Such speed was imperative if the citadels were to be captured before Zengi’s successor at Mosul, his son Sayf al-­Din Ghazi, could establish himself and feel strong enough to come to claim his father’s old possessions. Da’ud’s son, Qara Arslan, who had succeeded his father at Hisn Kayfa, had already been decisively defeated the year before Zengi’s death by Temürtash’s troops74 and although Qara Arslan’s fortunes also rose with the removal of Zengi and accession of Sayf al-­Din Ghazi,75 it is clear that Qara Arslan soon recognised the overall suzerainty of Temürtash.76 Temürtash received a temporary setback in 543/1148–9 when he endured a humiliating defeat at the hands of Sayf al-­Din Ghazi who came as far as Mardin to regain Zengi’s territories.77 This son of Zengi was not, however, destined to bother Temürtash for long since Sayf al-­Din died the following year (544/1149) at Mosul.78 The remaining few years of Temürtash’s reign were a period of genuine, if modest, power in the immediate area of Diyar Bakr. He obtained a recognition of his own suzerainty from the lord of Amid79 and acquired new territories, such as Sumaysat in 544/1149–5080 when the ruler of Edessa, Joscelin, was taken prisoner by Nur al-­Din. All the evidence points to Temürtash’s enjoyment of considerable local power before his death and to his being recognised as the most powerful ruler in Diyar Bakr. His brief but disastrous interlude at Aleppo when he was still a young man was his only real taste of power outside Diyar Bakr. An analysis of his whole policy testifies to a definite decision on his part to restrict his aspirations to the maintenance of his two cities of Mardin and Mayyafariqin and to the acquisition of whatever territories he could retain within a modest distance of his base at Mardin. A negative appraisal of his rule could pinpoint a lack of energy and ambition or e­ ven – ­as Ibn al-­Athir suggests81 – a life given over to pleasurable pursuits. More positively, however, his policies could be viewed as realistic and realisable. In view of the evidence presented in the above discussion and the

64  |  the m ed ie va l tur k s personality of Temürtash himself, who generally appears to have preferred negotiation to acts of sudden aggression, it is difficult to understand how the extraordinary but surely apocryphal story of Temürtash at Jerusalem continues to be treated seriously by Western historians of the Crusades. According to this anecdote, which comes from William of Tyre, and only from him, Temürtash made a sudden attack on Jerusalem with his bands of Turcomans, prompted by a desire to regain the lands once owned by his grandfather, Artuq. This story is treated seriously by Weil,82 Rohricht83 and Stevenson,84 and is allowed to remain unchallenged even by as recent a historian as Runciman.85 According to Stevenson,86 this anecdote from William of Tyre calls the invader of Jerusalem ‘Hiaroquin’. The identification of this name with Temürtash is Wilken’s.87 Whether or not such an identification is linguistically feasible, it is entirely out of character that Temürtash should make this sudden, hare-­brained and arduous foray towards Jerusalem. It is unwise, in any case, to treat as sacrosanct a narrative which appears only in this one source. If such a raid did indeed take place, then, directed as it was against ­Jerusalem – ­a prime bone of contention between Muslim and Frank in the twelfth ­century – ­it would surely have found some echo in Muslim and other Western sources.88 Temürtash’s only audacious ‘coup’ for which more than one account exists is an isolated reference to his capture of a caravan at Edessa in 533/1138–9.89 Whilst this incident, and his capture of a modest share of Joscelin’s p ­ ossessions – ­in which other neighbouring rulers also ­participated – ­testify to certain territorial ambitions they pale into insignificance beside a raid on Jerusalem itself. VIII. The Administration of Temürtash lbn al-­Azraq, and to a much lesser extent Ibn al-­Athir, make isolated references to aspects of social and economic life in the reign of Temürtash. It would, however, be foolhardy to make anything other than very tentative statements on the basis of such remarks as are found in the sources. Such wider issues as the inter-­relationship between towns and the nomadic groups present in the area of Diyar Bakr, as well as the treatment of the Christians, who probably outnumbered the Muslims in Temürtash’s reign, must remain almost entirely undiscussed through lack of information.

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 65 Certain concrete facts about Temürtash’s domestic policy emerge from lbn al-­Azraq’s history. He removed the remaining local resistance to him with his capture in 531/1136–7 of Hattakh, the last remaining Marwanid possession in Diyar Bakr.90 Individual members of the Marwanid family were, however, subsequently employed by Temürtash.91 There is evidence of an interest in civil as well as military building, which testifies to Temürtash’s desire to put down roots in the area of Mardin and Mayyafariqin. For Mardin, lbn al-­Azraq mentions the construction of a family tomb for the Artuqid family where the bodies of İl-­Ghazi and Sulayman were eventually housed.92 The mosque at Mayyafariqin was rebuilt after its dome collapsed in 547/1152–3.93 The most ambitious building project, however, was clearly the construction of the Qaraman bridge which was begun in 541/1146–7, completed in 548/1153–4 and was clearly a most impressive edifice.94 It is impossible to determine the exact religious allegiance of Temürtash, although his warm reception of an Isma‘ili shaykh is described in great detail by lbn al-­Azraq. Perhaps the elaborate deference which he accorded this shaykh testifies to the latter’s personal magnetism rather than to any deep-­ rooted religious inclination towards the Isma‘ili creed. This shaykh eventually lost his hold over Temürtash and left him.95 The evidence on commercial matters is also very slight. It is interesting to note that towards the end of his reign Temürtash minted coins.96 This one minting of copper coins is revealing on two counts. Firstly, it is an indication that some kind of trading took place with neighbouring areas.97 This was no doubt facilitated by the discovery in the reign of Temürtash of a copper mine north of Mayyafariqin.98 Secondly, it is a sign that Temürtash had at least begun to assume some of the trappings of a traditional Muslim ­ruler – ­rather than those of a semi-­nomadic chief. The fact that he did not strike gold coins which would have had a more than local currency is itself significant. Even in his arrogation to himself of the right to siqqa, Temürtash retained his customary local bias. The civilian inhabitants of the area of Mardin and Mayyafariqin were probably well-­treated. Ibn al-­Azraq’s biased testimony states that Temürtash behaved kindly towards the local ‘ulama’ and the important indigenous families.99 Ibn al-­Athir reveals that Temürtash appropriated in taxation only

66  |  the m ed ie va l tur k s a tenth of the revenue from the harvests in comparison with Zengi, who said that if it were necessary he would be prepared to levy as much as two-­thirds for his important military commitments.100 Allusions to Temürtash’s treatment of the Christian population are all too rare. As well as the possibility mentioned in Ibn al-­Azraq’s text that Ibn Mukhtar, one of his officials, was a Christian,101 a possibility which Cahen sees as a certainty,102 Michael the Syrian relates that Temürtash did not treat the Christians well but that at the end of his reign he repented.103 Michael the Syrian is as capable of biased testimony as Ibn al-­Azraq and it is difficult to assess the full significance of this one statement. The frequent changes of vizier and other officials faithfully recorded by Ibn al-­Azraq are not remarkable for the time. They are an indication of administrative malpractices prevalent among the Artuqids and other contemporary dynasties. The major achievement of Temürtash was his ability to hold on to Mardin and Mayyafariqin in this troubled period and to bring some measure of stability to these two cities. A rather belated recognition of this achievement was made by the caliph in 547/1152–3, the year before Temürtash died. Robes of honour and an edict entitling him to the land arrived for him.104 This event is mentioned without comment by lbn al-­Azraq, although it is very significant that Temürtash only received official recognition from Baghdad thirty years after his taking possession of Mardin. This may be an indication of the relatively unimportant status accorded him by the caliph and the sultan. It may also reveal that Temürtash enjoyed real power in Diyar Bakr only towards the very end of his reign, a fact which is demonstrably proved by historical events. As lbn al-­Azraq’s panegyric expresses it: After Zengi (had been killed), no amir other than Husam al-­Din (Temürtash) remained independent, unopposed, unchallenged and governing alone.105

To balance this glowing testimony to Temürtash’s achievements it should be added that whilst many medieval Muslim historians record the death of his father, İl-­Ghazi, his own death passes unmentioned except by a few sources.106

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 67 IX. Conclusions With the accession of Temürtash in 516/1122–3 at Mardin and 518/1124–5 at Mayyafariqin, the territorial ambitions of the Artuqids shrank to the purely local area of the Jazira. Temürtash was fortunate to survive for thirty years during a period which was dominated by the monumental figure of Zengi. Locally, Temürtash was dwarfed by his cousin Da’ud and enjoyed a brief interlude of genuine power in Diyar Bakr only in the last few years of his life, when both Zengi and Da’ud had died. Broadly speaking, his successors also retained the local emphasis of his polity. The death of İl-­Ghazi in 516/1122–3 could have ended Artuqid hopes of acquiring for themselves a portion of Islamic territory as their permanent domain. With the succession of Temürtash at Mardin and Mayyafariqin, a dynasty was firmly established and, against all odds, maintained itself in that same area until the early fifteenth century.

Appendix: The Genealogy of the Early Artuqids

70  |  the m ed ie va l tur k s Notes 1. C. Cahen, ‘Le Diyar Bakr au temps des premiers Urtukides’, Journal Asiatique CCXXVII (1935), 219–76. Little that is new is added in the relevant chapters of O. Turan, Doğu Anadolu Türk Devlelleri Tarihi (Istanbul, 1973). A. Sevim’s article ‘Temultaş in Halep hâkimiyeti’, Belleten XXV, Say1 100 (1961), 571– 81, is of limited relevance here since it does not deal with Temürtash’s later activities in Diyar Bakr. 2. H. F. Amedroz, ‘Three Arabic MSS. on the history of the city of Mayyafariqin’, Journal of the Royal Asiatic Society (1902), 785–812. 3. V. F. Minorsky, ‘Caucasica in the history of Mayyafariqin’, Bulletin of the School of Oriental and African Studies XIIJ/4 (1949), 27–35. 4. Ibn al-­Qalanisi, Dhayl ta’rikh Dimashq, ed. H. F. Amedroz (Leiden, 1908). 5. Op. cit. 6. lbn al-­Azraq, Ta’rikh al-Fariqi, ed. B. A. L. ‘Awad., rev. M. S. Ghorbal (Cairo, 1959). 7. C. Hillenbrand, The History of the Jazira 1100–1160: The Contribution of Ibn al-Azraq al-Fariqi, 2 vols (Edinburgh, 1979). 8. In this article, the following abbreviations have been used: Anonymous Syriac Chronicle: A. S. Tritton, ‘The First and Second Crusades from an Anonymous Syriac Chronicle’, with notes by H. A. R. Gibb, JRAS (1933), 69–101, 273–305. ‘Awad, Ta’rikh al-Fariqi, ed. B. A. L. ‘Awad, revised M. S. Ghorbal (Cairo, 1959). al-‘Azimi, ‘Chronique’: C. Cahen (ed.), ‘La chronique abrégée d’al-‘Azimi’, JA CCXXX (1938), 353–448. Bundari, Zubdat: Zubdat al-nusra wa nukhbat al-‘usra, ed. M. T. Houtsma (Leiden, 1889). Ibn al-‘Adim, Bughyat: Bughyat al-talab fi ta’rikh Halab, ed. A. Sevim (Ankara, 1976). Ibn al-‘Adim, Zubdat: Zubdat al-Halab min ta’rikh Halab, II, ed. S. Dahan (Damascus, 1954). lbn al-­Athir, Kamil, X: al-Kamil fi’l-Ta’rikh, X, ed. C. J. Tornberg (Leiden and Uppsala, 1864). Ibn al-­Athir, Atabegs: Ta’rikh al-dawla al-atabakiyya, ed. A. Tulaimat (Cairo, 1963).

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 71 Ibn al-­Azraq, Ms. A: British Library Ms. Or. 5803. Ibn al-­Azraq, Ms. B: British Library Ms. Or. 6310. Ibn al-­Furat, Duwal: Ta’rikh al-duwal wa’l-muluk, Vienna MS A.F. 811. lbn al-­Qalanisi, Dhayl: Dhayl ta’rikh Dimashq, ed. H. F. Amedroz (Leiden, 1908). Ibn Shaddad, Jazira: al-A‘laq al-khatira, Bodleian Ms. Marsh 333. Matthew of Edessa, Chronique: Patmut’iwn, tr. E. Dulaurier (Paris, 1858). Michael the Syrian, Chronique: Chronique de Michel le Syrien, tr. J.-B. Chabot (Paris, 1899–1914). Sibt b. al-­Jawzi, Mirat: Mirat al-zaman, ed. anon. (Hyderabad, 1951). Usama, Memoirs: Memoirs of an Arab-Syrian Gentleman, tr. P. K. Hitti (repr. Beirut, 1964). 9. Ibn al-­Qalanisi, Dhayl, 208; Anonymous Syriac Chronicle, 89; Michael the Syrian, Chronique, 218; Ibn al-­Athir, Kamil, X, 426. 10. Ms. A, f. 162b; Ms. B, f. 105a. 1.1 Ibn al-­Athir, Kamil, X, 389; Ibn al-­Qalanisi, Dhayl, 200. 12. ‘Awad, 212; lbn al-­Athir, Kamil, X, 93–4. 13. Ms. A, f. 161a; Ms. B, f. 101a. 1.4 lbn al-­Athir, Kamil, X, 418. If Ibn al-­Athir’s date is to be trusted, the granting of the city of Mayyafariqin to İl-­Ghazi by the sultan may well have been merely the formal recognition of a de facto situation which had already existed for several years. 15. This was a period when İl-­Ghazi was deeply involved in Aleppan affairs. 16. Ms. A, f. 162b; Ms. B, f. 105a. 17. A famous incident of this kind involved the widow of Malik-­Shah, Terken Khatun, who concealed the death of her husband in 485/1092–3 (Ibn al-­ Athir, Kamil, X, 142–5). She did not, however, resort to the same extremes as İl-­Ghazi’s widow. 18. Ibn al-­Azraq, Ms. A, f. 161a. 19. Michael the Syrian, Chronique, 218. 20. Ibn al-‘Adim, Zubdat, 200–2. 21. Ibn al-­Athir, Kamil, X, 417–8. 22. lbn al-‘Adim, Zubdat, 200. 23. Ibid., 200–2. 24. Ibn al-‘Adim, Bughyat, 228–9. 25. lbn al-­Athir, Kamil, X, 418. 26. lbn al-­Furat, Duwal, I, f. 179b (apud Cahen, ‘Diyar Bakr’, 241).

72  |  the m ed ie va l tur k s 27. lbn al-’Adim, Zubdat, 219; lbn al-­Athir, Kamil, X, 436; al-‘Azimi, ‘Chronique’, 394. 28. Matthew of Edessa, Chronique, 307. 29. Ibn al-’Adim, Zubdat, 200. 30. Ibn al-­Azraq, Ms. A, f. 162b; Michael the Syrian, Chronique, 211; Ibn al­‘Adim, Zubdat, 220. 31. Ms. A, f. 162b. 32. Balak had married the widow of the Seljuq ruler of Malatya and had become atabeg to her son (Turan, op. cit., 147). 33. Ibn al-­Qalanisi, Dhayl, 208. 34. Ibn al-‘Adim, Zubdat, 223. 35. Ibn al-­Qalanisi, loc. cit. 36. Cf. xxx, infra. 37. Cahen, ‘Diyar Bakr’, 241. 38. Ms. A, f. 163a. 39. Cf. lbn al-­Azraq’s description of the depraved son of Sulayman, Mahmud, where the author savours with obvious relish the details of this ‘black sheep’ of the Artuqid family (Ms. A, f. 163a). 40. Ibn al-­Adim, Zubdat, 200. 41. Cf. the attached genealogical table. 42. According to Ibn al-­Athir, Temürtash was seventeen in 515/1111–12 (Kamil, X, 418). 43. lbn al-­Qalanisi, Dhayl, 199; Ibn al-­Athir, Kamil, X, 373; Ibn al-‘Adim, Zubdat, 180; al-‘Azimi, ‘Chronique’, 197. According to the last two sources, Temürtash was left in Aleppo as a hostage. 44. Ibn al-­Athir, Kamil, X, 418. 45. G. Weil, Geschichte der Chalifen III (Mannheim, 1851), 240. 46. Matthew of Edessa, Chronique, 312. 47. lbn al-‘Adim, Zubdat, 220. 48. W. B. Stevenson, The Crusaders in the East (Cambridge, 1907), 111. 49. H. A. R. Gibb, ‘Zengi and the fall of Edessa’ in K. M. Setton and M. W. Baldwin (eds), A History of the Crusades. I: The First Hundred Years (Madison, Milwaukee and London, 1969), 452. 50. Ibn al-‘Adim, Zubdat, 220–3. 51. lbn al-‘Adim, Zubdat, 225. 52. Ibn al-‘Adim, Bughyat, 204–7. Ibn al-­Azraq’s own reticence on this whole incident may well spring from ignorance. More probably, it is the result of a

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 73 desire on his part to suppress evidence which is unfavourable to the Artuqid dynasty. Cf. Ms. A, f. 163a. 53. Ibn al-­Azraq, Ms. A, f. 162b. 54. Ibn al-­Athir, Atabegs, 81. 55. Ibn al-­Azraq, Ms. A, f. 163b. 56. Matthew of Edessa, Chronique, 146. 57. Ibn al-­Athir, Atabegs, 36–7. 58. Ibid., 38–9; Michael the Syrian, Chronique, 240; Ibn al-­Athir, Kamil, X, 467. 59. Ibn al-­Azraq, Ms. A, ff. 164a–b. These citadels included Qatalbas and Batasa, as well as the important town of Is‘ird. 60. Ibn al-­Azraq, Ms. A, f. 168a; Ibn al-­Qalanisi, Dhayl, 243; Ibn al-‘Adim, Zubdat, 253; Ibn Shaddad, Jazira, f. 133b; Ibn al-­Athir, Kamil, XI, 6–7; Usama, Memoirs, 185–6. 61. Al-‘Azimi, ‘Chronique’, 408. 62. Ibn al-­Azraq, Ms. A, f. 168a. 63. E.g. Bundari, Zubdat, 205. 64. Ibn al-­Azraq, Ms. A, f. 176b. 65. Ibn al-­Azraq, Ms. A, f. 169a; al-‘Azimi, ‘Chronique’, 417; lbn al-‘Adim, Zubdat, 271. 66. Ibn al-­Azraq, Ms. A, f. 170a. 67. Ibid. 68. Ibid. 69. Anonymous Syriac Chronicle, 280. 70. Ibn al-­Azraq, Ms. A, f. 171a. 71. Even a favourable source such as Ibn al-­Athir abounds in such anecdotes which testify to Zengi’s cruelty and ability to inspire terror. Bundari goes even further in his description of Zengi, calling him a ‘tyrant, striking at random, and a raging blast of calamities, tigerish in nature, lion-­like in ­malevolence . . . ­feared for his v­ iolence . . . ­the death of his enemies and of his subjects’ (Gibb, ‘Zengi’, 457, n. 10, quoting Bundari, Zubdat, 205). 72. lbn al-­Athir, Atabegs, 79. 73. Ms. A, f. 172b. 74. Ibid., f. 171a. 75. Qara Arslan acquired all the fortresses which Zengi had taken from his father Da’ud (Ibn al-­Azraq, Ms. A, f. 172b). Cf. also Michael the Syrian, Chronique, 268. 76. This is the interpretation which may be placed on Temürtash’s action in taking

74  |  the m ed ie va l tur k s Is‘ird and then handing it back to Qara Arslan a little later (lbn al-­Azraq, Ms. A, f. 173a). 77. Ibn al-­Athir, Atabegs, 91; Ibn al-­Azraq, Ms. A, f. 172b; Michael the Syrian, Chronique, 275. 78. Ibn al-­Athir, Atabegs, 92. 79. Ibn al-­Azraq, Ms. A, f. 176a. 80. Ibid., f. 175a; Michael the Syrian, Chronique, 297. 81. Ibn al-­Athir, Atabegs, 79. 82. Weil, op. cit., 296. 83. R. Roöhricht, Geschichte des Königreichs Jerusalem (1100–1291) (Innsbruck, 1898), 271. 84. Stevenson, op. cit., 171. 85. S. Runciman, A History of the Crusades. II: The Kingdom of Jerusalem and the Frankish East 1100–1187 (Cambridge, 1954), 337. 86. Stevenson, op. cit., 171, n. 2. 87. F. Wilken, Geschichte der Kreuzziige (Leipzig, 1807–32), IV, 17. 88. Röhricht draws attention to this fact but does not in so doing cast doubts on the veracity of the anecdote (op. cit., 271, n. 2). 89. Michael the Syrian, Chronique, 246; Ibn al-­Azraq, Ms. A, f. 169a; Anonymous Syriac Chronicle, 278. 90. Michael the Syrian, Chronique, 264; Sibt b. al-­Jawzi, Mir’at, 161; Ibn al-­Furat, Duwal, f. 93b. 91. ‘Awad, 254. 92. lbn al-­Azraq, Ms. A, f. 171a. 93. Ibid., f. 175a. 94. Ibid., ff. 171a–b, f. 179b. After the death of Temürtash his son, Najm al-­Din Alpı, completed the work on the bridge, which became the model for other bridges in the area, especially the one built by Fakhr al-­Din Qara Arslan on the Tigris at Hisn Kayfa (Ms. A, f. 179b). 95. lbn al-­Azraq, Ms. A, f. 171b. 96. Ibid., f. 172b. Cf. S. Lane-­Poole, The Coins of the Turkmán Houses of Seljook, Urtuk, Zengee, etc. in the British Museum (London, 1877), 139–40. 97. Copper coins were commonly those intended for local use. The lack of silver is not surprising; the Near East at this time was in the throes of a silver famine. 98. Ibn al-­Athir, Kamil, X, 215. 99. Ibn al-­Azraq, Ms. A, f. 176b. 100. Ibn al-­Athir, Atabegs, 79.

t h e e st ab li sh me nt o f a r tuq id p o we r in d iy a r b a k r  | 75 101. Ibn al-­Azraq, Ms. A, f. 173a. 102. Cahen, ‘Diyar Bakr’, 268. 103. Michael the Syrian, Chronique, 311. 104. lbn al-­Azraq, Ms. A, f. 175a. 105. Ibid. 106. Ibn al-­Qalanisi, Dhayl, 329; Bundari, Zubdat, 244; Ibn al-­Athir, Atabegs, 106; Sibt b. al-­Jawzi, Mir’at, 218–19.

3 The History of the Jazira, 1100–1250: A Short Introduction

T

his paper will not attempt to give a detailed dynastic history of the Jazira in the period 1100–1250.1 Instead, it will raise certain general questions and themes and try to pinpoint some of the elements which helped to form the prevailing atmosphere, both religious and cultural, of the period. But first a few words of introduction and definition. Medieval Muslim geographers traditionally divide Mesopotamia into two areas: al-Sawad and al-Jazira. The Jazira was bounded to the south-­west by Syria, to the north and north-­east by Armenia, to the east by Azerbaijan and to the south by Iraq. The Jazira takes its name, of course, from the fact that it is broadly speaking an ‘island’ between the two great rivers, the Tigris and the Euphrates. The area is traditionally sub-­divided into three territories: 1. Diyar Bakr to the north, with the major cities of Mayyafariqin and Amid. 2. Diyar Mudar to the west, with its principal towns of al-­Raqqa, Harran, Edessa and Saruj. 3. Diyar Rabi‘a, the eastern and largest province of the Jazira. Its major cities included Balad, Mosul, Mardin and Nisibin.2

Before raising more general questions, a brief historical overview of the period 1100–1250 in the Jazira would now seem appropriate. It is well known that many Turcoman groups who in the eleventh century had been directed by the Seljuq rulers of Persia out of their own territories towards north-­west 76

the hi stor y o f the j a zir a  | 77 Persia, the Caucasus and Anatolia stayed in Asia Minor after the battle of Manzikert in 1071. The difficulties which the Seljuqs encountered in exercising control over these far-­flung areas became exacerbated after the death of the great Seljuq sultan, Malikshah, in 1092, and were intensified even further in 1118 when Sultan Muhammad died. A number of small principalities arose in eastern Anatolia and the Jazira in the late eleventh and early twelfth centuries. These included the Seljuqs of Rum, the Artuqids of Diyar Bakr, the Shah-­i Arman and the Zengids of Mosul. The Artuqids managed, in spite of vigorous opposition, to keep hold of Diyar Bakr for 150 years. The founder of the dynasty, İl-­Ghazi, captured the two cities of Mardin and Mayyafariqin early in the twelfth century.3 After his death in 1122, his son Temürtash succeeded in holding on to the area for thirty years. Temürtash’s rule was dominated by the redoubtable ­Zengi – w ­ ho ruled Mosul and was always his military and political superior. After Zengi’s death, however, Temürtash was able to pass on a relatively stable principality to his successors, who held on to the area until the fifteenth century.4 Another branch of the Artuqid family ruled at Hisn Kayfa and Amid until the Ayyubid conquest. The rise of the Zengid dynasty (1127–1222) in the Jazira, centred on Mosul, was, however, destined to erode Artuqid power. The Zengids were of course involved, especially under the leadership of Nur al-­Din, in the Muslim struggle with the Crusaders in Syria and the Holy Land, a struggle which culminated in the recapture of Jerusalem by Saladin. Saladin’s successors, the Ayyubids, adopted a policy of greater détente towards the Crusaders but they moved into Diyar Bakr which they held until the Mongols came, taking Mayyafariqin and Mardin in 1259 and 1261.5 Further east, the period 1100–1250 witnessed an Indian summer for the ‘Abbasid caliphate of Baghdad before it was extinguished by the Mongols. Vigorous caliphs such as al-­ Muqtafi (1136–60) and above all al-­ Nasir (1180–1225) acted as territorial, not puppet, rulers and extended caliphal control over most of Iraq, thus placing themselves in direct opposition to the last Seljuq rulers of western Persia and Iraq and to the Khwarazmshahs who succeeded the Seljuqs.6 This paper will now attempt to answer three main questions à propos the Jazira in the period 1100–1250:

78  |  the m ed ie va l tur k s 1. What was the political orientation of the Seljuq ‘successor states’, the Artuqids and the Zengids? 2. What was the prevailing religious atmosphere in this period in the Jazira? 3. What can be gleaned from the sources about the cultural and economic life of the Jazira in this period? The Political Orientation of the Seljuq ‘Successor States’, the Artuqids and the Zengids Traditionally, Seljuq decline has been dated to the death of Sultan Muhammad in 1118. Thereafter, fratricidal rivalries, combined with the ambitions of Turcoman chiefs, especially in the more outlying provinces of the Seljuq empire, accelerated the process by which Seljuq territories became divided up and ruled in an independent or semi-­independent fashion. The late Seljuq empire was governed increasingly in two halves. One was dominated in the first half of the twelfth century by Sultan Sanjar whose power was centred in Khurasan; the other staggered precariously on in western Iran and Iraq. In the Jazira, the Artuqids of Diyar Bakr and the Zengids of Mosul were able to exploit the weakness of the Seljuqs to consolidate their own hold on their territories, but it is noteworthy that in many of their activities and, above all, in their religious and cultural orientation, the Artuqids and Zengids still looked east, to Baghdad and beyond. This ‘umbilical cord’ to the Great Seljuqs was of course not always tightly drawn, nor was it always intact, but the Artuqids and the Zengids can confidently be called ‘Seljuq successor states’. Both İl-­Ghazi and Zengi economised their efforts in the cause of fighting the Crusaders in Syria because of their political involvement in Iraq in the continuing power struggles between the Seljuq sultans and the ‘Abbasid caliphs.7 Moreover, the administration which evolved at Mardin and Mosul was a microcosm of the Great Seljuq system of Persia. Of course, there is a lack of clear definition in the terminology used in the sources to denote government posts, and it could be argued that in certain areas a given function might denote different duties. But there is no reason to believe that a wazir or a hajib or a treasurer did not carry out broadly similar functions at Mardin or Mosul to those they performed at Hamadan or Isfahan.8

the hi stor y o f the j a zir a  | 79 In these disturbed times, similarity of practice would have been encouraged by the fact, constantly alluded to in the sources, that high government officials were often peripatetic, not always of course from choice. This was after all the period par excellence of musadarat, the process by which a high official was mulcted for his ill-­gotten gains before being dismissed from his post. If such officials survived the tortures inflicted on them, they would often move on to a neighbouring ruler to offer their services to him. This practice is mentioned a number of times by Ibn al-­Azraq al-­Fariqi when writing about the twelfth-­century Jazira.9 He himself of course had been a peripatetic administrator too, serving at one time the king of Georgia amongst others.10 The same weaknesses which undermined the Seljuq government of Persia recur in microcosm in Artuqid and Zengid rule. Familiar centrifugal forces, such as the divisive system of inheritance, undermined the security of the Jazira upon the death of a strong ruler.11 What of the Prevailing Religious Atmosphere in the Jazira in the Period 1100–1250? Much has been made of the Sunni revival which took place in the Islamic world in the eleventh and twelfth centuries. Scholars such as Maqdisi have paved the way in this field, but many questions about this phenomenon still remain unanswered.12 The revival of Sunnism will be discussed here in the context of the twelfth century and with particular reference to the Artuqids, Zengids and Ayyubids. There is no reason to doubt that Maqdisi is right in placing considerable importance on Hanbalite circles in Baghdad for the revitalisation of Sunni Islam which took place as early as Buyid times.13 But certainly the whole process seems to have gained momentum with the coming of the Seljuq Turks into the Islamic world. Three separate factors in the twelfth century seem to combine to give Sunni Islam a much-­needed shot in the arm. 1. The indigenous influence of Hanbalite and other pietist circles of Sunni ‘ulama’, who in the face of the Shi‘ite threat both inside the Seljuq territories from the Isma‘ilis and outside from the Fatimids felt the need to redefine and sharpen their faith.14 2. The influx of the Seljuq Turkish leaders, recently converted to Islam, who

80  |  the m ed ie va l tur k s perhaps did not understand the dogmatic intricacies of kalam but made up for that lack of theological expertise by fanatical ardour and desire to prove themselves as good Muslims as anyone else. 3. The external threat of the Crusaders, many of whom came to the Levant with a brand of Christian fanaticism unknown to Eastern Christians and whose presence was eventually destined to help reactivate and redefine the concept of jihad amongst the Muslims in closest contact with them. The obvious question to ask here is how far the Artuqids and the Zengids of the Jazira became influenced by this increased religious awareness, indeed fanaticism, which mounted in fervour as the twelfth century proceeded and which reached its climax in the career of Saladin. To take the Artuqids first, it is clear from their titulature (notably in the use of terms such as ghazi and alpı) that the concept of jihad was known to them; but it is doubtful whether in the career of Najm al-­Din İl-­Ghazi, who set up the Artuqid dynasty, religion really played a genuine role in his political activities. As already mentioned, he was involved as much in Seljuq politics at Baghdad as he was in campaigns against the Crusaders or the Christian king of Georgia, and he seems to have operated a flexible approach both to religious groups and to political factions.15 His son Temürtash, who succeeded in maintaining his hold over Diyar Bakr for more than thirty years, seems to have been of a pacific disposition, and little is indicated in the sources about his religious attitude.16 On the other hand, the increasingly genuine fervour for jihad which emanated from nearby northern Syria under the Zengid Nur al-­Din seems to have infiltrated the Jazira as the century progressed.17 There are, for example, references in Michael the Syrian’s work to how the Artuqid Qara Arslan forbade the construction of new churches and even to his refusal to repair old churches.18 Moreover, in 1172 the main Jacobite church at Mardin was transformed into a mosque and the Church of the Forty Martyrs was plundered.19 In a predominantly Christian area such as the Jazira, this more aggressive policy must have been a reflection of much greater religious fanaticism on the part of Muslim overlords. As for the Zengids, their path to ‘genuine’ religious conviction about jihad and their contribution to the Muslim reunification and reconquest of Crusader territories have been studied lucidly and convincingly by Sivan.

the hi stor y o f the j a zir a  | 81 Sivan doubts whether Zengi was anything other than a Realpolitiker. Indeed, had Zengi been less embroiled in political activities in Baghdad he would probably have been more successful against the Crusaders.20 While Zengi’s son, the great Nur al-­Din, may well not always have been motivated by genuine religious belief, certainly by the end of his career his outward stance of piety and jihad seems to have been backed up by real religious fervour from within, as indeed was the case, it would appear, with Saladin.21 After Saladin’s death and the anticlimax which followed the recapture of Jerusalem, there was of course under the Ayyubids a let-­up in religious zeal, an inevitable slackening-­off of effort which was to prevail generally until the coming of the Mamluk dynasty in 1250.22 So the twelfth and the early thirteenth centuries saw a rising curve in publicly expressed religious fervour on the part of political leaders in the Jazira, a curve which reached its high point in the careers of Nur al-­Din and Saladin. From the point of view of actual religious practice, the leaders of the Seljuq successor states under discussion went through the motions of religious conviction publicly, whether or not they believed sincerely in what they were doing. This applies equally to the Artuqids, Zengids, Ayyubids and even Badr al-­Din Lu’lu’. At minimum, a Turcoman dynast built a madrasa, endowed a waqf and ordered his own dynastic mausoleum so that he and his family would be remembered. It did not matter how great or insignificant such a ruler was. Time after time, the sources record in the obituaries of petty Turcoman dynasts that they had been involved in some activity concerned with the building of madrasas, mosques and mausolea. No wonder, since such a fine example had been set by the Seljuq sultan Tughril to whom Ibn Khallikan attributes the following sentiments: ‘I should be ashamed to appear before God, were I to build for myself a dwelling and not erect a mosque beside it.’23 It was not only Nur al-­Din who frequented the ‘ulama’ and the Sufis and enjoyed discussion with them, for Artuqid and other Zengid rulers also apparently honoured and encouraged such members of the religious classes. If one can believe the pro-­Artuqid account of Ibn al-­Azraq, as early as the reign of the second Artuqid ruler, Temürtash, religious leaders were positively welcomed at Mardin.24 Towards the end of his reign he received

82  |  the m ed ie va l tur k s a ceremonial robe of honour from the caliph at Baghdad, and Temürtash is recorded as having built a library in which he placed religious books and established a waqf on it.25 There are many references in this period to travelling religious figures who spread their ideas, usually from Baghdad to the provinces. Ibn Khallikan records, for example, how al-­Suhrawardi went to Mosul and Damascus and was warmly received by Nur al-­Din.26 There is remarkable consistency in the titulature of Turcoman rulers of the Jazira in the period 1100–1250, with terms like ghazi and mujahid predominating. The reality of the jihad, the defence of the frontiers and the fight against the infidel, must have become somewhat remote to someone of the diplomatic skills of Badr al-­Din Lu’lu’. Nevertheless, he is still recorded in the various extant inscriptions in his name as having titles such as mujahid, murabit, muthagir, qatil al-kufrah wa’l-mushrikin and so on.27 No discussion of the religious atmosphere of the Jazira in the period 1100–1250 would be complete without a mention of the Isma‘ilis. The difficulty here, of course, is one of finding reliable, that is, unprejudiced, sources. It would appear, however, that the Isma‘ili presence was a constant source of danger to many Sunni rulers in northern Syria and the Jazira between 1100 and 1250. Isma‘ili activity is reported to have been directed at certain times towards the early Artuqid rulers, such as İl-­Ghazi and Temürtash, who remained generally unaffected or pragmatic in their approach to the Isma‘ilis. Jihad was encouraged from caliphal circles against the Isma‘ilis. Theologians with sound views and scholarly reputations would travel easily from one seat of learning to another, receiving an enthusiastic welcome. As already mentioned, the Hanbalites were strong not only in Baghdad but also in Damascus, thanks to prominent families such as the Banu Qudama.28 Cultural Life in the Jazira in the Period 1100–1250 Here there is a problem with the sources, which are predictably silent about major trends which must have been making themselves felt in this period. For example, the Muslim sources of course say nothing about the relationship between the Muslim minority in the Jazira, namely the Turcoman chiefs and their tribal followers, and the Christian majority who already lived in the area. Nor are the Christian sources very explicit on this issue. Questions such as the rate of conversion cannot therefore be answered.29

the hi stor y o f the j a zir a  | 83 The Muslim Turcoman rulers in the Jazira seem to have had a practical attitude to the responsibilities of ruling their predominantly Christian subjects. Taxation was organised by and the administration left to local families who had the expertise necessary, such as the Banu Nubata of Mayyafariqin. In their turn, the Artuqid and Zengid rulers, as well as building monuments with obvious religious functions, mended bridges and repaired roads, motivated no doubt by practical considerations. Again, the pattern is one set by the Seljuqs. Ibn Khallikan says of Malikshah: Successful in war, Malikshah also worked zealously to spread the benefits of civilisation: he dug numerous canals, walled a great number of cities, built bridges and constructed ribats in the desert.30

It is not easy to assess, however, to what extent one may see the Artuqid and Zengid rulers as patrons of the arts in general. To be sure, it is a very attractive hypothesis that a series of small dynasties in a given geographical area should stimulate artistic activity and vie with each other in patronage. This hypothesis certainly holds good for the welter of small dynasties which grew up in post-­Mongol Persia. It is true that, in the case of twelfth-­century Jazira, the Turcoman rulers there reveal themselves to be indefatigable builders, and the atmosphere was conducive to easy travel for craftsmen from one centre of power to another in the area. The phrase ‘centre of power’ is used here advisedly, since it would surely be inappropriate to apply the term ‘court’ to the rudimentary ‘palace’ life of early Artuqid times, say 1100–50. The Artuqid rulers themselves would lodge within the citadels at Mardin, Mayyafariqin and Hisn Kayfa, but it is clear that, either because their nomadic tribesmen were not wanted in the city, or because of personal preference, the rank and file of the nomads remained outside the city walls. Ibn al-­Azraq speaks of the hardships endured by the citizens of Mayyafariqin when Turcomans came into the ruins of the city and set up tents there. He also speaks of the heavy burden of billeting soldiers in the houses of the inhabitants of the city.31 There must have been severe tension between nomads and city-­dwellers, if certain anecdotes told by Ibn al-­Athir are to be believed.32 It is very difficult to find a satisfactory definition of ‘court’ which would be relevant in this context. However, a working definition might be ‘a group

84  |  the m ed ie va l tur k s of people following a predictable routine in the same place or places over a prolonged period of time’. If such a definition is applied to the Artuqids, it is not likely that their early leaders had such an ‘institution’. Of course, it is easier to envisage ‘court’ life in some areas than in others. Zengid Mosul or Damascus is a much more likely place to have exercised such pretensions than Artuqid Mardin. On the other hand, many of the rulers in this period were away on military campaigns for much of the year and would have had little time for cultural pursuits. It is inevitable that the energies of twelfth- and thirteenth-­century rulers should be directed to practical matters such as the mending of roads and bridges. The probable absence of a proper ‘court’, whatever that means, does not, however, preclude the likelihood that relatively stable rule in the hands of one dynasty controlling a specific area, such as the Artuqids in Diyar Bakr or the Zengids in Mosul, caused these territories to flourish. Ibn Shaddad in his historical geography paints, for example, a very glowing picture of Hisn Kayfa in the Artuqid period: The citadel had within its enclosure a green maidan and fields where enough wheat, barley and grain were grown to feed the inhabitants from year to year. In the rabad to the north were bazaars, caravanserais, madrasas, hammams and tombs.33

The preceding discussion has concentrated on the early Artuqids. Clearly, as the twelfth century progressed, Artuqid power, although considerably eroded first by the Zengids and then by the Ayyubids, remained stable within the limited confines of Diyar Bakr. What is not clear, however, is the point at which the term ‘court’ may reasonably be applied to the later Artuqids. The evidence which points to the production of manuscript painting and highly intricate automata in the Artuqid domains in the decades around 1200, both of them crafts implying established workshops, clearly indicates that courts of some kind did exist in this period, courts with enough prestige to attract people of the calibre of al-­Jazari and Usama b. Munqidh.34 Further research on the detailed history of the Jazira in the second half of the twelfth century should throw some light on the process by which the Artuqid Turcoman rulers changed from nomadic leaders to petty dynasts with an interest in fostering science and the arts.

the hi stor y o f the j a zir a  | 85 What happened in the countryside of the Jazira in this period is alluded to only rarely in the sources. But it would appear that the population of one area were on occasion taken away and forced to cultivate the land elsewhere. Such moves are recorded for two Artuqid leaders, Da’ud and Balak.35 References abound in the first half of the twelfth century to Turcomans pillaging the countryside and sometimes their leaders attempted to restrain them, but tribal movements in this period are so poorly documented that generalisations are impossible. Conclusion This discussion has tried to show that the political orientation of the immediate successor states to the Seljuqs, namely the Artuqids and the Zengids, was still decisively to the east. In the religious sphere, the period witnessed a gradual crescendo of religious fervour stimulated by the Crusaders, which reached a climax under Nur al-­Din and Saladin. Throughout the period 1100–1250 the rulers cultivated their public image as impeccably orthodox Muslims, an inheritance which they had received from the Great Seljuqs. The hostility felt towards the Isma‘ilis by the Artuqids and Zengids was a legacy from the same source. In the cultural domain the rulers of the Seljuq successor states maintained a policy of building widely for the public good, with a particular emphasis on religious structures. The evidence of the sources is explicit in indicating that it was architecture rather than the other arts which most readily attracted their patronage. The semi-­nomadic lifestyle of these rulers and their seemingly endless military campaigns often kept them away from their home base for much of the year. In such a situation the concept of a court is markedly less appropriate than, say, for Mamluk Cairo or ‘Abbasid Baghdad. By the end of the twelfth century, however, it would appear that courts of some kind did exist in the Artuqid territories of the Jazira as well as at Zengid Mosul. Notes   1. The detailed dynastic history of the Jazira between 1100 and 1250 may be pieced together from a variety of sources. These include Bosworth (1968); Cahen (1935); Hillenbrand (1979); Turan (1973); Elisséeff (1967); Cahen (1979a); Cahen (1979b).

86  |  the m ed ie va l tur k s   2. Cf. Canard (1965); Le Strange (1905), 86–114.   3. For the career of İl-­Ghazi, cf. Cahen (1935), 227–40; Hillenbrand (1981a); Sevim (1962).   4. Cf. Cahen (1935) and (1979a); Hillenbrand (1981b).   5. Cf. Cahen (1979b).   6. For some information on the reign of the caliph al-­Muqtafi, cf. Mason (1972), 13–66. For a detailed analysis of the caliph al-­Nasir, cf. Hartmann (1975).   7. For İl-­Ghazi, cf. Hillenbrand (1981a), 256–9; for Zengi, cf. Ibn al-­Athir (1963), 22–66.   8. For an analysis of Seljuq administration, cf. Lambton (1968); Horst (1964); Klausner (1973); Uzunçarşılı (1941). For a discussion of the functions of a provincial department of finance in Seljuq times, cf. Horst (1964), 51.   9. For several examples, cf. Ibn al-­Azraq, f. 164b. 10. Ibn al-­Azraq, f. 161b. 11. For one such example, cf. Ibn al-­Azraq, ff. 162b–163a. 12. Maqdisi (1963). 13. The formal reading of the Qadiriyya creed in palaces and mosques during the eleventh century was surely Hanbalite-­inspired. The first such reading is recorded for 1018. It is known that the creed was read publicly again in 1041 and 1055. Such a practice helped to strengthen Sunni views among the masses. Cf. Laoust (1979). 14. For an account of the Isma‘ilis, cf. Hodgson (1955). Ibn Hubayra, vizier to two caliphs, al-­Muqtafi and al-­Mustanjid, founded a madrasa in Baghdad to teach Hanbalite fiqh and in this he was typical of prominent figures of the time. Hanbalism was also strong in the provinces, especially in Damascus through the influence of Abu’l-­Faraj al-­Shirazi and the Banu Qudama. Cf. Laoust (1979). 15. Hillenbrand (1981a), 282–92. 16. Hillenbrand (1981b), 150. 17. Sivan (1968), 59–91. 18. Michael the Syrian (1899–1914), 307. 19. Michael the Syrian (1899–1914), 337, 340, 347, 350. 20. Sivan (1968), 44–5. 21. For a detailed bibliography on Saladin, cf. Lyons and Jackson (1982), 435–43. 22. Cf. Cahen (1979b). 23. Ibn Khallikan (1843–71), III, 227. 24. Ibn al-­Azraq, ff. 171a–b. 25. Ibn al-­Azraq, f. 171a.

the hi stor y o f the j a zir a  | 87 26. Ibn Khallikan (1843–71), II, 151. 27. Van Berchem (1978), 662. 28. Cf. n. 14. 29. For a general discussion of this question, cf. Vryonis (1971), 155–94. 30. Ibn Khallikan (1843–71), III, 441. 31. Ibn al-­Azraq, f. 161a. 32. Ibn al-­Athir (1864), 231–2. 33. Ibn Shaddad, f. 127a. 34. Hill (1974); Hitti (1964). 35. For Balak, cf. Michael the Syrian (1899–1914), 205–6; Ibn al-­Azraq, ff. 168a and 169a.

Bibliography Bosworth 1968: Bosworth, C. E., ‘The political and dynastic history of the Iranian world (A.D. 1000–1217)’ in J. A. Boyle (ed.), The Cambridge History of Iran, V: The Seljuq and Mongol Periods (Cambridge, 1968), 1–202. Cahen 1935: Cahen, C., ‘Le Diyar Bakr au temps des premiers Urtukides’, Journal Asiatique 227 (1935), 219–76. Cahen 1979a: Cahen, C., ‘Artukids’, Encyclopaedia of Islam, 2nd edn. Cahen 1979b: Cahen, C., ‘Ayyubids’, EI 2. Canard 1965: Canard, M., ‘Al-­Djazira’, EI 2. Elisséeff 1967: Elisséeff, N., Nur ad-Din. Un grand prince musulman de Syrie au temps des Croisades, 3 vols (Damascus, 1967). Hartmann 1975: Hartmann, A., An-Nasir li-Din Allah (1180–1225): Politik, Religion, Kultur in der späten ‘Abbasidenzeit (Berlin and New York, 1975). Hill 1974: Ibn al-­Razzaz al-­Jazari, The Book of Knowledge of Ingenious Mechanical Devices (Kitab fi ma‘rifat al-hiyal al-handasiyya), translated and annotated by D. A. Hill (Dordrecht-­Boston, 1974). Hillenbrand 1979: Hillenbrand, C., The History of the Jazira 1100–1150: The Contribution of Ibn al-Azraq al-Fariqi, 2 vols, unpublished PhD thesis (University of Edinburgh, 1979). Hillenbrand 1981a: Hillenbrand, C., ‘The career of Najm al-­Din İl-­Ghazi’, Der Islam 58/2 (1981), 250–92. Hillenbrand 1981b: Hillenbrand, C., ‘The establishment of Artuqid power in Diyar Bakr in the twelfth century’, Studia Islamica 54 (1981), 29–53. Hitti 1964: Hitti, P. K., Memoirs of an Arab-Syrian Gentleman (Beirut, 1964). Hodgson 1955: Hodgson, M. G. S., The Order of the Assassins (The Hague, 1955).

88  |  the m ed ie va l tur k s Horst 1964: Horst, H., Die Staatsverwaltung der Grosselg˘uqen und Hōrazmšahs 1038–1231. Eine Untersuchung nach Urkundenformularen der Zeit (Wiesbaden, 1964). Ibn al-­Athir 1864: Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, X, ed. C. J. Tornberg (Leiden and Uppsala, 1864). Ibn al-­Athir 1963: Ibn al-­Athir, Ta’rikh al-dawla al-atabakiyya: muluk al-Mausil, ed. A. Tulaimat (Cairo, 1963). Ibn al-­Azraq: Ibn al-­Azraq, Ta’rikh Mayyafariqin wa-Amid, B.M. Or. 5803. Ibn Khallikan 1843–71: Ibn Khallikan, Kitab wafayat al-a‘yan: Ibn Khallikan’s Biographical Dictionary, 4 vols, tr. Baron W. M. de Slane (Paris, 1843–71). Ibn Shaddad: Ibn Shaddad, Al-a‘laq al-khatira fi dhikr umara’ al-Sham wa’l-Jazira, Bodleian Ms. Marsh 333. Klausner 1973: Klausner, C. L., The Seljuk Vezirate: A Study of Civil Administration 1055–1194 (Cambridge, MA, 1973). Lambton 1968: Lambton, A. K. S., ‘The internal structure of the Seljuq empire’, in J. A. Boyle (ed.), The Cambridge History of Iran, V: The Seljuq and Mongol Periods (Cambridge, 1968), 203–82. Laoust 1979: Laoust, H., ‘Hanabila’, EI 2. Le Strange 1905: Le Strange, G., The Lands of the Eastern Caliphate (Cambridge, 1905). Lyons and Jackson 1982: Lyons, M. C. and Jackson, D. E. P., Saladin: The Politics of the Holy War (Cambridge, 1982). Maqdisi 1963: Maqdisi, G., Ibn ‘Aqil et la résurgence de l’Islam traditionaliste au XIe siècle (Damascus, 1963). Mason 1972: Mason, H., Two Statesmen of Mediaeval Islam (The Hague and Paris, 1972). Michael the Syrian 1899–1914: Michael the Syrian, Chronique de Michel le Syrien, ed. and tr. J.-B. Chabot (Paris, 1899–1914). Sevim 1962: Sevim, A., ‘Artuk Oğlu Ilgazi’, Belleten 26/101 (1962), 649–91. Sivan 1968: Sivan, E., L’Islam et la croisade. Idéologie et propagande dans les réactions musulmanes aux croisades (Paris, 1968). Turan 1973: Turan, O., Doğu Anadolu Türk devletleri tarihi (Istanbul, 1973). Uzunçarşılı 1941: Uzunçarşılı, I. H., Osmanli devleti teşkilatina medhal (Istanbul, 1941). Van Berchem 1978: Van Berchem, M., Opera minora (Geneva, 1978), II, 659–72. Vryonis 1971: Vryonis, S. Jr, The Decline of Medieval Hellenism in Asia Minor and the Process of Islamization from the Eleventh through the Fifteenth Century (Berkeley, Los Angeles and London, 1971).

4 Malazgird

T

he battle. The most important event with which the name of the town is connected is the battle of Manzikert fought in Dhu’l-­Qa‘da 463/August 1071 between Alp Arslan and Romanus IV Diogenes. This event is treated by a variety of sources, Byzantine Greek, Armenian, Syriac and Arabic. The most valuable account is surely that of Attaliates, who was present at the battle itself as well as being an adviser of the emperor, whereas that of Psellus, tutor of Michael VII Ducas who was to replace Romanus on the imperial throne after the battle (see below), is hostile. Although Cahen (1934, see Bibl.) was critical of Attaliates’ detailed testimony, more recently Vryonis and Cheynet (see Bibl.) have reinstated him as the prime source for the battle. A relevant Western source is the Gesta Roberti Wiscardi of William of Apulia (see Bibl.); other late Christian writers such as Michael the Syrian and Matthew of Edessa are strongly anti-­Byzantine, viewing the Seljuq invasions of Anatolia as divine retribution for the emperor’s treatment of non-­ Melkite religious minorities of the empire. There is no contemporary Muslim account, the earliest extant one being that of Ibn al-­Qalanisi (d. 555/1160), but the long description by Sibt b. al-­Jawzi (probably deriving from Ghars al-­ Ni‘ma Muhammad b. Hilal al-­Sabi’’s lost ‘Uyun al-tawarikh) is, among other Muslim sources, detailed and valuable. The policy adopted by Romanus when he became emperor in January 1068 was to take the offensive against the Muslim enemy beyond the Byzantine frontiers rather than to wait for raids to take place, and the campaign which culminated at Malazgird was the last of three conducted by the emperor himself, for which he left Constantinople in spring 463/1071, aiming to 89

90  |  the m ed ie va l tur k s securing against the Seljuqs the Armenian fortresses of Akhlat and Malazgird. Alp Arslan was for his part besieging Edessa in the spring, but on hearing of the arrival of the Byzantine army in the east, decided to move in that direction, probably via Mosul and Khuy, to assemble reinforcements. Romanus detached a contingent of his army under the Norman Roussel of Bailleul to take Akhlat, whilst he took and garrisoned Malazgird itself. Preliminary skirmishes took place, during which time some of the Uze (that is, Ghuzz) mercenaries in the Byzantine army deserted to the enemy, whilst Roussel and the Georgian Joseph Trachaniotes fled westwards from Akhlat, deserting the emperor, at the approach of Alp Arslan. The Muslim side offered peace, but Romanus refused any terms, feeling that he had numerical superiority and being unwilling to throw away the immense effort put behind his campaign. The Muslim sources emphasise Alp Arslan’s pessimism before the battle, but the fact that the battle took place on a Friday meant that the force of universal Muslim prayer was felt as an advantage (cf. al-­Husayni, Akhbar al-dawla al-saldjuqiyya, 47–9, which purports to give the text of special prayers offered up throughout the Sunniworld at the caliph al-­Qa’im’s orders). Although the figures in the Muslim sources for Romanus’s army (from 200,000 to 400,000) must be exaggerated, the emperor must, despite defections, have had superiority in numbers; Cheynet estimates his army at probably 60,000, with much baggage and impedimenta. Its morale, however, was not high, and its composition very heterogeneous; amongst foreign mercenaries are mentioned Franks, Arabs, Rus, Pechenegs, Georgians, Abkhazians, Khazars, Ghuzz, Qipčaq, Scyths, Alans and Armenians. Alp Arslan is generally credited in the Muslim sources with having 15,000 troops at the battle. The exact date of the battle has not hitherto been established with certainty, but the fact that it was a Friday in Dhu’l- Qa‘da seems to limit the possibilities to 20 Dhu’l- Qa‘da/19 August or the next week; in fact, astronomical indications, confirming Attaliates’ information that the night before the battle was moonless, point to 27 Dhu’l-­Qa‘da 463/26 August 1071, as is shown in a recent popular book on the battle, A. Friendly, The Dreadful Day, London 1981, 178. The exact location of the field of battle is likewise uncertain, though it was along the road between Malazgird and Akhlat; the al-­Rahwa (cf. Yaqut, ii, 880) of Sibt b. al-­Jawzi, Ibn al-­Jawzi and Ibn al-‘Adim seems most probable.

ma l a zgi rd  | 91 The course of the battle is described in most detail by Attaliates and by the anti-­Romanus, later writer Nicephorus Bryennius. A Byzantine return to camp at nightfall seems to have been interpreted as a retreat; the rearguard under Andronicus Ducas left the field, leaving the army’s rear unprotected; and in the later stages, the Seljuq forces lured the Greeks into ambushes. The emperor was captured and was honourably treated by Alp Arslan; several sources record the famous conversation in which the sultan asked Romanus what treatment should be meted out to him. A peace agreement was drawn up, the precise terms of which are not known, but which probably included a ransom, the cession of various frontier fortresses, and the provision of troops and annual tribute to the sultan; but since during Romanus’s brief captivity, Michael VII Ducas had been proclaimed emperor and Romanus was eventually blinded and killed by his supplanter (August 1072), it is likely that these terms were never put into force anyway. Romanus’s defeat seems to have sprung in part from inadequate intelligence about the movements of the Akhlat force, which was to rejoin him, and a poor choice of terrain, one favourable to the Seljuqs’ mounted archers; but internal dissensions within the Byzantine empire, moreover, had been reflected in the army itself, for Andronicus Ducas, cousin of the future emperor Michael VII, had been ill-­disposed towards Romanus. A Fatimid involvement in the campaign has recently been suggested by Hamdani (see Bibl.), but this remains speculative. Was the battle indeed ‘the greatest disaster of Byzantine history’ (Grousset)? In many ways, it was the decade of internecine strife within the empire after the battle which harmed the empire more and allowed Turks to infiltrate Byzantine territory. Byzantine prestige abroad was certainly harmed by the ignominy of Romanus’s capture, and some Crusader chroniclers (for example, William of Tyre) see Western European involvement in the Levant as dating from this time, with the Franks replacing the Greeks as upholders of Christianity against Islam there. The Muslim historians, for their part, tend to over-­dramatise the event, probably because it was the only major military confrontation during the infiltration process.

92  |  the m ed ie va l tur k s

Bibliography Primary Sources Arabic Ibn al-‘Adim, Zubdat al-halab min ta’rikh Halab, ed. S. Dahan, Damascus 1954, ii, 23–30. Ibn al-­Athir, al-Kamil fi’l-ta’rikh, x, ed. C. J. Tornberg, Leiden and Uppsala 1864, 43–6. Ibn al-­Azraq al-­Fariqi, Ta’rikh al-Fariqi, ed, B. A. L. ‘Awad, Cairo 1959, 186–90. al-­Bundari, Zubdat al-nusra wa-nukhbat al-‘usra, ed. M. T. Houtsma, Leiden 1889, ii, 36–44. al-­Husayni, Akhbar al-dawla al-saljuqiyya, ed. M. Iqbal, Lahore 1933, 46–53. al-­Makin b. al-‘Amid, Historia saracenica, ed. T. Erpenius, Leiden 1625, 555–6. Ibn al-­Jawzi, al-Muntazam, ed. S. Zakkar, Beirut 1995–6, viii, 260–5. Ibn al-­Qalanisi, Dhayl ta’rikh Dimashq, ed. H. F. Amedroz, Leiden 1908, 99. Sibt al-­Jawzi, Mir’at al-zaman, apud Ibn al-­Qalanisi, in op. cit., 100–5. Persian Mirkhwand, Rawdat al-safa, Tehran 1853–4. Zahir al-­Din Nishapuri, Saljuq-nama, Tehran 1332, 24–7. Eastern Christian Aristakes of Lastiverd, Récit des malheurs de la nation arménienne, eds M. Canard and H. Berberian, Brussels 1973, 124–8. Bar Hebraeus, The chronography of Gregory Abu’l-Faraj .  .  ., ed. and tr. E.  A.  W.  Budge, London 1932, 220–2. Matthew of Edessa, Patmut’iwn, tr. E. Dulaurier, Chronique de Mattieu d’Edesse (962–1136) . . ., Paris 1858, 163–70. Michael the Syrian, facs. ed. and tr. J.-B. Chabot, Chronique de Michel le Syrien, Paris 1899–1914, 168–70. Byzantine Attaliates, Historia, ed. I. Bekker, in Corpus scriptorum historiae Byzantinae, Bonn 1853, 144–69. Michael Psellus, Chronographia, tr. E. R. A. Sewter, The Chronographia of Michael Psellus, London 1953, 271–4.

ma l a zgi rd  | 93 Nicephorus Bryennius, Historia, tr. P. Gautier, Nicephore Bryenios Histoire, Brussels 1975, 104–20. Skylitzes, Ioannes Skylitzes continuatus, ed. E. T. Tsolakes, Thessalonika 1968. Zonaras, Ioannis Zonarae epitomae historiarum, ed. T. Büttner-­Wobst, in CSHB, Bonn 1897, 696–703. Western Christian William of Apulia, Gesta Roberti Wiscardi, in Monumenta Germaniae Historica Scriptorum 1851, ix, 239–98. Secondary Sources Cahen, C., La campagne de Mantzikert d’après les sources musulmanes, in Byzantion, ix (1934), 613–42. Cahen, C., Pre-Ottoman Turkey, London 1968, 26–30. Cheynet, J. C., Mantzikert: un désastre militaire?, in Byzantion, 1 (1980), 410–38. Grousset, R., Histoire de l’Arménie, Paris 1947, 624–30. Hamdani, A., A possible Fatimid background to the battle of Manzikert, in AÜDTCFD, vi (1968), 1–39. Hamdani, A., Byzantine–Fatimid relations before the battle of Manzikert, in Byzantine Studies, ii/2 (1974), 169–79. Honigmann, E., Die Ostgrenze des Byzantinischen Reiches, Brussels 1935, 189–90 and index Kafesoğlu, İ., İslam Ansiklopedisi, art. Malazgirt. Kaymaz, N., Malazgirt savaşı ile Anadolu’nun fethi ve türkleşmesine dair, in Malazgirt armağanı, Ankara 1972, 259–68. Laurent, J., Byzance et les Turcs Seldjoucides dans l’Asie occidentale jusqu’en 1081, Nancy 1913, 43–4. Mathieu, M., Une source négligée de la bataille de Mantzikert. Les ‘Gesta Roberti Wiscardi’ de Guillaume d’Apulie, in Byzantion, xx (1950), 89–103. Oman, C., A history of the art of war, London 1898, 216–21. Runciman, S., A history of the Crusades, Cambridge 1954, i, 62–5. Sevim, A., Malazgirt meydan savaşı ve sonuçları, in ibid., 219–30. Sümer, F., Malazgird savaşın katılan Türk beylerı, in Selçuklu Araştırmaları Dergisi, iv (1975), 197–207. Vryonis, Jr, S., The decline of medieval Hellenism in Asia Minor and the process of Islamization from the eleventh through the fifteenth century, Berkeley, etc. 1971, 96–104.

5 Marwanids

M

arwanids, a dynasty of Kurdish origin who, having ousted the Hamdanids, ruled Diyar Bakr from 380/990–1 to 478/1085. The founder of the dynasty, a Kurdish chief named Badh, seized the city of Mayyafariqin after the death of the Buyid ruler ‘Adud al-­Dawla (373/983), and then took Amid, Nasibin and Akhlat (Ibn al-­Athir, ix, 25; Ibn al-­ Azraq, 49–52). Badh successfully fended off attacks both from a Buyid army sent against him and from the Hamdanids but was killed by a coalition of Hamdanid and ‘Uqaylid forces after his unsuccessful attempt to take Mosul (380/990). The dynasty itself, however, takes its name not from Badh but from Marwan, a miller who had married Badh’s sister. It was their son Abu ‘Ali al-­Hasan b. Marwan who, having withdrawn after Badh’s death in 380/990 to Hisn Kayfa, married his uncle’s widow, routed the Hamdanids on two occasions and took possession of Mayyafariqin and Amid (Ibn al-­Azraq, 59–60; Ibn al-­Athir, ix, 50). After his murder at Amid in 387/997, his brother Mumahhid al-­Dawla Sa‘id ruled until 401/1011. These two precarious reigns paved the way for the accession of a third brother, Nasr al-­Dawla Ahmad, whose rule marks the apogee of Marwanid power. Nasr al-­Dawla was recognised as ruler of Diyar Bakr by the Buyid amir Sultan al-­Dawla, by the Fatimid caliph al-­Hakim, and by the Byzantine emperor, all of whom soon sent envoys and congratulatory messages to him (Ibn al-­Azraq, 103). Indeed, Nasr al-­Dawla in his long reign (401–83/1011– 61) was to practice a skilful policy of accommodation and self-­preservation with all three powers. He also had to contend with Bedouin Arab dynasties 94

ma r wa ni ds  | 95 such as the ‘Uqaylids and the Mirdasids, who wielded power in northern Syria and the Jazira, and to whom he was forced to cede Nasibin and Edessa respectively. The sixth/twelfth-­century chronicler of the Jazira, Ibn al-­Azraq al-­Fariqi, gives in his chronicle a very full account of Marwanid rule. Nasr al-­Dawla was fortunate to have the services of two capable viziers, Abu’l-­Kasim al-­ Husayn al-­Maghribi, who died in office (428/1037), and whose biography is given by Ibn Khallikan and the even more famous Fakhr al-­Dawla Ibn Jahir. Under Nasr al-­Dawla, Diyar Bakr enjoyed a high level of stability and commercial and cultural prosperity. The Marwanid court at Mayyafariqin was frequented by prominent ‘ulama’ and poets, such as the Shafi‘i ‘alim ‘Abd Allah al-­Kazaruni (d. 455/1063) (Ibn al-­Athir, ix, 52) and the poet al-­Tihami (d. 416/1025–6) (Ibn al-­Azraq, 82). Nasir-­i Khusraw visited Mayyafariqin in 438/1046 and was much impressed by it (Safar-nama, ed. Muhammad Dabir-­Siyaqi, Tehran 1335/1956, 8–11). Nasr al-­Dawla emerges as a flamboyant ruler with political acumen and extravagant tastes. His religious stance appears to have been a pragmatic one, suitable for the ruler of a vulnerable buffer state surrounded by greater powers of the most divergent confessional loyalties. It seems likely that he ruled a predominantly Christian population in the towns of Diyar Bakr and that he enjoyed a good relationship with Byzantium. Indeed, the emperor Constantine X asked him for help in procuring the release of the Georgian prince Liparit from the Seljuq sultan Tughril (Ibn al-­Athir, ix, 372–3). It is probable that Nasr al-­Dawla was persuaded for a short while from 430/1038–9 to give the khutba in favour of the Fatimid al-­Mustansir (Ibn Khaldun, ‘Ibar, iv, 318), but it is also noteworthy that in that same reign, ‘Abd Allah al-­Kazaruni went to Mayyafariqin and spread the Shafi‘i madhhab throughout Diyar Bakr (Ibn al-­Athir, ix, 52). In traditional fashion, Nasr al-­Dawla is praised for strengthening the frontiers and for building bridges and citadels, and these laudatory statements of Ibn al-­Azraq are confirmed by the evidence of Marwanid inscriptions found on the walls of Amid. Indeed, according to the evidence of an inscription dated 445/1053–4 on a marble slab in the Bab Hitta in Jerusalem, Nasr al-­Dawla was also responsible for establishing two houses for the use of pilgrims there (Burgoyne, 118–21). The sources comment on the immense

96  |  the m ed ie va l tur k s wealth accumulated by Nasr al-­Dawla. He is also said to have possessed 360 concubines who did not, however, prevent him from meticulous observance of the morning prayer. He was interested in gastronomical pleasures too, and sent his cooks to Egypt to learn the culinary arts of that country (Ibn al-­Athir, x, 11). When the Seljuq sultan Tughril advanced into Diyar Bakr (448/1056–7), he did not aim at abolishing the Marwanid state, so Nasr al-­Dawla recognised his suzerainty and kept his lands. Tughril wrote to him confirming his role as a frontier lord fighting the infidels and exhorting him to continue in this task (Ibn al-­Athir, ix, 275). On the death of Nasr al-­Dawla (453/1061), the power and prestige of the dynasty declined markedly. His son Nizam al-­Din Nasr succeeded him, at first only in Mayyafariqin and then two years later (having overcome his brother Sa‘id) in Amid too. On the death of Nizam al-­Din (472/1079) his son Nasir al-­Dawla Mansur, the last Marwanid ruler, came to power. The vizier Ibn Jahir, who had left Diyar Bakr for Baghdad, used his influence with Malikshah and Nizam al-­Mulk to persuade them to bring the Marwanid dynasty to an end and to seize their treasures. In 478/1085 Diyar Bakr fell to Ibn Jahir and direct Seljuq control was imposed (Ibn al-­Athir, x, 93–4). Ibn Jahir took their treasury for himself and the last Marwanid ruler Mansur was given Jazirat Ibn ‘Umar, where he lived on until 489/1096. Bibliography Primary Sources Ibn al-­Azraq al-­Fariqi, Ta’rikh Mayyafariqin wa-Amid, ed. B. A. L. Awad, Cairo 1959, passim. Ibn al-­Athir, al-Kamil fi’l-ta’rikh, ed. C. J. Tornberg, Leiden and Uppsala 1864, ix, 25, 49–52, 272–6, 372–3, 416; x, 11, 86, 93, 151, 174 Ibn Khaldun, Kitab al-i‘bar, ed. W. M. de Slane, Cairo 1847, iv, 315–21. Ibn Khallikan, Wafayat al-a‘yan, tr. W. M. de Slane, Paris 1843–71. Secondary Sources Amedroz, H. F., The Marwanid dynasty at Mayyafariqin in the tenth and eleventh centuries a.d., in Journal of the Royal Asiatic Society (1903), 123–54. van Berchem M. and J. Strzygowski, Amida, Heidelberg and Paris, 1910, 22–37.

ma r wa ni ds  | 97 Bosworth, C. E., in Cambridge history of Iran, Cambridge 1968, v, 24, 97–8. Burgoyne, M. H., A recently discovered Marwanid inscription in Jerusalem, in Levant, xiv (1982), 118–21. Gabriel, A., Voyages archéologiques dans la Turquie orientale, Paris 1940. Zambaur, E. K. M. von, Manuel, Salt Lake City 1982, 135. See also Encyclopaedia of Islam, 1st edn: Marwanids, Nasr al-­Dawla; and Encyclopaedia of Islam, 2nd edn: Diyar Bakr.

6 Mayyafariqin

The Islamic Period

T

he conquest and caliphal rule. In the wake of the conquest of the Jazira by ‘Iyad b. Ghanm, Mayyafariqin fell to him peacefully. The caliph ‘Umar b. al-­Khattab had made him governor of the Jazira in 18/639 (al-­ Baladhuri, Futuh, 179). From that time until the early ‘Abbasid period, the city was ruled as part of the Jazira, sometimes jointly with al-­Sham and on other occasions with Armenia and Azerbaijan. The names of individual governors of Mayyafariqin for this period are listed by the town chronicler Ibn al-­Azraq al-­Fariqi and copied by ‘Izz al-­Din Ibn Shaddad, the seventh/ thirteenth-­century geographer of the Jazira. During the reign of the ‘Abbasid caliph al-­Muhtadi (255–6/869–70), Mayyafariqin and Amid were seized by the Shaybanid ‘Isa b. al-­Shaykh (Ibn Shaddad, ms. Oxford, Marsh 333, fol. 10a). The Shaybanids continued to govern the area until its reconquest by the caliph al-­Mu‘tadid in 286/899. The grandson of ‘Isa b. al-­Shaykh, Muhammad, built the minaret of the Friday mosque in Mayyafariqin in 270/883–4 or 273/886–7. His name was inscribed on it (ibid., fol. 69a). Hamdanid and Buyid involvement in Mayyafariqin. Mayyafariqin fell under the sway of the Taghlibi Arab family, the Hamdanids, after the appointment of Nasir al-­Dawla al-­Hasan as governor of Mosulin 324/935. His brother, Sayf al-­Dawla ‘Ali (d. 356/967), ruled Aleppo and Diyar Bakr and showed a particular liking for Mayyafariqin. Sayf al-­Dawla repaired its walls and rebuilt 98

m a yya f a ri qi n  | 99 the old citadel, where he stayed when visiting the city. He also provided Mayyafariqin with a proper water supply. The entourage of Sayf al-­Dawla at Mayyafariqin included the famous preacher Ibn Nubata, and al-­Mutanabbi. The latter recited an elegy over ‘Abd Allah b. Sayf al-­Dawla, who died in the town in 338/949 (Ibn al-­Azraq, B. L. Or. 5803, fol. 113b; Ibn Shaddad, fols 77a–78a). During the rule of Sayf al-­Dawla, the Jazira was under frequent attack from the Byzantines, whose territorial possessions extended at times almost as far as Amid. The future Byzantine emperor John Tzimisces besieged Mayyafariqin in 348/959, and it was on this occasion that Ibn Nubata began to deliver sermons exhorting the citizens to engage in jihad (al-­Antaki, 774–7; Ibn al-­Azraq, Or. 5803, fol. 114b). Thereafter Sayf al-­Dawla began to strengthen the fortifications of the city (Ibn Shaddad, fol. 78b). When Sayf al-­Dawla died in 367/967, he was buried in the Hamdanid family turba at Mayyafariqin (Ibn al-­Azraq, Or. 5803, fol. 117a; Ibn Shaddad, fol. 78a), a detail which reveals the high esteem in which the city was held by his line. Moreover, Mayyafariqin was the residence of his wife and children (al-­Antaki, 807). After the death of Sayf al-­Dawla, ­Mayyafariqin – ­along with the rest of Diyar ­Bakr – ­fell to the Hamdanid ruler of Mosul, Abu Taghlib al-­Ghadanfar. The sister of Sayf al-­Dawla stayed on in Mayyafariqin and in 362/972–3 completed the task of improving the defences of the city. This was probably occasioned by another siege of Mayyafariqin conducted by John Tzimisces, now elevated to the purple, in 361/972 (Ibn al-­Azraq, Or. 5804, fol. 118b). In 368/978–9 Abu’l-­Wafa’, the general of the Buyid ruler ‘Adud al-­ Dawla, took Mayyafariqin on his behalf (Ibn Miskawayh, ii, 388–90). The name of ‘Adud al-­Dawla was recorded on the city walls (Ibn Shaddad, fol. 69b). The Marwanid dynasty, 372–478/983–1085. After the death of ‘Adud al-­Dawla in 372/983, Badh, the founder of the Kurdish dynasty of the Marwanids, seized Mayyafariqin (Ibn al-­Athir, ix, 25; Ibn al-­Azraq, ed. ‘Awad, 49–52). Badh’s successors were able to hold on to the city, making it their capital for over a century. The Marwanid period witnessed another cultural flowering in Islamic Mayyafariqin. The second Marwanid ruler,

100  |  the me d ie va l tur k s Mumahhid al-­Dawla, repaired the city walls and inscribed his name on them in many places (ibid., 86, 163; Ibn Shaddad, fol. 70a). An inscription of his is illustrated by Lehmann-­Haupt (Armenien, 424). The greatest of the Marwanid rulers, Nasr al-­Dawla (ruled 401–53/1011– 61), was responsible for much building activity in the city, including a new citadel with gilded walls and ceilings which was completed in Dhu’l-­Hijja 403/June–July 1013 and which stood on a hill, the site of the Church of the Virgin. The Christian relics were transferred to the Melkite church (Ibn al-­Azraq, ed. ‘Awad, 107–8). Nasr al-­Dawla also restored the old observatory (manzara), put a clock (bankam) in the Friday mosque, constructed and endowed a hospital (bimaristan), planted the citadel garden and built bridges, public baths and a mosque in the suburb of al-­Muhaddatha (ibid., 123, 138, 141, 143, 145, 163–4, 168). The Marwanid capital attracted prominent religious figures (ibid., 82, 144, 166); from it, for example, ‘Abd Allah al-­Kazaruni spread the Shafi‘i madhhab in Diyar Bakr (Ibn al-­Athir, ix, 52). Shaykh Abu Nasr al-­Manazi, a high official at the time of Nasr al-­Dawla, collected books and established waqfs for libraries in the mosques of Mayyafariqin and Amid (Ibn al-­Azraq, ed. ‘Awad, 131). Nasr al-­Dawla died in 453/1061 and was buried in the turba of the Banu Marwan at Mayyafariqin (ibid., 177). After this, the Marwanids held on to Mayyafariqin until the town and the rest of Diyar Bakr were taken by the Seljuqs in 478/1085 during the campaign conducted by Ibn Jahir, the erstwhile vizier of the Marwanids. Ibn Jahir had persuaded the Seljuq sultan Malikshah to authorise him to besiege Mayyafariqin, and the vizier was able to carry off vast treasures belonging to the Marwanids (Ibn al-­Athir, x, 86–8, 93–4; Ibn al-­Azraq, ed. ‘Awad, 208–12). In the brief quarter-­ century following the death of Malikshah in 485/1092, Mayyafariqin changed hands many times and was ruled by a succession of Seljuq princes and other local rulers, including Tutush, Duqaq, Qilij Arslan and Suqman al-­Qutbi of Akhlat. The Artuqids. After the death of the Seljuq sultan Muhammad b. Malikshah in 512/1118, Mayyafariqin fell under the sway of the Turcoman Artuqids. According to its chronicler Ibn al-­Azraq, the town was seized in 512/1108–19

ma yya f a r iq in  | 101 by the Artuqid Najm al-­Din İl-­Ghazi, who had already taken Mardin around 502/1108–9 (ms. B.L., Or. 5803, fol. 161a). Ibn al-­Athir puts the Artuqid capture of Mayyafariqin three years later, in 515/1121–2 (x, 418), but this is one instance where the dating of the local historian is more likely to be accurate. After İl-­Ghazi’s death in 516/1122, his son Temürtash was able to hold on to Mardin and Mayyafariqin for thirty years and to withstand Zengi’s attempts to extend his sphere of influence in Diyar Bakr (Ibn al-­Azraq, Or. 5803, fols 169a, 171a; Ibn al-‘Adim, 271; Ibn al-­Athir, Atabegs, 79). Temürtash’s most ambitious project was the building of the Qaraman bridge over the Satidama river (the Batman-­Su) five miles east of Mayyafariqin. The work was begun in 541/1146–7 and was completed by his son Najm al-­Din Alpı in 548/1153–4. The stone arch of the bridge measured more than sixty spans and was ‘one of the marvels of the age’ (Ibn al-­Azraq, Or. 5803, fols 171b, 179b). The bridge is described fully by Gabriel (Voyages, 236), who notes that Sauvaget read the name Temürtash and the year 542/1147–8 on the bridge (ibid., 345). A copper mine was discovered in the time of Temürtash in the area north of Mayyafariqin (Ibn al-­Athir, x, 215) and it is noteworthy that Temürtash is known to have minted copper coins (Ibn al-­Azraq, Or. 5803, fol. 172b). The Artuqids held on to Mayyafariqin after the death of Temürtash in 548/1152; but, unlike the Hamdanids and Marwanids, they preferred generally to live at Mardin. Continuity in administration of Mayyafariqin was provided by the Nubata family, who are often mentioned as holding the office of qadi (ibid., fols 161a, 162b, 169b). The third Artuqid ruler of Mayyafariqin, Najm al-­Din Alpı, was responsible for a major reconstruction of the Friday mosque. The minbar and arcades of the mosque had collapsed in 547/1152–3, the last year of the reign of his father, Temürtash (ibid., fol. 175a). According to Ibn Shaddad, Najm al-­Din Alpı pulled down the rest of the building (fol. 104b) and it was rebuilt with substantial changes by the year 552/1157–8 (Ibn al-­Azraq, Or. 5803, fol. 175b). There is an inscription in the name of Najim al-­Din Alpı at the base of the dome (Gabriel, Voyages, 227). The Ayyubids. After its conquest by Salah al-­Din in 581/1185, the city walls were decorated with a fine commemorative inscription. This was discovered

102  |  the me d ie va l tur k s by Gertrude Bell and analysed by Van Berchem (in Diez, Baudenkmäler, 108) and by Flury (Schriftbänder, 44–8). It is apparently the only Kufic inscription in the name of Salah al-­Din. Minorsky (EI1, art: Mayyafariqin) stated on the authority of Gertrude Bell that this ruler built a mosque at Mayyafariqin for which the columns of the Byzantine basilica were used. There would appear to be no evidence in the sources for this. Possibly the mosque in question was the one outside the walls, of which only the Ayyubid minaret remains (Gabriel, Voyages, 210, 228). Salah al-­ Din entrusted Mayyafariqin to his brother Sayf al-­ Din in 591/1195 and the city was ruled by this branch of the Ayyubid family until the Mongols conquered the city in 658/1260. In addition to the literary record, there is architectural, epigraphic and numismatic evidence of this short-­lived Ayyubid dynasty at Mayyafariqin. Awhad Najm al-­Din Ayyub (596–607/1200–10) left an inscription dated Ramadan 599/May–June 1203 on a tower of the eastern inner wall (illustrated and described by Lehmann-­ Haupt, Armenien, 425–6) and the name of his successor, Ashraf Musa (607– 12/1210–20), is inscribed on a tower to the north (ibid). Musa’s brother, Muzaffar Shihab al-­Din Ghazi (617–42/1220–44) built a fine mosque of red baked brick with an inscription dated 624/1227 which was seen and analysed by Taylor (ibid., 428). The inscription on the mihrab is given by Gabriel (Voyages, inscription no.  124). The coins struck by the Ayyubids of Mayyafariqin have aroused a certain interest amongst scholars (Grabar, 167–78; Lane-­Poole, iv, 122–30; Lowick, 164–5). A series of them minted between 582/1186–7 and 612/1215–6 represent crowned human figures. Some have long locks of hair; others are wearing caps with tassels; sometimes these figures are enthroned (ibid.). The Mongols devastated the area around Mayyafariqin as early as 628– 9/1231. In 638/1240–1 a Mongol embassy reached the town and demanded that it should surrender and that its fortifications be destroyed. On this occasion, Muzaffar Shihab al-­Din Ghazi succeeded in deflecting the attentions of the embassy elsewhere. His son, Kamil Muhammad (642–58/1244–60), defied the Mongols in a brave stand at Mayyafariqin, but the city fell in 658/1260 to the Mongol army of Hülegü under the command of Yashmut and it was then that this last Ayyubid ruler was killed (Rashid al-­Din, 77–81; Ibn Shaddad, fol. 120a).

ma yya f a r iq in  | 103 Descriptions of Mayyafariqin in the Muslim geographers. There is some disagreement in the classical Muslim geographical works on the placing of Mayyafariqin. Al-­Muqaddasi (137) puts it in Diyar Bakr, al-­Istakhri (188) considers it to be part of Arminiyya, whilst Ibn al-­Faqih (133) places it in Diyar Rabi‘a. Ibn Shaddad lists Mayyafariqin as one of four amsar of Diyar Bakr, the other three being Amid, Arzan and Mardin (fol. 65a). Al-­Istakhri (76) describes Mayyafariqin as having an encircling wall and an abundant water supply, but he comments on the town’s unhealthy climate. Al-­Muqaddasi (140) mentions the fortifications, including battlements, an encircling wall and ditch; he also notes that the water there is muddy in winter. According to Qudama (246), the combined revenue of Arzan and Mayyafariqin in ‘Abbasid times was 4,100,000 dirhams. Nasir-­i Khusraw visited the town in 438/1046–7. He was impressed by the excellent condition of its walls, which seemed as if they had only just been completed (tr. Schefer, 24–5). Yaqut (d. 626/1229) praised the city, especially its surrounding wall of white stone and its prosperous suburb (rabad) (Mu‘jam, iv, 703–7). When Ibn Shaddad visited Mayyafariqin in the seventh/thirteenth century, he found thriving khans and markets, as well as two madrasas, one Hanbali, the other Shafi‘i (fol. 71a). Both Ibn Shaddad and Yaqut mention eight city gates at Mayyafariqin, seven of which probably corresponded to those of Byzantine Martyropolis (Gabriel, Voyages, 218). The eighth–ninth/fourteenth–fifteenth centuries. In the Ilkhanid period (654–754/1256–1353), Mayyafariqin shared the fate of the rest of Diyar Bakr and was ruled by Mongol amirs. After the collapse of the Ilkhanid state, after 736/1336, Diyar Bakr fell into disarray and became the arena for power struggles between rival Turcoman (the Aq Qoyunlu and Qara Qoyunlu confederations), Kurdish and Arab groups, before falling victim to the depredations of Timur who attacked the area (but not, apparently, Mayyafariqin) in 796/1394 and 803/1400–1 (Ibn ‘Arabshah, 65–6, 164–5). Thereafter, Mayyafariqin was in the hands of one branch of the mostly nomadic Sulaymani Kurds until it was taken in 827/1427 by the Aq Qoyunlu leader Qara ‘Uthman (d. 839/1435), who appointed his son Bayazid governor of the town and other citadels in the area (Tihrani, 95).

104  |  the me d ie va l tur k s The Safawid and Ottoman periods. The Safavid Shah Isma‘il I occupied the whole of Diyar Bakr in his campaign against the last Aq Qoyunlu ruler Murad in 913/1507–8. He then allotted Diyar Bakr to Khan Muhammad Ustajlu (Iskandar Beg Munshi, ‘Alam-ara, i, 32–3). After Isma‘il’s defeat at Čaldiran in 920/1514, Mayyafariqin was seized by the Kurdish chief Sayyid Ahmad Beg Ruzaki. The city fell under Ottoman control in 921/1515 after the battle of Kosh Hisar, when the Safavids were forced to cede Diyar Bakr to the Ottomans. In his history of the Kurds the tenth/sixteenth-­century writer Sharaf al-­Din Khan Bidlisi lists the governors of Mayyafariqin in his own time (Sharaf-nama, 270–2). Information on Mayyafariqin in the Ottoman period is scanty. The Portuguese traveller Tenreyro went there in 936–7/1529 and found it ‘almost deserted’ (Itinerario, 406). Ewliya Čelebi (d. c.  1095/1684) visited the town (Seyahat-nama, iv, 76–8) and gave a long laudatory description of the Satidama bridge. Von Moltke, who passed through the city in the nineteenth century, while noting the well-­preserved state of its walls and towers, commented on the ruined condition of the rest of the city which he said had been caused by Ottoman–Kurdish struggles in the area (Lehmann-­Haupt, 394, 419). Indeed, the city was to remain de facto in Kurdish hands until the beginning of this century. Christianity in Mayyafariqin during the Islamic period. The Arabic sources record little of the transition from Christianity to Islam within Mayyafariqin, a major centre of Oriental Christianity. Isolated references indicate, however, that Christianity continued to prosper after the Muslim conquest until recent times. This evidence is of course corroborated and expanded by surviving Christian architecture in the area. Al-­Muqaddasi (146) records without comment that in the monastery of Thomas (dayr Tuma) one farsakh from Mayyafariqin there was a mummified corpse; it was allegedly that of one of the disciples [of Jesus]. Ibn al-­Azraq mentions the existence of a Melkite church in the Marwanid period and that Christians held office in the Marwanid government (ed. ‘Awad, 149, 164). The Jacobites had a bishopric in Mayyafariqin by the fifth/eleventh century (Vryonis, 53), although this is not mentioned in the detailed chronicle of Ibn al-­Azraq. Ibn Shaddad does, however, mention an incident in which a Seljuq governor, Qiwam or Qawam

ma yya f a r iq in  | 105 al-­Mulk Abu ‘Ali al-­Balkhi, became exasperated by the naqus from a monastery in Mayyafariqin and, refusing a large sum of money offered him by the Christians if he would leave the building intact, destroyed it (fol. 70b). The same author records that in his own time (the seventh/thirteenth century) there were monasteries on a hill to the north of Mayyafariqin (ibid.). In 936–7/1529 Tenreyro describes ‘beautiful monasteries and churches without roofs, containing sumptuous monuments with inscriptions in Greek letters. On the walls were pictures of apostles and other saints, painted in very fine colours and gold’. He remarks that the town had only a small number of inhabitants who were Jacobite Christians and spoke Arabic (Itinerario, 376). Mayyafariqin in recent times. In 1891 the population of the town was 7,000 divided about equally between Muslims and Christians (Cuinet, ii, 470–2). During its occupation by the Sulaymani Kurds, the name Mayyafariqin had been eclipsed by Silvan (cf. EI 1, art: Maiyafariqin, and Minorsky’s etymology there of Silwan). According to the 1945 census, the population was 2,155. The most recent information indicates that according to the 1980 census, the population of the administrative unit (idari birim) of Silvan was 43,624 (Türkiye istatistik yıllığı, 39). Bibliography (for earlier bibliography, see Encyclopaedia of Islam, 1st edn, art. Maiyafariqin) Primary Sources Abu Bakr Tihrani-­Isfahani, Kitab-i Diyarbakriyya, eds N. Lugal and F. Sümer, Ankara 1962–4. al-­Baladhuri, Kitab futuh al-buldan, tr. P. K. Hitti, Beirut 1966. Ewliya Čelebi, Seyahat-nama, Istanbul 1314–8. Ibn al-‘Adim, Zubdat al-halab min ta’rikh Halab, ed. S. Dahan, Damascus 1954. Ibn ‘Arabshah, ‘Aja’ib al-maqdur fi nawa’ib Timur, tr. J. H. Sanders, London 1936. Ibn al-­Athir, al-Kamil fi’l-ta’rikh, ix and x, ed. C. J. Tornberg, Leiden and Uppsala 1864. Ibn al-­Athir, Ta’rikh al-dawla al-atabakiyya, ed. A. Tulayma, Cairo 1963. Ibn al-­Azraq al-­Fariqi, Ta’rikh Mayyafariqin wa-Amid, ed. B. A. L. ‘Awad, Cairo 1959 (Marwanid section). B. L. ms. Or. 5803 (covers early Islamic period up to 572/1176–7). Ibn al-­Faqih, Kitab al-buldan, ed. M. J. de Goeje, Leiden 1885.

106  |  the me d ie va l tur k s Ibn Hawqal, Kitab surat al-ard, ed. J. H. Kramers, Beirut 1990s. Ibn Khallikan, Wayafat al-a‘yan, tr. W. M. de Slane, Paris 1843–71. Ibn Miskawayh, Tajarib al-umam, eds and trs D. S. Margoliouth and H. F. Amedroz, Oxford 1920–1. Ibn Qudama, Al-‘Umada, tr. H. Laoust as Le précis de droit d’Ibn Qudama, Beirut 1950. Ibn Rusta, Kitab al-a‘laq al-nafisa, ed. M. J. De Goeje, Leiden 1967. Ibn Shaddad, al-A‘laq al-khatira fi dhikr umara’ al-Sham wa’l-Jazira, Oxford, Bodleian ms. Marsh 333. Iskandar Beg Munshi, Ta’rikh-i ‘alam-ara-yi ‘Abbasi, tr. R. M. Savory, Boulder, CO 1978. al-­Istakhri, Kitab masalik al-mamalik, ed. M. J. de Goeje, Leiden 1870s. al-­Muqaddasi, Ahsan al-taqasim, partial tr. A. Miquel, Damascus 1963. Nasir-­i Khusraw, Safarnama, ed. and tr. Ch. Schefer, Paris 1881. Rashid al-­Din, Jami‘ al-tawarikh, iii, ed. A. A. ‘Alizada, Baku 1957. Sharaf al-­Din Khan Bidlisi, Sharafnama, Arabic tr. M. J. B. Ruzhbiyani, Baghdad 1953. Yahya b. Sa‘id al-­Antaqi, Annales, in Patrologia Orientalis, xviii, Paris 1924. Yaqut, Mu‘jam al-buldan, ed. F. al-­Jundi, Beirut 1990. Secondary Sources Altun, A., Andolu ’da Artuklu devri Türk mimarisi’nin gelişmesi, Istanbul 1978. Amedroz, H. F., Three Arabic manuscripts on the history of the city of Mayyafariqin, in JRAS (1901), 785–812. Bell, G. L., The churches and monasteries of the Tur ‘Abdin and neighbouring districts, articles reprinted with an introduction by M. Mango, London 1982. Cahen, C., La Jazira au milieu du treizième siècle d’après ‘Izz ad-Din Ibn Chaddad, in REI, viii (1934), 109–28. Cahen, C., Le Diyar Bakr au temps des premiers Urtuqides, in JA, ccxxvii (1935), 219–76. Canard, M., Histoire de la dynastie des Hamdanides, i, Paris 1953. Canard, M., La date des expéditions mésopotamiennes de Jean Tzimisces, in Mélanges Henri Grégoire, ii, Annuaire de l’Institut de Philologie et d’Histoire orientales et slaves, x (1950), 99 ff. Cuinet, V., La Turquie d’Asie, Paris 1891–4. Eddé-­Terrasse, A.-M., ‘Izz al-Din Ibn Shaddad. Déscription de la Syrie du Nord, Damascus 1984.

ma yya f a r iq in  | 107 Flury, S., Islamische Schriftbänder, Basel 1920. Gabriel, A., Voyages archéologiques dans la Turquie orientale, Paris 1940. Grabar, O., On two coins of Muzaffar Ghazi, ruler of Maiyafariqin, in Amer. Num. Soc., Museum Notes 5 (1952), 167–78. Hartmann, R., Zu Ewlija Tschelebi’s Reisen im oberen Euphrat – und Tigrisgebiet, in Isl., ix (1919), 184–244. Hillenbrand, C., The history of the Jazira 1100–1150; the contribution of Ibn al-Azraq al-Fariqi, PhD thesis, University of Edinburgh 1979, unpubl. Lane-­Poole, S., Catalogue of oriental coins in the British Museum, London 1879–89. Lehmann-­Haupt, C. F., Armenien einst und jetzt, Berlin 1910. Lowick, N., The religious, the royal and the popular in the figural coinage of the Jazira, in The art of Syria and the Jazira 1100–1250, ed, J. Raby, Oxford 1985, 159–74. Markwart, J., Südarmenien und die Tigrisquellen nach griechischen und arabischen Geographen, Vienna 1930. Sinclair, T., Early Artuqid mosque architecture, in The art of Syria and the Jazira 1100–1250, 49–68. Tenreyro, A., Itinerario de Antonio Tenreyro, Coimbra 1725. Türkiye istatistik yıllığı 1985, Ankara 1985. Vryonis Jr., S., The decline of Medieval Hellenism in Asia Minor and the process of Islamization from the eleventh through the fifteenth century, Berkeley, Los Angeles and London 1971. von Moltke, H., Briefe über Zustände und Begebenheiten in der Türkei aus den Jahren 1835 bis 1839, Berlin 1917. Woods, J. E., The Aqquyunlu: clan, confederation, empire, Minneapolis and Chicago 1976.

7 Mu‘in al-­Din Parwana: The Servant of Two Masters?

T

he history of the Seljuqs of Rum is still relatively little studied, although the dynasty survived much longer1 than its better-­known namesake, the Great Seljuqs in Iran. From the annals of the later history of the Seljuqs of Rum one name does at least create small ripples in one’s memory and usually merits perhaps a paragraph in general histories of the Islamic world: that of Mu‘in al-­Din Sulayman Parwana. It is probable, however, that the best-­ known facts about this personality are his unusual t­ itle – ­Parwana – and that according to popular belief, he was arrested by the Mongol khan, Abaqa, put to death and then eaten by his captors. Claude Cahen in his Pre-Ottoman Turkey2 provides the longest analysis in a European language of the career of Mu‘in al-­Din ­Parwana – ­some twelve ­pages – b­ ut concludes rather lamely with the following words: ‘The obvious strength of his personality makes it all the more to be regretted that the nature of the surviving documents renders it impossible to form any true appreciation of it.’ The Turkish scholar, Kaymaz, wrote a short but uncritical book on the Parwana in 1970 but failed to address many of the issues raised by the career of this enigmatic, although important, figure in thirteenth-­century Anatolian history.3 The object of this paper is to look again at the Parwana, to trace the major phases of his career and to highlight and discuss some of the wider questions his career poses both in the context of Anatolian history and against the wider backcloth of Mamluk–Mongol relations in the second half of the thirteenth century. 108

th e se rv a nt o f tw o ma ste r s ?  | 109 First, a brief word about the sources. The major text is of course Ibn Bibi’s history of the Seljuqs of Rum, a work completed in 680/1281.4 His detailed information can be a­ ssessed – a­ t least for those episodes which deal with Mamluk and Mongol involvement in Anatolian a­ ffairs – a­ gainst the testimony of Mamluk writers, such as Ibn Shaddad, Ibn al-­Dawadari, Mufaddal b. Abi’l-­Fada’il and others,5 as well as against the account given in the World History of Rashid al-­Din who views the Parwana from the vantage-­point of the Mongol court at Tabriz.6 Before tracing the major phases of the Parwana’s political career,7 it is important to set the scene. The second half of the thirteenth century in Anatolia may be viewed as tumultuous, even in the context of the area’s earlier turbulent history. The authority of the Seljuq state of Rum had crumbled after the battle of Köse Dagh in 641/1243. Thereafter the Mongols were a constant interfering presence within Anatolia, and in their wake more Turcoman groups came into an already disturbed territory. The career of the Parwana has to be assessed against such a backcloth, a highly disordered situation on which the sources throw only dim light. Mu‘in al-­Din Sulayman Parwana was the de facto ruler of the Seljuq state in Anatolia during most of the period of the Mongol Protectorate. His father, Muhadhdhab al-­Din ‘Ali al-­Daylami (d.  642/1244), the vizier of Kaykhusraw II, interceded with the Mongols after Köse Dagh and secured the survival of the Seljuqs of Rum, albeit as Mongol vassals.8 His son, Mu‘in al-­Din Sulayman, began his career as commander of Tokat. In 654/1256, through the good offices of the Mongol general Bayju, he was given the titles amir Hajib and parwana9 (the ‘personal assistant of the sultan’10). Thereafter he became the real ruler of the Seljuq state for over two decades though nominal authority remained with puppet Seljuq sultans. When the Seljuq sultan Qilij Arslan IV came of age, he grew weary of the Parwana’s authority. Aware of this, the Parwana arranged for his murder at Aksaray in 663/1265 and for the installation of his infant son, Kaykhusraw III as sultan.11 The next decade was turbulent. Certain Anatolian amirs, allegedly with the complicity of the Parwana himself (the sources mention letters supposedly written by him to Baybars), invited the Mamluk sultan to invade their land and rid it of the Mongol yoke. Baybars’ Anatolian expedition and his victory over the Mongol army at Albistan in 675/1277 are well documented

110  |  the me d ie va l tur k s in Mamluk sources. The Parwana still fought on the Mongol side but was already under some suspicion. His son and grandson were taken prisoner by Baybars. When the Mamluk sultan made his triumphal entry into Kayseri, the ­Parwana – w ­ ho had retreated to his stronghold at T ­ okat – ­did not come to acknowledge him. Baybars then withdrew to Syria on the news that Abaqa himself was on his way to Rum with a large army.12 The Parwana joined Abaqa on his visit to the battlefield at Albistan and was there blamed for the Mongol massacre.13 He then accompanied Abaqa eastwards, where the Ilkhan, inflamed by his grief-­stricken relatives, ordered the Parwana’s execution (1 Rabi‘ I 676/2 August 1277).14 The Armenian historian, Hayton, is the source of the widespread but at least partly apocryphal story that ‘in accordance with Tartar custom’, the Parwana’s corpse was cut in half and then eaten.15 Having given the broad outlines of the Parwana’s turbulent career and grisly demise, it is now time to highlight some of the wider questions. In the context of a short paper such as this, only a few issues out of many can be addressed. What is the interest to the historian of medieval Anatolia of an examination of the career of the Parwana? Certainly, it represents a key phase in the history of the Rum Seljuqs. After the establishment of the Mongol Protectorate in Anatolia it might have been expected that the Seljuq state had little room for independent manoeuvre, that its sole function was to provide tribute and grazing lands for the Ilkhanids. Yet it is clear that for at least the middle part of his period in power the Parwana succeeded in acquiring for himself a considerable degree of political autonomy. This was achieved by adroit and assiduous courting of his Mongol overlords. The Parwana, from the very outset of his career, made great play of visiting the Mongol court at Tabriz regularly, of receiving their ratification for his activities, of cultivating good relations with the Mongol representatives at the Seljuq court in Konya, and of fighting harmoniously side by side with the Mongol army. The Parwana’s close relationship with his Mongol overlords is confirmed by an inscription (663/1265) on the Sultan Muhammad mosque at Merzifon in the name of the Parwana who has the Mongol title iftikhar-i nuyan, a unique occurrence in Anatolian epigraphy.16 Above all, from the time of the accession of Abaqa (663/1265) it is clear that the Parwana enjoyed a special

th e se rv a nt o f tw o ma ste r s ?  | 111 relationship with him. This was never more apparent than just before the Parwana’s death when in spite of the evidence of his alleged duplicity sent by Baybars to Abaqa, the Mongol leader was still reluctant to impose the death penalty on the Parwana. Having established a basis of mutual cooperation and trust between the Mongol leadership and himself, from the beginning of his career, the Parwana was able to exercise considerable power locally within Anatolia and indeed to prolong Iranian political and cultural traditions within that area. Whilst the price the Parwana had to pay for this degree of autonomy was a lasting Mongol military presence on Anatolian soil, this did not prevent him from governing in the way he chose. He was therefore quick to establish a network of loyal followers, members of his own family or long-­standing associates, in key provinces and especially in border areas. From the time of the murder of Qilij Arslan IV the Parwana was ­able – ­during extremely troubled ­times – ­to maintain a relatively stable state. It is important to stress that the Parwana belonged within the Persian bureaucratic tradition. He was not really a military man, and he was certainly not a Turkish professional soldier, although circumstances led him to participate in military engagements. He and his close colleague, Fakhr al-­Din ‘Ali, were prominent amongst Persian notables who clearly saw themselves as local Anatolian rulers, lords of small territories which they carved out for themselves far away from the Mongol court at Tabriz.17 As with the Timurids later on, it was possible in Anatolia for small but thriving centres of Persian culture to develop, patronised by loyal rulers. Whilst distance helped the Parwana to create a small domain of his own around Tokat and Sinope, his frequent visits to the Ilkhanid court at Tabriz facilitated scholarly and religious contacts with a Persian-­speaking cultural milieu. Indeed, the Parwana settled the Persian mystic, Fakhr al-­Din ‘Iraqi (d. 688/1289) not far from Tokat18 and the Parwana and his wife endowed his ­domain – ­Tokat, Merzifon, ­Sinope – ­with mosques, khans and madrasas.19 It is possible that the importance of the Parwana’s relatively well-­publicised close friendship with Jalal al-­Din Rumi and his interest in the Mevlevi order may have been exaggerated. Certainly, the Parwana plays a significant if stylised role in Rumi’s work, Fihi ma fihi, where he represents the enquiring disciple who seeks enlightenment from Rumi, the master.20 Nevertheless, the

112  |  the me d ie va l tur k s reproach addressed to the Parwana by R ­ umi – t­ hat by making common cause with the (infidel) Mongols the Parwana had ruined the Dar al-Islam – has a ring of obvious historical authenticity to it and indeed might have been a contributory factor in the Parwana’s alleged change of attitude towards the Mamluks in the 1270s.21 The close ties between the Parwana and Rumi are corroborated by Aflaki, the fourteenth-­century biographer of the Mevlevi order, who writes that Rumi often held spiritual concerts in the Parwana’s house and that the Parwana and his wife provided funds for the building of Rumi’s mausoleum.22 Thus we see the Parwana assuming the trappings of a Perso-­Islamic ruler, frequenting the religious classes and sponsoring the building of religious monuments. The demise of the Parwana marked the end of even semi-­independent Seljuq rule in Anatolia, which was henceforth governed directly by the Mongols though his descendants held on to Sinope and Tokat for several generations.23 It is easy and tempting to dismiss the Parwana’s career as the rise and fall of a petty despot, a servant who in pursuit of his own personal ambition served two masters and fell foul of both of them. Given the prolonged period of his alleged double game of courting both Mongols and Mamluks it is surprising that the Parwana stayed in power as long as he did. Indeed, this political longevity is palpable evidence of his political skills. Ambitious he was, but he can also be seen as the representative of the last generation of Persian bureaucrats in Anatolia, men of Persian stock who brought Persian court traditions and religious modes into Anatolia and ­created the cultural environment in which figures such as Rumi could flourish. This late flowering of Persian traditions on Anatolian soil was of course destined to be short-­lived. Internal dissensions within the Seljuq state, above all the perennial problem of the centrifugal forces of Turcoman activities, obliged the Parwana to continue to rely on outside support from one of the two super-­powers. After the myth of Mongol invincibility had been exploded in the 1260s, after the Mongols no longer represented a monolithic entity to the outside world and especially after Abaqa’s reluctance to tackle Baybars directly,24 it is perhaps not surprising that the Parwana should have allegedly dabbled with the idea of courting a new overlord in Egypt. Perhaps he hoped to gain Baybars’ protection and thereby to strengthen and consolidate his

th e se rv a nt o f tw o ma ste r s ?  | 113 semi-­independent state in Anatolia, threatened as it was by both Turcoman and Mongol encroachments. The Parwana’s own role in the detailed but confused accounts of the events preceding Baybars’ Anatolian expedition remains ambiguous. Why did he backtrack at the eleventh hour, when Baybars finally came to Kayseri and invited him to declare allegiance to him? Did he really have a change of heart and revert to his former allegiance to Abaqa? Or was it an elaborate trap for Baybars so that Abaqa could come quickly from Tabriz and attack the Mamluk sultan far from home? Or was it that the Parwana withdrew to Tokat to wait and see which super-­power would emerge triumphant from the military encounter which he thought imminent and that he would then make his dispositions accordingly? Or was the Parwana blamed retrospectively in the sources for the plotting of other Anatolian amirs who entered into negotiations with Baybars? Through the opacity of sources the Parwana remains a shadowy but powerful personality, perhaps playing a dangerous double game, perhaps the victim of other people’s conspiracies. At first sight, the evidence seems to indicate the following: that the Parwana was involved in inviting Baybars to Anatolia; that he panicked after the battle of Albistan, which he may well have thought the Mongols would win; that he did renege on his agreement with Baybars; and that he did then call on Abaqa to come to Rum to take on Baybars. Right until the end he remained confident that he would be able to ingratiate himself once more with his old ally Abaqa and thus he placed himself voluntarily in Abaqa’s power when the latter arrived in Anatolia. There is indeed strong evidence from the sources to support the hypothesis that the Parwana shifted allegiance from Abaqa to B ­ aybars – ­at least for a while.25 This hypothesis is, however, by no means the only possible explanation of the Parwana’s conduct.26 It was, after all, a common practice at that time for Anatolian amirs to denigrate one another by accusations of complicity with Baybars. It is possible to argue that the Parwana was framed, that other Anatolian amirs and leaders were in league with Baybars and that the Parwana remained loyal to the Mongols all along. How else can one explain his fighting against the Mamluks at Albistan and the capture of some of his family by the Mamluks? How else can one interpret his waiting for Abaqa’s arrival and his otherwise foolish action in joining Abaqa voluntarily

114  |  the me d ie va l tur k s when he could easily have escaped beforehand, as so many other Anatolian leaders had done already, to the safety of Mamluk protection in Syria? Whatever the true motivation of the Parwana’s dangerous gamble, he aroused the ire of both Baybars and Abaqa and provided Abaqa with a convenient scapegoat for the ignominious Mongol defeat at Albistan. Within the context of Anatolian history the death of the Parwana was viewed as marking the end of a generation. The era of Anatolian Persians whose education and attitudes had been formed by the Golden Age of the Seljuqs of Rum was now over. The last word should perhaps go to the poet Sahib Diwan Shams al-­Din who laments the Parwana’s death with the words: ‘Solomon is no more and the devils are set free’.27 Notes  1. 470/1077–707/1307.   2. C. Cahen, Pre-Ottoman Turkey (London, 1968), 269ff.   3. N. Kaymaz, Pervane Mu‘inü’d-din Süleyman (Ankara, 1970).   4. Ibn Bibi, tr. H. W. Duda as Die Seltschuken-Geschichte des Ibn Bibi (Copenhagen, 1959), 260, 267, 271, 274, 277, 280–307.  5. Ibn Shaddad, Ta’rikh al-Malik al-Zahir, ed. A. Hutait (Wiesbaden, 1983), 34–5, 77–9, 107–8, 122–9, 153ff.; Ibn al-­Dawadari, ed. U. Haarmann, as Die Chronik des Ibn ad-Dawadari Pt 8 (Freiburg, 1971) 164, 177, 188–207; Mufaddal b. Abi’l-­Fada’il, Al-Nahj al-sadid wa’l-durr al-farid fi ma ba‘d Ta’rikh Ibn al-‘Amid, ed. and tr. E. Blochet in Patrologia Orientalis XIV (Paris, 1920), 389–439; Ibn al-­Suqa‘i, Tali Kitab Wafayat al-A‘yan, ed. and tr. J. Sublet (Damascus, 1974), 79–80, 102–3.   6. Rashid al-­Din, Jami‘ al-tawarikh, ed. A. A. ‘Alizada, III (Baku, 1957), 144–7.  7. Cf. Encylopedia of Islam, 2nd edn, art: Mu‘in al-­Din Parwana (C. Hillenbrand) (forthcoming).   8. Cf. Cahen, op. cit., 269ff.   9. Ibn Bibi, op. cit., 274 and 277. 10. Cf. the discussion in Cahen, op. cit., 221. 11. Ibn Bibi, op. cit, 288–90; al-­Maqrizi, tr. M. Quatremère (Paris, 1840), I, pt 2, 57. 12. Ibn Shaddad, op. cit., 169–77; Ibn al-­Dawadari, op. cit., 198–204; Mufaddal, op. cit., 421–33; al-­Maqrizi, op. cit., 140–2. Cf. also Ibn Bibi, op. cit., 303–5; Rashid al-­Din, op. cit., 144.

th e se rv a nt o f tw o ma ste r s ?  | 115 13. Ibn al-­Suqa‘i, op. cit., 80; Ibn Shaddad, op. cit., 181–2; Ibn al-­Dawadari, op. cit., 205; Mufaddal, op. cit., 434–5; al-­Maqrizi, op. cit., 145. 14. Rashid al-­Din, op. cit., 146–7; Ibn Shaddad, op. cit., 183; Ibn Bibi, op. cit., 307–8; al-­Maqrizi, op. cit., 163; Ibn al-­Suqa‘i, op. cit., 80; Ibn al-­Dawadari, op. cit., 207; Mufaddal, op. cit., 438–9. 15. Hayton, ‘Flos Historiarum Terre Orientis’, RHC, Documents arméniens (Paris, 1906), 308–9. Hayton writes as follows: ‘He (Abaqa) had him (the Parwana) . . . cut in half according to the custom of the Tartars, and ordered that what (there was) from the flesh of that traitor should be placed amongst all the food which was to be eaten. Abaqa himself ate from it and gave (it) to all his nobles to eat. And Abaqa exacted such a vengeance from that Parwana, the most wicked traitor.’ This account is dismissed by Spuler as legendary (Die Mongolen in Iran (Berlin, 1955), 378). Yet Spuler himself indicates that the Mongols used a wide variety of methods for exacting the death penalty, including the cutting of the body into two or several pieces (ibid., 377). Thus at least half of Hayton’s statement is accurate. The charge of cannibalism is more difficult to prove. There is little concrete evidence on which to generalise about the Mongols’ treatment of the corpses of criminals (and especially of political traitors of the lofty status of the Parwana). Hayton’s account may well not be as fanciful as it first appears. 16. J. M. Rogers, Patronage in Seljuk Anatolia, 1200–1300, unpublished DPhil thesis (University of Oxford), 276. 17. For details of the titulature and building activities of Fakhr al-­Din ‘Ali, cf. Rogers, op. cit., 167–9. 18. Aflaki, Manaqib al-‘Arifin, tr. C. Huart, Vol. I (Paris, 1918), 313. 19. Cf. Rogers, op. cit., 97, 169, 276–7; A. Gabriel, Monuments turcs d’Anatolie (Paris, 1931), II, 103, 106–7; Aflaki, op. cit. (Paris, 1922), 74, 283. 20. A. J. Arberry, tr. of Jalal al-­Din Rumi, Fihi ma fihi (London, 1975). 21. ‘Having made common cause with the Tartars, you are giving them assistance so as to destroy the Syrians and Egyptians and to ruin the realm of Islam. God therefore made that very means which would have secured the survival of Islam into the means of its diminishment’ (Jalal al-­Din Rumi, op. cit., 17). 22. Cf. Aflaki, op. cit., 267. The Parwana’s name also appears in inscriptions on a madrasa (661/1263) in Sinope (cf. RCEA XII (Cairo, 1943), 76) and a mosque (663/1265) in Merzifon (ibid., 94–5). 23. Cf. the genealogical tables found in C. E. Bosworth, Islam Devletleri Tarihi, trs E. Mercil and M. Ipsirli (Istanbul, 1980), 316.

116  |  the me d ie va l tur k s 24. Cf. the recent discussion by D. O. Morgan, ‘The Mongols and the Eastern Mediterranean’, Mediterranean Historical Review 4/1 (June 1989), 198–211. 25. Mamluk writers are unanimous on this, although their accounts may well stem from a single source, Ibn Shaddad. Their testimony is not fully supported by Rashid al-­Din who suggests that the charge made against the Parwana was insubordination rather than treachery (op. cit., 146–7). 26. Cahen’s discussion on the Parwana’s motivation is cautious but unnecessarily obscure. He concedes that it is ‘possible’ that the Parwana sought some kind of safeguard for himself from Baybars. So far so good. It is, however, rather difficult to understand the enigmatic statement that the Parwana was ‘anxious to bring Baybars’ forces nearer and yet not to compromise himself with him’ (op. cit., 285–6). Ibn Bibi, on the other hand, mentions the friendship between Baybars and the Parwana and describes the latter as fleeing from the battlefield of Albistan like a coward (op. cit., 304). 27. Quoted by Ibn Bibi in his obituary of the Parwana (op. cit., 308).

8 Mu‘in al-­Din Sulayman Parwana

M

u‘in al-­Din Sulayman Parwana, the de facto ruler of the Seljuq state in Anatolia (Rum) during most of the Mongol Protectorate. His father, Muhadhdhab al-­Din ‘Ali al-­Daylami (d. 642/1244) the vizier of Kaykhusraw II, interceded with the Mongol general Bayju after the battle of Köse Dagh (641/1243) and secured the survival of the Seljuqs of Rum, albeit as Mongol vassals. His son, Mu‘in al-­Din Sulayman, is first mentioned in the sources as commander of Tokat. In 654/1256, through the good offices of Bayju, he was given the titles amir hajib and parwana (the latter term denoting the ‘personal assistant of the sultan’, Cahen, 221). Thereafter he became the real ruler of the Seljuq state for over two decades, through nominal authority remained with the puppet Seljuq sultans. In Hülegü’s time, the Seljuq realm was partitioned, the eastern provinces being given to the pro-­Mongol Qilij Arslan IV, for whom the Parwana acted as vizier, and the remainder to Kayka’us II. After the latter had fled to Byzantium, Qilij Arslan IV became sole sultan in Rum, entering Konya on 14 Ramadan 659/13 August 1261. The Parwana visited the new Ilkhan, Abaqa, in 663/1265 and was officially awarded the port of Sinope which he had taken from the Greek ruler of Trebizond two years earlier. When Qilij Arslan IV came of age he grew weary of the Parwana’s authority. Perceiving this, the latter arranged for his murder at Aksaray in 663/1265 and the installation of his infant son, Kaykhusraw III, as sultan. The Parwana later forged marriage ties with the new s­ ultan – ­an inscription on the Sünbül Baba Zaviyesi at Tokat mentions the Parwana’s daughter as the wife of Kaykhusraw III (Rogers, 35; Gabriel, 103). 117

118  |  the me d ie va l tur k s The next decade (663–73/1265–75) was turbulent. Certain Anatolian amirs, probably with the complicity of the Parwana himself (the sources mention letters allegedly written by him to Baybars), invited the Mamluk sultan to invade their land and rid it of the Mongol yoke. Baybars’ Anatolian expedition and his victory over the Mongol army at Albistan (Dhu’l Qa‘da 675/April 1277) are well documented in Mamluk sources. The Parwana still fought on the Mongol side but was already under some suspicion. His son and grandson were taken prisoner by Baybars. When the Mamluk sultan made his triumphal entry into Kayseri, the P ­ arwana – w ­ ho had retreated to his stronghold at ­Tokat – ­did not come to acknowledge him. Soon afterwards, Baybars withdrew to Syria on hearing the news that Abaqa himself was on his way to Rum with a large army. The Parwana joined Abaqa on his visit to the battlefield at Albistan and was there blamed for the massacre of the Mongol army. He then accompanied Abaqa eastwards, where the Ilkhan, inflamed by his grief-­stricken female relatives, ordered the Parwana’s execution (1 Rabi‘ I 676/2 August 1277). The Armenian historian, Hayton, is the source of the widespread story that ‘in accordance with Tartar custom’, the Parwana’s corpse was cut in half and then eaten. The career of the Parwana represents a key phase in the history of the Rum Seljuqs as well as shedding light on the wider power struggle between Mamluks and Mongols. Internally, the Parwana attempted to maintain stability both amongst the Turkish amirs and between them and the ever-­ increasing number of Mongols resident on Anatolian soil. The Turcomans, especially the Qaramanids, were kept under control only with great difficulty (there were rebellions in 659/1261 and 675/1277). As for the Parwana’s relationship with his Mongol overlords, it is clear that at the beginning he was unreservedly on their side and that he cultivated his special links with Abaqa by frequent visits to Tabriz. Seljuq weakness after Köse Dagh allowed talented ministers, such as the Parwana and his friend and rival, Fakhr al-­Din ‘Ali, both of whom were of Persian stock and were directly nominated by the Mongols, to exercise semi-­independent power within Anatolia. Later on, however, the Parwana co-­existed uneasily with Abaqa’s representatives in Rum and especially with Ajay, Abaqa’s brother, whose removal he requested in 672/1273–4. Clearly, the Parwana had begun to resent increasing Mongol encroachment on

mu ‘ i n a l -di n su la ym a n pa r wa n a  | 119 Seljuq territory and the Mongols’ systematic exploitation of its economic resources. The Parwana’s own role in the detailed but confused accounts of the events preceding Baybars’ Anatolian expedition remains ambiguous. Certainly, the myth of Mongol invincibility had been exploded, and it seems likely that the Parwana’s loyalty shifted, for a while at least, to Baybars, whose power was increasing in the 1270s in the face of Mongol disunity and their apparent inactivity against Egypt. It is possible too that the Parwana may have been influenced in his more pro-­Mamluk stance by the reproach addressed to him by Jalal al-­Din Rumi (Fihi ma fihi, 17) that by making common cause with the Mongols he had ruined the Dar al-Islam. Once Baybars had reached Rum, the Parwana seems to have reneged on his arrangement and to have invited Abaqa to attack Baybars. Whatever the true motivation of the Parwana’s dangerous double game, he aroused the ire of both Baybars and Abaqa and provided a convenient scapegoat for the Mongol defeat at Albistan. The Parwana is said to have enjoyed a close relationship with Rumi. Some of the discourses in Fihi ma fihi are addressed directly to the Parwana. According to Aflaki, it was often in the Parwana’s house that Rumi held spiritual concerts and the Parwana and his wife provided funds for the building of Rumi’s mausoleum (Huart, ii, 80, 92–3, 267). The Parwana also settled the Sufi Fakhr al-­Din ‘Iraqi near Tokat. He was responsible for the construction of a limited number of religious buildings. An inscription (661/1263) on the madrasa of ‘Ala’ al-­Din at Sinope celebrates the Parwana’s conquest of the city (RCEA, xii, 76). The Sultan Muhammad mosque at Merzifon bears a foundation inscription (663/1265) in the name of the Parwana, who has the Mongol title iftikhar-i nuyan (ibid., 94–5). Its unique occurrence in Anatolian epigraphy demonstrates the Parwana’s close relationship with his Mongol overlords (Rogers, 276). The demise of the Parwana marked the end of even semi-­independent Seljuq rule in Anatolia, which was henceforth governed directly by the Mongols. Since they were primarily interested in eastern Anatolia, the Turcoman amirates were able to prosper in the west. The sons of the Parwana (the Parwana-­oghulları) continued for several generations (676–722/1277– 1322) to hold on to Sinope and Tokat, the area which the Parwana had made his personal domain.

120  |  the me d ie va l tur k s Bibliography Primary Sources Aflaki, Manaqib al-‘arifin, tr. Cl. Huart, Paris 1918, 1922. Jalal al-­Din Rumi, Fihi ma fihi, tr. A. J. Arberry, London 1975. Hayton, Flos Historiarum Terre Orientis in Recueil des historiens des croisades, Documents arméniens, Paris 1906, 308–9. Ibn Bibi, tr. H. W. Duda as Die Seltschuken-Geschichte des Ibn Bibi, Copenhagen 1959, 260, 267, 271, 274, 277, 280–307. Ibn al-­Dawadari, ed. U. Haarmann as Die Chronik des Ibn ad-Dawadari, Freiburg 1971, Pt viii, 164, 177, 188–207. Ibn al-­Suqa‘i, Tali Kitab Wafayat al-a‘yan, ed. and tr. J. Sublet, Damascus 1974, 79–80, 102–3. Ibn Shaddad, Ta’rikh al-Malik al-Zahir, ed. A. Hutait, Wiesbaden 1983, 34–5, 77–9, 107–8, 122–9, 153 ff. Maqrizi, Kitab al-Suluk, ed. M. M. Ziyada, Cairo 1956, Pt i, 625 ff. Mufaddal b. Abi’l-­Fada’il, al-Nahj al-sadid wa’l-durr al-farid fi ma ba‘d ta’rikh Ibn al-‘Amid, ed. and tr. E. Blochet, in PO, xiv, Paris 1920, 389–439. Rashid al-­Din, Jami‘ al-tawarikh, ed. A. A. ‘Alizada, Baku 1957, iii, 144–7. Répertoire Chronologique d’Épigraphie Arabe, xii, Cairo 1943. Secondary Sources Bosworth, C. E., Islam devletleri tarihi, trs E. Merçil and M. Ipșirli, Istanbul 1980, 316. Cahen, C., Pre-Ottoman Turkey, London 1968, 269 ff. Edhem, H., Merzifon’da Perwane Mu‘inü ’d-Din Süleyman namına bir kitabe, in TOEM, sene 8, 42–52. Encyclopaedia of Islam, 1st edn, art: Mu‘in al-­Din Sulaiman Parwana (J. H. Kramers). Gabriel, A., Monuments turcs d’Anatolie, Paris 1931, ii, 103, 106–7. Kaymaz, N., Pervane Mu‘inü ’din Süleyman, Ankara 1970. Rogers, J. M., Patronage in Seljuk Anatolia 1200–1300, DPhil thesis, University of Oxford 1972, unpubl. Spuler, B., Die Mongolen in Iran, Berlin 1955. Thorau, P., Sultan Baibars I von Ägypten, Wiesbaden 1987, 282–9.

9 1092: A Murderous Year

T

he year 485/1092 was a momentous one in Islamic history. It witnessed the deaths in quick succession of the two most famous men in the eastern Islamic world, in the great Seljuq empire which had flourished since the middle of the fifth/eleventh century. Those two men, as any student of Islamic history knows, were the great vizier, Nizam al-­Mulk and the Seljuq sultan himself, Malikshah. The received version of events goes as follows: Nizam al-­Mulk was the first important victim of the Assassin sect who as the avowed enemies of the Sunni Seljuq world aimed at the removal of major political and religious figures in their attempt to topple the Sunnis and usher in the Isma‘ili imamate. Nizam al-­Mulk fell, then, slain by an Assassin’s knife. Shortly afterwards the sultan Malikshah went hunting outside Baghdad, ate some rotten meat and died a few days later. With the removal of these two key figures, the Seljuq empire fell into disarray and was never again to enjoy the unity and prosperity of the two decades of Malikshah’s rule (1072–92). As for the Assassins, the impetus of their movement had been lost by the middle of the fifth/twelfth century, not before. This received version of the deaths of Nizam al-­Mulk and Malikshah is very difficult to dislodge from one’s mind. As Bernard Lewis writes: ‘Tactical and propagandist motives combine in the murder of major figures such as the great vizier Nizam al-­Mulk.’1 As if to confirm this point, the illustration of Nizam al-­Mulk’s death on the front cover of the recent reprint of his book The Assassins, confidently juxtaposes the Assassin group with the actual murder. As for the death of Malikshah, this is tacitly accepted by most modern scholars as having been due to natural causes.2 121

122  |  the me d ie va l tur k s Some scholars have already cast tentative doubts on whether Nizam al-­Mulk’s assassination was indeed perpetrated by the Assassins3 and a few have also questioned, still more haltingly, the circumstances surrounding the death of Malikshah.4 They have not, however, scrutinised all the evidence and allowed it to lead them away from the well-­trodden path. That will be the aim of this paper. But first a few words to set the scene. Throughout the reign of Malikshah and earlier, during the time of his father Alp Arslan, the Seljuq empire had been ruled by the vizier, Nizam al-­Mulk. Inevitably, such a man would attract rivals who would try regularly to unseat him. It is reasonable to assume that Nizam al-­Mulk had reached supreme power by masterminding the grisly5 demise of his predecessor, al-­Kunduri, who is alleged to have warned Nizam al-­Mulk that a similar fate awaited him.6 Once in power, Nizam al-­Mulk stayed there by filling a vast network of administrative posts with his sons, grandsons, sons-­in-­law and other relations and by surrounding himself with a sizeable personal army of Turkish mamluks, the Nizamiyya.7 The young sultan Malikshah who came to power in his late teens does not seem to have found the authority of Nizam al-­Mulk irksome, at least at the beginning of his reign, but the longer the great vizier lived (and he lived a long time) the more likely it became that Malikshah might wish to involve himself more personally in governing the empire and the more likely it would also be that other ambitious and talented rivals would wish that the vizier would retire and leave room for younger men. If he would not fall, perhaps he might need to be pushed. Let us now examine the events of the three years 485–7 and see where such an examination leads us. In Ramadan 485/October 1092 Malikshah and his court, including Nizam al-­Mulk with his entourage, were on their way from Isfahan to Baghdad. They stopped at a place called Sahna, near Nihawand, and Nizam al-­Mulk, en route in a large mêlée of people to his women’s quarters, was struck down and killed by an assailant’s dagger. The date was probably 10 Ramadan/14 October 1092.8 The assassin himself was killed at once. Malikshah continued his journey, arriving in Baghdad in Shawwal 485/November 1092. En route he appointed as his vizier Taj al-­ Mulk, a long-­standing rival of Nizam al-­Mulk. Malikshah became ill after a hunting trip outside Baghdad and died, probably on the 16 Shawwal 485/19

a mu rd er ou s ye a r  | 123 November 1092.9 Shortly afterwards, his widow Terken Khatun persuaded the ‘Abbasid caliph, al-­Muqtadi, to pronounce the khutba in the name of her young son, Mahmud. She then went back to Persia to bury Malikshah.10 The next year, 486/1093, began with the savage murder by the Nizamiyya of Taj al-­Mulk on 12 Muharram/12 February. These mamluks espoused the cause of Barkyaruq, Malikshah’s son from another wife, Zubayda Khatun and, according to some sources, they actually pronounced the khutba in Barkyaruq’s name.11 Meanwhile in Baghdad the young son of the marriage between the caliph and Malikshah’s daughter died some time in Jumada I/ June 1093.12 The catalogue of deaths continued into 487/1094. Barkyaruq arrived in Baghdad in Muharram and had the caliph pronounce the khutba in his name on 14 Muharram/3 February 1094. The following day, 15 Muharram/4 February 1094 the caliph al-­Muqtadi was dead.13 Later that year, in Ramadan/September–October 1094 Terken Khatun, Malikshah’s widow, and her young son, Mahmud, also died. Barkyaruq was now Seljuq sultan and his appointee, al-­Mustazhir, was the new ‘Abbasid caliph.14 Even if this catalogue of names and events is difficult to follow, it will be obvious that almost all the prominent figures in this drama had died during the short space of under three years. This chain of events had begun with the two most important deaths, those of Nizam al-­Mulk and Malikshah, and it is to those that we now return. Who Killed Nizam al-­Mulk? Let us examine the various possibilities. 1. The Assassins By this I mean the Assassins acting independently out of political motives and not being hired out by any other party. What is the evidence for the Assassins’ perpetrating the deed? Firstly, all the primary sources say that Nizam al-­Mulk was murdered by a Daylami youth, most of them adding that he was from the Assassins (batiniyya) at Alamut.15 Moreover, the Mongol historians, al-­ Juwayni and Rashid ad-­Din, had access to Assassin sources at the time of the Mongol sacking of Alamut and above all to the life of Hasan-­i Sabbah, the Sarguzasht-i Sayyidna, which they allegedly quote.16 These two authors state firmly that it was the Assassins who carried out the murder. Al-­Juwayni no

124  |  the me d ie va l tur k s doubt quoting or misquoting his Isma‘ili source, writes: ‘Hasan-­i Sabbah spread the snare of artifices in order at the first opportunity to catch some splendid game, such as Nizam al-­Mulk, in the net of destruction and increase thereby his own reputation.’ He then goes on to name the assassin, Bu Tahir Arrani, who came to his victim in the guise of a Sufi, and confidently concludes: ‘He (Nizam al-­Mulk) was the first person to be killed by the fida’is.’17 It is true, moreover, that Nizam al-­Mulk had fought long and hard against the Assassins, that earlier in 485/1092 two expeditions had been sent out against them by Malikshah, one to Alamut and the other to Quhistan,18 and that Nizam al-­Mulk’s own obsession with the Assassins’ threat to the Sunni world had led him to vituperate against them in his Siyasatnama and to encourage al-­Ghazali to write a refutation of their doctrines.19 As al-­Juwayni put it: ‘He (Nizam al-­Mulk) . . . strove to excise the pus of the Sabbahian rebellion.’20 So, the Assassins, it could he argued, needed to rid themselves of this dangerous enemy of theirs.21 2. Malikshah There is evidence in the sources which could be interpreted as suggesting Malikshah’s deep complicity in Nizam al-­Mulk’s murder. As he was the supreme sultan, his complicity should mean that he instigated the murder himself. Malikshah had good reason to wish to be rid of Nizam al-­Mulk. Nizam al-­Mulk had waxed exceeding proud. According to as-­Subki, ‘Nizam al-­Mulk’s vizierate was not a vizierate; on the contrary, it was above the sultanate’.22 A number of the sources cite various, probably semi-­apocryphal stories of Malikshah’s resentment at sharing his throne with Nizam al-­Mulk and his family network.23 Nizam al-­Mulk had allegedly courted the caliph, al-­Muqtadi, on Malikshah’s first visit to Baghdad (in 480/1087) and Nizam al-­Mulk had even been invited to be seated in the caliph’s presence.24 Ibn Khallikan states explicitly that some reports said that the assassin of Nizam al-­Mulk had been hired by Malikshah who was tired of seeing Nizam al-­ Mulk live so long.25 Hamdallah Mustawfi alleges that Malikshah had actually gone so far as to dismiss Nizam al-­Mulk.26 Other sources, however, mention that Malikshah had decided to remove Nizam al-­Mulk from office but that he could not do it because of the immense military strength of Nizam al-­ Mulk’s entourage.27 The only alternative therefore was to kill him. The timing

a mu rd er ou s ye a r  | 125 was also crucial. It would appear that Malikshah was resolved, on arrival at Baghdad, to bring the caliphate more firmly under his own control. This he would do by deposing the caliph, al-­Muqtadi, and appointing in his place his own grandson, Ja‘far, who was the offspring of the unhappy marriage between al-­Muqtadi and Malikshah’s daughter, Mah Malik Khatun.28 This course of action was strongly opposed by Nizam al-­Mulk. So Nizam al-­Mulk had to be disposed of before the Seljuq court reached Baghdad. Bundari on the authority of ‘Imad al-­Din al-­Isfahani from the Seljuq vizier Anushirwan b. Khalid writes cryptically: ‘It was as if Nizam al-­Mulk’s assassination was secretly authorised by the sultan and was a matter which had been plotted and pre-­arranged.’29 3. Taj al-­Mulk This man, the avowed enemy of Nizam al-­Mulk and Nizam al-­Mulk’s short-­ lived successor in the vizierate, certainly stood to gain by Nizam al-­Mulk’s death.30 Nizam al-­Mulk blocked his path, just as al-­Kunduri had blocked that of Nizam al-­Mulk many years earlier, in 456/1064. Taj al-­Mulk is explicitly mentioned by some sources as being involved in Nizam al-­Mulk’s murder.31 He is seen either as being directly responsible or as the one who poisoned Malikshah’s mind against Nizam al-­Mulk. Perhaps the most telling evidence against him is his own violent death, only two months or so later, which, according to most of the sources, was at the hand of the mamluks of Nizam al-­Mulk, who wished to avenge the death of their master by killing the man who had been responsible for Nizam al-­Mulk’s death.32 And the plot thickens still further, for it was even said that they were incited by Terken Khatun’s rival, Zubayda Khatun.33 4. Terken Khatun This Karahanid princess, one of Malikshah’s wives, is seen by most of the sources as playing a major role in the events of 485/1092 and thereafter. As will be seen from the genealogical table,34 two of her sons from Malikshah had predeceased their father. The only hope of power for this valide sultan lay in her third son, the young boy, Mahmud. Before Nizam al-­Mulk’s murder Terken Khatun had tried to persuade Malikshah to declare Mahmud his heir instead of Barkyaruq, the elder son of Malikshah by another wife, the Seljuq

126  |  the me d ie va l tur k s princess, Zubayda Khatun. These attempts by Terken Khatun had been foiled, largely because of Nizam al-­Mulk’s continuing support of Barkyaruq and his firm opposition to Terken Khatun and her young son Mahmud. Not surprisingly, Terken Khatun harboured a violent animosity towards Nizam al-­Mulk.35 These four, then, are the major murder suspects. What may be concluded from the evidence, detailed very fully in the sources and summarised here? All the primary sources, including al-­Juwayni and Rashid al-­Din, are Sunni and heavily biased against the Assassins. Yet there is something half-­hearted, tame and inconsistent about their stereotyped phrases attributing Nizam al-­Mulk’s murder to the Assassins. The only sources which attribute the murder solely to the Assassins are the Isma‘ili ones, doctored or not by Rashid al-­Din and al-­ Juwayni, and certain early Sunni accounts such as those of Ibn al-­Qalanisi36 and Ibn al-­Azraq in Mayyafariqin37 which do not emanate directly from the eastern Islamic world. All the other sources, although they say almost perfunctorily and certainly elliptically that the Assassins were responsible for the murder, go on to mention in detail one person or several people from amongst the suspects already discussed above and to implicate them. One may safely infer that the Assassins were not the instigators of Nizam al-­Mulk’s murder, especially ­since – ­as will be argued ­shortly – ­it was followed by a spate of other most opportunely timed deaths for which no source at all blames the Assassins. What of the other three major suspects? With or without hiring an actual Assassin from Alamut, it would appear that all three had good motives. Of course it is quite conceivable that no single person was exclusively responsible and that two or even three of these suspects conspired together to get rid of Nizam al-­Mulk. The duo of Taj al-­Mulk and Terken Khatun are probable culprits although Malikshah himself cannot be excluded. Such personages did not need to hire a heretic from Alamut to perform the crime. Indeed, the sultan had only just sent armies there to attack the citadel and contact with Alamut asking for assistance in Nizam al-­Mulk’s murder just after the Seljuqs’ ignominious retreat from Alamut that same year38 seems unlikely.

a mu rd er ou s ye a r  | 127 Who Killed Malikshah? 1. Terken Khatun Telltale evidence for her possible involvement in the killing of Malikshah may be found in her actions after Malikshah’s death. Her cajoling and bribes to induce Malikshah’s troops to support her, and her persuading the caliph to pronounce the khutba in favour of her infant son, Mahmud, could be construed as the actions of someone who, even after the death of Nizam al-­ Mulk, had not been able to persuade her husband to change his mind about his heir. She therefore decided to dispose of Malikshah too, at a time when he was far from his home base in Isfahan and when the only one of his sons who had accompanied him to Baghdad was her own son Mahmud. Alternatively, she may have planned to murder both Nizam al-­Mulk and Malikshah all along. The method of murder is significant. Very few Seljuq sources are close to the events they record: a relatively early account, that of Ibn al-­Azraq39 writing in the second half of the sixth/twelfth century, mentions elliptically that Malikshah was poisoned. A century or more later, the historians of Iraq and Syria40 relate an elaborate story of rotten meat and a fever, details quite compatible with poisoning although Ibn Khallikan gives one account of Malikshah’s being poisoned with a toothpick.41 Terken Khatun is reported by some sources as having behaved suspiciously after Malikshah’s death. Ibn Khallikan notes that Malikshah’s funeral was conducted in a most private manner, no prayer said over the grave, no condolences paid, no hair cut off the tails of horses. Indeed, ‘one would have thought he had been snatched bodily from this world’.42 Ibn al-­Athir reports that Terken Khatun took Malikshah’s body to Isfahan having won over Malikshah’s soldiers with money, and having concealed the news of his death until she had arrived safely and secured her position there.43 2. Taj al-­Mulk It is conceivable that this man, either single-­handedly or in conspiracy with Terken Khatun, brought about Malikshah’s murder. It is possible that just as Nizam al-­Mulk had acquired his unassailable hold of the Seljuq empire when Malikshah was in his teens, so too Taj al-­Mulk hoped to be the de facto ruler under cover of Terken Khatun’s son, Mahmud.

128  |  the me d ie va l tur k s 3. The ‘Abbasid caliph al-­Muqtadi Whilst it is extremely unlikely that al-­Muqtadi would have been involved in the murder of his ally and supporter, Nizam al-­Mulk, there are, however, stories of his complicity in Malikshah’s sudden death. The details vary but the general impression is that Malikshah was in Baghdad on extremely serious business, namely to oust al-­Muqtadi from the caliphal office and to make his (Malikshah) own grandson, Ja‘far, caliph. This démarche was the last in a series of acrimonious exchanges between the caliph and the sultan which had occurred when Malikshah visited Baghdad. Malikshah’s daughter had married the caliph, al-­Muqtadi, in 480/1087, and had borne him a son44 but she had soon complained of bad treatment from her husband and had gone back to her father in Isfahan, where she died in 482/1090.45 This can scarcely have improved Malikshah’s feelings towards the caliph. According to al-­Subki,46 Malikshah left Isfahan for Baghdad in Ramadan 485, determined to remove the caliph. On arrival in Baghdad he found that al-­Muqtadi had already appointed another son of his, al-­Mustazhir, as his wali ‘ahd. Thereafter, a violent row must have erupted. Malikshah demanded that al-­Muqtadi should remove himself as soon as possible from Baghdad, leaving the way clear for Seljuq control of the caliphate within the Seljuq family. This was indeed a momentous design. According to some sources, Malikshah contemptuously asked the caliph to vacate Baghdad and go wherever he chose. Having failed to dissuade Malikshah from this course of action, the caliph asked for a period of ten days in which to prepare his departure.47 Before that elapsed, Malikshah was dead, at the tender age of thirty-­eight. These three people, Terken Khatun, Taj al-­Mulk and al-­Muqtadi, are the three major suspects in the murder of Malikshah. It is an open question which one person or combination of people performed the deed. One thing is clear: Malikshah died in his prime and not of natural causes. Postscript Within two years of the violent deaths of Nizam al-­Mulk and Malikshah most of the key figures in this power struggle had also been removed. By 487/1094, Barkyaruq, no doubt with the support of his mother Zubayda Khatun and the mamluks and family of Nizam al-­Mulk, had disposed of

a mu rd er ou s ye a r  | 129 al-­Muqtadi, Ja‘far, Taj al-­Mulk, Terken Khatun and Mahmud. The way was now clear, for a short time at least, for Barkyaruq to assume the sultanate, supported by the scions of Nizam al-­Mulk.48 This bloodbath heralded the shattering of the tenuous unity built up within the Seljuq empire not by the first three sultans but by the personal empire of Nizam al-­Mulk himself. General Observations As well as being an episode full of interesting historiographical problems and unanswered questions, the events surrounding the deaths of Nizam al-­Mulk and Malikshah raise a number of important issues. It would appear likely that the Assassins were not involved in Nizam al-­ Mulk’s murder. Later, Sunni writers from a milieu ruled by staunchly Sunni Turkish dynasties such as the Ayyubids would wish to shift the blame for Nizam al-­Mulk’s murder squarely onto the Assassins. Yet, the practice of such writers of compiling reports from earlier historians, reports which modify or contradict Assassin involvement in Nizam al-­Mulk’s murder, led them, in spite of themselves, to give a truer picture of the events. As for Assassin sources quoted by Mongol h ­ istorians – i­f indeed they are cited c­ orrectly – i­t is highly probable that the Assassins would wish to claim credit for the most devastating blow to Seljuq power, namely, the murder of Nizam al-­Mulk. It is even more possible that Mongol Sunni historians doctored Assassin accounts. As for the thorny question of who killed Nizam al-­Mulk, the burden of evidence from the sources points to Taj al-­Mulk, especially since the Nizamiyya murdered him as soon as they could lay their hands on him after his return to Persia from Baghdad. Taj al-­Mulk may well have been aided and abetted by Terken Khatun who also wanted Nizam al-­Mulk out of the way. As for Malikshah’s demise, this was probably at the instigation of the caliph, although once again the complicity of Terken Khatun cannot be ruled out. This episode demonstrates clearly the strong pressures working against any attempt at centralised government in the Seljuq empire, even in its heyday. The thirty-­year rule of one man, Nizam al-­Mulk, was a remarkable achievement, made possible only by his own subtle and ruthless personality and by a Mafia-­type family network. And it is all too often the fate of a Godfather to die a violent death.

130  |  the me d ie va l tur k s It is also apparent from this episode that the usual generalisation that the Seljuqs elevated the status of the caliphate from its humiliation under the Buyids requires modification, even during the heyday of Seljuq power, which is generally accepted as being the twenty years of Malikshah’s rule. Considerable tensions between caliph and sultan complicated the task of Nizam al-­Mulk whose own attitude towards the caliph in the latter part of his life seems to have been one of alliance and conciliation, perhaps in part as a response to his awareness of Malikshah’s increasing resentment at being under the thumb of his vizier. The caliph’s own attitude remains unclear, but it is certainly likely that al-­Muqtadi may well already have had the same political and territorial aspirations as were to become apparent in his successors in the caliphal office, al-­Mustazhir, al-­Mustarshid and above all al-­Muqtafi. As for Malikshah’s intentions vis-à-vis the caliphate, it would appear possible that he had intended to bring it definitively under sultanal control, within the Seljuq family itself, through his own grandson, the ‘Abbasid Ja‘far;49 indeed, it is possible to postulate that he was aiming at merging the two institutions, or even at abolishing the caliphate altogether. This is a timely reminder of the high stakes for which the various suspects in this drama were playing. 1092 was therefore not only the year in which two violent murders, emanating from within the Seljuq court itself, swiftly removed the two most powerful men in the Islamic world. 1092 also heralded the rapid disintegration of the Seljuq empire and revealed clear signs that the Seljuq sultans and the caliphs were on a collision course which was further to weaken the unity of the eastern Islamic world. To conclude, I will return to the question I originally posed: ‘Who killed Nizam al-­Mulk and Malikshah?’ As already mentioned, Malikshah was probably murdered by the caliph. His manner of death, whether of natural causes or from poisoning, did not fit well with the Assassins’ m ­ ethods – v­ iolent death in public ­places – ­and his murder was not attributed later on to the Assassin sect. As for the assassination of Nizam al-­Mulk this is a more weighty problem. It is of course conceivable that he was murdered by a member of the Assassin sect acting on the instructions of Hasan-­i Sabbah, as the sources state. I hope, however, that some doubts have been aroused about this. The event certainly does not deserve the confident generalisations which have

a mu rd er ou s ye a r  | 131 surrounded it and have endowed it with the status of the first important political act of terrorism perpetrated by the Assassins. Al-­Kunduri, the man whom Nizam al-­Mulk ousted to reach supreme power, was right to predict that Nizam al-­Mulk would fall into the pit which he himself had dug. Key to abbreviations used in the footnotes Akhbar = al-­Husayni, Akhbar ad-dawla as-saljuqiyya, ed. M. Iqbal (Lahore, 1933). al-­Bundari = al-­Bundari, Zubdat an-nusra wa-nukhbat al-‘usra, ed. M. T. Houtsma in Recueil de Textes relatifs à l’Histoire des Seljoucides II (Leiden, 1889). 2 EI = Encyclopaedia of Islam, 2nd edition. H. Mustawfi = Hamdallah Mustawfi, Ta’rikh-i Guzida, tr. C. Defrémery, Journal Asiatique (April–May 1848), 417–63; (September 1848), 259–79. Ibn al-­Athir = Ibn al-­Athir, al-Kamil fi’l-ta’rikh, ed. C. J. Tornberg (Leiden and Uppsala, 1851–76). Ibn al-­Azraq = Ibn al-­Azraq, Ta’rikh al-Fariqi, ed. B. A. L. ‘Awad (Cairo, 1959). Ibn al-­Jawzi = Ibn al-­Jawzi, al-Muntazam fi ta’rikh al-muluk wa’l-umam (Hyderabad, 1938–41) (editor unidentified). Ibn Khallikan = Kitab wafayat al-a‘yan: Ibn Khallikan’s Biographical Dictionary, tr. Baron W. M. de Slane (Paris, 1843–71). Ibn al-­Qalanisi = Dhayl Ta’rikh Dimashq, ed. H. F. Amedroz (Leiden, 1908). Juwayni = al-­Juwayni, Ta’rikh Jahan Gusha, tr. J. A. Boyle, The History of the World Conqueror (Manchester, 1958). Rashid al-­Din = Rashid al-­Din, Jami‘ al-tawarikh, Edinburgh University Library, Ms. Arab. 20. Rawandi = al-­Rawandi, Rahat al-sudur wa-ayat al-surur, ed. M. Iqbal (London, 1921). al-­Subki = al-­Subki, Tabaqat ash-shafi‘iyya al-kubra (Cairo, 1905–6).

132  |  the me d ie va l tur k s Notes  1. The Assassins (London, 1985), 134. For a similar view, cf. D. Morgan, Medieval Persia (London, 1988), 32.   2. Morgan, op. cit., 32.   3. Cf. C. E. Bosworth, ‘The political and dynastic history of the Iranian world (AD 1000–1217)’, in The Cambridge History of Iran (Cambridge, 1968), V, 101. C. Cahen, ‘The Turkish invasion: the Selchükids’, in M. W. Baldwin and K. M. Setton (eds), A History of the Crusades (Wisconsin, 1969), I, 154. Cahen’s phrasing on this point is somewhat confusing: ‘The Assassins, whose first victim of ­note – ­if indeed they were the perpetrators of the crime which was incited by his other ­foes – w ­ as the vizier, Nizam al-­Mulk’ (ibid., 161). Bowen (following Rashid al-­Din) tentatively suggests that Nizam al-­Mulk was killed by a combination of the vizier’s enemies at court and the Assassins. Cf. EI 2, s.v. Cf. also K. Rippe, ‘Über den Sturz Nizam-­ul-­Mulks’, in Fuad Köprülü Armajani (Istanbul, 1953), 428.   4. For a recent treatment cf. EI 2, Malik-­Shah (C. E. Bosworth). Although Houtsma draws on only a few sources, he firmly believes that Malikshah was murdered at the instigation of the caliph. He does not, however, employ the identical critical faculty when discussing the murder of Nizam al-­Mulk which he attributes in too facile a manner to the Assassins. Other possibilities mentioned in the sources (such as the involvement of Taj al-­Mulk and Malikshah in Nizam al-­Mulk’s murder) are dismissed by Houtsma as unsubstantiated rumours. He prefers instead to rely on some of the evidence of Ibn al-­Athir (X, 213 and 216). If the stories about the caliph’s role in the death of Malikshah merit credence, so too do the other anecdotes about Taj al-­Mulk and Malikshah; it is illogical in this mass of rumours to select only a few. Cf. M. T. Houtsma, ‘The death of Nizam al-­Mulk and its consequences’, Journal of Indian History 2/3 (1924), 147–60.   5. Al-­Kunduri died in 456/1064. His scrotum was stuffed with straw and sent to Nizam al-­Mulk at Kirman, where it was committed to the earth (Ibn Khallikan, III, 295).   6. ‘You have acted wrongly in teaching the Turks to put to death their wazirs and the chiefs of the civil administration: he that digs a pit shall fall into it’ (ibid., 294).   7. Cf. al-­Subki, III, 144; Akhbar, 67. According to Akhbar, Nizam al-­Mulk had more than 20,000 mamluks.

a mu rd er ou s ye a r  | 133  8. Cf. Akhbar, 66–7; H. Mustawfi, 459–60; Ibn al-­Qalanisi, 121; Ibn al-­Athir, X, 137; al-­Subki, III, 142; Ibn Khallikan, 414–5; Ibn al-­Azraq, 229; Rawandi 135; Rashid al-­Din, under year 485; Bundari, 62; Ibn al-­Jawzi, IX, 66–7.  9. Cf. Akhbar, 71; H. Mustawfi, 461; Ibn al-­Qalanisi, 121; Bundari, 70; Ibn Khallikan, III, 444–5; Ibn al-­Azraq, 229, Ibn al-­Athir, X, 142; Rashid al-­Din, under year 485; Ibn al-­Jawzi, IX, 62. 10. Cf. H. Mustawfi, 462; Rashid al-­Din, under year 485; Ibn al-­Athir, X, n. 145; Rawandi, 139–40. 11. Ibn al-­Athir, X, 145; Rashid al-­Din, under year 485. 12. Ibn al-­Athir, X, 154. 13. Ibn al-­Athir, X, 154; Akhbar, 75; Bundari, 68. 14. Akhbar, 75; Ibn al-­Athir, X, 154. 15. Cf. note 8. 16. Cf. the discussion by Lewis, op. cit., 146. 17. Juwayni, 676–7. 18. Ibid. 19. For a recent comprehensive treatment of the intellectual life of Baghdad in the time of al-­Ghazali and Nizam al-­Mulk, cf. E. Glassen, Der Mittlere Weg. Studien zur Religions-politik und Religiosität der späten Abbasiden-Zeit (Wiesbaden, 1981), 63–84, 131–76. 20. Juwayni, 676–7. 21. This is the view accepted by al-­Subki, who does, however, present other possible suspects (op. cit., 143). 22. Ibid., 139. 23. H. Mustawfi, 457–8, 461; Rawandi, 134; Ibn al-­Athir, X, 138. 24. Al-­Subki, III, 142; Ibn Khallikan, I, 413. 25. Ibn Khallikan, I, 413. 26. H. Mustawfi, 458. For a long discussion of this whole question, cf. Rippe, op. cit. 27. Akhbar, 67; al-­Subki, III, 142. 28. Al-­Subki, III, 143. 29. Al-­Bundari, op. cit., 63. 30. For details on the life of Taj al-­Mulk, cf. Bundari, 61; Ibn al-­Jawzi, IX, 46, 62–3, 74; Rawandi, 133–5; Ibn al-­Athir, X, 120, 142, 145–7. 31. Ibn al-­Athir, X, 145; Ibn Khallikan, I, 415. 32. Ibn al-­Athir, X, 147; Ibn Khallikan, I, 415; Bundari, 62–3; Akhbar, 67. 33. Rashid al-­Din, under year 485.

134  |  the me d ie va l tur k s 34. A Partial Genealogy of the Seljuqs Chagri Beg Alp Arslan

Yaquti

Terken Khatun

=

Malikshah

Da’ud (d. 474/1082)

Mahmud

=

Ahmad (d. 481/1088)

Zubayda Khatun Barkyaruq

Sanghar

Muhammad

35. H. Mustawfi, 456; Rawandi, 134. 36. Ibn al-­Qalanisi, op. cit.,121. 37. Ibn al-­Azraq, op. cit., 229. 38. Juwayni, 674–5. 39. Ibn Azraq, op. cit., 229. 40. Cf. the sources quoted in n. 9. 41. Ibn Khallikan, III, 445. 42. Ibid. 43. Ibn al-­Athir, X, 142–3. 44. Ibn Khallikan, III, 444. 45. Al-­Subki, III, 143. 46. Ibid. 47. Al-­Bundari, 70; Ibn Khallikan, III, 445; al-­Subki, III, 144; Ibn al-­Jawzi, IX, 62. 48. The chronology of deaths may be tabulated as follows: 485/1092 10 Ramadan/14 October 16 Shawwal/19 November

. . . . . .

Nizam al-­Mulk murdered Malikshah murdered

486/1093 12 Muharram/12 February Jumada 1/30 May–28 June

. . . . . .

Taj al-­Mulk murdered, aged 47 Ja‘far died, aged 6

487/1094 15 Muharram/4 February Ramadan/14 Sept–13 Oct.

} 36 days

. . . al-­Muqtadi murdered, aged 38 . . . Terken Khatun and Mahmud died Mahmud was aged 7; Terken Khatun’s age is unknown

a mu rd er ou s ye a r  | 135 49. It is tempting to read too much into al-­Subki’s words on this subject. What is clear, however, is that there was considerable tension between Malikshah and al-­Muqtadi (op. cit., 143).

10 Ibn al-‘Adim’s Biography of the Seljuq Sultan, Alp Arslan Introduction

I

bn al-‘Adim’s massive biographical dictionary, the Bughyat al-talab fi ta’rikh Halab has been unexploited by scholars until relatively recently. As late as 1973, Khalidi’s review of Islamic biographical dictionaries touched on nineteen such works from the early eleventh century to the late eighteenth century but did not mention the Bughya.1 This neglect of Ibn al-‘Adim’s magnum opus is now being rectified,2 and the Bughya, now available in both facsimile3 and printed editions,4 will no doubt soon assume its rightful place as prime source for the history of Syria, especially between the eleventh century and the author’s own time (he died in 660/1262). Among much else the Bughya is a source for early Seljuq history, in conjunction with chronicles and other surviving material. In particular, Ibn al‘Adim compiled a long biographical notice on the second Seljuq sultan, Alp Arslan (ruled 1063–72), to which discussion in this paper will be devoted. By focusing on one biography alone an attempt will be made to shed light on Ibn al-‘Adim’s aims and methods more generally and to evaluate the ‘historical material’ which he includes. It may well be that by concentrating on one such biography as a model, one will be able to assess the wider benefit to be gained by scholars of Seljuq history in trawling the remaining two thousand or so biographical notices contained in the Bughya. The paper will fall into two parts: firstly, a general discussion of the structure, sources and content of Ibn al-‘Adim’s biography of Alp Arslan, and secondly a longer analysis of some of the anecdotal material given by Ibn al-‘Adim. Thereafter a few tentative conclusions will be drawn. 136

ib n a l- ‘ ad im ’s b io gr a phy o f a l p a r s l a n  | 137 I. The Structure, Sources and Content of Ibn al-‘Adim’s Biography of Alp Arslan It is misleading in fact to speak of a structure in this context. Certainly, the format of the biography is only loosely chronological. It is true that Ibn al-‘Adim begins with a short introduction in which he gives vague details of Alp Arslan’s names, genealogy and birth, and the notice ends with several accounts of the sultan’s death. In between these extremities of the biographical notice, there is a veritable jigsaw of jumbled and disordered accounts and snippets of information which bear on certain aspects of the career and person of Alp Arslan. A limited number of events are given in several versions, often with only small variations in detail. The career of Alp Arslan can be pieced together from this structure only by those already familiar with the basic outlines found in other sources. In this compositive narrative Ibn al-‘Adim draws on sixteen named sources which he quotes often at length. The majority of attributed written excerpts come from the lost works of twelfth-­century Syrian h ­ istorians – I­ bn Zurayq, al-­Halabi, al-­Shaybani and ‘Ali b. M ­ unqidh – ­details of whom were unearthed painstakingly by Cahen in his book La Syrie du Nord à l’époque des Croisades.5 Important material is also quoted from further afield but only ­occasionally – ­such material is taken from a lost work of the Iraqi historian, Ghars al-­Ni‘ma the son of Hilal al-­Sabi’6 and from the sadly nonextant history of the early Seljuqs, the Maliknama, probably commissioned by Alp Arslan whilst he was in Khurasan7. Ibn al-‘Adim also uses oral sources, notably his own father, a well-­respected religious scholar in his own right, and the Baghdadi faqih, ‘Umar b. Tabarzad,8 with whom he studied when the latter came on a visit to Aleppo. There is therefore a predominantly Syrian dimension to Ibn al-‘Adim’s choice of sources, although he is able on occasion to draw on material from the eastern Islamic world. One narrative is particularly noteworthy in this c­ ontext – ­it is an extract which by Ibn al-‘Adim’s own admission has been taken from al-­Husayni’s mukhtasar of a work known as the Zubdat al-tawarikh, a dynastic history of the Seljuqs emanating from Transoxiana.9 Fortunately, al-­Husayni’s work has survived, and a comparison between the extract given by Ibn al-‘Adim and al-­Husayni’s own text reveals the wording to be absolutely identical.10

138  |  the me d ie va l tur k s This may well indicate that there is a high level of accuracy generally in the borrowings made by Ibn al-‘Adim from the historical chronicles which are no longer extant. What concrete information, ‘hard facts’, can be extracted from Ibn al‘Adim’s biographical notice on Alp Arslan? As far as the sultan’s life and career are concerned, the following outlines are given. Alp Arslan was born in 424 AH, he was appointed heir to Chagri Beg and fought against the Ghaznavids in 451, he became sultan in 455 or 457, he emerged triumphant from succession struggles with two rival Seljuq contenders, Qutulmush and Qavurt, he laid siege to Aleppo in 463, defeated the Byzantine emperor at the battle of Manzikert that same year and was killed in Transoxiana in 465. A very skeletal outline indeed for a career which covers twenty-­three pages in Sevim’s partial edition of the Bughya. Ibn al-‘Adim gives no overall assessment of Alp Arslan and he provides only a few stereotyped phrases about the sultan’s justice and authority. As a predictable consequence of his aim to include within his biographical dictionary only those persons who lived in or who visited Aleppo, Ibn al-‘Adim concentrates in his biography of Alp Arslan on the sultan’s involvement with Aleppo, focusing in particular on his siege of the city in 46311. Extensive coverage is also given to Alp Arslan’s victory at Manzikert12. The eastern Islamic sphere of Alp Arslan’s operations, the Iraq–Iran–Central Asian context, receives much less attention.13 Generally speaking, the whole of Ibn al-‘Adim’s account is impressionistic and seems to assume that, as regards the city of Aleppo at least, its readers know the basic facts already. It may well be implied that if they do not, they may consult the author’s own chronicle of Aleppan history, the Zubdat al-halab fi ta’rikh Halab for a clearer account and more chronologically structured information.14 What then is the value of the Bughya? A comparison of Ibn al-‘Adim’s coverage of the same events in both his works reveals that at times he fleshes out his accounts in the Bughya with more detailed anecdotal material than in the Zubda. II. An Analysis of Some of the Longer Anecdotes These are mostly stories which have been preserved in the region of Aleppo and have to do with Alp Arslan’s siege of the city. True to his Aleppan perspective, Ibn al-‘Adim gives great prominence to this event, although he

ib n a l- ‘ ad im ’s b io gr a phy o f a l p a r s l a n  | 139 never explains its significance in the general span of the sultan’s career, nor does he describe the siege in any systematic or detailed fashion. Instead, he produces a number of ‘stories’. One of those concerns a redoubtable woman, the mother of Mahmud b. Nasr b. Salih, the Mirdasid ruler of Aleppo.15 According to the main line of the ‘story’, after Alp Arslan’s protacted siege, this woman and her son went out to Alp Arslan to surrender to him the keys of the city. They made signs of obeisance to him. When Alp Arslan questioned the lady he was so impressed with her fine responses that he returned the city to the care of her son. A second ‘story’ from the siege of Aleppo is a local joke, at Alp Arslan’s expense, which Ibn al-‘Adim tells on the authority of his own father who in turn heard it ‘from his ancestors’. Alp Arslan set up a mangonel on one of the towers in Aleppo. The people of the city then fastened a large piece of satin to the tower, mocking him and saying that the tower had a headache because of the erecting of the mangonel.16 Amongst the anecdotes selected by Ibn al-‘Adim for his biography of Alp Arslan are two which are hostile to his subject. The first concerns the sultan’s drunkenness during the siege of Aleppo. The story is recounted with great gusto and relish. Unlike most of the other anecdotes chosen by Ibn al-‘Adim, this one is told only once17 and is quoted from the lost history of ‘Ali b. Munqidh (d. 1150)18 who allegedly heard it from his father who had been told it by Ibn Badi‘, the wazir of the Seljuq ruler, Tutush (d. 1095). Thus, the chain given goes back to a generation or so after the period of the siege of Aleppo. A summary of the story is as follows: Alp Arslan took some wine on board at Aleppo and was soon totally inebriated. He began calling for his attendants to bring in the ruler of the city, Mahmud, so that he could behead him. They did not carry out his orders but instead told the Seljuq wazir, Nizam al-­Mulk, what was happening. Nizam al-­Mulk incautiously went to Alp Arslan and began remonstrating with him. Instead of heeding his wazir, however, Alp Arslan, in whom ‘the inebriation had reached an extreme’, struck him with the washbasin, making a mark on his face. The wazir retired, hurt. In view of his own inability to deal further with the situation Khwaja Buzurg (the name by which Nizam al-­Mulk is known in this text) resorted to the power behind the throne, the khatun, who was providentially lurking behind a canopy, and he urged her to act quickly. One sentence from her

140  |  the me d ie va l tur k s was enough to do the trick. When her husband saw her, he asked, ‘What has brought you here, Khatun?’ She replied, ‘Go to sleep; you’re drunk.’ In the morning, when Alp Arslan was presumably extremely sensitive to the sound of his wife’s voice, she began haranguing him about his conduct the previous night. He, however, denied all knowledge of what had happened, exclaiming: ‘By God, I don’t know anything about all this.’ Later that morning, when Nizam al-­Mulk came to see Alp Arslan, the sultan said: ‘What is that mark on your face, Hasan?’ The wazir replied: ‘O sultan of the world, this is a mark which I got yesterday. I was outside my tent and the tentpole hit me.’ The narrator (Tutush’s wazir) comments at this point that the people approved of Nizam al-­Mulk’s behaviour in withholding from the sultan what had really happened. The sultan then moved away from Aleppo, retracing his steps to encounter none other than the Byzantine emperor (al-­Manzikert). Why does Ibn al-‘Adim include this anecdote in his biography of Alp Arslan? The Bughya appears to be the only work which gives it. Of course, the drunken propensities of the Turcomans in general are mentioned frequently in the medieval sources but the picture of the Seljuq sultan is usually carefully whitewashed in the chronicles to portray them as good Sunni rulers, promoting orthodoxy at home and jihad abroad against the heretic and infidel. There are other angles to the story. Its narrator, the brother of the more famous Usama, comes from the celebrated and highly cultivated Banu Munqidh. Its tone and the stylised nature of its dialogue are strongly reminiscent of similar stories, told by Usama to highlight the self-­evident social and cultural superiority of the Muslims over the Crusaders.19 Here, however, the target is the Turk. The author’s technique is simple, the attitude patronising. The sultan is first brought to her by his wife whose treatment of him is devastatingly effective. His humiliation is completed by the superiority of the Persian wazir who announces with a straight face that he has been hit by his tent pole, whilst privately he and the readers of this story have the last laugh at Alp Asrlan’s expense.20 The second negative anecdote about Alp Arslan demonstrates an even blacker humour. Unlike the preceding story which is given quasi-­historical authenticity by being placed at the siege of Aleppo, the second anecdote occurs ‘when Alp Arslan was a young man’.21 Presumably with the aim of giving even greater verisimilitude and to protect himself against the allegation

ib n a l- ‘ ad im ’s b io gr a phy o f a l p a r s l a n  | 141 of passing on mere tittle-­tattle about great men, Ibn al-‘Adim prefaces this anecdotes with a long isnad, the last name of which is Ghars al-­Ni‘ma who heard it from ‘a certain Khurasani’. The story goes as follows: in his youth Alp Arslan went hunting. He saw a weak old man in whose head was a thorn which had cut him and was causing him pain. Alp Arslan said to the old man, ‘Do you want me to give you relief from the pain, labour, illness, elderliness and old age in which you find yourself?’ The man with the thorn, thinking that Alp Arslan would give him relief from his suffering replied, ‘By God, yes, my master.’ Then Alp Arslan shot an arrow at him, killing on the spot. This macabre little anecdote ends with the comment (presumably from Ghars al-­Ni‘ma): This came from Alp Arslan in the state of adolescence and ignorance; the intoxication of youth drove him to it. As for when he was in his prime and firmly established in the kingdom, he was one of the most just and best behaved kings, one of those most desirous of jihad, the support of religion.

The deliberate inclusion of these two pejorative anecdotes should be set against the usual stereotyped panegyrics conventionally found in medieval Islamic biographical notices. This inclusion of negative material may be interpreted as an attempt by Ibn al-‘Adim to detract from the overall favourable picture given of the sultan’s achievements. These are in no doubt. Nevertheless, it is probably local, ethnic, religious and intellectual pride which dictates that the audience who read the Bughya should be reminded that Alp Arslan was a nomadic Turkish barbarian upstart, a newcomer to Muslim ways and to Islamic government and culture. One other important category of material concerned dreams. Accounts of dreams are of course a common feature of many medieval Islamic biographies. These usually occur towards the end of the biography and in this respect the dream described by Ibn al-‘Adim in his notice on Alp Arslan is no exception. Almost at the very end, during one of his accounts of the death of Alp Arslan, Ibn al-‘Adim includes a dream narrative. He is at pains to provide it with solid credentials, giving an isnad of four names beginning with ‘Umar b. Tabarzad al-­Baghdadi and culminating in the poet Abu’l-­Hasan al-­ Basrawi. The latter dreams that he saw one Abu Tahir who tells him that God has ‘struck Nishapur with seventy-­two sticks’. When the dreamer awakes he

142  |  the me d ie va l tur k s seeks an explanation of the dream from experts, one of whom tells him to count seventy-­two days and then see what happens in Nishapur. Thereafter the narrator points out that Alp Arslan had crossed the Jayhun (just before he was killed) on the seventy-­third day after the dream; this he finds truly amazing.22 Why did Ibn al-‘Adim choose to include this dream in his biography of Alp Arslan? It certainly adds nothing to the reader’s knowledge of the sultan’s life or death. However, the dream provides a link between this world and the Hereafter which Alp Arslan was shortly to enter. Ibn al-‘Adim’s readers may reflect that the life of the great sultan Alp Arslan, conqueror of the Byzantine emperor and ruler of vast realms, is also subject to God’s inexorable and preordained decree. Thus, the predilection of the medieval Arabic biographer for giving allegorical slants to his anecdotal material is clearly shown here. As with the dreams about Alp Arslan recorded in Ibn al-­Banna’s diary23, Ibn al‘Adim’s dream narrative is an allegorical prediction which is aimed to excite wonder and awe in the reader and to remind him too of his own mortality. III. Conclusions This brief study of some aspects of a single biography in the Bughya has raised a number of issues about the aims and methods of medieval Islamic biographers. Given that Ibn al-‘Adim penned many hundreds of biographies, it is perhaps dangerous or misleading to generalise too sweepingly from an analysis of only one of them. Nevertheless, it is possible to point out certain obvious trends and characteristics. What is the value to the historian of the Seljuqs of this account of the life of Alp Arslan? Obviously, the information is fragmentary at times, detailed and overlaid with local folklore on other occasions. There is no attempt to present a comprehensive and structured description with a chronological framework covering the life of Alp Arslan. That must be sought elsewhere. Instead, despite an encyclopedic range of sources, mostly of Syrian provenance, Ibn al-‘Adim’s biography of Alp Arslan provides disappointingly scanty and narrowly focused information. Parts of the biography are clearly intended to entertain rather than to inform. Others are aimed at moral edification. Side by side with a few conventional panegyrics about Alp Arslan’s courage and justice Ibn al-‘Adim chooses to include material which reveals something of Alp Arslan not as a remote ruler but as a ­person – ­drunk,

ib n a l- ‘ ad im ’s b io gr a phy o f a l p a r s l a n  | 143 impetuous and capable of motiveless malignity. Whilst Ibn al-‘Adim provides no real overall evaluation of his subject, the anecdotes selected do betray his lurking intention to belittle Alp Arslan’s achievements. Notes  1. T. Khalidi, ‘Islamic biographical dictionaries: a preliminary assessment’, The Muslim World 63 (1973), 53–65.  2. Most recently, Ibn al-‘Adim has been the subject of a detailed study by D. Morray, Ibn al-‘Adim and Bughyat al-talab fi ta’rikh Halab, unpublished DPhil thesis (University of Oxford, 1992).   3. Ed. F. Sezgin, 11 vols (Frankfurt, 1989).  4. Ed. S. Zakkar, 11 vols (Damascus, 1988). The following discussion in this paper is based on A. Sevim’s partial edition of the Bughya, entitled Biyografilerle Selçukular Tarihi Ibnu’l-Adim Bugyetu’t-taleb fi Tarihi Haleb (Ankara, 1982), 16–39.   5. (Paris, 1940), 41–4, 53.   6. Cf. the discussion by Cahen in ‘The historiography of the Seljuqid period’, in B. Lewis and P. M. Holt (eds), Historians of the Middle East (London, 1962), 61.   7. Cf. C. Cahen, ‘Le Maliknameh et l’histoire des origines seldjukides’, Oriens ii (1949), 31–65.  8. Cf. Ibn Khallikan, Wafayat al-a‘yan, tr. W. M. de Slane (Paris, 1843–71), vol. II, 387–8.   9. pp. 34–5. 10. Akhbar al-dawla al-saljuqiyya, ed. M. Iqbal (Lahore, 1933), 29–30. 11. pp. 17–19, 21–5, 27–9. 12. pp. 19, 25–6, 29–31. 13. Cf. for example, Ibn al-‘Adim’s sketchy account of the origins of the Seljuqs from a ‘village called al-­Nur’ (16). The latter part of the biographical notice (32–9) deals with material from the eastern Islamic world. 14. Ed. S. Dahan (Damascus, 1951–68), vol. II, 19–25. 15. For a detailed discussion of the Mirdasids, cf. S. Zakkar, The Emirate of Aleppo, 1004–1094 (Beirut, 1971). 16. p. 17. 17. pp. 27–9. 18. Cf. Cahen, La Syrie du Nord, 44–6. 19. Cf. the section of the Memoirs which deals with the Frankish character in P. K. Hitti, Memoirs of an Arab-Syrian Gentleman (Beirut, 1964), 161–70.

144  |  the me d ie va l tur k s 20. It is intersting to speculate on the work from which Ibn al-‘Adim drew this story. It is conceivable that he came across it through one of his (now lost) sources, the ‘Uyun al-tawarikh of Ghars al-­Ni‘ma b. Hilal al-­Sabi’. The latter drew in turn on the works of his father, one of which was entitled the Kitab al-Wuzara’. The relationship between chief minister and ruler had been of course a favourite topic of the Fürstenspiegel genre, which extolled amongst the desiderata of the ideal minister wisdom, mature age and the ability to keep secrets, to remove any bad habits from the ruler and to avoid intoxicating drink. Cf. Nasihat al-muluk, tr. F. R. C. Bagley (Oxford, 1971), 106–­13 – ­this part of the work was probably written by al-­Ghazali – ­and Kayka’us b. Iskandar, Qabusnama, tr. R. Levy (London, 1951), 218. 21. pp. 34–5. 22. pp. 37–8. 23. Cf. G. Makdisi, ‘Autograph diary of an eleventh-­century historian of Baghdad’, in History and Politics in Eleventh-century Baghdad (Aldershot, 1990), 29–30, 249–50.

11 The Power Struggle between the Seljuqs and the Isma‘ilis of Alamut, 487–518/​ 1094–1124: The Seljuq Perspective

T

he history of the Seljuqs’ struggle against the Isma‘ilis of Alamut has already been written several times, but there are always new insights to be gleaned from a detailed reading of the primary sources and I hope to supplement, and on occasion, modify, what has already been said by scholars such as Hodgson, Lewis and Daftary.1 As for the events of this period viewed from the Seljuq side, Sanaullah’s lengthy study, covering precisely these crucial years, regrettably does not tackle at all the Seljuqs’ relationship with Alamut.2 On the other hand, Kafesog˘lu’s emotional and simplistic condemnation of this period does not stand up to scrutiny: With the exception of Sultan Sanjar, the general history of the Seljuk state of Iraq and Khurasan is a chronicle of brave but inept rulers devoid of political sense and unworthy of their ancestors, ambitious and devious state officials, and batini crimes.3

His statement does, however, provide a convenient starting point for an analysis of the Seljuqs’ activities between 487/1094 and 518/1124. The Reign of Barkyaruq, 487–98/1094–1105 The year 485/1092 was a fateful one for the Seljuq state, removing in quick succession its twin pillars, the vizier, Nizam al-­Mulk, and the sultan, Malikshah, after a long period of governmental continuity. Almost immediately, Malikshah’s sons and relations made the traditional bid for power. Although by 488/1095 his eldest son, Barkyaruq, had achieved overall recognition in the western sultanate of Iraq and western Iran, his hold on 145

146  |  the me d ie va l tur k s power was always tenuous and, for the remainder of his reign, his energies were almost exclusively directed towards defending his position against all comers.4 In particular, until 497/1103–4, Barkyaruq was locked in a debilitating and expensive power struggle with his half-­brother, Muhammad, who first rose in opposition against him in 492/1099. During Barkyaruq’s rule, only Khurasan enjoyed some political continuity under his half-­brother, Sanjar, whom he appointed as his deputy to rule over the eastern provinces in 490/1097.5 In general, however, there is no doubt that the population of areas under Seljuq control in this period suffered at the hands of the Seljuq princes and governors: under the year 495/1101–2, Ibn al-­Qalanisi, in distant Damascus, comments on the situation in the Seljuq territories of Khurasan, Iraq and Syria, singling out ‘the protracted discord, enmity, wars, corruption and mutual fear’ experienced by the population, as a result of their governors being ‘too preoccupied with dispute and fighting to pay attention to them and to keep an eye on their affairs’.6 As well as misrule, the period of Barkyaruq’s reign is characterised by decentralisation, with the various Seljuq princes and Turcoman amirs often changing allegiances and jockeying for power, and by increasing fragmentation of the empire into semi-­autonomous principalities on its peripheries. Small wonder, therefore, that both the Crusaders and the Isma‘ilis were able to make inroads into Seljuq territory with such ease during these years. As is well known, it was during Barkyaruq’s reign that the Isma‘ilis of Alamut achieved their greatest successes in Seljuq territory, in terms of citadels taken, assassinations of public figures and infiltration of urban and court circles.7 In 489/1096, the Isma‘ilis seized Girdkuh near Damghan, a citadel on the route from Khurasan to western Iran. They also captured Shahdiz outside Isfahan, which was the key to control of the city, and many other fortresses in different parts of Iran.8 Kashani’s detailed list of the assassinations allegedly carried out by the Isma‘ilis and based on material found at the time of the Mongol conquest of Alamut should not be accepted uncritically, even though he supplies dates and the names of both victims and assassins. Indeed, there are wide discrepancies, for example, between some of his dates and those of Ibn al-­Athir and other chroniclers.9 It is clear, however, from all accounts that the crucial years in which assassinations occurred in the greatest number form a cluster

p ow e r st ru g gl e be twe en the sel j uq s a nd is ma ‘ il is  | 147 between 488/1095 and 493/1100, peaking around 490/1097. This timing is very t­ elling – ­predictably, it is at a moment of extreme disarray and weakness on the Seljuq side. As well as the seizure of citadels and the removal of prominent military and religious figures, there were rumours of the Isma‘ili ‘contagion’ infiltrating the Seljuq army and court circles. This moment of extreme vulnerability will be examined in greater detail below. Most of the sources are harsh on Barkyaruq, in spite of the genuine difficulties that any Seljuq ruler would have had to face in establishing himself in an empire, bereft of the towering figure of Nizam al-­Mulk. To be sure, the latter’s sons were in plentiful supply and a number of them were to fill important positions in this period. They were not, however, of the calibre of their father and, above all, lacked his ability to stay in office for a very long time.10 Not only did Barkyaruq lack good administrative support, he seemed also to have alienated the Nizamiyya, the late vizier’s crack troops, who defected to the side of Muhammad and Sanjar.11 It is as if most of the chroniclers, normally so reticent, are looking in their reflections on this dire period of Seljuq weakness to find a scapegoat. It is not difficult to load blame onto Barkyaruq. In addition to a doleful tale of successive intrusions from the Crusaders and the dreadful havoc wrought by the Isma‘ilis, the chroniclers provide smears and innuendoes. Al-­Husayni describes Barkyaruq as a drunkard,12 whilst al-­ Bundari speaks of his having a ‘bad character’, ‘associating with youths’ and indulging in frivolous music-­making.13 It was also Barkyaruq’s misfortune to be afflicted with smallpox and piles,14 ailments which scarcely enhanced his public image. There are no concessions made here to Barkyaruq’s extreme youth and inexperience, although these same two qualities of youth and inexperience in his two half-­brothers, Muhammad and Sanjar, are tacitly condoned. But of course the chroniclers are writing with the benefit of hindsight, and Barkyaruq’s brothers both had indisputably more successful reigns. The worst smear on the part of the chroniclers of Seljuq history is the insinuation that Barkyaruq himself had Isma‘ili sympathies. To what extent is this true? After all, he himself had been the target of a failed assassination attempt.15 It is worth looking closely at the evidence provided by Ibn al-­Jawzi and Ibn al-­Athir on this matter. Both choose the year 494/­1101 – j­ust after the most widespread spate of reported ­assassinations – ­in which to reflect

148  |  the me d ie va l tur k s more generally on the rise of the Batinis.16 The coverage of Barkyaruq’s alleged Isma‘ili proclivities is fuller in Ibn al-­Athir’s account and will be followed here. This historian mentions that there was considerable resentment in Isfahan at the fact that most of the amirs killed in the recent wave of murders, such as Arghush and Surmuz, belonged both to the Nizamiyya and to Muhammad’s side. This threw suspicion onto Barkyaruq. The details given by Ibn al-­Athir indicate that an atmosphere of paranoia and intrigue must have pervaded both the army and the court. The situation deteriorated further when, emboldened by Muhammad’s defeat of Barkyaruq, the Isma‘ilis infiltrated the latter’s army and nearly gained the upper hand. Barkyaruq was then called upon by his close associates, who had taken to wearing chain mail under their clothes, to act before it was too late, especially in view of the fact that he himself had ‘Isma‘ili leanings’. Barkyaruq therefore rooted out and killed many suspects, including the alleged leader of the Isma‘ili troops, Muhammad b. Dushmanziyar, from Yazd. He had acted vigorously, if rather belatedly and under pressure. But this does not necessarily imply that he was a secret Isma‘ili sympathiser. He was short of troops and money and may well have been ill, weary of warfare and ill-­advised. The evidence against him remains opaque and ambivalent. Also in 494/1101, Sanjar took action against the Isma‘ilis, sending an expedition under the leadership of an amir of his, Boz-­Qush, to Kuhistan (Arabic, Quhistan). Boz-­Qush was persuaded by bribes to lift his siege of the Isma‘ili stronghold of Tabas Masinan, thus allowing the Isma‘ilis to recoup their losses thereafter.17 On a second campaign in 497/1104, Boz-­Qush granted the inhabitants of Tabas safe conduct. Ibn al-­Athir comments that ‘many people were angry with Sanjar about this’.18 Posterity has perhaps judged Barkyaruq a little harshly. Certainly, moments of compromise with the Isma‘ilis are condoned in his half-­brother, Sanjar, but condemned in him. Even Lewis follows the line laid down by medieval chronicles, speaking of Barkyaruq’s ‘complacency or worse’19 vis-­ à-vis the Isma‘ili threat. The evidence from the sources suggests that neither Barkyaruq nor Sanjar nor Muhammad was strong enough singly to resist the necessity of using whatever troops were available to them, even if they were ‘Isma‘ili’. All three were accused at some time of using the Isma‘ilis to dispose of their enemies. The reality was that none of the three brothers, including

p ow e r st ru g gl e be twe en the sel j uq s a nd is ma ‘ il is  | 149 Muhammad, considered the extirpation of ‘heretical’ Isma‘ilis to be a major part of their military strategies. Their youth, inexperience and preoccupation with each other allowed the Isma‘ilis to expand during Barkyaruq’s reign. Barkyaruq’s massacre of some eight hundred suspected Isma‘ilis in Isfahan was his sole decisive measure aimed at stemming their growth and infiltration. The Reign of Muhammad Tapar, 498–511/1105–18 Muhammad’s reign witnessed not only a diminution in the number of assassinations attributed to Alamut but also a greater political stability. This was possible both because of the harmonious relationship which he enjoyed with his uterine brother, Sanjar, and because of the continuing support of the Nizamiyya for Muhammad. Although his rule was not without crises, the spectre of the almost uninterrupted wars of Barkyaruq’s reign had gone. In view of this improved political landscape it might be expected that Muhammad would be able to turn his attention to the ‘enemy within’, the Isma‘ilis of Alamut. What measures did he in fact take against them during the thirteen years of his reign? His first move came quickly, as early as 500/1106–7, when he recaptured the citadel of Shahdiz which had fallen into Isma‘ili hands. This victory is seized upon eagerly by the chroniclers, who were no doubt weary of recounting Seljuq weakness and Isma‘ili success. The Damascene historian, Ibn al-­Qalanisi, who includes in his city chronicle snippets or longer extracts from other parts of the Islamic world, triumphantly interrupts his narrative for the year 500 AH to provide a lengthy text purporting to be a fath-nama written by Muhammad’s scribe, Abu Nasr b. ‘Umar al-­Isfahani, after the re-­conquest of the citadel of Shahdiz and intended to be read from the minbars of the empire.20 Couched in high epistolary prose, this victory proclamation finally affords the Sunni world a pretext to exult at the ‘shedding of a river of blood of the heretical Batiniyya’.21 Muhammad thus became the hero of the hour. He had acted early in his reign against the Isma‘ilis and he had won. Here begins the process whereby Muhammad’s activities are embellished by the Sunni chroniclers and result in the creation of his image as ‘the strong he-­camel’ of the Seljuqs.22 Yet it would be stretching historical credibility to attribute Muhammad’s prompt action against Shahdiz to ideological fervour. In this, the only anti-­Isma‘ili initiative of his whole reign in which he participated personally, his motives

150  |  the me d ie va l tur k s were, in part anyway, more practical. He was forced to deal with the problem of the Isma‘ilis in Isfahan since this was his power base and the Isma‘ilis were well entrenched there, with widespread support. They had even been reaping the benefits of tax revenues from the outlying areas around Isfahan. Here, it could be argued, he had no choice but to act. The Seljuq armoury and treasury were there, and the citadel at Shahdiz held the key to domination of the city.23 Isfahan was the traditional centre of Seljuq rule and Muhammad’s own power and prestige were linked with possessing it. So Muhammad began with a siege of the citadel of Shahdiz, an action prolonged by what Ibn al-­Athir presents as delaying ploys by the Isma‘ilis inside. Finally, however, Muhammad successfully seized and destroyed the citadel and killed the Isma‘ili leaders, Ahmad Ibn ‘Attash and his son. An important, indeed essential, victory had been achieved. The fall of Shahdiz seems to mark the end of extensive Isma‘ili power in Isfahan, although the burning of the Friday mosque in 515/1121–2 was blamed on the Isma‘­ilis – a­ n unlikely deed for them to have perpetrated, since such tactics were not a usual feature of their anti-­Sunni strategy.24 Henceforth, until his death in 511/1118 Muhammad kept Isfahan as the main centre of his power. The same year Muhammad moved against his own vizier, Sa‘d al-­Mulk, seized his possessions and crucified him at the gate of Isfahan. Sa‘d al-­Mulk was accused of treachery against the sultan but four other close associates, suspected of Batini beliefs, were killed with him. Given that Muhammad had endured the cost of the lengthy siege of Shahdiz and that the mulcting or killing of ministers of state often occurred at moments of economic difficulty, Muhammad’s killing of these five could have had a motive much more mundane than zeal against the Isma‘ilis.25 The finger of suspicion was pointed at many. When a suspected Isma‘ili, one Surkhan b. Kaykhusraw al-­Daylami, was brought before Muhammad, Ibn al-­Athir uses the opportunity to point out Muhammad’s impeccable Sunni credentials, putting the following pious words into the sultan’s mouth: I have made a vow to God that I will not kill a prisoner. (But) if it is established that you are a Batini, then I will kill you.26

The next move initiated by Muhammad against the Isma‘ilis struck at the heart of their power, Alamut, to which the sultan sent an army, probably in

p ow e r st ru g gl e be twe en the sel j uq s a nd is ma ‘ il is  | 151 503/1109–10, under the command of his vizier, Ahmad b. Nizam al-­Mulk.27 According to Ibn al-­Athir, the siege was eventually lifted because ‘winter took them by surprise and they returned without achieving their goal’.28 The same admission of honest failure is made by Juwayni.29 Al-­Husayni, however, obscures this fact by alleging that the Seljuq army had ‘inflicted a defeat on the Batinis’.30 The most ambitious anti-­Isma‘ili offensive of Muhammad’s reign was conducted by Anushtegin Shirgir against Alamut in 511/1118.31 The siege was long, lasting around nine months.32 It is likely that a familiar device is used by the chroniclers to obscure the unpalatable fact that the siege had failed. Certainly, Juwayni argues that in the reign of Malikshah, the campaign led by Qizil Sarigh into Kuhistan in 485/1092 had failed because of the untimely news of Malikshah’s death: Before he could take the place, however, he received the news of Malik Shah’s death, whereupon he raised the siege and his army dispersed.33

An identical argument is used by Juwayni for the lifting of the siege by Shirgir in 511/1118. Pious deception dictates that the sources are unanimous in their assurance that, as al-­Husayni puts it, Shirgir ‘was on the point of taking Alamut’ when the news of the sultan’s death arrived.34 Thereafter, there was no hope of continuing the siege; the commanders left immediately and their supplies were taken by the inhabitants of Alamut. Once again, fate had intervened and snatched the victory from the Seljuqs: there is no reference to the tenacity of the Isma‘ilis in withstanding the latest siege of their stronghold. Tailpiece After Muhammad’s death, the Isma‘ilis were given a respite by the ensuing power struggles of the Seljuqs. Certainly they remained entrenched in Alamut. In Kuhistan they regained enough strength to force Sanjar to come to terms with them.35 Sanjar, who had not seen the need to come in person to western Iran during the rule of his brother Muhammad to join forces against the Isma‘ilis and root them out, came in great pomp and with a massive army, in a show of strength in 513/1119–20 against his nephew, Mahmud b. Muhammad.36 Well before the death of Hasan-­i Sabbah in 518/1124 Sanjar had become the uncontested leader of the Seljuqs.37 Yet he does not seem to

152  |  the me d ie va l tur k s have felt that this position of responsibility made it incumbent upon him to grapple seriously with the Isma‘ili problem. Some Historiographical Considerations The treatment of this period by Ibn al-­Athir is extensive and thorough and raises a number of interesting issues. Periodically in his Universal History Ibn al-­Athir breaks with the traditional annalistic format to reflect more generally on an historical topic of significance. Like Ibn al-­Jawzi, he accords the emergence of the Isma‘ilis of Alamut a special excursus under the year 494/1101.38 Individual incidents which involve the Seljuq–Isma‘ili encounter are dealt with in their proper chronological place. When recounting the activities of the Isma‘ilis in Isfahan, Ibn al-­Athir speaks about them with a kind of invective reminiscent of Nizam al-­Mulk or Juwayni: When this affliction became widespread among the people in Isfahan, Almighty God allowed their covers to be ripped asunder and vengeance to be exacted from them.39

According to Ibn al-­Athir, after Muhammad’s accession there was nothing more pressing for the new sultan’s attention than to deal with the Isma‘ili threat: When Muhammad became sultan and nobody opposed him any longer, there was no matter more important for him than attacking the Batiniyya, fighting them and exacting justice for the Muslims against their tyranny and oppression.40

Ibn al-­Athir also devotes several pages to an obituary of Muhammad. It is a masterpiece of exaggeration and pious deception. He is described with the usual array of laudable e­ pithets – ­just, virtuous and courageous. An example of his justice was his removal of illegal taxes and duties (mukus and dara’ib) in all his territories. The rest of the obituary is devoted to the sultan’s activities against the Batiniyya: Amongst the finest of his deeds (mahasin) was what he did with the Batiniyya, as we shall relate. We have already mentioned what he undertook in the way of besieging their citadels. Here we will mention his increased concern

p ow e r st ru g gl e be twe en the sel j uq s a nd is ma ‘ il is  | 153 about them, for, when he realized that the well-­being of the faithful and the country lay in the wiping out of their traces and the destruction of their fortresses and citadels, he devoted himself indefatigably to attacking them.41

High-­sounding words indeed. But, as we have seen, it is debatable whether Muhammad had the eradication of the Isma‘ilis as his prime concern. Nor did he ever fight personally against them at Alamut. His only personal involvement was to recapture his own power base in Isfahan, an action that could quite well have been prompted by self-­interest. The subsequent evidence in the obituary which Ibn al-­Athir produces for Muhammad’s anti-­Isma‘ili zeal consists of a detailed account of the siege of Alamut in 511/1118. Muhammad delegated the conduct of this enterprise to Anushtegin Shirgir, the governor of Sawa, and it is to this man’s activities rather than those of his ostensible subject, that Ibn al-­Athir devotes the remaining section of his obituary. Historical veracity compels Ibn al-­Athir to admit Seljuq failures in the past. There is a brief but vague reference to unsuccessful attempts by Muhammad to take Alamut: The sultan sent troops against him (Hasan-­i Sabbah), as we have already mentioned, and they returned without achieving (their) goal.42

There then follows an account of the successful siege of Alamut, lifted on the news of Muhammad’s death. An inadequate obituary, by any standards. So how much does Muhammad deserve the eulogies given to him by Ibn al-­Athir? How hard did he fight a jihad against the ‘heretics’ and rescue the fortunes of the Seljuq house? It would appear that the image of him created by Ibn al-­Athir and others is a pious figment. The contemporary epigraphic evidence on Muhammad’s titulature is conventional. The undated inscription at Gulpayagan in his name gives him the following laudatory and quite lengthy epithets: ‘the just king (malik), the exalted sultan, the most venerated king of kings (Shahanshah), the client (mawla) of the Arabs and non-­Arabs . . . ­the associate (qasim) of the Commander of the Faithful’.43 But there is no reference here, as Ibn al-­Athir would wish us to believe, to Muhammad being a fighter of jihad against heretics in general or the Isma‘ilis of Alamut in particular. The perception of Sultan Muhammad as a zealous ‘smiter of heretics’ seems to have accrued to him in the strongly Sunni, jihad-­pervaded milieu of

154  |  the me d ie va l tur k s Zengid and Ayyubid times, when the achievements of the Turkish military leaders in this sphere were attributed by the chroniclers of that period retrospectively to their Turkish predecessors, the Great Seljuqs. All this goes some way to explaining the attitude of Ibn al-­Athir towards Sultan Muhammad and his careful but easily overlooked inflating of Muhammad’s role in the Seljuq attacks on the Isma‘ilis of Alamut. His death marked the end of effective Seljuq rule in the western sultanate and his reign signalled a slowing down of Isma‘ili successes. He was the last of the line and as such, was a suitable subject for historiographical remoulding. There is disagreement in the sources on the actual number of Seljuq campaigns against Alamut. According to Juwayni and other Persian historians,44 the Seljuqs sent campaigns against Alamut for eight successive years, whilst other sources mention only two specific attempts; Ibn al-­Athir hints rather vaguely that there may have been other unsuccessful ventures.45 On the one hand, the version offered by Juwayni implies firm Seljuq commitment to the destruction of the Isma‘ilis’ centre; on the other hand, the account given in the remaining sources suggests a more piecemeal effort on the part of the Seljuqs. What interpretations can be made on the basis of these conflicting views? It could be argued that it is certainly more plausible to give credibility to the latter interpretation, since the main body of sources for this period, and above all, Ibn al-­Athir, are at pains to praise the Seljuqs’, and notably Sultan Muhammad’s achievements against the Isma‘ilis. Why would they omit to mention such a major fact as eight successive campaigns against Alamut? This number of campaigns would be a clear sign of commitment to jihad and the eradication of ‘heresy’. The other line of argument would be that the main body of sources would naturally wish to minimise the fact that the Seljuqs failed in their attempts at destroying the Isma‘ilis’ centre at Alamut and therefore did not reveal the true extent of the Seljuqs’ unsuccessful efforts against them. Juwayni’s possible desire to exaggerate the number of Seljuq campaigns and to dwell on their failure is also explicable. After all, it was the Mongol destiny to achieve the definitive conquest of Alamut in the fullness of time: Yet when the time was not yet ripe, during the reign of Hasan-­i-Sabbah, that same castle of Alamut, whose garrison and resources were then but small, had during a period of 11 years been several times besieged by Muhammad,

p ow e r st ru g gl e be twe en the sel j uq s a nd is ma ‘ il is  | 155 son of Malik-­Shah, son of Alp-­Arslan, (as may be read in books of history), and all to no avail.46

Thus, Juwayni may well have had his own reasons for inflating the actual number of failed Seljuq expeditions against the Isma‘ilis at Alamut, and for suggesting that there were eight rather than two. As Morgan rightly argues, it was hardly coincidental that Juwayni makes the Mongol destruction of Alamut the climax of his book. Juwayni ‘was able to discern some silver linings in the Mongol clouds’.47 One such ‘silver lining’ was indeed Hülegü’s virtual extermination of the Isma‘ilis of Alamut. Juwayni, as befits a public servant of the Mongols, omits all reference to their sack of Baghdad and the Mongols’ murder of the last ‘Abbasid caliph in 656/1258 but deems it historiographically opportune to exaggerate the number of Seljuq failures at Alamut in order to pinpoint with greater force the magnificent achievement of the Mongols in ridding the Sunni world of its longstanding enemy. In his obituary notice of Muhammad, Ibn al-­Athir mentions, in passing, an interesting point. He writes that when the Batiniyya were suffering extremes of hardship, the news reached them of the death of Sultan Muhammad. At this point, the morale of the Batiniyya rose. He then adds the following little detail: The news reached the troops besieging them [the Batiniyya] a day after they [the Batiniyya] learned of it.48

This cryptic detail remains unexplained. One may wonder how this occurred and why Ibn al-­Athir mentions it. Unwittingly, however, Ibn al-­Athir points to the more effective communications system of the beleaguered Isma‘ilis who heard this important piece of news before the Seljuq troops did. This detail is supported by the archaeological evidence which reveals the intervisibility of the network of Isma‘ili fortresses in Daylam and which would permit fire, light or smoke signals to be exchanged, to say nothing of messages by carrier pigeons. It is known from archaeological evidence, for example, that Ghutinar had intervisibility with Alamut, Lamasar and Maymundiz.49 Did Hasan-­i Sabbah and other prominent Isma‘ili figures in this period have some kind of standing army or irregular troops on which they themselves could draw or which could even be hired out to other military leaders?

156  |  the me d ie va l tur k s There are some tantalisingly brief but revealing references in the sources which support the latter hypothesis. As early as 493/1100, some 5,000 ‘Batini foot soldiers’ are mentioned as being part of the troops of Amirdad Habashi, the lord of the stronghold of Girdkuh, in a battle against Barkyaruq. These foot soldiers apparently belonged to Isma‘il al-­Kalkali, the lord of Tabas.50 Al-­Husayni has Sanjar fighting on Barkyaruq’s side in this battle,51 but according to Ibn al-­Athir, Sanjar fought with Habashi’s troops.52 For the year 497/1103–4 Ibn al-­Athir mentions a great gathering of Isma‘ilis from the area of Bayhaq who were on the rampage; this was caused, he adds, by the neglect of the three sultans.53 In 501/1107–8, Sadaqa gathered 20,000 horsemen of the ‘scattered bands of Kurds, Turks, Daylamis and Arabs’.54 Of course, Daylamis were not necessarily synonymous with Isma‘ilis. Under the year 513/1119–20, Ibn al-­Athir mentions the name of Ghuzzog˘lu as being the leader of the ‘Isma‘ili Turks’.55 As for Sanjar’s mighty army which accompanied him in 513/1119–20 across Khurasan to Rayy to terrify his nephew, Mahmud, into submission, this included five kings and ‘thousands of Batiniyya and infidels’.56 It is unclear who these troops were or what their status was in the Seljuq armies, but it is likely that they were hired for specific engagements and could appear on either side of an inter-­Seljuq power struggle. Clearly, however, ideological questions played little part in all this, although the Isma‘ili presence is deemed worthy of mention by the chroniclers. Their involvement in Seljuq military activities suggests that, even in Hasan’s lifetime, the Isma‘ili policy of the assassination of major figures was giving way to the strategies and survival tactics of other small territorial leaders involved in the power struggles of the day. By 518/1124 the Isma‘ilis were one group among many, vying for power in localised contexts. The references to their military participation in different power struggles of the time suggest that their presence was more tolerated in the Iranian context than later medieval chroniclers, with their stylised anti-­Isma‘ili diatribes, would have us believe. Conclusions It is apparent from the preceding discussion that there are still new insights to be gleaned from a subject which might be regarded as almost too familiar in scholarly writings, especially when those scholars who have worked on

p ow e r st ru g gl e be twe en the sel j uq s a nd is ma ‘ il is  | 157 this topic are of a high calibre and write persuasively. Yet there are still unanswered questions, both about individual episodes and the grander sweep of events. It is all too easy to accept at face value what medieval chroniclers say. Much more work needs to be done on the historiographical jigsaw on which the general history of the Seljuq–Isma‘ili struggle has been based. The Sunni sources of the sixth/twelfth and seventh/thirteenth centuries generally try to inflate the Seljuq achievement against the Isma‘ilis of Alamut. This is especially the case with Sultan Muhammad. Juwayni, who draws on material found at Alamut, suggests greater activity on the part of the Seljuqs at certain stages against the Isma‘ilis and a larger number of actual campaigns against Alamut. There may well be propaganda motives behind t­ his – n ­ amely to glorify the Mongol conquest of Alamut. The combined forces of Muhammad and Sanjar might have been able to remove the Isma‘ilis from Alamut and thus to eradicate their major military centre. The threat of undercover Isma‘ili operations, aimed at conversion and subversion, at court and in the major cities of the Seljuq empire, would have remained. This more ‘pernicious contagion’ would be more difficult to isolate and contain. It is, however, difficult to prove its existence definitively from the primary sources, which resort to hysterical and far-­fetched stories. Muhammad, Barkyaruq and Sanjar never formed a united front against the Isma‘ilis of Alamut in the period of Isma‘ili growth in Iran, 485–98/1092–1105. Muhammad and Sanjar could, on occasion, unite against Barkyaruq but never in the latter’s reign did all three brothers attack their common enemy within. This lack of solidarity and common cause was scarcely surprising: all three had expectations of patrimonial inheritance and would not cede their share without military struggle. Three chunks of patrimonial inheritance were to be seized and retained from the Great Seljuq empire ruled by Malikshah. Sanjar was relatively secure in the east. But the western territories (western Iran and Iraq) were the scene of a grim and protracted struggle between Barkyaruq and Muhammad which lasted almost until Barkyaruq’s death in 498/1105 and which weakened both sides. Five battles between Barkyaruq and Muhammad must have demanded large resources and depleted their finances. To continue to be able to pay their troops was the prime task. Any concerted offensive against the Isma‘ilis was, therefore, not a priority or even a major objective for either brother at this

158  |  the me d ie va l tur k s point. Could Muhammad and Sanjar together have defeated the Isma‘ilis at Alamut? This is a difficult question to answer. Numerically, they would or could have been large but they lacked the aura of invincibility of the Mongols when they finally came to Alamut and the weapon of terror that such invincibility instilled in the future victims. Later on, after 498/1105, when Muhammad was uncontested ruler in the western sultanate, he must have had severe financial problems, at least for parts of his reign. There were other demands on his treasury, such as financing campaigns against the Crusaders, as he was later pressurised to do. There was clearly a limit to the amount of time and financial resources he could afford to invest in military campaigns and sieges against Alamut and other Isma‘ili centres. After all, he had to take steps to keep himself in power and there were also other calls on his attention between 498/1105 and 511/1118. However many campaigns were sent to Alamut, it is telling that Sultan Muhammad never went there in person to attack the enemy. He always delegated this task, either to one of the sons of Nizam al-­Mulk (who had good reasons for hating the Isma‘ilis), or to a competent military commander, notably Shirgir. This stance by Sultan Muhammad, during the thirteen-­ year period of his reign, suggests, at the very least, that he did not make the extirpation of the Isma‘ili threat his top priority. The sources do not, however, indicate that there were more pressing needs than this threat posed by the ‘heretical’ enemy from within. It is clear that the Crusaders, who were actively consolidating their position in Syria and the Holy Land, and who might have constituted an alternative and valid outlet for Muhammad’s energies, were only peripheral to his activities. So, the most that can be said of Sultan Muhammad’s efforts against the Isma‘ilis is that they were piecemeal and delegated to his associates. Their efficacy was limited. Certain facts can be summarised at the end of this discussion. The Seljuqs did not eradicate the overall Isma‘ili threat but they did remove the major threat of Isma‘ili power at Isfahan. It is clear that they devoted some, but not all, of their energies to attacks on the Isma‘ilis. Much of the ‘blame’ for the Seljuqs’ failure to wipe out the Isma‘ilis within their own territories lies with them and is not to be attributed to Isma‘ili dynamism alone. The possibility remains that the combined forces of Muhammad and

p ow e r st ru g gl e be twe en the sel j uq s a nd is ma ‘ il is  | 159 Sanjar could have definitively wiped out the Isma‘ilis at Alamut in the way that Hülegü’s troops eventually did. The factors which contributed to the Seljuqs’ failure to do this include the atomisation of Seljuq power, the lack of concerted Seljuq military activity, the problem of financial r­ esources – e­ ven under Sultan M ­ uhammad – ­and the probability, barely addressed in Seljuq sources, that the new Isma‘ili preaching was not confined to known Isma‘ili strongholds but that it had infiltrated much more deeply into many areas of Iran, both urban and rural, both in the Seljuq armies and at court. As such, it was much more difficult to stamp out. Altogether, then, at this dynamic stage of Isma‘ili propaganda, under the leadership of Hasan-­i Sabbah, there was a lack of real concerted will or effort in the Seljuq empire to deal with the Isma‘ilis. Notes   1. See M. Hodgson, The Order of Assassins (The Hague, 1955), 73–99; B. Lewis, The Assassins (London, 1967), 43–63, and F. Daftary, The Isma‘ilis (Cambridge, 1990), 351–71.   2. M. F. Sanaullah, The Decline of the Seljuqid Empire (Calcutta, 1938).  3. G. Leiser, A History of the Seljuks: Ibrahim Kafesog˘lu’s Interpretation and the Resulting Controversy (Carbondale and Edwardsville, 1988), 56.  4. Sadr al-­ Din ‘Ali al-­ Husayni, Akhbar al-dawla al-Seljuqiyya, ed. M. Iqbal (Lahore, 1933), 76; Ibn al-­Qalanisi, Dhayl ta’rikh Dimashq, ed. H. F. Amedroz (Leiden, 1908), 137, 139–40, and Abu’l-­Qasim Kashani, Zubdat al-tawarikh; bakhsh-i Fatimiyan va Nizariyan, ed. M. T. Danishpazhuh, 2nd edn (Tehran, 1366/1987), 115.   5. Ibn Khallikan, Wafayat al-a‘yan, ed. I. ‘Abbas (Beirut, 1968–72), vol. 2, 428.   6. Ibn al-­Qalanisi, Dhayl, 140.   7. See Hodgson, Order, 75–7. Kashani (Zubdat al-tawarikh, 115–16) comments, for example, that the Crusaders were able to take advantage of the strife between the Seljuq princes of Syria, Ridwan and Duqaq.  8. See the names of citadels provided by Ibn al-­Athir, al-Kamil fi’l-ta’rikh, ed. C. J. Tornberg (Leiden, 1851–76), vol. 10, 215–18.  9. Kashani, Zubdat al-tawarikh, 154–7. The death of Ahmadil, for example, is recorded by Kashani for the year 493/1100, whereas for Ibn al-­Athir (al-Kamil, vol. 10, 361), who is usually accurate with his chronology, it occurs as late as 510/1116–17.

160  |  the me d ie va l tur k s 10. See, for instance, the list given in C. L. Klausner, The Seljuk Vezirate: A Study of Civil Administration, 1055–1194 (Cambridge, MA, 1973), 106. 11. Ibn al-­Athir, al-Kamil, vol. 10, 220. 12. Al-­Husayni, Akhbar, 78. 13. Al-­Fath b. ‘Ali al-­Bundari, Zubdat al-nusra, ed. M. T. Houtsma (Leiden, 1889), 84. Not all the chroniclers, however, indulge in this vilification of Barkyaruq. Al-­Rawandi, for example, praises him in conventionally glowing terms, as befits a scholar who wishes to gain favour at Konya with the Seljuqs of Rum; see al-­Rawandi, Rahat al-sudur, ed. M. Iqbal (London, 1921), 138. For an as yet unknown reason, moreover, the artist of the Edinburgh manuscript of Rashid al-­Din honours Barkyaruq alone among the Seljuq sultans with a hanging crown; see D. Talbot Rice, The Illustrations to the ‘World History’ of Rashid al-Din (Edinburgh, 1976), 171. 14. Al-­Bundari, Zubdat al-nusra, 89, and Sibt b. al-­ Jawzi, Mir’at al-zaman (Hyderabad, 1951), vol. 8, 13. 15. Ibn al-­Jawzi, al-Muntazam, eds M. and M. ‘Ata (Beirut, 1992), vol. 17, 488. 16. Ibn al-­Athir, al-Kamil, vol. 10, 213–20, and Ibn al-­Jawzi, al-Muntazam, vol. 17, 62–5. It should be recalled that the great propaganda war against the Isma‘ilis was being waged in Baghdad during his reign. Al-­Ghazali was commissioned by the caliph al-­Mustazhir to write a systematic refutation of Isma‘ili doctrines and a defence of ‘Abbasid legitimacy; see al-­Ghazali, Fada’ih al-Batiniyya, ed. ‘Abd al-­Rahman Badawi (Cairo, 1964). However, it is doubtful whether al-­Ghazali’s polemical skills had any impact on Seljuq power struggles. 17. Ibn al-­Athir, al-Kamil, vol. 10, 221. 18. Ibid., vol. 10, 253. 19. Lewis, The Assassins, 52. 20. Ibn al-­Qalanisi, Dhayl, 152–6. The fath-nama is also quoted by Sibt b. al-­Jawzi, Mir’at, vol. 8, 20. 21. Ibn al-­Qalanisi, Dhayl, 155. 22. Al-­Bundari, Zubdat al-nusra, 118. 23. For the siege of Shahdiz, see al-­Rawandi, Rahat al-sudur, 161; al-­Husayni, Akhbar, 79; Ibn al-­Athir, al-Kamil, vol. 10, 299 ff.; Ibn al-­Jawzi, al-Muntazam, vol. 17, 102; Sibt b. al-­Jawzi, Mir’at, vol. 8, 20, and al-­Bundari, Zubdat alnusra, 90–1. 24. Ibn al-­Athir, al-Kamil, vol. 10, 420. 25. Al-­Bundari, Zubdat al-nusra, 90. 26. Ibn al-­Athir, al-Kamil, vol. 10, 313.

p ow e r st ru g gl e be twe en the sel j uq s a nd is ma ‘ il is  | 161 27. Al-­Husayni, Akhbar, p. 81; Ibn al-­Athir, al-Kamil, vol. 10, p. 335; Juwayni, The History of the World-Conqueror, tr. J. A. Boyle (Manchester, 1958), vol. 2, 680. 28. Ibn al-­Athir, al-Kamil, vol. 10, 335. 29. Juwayni, History, tr. Boyle, vol. 2, 680. 30. Al-­Husayni, Akhbar, 81. 31. Al-­Husayni, Akhbar, 79, 82; al-­Bundari, Zubdat al-nursra, 117, and Ibn al-­ Athir, al-Kamil, vol. 10, 369–70. 32. Juwayni, History, tr. Boyle, vol. 2, 681. 33. Ibid., vol. 2, 675–6. 34. Al-­Husayni, Akhbar, 79. 35. Juwayni, History, tr. Boyle, vol. 2, 682. 36. Sibt b. al-­Jawzi, Mir’at, vol. 8, 77. 37. Ibn Khallikan, Wafayat, vol. 2, 428. 38. Ibn al-­Athir, al-Kamil, vol. 10, 213 ff. 39. Ibid., vol. 10, 214. 40. Ibid., vol. 10, 299. 41. Ibid., vol. 10, 369. 42. Ibid. 43. E. Herzfeld, Matériaux pour un corpus inscriptionum arabicarum (Cairo, 1955), vol. 1, 118–19. The numismatic evidence that survives is sparse but confirms the titulature found on architectural monuments. Inevitably, the confined space on coins does not allow for the full panoply of epithets often found on buildings; see, for instance, G. Hennequin, ‘Monnaies Salg˘uqides inédites ou peu courantes’, Annales Islamologiques 19 (1983), 76–92. It is, perhaps, significant too that Ibn Khallikan (Wafayat, vol. 5, 72) who uses many of the same sources as Ibn al-­Athir and shares many of the latter’s attitudes towards the Seljuq period, devotes a long obituary to Sultan Muhammad but says little of Muhammad’s fight against the Isma‘ilis of Alamut, commenting only that Muhammad fought against the ‘heretical party’. 44. Juwayni, History, tr. Boyle, vol. 2, 681, and Kashani, Zubdat al-tawarikh, 166; see also Rashid al-­ Din, Jami‘ al-tawarikh; qismat-i Isma‘iliyan, eds M. T. Danishpazhuh and M. Mudarrisi Zanjani (Tehran, 1338/1959), 131. 45. Ibn al-­Athir, al-Kamil, vol. 10, 369. 46. Juwayni, History, tr. Boyle, vol. 2, 637. 47. D. Morgan, The Mongols (Oxford, 1990), 17–18. 48. Ibn al-­Athir, al-Kamil, vol. 10, 370. 49. W. Kleiss, ‘Assassin castles in Iran’, in R. Hillenbrand (ed.), The Art of the Seljuqs

162  |  the me d ie va l tur k s in Iran and Anatolia (Costa Mesa, CA, 1994), 316. Kleiss regards the identification of Ghutinar as an Isma‘ili castle as probable but not proven. 50. Al-­Husayni, Akhbar, 87; Ibn al-­Athir, al-Kamil, vol. 10, 201, and al-­Bundari, Zubdat al-nusra, 259–60. 51. Al-­Husayni, Akhbar, 87. 52. Ibn al-­Athir, al-Kamil, vol. 10, 201. 53. Ibid., vol. 10, 270. 54. Al-­Husayni, Akhbar, 55. 55. Ibn al-­Athir, al-Kamil, vol. 10, 393. 56. Sibt b. al-­Jawzi, Mir’at, vol. 8, 77.

12 Some Reflections on Seljuq Historiography

Introduction

I

n the eleventh century the Seljuqs, a Turkic dynasty of nomadic origin, conquered large areas of the eastern Islamic world, parts of central Asia, Iran, Iraq and Syria, as well as new lands in Anatolia. The Seljuq rulers quickly presented themselves as upholders of Sunni Islam. Their empire remained broadly unified until 1118; thereafter, centrifugal forces inherent in the nomadic heritage of the Seljuqs fragmented their polity. Seljuq history is interwoven with the history of several other empires and d ­ ynasties – ­Byzantium, Seljuq successor states in Syria such as the Zengids and Ayyubids, the Anatolian Turcoman dynasties, the Fatimids of Egypt who were the main opponents of the Seljuqs in the eleventh century, the Crusader states, and other groups in central Asia. In Anatolia the Seljuqs formed part of the mosaic of Turcoman dynasties vying for power on the eastern Byzantine borders and they interacted with Byzantium from the eleventh to the early fourteenth century. Given the time span of Seljuq power and their vast empire, it is not surprising that their history should have been written from a number of perspectives in widely varying territories. Definitions and Parameters ‘Seljuq historiography’ will be defined in this chapter as the historiography of the Great Seljuqs (c. 1030–1194), who provided the principal inspiration for subsequent historiographical traditions in Anatolia and Syria. Of course, the historiography of the Seljuqs of Anatolia (c. 1077–1307) is inextricably linked 163

164  |  the me d ie va l tur k s to the eastern tradition, as indeed is the historiography of the Seljuq successor states in Syria and that of the Mamluks of Egypt. Indeed, Seljuq ­traditions – ­cultural, military and ­governmental – ­were translated wholesale into Anatolia and Syria. So too was the art of writing history. ‘Historiography’ will be taken to denote the various Islamic genres of writing which touched more or less directly on the history of the Seljuqs – ‘Universal Histories’, dynastic histories and town chronicles.1 The term ‘Seljuq historiography’ does not, however, mean historical writing about the Seljuqs by the Seljuqs themselves in some form of Turkic language. The written history of the Seljuq Turks exists only in Arabic and Persian sources, composed by Arabic and Persian court officials or religious scholars with their own concerns and preoccupations and within their own historiographical traditions.2 Unfortunately, there are no extant Islamic sources from the period 1050 to 1150 which record the history of the Seljuqs. As was often the case in the medieval period, there is a sizeable time lag between events and their first appearance in the historical record. The Persian Historiographical Tradition The first historical works in New Persian appeared suddenly in the middle of the tenth century in Khurasan and Central Asia, areas under the rule of the Samanid dynasty.3 They were written by court officials, religious scholars and others. It was usually the case that such historians engaged in some other activity as well as history.4 An important aspect of the Persian historiographical tradition is its obvious didacticism. Scholars have been dismissive about the value of Persian historiography on the Seljuqs, criticising it for presenting difficulties for those in search of ‘factual data’ with which to reconstruct the past.5 But this is to misinterpret the purpose of such a tradition, which aims to present exemplary history and in particular to show the nature of true kingship. The influence of the Fürstenspiegel genre is very strong.6 This dates from the Sasanian period in the first instance and in fact even further back, to Baylonian times. But here it is given an Islamic dress. Kings are chosen by God for the good of mankind. They are endowed with divine charisma (farr-i ilahi) and they dispense justice in accordance with God’s decrees. This ‘rhetorical’ history seeks out striking examples to reinforce these themes and uses direct discourse and interpolated material, such as Qur’anic and Hadith quotations. Those who write this kind

so m e r ef lecti ons on sel ju q histo r io gr a p h y  | 165 of history are concerned to show the latent patterns of the events themselves and to elucidate their underlying ‘meaning’. It is difficult to identify the audience for such works. The device of the sultan being given counsels in Persian on how to rule stretches credulity in the case of the early Turkish-­speaking Seljuq sultans, although as time went on they may well have understood something of the works dedicated to them. It is more likely that these works were intended for a Persian-­speaking intellectual elite, in Iran and later in Anatolia. The Persian historical sources on the Seljuqs were clearly intended for declamation, not just for reading, and they are literary in style and structure, forming part of the belles-­lettres tradition. Meisami suggests convincingly that such works were targeted at those who did not know Arabic and that the promotion of Persian historical writing may well have been a means for sovereigns to legitimise their claims to rule in the eastern Islamic world.7 It is important to stress the approach and value of the Persian historiographical tradition both for its own intrinsic importance within the eastern territories of the Seljuq empire and also because it was this Persian ­tradition – ­not the Arabic ­one – w ­ hich was transplanted lock, stock and barrel into Seljuq Anatolia and which persisted there until Ottoman times. In his history of the early Ottoman sultans entitled Hasht Behesht (Eight Paradises)8 composed in Istanbul in the early sixteenth century, Idris Bitlisi writes in the same spirit of didacticism and in the same high-­flown, sophisticated Persian style that Ravandi had used several centuries earlier, first in Iran and then in Konya. The genres used by Seljuq historians writing in Persian fall into two broad ­categories – ­dynastic histories and local chronicles. The key sources which deal with the Seljuq rulers of Iran and ­Iraq – ­Nishapuri and ­Ravandi – ­are dynastic histories which cover the whole sweep of events from the legendary origins of the dynasty until its collapse. Local chronicles of individual cities, such as the work of Ibn Funduq,9 are of some additional interest but inevitably they are rather narrow in their focus. A rarity in Seljuq historiography, but largely outside the time frame of this chapter, is Ibn Bibi’s history of the Seljuqs of Anatolia, written in 1281, which takes the form of memoirs and deals with the years 1192–1280.10

166  |  the me d ie va l tur k s Two Key Persian Authors on the Seljuqs Nishapuri (d. 1187) The work of Nishapuri, the Saljuqnama,11 became the Urtext of all later Persian histories of the Seljuqs, such as those of Ravandi, Hamdallah Mustawfi, Mirkhwand and others. Nishapuri worked at the Seljuq court where he was the tutor to two Seljuq princes. He wrote his strongly pro-­Seljuq chronicle around 1175, probably shortly after the accession of the last Seljuq ruler of Iran, Tughril. His dilemma was that of the highly educated Persian bureaucrat attempting to come to terms with the Turkish presence. He charts the history of the Seljuq dynasty from their rise to power in Khurasan until his own day, when Seljuq decline is all too apparent. Despite its straightforward and unembellished style, the work is heavily impregnated with didactic elements. This is particularly true of the short preface in rhyming prose in which Nishapuri reflects on kingship in general and the government of the Seljuqs in particular. Although the Seljuqs were Turkish nomads, he places them within the ethos of Perso-­Islamic kingship. According to Nishapuri: ‘Kings are shepherds of the flock and the protectors of creatures from all kinds of calamity and dread.’12 Present rulers should learn from studying the past: ‘There were no sultans greater or kinder to the flock or more worthy in their regard for the people than the kings of the House of Seljuq.’ Writing in the period of Seljuq decline, Nishapuri laments the lapse in piety and good government which characterised the rule of earlier Seljuq rulers: ‘Were the kings of this age to emulate their exemplary conduct, this would result in the strengthening of (both) religion and rule and the establishment of the realm on firm foundations.’13 The work ends in panegyric mode; Nishapuri expresses the hope that ‘all the regions of the world will be under the command of the Seljuq sultan’.14 Ravandi (flourished until the early thirteenth century) Ravandi lived at the end of the Seljuq period. His only surviving work is the Rahat al-Sudur, a dynastic history of the Great Seljuqs.15 Ravandi would have wished to dedicate his work to a Persian Seljuq ruler but after the demise of the dynasty in Iran proper in 1194, he looked for patronage from Konya, wanting his book to be in the ‘name of a Seljuq sultan’.16 Indeed, he went there personally to present his work. The reorienting of Ravandi’s

so m e r ef lecti ons on sel ju q histo r io gr a p h y  | 167 history towards Anatolia is a clear sign that early thirteenth-­century Persian writers considered the Anatolian Seljuq dynasty to be the new champions of Sunni Islam, and Konya the centre for the continuation of Persian scholarly traditions. Ravandi’s work is inflated and derivative; much of it is based on that of Nishapuri. In the past scholars have not rated it highly. They do not like its padding and excess material.17 Indeed, entire chapters are devoted to non-­ historical information. The work is a ‘compilation’ and the crucial fact is that it does not have ‘history’ as its prime aim. It is divided into three p ­ arts – a­ long introduction, a historical section and finally the non-­historical chapters on topics as varied as chess, horsemanship and calligraphy. The style is full of rhetorical devices. There is no denying that these get in the way of ‘straight history’. Interpolations include Qur’anic quotations, hadith, proverbs, poetry (especially the Shahnama) and edifying anecdotes. Ravandi’s angle is apparent from the very beginning. The introduction praises the Seljuq Turks, devout Muslims who defend the True Faith with the sword. Sultan Kaykhusraw is praised as ‘the fruit of the tree of Seljuq: a tree whose root is the strengthening and propagation of the faith’.18 The principles are those of the ‘Mirrors for Princes’ genre. The historical part serves to legitimise Seljuq rule.19 Ravandi’s work is better structured than that of Nishapuri, who serves up a continuous and plain account. Ravandi uses events as paradigms. For example, when the early Seljuq vizier, al-­Kunduri, comes to a grisly end, this may be interpreted as a presage of the ultimate decline of the Seljuq dynasty.20 The exemplary nature of Ravandi’s account of the battle of Manzikert is clear from the issues which he selects for emphasis: a slave captures the Byzantine emperor, a small force successfully takes on the enormous Byzantine army, the eventual blinding of Romanos IV Diogenes can be seen as a metaphor for the darkness of unbelief. Above all, God’s inexorable purpose for the world involves the ultimate victory of Islam. All in all then, the work is intended to show the sultan how to rule,21 and it deploys both precept and example for this purpose. Thus the book has a sermonising tone, and Ravandi knows just how to milk an event so as to point a moral and adorn a tale. Present rulers would do well to derive lessons from a study of the past:

168  |  the me d ie va l tur k s When they become aware of that and they read and come to know about the life and behaviour of each one [of the past rulers], they will choose that which is the epitome of the virtuous conduct of those who have gone before (them).22

The conclusion of the book looks forward to a revival of Seljuq fortunes in Anatolia. Ravandi’s history is, then, above all, a well-­structured ­work – a­ imed at edifiying and at showing recurring patterns and eternal truths. Despite its intrinsic linguistic difficulty, it probably enjoyed popularity amongst the Persian intelligentsia in Seljuq Anatolia and was translated into Turkish in the reign of the Ottoman sultan Murad II (1421–51). The Arabic Historiographical Tradition The thirteenth century witnessed a surge of historical material about the Seljuqs,23 much of which is still extant. The Arabic historiographical tradition comprised a variety of genres – ‘Universal Histories’ from the Creation until the time of the author, dynastic histories and town chronicles.24 This Arabic tradition is much more diverse and territorially widespread than the Persian one, owing to the predominance of Arabic within the medieval Islamic world as the language of prestige. It allowed both a plain and an ornate style. The ‘Universal Histories’ and local chronicles are written in an unpretentious style and annalistic format, and this pattern endured for centuries. The tradition of high-­flown rhymed prose in Arabic historiography, as epitomised in the writings of Saladin’s biographer, ‘Imad al-­Din al-­Isfahani, was short-­lived.25 The great Mamluk historians of the fourteenth and fifteenth centuries followed the simpler way of writing. Three Key Arabic Authors on the Seljuqs Al-­Bundari (flourished in the 1220s) ‘Imad al-­Din al-­Isfahani, famous as a historian of Saladin and the conquest of Jerusalem, made in 1183 a revised Arabic version of the memoirs of the years 1072–1134 written in Persian by Anushirwan (d. 1138/9), the vizier of the ‘Abbasid caliph, al-­Mustarshid, and later of the Seljuq sultan Mas’ud.

so m e r ef lecti ons on sel ju q histo r io gr a p h y  | 169 This Arabic reworking, written in high rhetorical rhymed prose, has come down to us in an abridged version made by al-­Bundari, a fellow Persian who, like ‘Imad al-­Din, had moved to Syria. The work of al-­Bundari contains important information but it is difficult to read because of its retention of much of his predecessor’s style.26 Akhbar al-dawla al-saljuqiyya The authorship and provenance of the only Arabic dynastic history of the Seljuqs, entitled Akhbar al-dawla al-saljuqiyya, remain controversial.27 It is an interesting and often racy chronicle covering all periods of Seljuq rule. The author’s attitude to the Seljuqs is generally favourable. His sources must have been very varied. This is mirrored in the wide range of Arabic styles found in his text. At times the writing is simple and unvarnished; on other occasions the language is full of hyperbole. It is clear that the author is taking passages from earlier sources and piecing them together. His aim was to emphasise the Islamic credentials of the Seljuqs. Ibn al-­Athir (d. 1233) Mention should be made, above all, of Ibn al-­Athir and his Universal History, an annalistic work which provides, among much e­ lse – ­for the entire Islamic world was the author’s ­purview – ­the most coherent account of the Great Seljuq sultans. It also represents a high point of Muslim historiographical achievement. His presentation of information is clear and balanced, and his narrative is enlivened by a notable breadth of vision.28 The Differences between the Arabic and Persian Historiographical Traditions The Arabic tradition on the Seljuqs, as exemplified in the Universal History of Ibn al-­Athir or the local chronicles of Syria, with its sober and unpretentious style, makes it much easier to use than the Persian works in the form and style of Ravandi. The generally annalistic framework of the Arabic tradition makes the extraction of ‘facts’ much easier for historians and there is less poetry to obstruct or demoralise researchers. This is not, of course, to suggest for one moment that the poetry is not serving an important function in the view of the medieval author who chooses to place it at key points in his text. The Persian

170  |  the me d ie va l tur k s tradition, an ‘ethical-­rhetorical historiography’, is without doubt literary in nature, a branch of belles-­lettres (adab) which aims to entertain and instruct and which was designed to be read aloud and relished as a literary performance.29 Two Case Studies In keeping with the focus of this volume, two specific episodes of Byzantine history, as seen through the eyes of the Islamic sources, will now be ­considered. They will, it is hoped, serve as models on the basis of which the g­ eneralisations about Seljuq historiography made in this chapter can be tested. The Battle of Manzikert, 1071 The battle of Manzikert is recounted in numerous Islamic sources which come from a wide geographical area and chronological span.30 Just two of these accounts will now be considered. i. The narrative of Ibn al-­Qalanisi (d. 1160) The first extant Muslim account of the battle, that of Ibn al-­Qalanisi writing from Damascus, dates from the middle of the twelfth century.31 It is short and unadorned. It may be summarised as follows. The Seljuq sultan Alp Arslan defeated the Byzantine emperor at Manzikert. The Byzantine army was vast in size, whilst the Seljuq forces were smaller.32 The emperor was taken prisoner. He was subsequently released after a peace treaty had been signed. He was sent back to Byzantium, where he was blinded. It is difficult to assess whether this account is fragmentary because the author lacked more detailed information or because the battle, although worthy of some mention, was outside the focus of a local town chronicle. The propaganda potential of this narrative is left unexploited. Certain key elements which will recur in later accounts are, however, already established here: the superior numbers of the Byzantine army, the victory of Alp Arslan, the capture and release of Romanos Diogenes and his subsequent fate in Byzantium. It is noteworthy that the date is not mentioned. The style of the passage is unpretentious. Ibn al-­Qalanisi shows no special interest in the course of the battle itself. His account, short as it is, focuses rather on the events before and after it. Above all, the notion of Manzikert as a hinge of Islamic history is simply not there.33

so m e r ef lecti ons on sel ju q histo r io gr a p h y  | 171 ii. The narrative of Ibn al-­Jawzi (d. 1200) The account of Manzikert given by the Baghdadi scholar Ibn al-­Jawzi is full of new ‘details’ and elaborates some of the core themes found in the account of Ibn al-­Qalanisi.34 Fully fledged Islamic themes are now incorporated into the account. Alp Arslan addressses his troops like Elizabeth before the coming of the Armada.35 His speech to the Seljuq troops is a carefully constructed set piece, positioned for dramatic effect before the battle itself: When it was the time for the prayer on the Friday, the sultan prayed with the troops. He called on Almighty God and made humble supplication. He wept and abased himself and he said to them [the troops]: ‘We are with reduced numbers of men. I want to throw myself at them [the Byzantines] when prayers are being said for us and for the Muslims on the pulpits. I will either achieve my aim or I will go as a martyr to Paradise. So he amongst you who wants to follow me let him follow me and he who wants to leave let him leave.’

The speech ends with the pronouncement: ‘Today I am only one of you and I am fighting alongside you. He who follows me and gives himself to Almighty God, Paradise and booty will be his. He who leaves [the battlefield] his due will be the Fire and ignominy.’36 The role of the speech is to point to the significance of the battle in Islamic terms and to link it to God’s wider purpose for the ­world – ­the ultimate victory of Islam over Christianity. In the manner of Thucydides, the Muslim chronicler ‘puts into the mouths of the speakers the sentiments he thinks proper to their situation’.37 Alp Arslan is presented as the paradigm of the Muslim warrior of the faith fighting in the path of God, in the hope of the reward of Paradise. The awesome impedimenta of the Byzantine army are ­stressed – 4­ 00 waggons carry weapons, saddles, ballistas and ­mangonels – ­in order to enhance the victory of the depleted Seljuq forces. Ibn al-­Jawzi gives a precise place and date for the battle during which Romanos is captured by a slave and brought to Alp Arslan. He is released after terms for peace have been arranged. The natural antithesis between the mighty Byzantine emperor and the Muslim slave who captures him is exploited for all it is worth. Romanos and Alp Arslan converse in a stylised discourse conducted in high Arabic. Alp Arslan, the unlettered Turkish nomad, treats the Byzantine emperor honourably

172  |  the me d ie va l tur k s and releases him. Thus the eventual victory of the Seljuq sultan is not just a military triumph over the powerful Byzantine empire. The symbol of the captured Byzantine emperor is very potent: Christianity is subjugated to Islam. It is interesting to note that this speech, put into the mouth of the Turkish sultan Alp Arslan for the year 1071 by a late twelfth-­century Baghdadi Arabic author, is a prototype for the Ottoman gazanama genre.38 This was a popular genre in Ottoman times, depicting a campaign of the ruler in which he is shown as the warrior sovereign, the sultan of the jihad fighters, extending the frontiers of the Ottoman empire and the House of Islam. The sultan’s public prayers before the battle inspire his troops to victory.39 Here, then, are the germs of what in Ottoman times became a fully fledged historiographical genre intended to entertain and edify. It is difficult to say how far back this historiographical tradition may go. Certainly speeches with Qur’anic resonances, positioned at moments of high psychological tension, have a long pedigree in Islamic historical texts. It is significant, moreover, that Ibn al-­Jawzi was a famous preacher and prolific writer of religious treatises, and it is always possible that the speech may have been inserted by him for the first time into the account of the battle of Manzikert. In this way, the Islamic credentials of the Seljuq Turks are reinforced; these relative newcomers to the faith are placed firmly within an Islamic context. The Turkish nomad is transmuted into a Muslim leader fighting jihad. The chronology of events is ­vague – ­the two sides seem to have met on a Wednesday but the battle itself is placed firmly on a Friday on which Alp Arslan is mentioned as praying with his troops before the outset of the battle. Unusually, however, amongst the Muslim sources as a whole, Ibn al-­Jawzi manages to provide a date which corresponds to an actual Friday (27 Dhu’l-­Qa‘da 463/26 August 1071). Friday, the Muslim day of prayer, becomes a topos in subsequent accounts of the battle. It should be noted that the account of Ibn al-­Jawzi is reproduced in its broad outlines by most later Muslim authors who mention the battle.40 It might be said, therefore, that his is the canonical narrative. iii. General comments The preceding Islamic accounts of Manzikert, and others which have not been mentioned here, reveal the following points. The Islamic accounts of the battle are very vague about the military details and the course of the battle

so m e r ef lecti ons on sel ju q histo r io gr a p h y  | 173 itself. The date of it remains highly debatable. If we therefore attempt to reconstruct the battle from Muslim sources alone, we have very little concrete information. So far as raw data are concerned, the Muslim sources have little to offer about Manzikert as a battle. This rare interest41 on the part of Islamic historians in the relationship between the Islamic world and Byzantium stems rather from the propaganda potential of the battle. Twelfth- and thirteenth-­century Muslim historians eagerly seize on Manzikert as a focus for their accounts of the history of the Near East under the domination of the Turks, whom they wish to endow with good Islamic credentials. A subsidiary but significant factor for Muslim writers who have lived through the heyday of Muslim jihad against the Crusaders is the association of this victory at Manzikert over a Christian enemy, the Byzantine emperor, with subsequent Muslim victories over the Crusaders in Outremer. In other words, Manzikert was perceived to be the first step in the process by which Turkish-­led dynasties defeated the Christians and proclaimed the triumph of Islam. The Fall of Constantinople to the Latins in 1204 The unusually wide horizons of perhaps the greatest of the medieval Muslim chroniclers, Ibn al-­Athir, writing from Syria, prompted him to include under the year 600 AH (1203/4) an interesting and very rare account of the Fourth Crusade.42 In it the destinies of three groups which concern Ibn al-­Athir – t­ he Byzantines, the Crusaders and the Seljuqs of ­Anatolia – ­converge. Ibn al-­Athir relates that these Crusaders came originally to help Jerusalem but that they made their way to Constantinople because of the internal power struggle being enacted there. The Byzantines, in great distress at their treatment by the Crusaders, contacted the Seljuq sultan Sulayman b. Qilij Arslan, lord of Konya, but ‘he found no way of acceding’ to their request for help. Ibn al-­Athir possesses some concrete information about the Latin conquest of Constantinople, although his sources are not mentioned, and he presents it in his usual unvarnished way. He mentions specifically the name of three Crusader l­eaders – ­the blind Doge of Venice whose ‘horse had to be led’ for him,43 the Count of Flanders and the Marquis of Monferrat. He describes the three leaders casting lots for possession of Constantinople, and the conquest by ‘the patriarch of Byzantium, Lashkari’, of the lands east of Constantinople.44

174  |  the me d ie va l tur k s On the surface this account, perhaps surprising for its inclusion at all, since Seljuq or indeed any Muslim involvement is minimal, is straightforward. Yet Ibn al-­Athir, who rarely rises to great heights of invective or emotional o­ utrage – ­a striking and well-­known exception is his famous account of the Mongol ­invasion – ­is making here a few telling propaganda points as well as revealing a nuanced view of Byzantium and Christendom. Emphasising the plunder and killing of the population of Constantinople by the Franks, he writes the following: A group of the notables of Byzantium entered the great church which is called Sophia. The Franks came to it and a group of priests, bishops and monks came out to them with the Gospel and the Cross in their hands to entreat thereby to spare them. They did not heed them and they killed [them] and plundered the church.45

There are strong echoes here of the accounts in the Islamic sources of the fall of Jerusalem to the Crusaders in 1099. But here in 1204 the victims are the Byzantines who, although they are fellow-­Christians, are themselves victims of Crusader treachery and brutality. For Ibn al-­Athir the familiar Christian enemy, Byzantium, has been ousted from the firing line of Islamic anti-­Christian invective (directed before the coming of the Crusades at Byzantium) by a new and even more devilish foe, the Crusaders. Byzantium, a long-­familiar entity in the Near East and a frequent player in Seljuq affairs, is treated almost sympathetically here. It should also be noted that mention of Byzantium in the Seljuq sources is not usually accompanied by the traditional minatory tags, such as ‘May God curse them’ or ‘May God send them to perdition’ which are employed regularly for the Crusaders from the second half of the twelfth century onwards. General Observations For the historian of Byzantium the Islamic sources which deal with Seljuq history are perhaps disappointing. Their concentration on Manzikert, with only occasional references to other Byzantine involvement in Syria and Anatolia, is a clear indication of the introverted priorities of Islamic historians. Important and protracted historical developments are not the concern of such writings. Complex ­processes – ­such as the impact of the Turcoman nomads on agriculture, the level of Seljuq Islamisation, the situation of Christians under Seljuq

so m e r ef lecti ons on sel ju q histo r io gr a p h y  | 175 rule in ­Anatolia – ­are left undiscussed. Chance warlike encounters, such as Manzikert probably was, are given retrospective validity by being portrayed as part of an overall strategy, a deliberately expansionist policy on the part of the Seljuqs. Border skirmishes, raiding and nomadisation are cast as jihad. Thirteenth-­century Muslim writers living under Turkish rule, such as Ibn al-­Athir, Ravandi and al-­Husayni, hark back to the Seljuqs as the illustrious predecessors of their present regimes, just rulers with impeccable credentials, masters of their own destiny, zealous, newly converted Muslims who wage jihad against the infidel and heretic alike. Given this casting of the Turkish Seljuq sultans in ‘heroic’ Islamic mould at the hand of Arabic and Persian court officials and religious scholars, is it then possible to glean any insights at all from the sources into the effects of the Seljuq conquest on the Near East and, more especially in the context of this volume, on Anatolia? In other words, do the sources tilt towards a corroboration of the views of Cahen or Vryonis, Wittek or Lindner? It is more in the evaluation of short anecdotes and asides, in hints rather than in overall conceptions, that some tentative comments may be made. The lifestyle of the Turcomans, be it in Iran, Syria or Anatolia, is not ignored by Muslim writers even within the framework of their overarching historiographical agenda of legitimising and Islamising the rule of the Seljuq sultans. Beneath the rosy surface there are suggestions that the coming of the nomadic Turks was difficult to bear;46 their military might was a fact of life, a ‘necessary evil’ to be channelled and utilised by those who knew better, the bureaucrats and religious leaders, in order to support the Islamic world and defend it by force of arms against all comers. The innate superiority and hostility felt by Arabs and Persians, especially initially, towards their Turkish overlords emerges in recurring clichés about Turkish drunkenness and brutality; the topoi of Turks building minarets out of the skulls of their enemies and then making the call to prayer from these structures, or drinking wine from goblets made of these skulls, spring to mind.47 There are also stories of Turks being kept outside city walls because of their disruptive impact on those inside. The evolution of the Turkish leadership from nomadic chiefs to sedentary rulers must have been slow, difficult and complex. In general, however, the sources elide these problems and negative comments about the Seljuqs have to be searched for laboriously in an otherwise ­favourable – ­if ­stereotyped – ­picture of them in both the Arabic and Persian historiographical traditions.

Ibn Bibi

Aqsarayi Ahmad of Nig˘de Author of Tarikh-i al-i Saljuq

Turkish translations of Ravandi and Ibn Bibi

1250–1300

1300–1350

1350–

Ibn al-­Furat al-Maqrizi al-‘Ayni

al-­Shaybani Bundari Ibn Abi Tayyi’ Ibn al-Athir

al-Anawi

1200–1250

bn al-‘Adim Ibn Shaddad Ibn Wasil

Ibn al-­Qalanisi al-­Halabi ‘Imad al-Din al-Isfahani

1150–1200

Syria

Ibn Munqidh

Ibn Zafir

Egypt

1100–1150

1050–1100

Anatolia

Mirkhwand Ahmad Ghiffari

Rashid al-Din al-Yazdi Hamdallah Mustawfi

Ravandi

Sibt b. al-Jawzi

al-­Hamadhani Anushirwan

Author of the Maliknama

Iran

Ibn Funduq Nishapuri

Hilal al-­Sabi Ghars al-­Ni‘ma

Iraq

Ibn al-Azraq

Jazira

Appendix: Select Table of Medieval Historical Writing on the Seljuqs in Arabic and Persian

al-Husayni

Gardizi Bayhaqi

Central Asia

Notes: 1. Italics denote extant sources and roman script indicates lost works. 2. No sources on the history of the Seljuqs survive from the period 1050–1150 except Ghaznavid ones which cover only the rise of the Seljuqs. Lost sources can, to some extent at least, be reconstructed through quotations and borrowings from them in later writings. 3. The earliest extant dynastic histories of the ­Seljuqs – ­Nishapuri, al-­Bundari, Ravandi and the work attributed to al-­Husayni – ­date from the second half of the twelfth and the early thirteenth centuries. 4. Geographical borders were fluid: a. ‘Imad al-­Din al-­Isfahani, a Persian, lived in Iran before moving for most of his career to Syria. b. Ravandi, a Persian, wrote his work in Iran but dedicated it to the Seljuq sultan of Konya. 5. Linguistic borders were fluid: a. Writers composed works in Arabic, based on earlier Persian sources (for example, ‘Imad al-­Din al-­Isfahani) b. Writers composed in both Arabic and Persian (for example, Rashid al-­Din). c. The thirteenth-­century Syriac writer, Bar Hebraeus, who also knew Arabic, used the Persian source, the Maliknama, in his Chronography. 6. No medieval historical works on the Seljuqs appear to have been written in any form of Turkish. Turkish translations began in the fifteenth century. 7. The Persian historiographical tradition formed the framework for the writing of history in Seljuq Anatolia. Little or nothing on history seems to have been written in Arabic in Anatolia. 8. Schools of historiography existed but they were not necessarily hermetically sealed (for example, eleventh-­century Iraq, thirteenth-­century Syria, fourteenth-­century Anatolia, fifteenth-­century Egypt).

178  |  the me d ie va l tur k s Notes   1. Auxiliary historical works such as biographical dictionaries, ‘Mirrors for Princes’, geographical works and encyclopedias, and even poetry, are often valuable resources for Seljuq history.   2. Comparatively little work has been done on Seljuq historiography. As Humphreys rightly points out, ‘This literature has not been adequately studied’; see R. S. Humphreys, Islamic History: A Framework for Inquiry (Princeton, 1991), 165. The pioneer work of Cahen on this subject is still useful for the names of writers and their works; see C. Cahen, ‘The historiography of the Seljuqid period’, in B. Lewis and P. M. Holt (eds), Historians of the Middle East (London, 1962), 59–78. See also idem, ‘History and historians’, in M. J. L. Young, J. D. Latham and R. B. Serjeant (eds), The Cambridge History of Arabic Literature: Religion, Learning and Science in the ‘Abbasid Period (Cambridge, 1990), 165–8. The research of Kafesoğlu was also a landmark; see I. Kafesoğlu, ‘Türkiye’de Selçuklu tarihçiliği’, in Cumhuriyetin 50. yilina armağan (Istanbul, 1973), 83–92: tr. G. Leiser as ‘Seljuk historiography in Turkey’, International Journal of Turkish Studies 3 (1985–6), 131–6.   3. In the field of Persian historical writing important new research has been carried out very recently by Julie Meisami. See, for example, J. S. Meisami, ‘Why write history in Persian? Historical writing in the Samanid period’, in C. Hillenbrand (ed.), Studies in Honour of Clifford Edmund Bosworth II: The Sultan’s Turret (Leiden, 1999), 348–74; and above all, eadem, Persian Historiography to the End of the Twelfth Century (Edinburgh, 1999). See also M. A. Waldman, Towards a Theory of Historical Narrative: A Case Study in Perso-Islamicate Historiography (Columbus, 1980), and K. A. Luther, ‘Islamic rhetoric and the Persian historians, 1000–1300 AD’, in Studies in Near Eastern Culture in Memory of Ernest T. Abdel-Massih (Ann Arbor, 1990), 90–8.   4. See Meisami, ‘Why write history in Persian?’, 348.   5. The comments of Gibb are typical of this negative judgement: ‘History became a work of a­ rtifice . . . ­Their [the works of history] bombast and lack of judgement make the most unfavourable impression’ Encyclopaedia of Islam, 1st edn, Supplement, art: ta’rikh, 239. For similar views, see also G. von Grunebaum, Medieval Islam (Chicago, 1962), 282–3, and I. Afshar, in the introduction to his edition of Nishapuri, Saljuqnama (Tehran, 1953), 6.  6. For a recent overview of the Fürstenspiegel genre, see C. E. Bosworth, ‘Administrative literature’, in The Cambridge History of Arabic Literature: Religion, Learning and Science, 165–8.

so m e r ef lecti ons on sel ju q histo r io gr a p h y  | 179  7. Meisami, Persian Historiography, 364.   8. This vast work, available in many manuscripts, has still not been edited.   9. Ibn Funduq, Ta’rikh-i Bayhaq, ed. A. Bahmanyar (Tehran, 1965). 10. Ibn Bibi, Al-awamir al-‘ala’iyya fi’l-umur al-‘ala’iyya, published as Histoire des Seldjoucides d’Asie Mineure d’après Ibn Bibi (Recueil de textes relatifs à l’histoire des Seldjoucides), III, ed. M. Th. Houtsma (Leiden, 1902); tr. H. Duda, Die Seltchukengeschichte des Ibn Bibi (Copenhagen, 1959). 11. Nishapuri, Saljuqnama, ed. I. Afshar (Tehran, 1953). This is not always an accurate edition. A new edition is being prepared by A. Morton. 12. Ibid., 9. 13. Ibid., 10; the second quotation is translated by J. Meisami in Persian Historiography, 230–1. 14. Ibid., 83. 15. Ravandi, The Rahat-us-Sudur wa ‘Ayat-us-Surur, ed. M. Iqbal (London, 1921). 16. Ravandi, Rahat, 62. Originally he had dedicated the work to Sultan Sulayman II (d. 1204) but with the latter’s death he shifted his panegyrics to his successor, Kaykhusraw (Ibid., 19–38). 17. See the comments of Iqbal in the introduction to his edition. 18. Ravandi, Rahat, 29. 19. Especially vis-­ à-vis their predecessors, the Ghaznavids, and the ‘Abbasid caliphate. 20. Ravandi, Rahat, 117–18. 21. J. Meisami, ‘Ravandi’s Rahat al-sudur: history or hybrid’, Edebiyat, n.s. 5, 201. 22. Ravandi, Rahat, 65. 23. See the table in the appendix to this chapter. 24. Mention should also be made of the various biographical dictionaries of the thirteenth and fourteenth centuries. Noteworthy in the Seljuq context is the vast biographical dictionary of Ibn al-‘Adim, Bughyat al-talab, which contains 8,000 entries. This author quotes frequently from lost sources of the Seljuq period, such as al-­Hamadhani. 25. For a recent study of ‘Imad al-­Din al-­Isfahani, see L. Richter-­Bernburg, Der syrische Blitz. Saladins Sekretär zwischen Selbstdarstellung und Geschichtsschreibung (Beirut, 1998). 26. Al-­Bundari, Zubdat al-nusra wa nukhbat al-‘usra, ed. M. T. Houtsma (Leiden, 1889). 27. Akhbar al-dawla al-saljuqiyya, ed. M. Iqbal (Lahore, 1933). For a long time the work under this title was attributed to al-­Husayni but it seems that this valuable

180  |  the me d ie va l tur k s source in its present form is in fact an abridgement of a longer history called Zubdat al-tawarikh which was written by al-­Husayni. But more research needs to be done on this historiographical problem. 28. Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, ed. C. J. Tornberg (Leiden, 1853), XII, 124–6. Regrettably Ibn al-­Athir does not specify what his sources are, although some of them can be deduced. For example, he used a lost Persian source, the Maliknama. It is not clear whether Ibn al-­Athir knew Persian or not. Certainly, as Richards points out, Cahen gives no reasons for his assertion that ‘he [Ibn al-­Athir] was apparently ignorant of Persian’ (see Cahen, ‘Historiography’, 71; D. S. Richards, ‘Ibn al-­Athir and the later parts of the Kamil: a study of aims and methods’, in D. O. Morgan (ed.), Medieval Historical Writing in the Christian and Islamic Worlds (London, 1982), 88). 29. Some of the Persian historiographical works written in Seljuq Anatolia, such as those of al-­Anawi and Aqsarayi, remain relatively unexploited presumably because of their inherent linguistic difficulty. 30. See the contribution by S. Vryonis Jr in Antony Eastmond (ed.), Eastern Approaches to Byzantium (Aldershot, 2000) (Chapter 1). Both he and I are preparing books on the battle of Manzikert. See also EI 2, art: Malazgird (C. Hillenbrand); S. Vryonis Jr, ‘A personal history of the history of the battle of Manzikert’ in N. Oikonomidès (ed.), Asia Minor (Athens, 1998), 225–44. 31. The sources on which Ibn al-­Qalanisi drew for this account are unknown but it is likely that he had access to the work of the Baghdadi historian, Ghars al-­Ni‘ma b. Hilal al-­Sabi’ (d. after 1077). 32. The figures given are absurdly inflated for both sides: 600,000 in the Byzantine army and 400,000 Seljuq forces. 33. The nationalist interpretation of Manzikert as the decisive moment which shaped the history of Turkey is a largely twentieth-­century one. 34. Ibn al-­Jawzi, Al-Muntazam fi ta’rikh al-muluk wa’l-umam, ed. unidentified (Hyderabad, 1940), VIII, 261–3. 35. Ibid., VIII, 262. 36. Ibid. 37. K. Egan, ‘Thucydides, tragedian’, in R. H. Canary and H. Kozicki (eds), The Writing of History (Madison, 1978), 79–80. As Egan points out, ‘the relationship with what was actually said on particular occasions varies considerably’, ibid., 77. 38. See A. S. Levend, Gazavatnameler ve Mihalog˘lu Ali Beyin Gazavatnamesi (Ankara, 1956), 15–177.

so m e r ef lecti ons on sel ju q histo r io gr a p h y  | 181 39. For a discussion of the Ottoman gazanama tradition, see C. Woodhead, ‘Perspectives on Süleyman’, in M. Kunt and C. Woodhead (eds), Süleyman the Magnificent and His Age (London, 1995), 172–3. 40. For example, his grandson Sibt b. al-­Jawzi, Ibn al-­Athir and Ibn al-‘Adim. 41. An exception to this generalisation is Ibn Bibi, for whom Byzantine proximity and involvement in the affairs of the Anatolian Seljuqs occasions no surprise or explanation. It is a fact of life. 42. Ibn al-­Athir, Kamil, XII, 124–6. 43. Ibid., 125. 44. Ibid. 45. Ibid. 46. This negative picture is, of course, corroborated by the much more overtly hostile testimony of Georgian and Syriac sources. 47. Such a story is told of the Turkish commander, Tughtegin, who fought the Franks in the early twelfth century.

13 Women in the Seljuq Period

A woman is basically flesh, and flesh must be preserved. If it is not preserved, it begins to smell and then there is no cure for it. So treat women with respect and give them what they ­want – ­but lock your door of your house and keep other men away. Yusuf Khass Hajib, Wisdom of Royal Glory

T

he subject of medieval women is a daunting one. The sources on which to base generalisations do exist, but they are by their very nature often opaque or difficult to evaluate. It is easy to exaggerate or distort evidence from a wide variety of literary genres, geographical locations and historical periods. The Seljuq sultanate (1030–1190) was a key moment in medieval Iranian and Islamic history because it witnessed the arrival in the eastern provinces of the Muslim world of many Turkish nomads organised along tribal lines. Their irruption into Iran and subsequent infiltration into Anatolia and parts of Syria brought new ways of life to the Islamic world.1 The Seljuq period saw the encounter between the long-­Islamicised eastern Iranian provinces with their high level of culture and scholarly achievement and the nomadic lifestyle of the Seljuq Turks only recently converted to Islam. What would be the impact of the Islamic world on the freer way of life of the nomadic Turkish women who came with the conquerors? Would the greater independence of nomadic women have any influence at all on the lives of urban and rural Muslim women? This chapter assesses some of the existing evidence on which to answer these two questions and to highlight some facets of women’s lives in the 182

w ome n i n the se lj uq p e r io d  | 183 Seljuq period. The evidence discussed derives from the Seljuq world generally because the Seljuq sphere of influence was not restricted to Iran itself but also included parts of Syria and Anatolia where the Rum Seljuqs ruled until the Mongol conquest (1243) and beyond. The evidence used in this chapter comes from a wide variety of s­ources – ­Seljuq legal and literary works, historical chronicles, biographical dictionaries, monumental inscriptions and works of art. Inevitably, the picture emerging will remain impressionistic. Not all facets of every social class of women in the Seljuq period can be addressed here, because evidence on some aspects of women’s lives, especially those not from court circles, is virtually non-­existent.2 Women’s Position: The Theoretical View of the Fuqaha’ and the Intelligentsia The general position of women as laid down in the Shari‘a is well known. Works of Islamic jurisprudence (fiqh) followed a rigid format established over several centuries. By the Seljuq period, they contained a complete Book of Marriage. Al-­Ghazali’s magnum opus, Ihya’ ‘ulum al-din (Revivifying of the Sciences of Religion), and his so-­called summary of it in Persian, Kimiya-yi sa‘adat (Alchemy of Happiness), both contain lengthy discussions of marriage.3 He discusses the importance of physical beauty in relation to other attributes that he deems necessary and desirable in a model wife. His basic message is clear: chastity before beauty. Enumerating the eight qualities to be sought in a ­wife – p ­ iety, good character, beauty, light dowry, fertility, virginity, good lineage, and a not-­too-­close kinship with the would-­be ­husband – a­ l-­Ghazali puts beauty only third after piety and good character, and he cites the Prophet Muhammad as saying that a wife should be sought for her religion, not her beauty. Nevertheless, beauty is a desideratum in a wife since it evokes love and affection in a man. The woman’s role in marriage should be in accordance with the Qur’an and sunna. She should be well treated and taught her duties as a Muslim wife and mother. Al-­Ghazali’s tone is firm but not overtly hostile to women.4 Another legist of the Seljuq period, Ibn al-­Jawzi, devotes a complete work (Kitab ahkam al-nisa’, Book of Rules for Women) to a discussion of the statutes governing women’s conduct within Islam.5 Chapter 24, on the topic of

184  |  the me d ie va l tur k s women’s Friday prayer, reveals the view common to many medieval fuqaha’ (religious lawyers) that women are distracting in the mosque. Although Ibn al-­Jawzi is forced to concede that women are entitled to pray in the Friday ­assembly – ­there are, after all, irrefutable hadiths confirming ­this – ­he then adduces some arguments raising doubts about the validity of this practice. On the question of the rows of worshippers in the mosque, he states that any prayer performed by men who line up behind women is invalid. Since it can happen that men arrive late at the mosque and are therefore forced to place themselves behind rows of women (who have arrived on time), it would be easier to suggest that women should not go to the mosque. On the permissibility of women’s presence in the mosque, Ibn al-­Jawzi judges that if they are fearful of creating disorder in the minds of men, it is preferable for them to remain at home. Aware, however, of the hadith in which the Prophet states that the mosque should not be forbidden to women, Ibn al-­Jawzi concludes that the Friday prayer in the mosque is not a duty for women. The placing of his subsequent chapter, entitled ‘Why women must avoid going out’, is revealing.6 An amplified picture of the ideal role to be played by Muslim women in the Seljuq period emerges from contemporary works of adab (secular literature). These are permeated with perennial Near Eastern attitudes to women. Certain familiar stereotypes and themes appear, such as the dangers of women stepping outside their traditional domestic role and meddling in affairs of state. Such attitudes are found especially in the Mirrors for Princes literature, a genre that gives advice to the ruler on how to rule, usually through a series of edifying anecdotes.7 The Siyasatnama (Book of Government) of the Seljuq wazir, Nizam al-­Mulk, is typical. In the work’s second half, the author analyses what has gone wrong with Seljuq government and suggests some solutions. Chapter 42, ‘On the subject of those who wear the veil’, is devoted to women. The tone is profoundly negative. The chapter begins with an outspoken pronouncement that combines Islamic legal ideas with what some might call more universal, deep-­seated prejudices about women and with the author’s own practical experience of dealing first-­hand with women at the Seljuq court: ‘The king’s underlings must not be allowed to assume power, for this causes the utmost harm and destroys the king’s splendour and majesty. This particularly applies to women for they are wearers of the

w ome n i n the se lj uq p e r io d  | 185 veil and do not have complete intelligence.’8 Nizam al-­Mulk goes on to stress that women’s role is to procreate and to adopt a chaste and abstemious bearing. ‘But when the wives of the king begin to play the role of rulers, they base their orders on what interested parties tell them because they are not able to see things with their own eyes.’9 The whole fabric of the state is thus undermined, and he concludes with the harsh statement: ‘In all ages nothing but disgrace, infamy, discord and corruption have resulted when kings have been dominated by their wives.’10 Although the edifying anecdotes that then follow are set in the dim and distant ­past – ­involving, as they traditionally did in the Mirrors genre, pre-­Islamic Iranian figures such as Khusraw and Kayka’us as well as Adam and Eve and Alexander the ­Great – ­the chapter is given an Islamic coating with the Qur’anic quotation (4: 34), ‘Men are in charge of women’, and the hadith, ‘Consult women but whatever they say do the opposite and that will be right’. Nizam al-­Mulk comments sagely, ‘If women had been able to control themselves, God would not have set men over their heads.’11 Another Persian Mirror for Princes (dating from the mid-­twelfth century) is the anonymous Bahr al-fava’id (Sea of Precious Virtues). Book 7 in chapter 1 deals with the subject of children’s education. Its author states that children (male) should be preserved from seven evils, the third of which is ‘association with women’. He argues that constant contact with women makes young men effete. Such contamination can begin in early infancy. Indeed, even the suckling woman must be devout, for she can influence the baby at this stage through her nourishment.12 In chapter 11, book 16, the author exhorts men to teach their wives necessary religious ­knowledge – ­the rules of prayer, purification and menstruation. Men must also treat their wives equally if they possess more than one. In sexual matters, the man should think not only of his own pleasure but should also seek to satisfy the woman since not to do so is weakness in a man.13 A third Mirror for Princes, existing in both Persian and Arabic versions, is Nasihat al-muluk (Counsel for Kings), usually attributed to al-­Ghazali.14 This Mirror clearly springs from a Seljuq milieu and has pungent words to say on the topic of women, encapsulating no doubt much folk wisdom and familiar stereotypes. Some comments take the form of jokes, such as the following: ‘A sage wished his short wife were tall. People asked him, “Why did

186  |  the me d ie va l tur k s you not marry a wife of full stature?” He answered, “A woman is an evil thing and the less there is of an evil thing the better.”’15 Alternatively, women are likened to ten species of creatures; these include the pig and the ape: ‘The woman who resembles the pig in character knows full well how to eat, break crockery, and cram her stomach . . . She always wears filthy clothes and an unpleasant smell emanates from her.’ As for the ape, ‘the woman who has the character and peculiarities of the ape concerns herself with clothes of many c­ olors – ­green, red and yellow, with trinkets and ­jewels – ­pearls or rubies, and with gold and silver.’ The climax of the argument concerns the virtuous woman who ‘has the peculiarities of the s­ heep . . . ­in which everything is useful.’16 Women’s Roles: Evidence in Medieval Historical Sources Descending from the ideal to reality, what did women in the Seljuq period actually do, as opposed to what Islamic law and the men of the turban prescribed that they should do? Certain aspects of women’s lives may be taken for granted in this context: be they urban or rural Persians or nomadic Turks, their principal roles were those of wife and mother. Urban Persian women in the Seljuq period passed their time carrying out domestic responsibilities in the house and in spinning, weaving and needlework. Some of them worked as craftswomen, servants, nurses, midwives, singing-­girls, dancers and prostitutes. Rural women (in areas such as Gilan and Tabaristan) performed similar duties in addition to undertaking considerable physical labour on the land.17 Nomadic Turkish women also worked hard looking after their husbands, children and extended families; cooking; preparing furs and other garments; loading baggage carts; and tending animals. They were by all accounts toughened by the rigours of their difficult existence. The writings of travellers, both Muslim and European, though later than the Seljuq period, give a general picture of Turkish and Mongol nomadic life and suggest, as is well known, that women in this milieu enjoyed a greater degree of independence than their urban counterparts.18 Ibn Battuta, a careful and well-­informed observer who visited Turks in the Caucasus and Crimea, comments, ‘A remarkable thing which I saw in this country was the respect shown to women by the Turks.’19 In a mosque in Shiraz in Iran, Ibn Battuta was surprised at the

w ome n i n the se lj uq p e r io d  | 187 impressive number of women who (despite the reluctance of the fuqaha’ already discussed) were there at the Friday prayer.20 Medieval Islamic painting provides artistic impressions of the physical appearance of Turkish women, both high-­class and lowly, urban and nomadic.21 These women are, however, seldom described in historical sources. A rare exception is the Nestorian theologian and physician of Baghdad, Ibn Butlan (d.  1066), who wrote a treatise on how to buy slaves and how to detect bodily defects in them. Describing different kinds of slave-­girls, he extols the physical qualities of Turkish women in particular: Fair skinned, the Turkish women are full of grace and animation. Their eyes are small but enticing. They are thick-­set and are inclined to be of short stature. There are very few tall women among them. They are prolific in breeding and their offspring are but rarely ugly. They are never bad riders. They are generous; they are clean in their habits; they cook well; but they are unreliable.22

Thus speaks the shrewd observer who combines medical knowledge with some commonly held beliefs and prejudices. It would appear that women of the Seljuq Turcomans remained unveiled in accordance with the traditions of their nomadic heritage. The evidence of Islamic painting indicates that women at the court also were not always veiled.23 The historical chronicles most frequently mention women in the context of marriage alliances (and women are alluded to only rarely). Here women were used in a complex network of political ties and family relationships. The same emphasis emerges from a collection of biographies of women of the ruling elite, by the Baghdad historian Ibn al-­Sa‘i.24 This work includes eleven biographies from the Seljuq period, all dealing with the wives or concubines of caliphs or sultans. Within the rigid limitations of its genre it sheds light on some aspects of the lives of royal women. The Seljuq sultans contracted marriages with high-­ranking women of their family, usually cousins. They also married women of the ruling families of neighbouring Turkish dynasties such as the Karakhanids and Artuqids. From the beginning of Seljuq rule, women were used to cement alliances for high political stakes, especially in the fragile relationship between caliph and sultan. The most notorious example was the marriage of Sultan

188  |  the me d ie va l tur k s Tughril and the daughter of the caliph al-­Qa’im in 455/1063. The betrothal had taken place two years earlier, although the caliph had erected all kinds of obstacles to the match and had employed delaying tactics to what he perceived as a highly distasteful and politically dangerous marriage. On the personal level, al-­Qa’im’s daughter, whose age is not mentioned, was probably not enthusiastic at the prospect of enforced proximity to Tughril, a seventy-­year-­old Turkish nomad, even if he was ‘sultan of the East and West’. She clearly considered him to be beneath her and did not rise for him when he paid her ceremonial visits in the period preceding the wedding. The marriage proved short-­lived; she was reprieved from her ordeal by Tughril’s death soon afterwards.25 Sometimes women were married off several times by their male relatives in the course of a short period. One Seljuq princess, for example, the daughter of Ridwan, married in succession three Turcoman amirs of the Artuqid family.26 Once married, the woman usually stayed with her husband’s family, even after his death. If the widow was no longer remarriageable, she would still remain with her husband’s clan. Occasionally it was possible for a high-­ ranking woman to return to her own family after her husband’s death or even before, if she could put enough pressure on those empowered to permit her to do so. Arslan Khatun, Sultan Tughril’s niece, for example, left Baghdad with her uncle, having complained that her husband, the caliph, had repudiated her.27 In the case of a royal marriage, the bride’s entry into the groom’s city was a grand event. She was often carried on a bridal litter. Ibn Bibi, historian of the Seljuqs in Anatolia, describes how the bride of ‘Ala al-­Din Qayqubad I (r. 616–34/1219–37) was brought from Syria to her future husband: ‘The amirs and army commanders of Syria set up as a bridal litter seven pavilions of gold and silver; they decorated it with different jewels and attached it to the back of mules.’28 A caliph’s betrothal was also an opportunity for munificence. The Baghdad historian Ibn al-­Jawzi describes how the caliph al-­Muqtafi distributed jewels, camphor and amber on the occasion of the drawing up of his marriage contract with Fatima Khatun, daughter of the Seljuq sultan Muhammad, in 531/1136.29 A royal bride often received lavish gifts at the time of the wedding. When the caliph al-­Mustazhir married Sayyida Khatun,

w ome n i n the se lj uq p e r io d  | 189 daughter of the Seljuq sultan Malikshah, he gave his bride clothes, money, jewels, slavegirls, servants and riding animals.30 Marriage consummation is a subject barely alluded to in the sources. It is noteworthy to mention in this context the delightfully overblown tone of the account of the wedding of the Seljuq sultan of Rum, ‘Izz al-­Din Kayka’us I (r. 607–16/1210–19) to an unnamed royal woman from Syria. The ceremony’s climax is recorded in unusual detail and even with relish by Ibn Bibi: The wives of the amirs who were present went out of the city to receive the bridal litter . . . Noblewomen stood on both sides to attend the illustrious bridal litter. They went to the royal bedchamber and lowered the princess onto the bridal throne of mercy and good fortune. The illustrious king then went into the bridal chamber. The noblewoman with cheeks like Venus went behind the curtain of the chamber. The Sun of Sultans placed his foot on the bed with the Moon of Noblewomen. Kneeling down, the servants of the princess loosened the footwear from their mistress’ feet. The sultan lifted the cap of royal dignity from his head and loosened the fastener from the royal belt. And in pursuance of the authority given by the shari‘a he removed the delectable seal from that illustrious letter.31

Sometimes powerful emotions of love and grief emerge from the stereotyped, laconic historical chronicles and biographical dictionaries to suggest close and harmonious bonds between caliphs or sultans and their wives or concubines. The devotion of the caliph al-­Nasir to his wife, Saljuqa (or Saljuqi) Khatun, especially impressed the chroniclers and biographers. According to Ibn al-­Sa‘i, Al-­Nasir felt such sadness and grief at her death that he was prevented from eating and drinking for days. Her house, with all the clothes and household effects, was left exactly as it was for a number of years; it was not opened up and nothing was taken out of it.

Devotion to his wife’s memory then prompted al-­Nasir to endow next to her funerary turba a splendid ribat with a spacious courtyard and a beautiful garden overlooking the Tigris, which watered it.32 Nor were daughters always the curse that popular tradition made them. Although the eleventh-­century writer Yusuf Khass Hajib expresses the commonly held view about daughters, that ‘the best is that they are not born at

190  |  the me d ie va l tur k s all, or else that they do not survive’,33 they were clearly cherished by the ruling families of the time. The second Artuqid ruler of Mardin, Temürtash, for example, is described by his chronicler, Ibn al-­Azraq, as ‘grieving at the death’ in childbirth of his daughter Nura Khatun in 546/1151.34 How much freedom of movement did Seljuq women enjoy outside the home? What were the accepted norms? These questions are difficult to answer on the basis of the sources. According to the evidence of Islamic painting, urban women were seen occasionally in the presence of the qadi (judge) in the Shari‘a court or in specially segregated sections of the mosque. Elite women were seen at court, where there were also female entertainers.35 High-­ranking women, such as the wives of amirs and notables, were expected to leave the city walls, albeit at night sometimes, to receive an incoming bride of a similar status to their own.36 Such women are also mentioned as leaving cities to undertake peace negotiations after sieges. A fine example was the redoubtable battleaxe, the mother of the Mirdasid ruler of Aleppo, who made a truce with the Seljuq sultan Alp Arslan.37 Less acceptable was conduct unbecoming to Muslim women, conduct deviating from the norms laid down by Islamic law and social conventions. Historical sources occasionally mention instances of such conduct and sometimes hint at possible improprieties. One such instance occurred in 543/1149 in the Jazira when a certain royal woman, the khatun, daughter of Erzurum’s ruler and wife of Akhlat’s ruler, passed through the town of Hisn Kayfa on her way to perform the hajj (pilgrimage to Mecca). Shortly afterwards certain notables arrived, and the ruler of Hisn Kayfa contacted the amir of nearby Mardin. The result of their deliberations was that ‘they stopped the Khatun from going to the Hijaz, which is what the Shah-­i Arman [her husband] had asked them to do’.38 The woman was then questioned and sent back home. Her subsequent fate is not mentioned. Ibn Jubayr, a Muslim Spanish traveller who traversed the Islamic world in 579/1183, mentions a similar episode. It concerns three royal Turkish women, khatuns, who had actually been successful in reaching Mecca to perform the pilgrimage. He concentrates especially on one of these, Saljuqa bint Mas‘ud, who has already been mentioned. She was the daughter of the Seljuq sultan of Konya. Having dwelt on her high status, Ibn Jubayr takes time to mention that this princess ‘went with her personal servants and retinue to

w ome n i n the se lj uq p e r io d  | 191 Mecca on the Friday night. She returned at nightfall on Saturday, and the arrows of conjecture flew at random concerning the cause of the departure of this much indulged princess’.39 A whiff of impropriety still hovered about independent action, therefore, even during the hajj. The major nondomestic activities of élite women involved pious works and more especially the patronage of architecture. Such occupations, although probably relatively rare, are mentioned in historical sources and were clearly regarded as laudable in the eyes of Muslim men. The involvement of certain royal and other high-­ranking women, usually wives of amirs, in architectural patronage is also recorded in monumental inscriptions throughout the Seljuq domains in Anatolia, Syria, Iran and Central Asia. It would be erroneous, however, to place too great an emphasis on women’s role in this sphere, because statistically the number of monumental inscriptions that name women is still small. Of the thousand-­odd Arabic inscriptions from the entire Islamic world that survive from the years of Seljuq power, fewer than 100 relate to women, and of those only seven come from the Seljuq world.40 Of course, the evidence of inscriptions is patchy and incomplete. It is dangerous to generalise on the basis of chance survivals, especially since in the case of the Seljuq realms the buildings suffered from the onslaught of successive waves of Mongol destruction. Nevertheless, it is interesting to examine in closer detail those Seljuq inscriptions naming women. It is important to stress that these are found not only on humble tombstones but also on grander buildings such as mausolea, madrasas (religious schools), caravanserais, and on one occasion a mosque. Among these inscriptions is one dated 512/1118 in the name of the daughter of the Seljuq sultan Mahmud, Zumurrud Khatun, on the mausoleum of ‘Ali al-Rida’ at Mashhad. It was this Seljuq princess who undertook the monument’s renovation.41 A curiosity is the inscription dated 519/1125 on a tomb in Mecca that housed the earthly remains of a woman named al-­Jamaliyya Baltun bint ‘Abdallah, who was singled out ‘for having suckled the imam al-­ Mustazhir’, the ‘Abbasid caliph.42 In the same area, in the environs of Mecca at Mina, an inscription in the name of al-­Mustazhir’s wife records the fact that she had endowed that place as a waqf (pious bequest) for Sufi women.43 Perhaps the most grandiose of the Seljuq monuments bearing an inscription naming a woman is to be found on the caravanserai known as Ribat-­i

192  |  the me d ie va l tur k s Sharaf in Khurasan. The inscription dated 549/1154–5 mentions that the monument was erected (more accurately, repaired and refurbished) through the care of the khatun, the queen of women, ‘the glory of the family of Afrasiyab’. She is then named ‘Qutlugh Balka Sayyida Turkan, daughter of the khaqan [emperor or king]’.44 This was the wife of Sultan Sanjar himself. Another Seljuq royal woman from Rum, Mahperi Khatun, erected five madrasas and a mosque at Kayseri.45 In the sphere of architectural patronage, the extant epigraphic evidence is corroborated fully in historical sources, both dynastic chronicles and biographical dictionaries. A few notable examples of textual evidence will suffice here. Isma Khatun bint Malikshah, wife of the caliph al-­Mustazhir, was buried in a madrasa in Isfahan that she had endowed in favour of the Hanafi madhhab (school of law).46 An even more active patron of architecture was Banafsha bint Abdallah al-­Rumiyya, a specially favoured concubine of the caliph al-­Mustadi. She endowed a waqf on her house for it to be used as a madrasa and also built two bridges and a mosque in Baghdad.47 This prominent role of Seljuq royal women, proved both from epigraphic and textual evidence, in the sphere of architectural patronage raises some important issues.48 Monumental inscriptions are highly public in nature, designed to record for posterity the names of those mentioned in them. There was thus clearly no ignominy attached to the public display of the names of prominent Seljuq women. Moreover, the vast majority of monuments founded by such women are Islamic foundations, that is, religious buildings. Public welfare, in this case the safety of travellers (often regarded as a royal responsibility), is the function of the caravanserai known as Ribat-­i Sharaf, a grandiose structure, founded by a Seljuq royal woman whose name is writ large, literally and metaphorically, on it. Nor are these inscriptions mentioning women tucked away in remote corners of the Islamic world. A woman’s name was engraved on the holy mausoleum of ‘Ali al-Rida’ at Mashhad. Ibn Jubayr, describing the Seljuq princess of Rum, Saljuqa Khatun, comments, ‘She has provided many good works upon the pilgrim road, among them being public water works.’ He also points out that about a hundred camels were needed to transport her clothing and provisions for her journey to Mecca.49 As seen above, women’s names are also found in the heart of the religious centre of Islam, Mecca, for successive generations

w ome n i n the se lj uq p e r io d  | 193 of pilgrims to view when they arrived to perform the ceremonies of the hajj.50 Coupled with the costs of architectural patronage was largesse dispensed in acts of piety. The sources stress the virtue, piety and charity of Seljuq women.51 An especially noteworthy example appears to have been Shahan, the freedwoman of the caliph al-­Mustansir. According to her biography as recorded by Ibn al-­Sa‘i, she enjoyed the use of large sums of money because of her highly favoured status with the caliph. Quoting a verbal report from one of her household staff, Ibn al-­Sa‘i mentions that the monthly bill for work carried out for her by tinsmiths, cloth merchants, coppersmiths, jewellers and other craftsmen exceeded 100,000 dinars. Shahan was well known for her generosity to the poor, widows and orphans. Even after al-­Mustansir’s death she continued to receive generous grants of money from the new caliph, al-­ Musta‘sim, who established her comfortably in Banafsha’s house. Her retinue seems to have included numerous slave-­girls, servants and retainers.52 This information, as well as other references in historical sources to the independent financial status of Seljuq royal women, goes far to explain how they were able to expend wealth of their own on buildings and pious deeds. In the area of women’s intellectual life, the sources yield less evidence. For example, Ibn Khallikan’s biographical dictionary, Wafayat al-a‘yan, a vast collection of biographies of prominent persons in the Islamic world until his own time, includes only seven biographies devoted to women, of which only three date from the Seljuq period.53 Two of them were religious scholars. The first, Zaynab bint al-­Shari (d. 615/1218–19), came from Nishapur. Ibn Khallikan names the teacher under whom she studied Islamic law, and he mentions that she obtained certificates authorising her to teach. He describes her as a ‘woman of great instruction’.54 The second, Shuhda al-­Katiba, Fakhr al-Nisa’ (pride of women), was the wife of al-­Anbari, a prominent notable at the court of the ‘Abbasid caliph al-­Muqtafi. She is accorded high praise by Ibn Khallikan: ‘By her learning, she acquired an extensive reputation and ranked among the first scholars of the age; she wrote a beautiful hand and instructed great numbers in the Traditions, which she had received from the highest authorities.’55 The third biography of a Seljuq woman included by Ibn Khallikan is that of a poet, Takiya bint Ghayth. Although she lived in Ayyubid Damascus, she

194  |  the me d ie va l tur k s is worth mentioning because Ibn Khallikan portrays her as a formidable old woman with the characteristics of wisdom and wit found in the stereotypes of women in Arabic and Persian adab (secular) literature. Indeed, she is shown to have had the last word, even with princes. The Ayyubid prince of Damascus, on reading a poem of hers in which she describes a party with wine drinking, said, ‘The old lady knows all those things from her youth.’ Takiya, on hearing his comment, decided to teach him a lesson. She composed another ode, this time about war, which she described perfectly. She then sent it to the prince saying, ‘My acquaintance with that [wine] is like my acquaintance with this [war].’ These three female biographies concern women of intellectual achievement. They were not from the milieu of the palace but presumably from the middle class. Occasionally a Seljuq royal woman is also mentioned as having intellectual merit. For example, according to Ibn al-­Jawzi, the wife of the caliph al-­Muqtafi, Fatima Khatun, daughter of the Seljuq sultan Muhammad, could read and write.56 What of the historically recorded, rather than alleged, political role of women in the Seljuq period? How justified were the antagonism and suspicion of Muslim writers of Mirrors for Princes towards women meddling in politics? The relationship between the Seljuq wazir Nizam al-­Mulk and the wives of Alp Arslan and Malikshah may well have influenced the views on women expressed in his Siyasatnama. Certainly, it is likely that his attempts to discipline and restrict Malikshah’s redoubtable wife, Terken Khatun, caused her to machinate against him and plot his downfall. The fact that Seljuq royal women had their own financial resources, granted to them by their husbands or sons, allowed them to enjoy some political influence; they could draw on the support of their own slaves or suborn the military. After the murder of Nizam al-­Mulk in 1092, allegedly at the hand of the Assassins, Terken Khatun tried unsuccessfully to install her young son Mahmud on the throne. The evidence of the Seljuq historian al-­Husayni is telling: Terken Khatun had an immense influence over affairs during the days of Malikshah. [The first reason for this was that] she used to treat the soldiers kindly . . . The second reason was that she belonged to the family

w ome n i n the se lj uq p e r io d  | 195 of Turkish kings of Afrasiyab. The third reason was that the money was in her hands.57

In his biography of Terken Khatun, Ibn al-­Sa‘i writes, Sultan Mahmud ‘took power after him [Malikshah], under the administration of his mother. She had with her around 10,000 Turkish mamluks. She ran the kingdom and led the armies until she died in Ramadan 487. When she died, the position of her son Mahmud collapsed’.58 Royal women such as Terken Khatun played a dangerous game and often suffered a violent death. Later in the Seljuq period Inanch Khatun proposed marriage to Sultan Tughril II and lured him by mentioning her financial assets: ‘Now that God has made you ruler of your father’s territories, I am one of your servants. I have a large amount of treasures and tremendous riches.’59 Tughril took the bait and duly married her. Then he had her killed. The Seljuq period is punctuated with similar episodes demonstrating the political involvement of royal women and, above all, the ruler’s wife or concubine. The death of a sultan or amir was often concealed until the accession of the woman’s son was safely assured by winning the support of the military and gaining possession of the citadel and the treasury. One such instance involved the khatun, daughter of Tughtegin of Damascus and wife of İl-­Ghazi, the Artuqid ruler of Mayyafariqin. Her husband died outside the city in 1122. She and her son Sulayman brought the body as far as the gate of Mayyafariqin, whereupon they placed İl-­Ghazi upright on his horse and manoeuvred access into the city. İl-­Ghazi’s death was announced publicly only when they had taken firm control of the citadel and the khatun’s son had been declared the new ruler.60 The aim of such a subterfuge was not just to secure the succession for her son but, more important, to ensure that she herself would enjoy the power and privilege of being the ruler’s mother. A mother’s involvement in state affairs, however, was often far from maternal. The Seljuq prince Duqaq, for example, was allegedly poisoned by his own mother with a bunch of grapes.61 Royal women influenced the actual running of state affairs. A telling example is that of the mother of the Seljuq sultan Arslan (d. 1175). According to Rawandi, this woman even saw to the needs of the army and participated in military campaigns.62

196  |  the me d ie va l tur k s Rivalry, often intense, between wives and concubines contributed to the general male perspective of lawyers and administrators that women’s influence in the political sphere was pernicious. Of the eleven biographies of Seljuq women included by Ibn al-­Sa‘i in his book, six cover the concubines of caliphs and sultans. As well as sharing the virtuous attributes of piety and good works, such women were known to work vigorously and unscrupulously to further the aspirations of their sons and their own share of political power. In the later Seljuq period, for example, Inanch Khatun, wife of Pahlawan, complained bitterly to her husband that his two sons by a Turkish slave mother had been given preferential treatment. ‘How can it please you that the son of a slave-­woman should be higher in status than my two sons?’63 Another example of women’s involvement in politics concerns the fifth Artuqid ruler of Hisn Kayfa (d. 581/1185). In accordance with traditional practice, he had married Saljuqa Khatun, daughter of Qilij Arslan II, but he fell in love with a singing-­girl whom he also took to wife. Soon she controlled his kingdom and his purse. Saladin later made this uxorious vassal of his send the singing-­girl away.64 Although royal women in the Seljuq period were sometimes unscrupulous and ambitious, they were also sometimes praised for their qualities of good leadership, determination and high intelligence. Shahan, the freedwoman of al-­Mustansir, presided over a mazalim court (which handled the redress of grievances) in her own right and had the status and prestige to see that her judgements were implemented.65 Conclusion Inevitably, women in the Seljuq period led the same lives and shared the same fate as other women during the Islamic Middle Ages. Their domain was the home, their responsibilities were domestic and familial, and their lifestyle was circumscribed by Islamic social norms. Muslim men hedged them about with restrictions of movement and firm rules of behaviour. Urban women fared worse than their rural counterparts, for urban women were valued possessions to be guarded. As al-­Jahiz had observed in the ‘Abbasid period, ‘The only purpose of high walls, stout doors, thick curtains, eunuchs, handmaids, and servants is to protect them [women] and to safeguard the pleasure they give.’66

w ome n i n the se lj uq p e r io d  | 197 Yet it would appear that women learned strategies to loosen the constraints of their male-­dominated society. The epigraphic, artistic and textual evidence for the Seljuq period suggests that the evidence of the ‘ulama’’s writings should be treated with caution: they speak about what should be rather than what is. It seems irrefutable that women of rank in the Seljuq period were not simply underlings; especially after they had borne male children they were able to wield considerable political power and influence. Moreover, the stereotype is fleshed out in anecdotes from actual Seljuq historical chronicles. Young women are cited only as beloved daughters or as names in marriage contracts. But once liberated from her role as a sex object, a high-­standing Seljuq woman was free to develop other aspects of her life. Was there any perceptible shift in women’s roles with the advent of Seljuq nomads into the eastern Islamic world? It would certainly be misleading and dangerous to generalise here. The evidence discussed in this chapter, although restricted chronologically to the Seljuq period, is not hermetically sealed within that period. Earlier information from Buyid times and later data from Mongol sources reveal a much more nuanced chronology.67 Evidence from Ibn al-­Sa‘i, for example, indicates persuasively that the court women of the early and middle ‘Abbasid period are depicted as witty, cultivated and skilled in poetry and music. The shift towards piety and good works, so prevalent among elite Seljuq women, may well have occurred, or begun to occur, more than a century before the coming of the Seljuqs.68 The evidence adduced in this chapter strongly corroborates the view expressed by Ibn Jubayr, who came to the Seljuq domains from the extreme west of the Muslim world. He is struck by what to him anyway is a n ­ ovelty – ­Turkish women of high rank, in whom, as he himself describes it, ‘there is this most strange combination of pious works and regal pride’.69 Seljuq women often plotted, bribed, and murdered. They are portrayed as machinating and conniving, vigorous and effective. In any case, they were not to be ignored. They possessed, moreover, the social position and financial means necessary to exercise political power, to perform public acts of charity, to enhance the prestige of their menfolk and families, and to display the impeccable religious credentials associated with these latest converts to Islam.

198  |  the me d ie va l tur k s Notes An earlier version of this chapter appeared as ‘Seljuq Women’, in The Balance of Truth: Essays in Honour of Professor Geoffrey Lewis, ed. Çig˘dem Balim-­Harding and Colin Imber (Istanbul, 2000), 145–63.   1. For an overview of the Seljuqs in Iran, see Cambridge History of Iran (Cambridge, 1968), vol. 5, chaps 1, 2.   2. The evidence cited in this chapter has been deliberately chosen to supplement data given by A. K. S. Lambton in her discussion on Seljuq and Mongol women. See Lambton, Continuity and Change in Medieval Persia (Albany, 1988), 258–96.   3. Al-­Ghazali, Ihya’ ‘ulum al-din (Revivifying of the sciences of religion) (Cairo, 1967), vol. 2, 27–77; al-­Ghazali, Kimiya-yi sa‘adat (Alchemy of happiness), ed. H. Khedivjam (Tehran, 1976), vol. 1, 301–23.  4. Ibid.   5. Beirut, 1981.   6. See F. Mernissi, Sultanes oubliées (Paris, 1990), 113.   7. See C. E. Bosworth, ‘Administrative literature’, in Cambridge History of Arabic Literature, Religion, Learning, and Science in the ‘Abbasid Period (Cambridge, 1990), 165–7.   8. Nizam al-­Mulk, Book of Government or Rules for Kings, tr. H. Darke (London, 1978), 179ff.   9. Ibid., 179. 10. Ibid., 179–80. 11. Ibid., 186. 12. See The Sea of Precious Virtues (Bahr al-fava’id), tr. J. S. Meisami (Salt Lake City, 1991), 80. 13. Ibid., 163. 14. Al-­Ghazali, Ghazali’s Book of Counsel for Kings, tr. F. R. C. Bagley (London, 1971). 15. Ibid., 163. 16. Ibid., 165–6. 17. See F. B. Spuler, Iran in früh-islamischer Zeit (Wiesbaden, 1952), 281–2. 18. It is perhaps unwise to generalize about the lives of nomadic women across several centuries and in different geographical areas. Nevertheless, the lifestyles of many Central Asian nomadic ­pastoralists – b­ e they Turk or Mongol (and the sources often do not distinguish between them) – are unlikely to have

w ome n i n the se lj uq p e r io d  | 199 changed greatly from the eleventh to the fourteenth centuries. The evidence of fourteenth-­century European merchants and envoys who visited the Mongol court, for example, gives a general picture of the nomadic life they encountered en route. They note that nomadic women even participated in military campaigns. A vivid evocation of such women is to be found in the enigmatic, undated, and unprovenanced paintings associated with the name Siyah Qalam. See M. S. Ipsirog˘lu, Vollständige Faksimile Ausgabe der Blätter des Meisters Siyah Qalem (Graz, 1976), plates 18, 24. 19. Ibn Battuta, The Travels of Ibn Battuta, tr. H. A. R. Gibb (London, 1929), vol. 1, 146. 20. Ibid. 21. See A. S. Melikian-­Chirvani, Arts Asiatiques: Numéro spécial, Le roman de Varqe et Golshah (Paris, 1970); R. Ettinghausen, Arab Painting (Cleveland, 1962), 91. 22. Ibn Butlan, Risala fi shira al-raqiq wa taghlib al-abid, quoted in part in A. Mez, The Renaissance of Islam, trs S. K. Bakhsh and D. S. Margoliouth (Patna, 1937), 162. 23. See B. Faris, Signes musulmanes et vision chrétienne (Cairo, 1961), plate 2. 24. Jihat al-a’imma al-khulafa’, min al-dara’ir wa’l-ima, ed. M. Jawad (Cairo, 1960). 25. Ibn Khallikan, Wafayat al-a‘yan, tr. Baron W. M. de Slane (Beirut, 1970, 1843– 71), vol. 3, 227–8; al-­Husayni, Akhbar al-dawlat al-saljuqiyya, ed. M. Iqbal (Lahore, 1933), 80; Ibn al-­ Athir, Al-Kamil fi’l-ta’rikh, ed. C. J. Tornberg (Leiden, 1851–76), vol. 10, 16. See also G. Makdisi, ‘The marriage of Tughril Beg’, International Journal of Middle Eastern Studies 1 (1970), 259–75. 26. Their names were Balak, Sulayman, and Timurtash. See C. Hillenbrand, A Muslim Principality in Crusader Times (Leiden, 1990), 82, 145. 27. Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, vol. 10, 16. 28. K. Erdmann, Ibn Bibi als kunsthhistorische Quelle (Istanbul, 1962), 23. 29. Ibn al-­Jawzi, Al-Muntazam fi ta’rikh al-muluk wa’l-umam (Hyderabad, 1940), vol. 10, 67; also Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, vol. 11, 31. 30. Al-­Husayni, Akhbar al-dawlat, 81. 31. Erdmann, Ibn Bibi, 23. 32. Ibid., 115–19. 33. Ibid., 187. 34. Two daughters had now predeceased him. See Hillenbrand, Muslim Principality, 141. 35. See The Genius of Arab Civilisation, ed. J. R. Hayes (Oxford, 1978), 51. 36. See Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, vol. 9, 424.

200  |  the me d ie va l tur k s 37. See Ibn al-­Adim, Bughyat al-talab fi ta’rikh Halab, partial edition by A. Sevim (Ankara, 1982), 16–39. 38. Hillenbrand, Muslim Principality, 121. 39. Ibn Jubayr, The Travels of Ibn Jubayr, tr. R. J. C. Broadhurst (London, 1952), 189. 40. See Répertoire chronologique d’épigraphie arabe, vols 7–9 (Cairo, 1936–7). Most inscriptions mentioning women are found on tombstones. 41. Ibid., inscription no. 2978. 42. Ibid., inscription no. 3017. 43. Ibid., inscription no. 2977. 44. See A. Godard, ‘Khurasan Robat Sharaf ’, Athar-i Iran 4/1 (1949), 13. 45. See Ülkü Bates, ‘Women as patrons of architecture in Turkey’, in L. Beck and N. Keddie (eds), Women in the Muslim World (Cambridge, MA, 1978), 245. 46. Ibn al-­Sa‘i, Jihat al-a’imma, 109. 47. Ibid., 111–15. 48. See Bates, ‘Women as patrons’, 245ff. 49. Ibid., 190. 50. Perhaps the relationship between the women mentioned in the inscriptions and specific ‘Abbasid caliphs was a factor in their being ‘permitted’ in Mecca. The prestige of their architectural patronage no doubt redounded on the caliphs themselves. 51. See H. Kilpatrick, ‘Some late ‘Abbasid and Mamluk books about women’, unpublished paper, 1993. 52. Ibid. 53. De Slane’s English translation (Ibn Khallikan, Wafayat al-a‘yan) is more than 2,600 pages in length. For women in other biographical dictionaries, see Bulliet’s chapter in this volume [Women in Iran from the Rise of Islam to 1800, G. Nashat and L. Beck (eds)]. 54. Ibn Khallikan, Wafayat, vol. 1, 551. 55. Ibid., vol. 1, 625. 56. Ibid., vol. 1, 276. 57. See al-­Husayni, Akhbar al-dawlat, 74. 58. Ibid., 131–2. 59. Ibid., 184. 60. Hillenbrand, Muslim Principality, 44–5. 61. Ibn Khallikan, Wafayat, vol. 1, 274. 62. Rawandi, Rahat al-sudur, ed. M. Iqbal (London, 1921), 290–3.

w ome n i n the se lj uq p e r io d  | 201 63. Al-­Husayni, Akhbar al-dawlat, 175. 64. Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, vol. 11, 307–8; Sibt b. al-­Jawzi, Mir’at alzaman (Hyderabad, 1951), vol. 8, part 1, 360. 65. Ibn al-­Sa‘i, Jihat al-a’imma, 122. 66. See C. Pellat, The Life and Works of Jahiz (London, 1969), 257. 67. See B. Spuler, Die Mongolen in Iran (Berlin, 1955), 395–8. 68. Kilpatrick, ‘Some late ‘Abbasid and Mamluk books’. 69. Ibn Jubayr, Travels, 190.

14 Ravandi, the Seljuq Court at Konya and the Persianisation of Anatolian Cities

T

he historical process by which Hellenised Christian Anatolia was transformed into Islamic Turkey remains, for all its intrinsic importance and its absorbing interest, curiously opaque. Sp. Vryonis views the period 1000– 1400 as one of ‘unfathomable chaos’, broken only by the peaceful interlude of the thirteenth-­century Seljuq–Nicaean equilibrium.1 Certainly it is difficult to pinpoint with any precision from the existing sources the ways in which a Byzantine Christian substratum, the indigenous Kurdish and Armenian inhabitants of eastern Anatolia and the incoming Turkic tribesmen from Iran and further east were welded together to form Rum, Islamic Turkey. But there is another reason why the process of Islamisation has remained somewhat opaque, and that is because of the unacknowledged bias of some of the modern historians, whatever their ethnicity, who have worked on this period. That bias has resulted in the relative neglect of one discrete facet of this complicated process, namely, the role played by Persian traditions in the setting up of the Seljuq court at Konya and the creation of a truly Islamic urban context in Anatolia in the thirteenth century. Of course, the continuing influence of Byzantium on the Seljuqs of Rum, especially in the context of this volume, must not be forgotten or underestimated. This article, however, will attempt in a preliminary way to offset the comparative neglect of the Persian contribution to how Seljuq Islamic government and institutions in Anatolia took shape.2 It will be divided into three parts: a brief introduction to set the scene, an examination of the career and work of one Persian scholar, Muhammad b. ‘Ali Ravandi, who turned for patronage to the Seljuq sultan of Konya, and a wider discussion of the Persianisation of Anatolian cities in the thirteenth century. 202

t h e pers ia ni sa tio n o f a na t o l ia n citie s  | 203 Introduction It is clear that the initial Turkish presence in the eleventh century in Anatolia brought with it conquest, destruction and massacre. But over time this picture changed somewhat, for successive waves of Turkish invaders could not fail to cause major demographic shifts in the areas that they occupied. As a result, several embryonic Muslim states were established in eastern Anatolia in the twelfth century. These states, which included the Danishmendids of Cappadocia, the Artuqids of Diyar Bakr, the Mengücekids of Divriği and others, followed in their structure and governance a pattern similar to that of the Great Seljuq state based further east in Iran. That was the obvious model for them to follow, and it brought in its train Persian practices, Persian attitudes, Persian ­traditions – ­and, not least, Persians. This article, then, will concentrate on the Persian dimension of Anatolian cities and their culture. This is a dimension only partly discussed and evaluated by V. Gordlevski,3 Sp. Vryonis,4 Cl. Cahen,5 and other scholars. Greater Iran had long exerted a significant influence on its immediate n ­ eighbours – ­Georgia and ­Armenia – ­in many aspects, from historiography to concepts of divinely authorised rule.6 In distant Seljuq Anatolia it was Persian nomenclature, language and culture7 which were to influence court life as it developed into a fully fledged form in the thirteenth century. As with other twelfth- and thirteenth-­century Middle Eastern states ruled by a Turkish military elite, then, the Seljuqs of Anatolia modelled themselves closely on Persian–Islamic government traditions. These in turn drew on ancient pre-­Islamic Iranian models, whose touchstone of excellence was the Sasanian machinery of rule and the religious norms and forms of Islam. Together these created an enduring symbol of good government. It is a commonplace that Islam is an urban religion; and so it was the cities of Anatolia, not the countryside, that became the cradles of Muslim social development and the obvious forum in which Perso-­Islamic institutions were developed. How did this happen? Clues to this development can be found in an analysis of the career of a Persian polymath, Muhammad b. ‘Ali ­Ravandi – ­legal scholar, historian, courtier, calligrapher and gilder. Indeed, this man can be seen as an exemplar of the process of Persian cultural transmission in Anatolia.

204  |  the me d ie va l tur k s The Career of Ravandi Little is known of the life of Muhammad b. ‘Ali Ravandi, except for what he tells us himself in his work entitled Rahat al-sudur wa ayat al-surur.8 In this work Ravandi says that he comes from Ravand near Kashan, a town famed for producing skilled calligraphers. He studied fiqh in Hamadhan from 570/1174 to 580/1184 and he became a renowned gilder and calligrapher, mastering, as he boasts, seventy different scripts. He joined the entourage of the last Seljuq sultan of Iran, Tughril, and on the latter’s death in 590/1194 and with it the imminent extinction of his own prospects in western Iran, he promptly turned his attention to Rum, seeking the patronage of the Seljuq sultan of Konya. He would perhaps not have done this unless he had been confident that the culture which he himself epitomised would have been ­welcome – ­and understood and ­appreciated – ­there. The Rahat al-sudur was begun, according to Ravandi, in 599/1202–3. The work purports to be a history of the Seljuq dynasty of Iran and was written, as already mentioned, just after the demise of that dynasty. It seems indisputable that Ravandi would have wished to dedicate his work to Tughril, the last Seljuq sultan of Iran; but when that became impossible, and since he wanted to write it ‘in the name of a Seljuq sultan’, he sought instead within the same family a new patron for his work, namely Kaykhusraw b. Qilij Arslan, the Seljuq sultan who had acceded to the throne of the Rum Seljuq state in 600/1204.9 It is difficult to tell whether Ravandi ever actually went to Konya in person. In a sense, it is not a very important issue in the context of this discussion. What matters is Ravandi’s written testimony that he changed the orientation of his work to Anatolia and that he sought the patronage of the sultan of Konya rather than that, say, of the Khwarazmshah Muhammad b. Tekish in Central Asia. In the context of this article Ravandi’s work is important as a significant example of the process whereby Persian court traditions were transferred by Persian intellectuals on the move from Iran proper to Anatolia. This happened both before and especially after the break-­up of the Seljuq empire in Iran at the end of the twelfth century. The Rahat al-sudur is written in Persian (with a liberal sprinkling of quotations in Arabic) and was probably declaimed in the presence of the Seljuq sultan in Konya. Yet it contains not

t h e pers ia ni sa tio n o f a na t o l ia n citie s  | 205 a single word of Turkish. The ­implication – ­the flattering implication? – is that the Rum Seljuq sultan is thoroughly at home in Persian and Arabic. The work features a long panegyric extolling the Seljuqs in general as champions of Sunni orthodoxy and their military might: ‘in Arabia, Persia, Rum and Russia the sword is indeed in their hands. Ravandi also lauds in high-­flown prose and poetry the virtues of the sultan himself’.10 The Rahat al-sudur is clearly intended for a courtly milieu and one may justifiably wonder how much of this work would have been understood by the Turkish sultan and his military commanders. It was, of course, intended far more for the delectation of the sophisticated Persian bureaucrats, scholars and intellectuals who were transferring their services from the Seljuq courts of western Iran to the sultan of Konya. But the point is that the sultan gave all these men employment at his court.11 He did not staff his diwans with Turcoman warriors who probably spoke nothing but Turkish. Although the Rahat al-sudur gives a history of the Seljuq dynasty, the work is much more a Mirror for Princes, which aims at presenting a model of exemplary kingship through the use of poetry, edifying anecdotes and Arabic and Persian proverbial sayings.12 In a sense, it is not to the point whether the sultan understood these quotations, let alone whether he was capable of identifying them. For him, Ravandi may have been a trophy intellectual. But what counts is that he believed that the sultan was prepared to pay scholars like him for those wares, and this implies respect for those wares, if not a close understanding of them. That was perhaps a matter for the next generation. As with earlier Mirrors for Princes works, such as the Qabusnama13 and the Siyasatnama,14 history is used to admonish, educate and enlighten the ruler on what constitutes good and just government. Ravandi is writing during a period of great political disturbance and he uses the contemporary situation to point a moral and adorn a tale. The Arabic peroration at the beginning of the work is both sonorous and sombre in tone: ‘If you have noticed slackness in my expression, writing, skill and exposition, pray do not attribute it to my intention; indeed, my dance is to the measure [of the music] of the time.’15 Ravandi then turns to the Seljuq sultan of Konya, Kaykhusraw, exhorting him ‘to revive the dawla of the Seljuqs and uproot tyranny from the world’.16 At the end of the Rahat al-sudur Ravandi also includes a rather curious section of miscellaneous items intended to highlight some of the

206  |  the me d ie va l tur k s a­ ccomplishments of a courtier (adab-i nadim), including the manners and skills appropriate for the companions of the s­ultan – i­n other words, etiquette. Such material, of course, had long formed part of a literary tradition in both Arabic and Persian and reflects the ambience that must have prevailed in medieval Muslim courts. Here we see an example of how this tradition was planted in Anatolian soil. These miscellaneous items include discussions on chess, backgammon, riding, hunting, archery, as well as the ethics of wine-­ drinking, the principles of calligraphy, the keeping of government accounts and other topics. What is important here in Ravandi’s work is not the obvious fact that this ragbag of items of etiquette is derivative and ­plagiarised – ­much of it is drawn from Hanafi legal ­works – b­ ut that it is included here at all, and that Ravandi is bringing this tradition to Seljuq Anatolia, thereby helping to mould a Perso-­Islamic court milieu in Konya. Once again, then, their presence in the book implies a readiness on the part of the Seljuq Turks of Anatolia to buy into these cultural fashions and norms. The Persianisation of Anatolian Cities in the Thirteenth Century Persian scholars and administrators in the same mould as Ravandi, such as Ahmad b. Mahmud al-­Tusi al-­Qani‘i and Khwaja Dahhani, had gone west to seek their fortunes at the Seljuq court in Konya and other Anatolian towns. There must have been a steady trickle of such peripatetic individuals seeking patronage at the small Muslim courts of Anatolia in the late twelfth and early thirteenth centuries. However, with the advent of the Mongols, the trickle of Persian fortune-­hunters into Anatolia must have become a mighty torrent of refugees. The horrors of the Mongol invasions in central Asia and eastern Iran in the early decades of the thirteenth century sent those Iranians with any dynamism flooding westwards to Anatolia and in particular Konya. Moreover, the contribution made by such r­efugees – ­scholars, craftsmen, prominent religious ­figures – w ­ as well and truly in place before the Mongol invasion of Anatolia itself in 1243. Whilst there is no need to minimise the effect of the Mongol victory at Köse Dagh and what it brought in its wake, there can be no doubt that the Mongol invasion of Anatolia was not like the Mongol invasions of Transoxiana, Khurasan and Afghanistan. These earlier invasions snuffed out the long-­established Islamic culture, with its mosques, libraries and irrigation systems irrevocably destroyed and its populations

t h e pers ia ni sa tio n o f a na t o l ia n citie s  | 207 dramatically reduced. But Anatolia, located at the very extremity of a tidal wave of Mongol conquest which had already enveloped central Asia and Iran, got off relatively lightly. To some extent, the Mongols were a spent force in the westernmost areas of the Near East. Indeed, their first major reverse at Islamic hands, the battle of Goliath’s spring, lay only seventeen years in the future. So Anatolia was treated as a protectorate; a thin Mongol crust beneath which an Islamic culture could flourish. I would argue that it is the crucial generation of Muslims in the period 1220–50 which transformed Anatolian urban life, and above all Konya, and laid the foundations for the subsequent Islamisation of Anatolia to develop. The Seljuq state was to collapse by the end of the thirteenth century, but the crucial religious institutions were already securely in place. The presence of Persians in Anatolia during this vital period has long been known but not enough attention has been paid to it. The nationalist bias of some modern Turkish historiography has meant that key Turkish historians have skirted round this topic.17 Above all, the notion of a critical mass of thoroughly Islamised foreigners settling in quantity in a few important areas and thus radically changing the demography of those areas has not been sufficiently accepted, pondered and understood. The names of Persian craftsmen, like Ardishah, Khusraw, Hasan b. Firuz, ­Khurramshah – ­none of them Turkish ­names – ­are written on Anatolian buildings. Their nisbas indicate that they have come from Maragha, Tus, Arran, Tabriz, Marand and other distant places in Iran and even the Caucasus. The technique of glazed tilework, which was highly developed in Iran before 1220 but virtually non-­existent in Anatolia, suddenly flowers there (and not in devastated Iran) in the period after 1220. The extraordinary efflorescence of caravanserais, known in Iran since early Islamic times but highly developed there under the Great Seljuqs, which characterises Anatolia in this same period of 1220–50, reveals probable Persian influence.18 A network of caravanserais which linked the north–south corridor from Sinope to Alanya in order to facilitate the slave trade from southern Russia to Egypt, together with the building of a dry dock at Alanya for the trans-­ Mediterranean ­trade – ­all this largely constructed after 1­ 220 – a­ rgues for the existence of a functioning bureaucratic state, capable of raising revenue, self-­confident and effective.

208  |  the me d ie va l tur k s Bulliet has suggested in another context that the prevalence of learned religious figures in a particular time and place shows a cultural climate conducive to intellectual endeavour.19 Amongst the key elements of such a climate are large urban populations and the existence of important educational institutions. Precisely this situation obtained in Seljuq Anatolia, which houses the earliest substantial single body of medieval madrasas in the whole Muslim world; thirteen of these are in Konya.20 Urban Sufi institutions were also an important part of the religious infrastructure of Anatolian towns, again especially in Konya. In such urban Sufi ­buildings – ­notably in cloisters, zawiyas and khanqahs – such famous figures as Rumi could reside and teach. The plain but astonishing fact is that the most sophisticated and celebrated representatives of Persian Sufi Islam in the thirteenth century were active not in Iran but in Konya. The repercussions of this situation must have had a considerable impact on conversion to Islam in Anatolia. The principal Seljuq shrine in Anatolia developed around the grave of Rumi (d. 1273) and encouraged not only local visitation but also pilgrimage from afar off. The Seljuq sultans themselves were Turkish. But many of them bore (as their Great Seljuq predecessors did not) Persian names redolent of the Shahnama, such as Kaykhusraw, Kayqubadh and Kayka’us. Even more to the point, their administrators and the religious elite in the cities were Persian. These Persians were responsible for endowing awqaf and for building important monuments such as mosques, madrasas and caravanserais either on their own initiative or at the behest of the sultan. The foundation by rival viziers in S­ ivas – B ­ urujirdi and the Juwayni brothers21 – shows that such foundations might even have had a competitive edge. The walls of Seljuq Konya completed in 618/1221–2, whose decoration proclaims a deliberate eclecticism, show the diversity of the Seljuq polity. Alongside classical statuary, Persian influence is manifested in the use of tilework and in the profusion of quotations from the Shahnama, which, according to the court chronicler of the Anatolian Seljuqs, Ibn Bibi, were inscribed on the walls of Konya.22 The earliest illustrated Persian manuscript, signed by an artist from Khuy in north-­west Iran, was produced between 1225 and 1250, almost certainly in Konya23. What of religion itself? For Islam to flourish in Anatolian cities a critical mass of people for whom religious learning was a way of life was needed. As

t h e pers ia ni sa tio n o f a na t o l ia n citie s  | 209 with Egypt and Syria, which received such eastern religious institutions as madrasas and Sufi cloisters in that period, Anatolia received nothing less than a transfusion of religious expertise which enabled it to grow into an Islamic area, and to benefit from the presence en masse of precisely those religious and scholarly classes for whom there was temporarily no place of influence in Iran, whose new Mongol rulers were not Muslim. It is these classes that Bulliet calls ‘the functioning heart of the historic Muslim community’.24 The evidence of architecture shows clearly that in sharp contrast to Iran where the almost total lack of buildings, in the aftermath of the Mongol invasion from 1220 onwards is striking, the extant Seljuq monuments of thirteenth-­century Anatolia are numerous and richly varied.25 There is a marked efflorescence in the reigns of Kayqubadh I (1220–37) and Kaykhusraw II (1237–46) and for decades after the Mongol victory at Köse Dagh in 1243 there was scarcely any loss of impetus. Only after the 1270s was there any marked falling off in new foundations. This whole process can be traced in the epigraphic record. In the absence of contemporary written sources, the evidence of monumental inscriptions from Anatolia is of particular value. It might be argued that epigraphic data must be handled with caution since they are subject to chance survivals and non-­chronological rhythms. However, the evidence of monumental inscriptions can be extremely ­illuminating – o­ ften they tell us exactly when a monument was built, who built it and what kind of building it was. The overwhelming number of monumental inscriptions in Anatolia in the period 1071–1251 occur in a very narrow chronological band, mostly between 1206 and 1251, and most of these inscriptions are found on religious ­monuments – s­ eventeen mosques and eleven madrasas between 1206 and 1251, in addition to further inscriptions on eight citadels, to say nothing of yet more caravanserais throughout Anatolia from the far west to the far east and from the extreme north to the extreme south. So far as religious inscriptions are concerned, most of those on mosques (ten between 1206 and 1229) precede those on madrasas (eight between 1240 and 1251).26 These samples are admittedly small, but their evidence is telling. All these inscriptions are in Seljuq ­territory – ­Kayseri, Amasya, Antalya, Konya, Tokat, Sinope. It is worth discussing the significance of the s­ equence – ­mosque followed by madrasa – in more detail. In crude terms, mosques indicate an initial

210  |  the me d ie va l tur k s Muslim presence and are a label denoting Islam.27 V. Gordlevski points out that the moment of settling in a new place was marked by the building of mosques. I do not, however, share Cl. Cahen’s confident opinion that the very first Muslims in Anatolian towns ‘certainly practised their religion in small makeshift buildings, no doubt mainly churches or parts of churches that had been confiscated’. Cl. Cahen goes on to say that the appearance of mosques testifies to the ‘progress and officialization of Islam’. This judgement can be challenged for a number of reasons. Firstly, if there had been wide-­ scale appropriation of Christian buildings by the incoming Turks, then there would have been evidence for this. Moreover, any invaders, not being sure of their reception, would surely wait a while before utilising the monuments of the predominant faith of the newly conquered area for their own cult. It is much more probable in frontier territory, which Anatolia undoubtedly was, that the first Turkish Muslim worship in Anatolian towns would have taken place in simple enclosures or in structures made of wood. Cahen also ignores the propagandistic element in the Anatolian conquest. A mosque denotes the triumph of Islam. There is evidence from the earliest waves of Islamic conquest in the seventh century that once areas were taken and occupied, primitive mosque structures were erected. When mosques occur in any quantity, as in early thirteenth-­century Anatolia, they imply Islamisation, for the standard practice in Islamic lands was to renew or enlarge existing Friday mosques rather than to build new ones. Madrasas which inculcate Islamic law follow only later as a religious infrastructure develops. This is precisely the picture in early thirteenth-­ century Anatolia. There is clear evidence from the epigraphic record that the Seljuq rulers were busy setting up an urban Islamic government and religious-­educational system. In particular, the timing of the emergence of the madrasas – which coincides with the influx of Persian religious refugees from ­Iran – ­cannot be accidental. It is not only the presence of Islamic legal scholars to teach in the madrasas that is noteworthy. There is also epigraphic evidence as early as the 1220s pointing to Sufi influences, as well as to the actual building of Sufi cloisters and a Sufi tomb.

t h e pers ia ni sa tio n o f a na t o l ia n citie s  | 211 General Reflections By the thirteenth century, then, there was a solid Muslim presence in Byzantine Anatolia; this had been achieved by Turkic pastoralists and frontiersmen from the eleventh century onwards, and later by the cultivated Seljuq state of Konya. Inner Anatolia was now set to become Muslim gradually, and this process occurred under the leadership of the Turks. In Anatolia, as elsewhere, the Seljuq rulers drank in Persian cultural ways in their cities. This tendency to copy Iran in administration, religion and culture reached its height in the thirteenth century with the fuller development of the Seljuq state in Anatolia and the influx of Persian refugees to Anatolian cities. Thus ‘a second Iran’ was created in Anatolia.28 It is food for thought that, while it was the Turks who conquered and settled the land of Anatolia in the eleventh and twelfth centuries, it was the Persians who were instrumental in bringing to these territories a developed Islamic religious and secular culture. They did so with overwhelming success, especially in the crucial thirteenth century. This contribution should be considered as an important counterweight to the continuing influence exerted by the Seljuqs’ sophisticated neighbour, the Byzantine empire,29 on Muslim Anatolia. What happened next? Turkification in Anatolia in cultural terms had to wait until the fourteenth century and the emirates, and especially the slow growth of Ottoman power. Ravandi’s work was translated into Turkish only in the reign of the Ottoman sultan Murad II (1421–51) and it thus contributed to the beginnings of a similar rhetorical-­historiographical school of writing at the Ottoman court in Istanbul. But in the twelfth and thirteenth centuries it was Persian culture that provided a major formative influence on the newly emerging Islamic Anatolian courts and cities. Ravandi, la cour seldjoukide à Konya et la persianisation des villes anatoliennes (résumé). – Cet article aborde le sujet assez inconnu de la contribution iranienne à l’islamisation de l’Anatolie. L’article analyse la carrière de Muhammad b. ‘Ali Ravandi, qui était historien, savant, calligraphe et doreur. Ravandi est un exemple frappant de la tradition gouvernementale persane, imprégnée des idéaux et des usages de la culture sasanienne. Les réfugiés iraniens qui entrèrent en grand nombre en Anatolie à la suite des

212  |  the me d ie va l tur k s invasions mongoles de l’­Iran – ­les fonctionnaires, les poètes, les Sufis et, avant tout, les cadres r­eligieux – ­transformèrent de l’intérieur la culture urbaine de cette région. L’article se sert du témoignage des inscriptions qui figurent sur les monuments, qui, grâce aux nisbas qu’on y trouve, révèlent l’influence des artisans, des techniques et même des genres de bâtiments, tels que les caravansarays et, surtout, les madrasas, qui initièrent les générations suivantes d’une élite urbaine profondément islamisée.

Notes   1. A. Ireland, review of Sp. Vryonis Jr, The Decline of Medieval Hellenism in Asia Minor and the Process of Islamization from the Eleventh Century through the Fifteenth Century (Berkeley, 1986), in The Muslim World Book Review 14/3 (1994), 35.   2. Cl. Cahen led the way in advancing this view; cf. Cl. Cahen, ‘L’émigration persane des origines de l’islam aux Mongols’, in Les peuples musulmans dans l’histoire médiévale (Damascus, 1977), 452–6. This was a very general discussion but a useful beginning.   3. V. Gordlevski, Gosudarstvo Seldzhukidov maloi azii (Moscow, 1941).  4. Sp. Vryonis, op. cit.   5. Cl. Cahen, Pre-Ottoman Turkey (London, 1968); idem, La Turquie pré-ottomane (Istanbul and Paris, 1988).   6. Cf. S. H. Rapp Jr, ‘From bumberazi to basileus: writing, cultural synthesis and dynastic change in Medieval Georgia (K’art’li)’, in A. Eastmond (ed.), Eastern Approaches to Byzantium (Aldershot, 2001), 101–18; R. W. Thomson, ‘“The concept of history” in Medieval Armenian historians’, in A. Eastmond (ed.), op. cit., 89–100.   7. For the literary output of Persians in Anatolia, cf. Cl. Cahen, La Turquie préottomane, 331–6.   8. Ed. M. Iqbal (London, 1921).  9. For Ravandi’s own account of his life, cf. ibid., 38–44. Cf. J. S. Meisami, ‘Ravandi’s Rahat al-sudur: history or hybrid?’, Edebiyat 5 (1994), 183–215. For a recent, thorough analysis and appraisal of Ravandi, cf. eadem, Persian Historiography to the End of the Twelfth Century (Edinburgh, 1999), 237–56. For Ravandi’s turning to Konya, cf. ibid., 256. J. S. Meisami argues persuasively that Ravandi had written the bulk of the work for Tughril. 10. Ravandi, op. cit., 17–18, 29–30.

t h e pers ia ni sa tio n o f a na t o l ia n citie s  | 213 11. Cf. Cl. Cahen, ‘The Turks in Iran and Anatolia’, in R. L. Wolff and H.  W.  Hazard (eds), A History of the Crusades, vol. II (Madison, Milwaukee and London, 1969), 686. 12. For a comprehensive analysis of the Persian ‘Mirrors’ tradition, cf. C. Fouchécour, Moralia. Les notions morales dans la littérature persane du 3e/9e au 7e/13e siècle (Paris, 1986). 13. Kayka’us B. Iskandar, A Mirror for Princes: The Qabusnama, tr. R. Levy (London, 1951). 14. Nizam al-­Mulk, The Book of Government or Rules for Kings (London, 1978). 15. Ravandi, op. cit., 64. 16. Ibid., 38. 17. For the contribution of Persian scholars such as al-­Qani‘i and Dahhani, cf. Cl. Cahen, Pre-Ottoman Turkey, 248–58. For the views of a key modern Turkish historian, cf. G. Leiser, A History of the Seljuks: Ibrahim Kafesoğlu’s Interpretation and the Resulting Controversy (Carbondale and Edwardsville, 1988), 78. For a general discussion of this subject, cf. M. Strohmeier, Seldschukische Geschichte und türkische Geschichtswissenschaft (Berlin, 1984). For a recent analysis of modern Turkish writings on the Seljuqs, cf. O. A. Basan, The Great Seljuks in Turkish Historiography, unpublished PhD thesis (University of Edinburgh, 2002), 22–9. 18. Meinecke lists the nisbas of many craftsmen who were clearly Persians working in Anatolia; M. Meinecke, Fayencedekorationen seldschukischer Sakralbauten in Kleinasien (Tübingen, 1976), vol. I, 187–8. For caravanserais, cf. K. Erdmann, Das anatolische Karavansaray des 13 Jahrhunderts, vol. I–II (Berlin, 1961). 19. R. W. Bulliet, Conversion to Islam in the Medieval Period (Cambridge, MA, 1979), 14. 20. Cf. R. Hillenbrand, Islamic Architecture: Form, Function and Meaning (Edinburgh, 1994), 206–15. R. Hillenbrand remarks that ‘these Anatolian buildings have been quite unjustifiably neglected in the history of the genre’ (ibid., 206–7). 21. Cf. B. Brend, ‘The patronage of Fahr ad-­Din ‘Ali Ibn al-­Husain and the work of Kaluk Ibn ‘Abd Allah in the development of the decoration of portals in thirteenth century Anatolia’, Kunst des Orients XI/2 (1975), 1/2, 178. 22. Michael Rogers, citing the evidence of Ibn Bibi, drew attention to this in a paper entitled ‘The Anatolian Seljuks between East and West’, which he gave at the 33rd Spring Symposium of Byzantine Studies (University of Warwick, March 1999).

214  |  the me d ie va l tur k s 23. Cf. A. S. Melikian-­Chirvani, ‘Le roman de Varqe et Golsâh’, Arts Asiatiques XXII (Paris, 1970); F. Çagman-­Zeren Tanindi, Topkapı Sarayı müzesi islâm minyatürleri (Istanbul, 1979), 11; A. Ateş, ‘Un vieux poème romanesque persan: récit de Warqah et Gulshah’, Ars Orientalis 4 (1961), 143–52. 24. Bulliet, op. cit., 138. 25. Cf. J. M. Rogers, Patronage in Seljuk Anatolia, unpublished DPhil thesis (University of Oxford), 1972. 26. Cf. Répertoire chronologique d’épigraphie arabe, eds É. Combe, J. Sauvaget and G. Wiet, vol. X (Cairo, 1939); vol. XI (Cairo, 1952); vol. XII (Cairo, 1953). 27. V. Gordlevski, op. cit., 46; Cl. Cahen, Pre-Ottoman Turkey, 249. 28. This phrase was used by Cl. Cahen in his article ‘L’émigration persane’, 453. 29. For the chronicler of the Rum Seljuqs, Ibn Bibi, the relationship between the Seljuqs of Rum and Byzantium was a fact of life. For two examples, cf. Ibn Bibi, Die Seltschukengeschichte des Ibn Bibi, tr. H. W. Duda (Copenhagen, 1959), 27 and 38. V. Gordlevski highlights the influence of the Byzantine empire, speaking of ‘reflections of the brilliance and magnificence of the Byzantine court’ at Konya (op. cit., p. 157).

15 Artuqids

A

Turcoman dynasty in eastern Anatolia and Iraq, founded by Artuq, a chief of the Döger clan of the Oghuz tribe. Artuq fought for the Seljuq sultan Malikshah, established a power base in Upper Mesopotamia, and governed Palestine from 1086 until his death in 1091. In the early twelfth century, the Artuqids split into two branches: Artuq’s son, Suqman, took power in Hisn Kayfa in 1102, and his direct line ruled there until 1232. This branch was eventually extinguished by the Ayyubids. Another descended from Suqman’s brother, İl-­Ghazi; members of this branch governed Mardin and Mayyafariqin until 1408, when their rule was terminated by the Turcoman Qaraqoyunlu confederation. A third, short-­lived branch of the Artuqids (1185–1234) ruled at Kharput (mod. Harput, Turkey). Between 1100 and 1130 Suqman, his brother İl-­Ghazi, and their nephew Balak b. Bahram stood out as redoubtable opponents of the Franks. These nomadic Turcoman chiefs were still under the nominal suzerainty of the Seljuq sultan, whose main power base was in Persia and Iraq. The threat of punitive action from the sultan prevented them from operating as independently as they might have wished. Shortly before the arrival of the First Crusade (1096–9), Suqman and İl-­ Ghazi were governing the city of Jerusalem on behalf of the Seljuqs. Suqman rallied to Karbugha, the atabeg of Mosul, in his attempt to relieve Antioch in 1098. That same year the Fatimid vizier, al-­Afdal, captured Jerusalem from Suqman and İl-­Ghazi, who withdrew to Mesopotamia. Suqman then became active against the Franks. In 1101 he tried unsuccessfully to take the town of Saruj from Baldwin I, count of Edessa. In May 1104, he joined 215

216  |  the me d ie va l tur k s Jekermish, the governor of Mosul, in defeating the combined Frankish forces of northern Syria in a battle near Harran. Although disunity between the two commanders prevented an expected attack on Edessa, their victory at Harran halted Frankish momentum and was a boost to Muslim morale. In 1105 Suqman again proceeded to Syria, but he died suddenly, allegedly of quinsy, but possibly poisoned by Tughtegin, the atabeg of Damascus. Thus a vigorous opponent had been removed too early in the conflict with the Franks for him to make a real impact. İl-­Ghazi, however, came to the fore almost immediately after his brother’s death and had a much longer career. In April 1110, in the first military campaign sponsored by the Seljuq sultan Muhammad against the Franks, İl-­Ghazi, accompanied its commander Mawdud to Edessa. The following year Mawdud led a second expedition against the Franks with a new coalition, including Ayaz, İl-­Ghazi’s son, but not İl-­Ghazi himself. His neighbour, Suqman al-­Qutbi of Akhlat, died suddenly in the Syrian campaign, and İl-­Ghazi, attracted by the short-­term prospect of gain, attacked his funeral cortege on its route home, a deed that earned him opprobrium. In 1114, İl-­Ghazi again refused to join the next campaign from the east under Aq Sunqur al-­Bursuqi. In 1115 İl-­Ghazi’s disobedience to the Seljuq sultan became open hostility. When another campaign from the east came to Syria under Bursuq b. Bursuq, it faced a Muslim–Frankish alliance among İl-­Ghazi, Tughtegin, and Prince Roger of Antioch, forged to defend local territorial interests. Roger’s subsequent victory over Bursuq at Tell Danith made this the last expedition sent by the Seljuq sultan. The cessation of this threat, and the death of Sultan Muhammad in April 1118, emboldened İl-­Ghazi to take more personal initiatives against the Franks and to expand his power base. He followed his capture of Mardin and Mayyafariqin in 1118 by seizing Aleppo the following year. He then descended into the Orontes Valley. Tughtegin promised to join him, but in the event İl-­Ghazi did not wait for him. Nor did Roger of Antioch wait for other Frankish princes. In the ensuing battle in May 1119, İl-­Ghazi won a resounding victory at the battle of the Field of Blood (Lat. Ager Sanguinis), and the Franks sustained considerable losses. This triumph over Roger was the crowning moment of İl-­Ghazi’s career, gaining him great prestige in the Muslim world. The caliph bestowed on him the honorific

a rtu qi ds  | 217 title Najm al-­Din (Star of Religion); he had succeeded where semi-­official Seljuq campaigns had not, and the glory fell exclusively to him. Yet the short attention span of his Turcomans, who, once satisfied with the plunder from a battle, were reluctant to embark straightaway on further fighting, and his own prolonged alcoholic celebration (which lasted three weeks), prevented him from following up this triumph in the obvious way with an attack on Antioch. İl-­Ghazi, was, above all, opportunistic and would pursue whatever chances for plunder presented themselves. Accordingly, in 1121 he struck northwards into the Caucasus. There he was roundly defeated by the Christian Georgian king, David II, and was lucky to escape to Mardin. He died in November 1122, but there was a dramatic epilogue to follow. His widow ordered two men to hold his corpse upright on his horse as it entered Mayyafariqin so as to secure the citadel for his son Sulayman before his death was announced publicly. The Artuqid threat to the Franks of Syria was not, however, finished. İl-­Ghazi’s dynamic but over-­extended energies were now replaced by those of his intrepid nephew, Balak, who controlled Melitene (mod. Malatya, Turkey) and Kharput. Balak captured Count Joscelin I of Edessa in 1122 and King Baldwin II of Jerusalem in 1123; both languished in his pit at Kharput. The seizure of the Frankish king gave Balak great prestige. He then took Aleppo, but on hearing that Joscelin had escaped from prison, he hastened to Kharput, seized the castle, and threw all the inhabitants, except for Baldwin, over the battlements. Balak was killed by a stray arrow while besieging Manbij in May 1124. He had indeed been a formidable opponent of the Franks in the north. The inscription on his tomb speaks of him as a martyr and carries a quotation from the Qur’an (3: 169): ‘Think not of those who are slain in the way of God as dead. Nay, they are living’ [J. Sauvaget, ‘La tombe de l’Ortokide Balak’, Ars Islamica 5 (1938), 207–15]. The strength of the early Artuqids had lain in the raiding skills of their bellicose Turcoman cavalry, and İl-­Ghazi had relied latterly on his alliance with Tughtegin of Damascus. Between 1100 and 1130, the northern Syrian towns were under constant pressure; threatened by vigorous Frankish expansion, they turned for protection to Muslim commanders such as the Artuqids. But such protection was

218  |  the me d ie va l tur k s a mixed blessing. The proximity of the undisciplined Turcomans, whom the Artuqids could not always control, threatened the safety of the cities. Indeed, İl-­Ghazi’s troops had previously terrorised the citizens of Baghdad, and Balak’s lieutenant in Aleppo forcibly (and illegally) turned churches into mosques. With Balak’s death, the Artuqids relinquished Aleppo and ceased to operate as vigorous, independent opponents of the Franks in northern Syria. Henceforth their role would be to answer calls to arms from more powerful Muslim rulers in Syria and Egypt. Otherwise, they entrenched themselves in their lands in Upper Mesopotamia, involved in local power struggles. Once the career of the Muslim leader Zengi, governor of Mosul, was fully launched in the 1130s, both İl-­Ghazi’s son Temürtash, who succeeded him at Mardin, and his Artuqid cousins at Hisn Kayfa lived in Zengi’s shadow, threatened by his growing power and fearsome reputation. They had a stark choice: send troops to reinforce Zengi or risk his personal reprisals in their own territories. Their perceived misdemeanors were dealt with firmly on several occasions; for example, just before besieging Edessa in 1144, Zengi attacked the Artuqid ruler of Hisn Kayfa, Qara Arslan, who had allied himself with Joscelin II of Edessa. By the time of Nur al-­Din, Zengi’s son, the Artuqids had learned their lesson, and they knew their place. When called upon, they provided troops for Nur al-­Din, as, for example, at Artah in 1164. Saladin too relied on his Artuqid vassals, whom he kept under tight control and in whose domestic squabbles he sometimes intervened. In 1182, Saladin captured Amid, handing it to Nur al-­Din Muhammad, the Artuqid ruler of Hisn Kayfa (d.1185). When Saladin unsuccessfully attacked Kerak in Transjordan in 1184, Artuqid troops were with him. Under Saladin’s direct successors, the Ayyubids, Artuqid power was further reduced, and they lost Amid, Hisn Kayfa, and Mayyafariqin in turn. In the thirteenth century the sole remaining Artuqids of the Mardin branch were successively vassals to the Seljuqs of Rum, the Khwarazmshah Jalal-­Din, and finally the Mongols. Muslim medieval sources rarely mention the crucial fact that the Artuqids ruled predominantly Christian towns. However, as a small military elite, they established a modus vivendi with local notables, giving the citizens military protection in exchange for the payment of taxes. Beginning as nomadic

a rtu qi ds  | 219 chiefs, the Artuqids, as early as the second generation, became attached to the territories they governed. In administration, they ruled a Seljuq successor state in microcosm. Temürtash was more interested in peace than war, and he invited scholars to Mardin. That witty observer of Frankish mores, Usama b. Munqidh, speaks warmly of the three years he spent at the Artuqid court in Hisn Kayfa (1170–1173), and Temürtash’s son, Najm al-­Din Alpı (d. 1176), ordered a translation of the Syriac version of Dioscorides’s De materia medica. Artuqid copper coins, with figural imagery, based inter alia on ancient Greek, Roman and Byzantine models, have suggested a classical revival to some scholars. Moreover, the Artuqids were vigorous patrons of architecture, building mosques, bridges, tombs and caravanserais, and renovating the walls of their cities. After the Artuqids lost Aleppo, their territories were situated too far from Outremer to pose a serious threat to the Frankish territories. But their early leaders had indeed proved tough opponents for the Franks, and the simulated cavalry flights and sudden raids of their Turcomans were justifiably feared. However, İl-­Ghazi and Balak lacked the iron discipline of Zengi and they came too early to benefit from the heightened atmosphere of jihad (holy war) and the support of the religious classes later enjoyed by Nur al-­Din and Saladin. Bibliography Cahen, Claude, ‘Le Diyar Bakr au temps des premiers Urtukides’, Journal Asiatique 277 (1935), 219–76. Hillenbrand, Carole, ‘The career of Najm al-­Din İl-­Ghazi’, Der Islam 58 (1981), 250–92. Hillenbrand, Carole, ‘The establishment of Artuqid power in Diyar Bakr in the twelfth century’, Studia Islamica 54 (1981), 129–53. Hillenbrand, Carole, A Muslim Principality in Crusader Times (Leiden: Nederlands Historisch-­Archaeologisch Instituut te Istanbul, 1990). Ilisch, Ludger, Geschichte der Artuqidenherrschaft von Mardin zwischen Mamluken und Mongolen 1260–1410 a.d. (diss., Westfälische Wilhelms-­ Universität, Münster, 1984).

16 What’s in a Name? Tughtegin – ‘The Minister of the Antichrist’?1

No voice was heard but was afflicted with sorrow, mindful of his generous actions . . . This blessed amir [Tughtegin] had gone to the utmost lengths in enforcing justice and restraining from oppression’ (obituary of Tughtegin).2

Introduction3

T

he years immediately preceding the Crusader capture of Jerusalem in 492/1099 were a period of great turmoil for the Muslim rulers of Syria and Palestine. As is well known, their disunity and mutual rivalries eased the way for the Crusaders, who were able to establish four states in the region by 502/1109. The majority of the Muslim leaders during this key period of Crusader success were Turkish military commanders who had accompanied the Seljuq royal prince, Tutush, and his young sons, Duqaq and Ridwan, to Syria in the power vacuum left further east by the death of the brother of Tutush, the Seljuq sultan Malikshah, in 485/1092. The coming of the Turks in vast numbers to the Arab and Persian lands from the early fifth/eleventh century onwards proved to be a momentous new phase in the history of the Islamic world. Traditionally the Turks, many of whom crossed into Muslim territory under the leadership of the Seljuq family, are presented in the medieval Arabic and Persian sources as pious Sunni Muslims who had converted to Islam before their major invasions began. This portrayal of the Turks, who went on to govern the heartlands of the Dar al-Islam for many centuries, is not surprising. Their Arab and Persian bureaucrats, court chroniclers and poets needed to put a brave face on 220

wha t's in a na m e ?  | 221 political realities and to help to forge the alliance between the military force of the Turks and the long-­established prestige of the Sunni Arab and Persian religious classes. No doubt, over time, some of the Turkish sultans did indeed grow into the role given them by their biographers, court poets and ­chroniclers – ­namely, that of defenders of Sunni Islam and fighters of jihad against the infidel invaders. But alongside the panegyrics accorded them by such writers, there was always a tendency on the part of some writers, such as Usama b. Munqidh4 and Ibn al-‘Adim,5 to look with distaste at the mores of the alien Turkish newcomers, only a few generations removed from the nomadic life in the steppes, and to present them as a violent, drunken, barbarian ‘Other’. Yet, even familiarity and enforced coexistence with the Turkish-­speaking ­newcomers – m ­ ilitary commanders, their private armies and their Turcoman ­auxiliaries – d ­ id not create any real understanding or analysis of the nomadic background, traditions and Shamanistic beliefs and rituals of the Turks. So the stories in Muslim sources which shed a light on the continuing existence of such vestiges of the pre-­Islamic Turkish past remain unexplained in ethnographical terms. These stories come across in a different light altogether. They are used as religious and cultural ammunition against the Turks whose strange antics are savoured and mocked by a more sophisticated Arab and Persian readership. Thus the conquered can get their own back on the conquerors. As for the descriptions of the Turks written by Crusader chroniclers, they too devote regular attention to this topic, often giving detailed descriptions of Turkish military tactics. These historians, unlike most of their Arab or Persian counterparts, are, of course, not held back by any reasons of political expediency from writing what they wish to say about the Turks. So Crusader chroniclers, writing a language inaccessible to the enemy, are free to include less favourable and more sensational stories about the Turks. These anecdotes were intended to fascinate and shock their countrymen back in Europe and to reinforce religious and cultural stereotypes. This study examines in detail one such story about the Turks, told in both Muslim and Crusader sources; it concerns the killing of the Crusader ruler of Tiberias, Gervase of Basoches, by the Turkish military ruler of Damascus, Tughtegin, in 501/1108.6 The broad outline of Tughtegin’s career in the

222  |  the me d ie va l tur k s early years of the Crusader presence in the Levant has been traced already in secondary sources and need not be rehearsed again here.7 Suffice it to say that in the early years of the Crusader presence in Syria and Palestine, Tughtegin pursued a policy of limited and gradual local expansionism against the Franks, but on occasion he could also see the necessity of forging short-­ lived, opportunistic alliances with them when the small city-­states of Syria were threatened from outside.8 The Death of Gervase of Basoches, Lord of Tiberias, in 501/1107–8 The Crusader leader Gervase of Basoches received the fief of Tiberias in 498/1105; but he did not live long enough to enjoy this role to the full. His death in 501/1107–8 is recorded in a number of sources, both Muslim and Crusader. The key account of this event is that of Albert of Aachen, a chronicler who is ‘remarkably impartial towards the enemy, the Muslims’,9 and who is aware of the differences amongst the enemy between the ­Turks – ­former nomadic warriors who are now a ruling ­elite – ­and Saracens, the term he gives to the indigenous Muslim inhabitants of Syria and Palestine.10 Albert of Aachen’s account of Gervase’s killing is detailed and shocking. To his audience back home in Europe it might even have seemed amazing and almost unbelievable: Gervase was brought out in the middle of the town of Damascus, and after much mockery he was shot through by the Turks’ arrows and gave up the ghost. Gervase, eminent knight, having died in this way, Soboas,11 one of the most powerful of the Turks, ordered his head to be cut off, and the skin of his head, with his hair which was white and abundant and had not been cut for a long time, to be pulled off and dried, because it was wonderfully ornamental and so that it might always be borne aloft on the tip of his spear, as a token and memorial of the victory and to stir up the grief of the Christians.12

Similar details to those given by Albert of Aachen can be found in the account of another Crusader author, Guibert of Nogent, who relates that Gervase was killed by arrows, the crown of his head was removed and a cup was made from his skull.13

wha t's in a na m e ?  | 223 What of the Muslim accounts of this episode? Most of these go so far as to mention that Gervase was killed whilst he was Tughtegin’s prisoner, but they generally gloss over certain macabre details. One should begin with the narrative of Ibn al-­Qalanisi, the city chronicler of Damascus and the panegyrist of Tughtegin’s family. He is deliberately evasive. Unusually, he takes the trouble to avoid the customary practice of Muslim chroniclers, which is to vilify Frankish leaders. Instead, Ibn al-­Qalanisi praises Gervase in fulsome terms as: One of the chiefs of the Franks who were noted for knightliness, courage, gallantry, and prowess in combat, a man of the stamp of King Baldwin in his leadership of the Franks.14

This unusual outburst of admiration for a Frankish enemy from the pen of Ibn al-­Qalanisi must be, in part at least, a stratagem to divert the reader’s attention away from the horrors of the way in which Gervase was killed. It is possible too that Ibn al-­Qalanisi wishes here to enhance the prestige of Tughtegin who has managed to capture a personage of royal status, the nephew of none other than King Baldwin of Jerusalem himself. Writing later in the twelfth century at the court of Tughtegin’s successors, Ibn al-­Qalanisi would not have dared to divulge the details of what happened to Gervase. All that Ibn al-­Qalanisi says rather obliquely in his description of the events which, it is clear from other accounts of the episode, preceded the killing, is that Gervase offered a ­ransom – ­in the ­Frankish – ­Muslim conflict, often a more attractive option than k­ illing – ­for the release of himself and his companions. But Tughtegin refused to accept this. At this point in his account, Ibn al-­Qalanisi writes somewhat flatly: ‘Gervase and those of his companions who were in captivity with him were executed.’15 Another Muslim source, Sibt b. al-­Jawzi, suppresses details of the death altogether, stating simply that Gervase and his associates were sent as a gift to the sultan.16 In the account of Ibn al-­Athir, in which, incidentally, he puts Gervase’s death a year later, in 502/1108–9, he stresses Tughtegin’s inspirational leadership. Tughtegin was equipped with 2,000 cavalry and a large body of infantry whilst ‘the nephew of the Frankish king’ (he does not mention Gervase by name), had 400 cavalry and 2,000 foot soldiers.17 So Tughtegin had the upper hand as far as cavalry was concerned. Nevertheless, at the height of

224  |  the me d ie va l tur k s battle the Muslims began to flee. Tughtegin dismounted and, shouting to his men, he revived their courage. They therefore returned to the battle and routed18 the Franks. Having first praised Tughtegin’s leadership, Ibn al-­Athir then turns to Tughtegin’s treatment of Gervase. In the spring of 501/1108, he and his men fell into the hands of Tughtegin. Ibn al-­Athir relates that, having refused Gervase’s offer of a ransom of 30,000 dinars and the freeing of 500 Muslim captives, Tughtegin ‘was not satisfied with anything but his conversion and, when he would not agree, killed him with his own hand. He sent the prisoners he made to the caliph and the sultan’.19 So Ibn al-­Athir goes so far as to say that Tughtegin personally killed Gervase. Not all the Muslim sources shrink from recounting the gory details of Gervase’s death. Clearly, stories similar to that told by Albert of Aachen and Guibert of Nogent must have been circulating in the Levant. They are eventually mentioned, in rather laconic but frank fashion, by the much later Muslim historian, Ibn al-­Furat (d.  807/1405). He provides the gruesome detail that Tughtegin hollowed out Gervase’s skull while he was still alive and drank from it:20 It (Tiberias) remained in their (Frankish) hands until Zahir al-­ Din (Tughtegin) attacked the Count, the Lord of Tiberias, in Shawwal in the year 501 [that is, 14 May–11 June 1108]. He hollowed out the skull of his head while he was still alive, and drank wine in it, while he was looking at him [sic]. He lived for an hour, then he died.

This extraordinary account is provided with a useful footnote by Ursula and Malcolm Lyons, the editors and translators of this text.21 They suggest two possibilities to unravel the narrative; firstly, perhaps the skull was turned to face Tughtegin while he was drinking from it, or, secondly, Tughtegin was looking at Gervase’s body. In another part of his history, the Lyons point out that Ibn al-­Furat gives a different account in which Gervase’s skull is washed out with salt and water.22 Tughtegin then poured wine into it and drank from it, as did his associates. But what was behind this savage killing? Why did Tughtegin refuse much needed ransom money? Was Tughtegin’s judgement blinded by his intake of alcohol, as was often recorded in stories about Turkish leaders? Did

wha t's in a na m e ?  | 225 he have some personal grudge against Gervase? It should be remembered that Tughtegin had a fearsome reputation among Crusader writers. Walter the Chancellor describes him as ‘the investigator and discoverer of various tortures’.23 So this incident is by no means the only occasion when his cruelty is mentioned, but it is certainly the most memorable. The French scholar Mouton remarks in his short discussion of this episode that this treatment of Gervase was intended to frighten off other Crusader warriors from attacking the territory of Damascus. The communal drinking from his skull was a symbol of a total victory over an enemy.24 The following discussion will highlight two key aspects of the story, Gervase’s scalping and his skull being used as a drinking vessel. The Scalping of Gervase The practice of scalping the heads of enemies in war is long testified far into antiquity in Central Asia. Herodotus writes as follows about the Tauri tribe of the Scythians: Any one of them who takes a prisoner in war, cuts off his head and carries it home, where he sets it high over the house on a long pole . . . The heads are supposed to act as guardians of the whole house over which they hang.25

In probing the surface of the narrative of Albert of Aachen, a good starting point is an analysis of Tughtegin’s name. Given the understandable difficulties experienced by both Crusader and Muslim chroniclers in rendering the personal names of the other side, especially across the divide of different scripts, it is not surprising that extra curiosity is not evinced about the actual meaning of such ‘outlandish’ names. In the case of Tughtegin, Crusader sources make valiant attempts with such versions of his name as Doldequinus,26 Dochinus27 and Duodechinus.28 On one occasion, Albert of Aachen underlines his uncertainty about the name by calling Tughtegin ‘Dochinus or Duodechinus’.29 What does the name Tughtegin/Tuğ Teğin/Toğ Teğin30 in these various forms mean? The name has two parts. The first part – tugh/tuğ/toğ – has several meanings, one of which is ‘standard, flag, tail-­trophy’.31 Rásonyi defines the word, which he writes as tugh, more precisely, as ‘horsetail standards’. The tugh denotes royal authority. The second part of the name – tégin – is a

226  |  the me d ie va l tur k s very old honorific title meaning ‘prince’32 or ‘male relation of the khan’.33 So the whole composite name would seem to mean ‘prince bearing a horsetail standard’. By the sixth/twelfth century, it is clear that this grandiose personal name, despite its royal connotations, could be given to an individual who was not of royal Turkish lineage. Two striking examples spring to mind in this context: firstly, the Tughtegin we have been d ­ iscussing – ­the governor of Damascus and founder of the short-­lived Burid ­dynasty – ­and, secondly, Tughtegin, the brother of Saladin (Salah al-­Din).34 The first of these important figures was, as already mentioned, a Turk, a former slave, manumitted by the Seljuq prince Tutush. So the name must have suffered semantic depreciation and could be applied to non-­royal but prestigious Turkish military commanders.35 Saladin’s brother, on the other hand, was not a Turk; he was probably given the name Tughtegin as a conscious echo of an illustrious predecessor who had fought against the Franks.36 It will be argued below that Tughtegin of Damascus, coming from a fully Turkish milieu, probably did understand the connotations of his own name. However, it is difficult to know whether the same knowledge could be attributed to Saladin’s brother, a Kurd, who lived much later on in the sixth/twelfth century. In the context of this discussion it is important to go back to the origin of the tugh. One definition of it describes it as: ‘a scrap saturated with the enemy’s blood fixed on a pole’.37 Tughs had totemistic importance, and those who raised high the totem and rode with it into battle believed that they were gaining special powers from it. The use of the tugh, carried aloft by a mounted Turkish warrior, galloping at speed into battle, is vividly illustrated in the Rashid al-­Din Jami‘ al-tawarikh manuscript of 714/1314, in the Edinburgh University Library (see Figure 16.1).38 In the context of the story of Gervase’s death, it is important to mention the special features of tughs so clearly demonstrated in exhibits found in museums in Europe and Turkey (see Figure 16.2).39 Such tughs resemble not only a horse’s tail but also a human being’s flowing l­ocks – ­the very detail carefully emphasised in Albert of Aachen’s account, where Gervase’s hair is described as ‘abundant and had not been cut for a long time’, ‘wonderfully ornamental’ and ‘it might be always borne aloft on the tip of his spear’.40 In other words, Gervase’s hair is seen as an extraordinary trophy, a remarkable tugh.

wha t's in a na m e ?  | 227

Figure 16.1  Mounted Turkish warrior, galloping into battle, carrying a tugh (Rashid al-Din, Jami‘ al-tawarikh )

Figure 16.2 Ottoman tughs captured at the Battle of Vienna, 1683, gift of King John Sobieski to St Ann Church, Cracow

228  |  the me d ie va l tur k s It is probably no more than a coincidence that the very warrior who would probably have gone on to carry the tugh that was composed of the hair of the ill-­fated Gervase, should have been called Tughtegin. Be that as it may, the special powers inherited from such a venerable white-­haired old man would have been recognised by Tughtegin as an enhancement to his own status as a great warrior and commander. The Use of Gervase’s Skull as a Drinking Vessel The earliest known Turkish people, the Hsiung-­Nu, as well as the non-­ Turkish Scythians, made drinking bowls from the skulls of prestigious fallen enemies.41 Wine would be poured into such a ‘bowl’ and then mixed with the blood of a sacrificial victim and stirred with a sword. Esin writes tellingly that ‘amongst the Hsiung-­Nus, the Scythians and the proto-­Bulghars, the cup became occasionally the skull of the vanquished enemy and the ambrosia was spilled blood’.42 The Friulan Franciscan friar Odoric of Pordenone, describing the customs of Tibet in the early fourteenth century, writes that: ‘The son eats his father’s head and of the skull he makes a goblet, from which he and all the family always drink devoutly to the memory of the deceased father.’ He adds sententiously that the people in this region also have ‘many other preposterous and abominable customs’.43 As late as the early sixteenth century (916/1510), Shah Isma‘il I treated the corpse of his defeated Öz-­Beğ enemy, Shaybak Khan, in similar fashion. He gave orders for the skull to be removed and for it to be encased in red gold. He then drank wine out of ‘that golden skull’.44 With these examples alone, there is clear evidence of a continuing tradition of skulls being used as drinking vessels amongst the peoples of Asia. Concluding Remarks Reactions to this event on the part of the mediaeval chroniclers are varied. On the Crusader side, William of Tyre, writing many years later, does not mention it. Albert of Aachen, on the other hand, concentrates on the scalping in very graphic detail, no doubt in horror at the perceived barbarity of the deed and to underline to his Christian European readership the despicable mores of the infidel. On the Muslim side, Ibn al-­Qalanisi, writing in the employ

wha t's in a na m e ?  | 229 of Tughtegin’s descendants in Damascus, glosses over the event, as does Ibn al-­Athir, who, in any case, relies closely on the narrative of Ibn al-­Qalanisi for his version of the history of early sixth/twelfth-­century Syria. A much later writer, Ibn al-­Furat, gives an unvarnished account which in its turn focuses on the part of Gervase’s tale which involves his skull being used as a drinking vessel. Here the topos of Turkish barbarian warriors drunk after victory on the battlefield is no doubt too tempting for this author to omit. Stories about the Turks, such as the tragic end to the life of Gervase of Basoches, analysed above, were narrated to titillate and mock and were written with feelings of religious or cultural superiority on the part of Muslim and Crusader sources alike. To the Muslims, the Turks were both their overlords and also a barbarian ‘Other’. To the Crusaders, the Turks were another ‘Other’, even more alien than the infidel Arabs (or Saracens, as they were often called). However, both Muslim and Crusader narratives, when analysed, suggest, unwittingly but persuasively, through the tales they tell of Turkish customs dating back to the rituals of the Inner Asian nomads and to their pastoral hunter lifestyle, that the process of sedentarisation and Islamicisation amongst the Turks must have been much slower and more protracted than the Sunni Muslim sources would have us believe. Certainly the story of Gervase’s killing, featuring as it does the enactment of pre-­Islamic Turkish rituals in the very heart of ­Damascus – ­a Muslim city since the seventh century ad – is thought-­provoking. It highlights the difficulties that the sedentary Muslim Arabs (and the Persians further east) must have experienced in coming to terms with these new and powerful overlords. In the case of Gervase’s death, it is the precise account of a Crusader author, Albert of Aachen (whatever his motivation may have been), which is the most valuable testimony. Notes   1. This title is given to Tughtegin by the Crusader chronicler Walter the Chancellor in his Bella Antiochena, tr. T. S. Ashbridge and S. B. Edgington as Walter the Chancellor’s The Antiochene Wars: A Translation and Commentary (Aldershot, 1999), 165.   2. Ibn al-­Qalanisi, Dhayl ta’rikh Dimashq, ed. H. F. Amedroz (Leiden and Beirut, 1908) as History of Damascus 363–555 A.H. from the Bodleian MS. Hunt 125,

230  |  the me d ie va l tur k s being a continuation of the history of Hilal al-Sabi, 219; tr. H. A. R. Gibb as The Damascus Chronicle of the Crusades (London, 1932), 184–5.   3. It is an honour and great pleasure to write a contribution for the Festschrift of Dr Farhad Daftary, who has done so much valuable research on mediaeval Islamic history.   4. Usama b. Munqidh, Kitab al-I‘tibar, tr. P. K. Hitti as Memoirs of an Arab-Syrian Gentleman, or An Arab Knight in the Crusades: Memoirs of Usamah ibn-Munqidh (Kitab al-I‘tibar) (Beirut, 1964).  5. Ibn al-‘Adim, Bughyat al-talab fi ta’rikh Halab, ed. S. Zakkar (Damascus, 1988).   6. His full names were Abu Mansur Tughtegin b. ‘Abd Allah. His honorific titles (laqabs), acquired whilst he was ruler of Damascus, were Zahir al-­Din and Sayf al-­Islam. Most often in the Islamic sources which recount his career he is simply mentioned as Atabeg or Atabeg Tughtegin. Perhaps it is worth noting that Tughtegin’s father is here called ‘Abd Allah, an indication that he was a ­Muslim – t­his title was also often used to denote a Muslim ­slave – a­ nd thus showing that Tughtegin’s family had converted to Islam at least one generation back; cf. Ibn Taghribirdi, al-Nujum al-zahira fi muluk Misr wa’l-Qahira (Cairo, 1939), vol. 5, 234. Not all these names are mentioned in all the sources; see Ibn Khallikan, Wafayat al-a‘yan, tr. W. M. de Slane as Ibn Khallikan’s Biographical Dictionary (Paris, 1842–71), vol. 1, 274; Sibt b. al-­Jawzi, Mir’at al-zaman fi ta’rikh al-a‘yan (Hyderabad, 1951), 127.   7. See T. K. El-­Azhari, The Saljuqs of Syria during the Crusades 463–549 AH/1070– 1154 AD (Berlin, 1997), 171–229; J. al-­ Zanki, The Emirate of Damascus, PhD thesis (University of St Andrews, 1989); A.-M. Eddé, ‘Tughtigin’, in A. V. Murray (ed.), The Crusades: An Encyclopedia (Santa Barbara, CA, 2006), vol. 4, 1204–5; J.-M. Mouton, Damas et sa principauté sous les Saljoukides et les Bourides, 1076–1154 (Cairo, 1994). The achievements of Tughtegin and his family dynasty, the Burids, are also recorded in a monograph in Turkish but this book contains little proper analysis; C. Alptekin, Dimaşk Atabegleg˘i: TogTeginliler (Istanbul, 1985).   8. The most important source for the life of Tughtegin, ruler of Damascus for almost thirty years, is the town chronicle of that city written by Ibn al-­Qalanisi (d. 554/1160), Dhayl ta’rikh Dimashq. Roger le Tourneau’s French translation, Damas de 1075 à 1154 (Damascus, 1952) is fuller than Gibb’s. Despite his obvious bias towards his Turkish overlords and his convenient omissions and obfuscations of negative information, Ibn al-­Qalanisi is a keen observer of events

wha t's in a na m e ?  | 231 and he has a good grasp of dates. His evidence is invaluable as a starting point for writing the life of Tughtegin.   9. Albert of Aachen, Historia Ierosolimitana. History of the Journey to Jerusalem, ed. and tr. S. B. Edgington (Oxford, 2007), xxxv. 10. Ibid. 11. Edgington mentions this person as ‘unidentified’ in her translation; see Albert of Aachen, Historia Ierosolimitana, 770, n. 2. However, it would seem quite probable that this is a valiant but garbled attempt by Albert to write the Turkish word su-başi (police superintendent), the official who would customarily be charged with the carrying out of a public execution, which this event probably was. 12. Albert of Aachen, Historia Ierosolimitana, 770–1. 13. Guibert of Nogent, cited by Edgington, Albert of Aachen, Historia Ierosolimitana, 771. Surprisingly, in view of their general hostility to the Turks, the usual Eastern Christian sources say nothing about Tughtegin’s killing of Gervase. 14. Ibn al-­Qalanisi, Dhayl ta’rikh Dimashq, ed. Amedroz, 161; tr. Gibb, 86. 15. Ibn al-­Qalanisi, Dhayl ta’rikh Dimashq, tr. Gibb, 87. 16. Sibt b. al-­Jawzi, Mir’at al-zaman, 25. A contemporary of Ibn al-­Qalanisi, Muhammad b. ‘Ali al-‘Azimi, in his very fragmentary chronicle, Ta’rikh Halab, devotes only one short sentence to Tughtegin’s military encounter with Gervase; see C. Cahen (ed.), ‘La chronique abrégée d’al-‘Azimi’, Journal Asiatique 230 (1938), 379. 17. Ibn al-­Athir, al-Kamil fi’l-ta’rikh, tr. D. S. Richards as The Chronicle of Ibn al-Athir for the Crusading Period from al-Kamil fi’l-ta’rikh. Part 1. The Years 491– 541/1097–1146: The Coming of the Franks and the Muslim Response (Aldershot, 2006), vol. 10, 327. 18. Reading kasara rather than kabasa in the text; ibid. 19. Ibid., 142. 20. Ibn al-­Furat, Ta’rikh al-duwal wa’l-muluk, eds and trs (partial) U. and M. C. Lyons (Cambridge, 1971), vol. 1, 55. 21. Ibn al-­Furat, Ta’rikh al-duwal, vol. 1, 223, 55.16. 22. Ibid., 223; Mouton, Damas et sa principauté, 171. Salt seems to have been used in trepanning operations on the skull, as in the famous instance involving a woman possessed of a devil, mentioned by Usama b. Munqidh in his memoirs. In her case, it is a Frankish doctor who performs the deed: ‘he peeled off the skin at the middle of the incision until the bone of the skull was exposed and rubbed it with salt’. Of course, the woman died; Usama b. Munqidh, Kitab al-I‘tibar, tr. Hitti, 162.

232  |  the me d ie va l tur k s 23. Walter the Chancellor, Bella Antiochena, 160. Walter mentions a typical example of Tughtegin’s barbarous punishments in the case of Robert the Leper; see 159–61. 24. Mouton, Damas et sa principauté, 171. 25. Herodotus of Halicarnassus, The Histories, tr. A. de Sélincourt (London, 1954), 276. 26. William of Tyre, Historia rerum in partibus transmarinis gestarum, ed. R. B. C. Huygens as Willelmi Tyrensis archiepiscopi chronicon (Turnhout, 1986); trs E. A. Babcock and A. C. Krey as A History of Deeds Done Beyond the Sea (New York, 1943), 501, n. 84. 27. Albert of Aachen, Historia Ierosolimitana, 766, 834, 852, 854, 856, 874. 28. Ibid., 834, 840, 850. Albert also calls Tughtegin ‘Hertoldinus’. Edgington, ibid., 774, n. 2, rightly suggests that this other name given to Tughtegin is an attempt to reproduce in Latin Tughtegin’s honorific title, Zahir al-­Din. 29. Ibid., 834. 30. Sümer gives both the versions Tug˘ Teg˘in and Tog˘ Teg˘in; see F. Sümer, Türk devletleri tarihinde şahis adlari, ed. Turan Yazgan (Istanbul, 1999), 699, 737. 31. M. al-­Kashgari, Diwan lughat al-turk, trs R. Dankoff and J. M. Kelly as Compendium of the Turkic Dialects (Diwan Luyat at-Turk) (Cambridge, MA, 1982–5), vol. 3, 199; see also G. Clauson, An Etymological Dictionary of PreThirteenth-century Turkish (Oxford, 1972), 469; H.-W. Haussig, ‘Herkunft, Wesen und Schicksal der Hunnen’, in H. R. Roemer (ed.), History of the Turkic Peoples in the Pre-Islamic Period (Berlin, 2000) 278; B. Spuler, Die Mongolen in Iran (Leiden, 1985), 331; J. Sauvaget, ‘Noms et surnoms de Mamelouks’, Journal Asiatique 238 (1950), 50; A. von Le Coq, ‘Türkische Namen und Titel in Indien’, Aus Indiens Kultur (1927), 2; E. Esin, ‘Tös and Moncuk. Note on Turkish Flag-­pole Finials’, Central Asiatic Journal 16 (1972), 25, 32; Haussig, ‘Herkunft’, 277–8; C. Huart, art: Tugh, Encyclopaedia of Islam, 1st edn, vol. 4, 820; the bearer of the standard was called the tugh-beg˘i. C. E. Bosworth, art: Tugh, EI 2, vol. 10, 950. The other possible meaning of tuğ/toğ (‘full, satisfied, sated’) fits less well; see L. Rásonyi and I. Baski, Onomasticon Turcicum. Turkic Personal Names as Collected by László Rásonyi (Bloomington, IN, 2007), vol. 2, 764. I am grateful to Edmund Bosworth for giving me the information from this valuable new onomasticon. 32. Clauson, Etymological Dictionary, 483; Rásonyi, Onomasticon Turcicum, vol. 2, 727; E. G. Pulleybank, ‘The Hsiung-­nu’, in Roemer (ed.), History of the Turkic Peoples, 64.

wha t's in a na m e ?  | 233 33. Rásonyi, Onomasticon Turcicum, vol. 2, 727. 34. For the activities of this brother of Saladin in Damascus, see M. C. Lyons and D. E. P. Jackson, Saladin: The Politics of the Holy War (Cambridge, 1982), 39, 83. 35. Spuler, Die Mongolen in Iran, 330–1. 36. Indeed, Saladin’s brother also bore exactly the same honorific titles as Tughtegin of ­Damascus – ­Zahir al-­Din, Sayf al-­Islam. 37. W. Smith, Flags Through the Ages and Across the World (New York, 1975), 12; Z. Zygulski, Jr, Ottoman Art in the Service of the Empire (New York and London, 1992), 71–3. 38. Rashid al-­Din Tabib, Jami‘ al-tawarikh, Edinburgh University Library MS Arab 20, fol. 109r. 39. These are tughs now in Cracow, illustrated in Zygulski, Ottoman Art, 78. 40. Albert of Aachen, Historia Ierosolimitana, 771. 41. Pulleybank, ‘The Hsiung-­nu’, 52, 69. 42. E. Esin, ‘“And”: the cup rites in Inner-­Asian and Turkish art’, in O. Aslanapa and R. Naumann (eds), Forschungen zur Kunst Asiens. In Memoriam Kurt Erdmann (Istanbul, 1969), 227. 43. Odorico da Pordenone, Relazione del viaggio in Oriente e in Cina (1314?–1330) (Pordenone, 1982), 65; cf. also H. Yule, ed. and tr., Cathay and the Way Thither (London, 1916; repr., New Delhi, 1998), 254. 44. M. K. Youssef-­Jamali, The Life and Personality of Shah Ismail I (892–930/1439– 1524) (Isfahan, 1998), 213, quoting Rawdat al-safawiyya, fol. 78a.

17 Aspects of the Court of the Great Seljuqs

S

everal quite distinct approaches to the study of the Seljuq court might be proposed. One could view it from the perspective of an apparently immemorial, quasi-­mythical ancient Iranian tradition which found its most characteristic expression in Sasanian times but drew its inspiration from Parthian and especially Achaemenid models.1 Or one could highlight the very different traditions inherited from the nomadic cultures of the Eurasian steppe, stretching back in historical times to the Turkic culture exemplified by the Orkhon inscriptions of the seventh and eighth centuries ad. In still earlier times these nomadic cultures may be glimpsed in the perhaps distorting mirror of the Chinese sources, so here too there is a taproot that extends to a very distant past.2 Yet another approach to understanding the Seljuq court might be founded on a close study of ‘Abbasid precedent; the court life in the golden prime of Harun al-­Rashid described in admiring detail in a multitude of Arabic sources.3 It set a standard of luxury, ceremony and ostentatious display to which many a later Muslim court aspired. The Samanid and Ghaznavid courts in the eastern Iranian world took over many of the practices of the caliphal court;4 this was the most readily available model for the Seljuqs to follow, adapting it where appropriate to their own preferences and needs. The topic of the Seljuq court, then, is a complex one with multiple ramifications to earlier cultures. Moreover, the very concept of a court in this period requires close examination. In the present chapter, restrictions of space make it appropriate to limit the discussion to certain key themes. Accordingly, this chapter will highlight a few aspects of the many ceremonial, intellectual and leisure activities pursued by the Seljuq sultans. 234

aspe c ts of the cou rt of the gr e a t s e l j uq s  | 235 Most of the evidence, though not all of it, will be drawn from the heyday of their rule, from the reign of Sultan Tughril (d. 453/1063) until that of Sultan Muhammad (d. 512/1118).5 What is a Court in the Context of the Great Seljuqs? There has been much scholarly interest in the concept of medieval princely courts in recent years, from the context of Europe in the West to that of China in the East.6 Vale’s book on the cultural characteristics of thirteenthand fourteenth-­century north-­western European courts, for example,7 stresses the continuities from Carolingian times onwards. The book also examines such issues as the crucial elements of public display expected of rulers and the necessary material foundations of court life. In many European countries, the court of the medieval ruler was often peripatetic. How much more so was the case for nomadic military usurpers such as the Seljuq Turks, who travelled a great deal, often over vast distances, first to secure territories and then to defend and expand them. So the safest thing to say about the Seljuq ‘court’ is that it could be found wherever the sultan happened to be. Like other medieval courts, that of the Seljuqs followed a predictable routine wherever it went. The first two Seljuq sultans, Tughril and Alp Arslan, spent most of the time in the saddle, criss-­crossing enormous tracts of land in pursuit of territory and booty. Alp Arslan went almost immediately from eastern Turkey, where he had gained his famous victory over the Byzantine emperor in 463/1071 at Manzikert, to Central Asia where he was assassinated the following year. The twenty-­year reign of Sultan Malikshah represents an important transition in Seljuq history. At this point there seems to have been a greater tendency for the Seljuq sultan, and therefore the court attending him too, to stay in one place for at least certain periods of the year. With Malikshah there is a clear attachment to Isfahan,8 although he too continued to travel around the empire but to a lesser extent. Such travel was nothing new in Iran. Indeed, since ancient times, the courts of the rulers of Iran had moved around. The Achaemenid court, for example, moved according to the season, spending autumn and winter in one place and spring and summer in another. At various times Susa, Hamadan and Persepolis served as a royal centre.

236  |  the me d ie va l tur k s Sources For most of our knowledge of the Seljuq period, we are dependent on a good number of Arabic and Persian chronicles which provide detailed information on the dynasty. However, these sources, by their dating and their very nature, have to be treated with caution. They date largely from 150 to 200 years after the heyday of Seljuq rule in the eleventh century and they reflect the retrospective views of non-­Turkish Muslim ­chroniclers – ­usually religious scholars or b­ ureaucrats – ­who lived under the rule of later Turkish overlords in Syria, Egypt, Iraq and Iran and who, without a second thought, cast Seljuq rule into a mould of Perso-­Islamic statecraft. Moreover, these chronicles are written, not in Turkish, but in Arabic and Persian. The Seljuq Turks themselves have no authentic voice. Instead, they are seen in the Muslim sources, as it were, through a rosy Islamic prism.9 A very useful resource, however, for reconstructing the realities of Seljuq statecraft and court life is the ‘Mirrors for Princes’ literature, and the Seljuq period abounds in such material.10 The letters in Persian written by al-­Ghazali to contemporary rulers echo the same ethos.11 But the material in the ‘Mirrors for Princes’ genre is problematic in that it speaks more often of a model court rather than an actual one. It describes what ought to be rather than what the author has actually experienced. This is especially true of the Siyar al-muluk, the ‘Mirror for Princes’ work written by Nizam al-­Mulk, the most famous Seljuq vizier, dedicated to the third Seljuq sultan Malikshah. In it Nizam al-­Mulk harks back to the halcyon days of the Ghaznavid court in Khurasan and he sees a fall in standards in Seljuq protocols and ceremonies as well as in government administration more widely. To sum up this discussion of written sources, it is important to re-­ emphasise that they remain problematic. They date from later periods of Turkish rule. They are written by men who never milked a flock, never pitched a tent, never ate steak tartare, never strung a bow for battle, never galloped across the plain. So, as scholars and bureaucrats, they airbrush out the nomadic milieu. And they fit the Seljuqs onto the procrustean bed of Islamic ‘Mirrors for Princes’. When it comes to reconstructing the life of the Seljuq court, the scholar is faced with the need to seek supplementary information from material

aspe c ts of the cou rt of the gr e a t s e l j uq s  | 237 culture to amplify the fragmentary, often unexplained, snippets of relevant information scattered through the medieval Arabic and Persian written sources. With such concrete contemporary or near-­contemporary ­evidence – ­such as pottery, sculpture and p ­ ainting – i­t is important, where they exist, to consult artefacts as close as possible in date to the Seljuq period. It is worth pointing out that although the period of strong, unified Seljuq rule can be said to have lasted only until 485/1092, or at a pinch until 512/1118 when the Seljuq sultan Muhammad Tapar died, military commanders and atabegs founded smaller polities in eastern Turkey, Iraq and Syria in the twelfth and thirteenth centuries. These small p ­ rincipalities – t­hey can rightly be labelled Seljuq successor-­states – m ­ odelled themselves on the government and court life of the Great Seljuqs. For such Turkish states as ­these – ­the Artuqids in eastern Turkey, the Burids in Syria and the Atabegs of Mosul in ­Iraq – ­there is additional and most useful visual evidence which can be brought into play. Rank The importance of rank was maintained at the Seljuq court. It was represented visually by all kinds of symbols. Already in the Sasanian period, in the Letter of Tansar, the necessity of maintaining differences between the common people and the nobility and of making the upper classes dependent on the patronage of the ruler is emphasised. Indeed, the sovereign is explicitly advised as follows: There is no wickedness or calamity, no unrest or plague in the world which corrupts so much as the ascending of the base to the station of the ­noble . . . ­You must make the heads of their first families and their men of rank and their lords and nobles rely on your position and patronage; and through favours and kindnesses you must banish the causes of vexation and care from their hearts.12

The anonymous author goes on to say that the King of Kings has established ‘a visual and general distinction between men of noble birth and common people with regard to horses and clothes, houses and gardens, women and servants.’ Indeed, the nobles are clearly differentiated from ordinary folk by their dress, horses, lofty dwellings, headgear, hunting and other characteristics related to their status.13

238  |  the me d ie va l tur k s The same kind of ‘pecking order’ must have prevailed in Seljuq times when it is likely that the nobles had their own prescribed positions, according to rank and office held, and protocol dictated where they should stand on official occasions. An Artistic Overview of a Typical Medieval Turkish Court A telling visual evocation of rank and its importance can be seen in a well-­ known masterpiece of Arab painting which evokes with remarkable clarity many aspects of the court life of the Turkish-­dominated Seljuq successor-­ states, and indeed it can be regarded as a reliable guide to actual Seljuq practice. Here we see the easy interplay between the domestic and the public, the formal and the informal, business activities and the life of leisure, which characterised these medieval Middle Eastern courts. The painting depicts this entire world in microcosm and it can therefore serve to frame this entire paper. And, as the adage has it, a picture is worth a thousand words. This particular picture, datable c. 648/1250 and probably hailing from northern Iraq, is a f­rontispiece – a­ kind of medieval equivalent to the modern dust jacket, with the same purpose of advertising the book i­tself – ­to a medical treatise dealing with antidotes to snakebite, a text based on the work of the Greek physician Galen (Figure 17.1).14 So here is a work of Greek science rendered into Arabic and given a full-­ page introductory illustration that has nothing whatever to do with snakes. But the respect for Greek learning is clear and serves as a reminder of the ruler’s commitment to intellectual pursuits. In fact the format of this image also has classical roots, for it ingeniously adapts the form of the late antique diptych, the wooden or ivory panel whose flip side was coated with wax on which the letter was incised. In other words, it was an endlessly re-­usable envelope. In late Roman and early Byzantine times such diptychs bore an image of the ruler. And so does t­his – ­imitation is the sincerest form of flattery. The difference is that the classical letter has become the Arabic book. But in place of the rigid formality of its classical source, here informality is the keynote. We see the ruler in relaxed mode. He wears a fur sharbush on his head and sits cross-­legged on an ample bolster with a decorative L-­shaped wooden backrest. No formal throne here. His elegant caftan, of Central Asian type with voluminous sleeves, seems to be trimmed with ermine and

aspe c ts of the cou rt of the gr e a t s e l j uq s  | 239

Figure 17.1  Frontispiece of the Kitab al-Diryaq of Pseudo-Galen, probably Mosul, c. 1250 (Vienna, National Library)

240  |  the me d ie va l tur k s falls open to reveal a scarlet under-­robe. He holds a glass beaker with a ruby-­ coloured liquid in it, presumably wine. A golden ewer on a shelf behind him, flanked by vases with flowers, promises easy refills. In front of it is an open dish piled high with fruit. Meanwhile a gorgeously clad, sharbush-crowned kneeling ­retainer – ­no kitchen boy, ­this – ­prepares two skewers of shish kebab on a fancy barbecue. Two bodyguards holding up tasselled swords keep watch. Like the other courtiers dancing attendance on the ruler, they perch uncomfortably on one leg. This central panel is set within an architectural framework that suggests an open-­plan two-­storey palace. The four side panels contain two men apiece. Several of them bear their symbols of o­ ffice – t­he falconer, who was responsible for the hunt, is at nearside top left, while at top right the men holding a duck and a goblet respectively were charged with the provision of food and drink. Meanwhile at lower right is the hajib in pink with his staff of office and the polo-­master in green with his polo stick. These are the khassakiyya or court officials in the immediate retinue of the sultan, each with his own designated responsibility. Those responsibilities often had an army dimension and were not just confined to court life. Thus the polo-­master oversaw the cavalry, and the man in charge of the court buttery was also the army quartermaster. Practically all the men, from the ruler downwards, wear their hair in long plaits. The oblong panels above and below also have their parts to play in this microcosm of court life. Above is a hunting scene complete with a saluki and an onager being brought down by archers mounted on richly caparisoned horses with docked tails. Below, four heavily veiled ladies, one with a child, and mounted on camels, are led by an attendant on foot towards a group of horsemen with polo sticks. Thus the composite image contains references to hunting, polo, the harem and the family as well as evoking the functions of the principal court officials and depicting the ruler feasting at his ease. All this is put across in condensed form but with a wealth of corroborative detail in furnishings, accessories and costume so that we are vouchsafed a remarkably lifelike snapshot of the Seljuq court at play. The Nawba There were many visible ways in which the Seljuq sultan demonstrated his sovereignty over all his subjects, including those who ruled on his behalf in

aspe c ts of the cou rt of the gr e a t s e l j uq s  | 241 a subordinate capacity in other parts of his realm. Acknowledgement of the supreme authority of the sultan was demanded especially from members of his own family. From the very beginning of Seljuq power in the Muslim world, although the two brothers Chaghri and Tughril agreed to divide and share the lands they conquered, there would be only one overarching sovereign. This precept was enshrined in the conferring of the title of ‘Sultan of East and West’ on Tughril by the caliph al-­Qa’im (no doubt under duress) in 447/1055.15 This title had never been used in an Islamic context before and it may well have had its roots in the nomadic concept of world dominion in the ­steppes – ­a symbol of Turkish sovereignty, extending from where the sun rose to where it set.16 Long before the coming of the Seljuqs the nawba ceremony was an integral part of court protocol at the caliphal court in Baghdad.17 Amongst the multiple meanings of the word nawba the following definitions of the word are relevant in the context of this discussion: nawba could mean ‘a drum struck at stated hours’, ‘drums beating at the gate of a great man at certain intervals’ or ‘a musical band playing at stated times before the palace of a king or prince’.18 The nawba took place at the hours of prayer but its use was extended to other celebratory occasions, such as a military victory.19 The visual evidence from thirteenth-­century painting (Figure 17.2) indicates that wind and stringed instruments were also played in the nawba concert.20 The nawba was a crucial symbol of Muslim sovereignty.21 According to Hilal al-­Sabi’, who devoted a short chapter to the nawba in his book on court protocol written just before the Seljuqs took power in Baghdad, the nawba was originally the prerogative of the caliph only. He implies, but does not state explicitly, that this would take place five times a day. Gradually, however, with the weakening of centralised caliphal power, the right to have the nawba was extended to other rulers. The number of times a day this was permitted depended on the recipient’s status. Hilal al-­Sabi’ explains how under the Buyids their rulers were allowed to beat the drum at the gate of their palace in Baghdad: ‘A tent was pitched there for drummers, who beat at the time of each of the three prayers.’22 It is noteworthy that the Buyids were only to be allowed three nawbas, namely at the time of the early morning prayer and the two evening prayers.

242  |  the me d ie va l tur k s

Figure 17.2 Al-Jazari nawba, Kitab fi Ma‘rifat al-Hiyal al-Handasiyya, Iraq, 1206 (Istanbul, Topkapı Sarayı Library)

aspe c ts of the cou rt of the gr e a t s e l j uq s  | 243 Hilal al-­Sabi’ remarks too that it is important that, if the nawba ceremony takes place in honour of rulers other than the caliph, it should do so at a suitable distance from the caliphal palace. Although he does not state this overtly, Hilal al-­Sabi’ implies that only the caliph was entitled to have five nawbas each day. So the sounding of the nawba a certain number of times a day was jealously guarded as one of the signs of Islamic sovereignty; the more the nawbas, the greater the prestige of the ruler. The Seljuq viziers saw to it that the sultans embraced preceding court ceremonials. So they included the nawba as one of the marks of their sovereignty and they would also grant such a privilege to certain rulers subordinate to them. But the number of nawbas awarded could vary. There are isolated references to the nawba in the Arabic and Persian sources that deal with the Seljuqs, but when such references do occur the chronicler sees no need to give details of what nawba means. For example, Nishapuri writes simply that Sultan Muhammad struck five drumbeats (panj nawbat) in Hamadan.23 Sibt b. al-­Jawzi, however, is more explicit, mentioning under the year 457/1064–5 that amongst the honours Alp Arslan conferred on al-­Fadluya Shabankara’i, a Kurdish chief who had seized power from the Buyids in Fars,24 was the prerogative that drums (al-tubul) should be beaten outside his door at the times of prayer.25 There is evidence in an intriguing but unexplained passage in Akhbar al-dawla al-saljuqiyya that the supreme Seljuq sultan used the nawba prerogative as a mechanism for disciplining his subsidiary adjunct rulers when they came to visit him, as they were obliged to do, in order to pay him allegiance, or when he came, in threatening mode, to see them. Sultan Mahmud, for example, was summoned by his uncle Sultan Sanjar and was instructed by Sanjar’s vizier on how he should enter into the presence of the Great Sultan. According to al-­Husayni,26 the vizier instructed Mahmud as follows: He should give up the sultanate formalities (rusum) of the red military band fanfare (al-nawba al-hamra’), he should dismount with two black and white fanfares (fi nawbatayn sawda’ wa-bayda’), he should renounce the beating of five fanfares and he should kiss the ground in front of Sanjar when he went in to see him.

244  |  the me d ie va l tur k s So Mahmud followed these instructions. It is clear from this narrative that Mahmud had arrogated to himself the full five nawbas and that Sanjar was most displeased with this. He alone amongst the Seljuq rulers wanted this privilege. What is unusual in this passage is the mention of coloured nawba­s – ­red, black and white. As already mentioned, five nawbas were the highest sign of kingship and here Sanjar is denying Mahmud the privilege of the red nawba. Is this a usage inherited from Turkic steppe tradition? The Russian scholar Kononov mentions that one of the ways of determining orientation amongst the Turks, as well as the Indians and the Chinese, was a ‘finely elaborated system wherein the cardinal points are designated by colours’.27 Unfortunately he does not discuss what he calls this ‘colour geo-­symbolism’ in any detail. However, other specialists, such as Sinor and Pullyblank,28 have argued that in nomadic Turkic societies orientation labels played an important role in public claims to universal world domination and in this context colours were important. Red was associated with the south, white with the west, black with the north and blue with the east.29 The supreme ruler was in the north, facing the south (the colour ‘red’). And these colours continued to be used long afterwards to denote divisions of territory amongst the steppe peoples. Hence too, such terms as Qara Deniz (Black Sea) for the north and Aq Deniz (White Sea) for the Mediterranean to the west. To return to the quotation in the chronicle of al-­Husayni, the mention of red, black and white fanfares may well be a reference to this ancient Turkic tribal colour symbolism. Yet it has to be admitted that the Arab world itself had made public use of colours to fulfil certain social and political agendas. De Slane comments that ‘even before the promulgation of Islamism, red or scarlet tents indicated that their possessors were princes’.30 But how would the colours red, black and white be demonstrated publicly in the nawba? Would the musicians wear clothes of special colours? Would they have carried flags of different colours? Or would the horses be decorated in special coloured trappings? Perhaps it is too fanciful to suggest that horses of various colours would be used but Figure 17.3 certainly shows horses coloured red and blue.

aspe c ts of the cou rt of the gr e a t s e l j uq s  | 245

Figure 17.3  Illustration from Varqa va-Ghulshah by ‘Ayyuqi, probably Konya, c. 1250 (Istanbul, Sarayı Library)

Men of Letters at the Seljuq Court For the ‘Abbasid caliphs in Baghdad, as Kennedy writes, ‘the patronage of scholars was part of the exercise of élite power and the caliphs led the way’.31 After the fragmentation of the Abbasid realms in the ninth and tenth centuries, the little courts of upstart military barons and provincial governors that sprang up in the eastern Islamic world followed ‘Abbasid models closely. For such rulers too it was an intrinsic sign of the prestige of their court to have an entourage of public intellectuals, even if the rulers themselves were unschooled. The Turkish ruler Mahmud of Ghazna collected intellectual elites from the areas he conquered and could boast of scholars of the calibre of the great polymath al-­Biruni at his court in Ghazna. The Seljuqs and their advisers wished to carry on that tradition. And indeed al-­Ghazali was the jewel in the crown of the intellectual élite that flocked to the entourage of the Seljuq sultan Malikshah and his vizier Nizam al-­Mulk. A key role in medieval Muslim courts was played by poets, some of whom fulfilled the role of nadim (the boon companion of the ruler), whose task it was to amuse and entertain his royal master.32 Malikshah, for example, had ‘Umar Khayyam as one of his nadims.33 Like the Samanids and Ghaznavids34 before them, the Seljuq sultans attracted a good number of poets, seeking

246  |  the me d ie va l tur k s their patronage and the contents of their purse.35 A veritable galaxy of poets flocked to the Seljuq court and their presence proved to be beneficial to the sultans who wished their renown to be spread across their realms through poetry declaimed at their courts. As Nizami-­yi ‘Arudi puts it: ‘A king cannot dispense with a good poet, who shall conduce to the immortality of his n ­ ame 36 . . . ­his name will endure forever by reason of the poet’s verse’. Particularly noteworthy was the poet Mu‘izzi who wrote panegyrical poems in Persian addressed both to Nizam al-­Mulk and to Malikshah.37 In one poem he called Malikshah ‘the Khusraw of the age’.38 His verses addressed to the sultan could be very grandiose, praising his military victories, or sometimes more lyrical. When speaking of a visit paid by Malikshah to Iraq in 479/1087, he declares: ‘It is the season of the ‘id, the banks of the Tigris are joyful with the scent of sweet basil and the brightness of the goblets of ruby wine.’39 But it was not just Persian poetry that was declaimed at the Seljuq court. Arab literary figures came too. Take the Arab poet al-­ Tughra’i 40 ­ olymath – h ­ e com(d.  514/1120–1), for example. He was obviously a p posed at least six works on alchemy and served as the court astrologer, as well as vizier to the Seljuq sultan Mas‘ud. The masterpiece of al-­Tughra’i, a poetic ode entitled Lamiyyat al-‘ajam (The verses rhyming in lam of the non-Arabs), bemoans the corrupt times in which he lived and berates the inferior brand of politicians who now rule in the corridors of power. The work aroused great admiration for its use of rare words and was a deliberate parallel to the more famous Lamiyyat al-‘arab (The verses rhyming in lam of the Arabs) of al-­Shanfara; indeed, all the lines of both works rhyme in the letter lam.41 Another Arab poet, nicknamed Haysa Baysa (d. 574/1179), because he once uttered the phrase fi haysa baysa (in dire straits), cut a dashing figure at the Seljuq court, always wearing Bedouin dress and speaking ‘the purest Arabic’.42 He also wrote ornate epistles so convoluted that the recipients could not understand them, as well as an ode dedicated to Sultan Mahmud in which every line rhymed in ‘d’.43 Perhaps not surprisingly, his collection of poetry is still unedited. Yet another poet, al-­Abiwardi (d. 507/1113), served one of the sons of the great Seljuq vizier, Nizam al-­Mulk. His poetry, both panegyrical and pastoral, found favour until he incurred the displeasure of the Seljuq sultan Muhammad who had him poisoned.44 Seljuq poets had to

aspe c ts of the cou rt of the gr e a t s e l j uq s  | 247 produce celebratory odes on demand. On one occasion a scribe composed verses to celebrate the arrival of the head of a defeated enemy at Sanjar’s court.45 Not all the poets who frequented the Seljuq court spent their energies on eulogising the sultans and their viziers. Following long-­established ‘Abbasid traditions, Ibn al-­Qattan (d.  498/1105)46 and Ibn al-­Habbariyya (d.  504/1110–11),47 and no doubt others besides, wielded their pens and tongues in satirical verses. In his obituary of Ibn al-­Qattan, Ibn Khallikan breaks away from his customarily bland laudatory comments to describe the poet in juicy terms as ‘excessively licentious and dissolute, full of humour and pleasantry, pertinacious in flattering and in satirizing the proud and haughty’.48 Ibn Khallikan goes on to say that Ibn al-­Qattan was dreaded for the ‘virulence of his tongue’ and that nobody, not even the caliph, was immune from his attacks. He was locked into poetic skirmishes, often acrimonious no doubt, with Haysa Baysa. As for Ibn al-­Habbariyya, one of the poets in the retinue of Nizam al-­Mulk, his verses are described by Ibn Khallikan as satirical, humorous and obscene. He was not popular at court and he had to put up with much malice from the pages and other members of the entourage of Nizam al-­Mulk. The rivalry between poets at court must have been disagreeably vicious at times. To gain the sultan’s favour was thrilling; to fall from grace was terrible. When Anvari, another Persian poet, was banished in disgrace from the Seljuq court, he wrote strong words condemning the sycophancy of its poets, speaking of poetry as ‘male menstruation’.49 The bureaucratic class were also distinguished men of letters; two famous examples were al-­ Bakharzi (d.  467/1075) and ‘Imad al-­ Din al-­ Isfahani (d. 597/1201). Al-­Bakharzi wrote in prose as well as verse; his most famous work was a piece which he wrote to comfort his patron, the Seljuq vizier al-­Kunduri, after his castration.50 Stretching credulity to its limits, he quips: ‘He is now increased in virility since the removal of his testicles.’ Here, the poet has sacrificed verisimilitude for the sake of a pun. As for ‘Imad al-­Din al-­Isfahani, he is best known for his historical writings once he had moved to Syria and entered the service of Saladin, but his extremely informative work, the Kharidat al-qasr, is a massive anthology of sixth/twelfth-­century poets who wrote in Arabic.51

248  |  the me d ie va l tur k s The Sultan at Play: ­Hunting – ­The Sport of Kings As Daryaee has mentioned recently, the hunt was a favourite pastime of the nobility in Sasanian Persia. He points out that ‘the hunt mirrored warfare in the off season, and symbolically signified the battle readiness of the warriors’.52 He draws attention to similar ideas in Europe amongst the Germanic tribes for whom hunting also became the accepted activity of the warrior aristocracy.53 So too with the incoming Seljuq chiefs. Malikshah is singled out in the sources for his love of the hunt; indeed, in his eulogy of Nizam al-­Mulk, Ibn Khallikan writes sweepingly that the vizier had all the power, ‘whilst the sultan (Malikshah) had nothing more to do than show himself on the throne and enjoy the pleasures of the chase’.54 In his obituary of Malikshah, Ibn Khallikan elaborates on this theme, describing the sultan’s fondness of the chase as excessive.55 He is said to have killed 10,000 game beasts in one day.56 A more realistic figure is given by Nishapuri who reports that one day Malikshah shot seventy gazelles; in each hunting place he made towers (manarha) of the hooves of gazelles and wild asses. Fear of offending God led him to give a Maghribi gold coin in alms for every animal that he had killed.57 The sultan would hunt in a party with his vizier and his military commanders and when visitors came to his court he would invite them to attend too. Hunting duties went alongside religious obligations; in one trip in 479/1086–7, as Ibn al-­Athir relates, Malikshah went hunting in the desert, visited the shrines of ‘Ali and al-­Husayn, and then went into the hinterland, where he hunted many gazelles and other animals. Thereafter, he constructed a beacon of horns between the two shrines at Najaf and Kufa.58 Indeed, according to Rashid al-­Din, ‘at Isfahan and in those districts, wherever a hunting-­ground is found, he (the sultan) had left traces, and they remained for a long time’.59 A later Seljuq sultan, Mahmud, had an especially acute obsession with hunting and it apparently led to his neglecting the affairs of the realm. He had 400 hunting dogs with gold collars and ‘he spent all his time on baits and traps’. He was devoted to pigeons, caged birds, sparrowhawks, dogs, cheetahs and all sorts of other hunting animals.60 Drums also accompanied hunting expeditions. A loud drum roll made sure that any birds roosting in the trees flew into the air so that the hunters could shoot them more easily. Usama b.

aspe c ts of the cou rt of the gr e a t s e l j uq s  | 249 Munqidh devotes much space to his hunting experience and provides valuable detail on how the chase was organised in Syria, Mesopotamia and Egypt in the sixth/twelfth century.61 Final Reflections In a discussion of three themes connected with the court of the Seljuq sultans this chapter has shown that these newcomers to Iran, no doubt under the skilful tutelage of their sophisticated Persian viziers, equipped themselves with one of the key status symbols of the Perso-­Islamic ­ruler – a­ n entourage of distinguished literati and scholars. This formal intellectual milieu was something new for the Seljuq tribal chiefs. Hunting, ‘the sport of kings’, on the other hand, came easily to these sultans. Here, however, the tough exigencies of survival on the steppes were exchanged for lavish court hunting parties in a style which harked right back to pre-­Islamic Iran. As for the traditional Muslim nawba ceremony, in Seljuq times it was given a Turkic gloss, as was shown in the above preliminary analysis of a short enigmatic passage in Akhbar al-dawla al-saljuqiyya. In this passage the colour ‘red’ is used to accompany the privilege of five nawba­s – ­the prerogative of the supreme ­sultan – ­whilst the right of subordinate rulers to two nawbas is called ‘black’ or ‘white’. This system of colours would seem to denote cardinal points. Here the Seljuq Turks have adopted the nawba ceremony and tweaked it to fit their own ancient and familiar colour geo-­symbolism. Thus the gradual process of assimilation of the nomadic Turks was now fully in train. Notes   1. Cf. E. Diez, The Ancient Worlds of Asia (London, 1961), 78–99; T. Daryaee, Sasanian Persia: The Rise and Fall of an Empire (London, 2009).  2. E. Chavannes, ‘Notes additionnelles sur les Tou-­Kiue (Turcs) occidentaux’, in E. Chavannes (ed.), Documents sur les Tou-Kiue (Turcs) Occidentaux (Paris, 1900).   3. Cf. H. Kennedy, The Court of the Caliphs: The Rise and Fall of Islam’s Greatest Dynasty (London, 2004).   4. C. E. Bosworth, The Ghaznavids: Their Empire in Afghanistan and Eastern Iran 994–1040 (Edinburgh, 1963).

250  |  the me d ie va l tur k s   5. I will be writing about other themes concerned with the Seljuq court in a later publication.   6. Cf., for example, E. Norbert, The Court Society, tr. E. Jephcott (Oxford, 1983).   7. M. G. A. Vale, The Princely Court: Medieval Courts and Culture in North-west Europe, 1270–1380 (Oxford, 2001).  8. This has been persuasively argued in D. Durand-­Guédy, Iranian Elites and Turkish Rulers: A History of Isfahan in the Saljuq Period (London, 2010).   9. C. Hillenbrand, ‘Some reflections on Seljuq historiography’, in A. Eastmond (ed.), Eastern Approaches to Byzantium (Aldershot, 2000), 73–88. 10. Cf., for example, Pseudo-­Ghazali (sixth/twelfth century), Nasihat al-muluk, tr. F. R. C. Bagley as Ghazali’s Book of Counsel for Kings (Oxford, 1971); Abu ‘Ali Hasan Tusi Nizam al-­Mulk (d. 485/1092), Siyar al-muluk, ed. H. Darke (Tehran, 1347/1928), and tr. H. Darke as The Book of Government or Rules for Kings (Henley-­on-­Thames, 1978); Anon. (sixth/twelfth century), The Sea of Precious Virtues (Bahr-i fava’id): A Medieval Mirror for Princes, tr. J. Scott Meisami (Salt Lake City, UT, 1991). 11. Abu Hamid Muhammad al-­Ghazali, Fada’il al-anam, tr. A. Qayyum as Letters of al-Ghazzali (Lahore, 1976). 12. Anon., The Letter of Tansar, tr. M. Boyce (Rome, 1968), 27–8. 13. Ibid., 44. 14. The illustration is the frontispiece of the Pseudo-­Galen, Book of Antidotes. A picture of it can be found in The Genius of Arab Civilization: Source of Renaissance, ed. J. R. Hayes (Oxford, 1978), 115. The contents of the painting are described by Holter who makes no comment on their significance; K. Holter, ‘Die Galen-­ Handschrift und die Maqamen des Hariri der Wiener Nationalbibliothek’, Jahrbuch der Kunsthistorischen Sammlungen in Wien N.F.XI (1937), 1–15 and Plate 1. Grabar provides a short but useful analysis of the picture in Hayes, Genius, 114. 15. Sadr al-­Din Abu l-­Hasan ‘Ali al-­Husayni (seventh/thirteenth century), Akhbar al-dawla al-saljuqiyya, ed. M. Iqbal (Lahore, 1933), 18. 16. The Turkic empire in the sixth century was organised on a bipartite, east–west principle; cf. P. B. Golden, ‘War and warfare in the pre-­Cinggisid western steppes of Eurasia’, in N. di Cosmo (ed.), Warfare in Inner Asian History (Leiden, 2002), 112. 17. Cf. Encyclopaedia of Islam, 2nd edn, s.v. ‘Marasim. 3. In Iran’, VI, 521–9 (A. K. S. Lambton); s.v. ‘Tabl-­khana’, X, 34–8 (H. G. Farmer). 18. F. Steingass, Persian-English Dictionary (London, 1892), 1431.

aspe c ts of the cou rt of the gr e a t s e l j uq s  | 251 19. Cf. B. Spuler, Iran in früh-islamischer Zeit. Politik, Kultur, Verwaltung und öffentliches Leben zwischen der arabischen und der seldschukischen Eroberung 633 bis 1055 (Wiesbaden, 1952), 349–50. 20. Figure 2.2 illustrates a nawba band. 21. Cf. M. Quatremère, Histoire des Sultans Mamlouks de l’Égypte I (Paris, 1837), 139, n. 18. 22. Hilal al-­Sabi’, Rusum dar al-khilafa, tr. E. A. Salem as The Rules and Regulations of the ‘Abbasid Court (Beirut, 1977), 115. 23. Zahir al-­ Din Nishapuri (d. mid-­ sixth/twelfth century), Saljuqnama, ed. A. H. Morton (Chippenham, 2004), 41. 24. C. E. Bosworth, The New Islamic Dynasties (Edinburgh, 1996), 154. 25. Abu’l-­Muzaffar Yusuf b. Kizoghlu Sibt b. al-­Jawzi (d. 654/1257), Mir’at alzaman fi ta’rikh al-a‘yan, ed. A. Sevim (Ankara, 1968), 121. 26. Al-­Husayni, Akhbar al-dawla al-saljuqiyya, 89. 27. A. N. Kononov, ‘Terminology of the definition of cardinal points at [sic] the Turkic peoples’, Acta Orientalia Academiae Scientiarum Hungaricae 31 (1977), 61–2. 28. D. Sinor, ‘Some components of the civilization of the Turks’, in G. Jarring and S. Rosén (eds), Altaistic Studies (Stockholm, 1985), 147; E. G. Pullyblank, ‘The Hsiung-­nu’, in H. R. Roemer (ed.), History of the Turkic Peoples in the Pre-Islamic Period (Berlin, 2000), 52–75. 29. E. G. Pulleyblank, ‘The nomads in China and Central Asia in the post-­Han period’, in Roemer (ed.), History, 92. 30. Ahmad b. Muhammad Ibn Khallikan (d. 681/1282), Wafayat al-a‘yan wa anba’ ibna’ al-zaman, tr. Baron W. M. de Slane as Kitab Wafayat al-A‘yan: Ibn Khallikan’s Biographical Dictionary (Paris, 1843–71), I, 641, n. 2. 31. H. Kennedy, The Court of the Caliphs (London, 2004), 260. 32. Cf. J. Scott Meisami, Medieval Court Poetry (Princeton, NJ, 1987), 3–40. 33. G. E. Tetley, The Ghaznavid and Seljuk Turks: Poetry as a Source for Iranian History (London, 2009), 13. 34. Mahmud of Ghazna was said to have had four hundred poets in regular attendance at his court; Bosworth, Ghaznavids, 131. 35. Cf. ‘Ali Jawad al-­Tahir, Al-shi‘r al-‘arabi fi’l-‘Iraq wa balad al-‘ajam fi’l-‘asr al-saljuqi (Baghdad, 1958); R. A. Nicholson, A Literary History of the Arabs (Cambridge, 1969), 326, 329–36; Meisami, Medieval Persian Court Poetry; Tetley, Ghaznavid and Seljuk Turks. 36. Ahmad b. ‘Umar Nizami-­ yi ‘Arudi (second half of sixth/twelfth century),

252  |  the me d ie va l tur k s Chahar Maqala, tr. E. G. Browne as The Four Discourses (Cambridge, 1921), 45. 37. Tetley, Ghaznavid and Seljuk Turks, ch. 4. 38. Ibid., 96. 39. Ibid., 121. 40. Mu’ayyid al-­Din al-­Tughra’i (d. 514/1120–1), Diwan, eds ‘A. J. al-­Tahir and Y. al-­Juburi (Baghdad, 1976). 41. Al-­Fath b. ‘Ali al-­Bundari (d. seventh/thirteenth century), Zubdat al-nusra wa-nukhbat al-‘usra, in Recueil de textes relatifs à l’histoire des Seldjoucides, ed. M. T. Houtsma (Leiden, 1889), 110, 116, 132–3; cf. also ‘A. J. al-­Tahir, Lamiyyat al-Tughra’i (Baghdad, 1962). 42. Ibn Khallikan, Wafayat, tr. de Slane, I, 559; III, 337, 584–6; IV, 119–21; ‘Imad al-­Din al-­Isfahani, al-­Katib (d.  597/1200), Kharidat al-qasr, ed. M. B. al-­Athari (Baghdad, 1980), 202–366. 43. Ibn Khallikan, Wafayat, tr. de Slane, III, p. 337. 44. Ibid., III, 144–8. 45. Rashid al-­Din, Jami‘ al-tawarikh, partial ed. A. Ateş as Cami‘ al-tavarih (Metin), II. Cild, 5. Cuz, Selçuklar tarihi (Ankara, 1960); tr. K. A. Luther as The History of the Seljuq Turks from the Jami‘ al-tawarikh: An Ilkhanid Adaptation of the Saljuq-nama of Zahir al-Din Nishapuri (London, 2001), 86. 46. Ibn Khallikan, Wafayat, tr. de Slane, III, 583–9. 47. Al-­Bundari, Zubdat, 64–5, 103–5; Sibt b. al-­Jawzi, Mir’at, VIII/I, 58–62; Ibn Khallikan, Wafayat, tr. de Slane, III, 150–3. 48. Ibn Khallikan, Wafayat, tr. de Slane, III, 583. 49. B. Lewis, Music of a Distant Drum (Princeton, NJ, 2001), 16. 50. Ibn Khallikan, Wafayat, tr. de Slane, III, 294. 51. ‘Imad al-­Din al-­Isfahani, Kharidat al-qasr; this volume of the anthology deals with the poets of Iraq. 52. Daryaee, Sasanian Persia, 51–2. 53. For a fascinating and detailed account of hunting in seventh–­thirteenth-­century North Africa, cf. Al-Mansur’s Book on Hunting, trs Sir Terence Clark and M. Derhalli (Warminster, 2001). 54. Ibn Khallikan, Wafayat, tr. de Slane, I, 413. 55. Ibid., III, 441. 56. Ibid., III, 441. 57. Nishapuri, Saljuqnama, 30. 58. ‘Izz al-­Din Abu’l-­Hasan ‘Ali Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, ed. C. J. Tornberg

aspe c ts of the cou rt of the gr e a t s e l j uq s  | 253 (Leiden and Uppsala, 1851–76), partial tr. D. S. Richards as The Annals of the Seljuk Turks: Selections from al-Kamil fi’l-Ta’rikh of ‘Izz al-Din Ibn al-Athir (London, 2002), 228; cf. also Abu l-­Faraj ‘Abd al-­Rahman b. ‘Ali Ibn al-­Jawzi, Al-Muntazam fi ta’rikh al-muluk wa-l-umam (Hyderabad, 1357–8/1938–40), IX, 35, 70; Sibt b. al-­Jawzi, Mir’at, 243. 59. Rashid al-­Din, Jami‘, tr. Luther, 61. 60. Ibid., 99. 61. Usama b. Munqidh (d. 584/1188), Kitab al-i‘tibar, tr. P. K. Hitti as Memoirs of an Arab-Syrian Gentleman (Beirut reprint, 1964), 222–54.

18 Nizam al-­Mulk: A Maverick Vizier?

i. Introduction

N

izam al-­Mulk was a numinous figure whose memory as the ‘servant of the state’ has been cherished, despite his obvious flaws, by many generations of Iranians. But his career fits to a nicety Lord Acton’s axiom, expressed in 1887, that: ‘Power tends to corrupt, and absolute power corrupts absolutely. Great men are almost always bad.’ Nizam al-­Mulk demonstrated that axiom ever more clearly in the course of his astonishingly prolonged grip on power. In his time the vizierate was what you made of it, and just as the caliph in some essential senses was both emperor and pope, so too Nizam al-­Mulk was both head of the bureaucracy and ‘prime minister’, and on occasion also a university chancellor, head of the diplomatic service, and a military chief of staff. It must have seemed at the time that there was no nook or cranny of public life into which he did not stick his long nose, o­ r – ­to use the Persian e­ quivalent – ­he was diraz-dast (‘long-­handed’). In addition to these diverse and wide-­ ranging responsibilities he needed to have eyes in the back of his head to guard himself against constant plots and personal danger. The chronic uncertainty that this situation engendered helps to explain a policy that later generations as well as his contemporaries condemned, namely the way in which he stuffed public posts with members of his own family. Indeed, at one stage twelve of his sons held high government positions simultaneously.1 One would not be far wrong in regarding him as a capo mafioso well ahead of his time. But the remarkable thing is that no short description can do justice to so protean and complex a figure. Above all, he had the ­difficult – ­and at times it 254

ni z a m a l-m ul k : a m a ve ri ck viz ie r ?  | 255 must have seemed ­impossible – ­task of serving as the intimate adviser and first port of call to rulers who were powerful and wilful personalities and whose upbringing and mindset could scarcely have been more different from his own. They were Turks, not Persians, and belonged to a nomadic society, not a settled one. Their norms and forms were not his, and it required constant vigilance on his part to handle the difficulties that arose as a natural result of these perspectives. What manner of man was this, who inspired alongside the hatred of his rivals such universal respect? There can be no doubt that he himself saw his career and his responsibilities not simply in the context of his own troubled times but also from the wider perspective of Iranian history, and in particular that of the Sasanian empire. The accounts of the sayings of Buzurjmihr scattered throughout his Mirror for Princes work, the Siyasatnama, indicate clearly enough whom Nizam al-­Mulk saw as a model, as the following pronouncement makes clear: ‘The kings who have become great rulers and whose names will be blessed until the resurrection, have all been those who had good vazirs.’2 What was a good vizier? Nizam al-­Mulk himself outlines a suitable job description: ‘A vazir needs to belong to a pure religion, to have sound faith and to adhere to a good school of ­doctrine – ­either Hanafi or Shafi‘­i – ­and he must be efficient, shrewd, a fluent writer, and loyal to his king.’3 Lambton argues for a clear distinction between the ‘men of the pen’, to which group viziers, usually Persian, belonged, and the ‘men of the sword’ who in the Seljuq period were mostly of Turkish extraction.4 She points out, moreover, that the most important function of the vizier was the supervising of state finances. Given his religious education, his administrative skills  and his wide experience, Nizam al-­Mulk certainly conforms to the usual profile of a ‘man of the pen’ trained in Perso-­Islamic learning and statecraft. However, it has long been known that certain viziers exceeded their brief as dutiful bureaucrats and on occasion took over the whole running of the state in times when a caliph or sultan was a minor, or an adult who could easily be dominated. In his own career Nizam al-­Mulk went a good deal further than the idealised blueprint for a vizier that he advocates in the Siyasatnama.

256  |  the me d ie va l tur k s This paper can do no more than tilt at some of the ways in which the definition of the vizierate was stretched and tweaked by Nizam al-­Mulk. But the discussion below will aim to highlight in particular a lesser-­known facet of his c­ areer – ­his personal participation in military engagements. First, a few words to sketch the background. Nizam al-­Mulk is usually described as the vizier of the second and third Seljuq sultans, Alp Arslan and Malikshah, and he filled that office so well that he governed the Seljuq state for thirty years. But, cast in the mould of earlier Iranian bureaucrats such as al-­Bal‘ami,5 his task was rendered much harder than that of his predecessors by a new and most volatile ingredient: the nomadic Turks. Even allowing for negative, anguished exaggeration in Christian accounts6 and for laudatory, ‘whitewashing’ comments in Muslim ones, the impact of these nomads under their Seljuq chieftains must have been violent and disruptive.7 They entered the Muslim world, as the dynastic historian of the Seljuqs, al-­Husayni describes them: completely attached to warfare and its p ­ ractice . . . ­the Seljuqs wander about like desperadoes and ­outcasts . . . ­careless whether they suffer destruction and p ­ erish . . . ­They have horsemen who boldly face death. They are hardly to be considered as human beings.8

The problems posed by these unassimilated nomadic Turks were exacerbated by the peripatetic lifestyle that they continued to follow once they had entered the eastern Islamic world. Alp Arslan regularly criss-­crossed his vast domains, which stretched from Aleppo to Marv. Such traversing of tracts of territory between winter and summer pastures was a long-­established custom for the Seljuq Turks, as for other nomads, but, for the Iranian bureaucrat and scholar that Nizam al-­Mulk was, to accompany the sovereign such protracted journeys must have required enormous reserves of stamina. Yet he had to go with the sultans in order to keep control of what was going on. His strategic skills had to be exercised on the move. Nizam al-­Mulk had to contend with many conflicting groups and their relationship both with each other and with ­him – ­the relationships between the Seljuq family with their nomadic followers, between individual members of the Seljuq family, between the Seljuqs and the cities they governed, between the Seljuq sultan and the caliph, and between all these groups and

ni z a m a l-m ul k : a m a ve ri ck viz ie r ?  | 257 himself. This called for him to perform a diabolically difficult balancing act. ii. Nizam al-­Mulk – ­The Fighting Vizier In view of his prestigious reputation as an administrator, religious scholar of Shafi‘i fiqh, founder of madrasas and writer of a famous Mirror for Princes, it is perhaps surprising to think of Nizam al-­Mulk donning armour and leading troops into battle, but that is what he is recorded as doing at certain times. Indeed, it is mentioned specifically in a number of primary sources that during his period in power Nizam al-­Mulk actually fought in battles and laid sieges, either in the company of the sultan, or sometimes even on his own. Especially in the reign of Alp Arslan (ruled 455/1063–465/1073) when the sultan was moving so energetically across the length and breadth of Iran, consolidating his power in the border areas of his empire, Nizam al-­Mulk is often mentioned in the primary sources as fighting in person. This military activity is first mentioned for the year 456/1064 when, according to al-­Husayni, he donned armour (labisa al-silah) and lined up (cavalry) detachments (‘abba al-kata’ib) during the power struggle between Alp Arslan and his relative, Qutlumush, the cousin of Sultan Tughril.9 That same year, in the company of Alp Arslan’s young son, Malikshah, who was the heir to the throne, Nizam al-­Mulk performed feats of military distinction at a frontier town called Maryam-­nishin,10 preparing ships and river boats and labouring night and day until the town was taken,11 helped, so al-­Husayni relates, by the eruption of an earthquake during the night.12 The story of this episode as presented by al-­Husayni has clearly been embellished in order to laud the achievements of the young Seljuq prince Malikshah himself. It is unlikely that, aged only eleven at the time, he could have achieved the military feats attributed to him or even been allowed to attempt them.13 It is much more likely is that Nizam al-­Mulk was fully in control. In 459/1067 Nizam al-­Mulk ‘kindled the fires of war’ with his own army against Fadlun, the lord of Ganja.14 He is then recorded as being in Fars, conquering citadels one by one, ‘now by negotiation and now by fighting’.15 Ibn al-­Athir describes this achievement as enhancing the already high reputation that Nizam al-­Mulk enjoyed with Alp Arslan: ‘He (Alp Arslan) exalted him even higher and added to his authority.’16

258  |  the me d ie va l tur k s It is thus clear that Nizam al-­Mulk was given military responsibilities, such as recruiting armies and besieging fortresses, and even that he may indeed have assumed such responsibilities on his own initiative. In his obituary notice of Nizam al-­Mulk, al-­Subki states grandly that Nizam al-­Mulk participated in military operations in Rum, Aleppo, Khurasan and Transoxiana.17 Given this succession of military activities on the part of Nizam al-­Mulk, it is surprising that this aspect of his career as vizier has not received much attention from modern scholars of the Seljuqs. Of course, there has been some mention of his fighting.18 Klausner rightly points out that Nizam al-­ Mulk was militarily active especially in the early stages of expansion in the time of Alp Arslan, when raids were conducted on several fronts at the same time.19 Only Safi sees the importance of Nizam al-­Mulk having had military experience in helping him deal with the Turkish military commanders.20 But a number of other points should be made in this context. Why should an undoubtedly well-­educated, cultivated ‘man of the pen’ become involved in actual warfare itself? Why did Alp Arslan not delegate the task of leading the Seljuq forces to other Turkish military commanders rather than to his Persian civilian vizier? Clearly Nizam al-­Mulk was still young enough to take active command; he was forty-­six years old when he is first recorded as fighting21 and this active participation continued for at least a further eight years. It is probable that in the early days of Alp Arslan’s sultanate when his family rivals needed to be dealt with and border territories had to be secured, Nizam al-­Mulk, with his fabled sagacity and undoubted ruthlessness,22 felt it essential to be at the sultan’s side as much as possible. His going on campaign with the sultan in the latter’s early days in power must have brought the two men very close together, and Nizam al-­Mulk would thus have learned a great deal first-­hand about the milieu in which the Seljuqs lived and how they thought. Fighting alongside Alp Arslan was a sure method of showing solidarity with him and gaining his respect. Being so frequently with him would have enabled Nizam al-­Mulk to teach him how to conduct himself in court ceremonies and religious gatherings. Quite conceivably, Nizam al-­Mulk used this opportunity to learn Turkish. Later in the time of Malikshah he maintained this same close proximity to the sultan, especially when he began to fall from favour. Nizam al-­Mulk

ni z a m a l-m ul k : a m a ve ri ck viz ie r ?  | 259 obviously followed a strategy of keeping his friends close but his enemies closer. He even accompanied Malikshah when he indulged in his favourite pastime of hunting. According to Ibn al-­Jawzi, in 479/1086–7, Malikshah went hunting with Nizam al-­Mulk in the desert between Najaf and Kufa. They hunted gazelles and other animals and Malikshah built a tower of horns from the animals he had killed.23 One cannot escape the conclusion that Nizam al-­Mulk was keeping a watch on Malikshah at all times. How else can one explain why an elderly statesman with cultural ­tastes – N ­ izam al-­Mulk was by then sixty-­nine – ­would go hunting with Malikshah? Nizam al-­Mulk is also recorded as having his own private army, numbering thousands of mamluks.24 Their devotion to him seems undoubted, given their violent reaction to his death and the reprisals they soon took against his rival, Taj al-­Mulk, whom they obviously blamed for his death, tearing his body to pieces in a frenzied attack. That Nizam al-­Mulk possessed his own private army is scarcely surprising. In view of the largely peripatetic lifestyle that he perforce followed in the reign of Alp Arslan and in the early years of Malikshah’s rule, Nizam al-­Mulk often visited border areas where he could have been vulnerable to a fatal attack if he displeased the sultan; indeed it is worth remembering that he was murdered in a remote place whilst on a journey with Malikshah to Baghdad in 485/1092. The control wielded by Nizam al-­Mulk as vizier over the diwans and, above all, the state finances, could easily provoke the envy of the ruler, and the sources also speak of the vast personal fortune of Nizam al-­Mulk. Moreover, political rivals in Alp Arslan’s and Malikshah’s entourage could and often did create serious difficulties for Nizam al-­Mulk. It was dangerous to be a vizier. But Nizam al-­Mulk knew how to defend himself. So a large bodyguard, it may be assumed, was constantly watching over his personal safety, and that, in view of his unusually long period in power, these mamluks must have performed their task extremely well. This was proved by the fact that he is recorded as having survived two attempts on his life in the time when his relationship with Malikshah was deteriorating. It may also be assumed that Nizam al-­Mulk operated some form of surveillance system (barid), as he advocated in his advice to the sultan in the Siyasatnama.25

260  |  the me d ie va l tur k s iii. Other Fighting Viziers in the Seljuq Period Nizam al-­Mulk was not the only Seljuq vizier who entered the military arena. In Fairbanks’ analysis of the Ta’rikh al-vuzara’ of the late sixth/twelfth century Persian writer Najm al-­Din Qummi, he includes a very interesting chapter entitled ‘The Military and Other Roles of the Vizier’,26 in which he mentions some examples of fighting viziers. According to Qummi, the Seljuq sultan Mas‘ud (ruled 529/1134–547/1152) singled out an official called Kamal al-­Din Muhammad Khazin especially for his military abilities: ‘Kamal al-­Din Muhammad struck Sultan Mas‘ud as something astonishing. He was a man of the sword, not of the pen.’27 Mas‘ud later conferred the vizierate on Kamal al-­Din who clearly knew, at least for a while, how to govern the state: ‘He laid down law and order as no other vizier had done. Every day he was busy making light of the amirs who commanded the armies. He threw vinegar in the noses of all of them’.28 In the end, however, such a show of strength proved inadequate against a coalition of amirs who demanded the execution of Kamal al-­Din, and Mas‘ud was forced to hand him over.29 Further west, in thirteenth-­century Anatolia, another Seljuq fighting vizier comes to mind whose conduct in power resembles that of Nizam al-­ Mulk – ­Mu‘in al-­Din Parwana (d. 676/127).30 He too was not a professional military man. Like Nizam al-­Mulk, he belonged firmly within the Persian bureaucratic tradition, although he was at times drawn into military engagements. Like Nizam al-­Mulk in the Great Seljuq state, the Parwana became the real ruler of the Rum Seljuq state in Anatolia and stayed in power for more than twenty years. His political longevity argues strongly that he possessed great political and survival skills. He too, like Nizam al-­Mulk, came to a terrible end, indeed an even more ignominious one, being killed and then allegedly eaten by the Mongols. iv. Badr al-­Jamali – ­A Fighting Vizier in Fatimid Egypt It is interesting to compare Nizam al-­Mulk with his almost exact contemporary and counterpart in Fatimid Egypt, Badr al-­Jamali. Bianquis writes about the installation of what he calls the ‘military vizierate’ in Fatimid Egypt in

ni z a m a l-m ul k : a m a ve ri ck viz ie r ?  | 261 the reign of al-­Mustansir. This caliph (ruled 427/1036–487/1094) invited the Armenian general, Badr al-­Jamali, to come to Cairo in 464/1072–3 and, on his arrival, he granted him full civil and military powers. These Badr al-­Jamali, enjoyed right up until his death in 487/1094.31 The ways in which Nizam al-­Mulk exercised supreme power in the Seljuq state, breaking through the boundaries described in the blueprint for medieval Muslim government that the Mirrors for Princes literature presents, can usefully be compared to the conduct of Badr al-­Jamali who ruled the Fatimid state in a way similar to that of Nizam al-­Mulk. It is small wonder that the Fatimid caliph, al-­Mustansir, had called on the services of a military man in 466/1074; there had, after all, been forty-­ nine viziers since his accession in 427/1036. Like Nizam al-­Mulk, Badr al-­Jamali lived to a ripe old age (around eighty); he had ruled for twenty-­one years. He died two years after Nizam al-­Mulk. Just as Nizam al-­Mulk is clearly portrayed as being in charge of the Seljuq state in its entirety, so too is Badr al-­Jamali in the Fatimid context. Ibn Muyassar calls Badr al-­Jamali a ‘wazir of the sword’ and writes of him as follows: He had ruled as kings do in Egypt and, with him in charge, no power remained for al-­Mustansir who handed over affairs to him.32

The same source mentions how Badr al-­Jamali killed many people and made the state flourish.33 So both these men, one on Iran and the other in Egypt, transcended the limits normally laid down in medieval Muslim bureaucratic manuals and Mirrors for Princes literature for the vizierate. But their attitudes and approach must have been very different, for Badr al-­Jamali was a military man who became a vizier, and Nizam al-­Mulk was a vizier who, for a while at least, became a military man. v. Why was Nizam al-­Mulk Called Atabeg? The significance of the title atabeg in the career of Nizam al-­Mulk deserves analysis at this point, since it clearly sheds light on the success that Nizam al-­Mulk had achieved in reaching accommodation and a modus operandi with the alien Turkish newcomers.

262  |  the me d ie va l tur k s The Turkish term atabeg does not seem to have been used in the medieval Muslim sources before the coming of the nomadic Turks. What it meant for them in the days preceding their entry into the Muslim world remains unclear. According to Peter Golden:34 The atabeg institution, at least with that title, is not mentioned in Old-­ Middle Turkic materials until into the Seljuq era.

Golden argues most persuasively that if such an office had existed in Inner Asia in pre-­Seljuq times among the Türk Ashina royal clan or their Uyghur successors, the Tang court historians would have mentioned it. But the Russian historian Agadzhanov writes that the Seljuqs did retain some steppe traditions from the Oghuz tribes in Jangikent whose leaders were called yabghus.35 He especially mentions the role of atabegs who acted as the guardians, mentors and educators of their under-­age princes. And Golden also concludes that he would ‘not totally exclude the possibility’ of the existence of the atabeg institution, ‘at least among the Oghuz’.36 So when was the first time that the term atabeg was used in a Muslim context? Although it is usually thought that Nizam al-­Mulk was the first person to be described in this way in the Arabo-­Persian world,37 there is an earlier reference by al-­Husayni to an official in the entourage of Sultan Tughril’s brother, Chagri. This official was called Atabeg Amir Qutb al-­Din Kül-­sarigh;38 he was captured whilst fighting against the Ghaznavids.39 Even before Chagri’s death in 451/1059, Nizam al-­Mulk had been appointed as Alp Arslan’s vizier. Indeed, Chaghri had been so impressed by the abilities of Nizam al-­Mulk that he had told his son Alp Arslan to view this gifted Persian administrator as his father (surely an echo of the idea enshrined in the word atabeg). According to Ibn Khallikan, Da’ud (that is, Chagri Beg): ‘received such proofs of fidelity and attachment, that he gave him (Nizam al-­Mulk) over to his son, Alp Arslan, saying: “Consider him as a parent, and disobey not his counsels”.’40 Here it appears that the role of Nizam al-­Mulk was envisaged as that of a fatherly adviser to Alp Arslan, although the term atabeg is not used. By the time of Malikshah’s accession in 465/1073, however, the term atabeg is specifically mentioned in connection with Nizam al-­Mulk. Ibn al-­Athir records that Malikshah said to Nizam al-­Mulk:

ni z a m a l-m ul k : a m a ve ri ck viz ie r ?  | 263 I hand over to you all matters, great and small, for you are the father (anta al-walid) . . . He (Malikshah) gave him (Nizam al-­Mulk) titles amongst which was atabeg the meaning of which is ­father – ­commander.41

How can the term atabeg be viewed in this context? Ibn Khallikan attempts a definition as follows: This word means ‘a bringer-­up of princes’: ata in Turkish signifying ‘father’ and bek ‘emir’.42

Describing the role of an atabeg from the period of Alp Arslan at least, Cahen writes that this term denotes: a military chief in the ruler’s entourage, to whom the ruler entrusts the education and care of his son, ultimately marrying him to the pupil’s mother.43

Whilst this definition largely applies to later holders of the atabeg title in the Seljuq and Ayyubid periods,44 it certainly does not fit Nizam al-­Mulk, who could not be regarded primarily as a military chief and who did not marry Malikshah’s mother. Nevertheless, Nizam al-­Mulk, despite being a Persian civilian, seems to have accepted being designated as Malikshah’s atabeg. The timing of the appointment of Nizam al-­Mulk as Malikshah’s atabeg is surely ­significant – ­for it occurred just after he had performed admirably well in the military arena. The long years of travelling around the empire with Alp Arslan and fighting, raiding, laying sieges and commissioning troops, either with the sultan or in the company of the young Malikshah, had earned Nizam al-­Mulk what Safi rightly calls ‘the highest military honour the Turkish warlords had to offer’, that is, the title atabeg.45 In more popular parlance, Nizam al-­Mulk was being made ‘an honorary Turk’. From his point of view this was a triumph, validating his right to control the young sultan and indeed to take over and centralise the organisation of a new polity, the Seljuq state, based on a blend of Persian statecraft and Turkish military strength. A number of obscurities and misconceptions surround the ‘educational’ role of an atabeg. It is likely that the duties of such an office would have involved teaching the prince equestrian skills, bowmanship and other martial arts. Ironically, Nizam al-­Mulk was the only atabeg who had the q­ualifications

264  |  the me d ie va l tur k s necessary to educate a prince in the religious sciences and Sasanian statecraft rather than the arts of war. Whether he did so or not is unclear. After Malikshah’s death the role of atabeg was henceforth given to Turkish military chiefs alone. It became a force for fragmentation, creating centrifugal tendencies within the Seljuq state, especially during periods when central power was weak. Its educational aspects seem to have been more honoured in the breach than the observance. Instead, it seems to have been primarily an instrument of control and for the acquisition of power. vi. Concluding Comments This paper has focused on Nizam al-­Mulk during the formative early years of his vizierate, in the period of consolidation of Seljuq power under Alp Arslan and to a lesser extent under Malikshah. This was a period when Nizam al-­Mulk was already proving his mettle as an extraordinarily clever, flexible and versatile leader, ready to perform a military role as well as a civilian one. He gained the confidence and respect of the recently arrived Turkish leaders by spending extended time in their company in what must have been most arduous conditions and exhausting journeys. Nizam al-­Mulk’s truly awesome achievement in staying in power for over thirty years in a permanently unstable, unpredictable, violent and indeed murderous milieu benefited from the early experiences discussed in this chapter. He succeeded in holding together the Seljuq empire by a combination of subtlety and ruthlessness, by extraordinary stamina, by personal courage and a truly formidable intellect. As Ibn al-‘Adim said: In reality, the state came under his control (Nizam al-­Mulk’s) and was the sultan’s (only) in ceremonial and name.46

Nizam al-­Mulk was attempting an almost impossible task: to achieve balance and harmony between various perennially conflicting forces. These included the Turcomans and the sultans, the sultans and their relatives, the Seljuqs and their amirs, the cities and the nomads, and the Persian bureaucrats and the populations they ruled, from Jerusalem to Samarqand. Above all, he aimed to maintain stability in an empire which he almost single-­handedly built up to cover the land mass ruled by the ancient Sasanian empire. Whilst he was

ni z a m a l-m ul k : a m a ve ri ck viz ie r ?  | 265 alive, the unity of the Seljuq state remained intact. But once he was dead, nothing worked as well again. Notes  1. Nishapuri, Saljuqnama, ed. A. H. Morton (Chippenham, 2004), 31; Rawandi, Rahat al-sudur wa ayat al-surur, ed. M. Iqbal (London, 1921), 116.   2. Nizam al-­Mulk, Siyasatnama, tr. H. Darke as The Book of Government or Rules for Kings (Henley-­on Thames, 1978), 173.   3. Ibid., 173.  4. Encyclopaedia of Islam, 2nd edn, art: wazir (A. K. S. Lambton).  5. Cf. A. Peacock, Medieval Islamic Historiography and Political Legitimacy: Bal’ami’s Tarikhnama (London, 2007).   6. Cf. for example, Aristakes of Lastiverd, Aristakès de Lastivert. Récit des malheurs de la nation arménienne, trs M. Canard and H. Berbérian (Brussels, 1973), 75–87; Matthew of Edessa, Patmut’iwn, tr. E. Dulaurier as Chronique de Matthieu d’Edesse (962–1136) avec la continuation de Grégoire le Prêtre jusqu’en 1162 (Paris, 1858).   7. Cf. Ibn Hassul, Kitab tafdil al-atrak ‘ala sa’ir al-ajnad, ed. A. ‘Azzawi (Cairo, 1940), 40–1; Sharaf al-­ Zaman Tahir Marvazi, Taba’i‘ al-hayawan, tr. V. Minorsky (London, 1942).   8. Al-­Husayni, Akhbar al-dawlat al-saljuqiyya, ed. M. Iqbal, reprint (Beirut, 1984), 8: tr. C. E. Bosworth, as The History of the Seljuq State: A Translation with Commentary of Akhbar al-dawlat al-saljuqiyya (Abingdon, 2011), 13; cf. Q. Ayaz, An Unexploited Source for the History of the Saljuqs: A Translation and Critical Commentary on the Akhbar al-dawlat al-saljuqiyya, unpublished PhD thesis (University of Edinburgh, 1985).   9. Al-­Husayni, Akhbar, 31; tr. Bosworth, 27. 10. The location of this town is not known but its name is ‘probably a corruption of Marmarashen, a famous monastery in the district of Shirak (between the Araxes and Kur valleys)’; cf. tr. Bosworth, n. 135. 11. Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, tr. D. S. Richards as The Annals of the Seljuk Turks (London, 2002), 153. 12. Al-­Husayni, Akhbar, 36; tr. Bosworth, 30. 13. Malikshah is described as securing a rope to the parapet of the walls and then climbing up. Not the wisest move for an eleven-­year-­old heir apparent. 14. Al-­Husayni, Akhbar, 42–3; tr. Bosworth, 34; cf. also G. Le Strange, Description of the Province of Fars in Persia at the Beginning of the Fourteenth Century a.d. from the MS. of Ibn al-Balkhi in the British Museum (London, 1912), 10.

266  |  the me d ie va l tur k s 15. Sibt b. al-­Jawzi, Mir’at al-zaman fi ta’rikh al-a‘yan, ed. A. Sevim (Ankara, 1968), 137; Ibn al-­Athir, Al-Kamil, tr. Richards, 160. 16. Ibn al-­Athir, ibid. 17. Al-­Subki, Tabaqat al-Shafi’iyya al-kubra, eds M. Tanahi and A. al-­Hilw (Cairo, 1964), IV, 313. 18. C. L. Klausner, The Seljuk Vezirate. A Study of Civil Administration 1055–1194 (Cambridge, MA, 1973), 87–8; S. Fairbanks, The Ta’rikh al-vuzara’: A History of the Saljuq Bureaucracy, unpublished PhD thesis (University of Michigan, 1977), 151; O. Safi, The Politics of Knowledge in Pre-modern Islam: Negotiating Ideology and Religious Inquiry (Chapel Hill, NC, 2006), 59–60; A. K. S. Lambton, ‘The internal structure of the Saljuq Empire’, in J. A. Boyle (ed.), The Cambridge History of Iran, V: The Saljuq and Mongol Periods (Cambridge, 1968), 261. 19. Klausner, Vizierate, 40. 20. Safi, Politics, 60. 21. The year 456/1064. 22. The grisly death of al-­Kunduri, the predecessor of Nizam al-­Mulk, in the vizierate, is blamed squarely on the latter: cf. Ibn Khallikan, Wafayat al-a‘yan wa anba’ ibna’ al-zaman, tr. Baron M. de Slane as Kitab Wafayat al-a‘yan: Ibn Khallikan’s Biographical Dictionary (Paris, 1843–71), III, 294–5. 23. Ibn al-­Jawzi, Al-Muntazam fi ta’rikh al-muluk wa-l-umam, ed. F. Krenkow (Hyderabad, 1357–8/1938–40), IX, 35, 70. 24. The figures vary widely between 20,000 and 80,000 mamluks; cf. Ibn al-­Jawzi, Muntazam, 9, 66; al-­Husayni, Akhbar, 145. Ibn al-‘Adim claims more modestly that Nizam al-­Mulk had 7,000 male and 7,000 female mamluks; cf. Ibn al‘Adim, Bughyat al-talab fi ta’rikh Halab, partial ed. A. Sevim, as Biyografilerle Selcukular Tarihi Ibnu’l-Adim Bügyetu’t-taleb fi Tarihi Haleb (Ankara, 1982), 89. In any case, it may be assumed that it was an unusually large group. 25. Tr. Darke, 74–87. 26. Fairbanks, Saljuq Bureaucracy, 151–66. 27. Ibid., 151, citing the Ta’rikh al-vuzara’, Dar al-­kutub, Cairo, Persian ms 7, f. 85a. 28. Ta’rikh al-vuzara’, f. 85b; Fairbanks, Saljuq Bureaucracy, 152. 29. Fairbanks lists other viziers who accompanied their sultans into battle in the late Seljuq period and he points out that for the twelfth century ‘only slightly more than half the viziers had ever held positions in the bureaucracy which they were appointed to head’; cf. Fairbanks, Saljuq Bureaucracy, 152–3. 30. Cf. EI 2, art: Mu‘in al-­Din Parwana (C. Hillenbrand); C. Hillenbrand, ‘Mu‘in

ni z a m a l-m ul k : a m a ve ri ck viz ie r ?  | 267 al-­Din Parwana: the servant of two masters?’, in F. de Jong (ed.), Miscellanea Arabica et Islamica (Louvain, 1993), 267–75. 31. EI 2, art: wazir: The Fatimid caliphate (Th. Bianquis). 32. Ibn Muyassar, Akhbar Misr. Choix de passages de la Chronique d’Égypte d’Ibn Muyassar, ed. A. F. Sayyid (Cairo, 1981), 52. For a detailed account of the reign of al-­Mustansir, cf. K. Clarke, Aspects of the Reign of al-Mustansir Bi’llah, 427–487/1036–1094, unpublished PhD thesis (University of Edinburgh, 2006). 33. Note that Badr al-­Jamali was called wazir al-sayf wa’l-qalam (the vizier of the sword and the pen): H. Eisenstein, ‘Die Wezire Ägyptens unter al-­Mustansir ah 452–466’, WZKM LXXVII (1987), 49. 34. Personal communication, October 2012. 35. S. G. Agadshanow, Der Staat der Seldschukiden und Mittelasien im 11.–12 Jahrhundert, tr. R. Schletzer (Berlin, 1994), 34; Andrew Peacock remarks cautiously that ‘at least a nod towards the continuation of these steppe political traditions among the Seljuqs is suggested by the title yabghu’; A. C. S. Peacock, Early Seljuq History: A New Interpretation (Abingdon, 2010), 33. 36. Agadshanow, Staat, 37 and 145. It is also interesting that Golden in his extremely helpful personal message to me comments that: ‘The Georgians adopted the institution (at’abagi). It is noted under Queen T’amar (r. 1184–1212) in the K’art’lis Ts’khovreba, ed. S. Qaukhch’shvili (Tbilisi, 1955), I, 368, 370’. Tamar is credited with creating the office. 37. Lambton remarks that the first well-­attested mention of the term atabeg in a Muslim context is in connection with Nizam al-­Mulk; Lambton, CHI, V, 239–40. 38. Bosworth suggests that this early Turkish title apparently means ‘pale-­ complexioned’: cf. al-­ Husayni, Akhbar, tr. Bosworth, 138, n. 115, citing L. Rasonyi and I. Baski, Onomasticon turcicum. Turkic Personal Names, Indiana University Uralic and Altaic Studies 172 (Bloomington, IN, 2007), II, 389, 634–5. 39. Al-­Husayni, Akhbar, tr. Bosworth, 26. 40. Ibn Khallikan, Wafayat, tr. de Slane, I, 413. 41. Ibn al-­Athir, Al-Kamil, tr. Richards, 181–2. 42. Ibn Khallikan, Wafayat, tr. de Slane, I, 330. 43. C. Cahen, Pre-Ottoman Turkey, tr. J. Jones-­Williams (London, 1968), 37. 44. The term became known in later Chinese sources as a-t’a-bei; cf. E. Bretschneider, Medieval Researches from Eastern Asiatic Sources (London, 1967), I, 146.

268  |  the me d ie va l tur k s 45. Safi, Politics, 133. 46. Ibn al-‘Adim, Bughya, 67.

Bibliography Primary Sources Aristakes of Lastiverd, Aristakès de Lastivert. Récit des malheurs de la nation arménienne, trs M. Canard and H. Berbérian (Brussels, 1973). Husayni, al-, Akhbar al-dawlat al-saljuqiyya, ed. M. Iqbal, reprint (Beirut, 1984); tr. C. E. Bosworth, as The History of the Seljuq State: A Translation with Commentary of the Akhbar al-dawlat al-saljuqiyya (Abingdon, 2011). Ibn al-‘Adim, Bughyat al-talab fi ta’rikh Halab, partial ed. A. Sevim, as Biyografilerle Selcukular Tarihi Ibnu’l-Adim Bügyetu’t-taleb fi Tarihi Haleb (Ankara, 1982). Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, tr. D. S. Richards as The Annals of the Seljuk Turks (London, 2002). Ibn al-­Jawzi, Al-Muntazam fi ta’rikh al-muluk wa-l-umam, VIII, ed. F. Krenkow (Hyderabad, 1359). Ibn Hassul, Kitab tafdil al-atrak ‘ala sa’ir al-ajnad, ed. A. ‘Azzawi (Cairo, 1940). Ibn Khallikan, Wafayat al-a‘yan wa anba’ ibna’ al-zaman, 2 vols (Cairo, 1299/1881– 2); tr. Baron M. de Slane as Kitab Wafayat al-a‘yan: Ibn Khallikan’s Biographical Dictionary, 4 vols (Paris, 1843–71). Ibn Muyassar, Akhbar Misr. Choix de passages de la Chronique d’Égypte d’Ibn Muyassar, ed. A. F. Sayyid (Cairo, 1981). Matthew of Edessa, Patmut’iwn, tr. E. Dulaurier as Chronique de Matthieu d’Edesse (962–1136) avec la continuation de Grégoire le Prêtre jusqu’en 1162 (Paris, 1858). Nishapuri, Saljuqnama, ed. A. H. Morton (Chippenham, 2004). Nizam al-­Mulk, Siyasatnama, tr. H. Darke, as The Book of Government or Rules for Kings (Henley-­on Thames, 1978). Rawandi, Rahat al-sudur wa ayat al-surur, ed. M. Iqbal (London, 1921). Sibt b. al-­Jawzi, Mir’at al-zaman fi ta’rikh al-a‘yan, ed. A. Sevim (Ankara, 1968). Subki-­al, Tabaqat al-Shafi’iyya al-kubra, eds M. Tanahi and A. al-­Hilw (Cairo, 1964–76). Secondary Sources Ayaz, Q., An Unexploited Source for the History of the Saljuqs: A Translation and Critical Commentary on the Akhbar al-dawlat al-saljuqiyya, unpublished PhD thesis (University of Edinburgh, 1985).

ni z a m a l-m ul k : a m a ve ri ck viz ie r ?  | 269 Bretschneider, E., Medieval Researches from Eastern Asiatic Sources (London, 1967). Clarke, K., Aspects of the Reign of al-Mustansir Bi’llah, 427–487/1036–1094, unpublished PhD thesis (University of Edinburgh, 2006). Encyclopaedia of Islam, 2nd edn, art: Mu‘in al-­Din Parwana (C. Hillenbrand). EI 2, art: wazir: The Fatimid caliphate (Th. Bianquis.) EI 2, art: wazir: In Persia (A. K. S. Lambton). Eisenstein, H., ‘Die Wezire Ägyptens unter al-­Mustansir ah 452–466’, WZKM LXXVII (1987), 37–50. Fairbanks, S., The Ta’rikh al-vuzara’: A History of the Saljuq Bureaucracy, unpublished PhD thesis (University of Michigan, 1977). Hillenbrand, C., ‘Mu‘in al-­Din Parwana’: the servant of two masters?’ in Miscellanea Arabica et Islamica, ed. F. de Jong (Louvain, 1993), 267–75. Klausner, C. L., The Seljuk Vezirate: A Study of Civil Administration 1055–1194 (Cambridge, MA, 1973). Lambton, A. K. S., ‘The internal structure of the Saljuq Empire’ in The Cambridge History of Iran, V: The Saljuq and Mongol Periods, ed. J. A. Boyle (Cambridge, 1968), 203–82. Le Strange, G., Description of the Province of Fars in Persia at the Beginning of the Fourteenth Century a.d. f ro m t he MS. o f Ibn al -Bal khi in t he Brit ish Museum (London, 1912). Peacock, A. C. S., Early Seljuq History: A New Interpretation (Abingdon, 2010). Rasonyi L. and I. Baski, Onomasticon turcicum. Turkic Personal Names, Indiana University Uralic and Altaic Studies 172 (Bloomington, IN, 2007). Safi, O., The Politics of Knowledge in Pre-modern Islam: Negotiating Ideology and Religious Inquiry (Chapel Hill, NC, 2006).

19 The Life and Times of ‘Amid al-­Mulk al-­Kunduri

T

here is no doubt that the most famous vizier of the first Seljuq sultan Tughril has acquired a somewhat curious reputation.1 Two lurid episodes stand out from the story of his life and d ­ eath – fi ­ rstly, his castration ordered by Tughril (or his alleged self-­castration); and, secondly, his death at the behest of Nizam al-­Mulk, after which his body parts were sent to diverse places in the eastern Islamic world. These stories do him no favours at all. Indeed, much of the accepted view of Kunduri’s life comes through the lens of a retrospective and negative comparison between him and Nizam al-­Mulk, his immediate successor in the Seljuq vizierate. This perception is present already in most of the primary sources and is swallowed wholesale by none other than George Makdisi in his article on Kunduri in the second edition of the Encyclopaedia of Islam.2 In this piece, Makdisi, normally such a sound scholar, goes overboard; his presentation of Kunduri is little more than a tirade against him, a rant in which Nizam al-­Mulk emerges superior to Kunduri on every front. Makdisi writes: The main ambition of these two viziers was to manipulate power and influence through the sultans whom they served. Nizam al-­Mulk did this with consummate skill for three full decades, outmanoeuvring his rivals, and always keeping a step ahead of them in their plots and intrigues against him.

Makdisi then declares: ‘In comparison with the magisterial politics of Nizam al-­Mulk, Kunduri appears as a bungling fool.’3 Thereafter Makdisi embarks on an account of what he views as the egregious blunders committed by Kunduri during his career as vizier. Anyone looking for a detailed 270

t h e l ife a nd tim es of ‘a mi d a l-mul k a l - k und ur i  | 271 biography and balanced assessment of Kunduri will not find it in Makdisi’s article. This paper will not discuss one key facet of Kunduri’s c­ areer – ­his religious policies, and in particular his involvement in the inter-­sectarian disputes in eleventh-­century Nishapur. This is a subject that has already received a great deal of academic attention and has produced lively debate amongst scholars of medieval Islamic history, law and theology.4 Recent scholarship has provided a more nuanced and balanced picture of Kunduri’s attitudes and conduct in the complex religious and political milieu of Seljuq Nishapur and the events of the anti-‘Asharite mihna there, showing that he was by no means as fanatical and intransigent in his views as he has been painted.5 After giving a short summary of Kunduri’s life, this chapter has three aims: firstly, to examine aspects of the episode of Kunduri’s ­castration – ­a subject that has received little attention in the scholarly literature; secondly, to analyse Kunduri’s career on its own terms, without its being constantly seen through the prism of the achievements of his much more famous successor, Nizam al-­Mulk; and thirdly, to reflect more generally on the characteristics of the early Seljuq vizierate as exemplified in Kunduri’s career. A Short Summary of Kunduri’s Life Kunduri was born in 415/1024–5 into a landowning family of Kundur in Khurasan.6 He was educated alongside the poet Bakharzi in Nishapur in the circle of the Shafi‘i scholar al-­Muwaffaq al-­Nishapuri. Kunduri joined the service of the Seljuq sultan Tughril as a scribe in the diwan al-rasa’il when Tughril came to Nishapur.7 In 446/1054–5 Kunduri, a mere thirty or so years old, took over as probably Tughril’s fourth vizier.8 At such a young age his rise to power was meteoric. The primary sources chronicle in some detail a series of important occasions on which Kunduri played a key role. They include the following events. When Tughril arrived in Iraq in Ramadan 447/December 1055 Kunduri negotiated, both outside and inside Baghdad, on Tughril’s behalf with the entourage of the caliph al-­Qa’im; eight days before the end of Ramadan/15 December 1055, al-­Qa’im pronounced the khutba in Tughril’s name.9 In Muharram 448/April 1056, Kunduri then successfully brokered a marriage between the caliph and Tughril’s niece, Arslan Khatun.10 During Tughril’s

272  |  the me d ie va l tur k s visit to Baghdad in 449/1057–8, Kunduri orchestrated a grand ceremonial face-­to-­face meeting between his master and the caliph.11 In 453/1061 Kunduri drew up a contract for Tughril to marry the caliph’s daughter. The horrified caliph summoned Kunduri to protest but was finally forced to sign the marriage contract the following year.12 The bride was taken to the caliphal palace in 455/1063.13 When the childless Tughril died in Rayy later the same year, Kunduri was seventy leagues away besieging Tughril’s nephew Qutlumush in the fortress of Girdkuh near Damghan. Kunduri acted decisively, arriving in Rayy only two days later, and he himself conducted Tughril’s funeral.14 He then had Tughril’s nephew, Sulayman b. Chaghri, named in the khutba as his heir. This appointment was, however, only short-­lived. Sulayman’s brother, Alp Arslan, was in Qazwin with his vizier Nizam al-­Mulk, who arranged for the khutba to be said in Alp Arslan’s name. Hoping to stay in his post with the new sultan, Kunduri then gave his public support in Rayy for Alp Arslan and, according to some sources, Alp Arslan appointed him as his vizier.15 However, in 456/1064, after Nizam al-­Mulk had met Kunduri and had become jealous and intimidated by the latter’s grand retinue and large number of troops, he obviously recognised a serious rival and at once moved against him.16 He persuaded Alp Arslan to arrest Kunduri and to send him in custody to Marw al-­Rudh. Kunduri spent a year there in captivity before being killed. He died on 16 Dhu’l-­Hijja 456/29 November 1064. He was in his early forties and had served as vizier for eight years and a few months.17 This short summary of the major events in Kunduri’s career highlights some substantial successes, and shows that despite his relative youth he was a seasoned and supple politician, clearly a man to reckoned with. Nizam al-­Mulk, though, had no room for a colleague to share his throne. 1. Kunduri’s Castration The custom of male castration in ancient and medieval societies is, of course, well known. Castration could be religiously motivated as in Christian contexts. The famous example of ‘the towering genius’,18 Origen (c. 185–c. 254), comes to mind. Taking literally Jesus’ words in Matthew 19: 12: ‘And there be eunuchs, which have made themselves eunuchs for the kingdom of heaven’s sake’, Origen performed what has been called ‘a brave, irrevocable, and

t h e l ife a nd tim es of ‘a mi d a l-mul k a l - k und ur i  | 273 controversial act’;19 he castrated himself to devote himself exclusively to God. Eunuchs were used at the Sasanian court to guard the harem and possibly also as cupbearers.20 Medieval Muslim rulers also employed eunuchs to supervise the women’s quarters.21 Medieval Muslim ethnographic writings, such as the works of al-­Muqaddasi and al-­Mas‘udi, devote short chapters to the subject of eunuchs.22 However, amongst the terrible punishments inflicted on malefactors in medieval times in the Islamic world, there would appear to be little mention of the punishment of castration. The story of Kunduri’s castration is included in almost all the chronicles which deal with Seljuq history, and there is no good reason to doubt it. This event is shown to have happened early during Tughril’s time, except in two accounts when the story is improbably placed at the beginning of the reign of Alp Arslan.23 There are two main versions of the ­story – ­one in which Tughril had Kunduri castrated, and one (less frequently given) in which Kunduri castrates himself. Some chronicles include both versions; others relate only that the castration was carried out on Tughril’s orders. No version mentions only Kunduri’s self-­castration. An additional variant is that Alp Arslan ordered the castration shortly after he had become sultan. The exact timing of the event though is not as significant as the event itself. i. The Castration of Kunduri Ordered by Tughril In this version of the story Tughril had Kunduri castrated because when he sent him to ask on his behalf for the hand in marriage of an unnamed woman, Kunduri married her himself. Despite the castration, Tughril kept him in his service.24 The chroniclers do not give an exact date for this event but al-­Bundari says that it happened at the beginning of Tughril’s reign.25 Supposing this story is taken at face value, it is noteworthy that, despite Kunduri’s unwise action in stealing the woman from Tughril, who then had him castrated, the sultan still kept Kunduri in his service as his vizier. In this version of events, it is clear that Tughril still recognised and needed Kunduri’s outstanding professional abilities. Meting out such a punishment guaranteed that any further sexual misconduct would not recur; Kunduri would now focus single-­mindedly on the job in hand.

274  |  the me d ie va l tur k s ii. Kunduri’s Self-­castration Ibn al-­Athir also mentions the self-­castration as occurring in Tughril’s time and quotes a poem about the event by Bakharzi containing puns about testicles.26 A variant of the story, included in the account given by Ibn Khallikan, relates that after Alp Arslan had confirmed Kunduri in the post of vizier after Tughril’s death and had sent him to secure a marriage contract with the Khwarazmshah’s daughter, Kunduri’s enemies spread a rumour that he wanted to marry her himself. When Kunduri heard that Alp Arslan was angry about this, he was so afraid that he cut off his beard and ‘eradicated from his body every trace of manhood’.27 In this version it is obvious that, despite being in his prime, Kunduri preferred high office and the exercise of enormous power to continued sexual activity. iii. Reflections on the Castration Narratives The castration story is inserted into the chronicles without any comment at all. Yet its widespread inclusion in the primary Arabic and Persian sources suggests that its memory lingered on tenaciously in twelfth- and thirteenthcentury historiography. Perhaps the chroniclers of this period, who are so laudatory about the achievements of Nizam al-­Mulk, include the castration story (in one or both of its two main versions) in order to stress even more strongly the superiority of Nizam al-­Mulk, Kunduri’s successor as Seljuq vizier. However, the story is often placed somewhat uneasily in their obituary notices of Kunduri alongside the usual stereotypical panegyric phrases accorded to famous men. One does wonder whether the castration story has an additional function and whether it is being used here in a metaphorical sense, as an exemplary warning of the dire dangers inherent in the exercise of supreme power. It is in this metaphorical sense that its probable historical context in Tughril’s reign carries with it some significant historical connotations which would not have been lost on an educated Persian audience. There are unmistakable resonances here from the story of Abarsam, the close adviser of the first Sasanian ruler, Ardashir son of Babak.28 This story appears in Firdawsi’s Shahnama29 but it is told at some length before Firdawsi’s time by al-­Tabari in his World History.30 Firdawsi would not, however, have needed the narra-

t h e l ife a nd tim es of ‘a mi d a l-mul k a l - k und ur i  | 275 tive of al-­Tabari to inspire him to write about the Sasanians since he could draw on ‘an ancient and still living folk tradition’.31 In the story of Abarsam and, as already mentioned, in some of the accounts of Kunduri’s life, both these high-­ranking government advisers castrate themselves. Both serve the first ruler of a new dynasty. Both are associated, if only by gossip (especially in the case of Abarsam), with sexual misconduct, involving a woman in whom the ruler has an interest. Finally, and perhaps most significantly, both advisers continue to serve their masters afterwards. Abarsam gives Ardashir a sealed box containing his testicles and tells him that he has castrated himself so as to exonerate himself from any potential suspicion on the part of the king. He says to Ardashir: ‘I castrated myself so that no-­one could speak evilly of me and soak me in a sea of infamy.’32 Sasanian history and culture were very familiar to the cultivated Perso-­ Islamic elites in Khurasan, especially in view of the popularity of Firdawsi’s Shahnama and the oral traditions on which this poetic masterpiece were based. Moreover, the numerous exemplary anecdotes and models of good government provided in the Siyasatnama of Kunduri’s arch-­enemy, Nizam al-­Mulk, reveal just how deeply entrenched the memories of Sasanian statecraft and culture still were in Khurasan.33 So it is highly likely that Kunduri and his contemporaries, as well as some at least of the twelfth and thirteenth-­ century chroniclers, both those who wrote in Arabic and those who used Persian, would have known the story of Ardashir and Abarsam. It is also worthy of mention that at the beginning of his account of Tughril’s reign, Rawandi quotes a saying of none other than Ardashir b. Babak: ‘Every king must obtain a ­vizier . . . ­for his vizier is the prop of his kingdom.’ These arguments suggest that even in the twelfth and thirteenth centuries the exemplary nature of Sasanian models of good government had not been forgotten. While no Arabic or Persian source explicitly links the Sasanian and Seljuq castration stories, it is eminently plausible that Kunduri himself would have been familiar with an episode from early Sasanian history that had such sinister personal relevance to his own case. 2. Kunduri’s Achievements as Vizier Previous scholarship about Kunduri, notably the studies by Makdisi and Bosworth, has highlighted vigorously the series of mistakes Kunduri made

276  |  the me d ie va l tur k s during his short but eventful career in his handling of the complex internal rivalries within the Seljuq family. In the light of hindsight, Kunduri seems to have backed the wrong members of the family on several occasions. In particular, his decision after Tughril’s death to support his nephew Sulayman b. Chagri for the sultanate proved disastrous. Kunduri proved no match for Nizam al-­Mulk with his candidate for the sultanate, Alp Arslan, and even after Kunduri had belatedly declared his support for Alp Arslan, his future career was doomed. To balance, to some extent at least, the above-­mentioned negative judgement of Kunduri, the following discussion will focus instead on what may be viewed as his successes during the turbulent eight-­year period of his vizierate. There is no doubt in the sources about his administrative abilities. Al-­Bundari mentions that when Tughril came to Nishapur he needed a scribe who could manage both literary Arabic and Persian; al-­Bundari adds that by taking Kunduri into his service Tughril had acquired ‘a youth with the judgement of a middle-­aged man’.34 This makes excellent sense, since otherwise his promotion to the post of royal vizier at such an early age would simply not seem reasonable. Ibn Hassul, a scribe in Tughril’s administration from 434/1058, whom Kunduri called the best scribe of the age,35 writes in his Arabic epistle extolling the virtues of the Turks that Kunduri read and commented on his work and that he translated it into Turkish for the sultan and explained it to him.36 Kunduri’s vizierate was a very difficult time for the Khurasani Perso-­ Muslim bureaucratic elite. To be sure, they had been acquainted with Turks during the Ghaznavid period, but there were no precedents for how to manage nomadic Turks coming into Khurasan on the scale of the Seljuq mass migration. Tughril’s entry into Baghdad was fraught with difficulties and it required sophisticated and delicate handling as well as linguistic and diplomatic skills of the highest order to deal with them. These qualities Kunduri possessed in abundance, and he managed the extremely difficult task of negotiating the relationship between the ‘Abbasid caliph al-­Qa’im on the one hand and the untutored Turkish-­speaking Seljuq sultan Tughril on the other hand. When speaking or writing to the caliph elaborate Arabic protocols that befitted his office were needed, and Kunduri appears to have been the ideal man to write these.

t h e l ife a nd tim es of ‘a mi d a l-mul k a l - k und ur i  | 277 The build-­up to Tughril’s eventual entry into Baghdad was complicated. Preliminary meetings were held by Kunduri with the caliphal retinue, and indeed with the caliph himself. Kunduri needed to draw up documents and letters in Arabic and then to translate and explain these as well as the caliph’s responses to his master Tughril in Turkish, presenting their contents in an appropriate light so as to flatter the sultan’s vanity and not to antagonise him. Once in control in Baghdad, Kunduri the administrator came to the fore; as the author of the Pseudo-­Nishapuri writes: ‘Al-­Kunduri took all the official governmental correspondence under his control in such a manner that the dust of affliction settled nowhere.’37 In particular, Kunduri asked to see papers on the taxation revenues of Baghdad and he decided on a suitable revenue for the caliph.38 In other words, he controlled the caliphal purse strings. The same source relates that whilst in custody just before his death, Kunduri was questioned closely by Nizam al-­Mulk about the conditions in the provinces and about the contents of the tax-­registers of every city and district.39 More generally, it is clear from the sources that Kunduri’s administrative skills were fully recognised by Nizam al-­Mulk, who was envious of Kunduri’s ostentatious public persona but must also have feared him just because of his all-­round competence.40 Much then had been achieved in just over eight years by such a young but brilliant linguist and bureaucrat. Kunduri’s last hours are presented in the sources as a model for how to meet death with courage, dignity and piety. Al-­Bundari takes the trouble to point out the high status that Kunduri had enjoyed in his heyday by mentioning that his torso was wrapped in a cloth which was part of the covering of the Prophet’s burda and in a brocade; both these valuable textiles he had received as gifts from the caliph al-­Qa’im himself.41 However, Kunduri’s death at such a young age provides an opportunity for the chroniclers to launch into didactic mode about the fragility and dangers of power; indeed, according to the well-­known exemplary quotation given by Ibn Khallikan (and mentioned in different phrasing in other sources), Kunduri instructed his executioner to say to the vizier Nizam al-­Mulk: ‘You have acted wrongly in teaching the Turks to put to death their viziers and administrative heads. He who digs a pit shall fall into it.’42

278  |  the me d ie va l tur k s 3. The Characteristics of the Early Seljuq Vizierate as Exemplified in Kunduri’s Career Goitein rightly pointed out that in the period of Turkish and Mongol rule ‘the delegation of the highest civil authority to a competent official chosen from the conqueror’s cultivated subjects became imperative’.43 However, a Seljuq vizier had no prescribed role and no fixed tenure of office.44 The career of Kunduri epitomises many of the characteristics as well as the hazards of the vizierate. This was a dangerous position to fill. On the one hand, there was the potential for wielding enormous power and for the acquisition of great wealth. On the other hand, the office brought with it many dangers. It was not easy to enjoy a lasting and close relationship with the sultan who could often become jealous of the power wielded by his vizier. And the safety of a vizier was often endangered by rivalry from other contenders for his position. Many a vizier came to a violent end. All three famous early Seljuq ­viziers – ­Kunduri, Nizam al-­Mulk and Taj al-­Mulk – ­suffered this fate. Scholars such as Lambton have argued for a clear distinction between the ‘men of the pen’, to which group viziers, usually Persian, belonged, and the ‘men of the sword’ who in the Seljuq period were mostly of Turkish extraction.45 She points out, moreover, that the most important function of the vizier was the supervising of state finances. Given his religious education and his administrative skills, Kunduri certainly conforms to the usual profile of a ‘man of the pen’ trained in Perso-­Islamic learning and statecraft. However, the distinction between ‘men of the pen’ and ‘men of the sword’ in early Seljuq times was certainly not as clear-­cut as Lambton, or indeed Makdisi, suggested, and it is important to correct Makdisi’s confident but erroneous assertion in the last sentence of his Encyclopaedia of Islam2 article where he declares: ‘Not the least among al-­Kunduri’s mistakes was his personal engagement in battle, whereas Nizam al-­Mulk remained a man who manipulated the pen, leaving the sword for those better suited to the battlefield.’46 In fact this is totally wrong. Both these viziers, as well as some other later Seljuq viziers, entered the military arena. In his obituary notice of Nizam al-­Mulk, al-­Subki states grandly that Nizam al-­Mulk participated in military operations in Rum, Aleppo, Khurasan and Transoxiana,47 and according

t h e l ife a nd tim es of ‘a mi d a l-mul k a l - k und ur i  | 279 to Sibt b. al-­Jawzi, Kunduri besieged the citadel of Girdkuh in the year 455/1063.48 The problems posed by these unassimilated nomadic Turks were exacerbated by the peripatetic lifestyle that they continued to follow once they had entered the eastern Islamic world. Tughril was constantly on the move, but for the Iranian bureaucrat and scholar that Kunduri was, to accompany the sovereign such protracted journeys must have required enormous reserves of stamina. Yet he had to go with the sultan in order to keep control of what was going on. His strategic skills had to be exercised on the move. Kunduri had to contend with many conflicting groups and their relationship both with each other and with h ­ im – t­ he relationships between the Seljuq family with their nomadic followers, between individual members of the Seljuq family, between the Seljuq sultan and the caliph, and between all these groups and himself. This called for him to perform a diabolically difficult balancing act. Other areas of the role played by Kunduri as vizier included acting as a kind of tutor to the sultan. In his high-­flown panegyric of Kunduri, Ibn Hassul mentions the vizier’s vast range of knowledge in the religious sciences, Arabic grammar, astronomy and physics and he records that Kunduri gave Tughril lessons in good manners, telling him edifying tales about wise men from the past.49 Ibn al-‘Imrani writes that when Tughril met al-­Qa’im, Kunduri translated the caliph’s words from Arabic into Turkish for Tughril.50 Ibn al-­Athir is sometimes hostile to Kunduri, measuring his behaviour unfavourably against that of Nizam al-­Mulk. For example, under the year 451/1059–60, Ibn al-­Athir tells the story of a library in Karkh being burned and books looted; Kunduri came and selected the best of the books. They numbered over 10,000 volumes. Having banished the mob who were looting the books, Kunduri then sat down to choose some for himself. Ibn al-­Athir remarks that this was a misuse of his power and wicked behaviour on his part, concluding as follows: ‘What a difference between his conduct and that of Nizam al-­Mulk who built his madrasas, organized learning in all the lands of Islam, and made pious donations of books and other things.’51 Kunduri does not seem to have appointed a family network in important governorships and high administrative posts in the Seljuq state, in order to

280  |  the me d ie va l tur k s bolster his own position, as Nizam al-­Mulk was to do later. Kunduri hardly had the time to do so though. Concluding Remarks Kunduri established a key part of the role of the vizier in Seljuq ­times – ­namely to begin the protracted process of embedding the rule of the Seljuq Turks into what may be called Perso-­Islamic government and making the presence of the Turkish military leaders and their nomadic followers palatable to the Arab and Persian populations whom they dominated by their military might. Moreover, Kunduri was striving to create a pragmatic working relationship between the Turkish military usurpers and the caliph. Yet there was no fixed blueprint for the role of a Seljuq vizier, nor for how long the post should last, although there were ample precedents from Abbasid, Samanid and Ghaznavid times. The difference for Kunduri and for those who were to succeed him in the post of Seljuq vizier was the need to learn how to handle the Turkish leaders and to control the disruptive presence of the Turkish nomads. Above all, Kunduri had the ­difficult – ­and at times it must have seemed ­impossible – ­task of serving as the intimate adviser and first port of call to a ruler who was a powerful and wilful personality and whose upbringing and mindset could scarcely have been more different from his own. Tughril was a Turk, not a Persian, and he belonged to a nomadic society, not a settled one. Tughril’s Weltanschauung was not that of Kunduri, and it required constant vigilance on the latter’s part to handle the difficulties that arose as a natural result of these different perspectives. Despite the hostile comments made about Kunduri in the primary sources because of the mistakes he made about which Seljuq aspirants to the sultanate he supported, the chroniclers give a generally favourable image of him. Indeed Qummi waxes lyrical about him: ‘After ‘Amid al-­Mulk Kunduri there was no vizier more distinguished in the blessed Seljuq state than he was.’52 Ibn Khallikan praises Kunduri for his ‘acuteness of mind and abilities as a scribe, adding that ‘he administered the state with uncontrolled authority’.53 Al-­Bundari, drawing on a respected historiographical genealogy which went back through ‘Imad al-­Din al-­Isfahani to the earlier twelfth-­century Seljuq vizier Anushirwan b. Khalid, gives the following assessment of Kunduri: ‘He

t h e l ife a nd tim es of ‘a mi d a l-mul k a l - k und ur i  | 281 had sound judgement, he was eloquent, he gave good advice (and) he was dominant in his position, having command of his sultan.’54 However, in his ornately phrased obituary notice of Kunduri, al-­Bundari balances this favourable assessment with a reference to his irresponsibility and the weakness of his ‘long hand’;55 in other words, he lacked the firm grip of the true despot. In sum, it can be said with some confidence that Kunduri was very far from being a bungling fool, as Makdisi avers. He was clearly an outstanding linguist, translator and negotiator. He was the de facto ruler of the Seljuq state at perhaps its most critical ­stage – ­that of obtaining from the Sunni caliph the necessary credentials of rule and titulature for the Seljuqs as ‘sultans of the east and the west’. Through protracted and tortuous negotiations, and despite a highly volatile situation, Kunduri managed to effect a modus operandi between caliph and sultan. It was his misfortune to arouse the hostility of Nizam al-­Mulk because of their religious differences and their intense political rivalry. It is surely inappropriate and unfair to compare Kunduri’s eight-­year rule with the thirty-­year rule of Nizam al-­Mulk, as the primary sources, and indeed earlier secondary research which deals with the Seljuq period, have tended to do. Indeed, it can certainly be said that Nizam al-­Mulk built on the precedents set by Kunduri as well as creating his own strategies for staying in power. Kunduri was a young man though and the castration story, even if apocryphal, suggests his impetuosity and lack of self-­control. Nevertheless, in the arena of grandiose ceremonies and negotiating at the highest level in the land, his reputation must have been a hard act to follow. In a number of ways, he was indeed a trailblazer for Nizam al-­Mulk. Notes   1. This contribution should be read in conjunction with my forthcoming book chapter about Nizam al-­Mulk where there is also an analysis of aspects of the Seljuq vizierate: cf. C. Hillenbrand, ‘Nizam al-­Mulk: A maverick vizier?’, in E. Herzig and S. Stewart (eds), The Idea of Iran, Vol V: The Age of the Seljuqs (London, 2014), 24–35.   2. George Makdisi, art: Al-­Kunduri, Encycopaedia of Islam, 2nd edn, v, 387–8.  3. Ibid.  4. Cf. the impressive corpus of research published by Bulliet on Khurasan in

282  |  the me d ie va l tur k s Samanid, Ghaznavid and Seljuq ­times – ­for example: R. W. Bulliet, ‘Local politics in Eastern Iran under the Ghaznavids and Seljuks’, Iranian Studies xi/1 (1978), 35–56, and idem, ‘The Shaikh al-­Islam and the evolution of Islamic society’, Studia Islamica xxxv (1972), 53–67.   5. Peacock argues this case most cogently: cf. A. C. S. Peacock, Early Seljuq History: A New Interpretation (London, 2010), 99–127.   6. For other accounts of Kunduri’s career, cf. H Halm, ‘Der Wesir al-­Kunduri und die Fitna von Nisapur’, Die Welt des Orients vi/2 (1971), 205–33; C. E. Bosworth, art: Konduri, Mohammad b. Mansur, Encyclopaedia Iranica.  7. H. Bowen, ‘Notes on some early Seljuqid viziers’, Bulletin of the School of Oriental and African Studies xx/1 (1957), 110, citing Bakharzi, Dumyat al-Qasr wa-‘usrat ahl al-‘asr, ed. M. R. Tabbakh (Aleppo, 1930), 641. According to al-­Husayni, Kunduri’s first post was as court chamberlain (hajib al-bab); al-­ Husayni, Akhbar al-dawla al-saljuqiyya, tr. C. E. Bosworth as The History of the Seljuq State: A Translation with Commentary of Akhbar al-dawla al-saljuqiyya (London, 2011), 22.   8. Cf. Ibn al-­Athir, al-Kamil fi l-Ta’rikh, tr. D. S. Richards as The Annals of the Seljuk Turks (London, 2002), 57; Nishapuri, The Saljuqnama of Zahir al-Din Nishapuri, ed. A. H. Morton (Chippenham, 2004), 20. However, according to Rawandi, Kunduri was Tughril’s third v­ izier – ­cf. Rawandi, Rahat al-Sudur wa-ayat al-surur, ed. M. Iqbal (London, 1921), 98. Al-­Bundari calls him ‘the first vizier of the Seljuqs’ – cf. al-­Bundari, Zubdat al-Nusra wa-nukhbat al-‘usra, ed. M. T. Houtsma in Recueil de textes relatifs à l’histoire des Seldjoucides, Vol. II: Histoire des Seldjoucides de l’Irâq (Leiden, 1889), 10.   9. Ibn al-­Athir, Annals, tr. Richards, 100. 10. Ibid., 103–4. 11. Ibid., 114. 12. According to Ibn al-‘Imrani, Kunduri was in charge of organising the marriage contract. He took it with the caliph’s signature to Tughril and explained its contents to him. Ibn al-‘Imrani also gives the text of the document; cf. Ibn al-‘Imrani, Al-Inba’ fi ta’rikh al-Khulafa’, ed. Q. al-­Samarra’i (Leiden, 1973), 198. 13. Ibn al-­Athir, Annals, tr. Richards, 141–2. 14. Ibid., 142–3. 15. Ibid., 145; Sibt b. al-­Jawzi, Mir’at al-Zaman fi ta’rikh al-a‘yan, ed. A. Sevim (Ankara, 1968), 110. For the short-­ lived appointment of Kunduri as Alp Arslan’s vizier, cf. Ibn al-‘Imrani, p. 199; Ibn Khallikan, Kitab wafayat al-a‘yan.

t h e l ife a nd tim es of ‘a mi d a l-mul k a l - k und ur i  | 283 tr. Baron M. de Slane as Ibn Khallikan’s Biographical Dictionary, iii, reprint (Beirut, 1970). 16. Ibn al-­Athir, Annals, tr. Richards, 146; Sibt b. al-­Jawzi, Mir’at, 112. 17. Al-­Bundari, Zubda, 29; al-­Husayni, History, tr. Bosworth, 24. 18. P. Brown, The World of Late Antiquity: From Marcus Aurelius to Muhammad (London, 1971), 82. 19. T. D. Barnes, Constantine and Eusebius (Cambridge, MA, 1981), 83. 20. T. Daryaee, Sasanian Persia: The Rise and Fall of an Empire (London, 2009), 52. 21. For a detailed discussion about eunuchs in medieval Muslim contexts, see D. Ayalon, ‘On the eunuchs in Islam’, Jerusalem Studies in Arabic and Islam i (1979), 67–124; idem, ‘On the term khadim in the sense of ‘eunuch’ in the early Muslim sources’, Arabica xxxii (1985), 289–308. 22. Al-­Muqaddasi, Best Divisions for Knowledge of the Regions: Ahsan al-taqasim fi ma‘rifat al-aqalim, tr. B. Collins (Reading, 2001), 200; al-­Mas‘udi, Muruj al-dhahab, trs P. Lunde and C. Stone as The Meadows of Gold: The Abbasids (London and New York, 1989), 345–6; cf. also L. Richter-­Bernburg, art: castration, EIr, v, 70–3 and C. Pellat, art: Khasi, EI 2, v, 1087–92. 23. Al-­Husayni, Akhbar, tr. Bosworth, 23; Ibn Khallikan, Biographical Dictionary, tr. de Slane, 293–4. Bosworth rightly points out the improbability of such a placing for this event: C. E. Bosworth, ‘Kondori’. 24. Ibn al-­Athir, Annals, tr. Richards, 148; Sibt b. al-­Jawzi, Mir’at, 124. 25. Al-­Bundari, Zubda, 30. 26. Ibn al-­Athir, Annals, tr. Richards, 148. 27. Ibn Khallikan, Biographical Dictionary, tr. de Slane, iii, 293–4; see also al-­ Husayni, Akhbar, tr. Bosworth, 23. 28. Al-­Tabari writes this name as Harjand b. Sam; cf. al-­Tabari, Ta’rikh al-Rusul wa’l-muluk, tr. C. E. Bosworth as The History of al-Tabari. Volume V. The Sasanids, the Byzantines, the Lakhmids, and Yemen (Albany, 1999), 24, n. 85. 29. A. Ferdowsi, Shahnameh: The Persian Book of Kings, tr. D. Davis (New York, 2000), 556–8. 30. Al-­Tabari, History, tr. Bosworth, V, 24–6; see also al-­Dinawari, Kitab al-Akhbar al-Tiwal, eds ‘A. al-­Mu‘nim ‘Amir and J. al-­Din al-­Shayyal (Cairo, 1960), 43–5. 31. Davis, ‘Introduction’, in Ferdowsi, Shahnameh, tr. Davis, xiii. 32. Ferdowsi, Shahnameh, tr. Davis, 556–8. 33. See D. G. Tor, ‘The long shadow of pre-­Islamic Iranian kingship: antagonism or assimilation’ in T. Bernheimer and A. Silverstein (eds), Late Antiquity: Eastern Perspectives (Oxford, 2012), 145–163.

284  |  the me d ie va l tur k s 34. Al-­Bundari, Zubda, 10. 35. Pseudo-­Nishapuri, tr. K. A. Luther as The History of the Seljuq Turks from the Jami‘ al-tawarikh: An Ilkhanid Adaptation of the Saljuqnama of Zahir al-Din Nishapuri (London, 2001), 43. 36. Ibn Hassul, Kitab Tafdil al-Atrak ‘ala sa’ir al-ajnad, ed. A. ‘Azzawi (Cairo, 1940), 45. 37. Pseudo-­Nishapuri, History, tr. Luther, 46. 38. Ibid., 44. 39. Ibid., 47. 40. Ibid. 41. Al-­Bundari, Zubda, 29. 42. Ibn Khallikan, Biographical Dictionary, tr. de Slane, 294. 43. S. D. Goitein, ‘The origin of the vizierate and its true character’, Islamic Culture xvi (1942), 255–62, 380–92; reprinted in idem, Studies in Islamic History and Institutions (Leiden, 1968), 190. 44. S. Fairbanks, The Ta’rikh al-vuzara’: A History of the Saljuq Bureaucracy, unpublished PhD thesis (University of Michigan, 1977), 140. 45. A. K. S. Lambton, art: Wazir, EI 2, xi, 192. 46. Fairbanks, ‘The Ta’rikh al-vuzara’, 152–3. 47. Al-­Subki, Tabaqat al-Shafi‘iyya al-Kubra, eds M. M. Tanahi and ‘Abd al-­Fattah Muhammad al-­Hulw (Cairo, 1964), v, 313. 48. Sibt b. al-­Jawzi, Mir’at, 101. 49. Ibn Hassul, Kitab Tafdil al-Atrak, 49. 50. Ibn al-‘Imrani, al-Inba’, 189. 51. Ibn al-­Athir, Annals, tr. Richards, 130–1. 52. Najm al-­Din Qummi, Ta’rikh al-Wuzara’, eds H. M. Tabataba’i (Tehran, 1389), 118. 53. Ibn Khallikan, Biographical Dictionary, tr. de Slane, 291. 54. Al-­Bundari, Zubda, 10. 55. Ibid., 29. The Arabic term used here – yaduhu al-tula – is a calque of the Persian daraz-dast. The term was famously used of the Achaemenid ruler Artaxerxes I, known to the Romans as Longimanus. Once again, there are echoes of ancient Persian models here. What began as a physical description of Artaxerxes I, as shown in the rock relief on his tomb at Naqsh-­i Rustam, took on the metaphorical meaning of rapacity and oppressive rule.

t h e l ife a nd tim es of ‘a mi d a l-mul k a l - k und ur i  | 285 Bibliography Ayalon, David, ‘On the eunuchs in Islam’, Jerusalem Studies in Arabic and Islam i (1979), 67–124. Ayalon, David, ‘On the term khadim in the sense of ‘eunuch’ in the early Muslim sources’, Arabica xxxii (1985), 289–308. Bakharzi, Dumyat al-Qasr wa-‘usrat ahl al-‘asr, ed. Muhammad Raghib Tabbakh (Aleppo, 1930). Barnes, Timothy D., Constantine and Eusebius (Cambridge, MA, 1981). Bosworth, C. Edmund, ‘Kondori’, Encyclopaedia Iranica. Bosworth, C. Edmund, ‘The political and dynastic history of the Iranian world (a.d. 1000–1217)’, in J. A. Boyle (ed.), The Cambridge History of Iran, Vol V: The Saljuq and Mongol Periods (Cambridge 1968), 1–202. Bowen, Harold, ‘Notes on some early Seljuqid viziers’, Bulletin of the School of Oriental and African Studies xx/1 (1957), 105–110. Brown, Peter, The World of Late Antiquity: From Marcus Aurelius to Muhammad (London, 1971). Bulliet, Richard W., ‘The Shaikh al-­Islam and the evolution of Islamic society’, Studia Islamica xxv (1972), 53–67. Bulliet, Richard W., ‘Local politics in eastern Iran under the Ghaznavids and Seljuks’, Iranian Studies xi/1 (1978), 35–56. al-­Bundari, Zubdat al-Nusra wa-nukhbat al-‘usra, ed. M. T. Houtsma in Recueil de textes relatifs à l’histoire des Seldjoucides, vol II: Histoire des Seldjoucides de l’Irâq (Leiden, 1889). Christensen, Arthur, L’empire des Sassanides. Le peuple, l’état et la cour (Copenhagen, 1907). Daryaee, Touraj, Sasanian Persia: The Rise and Fall of an Empire (London, 2009). al-­Dinawari, Kitab al-Akhbar al-Tiwal, eds ‘Abd al-­Mu‘nim ‘Amir and Jamal al-­Din al-­Shayyal (Cairo, 1960). Fairbanks, S., The Ta’rikh al-vuzara’: A History of the Saljuq Bureaucracy, unpublished PhD thesis (University of Michigan, 1977). Ferdowsi, Abolqasem, Shahnameh: The Persian Book of Kings, tr. Dick Davis (New York, 2000). Goitein, S. S., ‘The origin of the vizierate and its true character’, Islamic Culture xvi (1942), 255–62, 380–92; reprinted in S. S. Goitein, Studies in Islamic History and Institutions (Leiden, 1968), 168–96.

286  |  the me d ie va l tur k s Halm, Heinz, ‘Der Wesir al-­Kunduri und die Fitna von Nisapur’, Die Welt des Orients, vi/2 (1971), 205–33. Hillenbrand, Carole, ‘Nizam al-­Mulk: A maverick vizier?’, in Edmund Herzig and Sarah Stewart (eds), The Idea of Iran, Vol. V: Early Islamic Iran (London, 2012), 24–35. al-­Husayni, Akhbar al-Dawla al-Saljuqiyya, tr. C. E. Bosworth as The History of the Seljuq State: A Translation with Commentary of Akhbar al-dawla al-saljuqiyya (London, 2011). Ibn al-­Athir, al-Kamil fi l-Ta’rikh, tr. D. S. Richards as The Annals of the Seljuk Turks (London, 2002). Ibn Hassul, Kitab Tafdil al-Atrak ‘ala sa’ir al-ajnad, ed. A. ‘Azzawi (Cairo, 1940). Ibn al-‘Imrani, al-Inba’ fi Ta’rikh al-Khulafa’, ed. Qasim al-­Samarra’i (Leiden, 1973). Ibn Khallikan, Kitab Wafayat al-A‘yan, tr. Baron Macguckin de Slane as Ibn Khallikan’s Biographical Dictionary, iii, reprinted (Beirut, 1970). Klausner, C. L., The Seljuk Vezirate: A Study of Civil Administration 1055–1194 (Cambridge, MA, 1973). Lambton, A. K. S., ‘The internal structure of the Saljuq empire’, in J. A. Boyle (ed.), The Cambridge History of Iran, Vol. V: The Saljuq and Mongol Periods (Cambridge, 1968), 203–82. Lambton, A. K. S., ‘Wazir’, Encyclopaedia of Islam, 2nd edn, xi, 192–4. Makdisi, George, ‘Al-­Kunduri’, EI 2, v, 387–8. Makdisi, George, ‘The marriage of Tughril Beg’, International Journal of Middle Eastern Studies i (1970), 259–75. al-­Mas‘udi, Muruj al-dhahab, trs Paul Lunde and Caroline Stone as The Meadows of Gold: The Abbasids (London and New York, 1989). al-­Muqaddasi, Best Divisions for Knowledge of the Regions: Ahsan al-taqasim fi ma‘rifat al-aqalim, tr. Basil Collins (Reading, 2001). Nishapuri, The Saljuqnama of Zahir al-Din Nishapuri, ed. A. H. Morton (Chippenham, 2004). Nizam al-­Mulk, Siyasatnama, tr. H. Darke as The Book of Government or Rules for Kings (Henley-­on-­Thames, 1978). Nöldeke, T., Das iranische Nationalepos (Berlin and Leipzig, 1920). Peacock, A. C. S., Early Seljuq History: A New Interpretation (London, 2010). Pellat, Charles, ‘Khasi’ EI 2, iv, 1087–92. Pseudo-­Nishapuri, tr. K. A. Luther as The History of the Seljuq Turks from the Jami‘ al-tawarikh. An Ilkhanid Adaptation of the Saljuqnama of Zahir al-Din Nishapuri (London, 2001).

t h e l ife a nd tim es of ‘a mi d a l-mul k a l - k und ur i  | 287 Qummi, Najm al-­ Din, Ta’rikh al-Wuzara’, ed. Husayn Mudarrisi Tabataba’i (Tehran, 1389). Rawandi, Rahat al-Sudur wa-ayat al-surur, ed. M. Iqbal (London, 1921). Richter-­Bernburg, Lutz, ‘Castration’, EIr, v, 70–3. Sibt b. al-­Jawzi, Mir’at al-Zaman fi ta’rikh al-a‘yan, ed. A. Sevim (Ankara, 1968). al-­Subki, Tabaqat al-Shafi‘iyya al-Kubra, eds Mahmud Muhammad Tanahi and ‘Abd al-­Fattah Muhammad al-­Hulw (Cairo, 1964). al-­Tabari, Ta’rikh al-Rusul wa’l-Muluk, tr. C. E. Bosworth as The History of al-Tabari, Volume V: The Sasanids, the Byzantines, the Lakhmids, and Yemen (Albany, 1999).

20 The Nizamiyya Madrasas

Introduction

T

he most famous of all Seljuq viziers, Nizam al-­Mulk, had a remarkable career. He held the office of the vizierate for a considerable ­time – m ­ ore than twenty-­nine ­years – ­and used it to fashion the Seljuq state, masterminding the turbulent transition of the Turkish nomads from steppe warlords to Perso-­Islamic rulers.1 He ruled an enormous empire in the name of Alp Arslan (ruled 455–64/1063–72), and even longer in the name of Malikshah (465–85/1072–92), the second and third of the Great Seljuq sultans of Iran. Nizam al-­Mulk controlled the Seljuq state, with his iron determination, his Mafioso ruthlessness, his vast wealth and the tentacles of his family connections. He was murdered in 485/1092 in his high seventies.2 In this paper one major facet of the rule of Nizam al-­Mulk will be ­discussed – h ­ is project of building the Nizamiyya madrasas. It will deal with the origins of these madrasas and the supervision that Nizam al-­Mulk exercised over them. The core of the paper, namely the great Nizamiyya in Baghdad, will then be discussed. A few words will be devoted to the topic of the students in the madrasas and the curricula which they followed. The paper concludes with an attempt to answer a fundamental question, namely why were the Nizamiyyas built? The Origins of the Nizamiyya Madrasa3 The institution of the madrasa has long been associated with the name of Nizam al-­Mulk. He came from Khurasan whose patrician families in 288

the ni za mi yya m a dr a sa s  | 289 Nishapur have been analysed so ably by Bulliet.4 This milieu was of great assistance to Nizam al-­Mulk in his grand enterprise of building up the Seljuq state. Nizam al-­Mulk knew the ethos of the patricians of Khurasan and how to enlist their support. He also knew how to balance the two rival legal factions, the Hanafis and Shafi‘is, within the Khurasanian cities. Al-­Kunduri, the predecessor of Nizam al-­Mulk as Seljuq vizier, had persecuted the Shafi‘is in the middle of the eleventh century and prominent scholars such as al-­ Juwayni were forced to flee from Nishapur.5 Bulliet’s approach makes it clear that this anti-­Shafi‘i policy was more about power politics within Nishapur than about legal or theological minutiae. After the murder of al-­Kunduri in 455/1063, Nizam al-­Mulk established in many cities of the Seljuq state his famous network of Shafi‘i madrasas, known in his honour as the Nizamiyyas. There is virtually no surviving material evidence of the Nizamiyyas. It is possible that a ruined building, situated in the remote village of Khargird to the south-­east of Nishapur, may have been a madrasa, but that is not certain. Its beautiful but undated and fragmentary inscription, now in the Tehran Museum, nevertheless bears the name of al-­Hasan b. ‘Ali b. Ishaq, which clearly indicates that it was founded by Nizam al-­Mulk himself.6 The lack of material evidence for the madrasas is in striking contrast to the wealth of information written about them in the medieval Arabic and Persian sources. Ibn al-­Athir (died 630/1233) writes fulsomely that The Nizamiyya madrasas are famous the world over; no city was without one, not even Jazirat ibn ‘Umar, a town lost in a corner of the earth to which nobody pays any attention. Nizam al-­Mulk built a beautiful big madrasa there which is now known as the Radi al-­Din madrasa.7

This is a grand way of saying that these Nizamiyyas were spread right across the Seljuq state. Other sources mention more precisely that they were found in Nishapur, Isfahan, Baghdad, Mosul, Herat, Balkh, Basra, Marv, Amul and Tus.8 The first Nizamiyya was established in Nishapur in 450/1058. Its first director was al-­Juwayni.9 A Nizamiyya was then founded in Isfahan in 455/1063, and the most famous of all the Nizamiyyas, the one in Baghdad, was completed by 459/1067. And as van Renterghem has suggested,10 there were many madrasas erected in Baghdad in the twelfth century, especially by

290  |  the me d ie va l tur k s ‘Abbasid and Seljuq officials and even by private citizens of all four Sunni madhhabs. How did the establishment of the Nizamiyyas occur? It seems likely that Nizam al-­Mulk did not set out with a pre-­conceived master plan to establish the network of Shafi‘i madrasas that proclaimed his name across the whole Seljuq empire. He was accustomed to the existence of madrasas which had long operated in Khurasan (some thirty-­eight madrasas are known to have existed in Nishapur alone)11 and he established his first Nizamiyya in that very city early on in his relationship with the Seljuq family. Thereafter it would appear probable that while he served as vizier to Alp Arslan, he perforce accompanied this most peripatetic sultan who crossed the length and breadth of the Seljuq territories in time-­honoured nomadic Turkish fashion. In the course of their travels the vizier would endow madrasas as occasion offered in the cities that they visited, but the exact order in which they were built is not clear from the sources. Nizam al-­Mulk’s Personal Oversight of his Madrasas It seems that, once he was in full control of the Seljuq state, Nizam al-­Mulk opted to stay for considerable periods in Isfahan, the city he preferred, especially in the second half of his life.12 Not surprisingly, he had chosen to build a Nizamiyya there, as already mentioned. Nizam al-­Mulk made sure that control of individual Nizamiyyas remained in the trusted hands of his sons or other family members.13 The appointment of the academic staff to the Nizamiyyas, however, remained entirely the prerogative of Nizam al-­Mulk himself. This was typical of his steely determination to keep a centralised grip on his madrasas. In modern parlance he was a ‘control freak’. It was simply common sense for him not to become embroiled in local rivalries and conflicts and to appoint men whose credentials he could trust and who came from his own Khurasanian milieu. Thus, for the post of chief professor at the Nizamiyya in Nishapur, he chose the great Shafi‘i/‘Ashari scholar of that city, Imam al-­Haramayn al-­Juwayni. Al-­Juwayni remained in the post until his death in 478/1085.14 For the prestigious Nizamiyya in Baghdad Nizam al-­Mulk appointed another Persian, Abu Ishaq al-­Shirazi, about whom more will be said later. Professors in charge of the Nizamiyyas did not always occupy

the ni za mi yya m a dr a sa s  | 291 the post for a long time. Sometimes they seem to have failed to get ‘tenure’, so to speak. Al-­Ghazali, on the other hand, taught in two Nizamiyyas, in Baghdad and Nishapur.15 He came to Baghdad in Jumada I 484/June–July 1091.16 His brilliance is described in rapturous terms by ‘Imad al-­Din, as cited by al-­Bundari, who writes that: ‘In knowledge he was an overflowing sea and a radiant full moon.’17 Ibn al-­Jawzi mentions that Nizam al-­Mulk gave al-­Ghazali the honorific title of Zayn al-­Din and says of the latter that ‘his speech was honeyed and his mind was brilliant’.18 On some occasions more than one professor was in post at the same time and they gave their classes on alternate days. Such was the case with Abu ‘Abdallah al-­Tabari and Abu Muhammad ‘Abd al-­Wahhab al-­Shirazi, who both arrived within a few months of each other with diplomas appointing them to the Nizamiyya in Baghdad in 483/1090.19 The Nizamiyya in Baghdad Let us now turn to the key monument under discussion in this paper, the Nizamiyya in Baghdad, called by Hamdallah Mustawfi the ‘mother of madrasas’.20 Ibn al-­Athir reports that in 457/1065 Nizam al-­Mulk ordered Abu Sa‘id al-­Nishapuri to begin work on this Nizamiyya.21 The choice of an architect hailing from Nishapur is surely significant. After all, Baghdad was the acknowledged capital of the Islamic world, culturally as well as politically, and it is next to impossible that the city which enjoyed the prestige of being the home of the ‘Abbasid caliphate would lack competent architects. But it is natural to assume in the case of Nizam al-­Mulk that he would choose someone from his own part of the world for this task because he was confident that such a man would understand his own vision for this building. After all, Nishapur already had its own Nizamiyya, built less than a decade earlier. It is clear therefore that Nizam al-­Mulk, whose patronage dominated the building programmes in the Seljuq state during his lifetime, was inspired by architectural antecedents in Khurasan, and especially in the case of his madrasas. And it was these antecedents that he wanted replicated in Baghdad. The splendid opening ceremony for this key building took place in 459/1067, presided over by the caliph himself, who was accompanied by his entourage. All the inhabitants of the city were assembled that day; their number filled the courtyard of the madrasa and the surrounding quarters of

292  |  the me d ie va l tur k s the city. The waqfs endowed to support the Nizamiyya included those for its library, as well as bazaars, caravanserais and bathhouses.22 However, the appointment of Abu Ishaq al-­Shirazi as the first professor of the Nizamiyya in Baghdad was not without its drama. There are signs that the building was completed in a hurry. According to Sibt b. al-­Jawzi, houses in Baghdad were pulled down in order to provide material for the building of the Nizamiyya.23 On the opening day the people were assembled to welcome the arrival of al-­Shirazi, but the great man did not show up.24 Ibn al-­Athir continues the story, relating that al-­Shirazi had been visited by a youth who had reproached him for agreeing to teach within a building that was ill-­ gotten property. So al-­Shirazi stayed away.25 Another scholar was appointed temporarily at the Nizamiyya, but after twenty days al-­Shirazi was persuaded to take up the post.26 Even at a distance Nizam al-­Mulk dictated what happened in his madrasas. For example, when al-­Shirazi died in 476/1084,27 Mu’ayyad al-­Mulk, the son of Nizam al-­Mulk,28 quickly appointed a new principal teacher at the Nizamiyya in Baghdad. Nizam al-­Mulk was angered by this and ordered the madrasa to be closed for a whole year as a sign of mourning.29 This demonstrates the control that Nizam al-­Mulk exercised over the appointments to the Nizamiyyas and their curriculum. He hired and fired the staff and kept a close eye on their activities. The Nizamiyya in Baghdad was not just a teaching institution; it also served as a place to which the city’s poor would come in order to receive alms. Ibn al-­Athir mentions that when Malikshah visited Baghdad in 484/1091–2 the notables and amirs of his army dispensed alms on behalf of the sultan to countless indigent people.30 What Did the Nizamiyyas Look Like? The key fact is that they were large. According to Ibn Jubayr: The madrasas there number around thirty and they are all in the eastern part (of Baghdad). There is not one madrasa amongst them that falls short of the finest palace. The greatest and most famous of them is the Nizamiyya that was built by Nizam al-­Mulk and was restored in 504/1110.31

However, the vicissitudes of time, fire, flood and invasion have ensured that no traces of any of the Nizamiyyas now remain.32 This makes it difficult to

the ni za mi yya m a dr a sa s  | 293 imagine what they might have looked like. In the case of the Nizamiyya in Baghdad, however, there are perhaps a few clues to be found through looking at the still ­extant – ­though now much denatured-­Mirjaniyya madrasa built by Amir Mirjan in Baghdad in 758/1357. According to the twentieth-­century Iraqi scholar al-­Alusi in his book entitled The Mosques and Monuments of Baghdad, the Mirjaniyya mosque was built on the model of the Baghdad Nizamiyya. The Nizamiyya had three storeys; on the ground floor was situated the room facing the qibla where the students were taught; on the first floor were rooms for the accommodation of the boarders and a library. Al-­Alusi goes on to say that the Nizamiyya was square in form and that in addition to the two storeys already mentioned it contained basements with a kitchen and bathhouse. Unfortunately al-­Alusi does not state his source for this information about the Nizamiyya but he concludes that this short description accords almost exactly with the layout of the Mirjaniyya madrasa.33 The Curriculum The medieval Arabic and Persian sources make it clear that the curriculum in the Nizamiyyas was not restricted to the teaching of Shafi‘i fiqh. Indeed, it is likely that the whole range of the religious sciences was included in the programme followed by the students.34 The Spanish traveller Ibn Jubayr attended a session in the Nizamiyya on Friday 12 Safar 580/25 May 1184; the speaker was Sadr al-­Din al-­Khujandi, the chief Shafi‘i scholar from Khurasan (sayyid al-‘ulama’ al-khurasaniyya wa-ra’is al-a’immat al-Shafi‘iyya), who discussed various branches of sciences (afanin min ‘ulum) in his address to an audience delighted and honoured to see him.35 Such a versatile scholar as al-­Ghazali, for example, who taught in the Nizamiyya in Baghdad from 484–8/1091–5,36 would have found it irksome just to limit his intellectual virtuosity to fiqh, which in one place in his magnum opus, the Ihya’ ‘ulum al-din, he describes as no more than a ‘discourse on menstruation’.37 The syllabus included the Qur’an, hadith, ‘Ashari theology, Arabic grammar, literature and arithmetic.38 Nizam al-­Mulk seems to have practised a ‘hands-­on’ approach to the teaching in his madrasas. He personally visited the Baghdad Nizamiyya in 479/1086, sat in the library and read and dictated sections of Prophetic tradition to the students.39According to ‘Imad al-­Din al-­Isfahani:

294  |  the me d ie va l tur k s In the time of Nizam al-­Mulk great attention was paid to instruction and education; parents busied themselves with the education of their children and would take them to the classes given by Nizam al-­Mulk. He would question then and often take over the responsibility for directing the career of the most intelligent ones.40

Why Were the Nizamiyyas Built? A number of sweeping statements have been made about why Nizam al-­Mulk built his madrasa. It is important to point out first that Nizam al-­Mulk himself did not come up with the idea of the madrasa as a separate building. Already in the sixth/thirteenth century Ibn Khallikan makes this mistake.41 It is, moreover, unlikely that Nizam al-­Mulk set out initially to establish a network of madrasas to combat the Isma‘ili threat posed by the Fatimids in Cairo, as some have claimed.42 This factor may have had some significance by the last decade of Nizam al-­Mulk’s life, but we should rather seek his motivation for embarking on the whole Nizamiyya enterprise inside rather than outside the Seljuq borders. Nor were his madrasas founded by the Seljuq state itself just because Nizam al-­Mulk, a vizier, founded them. As Makdisi rightly pointed out long ago, ‘any Muslim with full legal capacity could found a madrasa’.43 However, I do not agree with another statement by Makdisi, namely that Nizam al-­Mulk ‘founded his network of madrasas to implement his political policies throughout the vast lands of the empire under his sway’.44 Nor is Omid Safi’s statement that the Nizamiyyas were for the propagation of state-­defined ‘Islamic thought’ convincing.45 It would seem more generally that Nizam al-­Mulk worked out religious policies at micro-­level early on in Nishapur while he was working alongside Alp Arslan when the latter was still only governor of that city. Nizam al-­Mulk then applied such policies more widely as his network of Nizamiyyas grew. Bulliet convincingly argues that Nizam al-­Mulk aimed at creating a balance between rival religious factions within the Sunni fold. I do not believe that Nizam al-­Mulk wanted every Muslim to be a Shafi‘i Muslim, as he himself was. Bulliet goes on to argue that Nizam al-­Mulk had planted the seed of ‘state patriciate’ by means of his madrasas.46 But that little word ‘state’ begs the whole question: ‘Who or what is the state at this time?’ And if the state is to act it needs actual people as its instruments. Step forward Nizam al-­Mulk;

the ni za mi yya m a dr a sa s  | 295 his motto could very easily have been, foreshadowing Louis XIV, ‘L’état c’est moi’. And if the next question is: ‘Did this embodiment of the state use public funds or his own private purse for this purpose?’ the answer must remain rather blurred. Most probably Nizam al-­Mulk’s motives for building the Nizamiyyas were complex and overlapping, perhaps pragmatic initially, and then more carefully thought out. At all events, the evidence points clearly to a strongly accented local bias in the entire Nizamiyya enterprise. The various strands are easily enumerated: the development of the madrasa concept in Nishapur from at least the tenth century onwards; the fact that the earliest recorded Nizamiyya was built in that city; the choice of a Nishapuri architect to build the key Nizamiyya in Baghdad; the Khurasani provenance of one noted professor after another; and the fact that Nizam al-­Mulk’s career as a vizier began in that same city of Nishapur. He may well have been open-­minded about which madhhab people belonged to, but one thing is clear: he was determined to propagate as widely as possible the intellectual culture of Nishapur, the capital of his home province of Khurasan, and his original power base. Concluding Remarks The honorific title Nizam al-­Mulk itself – ‘the ordering of the realm’ – was taken by his admirers as a symbol of his power. To his biographer, al-­Subki, Nizam al-­Mulk was the government; ‘his vizierate was no vizierate, but it rather stood above the sultanate’.47 His role as vizier was even more significant than that of the Barmakid family of viziers under the ‘Abbasids, because Nizam al-­Mulk’s so-­called masters were Turkish warlords with no political experience. No wonder the Seljuq state at its height was punningly called al-dawla al-nizamiyya, the well-­ordered state, or alternatively the state of Nizam al-­Mulk. But history was to show that the centrifugal forces inherent in the Turkic concept of family rule worked against the centralised model of the Seljuq state that Nizam al-­Mulk had tried to impose. Nizam al-­Mulk was attempting to achieve balance and harmony between various conflicting forces:48 the Turcomans and the sultans; the sultans and their relatives; the Seljuqs and their amirs; the cities and the nomads; the different madhhabs within Sunni Islam; and above all, he sought to establish stability in an empire which he

296  |  the me d ie va l tur k s had almost single-­handedly built up and which eventually covered the land mass ruled by the ancient Sasanian empire. This harmonising and balancing he achieved through his various networks, one of which was the madrasa system. Whilst he was alive, the unity of the Seljuq state remained intact.49 But once he was dead, nothing worked as well again. However, there is an optimistic coda to this melancholy conclusion. In 625/1227, more than a century after the death of Nizam al-­Mulk, in the very city where his most celebrated Nizamiyya stood, there rose the greatest medieval madrasa of them all, the Mustansiriyya,50 whose global r­ each – f­or it catered for all four Sunni madhhab­s – ­would surely have pleased him. In spirit, if not in name, it perpetuated his ­legacy – ­and indeed went further still, opening its doors to students all over the Muslim world. Notes  1. For the life of Nizam al-­Mulk, see Ibn al-‘Adim, Bughyat al-talab fı ta’rikh Halab, partial ed. A. Sevim, as Biyografilerle Selcukular Tarihi Ibnu’l-Adim Bügyet ul-taleb fi Tarihi Haleb (Ankara, 1976), 59–94; al-­Subki, Tabaqat alshafi‘iyya al-kubra, ed. unidentified (Cairo, 1964–76), III, 135–45; Ibn al-­Jawzi, Muntazam fi ta’rikh al-umam wa’l-muluk, ed. unidentified, IX (Hyderabad, 1359/1940), 64–8; al-­ Bundari, Zubdat al-nusra wa-nukhbat al-‘usra, ed. M. T. Houtsma in Recueil de textes relatifs à l’histoire des Seldjoucides (Leiden, 1889), II, 56–7; Ibn al-­Athir, Al-Ta’rikh al-bahir fi’l-dawlat al-atabakiyya, ed. A. Tulaymat (Baghdad, 1382/1963), 9–10; Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, X, ed. C. J. Tornberg (Leiden, 1864), 207–10; Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, partial tr. D. S. Richards as The Annals of the Seljuq Turks (London, 2002), 255–8; Ibn Funduq, Ta’rikh-i Bayhaq, ed. A. Bahmanyar (Tehran, 1938), 73–83; Ibn Khallikan, Kitab wafayat al-a‘yan, tr. Baron M. de Slane as Ibn Khallikan’s Biographical Dictionary (Beirut, 1979), I, 413–5; see also EI 2, art: Nizam al-­Mulk (H. Bowen/C. E. Bosworth); N. Yavari, art: Nizam al-­Mulk in A Rippin (ed.), The Islamic World (Abingdon, 2008), 351–8.  2. According to Ibn al-‘Adim, Nizam al-­Mulk died aged seventy-­six years, ten months and ninteen days. He was carried on a litter because of his old age; Bughyat, 91. See also Ibn al-­Athir, Kamil, X, 204; tr. Richards, 253.   3. On the origins of the Nizamiyyas and discussions of the madrasa institution in general, see H. Halm, ZDMG, Suppl. III/1, XIX, Deutscher Orientalistentag (1977), 438–48; A. Talas, L’enseignement chez les Arabes. La Madrasa Nizamiyya

the ni za mi yya m a dr a sa s  | 297 et son histoire (Paris, 1939), especially 13, 26–31; R. Hillenbrand, Islamic Architecture: Form, Function and Meaning (Edinburgh, 1994), 173–250, and especially 175–83; R. Bulliet, The Patricians of Nishapur (Cambridge, MA, 1972), 249–55; G. Makdisi, ‘The madrasa as a charitable trust and the university as a corporation in the Middle Ages’, in Actes du Ve Congrès International d’Arabisants et d’Islamisants (Brussels, 1971), 329–37; A. L. Tibawi, ‘Origin and character of “al-­madrasah”’, BSOAS 25 (1962), 225–38; E. Herzfeld, Damascus: Studies in Architecture, II, Ars Islamica X (1943), 13–14.  4. Bulliet, Patricians, 56–7, 72–5.   5. R. Bulliet, ‘Local politics in Eastern Iran under the Ghaznavids and Seljuks’, Iranian Studies 11 (1978), 50.  6. E. Herzfeld, ‘Eine Bauschrift von Nizam al-­Mulk’, Der Islam XII (1921), 98–101; see also S. Blair, The Monumental Inscriptions from Early Islamic Iran and Transoxiana (Leiden, 1992), 149–52 and plates 99–101; Hillenbrand, Islamic Architecture, 180.   7. Ibn al-­Athir, Atabegs, 9; see also Kamil, tr. Richards, 257.   8. Al-­Subki, Tabaqat, III, 137.  9. Bulliet, Patricians, 254–5, citing ‘Abd al-­ Ghafir al-­ Farisi, Siyaq li-ta’rikh Naysabur, facsimile ed. R. Frye (Cambridge, MA, 1965), ff. 48b, 72a and 90a. 10. V. Van Renterghem, ‘Controlling and developing Baghdad: caliphs, sultans, and the balance of power in the Abbasid capital (mid-­5th/11th to late-­6th/12th c.)’, in C. Lange and S. Mecit (eds), The Seljuqs: Politics, Society and Culture (Edinburgh, 2011), 122. 11. Bulliet, Patricians, 249–54; by the time that Nizam al-­Mulk was born, there were already four madrasas in Nishapur; al-­Subki, Tabaqat, III, 137. 12. D. Durand-­Guédy, Iranian Elites and Turkish Rulers: A History of Isfahan in the Saljuq Period (London, 2010), 304. 13. Nizam al-­Mulk had twelve sons to each of whom he gave a job (shughl) and governorship: see Nishapuri, Saljuqnama, ed. A. H. Morton (Chippenham, 2004), 31; Rawandi, Rahat al-sudur, ed. M. Iqbal (London, 1921), 132. 14. The tradition of appointing local personages continued long after the death of Nizam al-­Mulk. Ibn Khallikan mentions a Shafi‘ite scholar named Qutb al-­Din al-­Nishapuri (d. 578/1183) who gave lessons in the Nizamiyya in Nishapur: tr. de Slane, III, 351–2. 15. Al-­Ghazali was later persuaded to come out of retirement in Tus to take over as director of the Nizamiyya in Nishapur in 499/1105–6; Bulliet, Patricians, 255. 16. Al-­Bundari, Zubdat, II, 36–44. Other Shafi‘i scholars who taught at the

298  |  the me d ie va l tur k s Nizamiyya in Baghdad included Ahmad al-­Ghazali (d. 520/1126), the poet and brother of Abu Hamid, gave lectures in the Nizamiyya in Baghdad, substituting for his brother; see Ibn Khallikan, Wafayat, tr. de Slane, I, 79; a scholar named al-­Mihani was appointed twice to the Nizamiyya in Baghdad; see Ibn Khallikan, Wafayat, tr. de Slane, I, 189–90; and al-­Usuli, who taught there for a month and then died; Ibn Khallikan, Wafayat, tr. de Slane, I, 80. 17. Al-­Bundari, Zubdat, 80. 18. Ibn al-­Jawzi, Muntazam, IX, 55. 19. Ibn al-­Athir, Kamil, tr. Richards, 247. 20. Hamdallah Mustawfi, Nuzhat al-qulub, ed. G. Le Strange (Leiden, 1915), 35. 21. Ibn al-­Athir, Kamil, X, 18. 22. Ibn al-‘Adim mentions that Nizam al-­Mulk endowed madrasas and plentiful waqfs and that he established libraries; Bughyat, 67. 23. Sibt b. al-­Jawzi, Mir’at al-zaman fi ta’rikh al-a‘yan, partial ed. A. Sevim (Ankara, 1968), 184. 24. Ibn Khallikan, Wafayat, tr. de Slane, II, 164; Sibt b. al-­Jawzi, Mir’at, 185. 25. Ibn al-­Athir, Kamil, X, 55; tr. Richards, 161–2. 26. Ibn al-­Athir, Kamil, X, 55; tr. Richards, 162. 27. Ibn al-­Jawzi, Muntazam, X, 7. 28. Ibn Khallikan, Wafayat, tr. de Slane, I, 5–6. 29. Ibn Khallikan, Wafayat, tr. de Slane, I, 11. 30. Ibn al-­Athir, Kamil, tr. Richards, 250. 31. Ibn Jubayr, Rihla, ed. M. J. de Goeje (Leiden, 1907), 229. 32. According to Hammer-­Purgstall, the Nizamiyya was burned by a fire caused by an unprecedentedly large comet in 499 AH/AD 1105; see J. Hammer-­Purgstall, Geschichte der Ilchane (Darmstadt, 1842), 128. Unfortunately Hammer-­ Purgstall does not mention his source. Sibt b. al-­Jawzi mentions the comet but not the fire; Mir’at al-zaman, I, ed. unknown (Hyderabad, 1370 AH/AD 1951), 16. Matthew of Edessa also refers to the comet which was both wonderful and terrifying: Matthew of Edessa, Patmut’iwn, tr. E. Dulaurier as Chronique de Matthieu d’Edesse (962–1136) avec la continuation de Grégoire le Prêtre jusqu’en 1162 (Paris, 1858), 262–3. 33. Quoted by Talas, L’enseignement, 27–8. For a detailed account of the Mirjaniyya, see T. J. al-­Janabi, Studies in Mediaeval Iraqi Architecture (Baghdad, 1982), 113–40. 34. Ibn al-‘Adim, Bughyat, 67. 35. Ibn Jubayr, Rihla, 220.

the ni za mi yya m a dr a sa s  | 299 36. Ibn Khallikan, Wafayat, tr. de Slane, II, 622. 37. Tibawi, ‘Origin and character’, 228. 38. Talas, L’enseignement, 37–8. 39. Ibn al-­Jawzi, Muntazam, IX, 36. 40. Al-­Bundari, Zubdat, II, 57. 41. Ibn Khallikan, Wafayat, tr. de Slane, I, 414. 42. Tibawi, ‘Origin and character’, 234; Talas, L’enseignement, xii and 18. 43. Makdisi, ‘The madrasa as a charitable trust’, 332. 44. Ibid., 334. 45. O. Safi, Politics of Knowledge in Pre-modern Islam: Negotiating Ideology and Religious Inquiry (Chapel Hill, 2006), 7. 46. Bulliet, Patricians, 74. 47. Al-­Subki, Tabaqat, III, 137. 48. Bulliet, ‘Local politics in Eastern Iran’, 52. 49. ‘This Hasan is a barrier against discords’; Ibn al-‘Adim, Bughyat, 67. 50. Described by Hamdallah Mustawfi as the ‘best building there’: Nuzhat, 35.

Abbreviations BSOAS: Bulletin of the School of Oriental and African Studies EI 2: Encyclopedia of Islam, 2nd edition GMS: Gibb Memorial Series ZDMG: Zeitschrift der Deutschen Morgenländischen Gesellschaft

Bibliography Primary Sources ‘Abd al-­Ghafir al-­Farisi, Siyaq li-ta’rikh Naysabur, facsimile ed. Richard Frye (Cambridge, MA: Harvard University Press, 1965). al-­Bundari, Zubdat al-nusra wa-nukhbat al-‘usra, ed. M. T. Houtsma in Recueil de textes relatifs à l’histoire des Seldjoucides (Leiden: Brill, 1889). Hamdallah Mustawfi, Nuzhat al-qulub, ed. Guy Le Strange (Leiden: Brill, 1915). Ibn al-‘Adim, Bughyat al-talab fı ta’rikh Halab, partial ed. A. Sevim, as Biyografilerle Selcukular Tarihi Ibnu’l-Adim Bügyet ul-taleb fi Tarihi Haleb (Ankara: Türk Tarih Kurumu Basimevi, 1976). Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, X, ed. Carolus Johannes Tornberg (Leiden: Brill, 1864), pp. 207–10; Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, partial tr. Donald Sydney Richards as The Annals of the Seljuq Turks (London: Routledge, 2002).

300  |  the me d ie va l tur k s Ibn al-­Athir, Al-ta’rikh al-bahir fi’l-dawlat al-atabakiyya, ed. Ahmad Tulaymat (Baghdad, Dar al-­kutub al-­haditha bi’l-­Qahira wa-­maktabat al-­muthanna bi-­ Baghdad, 1382/1963). Ibn al-­Jawzi, Muntazam fi ta’rikh al-umam wa’l-muluk, ed. unidentified, IX (Hyderabad: Da’irat al-­ma‘arif al-‘uthmaniyya, 1359/1940). Ibn Funduq, Ta’rikh-i Bayhaq, ed. A. Bahmanyar (Tehran: Islamiyya Press, 1938). Ibn Jubayr, Rihla, ed. M. J. de Goeje (Leiden: Brill, 1907). Ibn Khallikan, Kitab wafayat al-a‘yan, tr. Baron MacGuckin de Slane as Ibn Khallikan’s Biographical Dictionary (Beirut: Librairie du Liban, 1979). Matthew of Edessa, Patmut’iwn, tr. E. Dulaurier as Chronique de Matthieu d’Edesse (962–1136) avec la continuation de Grégoire le Prêtre jusqu’en 1162 (Paris: A. Durand, Libraire, 1858). Nishapuri, Saljuqnama, ed. A. H. Morton (Chippenham: Gibb Memorial Trust, 2004). Rawandi, Rahat al-sudur, ed. Muhammad Iqbal (London: Brill, 1921). Sibt b. al-­Jawzi, Mir’at al-zaman fi tar’ikh al-a‘yan, partial edition, Ali Sevim (Ankara: Türk Tarih Kurumu Basimevi, 1968); Sibt b. al-­Jawzi, Mir’at al-zaman, I, ed. unknown (Hyderabad: Matba‘at majlis da’ira al-­ma‘ruf al-‘uthmaniyya, 1370/951). al-­Subki, Tabaqat al-shafi‘iyya al-kubra, ed. unidentified (Cairo: Al-­ Matba‘a al-­ Hasaniyya al-­misriyya, 1964–76). Secondary Sources Blair, Sheila, The Monumental Inscriptions from Early Islamic Iran and Transoxiana (Leiden: Brill, 1992). Bulliet, Richard, ‘Local politics in Eastern Iran under the Ghaznavids and Seljuks’ Iranian Studies 11, 1/4 (1978), pp. 35–56. Bulliet, Richard, The Patricians of Nishapur (Cambridge, MA: Harvard University Press, 1972). Durand-­Guédy, David, Iranian Elites and Turkish Rulers: A History of Isfahan in the Saljuq period (London: Routledge, 2010). Halm, Heinz, ‘Die Anfänge der Madrasa’, ZDMG, Suppl. III/1, XIX, Deutscher Orientalistentag vom 28. September bis 4 Oktober 1975 in Freiburg im Breisgau. Vorträge, ed. W. Voigt, Wiesbaden: Franz Steiner Verlag (1977), pp. 438–48. Hammer-­Purgstall, Joseph, Geschichte der Ilchane (Darmstadt: Verlag Leskr, 1842). Herzfeld, Ernst, ‘Eine Bauschrift von Nizam al-­ Mulk’, Der Islam XII (1921), pp. 98–101.

the ni za mi yya m a dr a sa s  | 301 Herzfeld, Ernst, Damascus: Studies in architecture, II, Ars Islamica X (1943), pp. 13–70. Hillenbrand, Robert, Islamic Architecture: Form, Function and Meaning (Edinburgh: Edinburgh University Press, 1994). Makdisi, George, ‘The madrasa as a charitable trust and the university as a corporation in the Middle Ages’, in Actes du Ve Congrès International d’Arabisants et d’Islamisants (Brussels: Publications du Centre pour L’Étude des Problèmes du Monde Musulman Contemporain, 1971), pp. 329–37. Safi, Omid, Politics of Knowledge in Pre-modern Islam: Negotiating Ideology and Religious inquiry (Chapel Hill: University of North Carolina Press, 2006). Talas, Asad, L’enseignement chez les Arabes. La Madrasa Nizamiyya et son histoire (Paris: Librairie Orientaliste Paul Geuthner, 1939). Tibawi, A. L., ‘Origin and Character of “al-­ madrasah’”, BSOAS 1/3 (1962), pp. 225–38. Van Renterghem, Vanessa, ‘Controlling and developing Baghdad: caliphs, sultans, and the balance of power in the Abbasid capital (mid-­5th/11th to late-­6th/12th c.)’, in The Seljuqs: Politics, Society and Culture, C. Lange and S. Mecit (eds) (Edinburgh: Edinburgh University Press, 2011), pp. 117–38.

21 The Life and Times of the Artuqid Ruler Najm al-­Din Alpı (r. 548/1154–572/1176)

Introduction

T

his article looks at the third ruler of the branch of the Turcoman Artuqid dynasty in Mardin and Mayyafariqin, Najm al-­Din Alpı, who was the lord of these two cities for twenty-­two years. An extensive trawl through the twelfth- and thirteenth-­century Muslim chronicles and biographical dictionaries in search of information about Najm al-­Din Alpı shows that he is viewed only as a peripheral player in events in the Jazira and further afield. Indeed, in the coverage of his long reign in the works of major medieval chroniclers such as Ibn al-‘Adim, Ibn al-­Athir and others, Alpı is mentioned only once or twice and then just briefly.1 Fortunately Najm al-­Din Alpı is given pride of place in the Ta’rikh Mayyafariqin wa-Amid, a chronicle centred on Mayyafariqin, written by Ibn al-­Azraq al-­Fariqi, an exact contemporary of Najm al-­Din Alpı and a seasoned traveller, who worked as an administrator for him and for his father Temürtash before him.2 In 1935 Cahen wrote a detailed summary of the events of Alpı’s rule, based on two extant manuscripts of Ibn al-­Azraq’s text in the British Library: Or. 6310 and Or. 5803.3 The edited Arabic texts of the two earlier sections of this work which concern the first two rulers of the Artuqid dynasty, Najm al-­Din İl-­Ghazi and Husam al-­Din Temürtash, were published by Savran and then by me on the basis of the two manuscripts in the British Library.4 I also provided an English translation and commentary on the text. However, the lengthy section of Ibn al-­Azraq’s Arabic text which covers the rule of Najm al-­Din Alpı and which was 302

t h e li f e a nd ti me s o f na j m a l - d in a l pı  | 303 included by Savran in his nowadays unfortunately rare book, still remains little known.5 Ibn al-­Azraq’s work provides detailed information on the life and times of Alpı and it will form the major focus of this paper. The occasional comments of the Oriental Christian sources written by Michael the Syrian and Bar Hebraeus6 will be analysed too. Monumental inscriptions and coins dating from Alpı’s rule also shed valuable light on the cultural, economic and religious life of the period. This paper will look first at the reign of Najm al-­Din Alpı within his own territories. This will include a survey not only of significant political events but also of cultural and economic achievement. It will then turn to his relations with other local Turcoman states in the Jazira, and, above all, with the Zengids of Iraq and Syria. The Rule of Najm al-­Din Alpı in Mardin and Mayyafariqin It will be convenient to begin with a few biographical details of Najm al-­ Din Alpı provided in Ibn al-­Azraq’s text. He was born in 520/1126–7.7 In 541/1146–7 he married in Mardin the daughter of Ahmad b. Suqman, the lord of Akhlat, and the sister of the Shah-­i Arman,8 thus forging links with a nearby Turcoman dynasty, as his predecessors had done. Ibn al-­Azraq’ s chronicle finishes in 570/1174–5, two years before Alpı’s death.9 In neither of the historical works of Ibn al-­Athir does he mention the date of this event, a mute commentary on the insignificance of this ruler on the wider canvas of twelfth-­century Muslim history. According to Ibn al-­Azraq, Alpı took over power peacefully without opposition on the death of his father Temürtash in Mardin on 2 Dhu’l-­ Qa‘da 548/19 January 1154.10 When the announcement of Alpı’s accession was made publicly in the khutba, no blood was shed and no property was plundered.11 On taking the reins of power, Alpı received the title of al-‘Adil.12 Following a familiar pattern in medieval Muslim historiography, the main focus of Ibn al-­Azraq’s account of Alpı’s reign concerns only the small ruling Muslim groups. From the beginning of his rule, it was essential for Alpı to obtain the support of his two brothers and of the grandees in Mardin and Mayyafariqin. To that end, he appointed one of his brothers, Jamal al-­ Din, as governor of Hani and the other brother, Shams al-­Din, as governor

304  |  the me d ie va l tur k s of Dara.13 As his father and grandfather had done, Alpı made alliances with the local urban Muslim elites to whom he offered military protection and security for their possessions in exchange for their administering the cities, and, above all, organising the collection of taxes. The administrative and religious elites (akabir al-dawla) in Mayyafariqin, and especially the Banu Nubata, who had been famous there since Hamdanid times, are frequently mentioned by Ibn al-­Azraq.14 Indeed, he points out that the office of qadi in Mayyafariqin had remained in the hands of the Banu Nubata since the year 490/1096 and that ‘nobody said the khutba in Mayyafariqin except them (the Banu Nubata)’.15 No doubt, too, the rousing sermons of the most famous member of this family, the preacher Ibn Nubata al-­Fariqi (d. 374/984–5), had not been forgotten.16 Ibn al-­Azraq also fills his chronicle with administrative minutiae, regularly providing lists of the names of officials, envoys, preachers and scholars who came and went in Mardin and Mayyafariqin during Alpı’s reign. Their activities are recorded without much comment.17 On one occasion only is there is a detailed anecdote on this theme; this concerns a chamberlain (hajib) called Shams al-­Din, who had displeased Alpı and whose arrest Alpı ordered in Ramadan 558/August–September 1163. Ibn al-­Azraq relates that Shams al-­Din was strangled in the citadel of Mayyafariqin on Alpı’s orders. Alpı then removed all his possessions from his house, including riding animals, silken cloth and treasures, and appropriated them; just what the unfortunate official had done is not mentioned.18 Ibn al-­ Azraq records the births, marriages and deaths of individual members of the Artuqid family, such as the birth of Alpı’s son and heir, Qutb al-­Din İl-­Ghazi, on 12 Muharram 544/22 May 1149.19 Considerable attention is given later to the wedding of Qutb al-­Din to the daughter of Qara Arslan (Alpı’s distant cousin in Hisn Kayfa) in Muharram 560/ November–December 1164. The celebrations lasted over four days. Qara Arslan and his entourage arrived with a vast trousseau. The lady of Akhlat laid out a great carpet20 for a banquet and she placed ceremonial garments on all the grandees present; as Ibn al-­Azraq writes, ‘none of the attendants or amirs or chamberlains were without ceremonial garments on them’.21 Ibn al-­Azraq mentions details of other significant events at Alpı’s centre of power, such as the signing of marriage contracts, the size of dowries, and

t h e li f e a nd ti me s o f na j m a l - d in a l pı  | 305 births and deaths. A moment of prestige for Alpı came in 555/1160, two years after the accession of al-­Mustanjid to the caliphate. A caliphal envoy arrived in Mardin bearing a robe of honour and a diploma legitimising Alpı’s rule there. Alpı donned the ceremonial garment and the diploma was read out publicly to great acclaim.22 It is probably unlikely that the recipient of this document would have understood much, if any, of the high Arabic in which it was couched. However, either Ibn al-­Azraq, with his frequent travels in the Arab world, and especially his stays in Baghdad,23 or one of the Banu Nubata would have been able to translate the lofty epistolary style of the caliphal diploma for his master. Najm al-­Din Alpı inherited a tradition of Muslim learning that had flourished in the Jazira for hundreds of years. By the twelfth century there were a number of libraries in Artuqid territory.24 According to Ibn Khallikan, Abu Nasr Ahmad al-­Manazi (d. 437/1045–6), the well-­known vizier, writer and poet, had a large collection of books and had endowed a waqf on them in the mosque at Mayyafariqin.25 Najm al-­Din Alpı commissioned an Arabic translation of the Syriac version of the De Materia Medica written by the Greek physician Dioscorides; this richly illustrated text, now housed in the Museum of the Imam Rida Shrine in Mashhad, was translated by Mihran b. Mansur in Mayyafariqin.26 Ibn al-­Azraq does not draw specific attention to the fact that the majority of the inhabitants in the two major cities that Alpı ruled, Mardin and Mayyafariqin, were Christians. References to these populations are fleeting and hard to evaluate. According to Michael the Syrian (d.  1199), Alpı’s father, Temürtash, had treated the Christians harshly at the beginning of his reign but at the moment of his death he recommended to his sons that they should be kind to their Christian subjects.27 The same Christian source also mentions that under Alpı’s rule in 567/1172 the Artuqid governor of Mardin took possession of the courtyard of the main Jacobite church there and annexed it to the nearby mosque.28 However, Bar Hebraeus (d. 1286) paints a more favourable picture, saying that Alpı heeded his father’s words, writing that at the time of Alpı’s death in Mardin in the year 572/1176 ‘the Christians, churches and monasteries were on very friendly terms with him’.29 The twelfth-­century Turcoman rulers of the Jazira were committed to a policy of building widely for the public good. The geography of the area

306  |  the me d ie va l tur k s made it easy for craftsmen to travel from one small centre of power to another in search of new patrons, although sometimes they went far further afield; Michell mentions that the signature of an architect from Hisn Kayfa, Ahmad b. Husayn al-­Hisnkayfi, dated 808/1406, has been found on a mosque in Firuzabad in the Deccan in southern India.30 As in the small states of his neighbouring Turcoman rulers, regular efforts were made to maintain the architectural infrastructure of the two cities ruled by Alpı. Whilst Ibn al-­ Azraq mentions a number of public buildings in the context of Alpı’s r­eign – ­mosques, madrasas, citadels, shrines, bazaars, khans, turbas, qubbas and the l­ike – i­t is not clear how many of them were founded or financed by the Artuqid ruler himself.31 It is certainly likely that, in accordance with the pattern established by other Seljuq successor-­states, and by his own father, as well as his cousins in Hisn Kayfa, Alpı would have wished to present himself as a model orthodox Sunni ruler. He was, for example, responsible for a major reconstruction of the Friday mosque in Mayyafariqin. Its minbar and arcades had collapsed in 547/1152–3, the last year of the reign of his father, Temürtash.32 According to the geographer ‘Izz al-­Din Ibn Shaddad, Alpı pulled down the rest of the building;33 Ibn al-­Azraq records that work on rebuilding the mosque was begun in 550/1155 under the supervision of al-­Zahid b. al-­Tawil and completed by the year 552/1157–8.34 Alpı also followed the example of his father in the matter of bridge building, and indeed some of the finest medieval bridges in all Anatolia were built with Artuqid patronage. The hog’s-­back Qaraman (now called Malabadi) bridge that Alpı completed in the first months of 548/1153–4 over the Satidama river (now known as the Batman-­Su), twenty-­two kilometres35 east of Mayyafariqin, still survives; its centrepiece is a spectacular arch, some nineteen metres in height. Ibn al-­Azraq’s description of it is unusually long; he writes as follows: Al-­Sa‘id Husam al-­Din Timurtash had begun the building of the Qaraman bridge over the river Satidama and had built most of it until all that remained to be done was the completion of the joining of the arch. Then he died. Najm al-­Din (Alpı) set about finishing and restoring it. The joining up of the bridge was finished and it (the span) was around sixty carpenters’ (najjar) cubits.36

t h e li f e a nd ti me s o f na j m a l - d in a l pı  | 307 Ibn al-­Azraq continues: ‘There is nothing like it in the world. He (Alpı) (may God be pleased with him) is the first to build a bridge in Diyar Bakr in this time.’37 Thus Ibn al-­Azraq exhibits considerable pride in his account of the bridge, and posterity has agreed with his laudatory judgement of it.38 It is not surprising that other rulers in the Jazira followed the example set by Alpı; his cousin and rival, Qara Arslan, built a bridge over the Tigris at Hisn Kayfa,39 and Qutb al-­Din Mawdud b. Zengi, the ruler of Mosul from 544 to 565/1149/50–1169/70 AD, constructed another fine bridge at Jazirat ibn ‘Umar.40 The Malabadi bridge must have played an important role in the local economy, for it is recorded by Ibn al-­Azraq that a wayside inn (funduq) was built next to it.41 This suggests that it was positioned on a trade route, and it is likely that alongside the funduq was a customs house to oversee the collection of tolls and tariffs paid by those using the bridge. This was the only bridge over the river for miles on either side, so travellers had little option but to use it. It made possible the movement of local trade, as, for example, from Bitlis to Diyar Bakr.42 The Copper Coins Minted in Alpı’s Time Further clues to the nature of Artuqid rule under Alpı are provided by the coinage that he minted. His father Temürtash had exploited a recently

Figure 21.1  A copper dirham of Najm al-Din Alpı, 1154–76 (Universität Tübingen, Islamische Numismatik)

308  |  the me d ie va l tur k s

Figure 21.2  Another copper dirham of Najm al-Din Alpı (British Museum)

discovered copper mine within his territory,43 but that was late in his life, in the year 542/1147–8.44 It is probably no coincidence that Alpı chose to mint his own coins only a year after the completion of the bridge.45 This development occurred in the wake of the local prestige he would no doubt have acquired through finishing the bridge and he may well have been advised by Ibn al-­Azraq himself to mint coins in order to promote trade using money. Indeed, Ibn al-­Azraq records specifically that Alpı minted copper coins, known as al-fulus al-Najmiyya, in Dhu’l-­Qa‘da 550/January 1156 after Ibn al-­Azraq himself had gone to al-­Ma‘dan,46 just as he had done for Temürtash in 542/1147–8, to buy the copper for the coins.47 Ibn al-­Azraq mentions that the people in Mayyafariqin then used the coins in their commercial transactions.48 Many fine numismatists have discussed these interesting coins, such as the ones illustrated here.49 Suffice it to say here that they are of varied design. They draw their inspiration from Hellenistic, Roman and Byzantine models. But quite apart from the remarkable range of designs on these hefty copper Artuqid coins, the fact they were minted at a time when copper was king (because silver was in short supply), ensured their currency even beyond Artuqid territory. And the name of the Artuqid ruler who minted them was spread accordingly. As for the reason for the consistent use of non-­Islamic coin types as models, factors that need to be taken into account to explain this radical departure from the Islamic norm include the presence of large num-

t h e li f e a nd ti me s o f na j m a l - d in a l pı  | 309 bers of Christians within Artuqid territory and also the fact that this territory was on the marches, on the very periphery of the Dar al-Islam, with solidly Christian kingdoms to the north. And Christians had long been accustomed to figural coins. Nor, it seems, was it a step too far to depict overtly Christian religious images on these coins. More locally, within his own territories the evidence of Alpı’s coins shows that he was alert to his official status, and that, like his father before him, his coins bear images redolent of authority and sovereignty taken from non-­Islamic and principally pre-­Islamic models. Was this a rather crude attempt to claim for himself an unbroken legitimacy stretching back to the remote past? Alpı also took care to recall the memory of his own family on his coins, the prime vehicle for his public image. He continued a practice instituted by Temürtash whereby the prime position, at top centre, on the reverse of these copper coins is allotted to the name İl-­Ghazi, followed by two titles of Temürtash, with Temürtash’s own name being banished to the outer right margin: this is a visual acknowledgement of the key importance of İl-­Ghazi, the most famous Artuqid ruler. The coins record that on occasion, too, Alpı silently dropped his father’s name in his own pedigree, calling himself Alpı b. İl-­Ghazi b. Artuq. Yet he adopted three of his father’s laqabs, namely malik al-umara’, al-malik al-‘alim and al-‘adil, while adding to them the territorial title malik Diyar Bakr and the kunya Abu’l-­Muzaffar.50 Alpı’s Monumental Inscriptions Fuller epigraphic evidence about Alpı’s public persona and titulature can be found on two extant Arabic inscriptions in Mayyafariqin. A short inscription in the name of Najm al-­Din Alpı is found on the band at the base of the dome of the Friday mosque in Mayyafariqin.51 It is written in fine naskhi script. It is undated but it is likely to have been placed on the dome at the time of the completion of the rebuilding of the Friday mosque in 552/1157–8, as mentioned earlier.52 This inscription opens with Sura 9, verse 18, which speaks to the Muslim faithful about who is allowed to look after the mosque. Grand titles are then given to the ruler Najm al-­Din Alpı who has ordered the work on the mosque. His full name is given as Alp Inanch Qutlugh Beg Abu’l-­Muzaffar Alpı b. Temürtash b. İl-­Ghazi b. Artuq. The list of titles accorded to him concern his role as a staunch supporter of the caliph

310  |  the me d ie va l tur k s and a vigorous mujahid against infidels and heretics, and he is described as the Khusraw of Iraq, Syria, Diyar Bakr and Armenia. A longer inscription, consisting of eight lines written in stiff Kufic, is located on two flagstones on the tower flanking the eastern gate of the city.53 It is dated 561 AH.54 It reads: In the name of God the Merciful, the Compassionate. Our master, the amir, the most lofty commander (isfahsalar), the triumphant, victorious, learned, just lord, the jihad warrior, Najm al-­Din, the honour of Islam, the one chosen by the Imam,55 the protector of mankind, the light and crown of the state, the splendour and ornament of religion, the pride and glory of the community, the pole of princes and sultans, the helper of those fighting jihad, the tamer of infidels and polytheists, the leader of the armies of the Muslims, the instrument of the caliphate, the prince of Diyar Bakr, the amir of Iraq, Syria and Armenia, the champion (pahlawan) of the frontiers, the sphere of noble qualities, Alp Inanch Qutlugh Beg Abu’l-­Muzaffar Alpı b. Temürtash b. İl-­Ghazi b. Artuq, the flame of the Commander of the Faithful, ordered this work in the months of the year 561.

Pride in Alpı’s Turcoman Artuqid genealogy is demonstrated here by the way in which his Turkish names are written in ­full – A ­ lp Inanch Qutlugh ­Beg – ­extending as far back as his great-­grandfather, Artuq, the founder of the dynasty. Alpı’s religious credentials are shown in the emphasis on his relationship with the Sunni caliph and on his role as a frontier jihad leader. The extent of his territories is once again exaggerated to include Iraq, Syria and Armenia. It is probable that the epigrapher was using model titles from neighbouring Iran and Syria, such as those used by the Seljuqs and the Zengids. The inscription is written in lofty Arabic epistolary prose with grandiose rhyming trios of ­phrases – ­as, for example, Sharaf al-Islam, Ikhtiyar al-Imam, Mujir al-anam, and Nur al-Dawla wa-tajuha, Baha’ almilla wa-zaynuha, Fakhr al-umma wa-majduha. The titles used for Najm al-­Din Alpı are couched in fulsome panegyrical style, in the mould of that used for much more celebrated Muslim contemporaries of Alpı in Syria, such as Nur al-­Din and Saladin. The addition of Persian titulature from the east – isfahsalar and pahlawan – add to the inflationary style of address used here to laud a minor Turcoman ruler in the Jazira. Both the inscriptions

t h e li f e a nd ti me s o f na j m a l - d in a l pı  | 311 discussed here are similar in the choice and range of titulature and the style used. The Relationships of Najm al-­Din Alpı with his Neighbours Alpı’s father, Temürtash, had stayed in power for thirty years, a remarkable feat, given the turbulence of the times in which he lived and the looming proximity of the ferocious figure of Zengi, first in Mosul and then in Syria. But Temürtash had known how to stay quietly in his place and to keep a low profile. The same approach was adopted by Alpı who, like his father, recognised what Lane-­Poole strikingly called ‘the machine that was eventually to grind them (the Artuqids) to powder’,56 namely the military might of Nur al-­Din b. Zengi and thereafter Saladin. Indeed, it was Alpı’s misfortune that his tenure of power coincided almost exactly with that of Nur al-­Din, the dominant figure in the politics of the Middle East in the third quarter of the twelfth century. So even if Alpı had had serious territorial ­ambitions – ­and there is little evidence to suggest that he did, despite the rhetoric of his ­inscriptions – h ­ e would have found himself outclassed, outnumbered and outmanoeuvred at every turn. Luckily for him, his own home ­territory – ­the cities of Mardin and M ­ ayyafariqin – ­and their adjoining ­territories – ­lay safely at the outer periphery of Nur al-­Din’s sphere of operations, and so the two polities did not clash. But Mosul was the crucial backstop for Nur al-­Din’s activities in Syria, and its security required Nur al-­Din to keep a watchful eye on the principalities immediately to the north, particularly the Shah-­i Arman and the Artuqids. Alpı was no doubt well advised to play it safe, avoid adventures and consolidate his local power base. Not for him the swashbuckling campaigns of his grandfather İl-­Ghazi that took this warrior, the real founder of the dynasty, to Aleppo, Damascus and Jerusalem and made him a major player in the politics of the early twelfth-­ century Middle East. Alpı kept a low profile, dug in at the nerve centres of Artuqid power, and thus played his part in ensuring the remarkable longevity of his dynasty, which lasted until the early fifteenth century in this region. A rare involvement on the part of Alpı in events further afield is given by Ibn al-­Azraq in the build-­up to the Battle of Harim fought between the forces of Nur al-­Din and the Franks on 22 Ramadan 559/12 August

312  |  the me d ie va l tur k s 1164. According to his account, on Wednesday 12 Jumada II of the year 559/6 May 1164, Qara Arslan came to the citadel of Mardin and met Najm al-­Din Alpı. After three days they decided to go to attack the lands of the Danishmendid ruler Ya’qub Arslan. Qara Arslan went back to Hisn Kayfa and met Alpı thereafter on the banks of the Tigris. They then proceeded to Khartbayt (Harput) where its ruler agreed to join them in the raid. They crossed the Euphrates to Malatya and plundered some of the land. When Ya’qub Arslan heard this news, he fled to the extremity of his territory, pursued by the Artuqid leaders as far as the neighbourhood of Siwas. At that  point a messenger appeared from Nur al-­Din57 calling them to wage ghaza with him and reproaching them with the following words: ‘The Franks have come out (to fight us) whilst you are Muslims fighting each other.’ The envoy then served as peacemaker between the warring Turcoman leaders. Qara Arslan went with his troops and retinue to Syria for ghaza whilst Najm al-­Din at the end of Sha‘ban went to his own territory.58 The principal Arab chroniclers for this period, Ibn al-­Athir, Sibt b. al-­Jawzi and Ibn al‘Adim, all mention that ‘Alpı sent a contingent of troops’ to Nur al-­Din, and they all use the same wording.59 The implication of this phrase is unambiguous; Alpı did not lead his troops in person, unlike his Artuqid cousin Qara Arslan, who did actually go to support Nur al-­Din. Clearly, Alpı was less swayed by the rhetoric of Nur al-­Din’s talk of jihad. However, as one of what the sources call the ashab al-­atraf, the petty rulers of the Turcoman border states of the Jazira, under de facto vassalage to the Zengids of Mosul and Syria, Alpı did not dare to ignore the imperious command from Nur al-­Din to send him reinforcements. At this point Ibn al-­Azraq provides supplementary information, implicitly perhaps justifying Alpı’s absence from the battlefield. He mentions that, instead of going with his troops to Syria, Alpı made his way to the citadel of al-­Shaybaniyya in the province of Mardin, where he fasted for the month of Ramadan because it was a place in which there was cool water from the meadows and pure air.60 Alpı’s absence is unlikely to have been caused by worries about fighting jihad in Ramadan since he would have learned that the Prophet Muhammad himself had achieved a famous victory in Ramadan against the Meccans at the Battle of Badr. Nor was Alpı opposed

t h e li f e a nd ti me s o f na j m a l - d in a l pı  | 313 on principle to fighting, since he is recorded in Ibn al-­Azraq’s text before and after Harim as conducting or joining in local raiding expeditions at intervals against his neighbours in the Jazira. For example, Alpı attacked Qara Arslan’s territory in 552/115761 and he besieged Dara in Jumada I 566/January 1171.62 Ibn al-­Azraq’s reference to fresh air and pure water at al-­Shaybaniyya might well suggest that Alpı was either wounded or sick and that he needed to recover his health in a suitable place at the height of summer. Given the laconic nature of Ibn al-­Azraq’s chronicle, it is interesting to note that he explicitly mentions twice that Alpı had fallen ill. On the first such occasion, in Safar 557/January 1162, Alpı’s son, Qutb al-­Din İl-­Ghazi, who was living in Akhlat with his maternal uncle, the Shah-­i Arman, was summoned to his father’s bedside. It is conceivable that Alpı was very sick since it was at this point that Qutb al-­Din was appointed officially as his heir apparent, no doubt to secure the succession. However, Ibn al-­Azraq then writes that ‘God cured him (Alpı) in Mardin’.63 The same phrasing is used later about Saladin by his biographer Ibn Shaddad, who mentions that ‘Saladin was sometimes ill but he would recover’64 or that ‘God cured him’.65 A more difficult and delicate situation occurred in 565/1169 when, to quote Ibn al-­Azraq: Atabeg Qutb al-­Din Mawdud, lord of Mosul (that is, the brother of Nur al-­Din) and his sons came to the citadel in Mardin to visit Alpı because he was ill and then he was cured. Ibn al-­Azraq goes on to say that these Zengid visitors stayed for days, and, to quote him again: ‘Alpı lavished hospitality on them and what was needed was not little.66

This event smacks of an overlord visiting a vassal, who was expected to organise a magnificent feast and thus to show obeisance. It is likely too that Qutb al-­Din Mawdud was keen to know how ill Alpı was and to check on who would be his successor. Conclusions Najm al-­Din Alpı’s rule in Mardin and Mayyafariqin shows how a small frontier Turcoman Seljuq successor-­state was run in the twelfth century. In

314  |  the me d ie va l tur k s spite of tough opposition, the Artuqids succeeded in governing parts of the Jazira for centuries. This paper has looked at the third ruler of the branch of the dynasty that ruled in Mardin and Mayyafariqin, Najm al-­Din Alpı, who was the lord of those two cities for almost a quarter of a century from 1153. His is not the first name that comes to mind when one considers the great personalities of the Muslim Middle Ages or even of the Artuqid dynasty alone. It may be that Alpı had an insipid, quietistic and even valetudinarian personality; at the very least, he was often sickly. Cahen has a very negative view of Alpı, arguing that under the combined effect of the strength of the Zengids and a temperament ‘peu capable et maladif du Prince’, Alpı’s rule was internally weak.67 There is certainly some truth in this statement. Indeed, by the time of Nur al-­Din, Zengi’s son, the Artuqids had learned their lesson, and they knew their place. When called upon, they provided troops for Nur al-­Din. It is noticeable that in crisis situations, when his help was urgently requested by a mighty fellow-­monarch such as Nur al-­Din for whom he had the status of de facto vassal, Alpı saw the wisdom of sending auxiliary troops, but it is likely that he did always not take the field himself. Nevertheless, Alpı managed to stay in power for twenty-­two years, no mean feat in troubled times, especially given the surge of Muslim jihad activity against the Franks during the time of Nur al-­Din. In much of his narrative about Alpı, Ibn al-­Azraq’s tone is panegyrical and full of bland platitudes about Alpı’s rule, echoing the tone of his accounts of the reigns of İl-­Ghazi and Temürtash, Alpı’s grandfather and father. It is regrettable that, because Ibn al-­Azraq’s chronicle ends in 570/1174, two years before the death of Alpı, there is no obituary notice of him like the one accorded to his father Temürtash in the same chronicle. As is the case with other Turcoman rulers of the later twelfth century in eastern Turkey and the Jazira, Alpı, controlling relatively small territories, played a quiet game of consolidation; he nurtured limited ambitions and was thus able to achieve some form of stability in Mardin and Mayyafariqin, to which his loyalty belonged. Without the chronicle of Ibn al-­Azraq, we would know even less about the rule of Alpı, about his entourage and his cultural aspirations. The evidence presented in this chapter has revealed that in the two cities ruled by Najm al-­Din Alpı there was a thriving social and cultural life. And the extant Artuqid coinage, the impressive mosque at Mayyafariqin (Silvan), and the

t h e li f e a nd ti me s o f na j m a l - d in a l pı  | 315 Malabadi bridge continue to speak for him. This was the heritage that they bequeathed to their successors. As an unnamed poet quoted by Hamdallah Mustawfi wrote: These are our works, and they declare us, Wherefore, after we are gone, look at our works.68

Notes   1. For example, Ibn al-‘Adim 1997, II, 493; Ibn al-­Athir 1399/1979, X, 302.   2. In his article on the earlier history of Mayyafariqin in the period of Hamdanid and Marwanid rule, Amedroz provides a long narrative about the career of Ibn al-­Azraq; Amedroz 1902, 785–7.   3. Cahen 1935, 254–63, 272–3.   4. Savran 1987; Hillenbrand 1990.   5. Savran 1987, 111–96; Alpı’s rule takes up around 43 per cent of Ibn al-­Azraq’s account of the Artuqid dynasty.   6. Michael the Syrian 1899–1914; Bar Hebraeus 1932.   7. Hillenbrand 1990, 50.   8. Hillenbrand 1990, 113.   9. Michael the Syrian gives a date of July 1176 for Alpı’s death; Michael the Syrian 1899–1914, 368. 10. Savran 1987, 111–13; see also Ibn al-­Athir 1963, 106; Sibt b. al-­Jawzi 1951, 218–19. 11. Savran 1987, 111–13; see also Ibn al-­Athir 1963, 106. 12. Savran 1987, 111. 13. Savran 1987, 109. 14. Ibn al-­Azraq, BL. Or. 5803, f. 113b; ‘Izz al-­Din Ibn Shaddad, Marsh 333, ff. 77a–78a; Hillenbrand 1990, 37, 39, 47, 110, 125. 15. Savran 1987, 118–19. 16. Hillenbrand 1999, 101–2. 17. For example, under the year 562; Savran 1987, 169. 18. Savran 1987, 144. 19. Savran 1987, 90. 20. A carpet on which a meal was placed (simat, whence the English word ‘samite’). Samite was a heavy silk fabric, much prized in medieval times. 21. Savran 1987, 152. 22. Savran 1987, 129.

316  |  the me d ie va l tur k s 23. Ibn al-­Azraq mentions that he studied religious texts in Baghdad with famous ‘ulama’; Hillenbrand 1990, 98–9. 24. Eche 1967, 201–2, 295. 25. Ibn Khallikan 1843–71, I, 77. 26. Grube points out that the mention of Najm al-­Din’s name in the preface to some extent makes up for the missing colophon; Grube 1959, 171, n. 39; also Ward 1985, 80. 27. Michael the Syrian 1899–1914, 311. 28. Michael the Syrian 1899–1914, 337. 29. Bar Hebraeus 1932, 307. According to Sinclair, Bishop John of Mardin (d. 1166) restored or built fifty churches or monasteries in and near Mardin; Sinclair 1989, III, 389. 30. Michell 1992, 34. 31. Alpı’s wife, Sitti Radawiyya, founded a madrasa named after her in Mardin in the second half of the twelfth century: Beyazit 2009, 83. 32. Hillenbrand 1990, 134–5; Cahen 1935, 272. 33. ‘Izz al-­Din Ibn Shaddad, Bodleian, MS Marsh 333, f. 104b. 34. Savran 1987, 121. Bell calls this mosque ‘among the finest of Mohammadan monuments’; Bell, 194, 159. She also discusses the inscription in the name of Najm al-­Din Alpı on this monument. 35. Sinclair 1989, III, 293. 36. Savran 1987, 115; Sibt b. al-­Jawzi 1951, 219. For the use of the term ‘najjar cubits’, see the description of Saladin’s attack on the wall the Castle of Sorrows at Jacob’s Ford in 1178: ‘It was nine cubits thick, not ordinary cubits neither, but carpenters’ cubits half as long again’: Lane-­Poole 2002, 124. 37. Savran 1987, 115. 38. Le Strange 1966, 111; Ilter 1976, 99. 39. Savran 1987, 115. 40. For a recent discussion of this bridge see Nicolle 2013, 223–64. Taylor mentions a number of other bridges built in Diyar Bakr: Taylor 1865, 25, 28, 33, 46, 49 and 51. 41. Savran 1987,153. 42. Sinclair 1989, III, 295. 43. Hillenbrand 1990, 120. 44. Savran 1987, 72: Hillenbrand 1990, 120. 45. Hillenbrand 1990, 114, n. 41.

t h e li f e a nd ti me s o f na j m a l - d in a l pı  | 317 46. Amedroz writes this place name as al-­Mu’dan, Amedroz 1902, 787; Cahen writes it as al-­Ma’din: Cahen 1935, 271; Whelan calls it ‘al-­Ma’den’; Whelan 2006, 76. Clearly the existence of the mine gave this place its name, whatever its pronunciation may have been. 47. Savran 1987, 79; Hillenbrand 1990, 197, n. 2. 48. Savran 1987, 122. 49. For discussion of Artuqid coins, on which the scholarly literature is very substantial, see, for example, Lane-­Poole 1875; Brown 1974; Lowick 1985; Väth 1987; Spengler/Sayles 1992; Whelan 2006. 50. For examples of the copper coins minted in Alpı’s time, see pls 1–2. 51. RCEA, 1937, no. 3351, 102–3; Gabriel 1940, no. 123, 343–4. 52. Cahen 1935, 273. 53. Sauvaget translated this inscription into French twice, with slightly different wording; RCEA 1937, no. 3272, 49–50; Gabriel 1940, no. 113, 339– 40. 54. 1165–6 CE. 55. Literally: ‘the choice of the Imam’. 56. Lane-­Poole 1875, 7. 57. The message was sent jointly from Nur al-­Din and his ally Zayn al-­Din ‘Ali Küçük: Savran 1987, 149. 58. End of Sha‘ban 559/July 1164. 59. Ibn al-­Athir 1963, 123; Sibt b. al-­Jawzi 1951/1380, 248–50; Ibn al-‘Adim 1997/1418, II, 493. 60. Savran 1987, 149. 61. Savran 1987, 126. 62. Savran 1987, 184. 63. Savran 1987, 136. Cahen speaks about the ‘enfeebled state of Alpı’s health’ but does not elaborate on this statement; Cahen 1935, 262. 64. Baha’ al-­Din Ibn Shaddad 2001, 27. 65. Baha’ al-­Din Ibn Shaddad 2001, 64. 66. Savran 1987, 170 67. Cahen 1935, 263. 68. Mustawfi 1919, 235.

318  |  the me d ie va l tur k s Bibliography Primary Sources Baha’ al-­Din Ibn Shaddad 2001: Baha’ al-­Din Ibn Shaddad, Al-nawadir al-sultaniyya wa’l-mahasin al-yusufiyya, tr. Donald S. Richards as The Rare and Excellent History of Saladin (Aldershot 2001). Bar Hebraeus 1932: Bar Hebraeus, The Chronography of Gregory Abu’l-Faraj, the Hebrew Physician, Commonly Known as Bar Hebraeus, facsimile ed. and tr. Ernest A. W. Budge, 2 vols. (London 1932). Ibn al-‘Adim 1997: Ibn al-‘Adim, Zubdat al-halab min ta’rikh Halab, II, Suhayl Zakkar (ed.) (Damascus 1997/1418). Ibn al-­ Athir 1399/1979: Ibn al-­ Athir, Al-Kamil fi’l-ta’rikh, vols X–XI, Carl J. Tornberg (ed.), reprint (Beirut 1399/1979). Ibn al-­ Athir 1963: Ibn al-­ Athir, Ta’rikh al-dawla al-atabakiyya, ‘Abd al-­Qadir Tulaymat (ed.) (Cairo 1963). Ibn al-­Azraq al-­Fariqi, Ta’rikh Mayyafariqin wa-Amid:   a) British Library MSS Or. 5803 and Or. 6310;   b) Savran 1987: Ibn al-­Azraq al-­Fariqi, Ta’rikh Mayyafariqin wa-Amid; A Critical Edition of the Artukid Section in Ta’rikh Mayyafariqin wa Amid by Ibnu’lAzrak al-Fariqi Ahmad b. Yusuf b. ‘Ali (510–578?/1116–1182), ed. Ahmet Savran (Erzurum 1987);   c) Hillenbrand 1990: Carole Hillenbrand, ed. and tr., A Muslim Principality in Crusader Times: The Early Artuqid State (Istanbul 1990). Ibn Khallikan 1843–71: Ibn Khallikan, Wafayat al-a‘yan wa anba’ ibna’ al-zaman, 2 vols (Cairo, 1299/1881–2;) tr. Baron MacGuckin de Slane as Kitab Wafayat al-A‘yan: Ibn Khallikan’s Biographical Dictionary, 4 vols (Paris 1843–71). Ibn Shaddad: Ibn Shaddad, ‘Izz al-­Din, Al-A‘laq al-khatira fi dhikr umara’ al-Sham wa’l-Jazira, Oxford, Bodleian Ms. Marsh 333 Michael the Syrian 1899–1914: Michael the Syrian, facsimile ed. and tr. Jean-­ Baptiste Chabot as Chronique de Michel le Syrien (Paris 1899–1914). Mustawfi 1919: Mustawfi, Hamdallah, tr. Guy Le Strange as The Geographical Part of the Nuzhat al-Qulub composed by Hamd-allah Mustawfi of Qazwin in 740 (1340) (Leiden and London 1919). RCEA 1937: Étienne Combe, Jean Sauvaget, Gaston Wiet (eds), Répertoire chronologique d’épigraphie arabe IX (Cairo 1937). Sibt b. al-­Jawzi 1951: Sibt b. al-­Jawzi, Mi’rat al-zaman fi ta’rikh al-a‘yan, editor unidentified, VIII/1 (Hyderabad 1951).

t h e li f e a nd ti me s o f na j m a l - d in a l pı  | 319 Secondary Sources Amedroz 1902: Henry F. Amedroz, ‘Three Arabic manuscripts on the history of the city of Mayyafarikin’, in Journal of the Royal Asiatic Society 34/04 (1902), 785–812. Bell 1914: Gertrude Bell, Palace and Mosque at Ukhaidir: A Study in Early Mohammadan Architecture (Oxford 1914). Beyazit 2009: Deniz Beyazit, Le décor architectural artuqide en pierre de Mardin placé dans son contexte régionale. Contribution à l’histoire du décor géométrique et végétal du Proche-Orient des XII-XVe siècles (unpublished doctoral thesis, Université de Paris I-­Panthéon-­Sorbonne, 2009). Cahen 1935: Claude Cahen, ‘Le Diyar Bakr au temps des premiers Urtukides’, in Journal Asiatique 227 (1935), 219–76. Doğangün/Ural 2007: Adem Doğangün and Ali Ural, ‘Characteristics of Anatolian stone arch bridges and a case study for Malabadi Bridge’, 5th International Conference on Arch Bridges (place of publication unknown, 2007), 179–86. Eche 1967: Youssef Eche, Les bibliothèques arabes publiques et semi-publiques en Mésopotamie, en Syrie et en Égypte au Moyen Âge (Damascus, 1967). Gabriel 1940: Albert Gabriel, Voyages archéologiques dans la Turquie orientale I (Paris 1940). Grube 1959: Ernst J. Grube, ‘Materialien zum Dioskurides Arabicus’, in Richard Ettinghausen (ed.), Aus der Welt der Islamischen Kunst. Festschrift für Ernst Kühnel zum 75. Geburtstag am 26.10.1957 (Berlin 1959), 163–94. Hillenbrand 1981: Carole Hillenbrand, ‘The establishment of Artuqid power in Diyar Bakr in the twelfth century’, in Studia Islamica 54 (1981), 129–53. Hillenbrand 1999: Carole Hillenbrand, The Crusades: Islamic Perspectives (Edinburgh 1999). Ilter 1976: Fügen Ilter, ‘Eine Gruppe der frühtürkischen Brücken in Südanatolien’, in IVème Congrès International d’Art Turc (Aix-­en-­Provence 1976), 97–104. Lane-­Poole 1875: Stanley Lane-­Poole, Coins of the Urtuki Turkumans (Foes of the Crusaders) (Closter 1875). Lane-­Poole 2002: Stanley Lane-­Poole, Saladin and the Fall of Jerusalem (London and Mechanicsburg 2002). Lowick 1985: Nicholas Lowick, ‘The religious, the royal and the popular in the figural coinage of the Jazira’, in Julian Raby (ed.), The Art of Syria and the Jazira 1100–1250 (Oxford 1985), 159–74. Michell 1992: George Michell, Firuzabad: Palace City of the Deccan (Oxford 1992).

320  |  the me d ie va l tur k s Nicolle 2013: David Nicolle, ‘The Zangid bridge of G ˘ azirat ibn ‘Umar (‘Ayn Diwar/ Cizre): a new look at the carved panel of an armoured horseman’, in Bulletin d’Études Orientales 62 (2013), 223–64. Sinclair 1989: Thomas Sinclair, Eastern Turkey: An Architectural and Archaeological Survey, vol. 3 (London 1989). Spengler/Sayles 1992: William F. Spengler, Wayne G. Sayles, Turkoman Figural Bronze Coins and Their Iconography (Lodi 1992). Taylor 1865: John G. Taylor, ‘Travels in Kurdistan, with notices of the sources of the eastern and western Tigris, and ancient ruins in their neighbourhood’, in The Journal of the Royal Geographical Society 35 (1865), 21–58. Väth 1987: Gerhard Väth, Die Geschichte der artuqidischen Fürstentümer in Syrien und der Ğaziral Furatiya (496–812/1102–1409) (Berlin 1987). Ward 1985: Rachel Ward, ‘Evidence for a school of painting at the Artuqid court’, in Julian Raby (ed.), The Art of Syria and the Jazira 1100–1250 (Oxford 1985), 69–83. Whelan 2006: Estelle J. Whelan, The Public Figure: Political Iconography in Medieval Mesopotamia (London 2006).

22 What Is Special about Seljuq History?

Introduction

T

he arrival of the nomadic Turks in the Islamic world was a pivotal moment in medieval Iranian and Islamic history. From the 1020s onwards the eastern Islamic world experienced a wave of nomadic Turkish invasions from the Central Asian steppes. The Seljuq state which was then established marked the beginning of a Turkish military and political dominance in the Muslim world which was to last for many centuries (Figure 22.1).1 The Muslim Arabs and Persians had become acquainted with one kind of Turkish military presence since the eighth century. Turkish slave soldiers (mamluks), much admired for their skills in horsemanship and archery, had long formed part of the armies and bodyguards of caliphs and independent rulers (Figure 22.2). But that Turkish component in Islamic society had entrenched itself slowly and did not burst onto the scene with the sudden intensity, scale and alien character of the Seljuq invasions. What Kind of Nomadic Society Did the Seljuq Turks Bring with Them into the Sedentarised Eastern Islamic World? The nomadic Turks (Turcomans) had endured unremittingly hard lives for millennia. Their most striking characteristic was mobility. This was the key to their survival. It was by their mobility that they could attack the enemy, plunder and then depart at great speed. Their mobility was not confined to warriors alone; it entailed the movement of men, women, livestock, tents and baggage wagons. Their military strength lay in their mounted archers, who 321

322  |  the me d ie va l tur k s

Figure 22.1  Map of the Seljuq empire (Courtesy of the Metropolitan Museum of Art, New York (Court and Cosmos))

w h at i s sp eci a l a bo ut sel j uq h is to r y ?  | 323

324  |  the me d ie va l tur k s

Figure 22.2  ‘Turkish cavalry’, Rashid al-Din’s Jami‘ al-tawarikh (Edinburgh University Library Or. Ms. fol. 129v)

discharged their arrows at a distance and avoided direct encounters with the enemy (Figure 22.2). Already in the ninth century the Arabic writer al-­Jahiz praised the Turkish cavalry, saying: ‘The Turk has two pairs of eyes, one at the front and the other at the back of his head.’2 The nomads’ dress contrasted forcibly with that of the settled Muslims (Figure 22.3). The Turcomans wore furs, or shapeless sack-­like cloth garments. Both sexes wore trousers. The men commonly wore their hair down to their waists and greased it with rancid butter. Indeed, it was rumoured amongst Eastern Christians that the Seljuq army could be smelt three days’ distance away. The basis of the nomadic economy was a combination of pastoralism and raiding. The Turcomans’ major food supply came from their herds. Their wealth and status were usually assessed by how many horses they owned. The sedentary Muslim populations were disgusted when they saw what the Turcomans ate: wolves, foxes, dogs, mice, rats and snakes. The eleventh-­century Arab writer Ibn Hassul says: ‘They consume only meat, and do not wish for any substitute, even if it is dripping blood or filthy, and they do not wish for anything else.’3 Turcoman society was shamanistic, and even after the conversion of their tribal leaders to Sunni Islam the Turcomans retained many of their basic shamanistic practices. Power within the tribe

w h at i s sp eci a l a bo ut sel j uq h is to r y ?

|

325

Figure 22.3 Demon in Chains, in Siyah Qalam style, Iran or Central Asia, c. 1453 (The Cleveland Museum of Art)

lay with certain families. On the death of a senior male member of the ruling clan, his patrimony would be shared out amongst his male relatives. A Brief Overview of Seljuq History In the 990s Turcoman groups from Inner Asia moved towards the Muslim frontier near the Aral Sea. The chief of the Oghuz clan, Seljuq, converted to Islam. Some of the Seljuq Oghuz nomadic tribes crossed the river Oxus into Muslim territory in 1025, and the remainder did so ten years later. In 1036–37, under the leadership of two Seljuq brothers, Tughril and Chagri, these nomadic forces conquered eastern Iran. Probably now numbering around 4,000 tents, they were well and truly inside the Muslim world. This was a substantial number of people. Such tents were large, housing extended families and often their animals, too (Figure 22.4). At that point Chaghri decided to divide forces; he would stay in eastern Iran, while Tughril would move west and keep whatever territory he conquered. In 1043 Tughril seized Isfahan and in 1055 entered Baghdad; he had himself proclaimed ‘Sultan of east and west’4 by the Sunni caliph there and then persuaded the reluctant caliph to allow him to marry his daughter.

326  |  the me d ie va l tur k s

Figure 22.4  ‘Court tents’, Rashid al-Din’s Jami‘ al- tawarikh (Edinburgh University Library Or. Ms. fol. 125v)

Tughril was not destined to enjoy supreme power for long. Having definitively taken control of Baghdad, he died in 1063. The reigns of the next two Seljuq sultans, Alp Arslan and his son Malikshah (Figure 22.5), represent the high point of Seljuq rule in the east. These two sultans were ably supported, and indeed dominated, by the superbly gifted Persian vizier Nizam al-­Mulk. Territorially, the Seljuq period saw the first major expansion of the Islamic empire since the seventh-­century Arab conquests. Alp Arslan was constantly on the move and created a unified empire that stretched from Syria and Anatolia to Central Asia. In 1071 Alp Arslan moved to confront a huge Byzantine army near Lake Van. He won a famous victory, the battle of Manzikert;5 he actually captured the Byzantine emperor and then released him honourably. News of this extraordinary military victory reached Europe, causing unease at the growing Turkish power in Anatolia. Although the Turcomans had roamed across Anatolia as early as the 1030s, Manzikert has been viewed by many historians as the moment when the Turcomans were there to stay (Figure 22.6). Alp Arslan was murdered in 1072 in the east of his empire.

w h at i s sp eci a l a bo ut sel j uq h is to r y ?  | 327

Figure 22.5  ‘Malikshah’, Rashid al-Din’s Jami‘ al-tawarikh (Edinburgh University Library Or. Ms. fol. 138r)

His son Malikshah succeeded him as Seljuq sultan (Figure 22.5). The famous thirteenth-­century Arab historian Ibn al-­Athir praised his reign, saying: ‘His name was mentioned in sermons preached from the borders of China to the limits of Syria and from the extremities of the lands of Islam in the north to the bottom of Yemen.’6 Malikshah continued to go on military campaigns, but his twenty-­year reign represents an important transition in Seljuq history, for he seems to have stayed in one place for certain months of the year. That place was Isfahan, which he made his capital.7 But it would

328  |  the me d ie va l tur k s

Figure 22.6  Modern Turkish Manzikert celebratory stamp, 1971 (collection of the author)

appear that he did not live in a palace or a citadel. It is probable that he still lived in a tented encampment outside the walls of the city. His vizier Nizam al-­Mulk, however, was resident inside the city, conducting government business. Both men were killed within weeks of each other in 1092, and the Seljuq state was never the same again, although the dynasty lasted for another century in Iran. In the twelfth century, unassimilated Turcomans on the fringes of the Seljuq empire moved from summer to winter pastures, as they had always done. Closer to the centres of Seljuq power in Isfahan and Nishapur the unity of the state was increasingly undermined by fratricidal succession disputes and by the lingering traditional patrimonial division of heritage amongst the ruler’s male heirs on his death. Sultan Sanjar stayed in power in eastern Iran, ruling there from 1097 and serving as supreme sultan from 1118 to 1157. But his death ushered in the full decline of the Seljuq state in Iran (Figure 22.7). The last Seljuq sultan, Tughril III, died in 1194. The Realities of Early Turkish Rule The presence of the Turcomans was viewed as a necessary evil by the sedentary Muslim populations. The Seljuq leadership exercised control over them,

w h at i s sp eci a l a bo ut sel j uq h is to r y ?  | 329

Figure 22.7  Mausoleum of the Seljuq sultan Sanjar (photograph by Bernard O’Kane)

sometimes precariously. The Seljuqs also came to an understanding with the urban elites; the Seljuqs would provide military support, and the cities would give taxes and the bureaucratic expertise necessary to administer the Seljuq empire. Despite the disruption to agriculture caused by the Turcomans and their tendency to damage cities and monuments, Muslim writers valiantly made the best of the situation, for they saw that these nomads were here to stay.

330  |  the me d ie va l tur k s Seljuq Government Continuity of governmental practice was achieved in the Seljuq period; Persian elite families passed on their skills from one generation to the next and helped to ensure administrative stability in turbulent times. The chief architect of Seljuq administration was Nizam al-­Mulk, probably the most famous of all Persian viziers. During his extraordinarily long period in power (over thirty years), he worked in a travelling court. Wherever the Turkish Seljuq sultan was, Nizam al-­Mulk was there, too. He needed to be there with a restraining hand, teaching him the ways of Perso-­Islamic government.8 There were, however, new aspects of Seljuq government that came from the shamanistic steppe heritage. The first two Seljuq sultans retained their Turkish totemistic ­names – ­Tughril meaning ‘falcon’ and Alp Arslan meaning ‘hero lion’. Thereafter, most (but not all) subsequent Seljuq sultans were known by Muslim names. Even so, the title ‘Sultan of east and west’ bestowed by the ‘Abbasid caliph in Baghdad on the first Seljuq Sultan Tughril in 1055 reflects the nomadic steppe concept of world ­dominion – ­a symbol of Turkish sovereignty, extending from where the sun rises to where it sets. The earliest Seljuq coins show a bow and arrow to denote sovereignty. According to Ibn al-­Athir, the Byzantine emperor repaired the mosque in Constantinople in honour of the first Seljuq sultan Tughril, and in the mihrab he placed a bow and arrow.9 The Seljuq Turkish institution of atabeg, usually denoting a powerful Turkish military commander whose duty it was to supervise the upbringing of young Seljuq princes, also came from steppe tradition. But when it was introduced into the Seljuq government, it had dire consequences. The institution proved to be a serious centrifugal force, working against the centralised government model beloved of Nizam al-­Mulk. After 1092 atabegs seized power in their own right, and this contributed to the disintegration of the Seljuq state. What of Religion? Traditionally, the Seljuq Turkish elite are presented in the medieval Arabic and Persian sources as pious Sunni Muslims who had converted to Islam before their major invasions began. This portrayal of the Turks, who went

w h at i s sp eci a l a bo ut sel j uq h is to r y ?  | 331 on to govern the heartlands of the Islamic world for many centuries, is not surprising. Given the political realities, their Arab and Persian bureaucrats, court chroniclers and poets needed to put on a brave face and help forge an alliance between the military force of the Turks and the long-­established prestige of the Sunni Arab and Persian religious classes. No doubt, over time some of the Turkish sultans did indeed grow into the role accorded them by their chroniclers and court p ­ oets – ­namely, that of defenders of Sunni Islam and fighters of jihad against infidels and heretics. As an alien, invading military force, the Seljuq Turkish sultans generally adopted a publicly supportive and deferential stance towards the Sunni caliph in Baghdad, from whom they sought religious credentials. And the caliphs gave them grand robes and honorific titles. Both internally and externally, the Seljuq government took decisive steps against the Isma‘ili Shi‘ites, who were cursed from the pulpits and forbidden employment. A major religious development was the establishment of a network of Sunni madrasas by high-­ ranking Seljuq administrators across the empire. Nizam al-­Mulk set the example by funding and supervising at least ten such madrasas or religio-­legal colleges; the principal one of these was in Baghdad. Thus, a reliable, well-­ trained class of Sunni religious scholars could promote orthodox Islam across the empire. What of the Seljuq Court? Two traditions blend in the evolution of the Seljuq court ceremonial: the Perso-­Islamic heritage and the Turkish tradition of the Central Asian steppes. It is hard to assess in detail how this worked for the Seljuqs; the medieval Islamic sources present the Seljuqs’ lifestyle through the distorting mirror of the Muslim religious and bureaucratic elite. The Seljuq Turks have no real voice of their own. However, despite being military warlords and ‘Turkish nomadic outsiders’, the Seljuq sultans decked themselves in the usual trappings of Perso-­ Islamic rulers. Their courts employed many officials and servants. The most important of these was the chamberlain who controlled access to the sultan’s presence. The court also included a chief executioner, the holder of the royal parasol, scribes, religious scholars, Sufis, doctors, artists, interpreters, jesters, astrologers, cooks and many others. It was an intrinsic sign of a ruler’s

332  |  the me d ie va l tur k s prestige for him to have an entourage of public intellectuals, despite his being unschooled, even illiterate. A veritable galaxy of Persian and Arab poets flocked to the Seljuq courts. The job of the cooks must have been very onerous. Indeed, it was said that Sultan Alp Arslan used to slaughter fifty sheep a day. Doctors were essential since there was always the danger of infected water on campaign. The Seljuq sultan had many regular formal duties. He received envoys from Muslim and non-­Muslim lands. He inspected his troops. He held circumcision celebrations for his male relatives. And he gave banquets to consolidate his relationships with tribal chiefs on whose auxiliary military support he relied. When the court went outside the royal enclosure, the ceremonial parasol would be held over the sultan’s head. The giving and receiving of gifts were important facets of court ceremonial. The first Seljuq sultan, Tughril, received a present from a vassal that included multi-­coloured garments, horses and three well-­nourished sparrow hawks. Exotic animals, such as the giraffe, were also welcome gifts and would be kept in a menagerie. Malikshah is singled out by Ibn Khallikan in his famous biographical dictionary for his love of the hunt: he writes that ‘all power was concentrated in the hands of the vizier, whilst the sultan (Malikshah) had nothing more to do than show himself on the throne and enjoy the pleasures of the chase’.10 Indeed, Malikshah is reported to have shot seventy gazelles in one day; in each hunting place he made towers of the hooves of gazelles and wild asses. The sultans also liked to listen to music, to watch dancing, to play chess and backgammon, and to hear tales of exotic lands. Wine-­drinking was an integral part of Seljuq court life, and some Seljuq rulers, such as Malikshah’s son Barkyaruq, are described in the sources as being addicted to wine. The women of the Seljuq ruling family were involved in court politics, and they acted as patrons of religion and learning. Many of them had their own viziers. The power of queen-­mothers increased further after the fragmentation of Seljuq territories, and some ruled after the death of their husbands. A famous example of this was Terken Khatun, Malikshah’s wife, after his death in 1092. She was reported to have owned 10,000 mamluks. She ruled the kingdom and led the armies. No wonder Nizam al-­Mulk in his famous Book of Government railed against royal women who meddled in politics: ‘In

w h at i s sp eci a l a bo ut sel j uq h is to r y ?  | 333 all ages nothing but disgrace, infamy, discord and corruption have resulted when kings have been dominated by their wives.’11 Overall, seeing nomadic Turkish chieftains sitting on thrones and wearing crowns must have been a novel and not altogether enjoyable experience for their Persian and Arab subjects, long used to seeing ethnic Turks only as slave bodyguards, military commanders and eunuchs guarding the harem. The Seljuqs of Rum (Anatolia) The year of 1194 was not the end of the Seljuq Turkish presence in the Middle East. The initial disruptive Turcoman presence in eleventh-­century Anatolia had been followed by the establishment there of several embryonic Seljuq successor states, such as the Danishmendids and the Artuqids, which governed in a manner similar to that of the Great Seljuqs. Above all, the Seljuq state of Rum, founded in 1077 by Malikshah’s cousin Sulayman, survived until 1307. The Rum sultans made Konya their capital. They were Turks. But, unlike their Great Seljuq predecessors, many of them bore Persian names such as Kaykhusraw, Kayqubadh and Kayka’us. Moreover, their administrators and the religious elite in their cities were Persian. And with the terrifying thirteenth-­century Mongol invasions of the eastern Islamic world, a mighty torrent of Persian refugees poured into Anatolia, particularly Konya. The Rum Seljuq s­tate – ­with a diverse population of Byzantine Greeks, Armenians, Kurds, Turcomans and P ­ ersians – ­was especially successful in the period from 1220 to 1250 and laid the foundations for the subsequent Islamisation of Anatolia. Indeed, Seljuq Anatolia houses the earliest substantial single body of medieval madrasas in the whole Muslim world; thirteen of these are in Konya. There were also Sufi cloisters in Anatolian towns, especially in Konya, where such famous figures as Jalal al-­Din Rumi could reside and teach. Concluding Reflections Empires built by usurping nomadic warlords are notoriously ephemeral, and nomadic traditions sit lightly on the civilisations they conquer. As with the nomadic Mongols in Iran and China, it was to be expected that indigenous Perso-­Islamic statecraft would eventually prevail over the incoming Seljuq

334  |  the me d ie va l tur k s Turks. But the military power of the Turks was an essential ingredient, too. Thus, in the Seljuq period a fascinating symbiosis of ancient Persian culture, Sunni Islamic norms and Turkish steppe tradition came into being. By the thirteenth century, there was a solid Muslim presence in Byzantine Anatolia; this had begun with Turcoman raids from the eleventh century onwards and had been put on a firmer footing by the Seljuq state of Konya. Byzantine Christian Anatolia further west was now set to be Islamised under Turkish leadership. The Turkish conquest of Constantinople in 1453 was the culmination of this process. The land now known as Turkey was soon to be born. Notes   1. For an overview of Seljuq history see Peacock 2015 and Bosworth 1968.   2. Pellat 1969, 93.   3. Ibn Hassul 2015, 77.   4. This was a new regnal title for the Muslim world, but an ancient one for the Central Asian steppe peoples.   5. For a detailed discussion of this pivotal battle, see Hillenbrand 2007.   6. Ibn al-­Athir 1965–7, vol. 10, 211.   7. Durand-­Guédy 2010.   8. Nizam al-­Mulk wrote a famous Mirror for Princes on good Perso-­Islamic government: Nizam al-­Mulk 1960.  9. Ibn al-­Athir 1965–7, vol. 10, 28, explicitly states that the Umayyad general Maslama (d. 738 AD) had built the mosque. Ibn al-­Athir is probably recycling a folkloric memory, for Maslama was never able to enter the city. 10. Ibn Khallikan 1843–71, vol. 1, 413. 11. Nizam al-­Mulk 1960, 179–80.

Bibliography Bosworth, C. E., ‘The political and dynastic history of the Iranian world (ad 1000– 1217)’, in J. A. Boyle (ed.), The Cambridge History of Iran, Vol. 5, The Saljuq and Mongol Periods (Cambridge, 1968), 1–202. Durand-­Guédy, D., Iranian Elites and Turkish Leaders: A History of Isfahan in the Saljuq Period (London, 2010). Hillenbrand, C., ‘Sources in Arabic’, in M. Whitby (ed.), Proceedings of the British Academy (London, 2007), 283–340.

w h at i s sp eci a l a bo ut sel j uq h is to r y ?  | 335 Ibn al-­Athir, Al-Kamil fi’l-ta’rikh, ed. C. J. Tornberg (Beirut, 1965–7). Ibn Hassul, ‘The superiority of the Turks over other regiments’, in The Turkic Peoples in Medieval Arabic Writings, tr. Y. Frenkel (London, 2015), 68–82. Ibn Khallikan, Wafayat al-a‘yan, tr. W. M de Slane as Ibn Khallikan’s Biographical Dictionary, 4 vols (Beirut, 1970). Nizam al-­Mulk, The Book of Government, or Rules for Kings: the Siyasat-nama or Siyar al-Muluk of Nizam al-Mulk, tr. H. Darke (London, 1960). Peacock, A., The Great Seljuk Empire (London, 2015). Pellat, C., The Life and Works of Jahiz: Translations of Selected Texts (Berkeley, 1969).

23 ‘The View from Above’: Muslim Perceptions of the Turks of Syria and the Jazira in the Period 1070–1176 Introduction

I

n this paper an attempt will be made to examine the interaction of the Muslims of Syria and the Jazira with the incoming Turks in the eleventh and twelfth centuries. The discussion will focus especially on the period from the 1070s until the accession of Nur al-­Din in the middle of the twelfth

Figure 23.1  Conversation among Turkish nomads (after İpşiroğlu, Bozkir Rüzgâri Siyah Kalem (Istanbul, 1985), pl. 18)

336

‘the v ie w f ro m a bo ve ’  | 337 century. The discussion will embrace the religious beliefs, the traditions and, above all, the conduct of the Turks in this time and place. The Historical Context Much is known and much has been written about the impact of the Turks in the Muslim world, first in ninth-­century Iraq when they served as the unruly praetorians of the increasingly enfeebled ‘Abbasid caliphs, then when they invaded the eastern Islamic world, founding states serviced principally by Persian officials, and finally when they set up long-­standing empires in the later Middle Ages, ruled first by the Mamluks of Egypt and Syria, and then by the Ottomans with their base in Istanbul, spanning two continents and soon expanding into three. That, beyond doubt, was the finest hour of the Turks, and the Ottoman empire lasted until less than a century ago. Such, in very simplified form, is the chronological perspective, and it is familiar to every student of Islamic history.1 The arrival of the nomadic Turks in the Islamic world had been a pivotal moment in medieval Iranian and indeed Islamic history. From the 1020s onwards, the eastern Islamic world experienced a wave of nomadic Turkish invasions from the Central Asian steppes. The Seljuq state was established soon after; it marked the beginning of a Turkish military and political dominance in the Muslim world which was to last for many centuries. Indeed, as Marshall Hodgson rightly stressed, ‘The Turks formed the most important military classes of the Muslim regions wherever the central Eurasian steppe was reasonably accessible; that is, from the Nile valley to the Afghan mountains.’2 The Sunni Seljuq sultans ruling Iran and Iraq valued Syria as a march province between themselves and the Shi‘ite Fatimid caliphs of Egypt. Some key parts of Syria fell under Seljuq domination in the 1070s. In the 1080s Damascus and Aleppo were the major centres of Seljuq power. Fatimid control over Syria was threatened and reduced by Turcoman military barons sent there by the Seljuq sultans, and also by the infiltration of nomadic Turcoman leaders, their tribesmen and their flocks. The Syrian coast south of Tripoli, however, remained in Fatimid hands. But the Levant remained prey to several conflicting factions. So the years immediately preceding the Crusader capture of Jerusalem in 492/1099 were a period of great turmoil

338  |  the me d ie va l tur k s for the Muslim rulers of Syria and Palestine. As is well known, their disunity and mutual rivalries eased the way for the Crusaders, who were able to establish four states in the region by 502/1109. The majority of the Muslim leaders during this key period of Crusader success were Turkish military commanders who had accompanied the Seljuq royal prince, Tutush, and his young sons, Duqaq and Ridwan, to Syria in the power vacuum left further east by the death of the brother of Tutush, the Seljuq sultan Malikshah, in 485/1092. Traditionally, the Turks, many of whom had crossed into Muslim territory under the leadership of the Seljuq family, are presented in the medieval Arabic and Persian sources as pious Sunni Muslims who had converted to Islam before their major invasions began. This portrayal of the Turks, who went on to govern the heartlands of the Dar al-Islam for many centuries, is not surprising. Their Arab and Persian bureaucrats, court chroniclers and poets needed to put a brave face on political realities and to help to forge a much-­needed alliance between the military force of the Turks and the long-­established prestige of the Sunni Arab and Persian religious classes. The stability of the realm depended on that alliance, even though it was essentially a ‘marriage of convenience’.3 No doubt, over time some of the Turkish sultans did indeed grow into the role given them by their biographers, court poets and c­hroniclers – ­namely that of defenders of Sunni Islam and fighters of jihad against the infidel. But alongside the panegyrics accorded them by such writers, there was always a tendency on the part of some of them, such as the Arab aristocrat Usama b. Munqidh (d. 1188) and the chronicler Ibn al-‘Adim (d. 1262), to look down on and sneer at the mores of the alien Turkish newcomers, only a few generations removed from the nomadic life of the steppes, and to present them as a violent, drunken, barbarian ‘Other’. Yet even familiarity and enforced coexistence with the Turkish-­speaking ­newcomers – m ­ ilitary commanders, their private armies and their Turcoman ­auxiliaries – ­did not create among the Arabs any real understanding or analysis of the nomadic background, traditions and shamanistic beliefs and rituals of the Turks. So those stories in the Muslim sources which shed light on the continuing existence of such vestiges of the pre-­Islamic Turkish past remain unexplained in ethnographical terms. These stories come across in a

‘the v ie w f ro m a bo ve ’  | 339 different light altogether. They are used as religious and cultural ammunition against the Turks whose strange antics are savoured and mocked by a more sophisticated Arab and Persian readership. Thus, the conquered can get their own back on the conquerors. As for the descriptions of the Turks produced by Crusader chroniclers, they too devote regular attention to this topic, often giving detailed descriptions of Turkish military tactics.4 Such historians, unlike their Arab or Persian counterparts, are, of course, not held back on occasion for reasons of political expediency from writing what they wish to say about the Turks. So Crusader chroniclers are free to include, writing in a language inaccessible to the enemy, less favourable and more sensational stories about the Turks. Such anecdotes are intended to fascinate and shock their countrymen back in Europe and to reinforce religious and cultural stereotypes.5 The Problematic Nature of the Medieval Arabic and Persian Sources about the Turks in Syria in the Eleventh and Twelfth Centuries The Turks themselves in this period are for the most part seen through the lens of the Arab and Persian writers who recount their arrival in and impact on the Muslim world. Some of this history also figures in the accounts of Eastern Christians and occasionally Crusader writers. Knowledge of this subject is inevitably fragmentary, and what has survived is hard to decode; in pre-­modern times there was little archival material, and few diaries or collections of private papers. There are many unanswered questions in the Arabic and Persian literary sources about the relationship between Islamic faith and practice and the religious beliefs and customs of the Turkish newcomers. The spirit of anthropological enquiry is absent (except with much earlier Muslim geographical writers such as Ibn Faqih (fl. 902));6 there is little curiosity, and no interest in the social structure or religious beliefs of the Turks. The Arabic texts, in short, reveal a lack of sympathy, an inability on the part of the writer to put himself in the place of the Turks. Muslim writers take refuge in broad ­generalisations – ­for example, condescension about the dull intellect of the Turks, respect for their military p ­ rowess – b­ ut the underlying reasons for these comments do not emerge clearly. The sources express disapproval about the excesses of the ­Turks – ­for example, their drunken sprees after victories in battle. Important

340  |  the me d ie va l tur k s issues such as the Turkish style of rule, their political institutions, such as the atabegate, their succession policies or the role of women are not addressed. Al-­Ghazali viewed the Turkish military presence as a necessary evil, whilst the most famous vizier of the Seljuq state, Nizam al-­Mulk, who knew the Turks extremely well and from the inside, maintains a prudent silence about them in his ‘Mirror for Princes’ work, the Siyasatnama. Verisimilitude is not a priority for medieval Muslim historians. In the interests of their overall ­aim – t­ o demonstrate that all history is part of God’s overarching purpose that the whole world should embrace ­Islam – t­hey exploit rhetorical devices, poetry, speeches and Qur’anic quotations to the full. They quote high Arabic epistolary prose allegedly composed by major Muslim leaders of the Crusading p ­ eriod – m ­ en who were in reality Turkish military commanders whose grasp of Arabic must have been extremely rudimentary. What matters here is not the implausibility of this alleged authorship but the ideological message being imparted by such letters and speeches. Rather than focusing on the process and extent of the Islamisation of the incoming Turks, Muslim sources prefer simply to recount events. Warfare looms large in these accounts. Moreover, Muslim authors always had to bear in mind the interests and prejudices of the patrons who employed them. So serious constraints limited their freedom of expression. Indeed, they focused principally on the role of the Seljuq sultans and other Turkish rulers as promoters of Sunni Islam. Little is said about the relationships between Turkish leaders and the Turcomans, although, as Andrew Peacock has rightly said, ‘At the heart of Seljuq history lie the tribes, who are the key to understanding the early development of the Seljuq empire.’7 There is one further element that obscures a clear understanding of how the Turks (in their various incarnations) interacted with the local Muslim and Christian population in Syria in the eleventh and twelfth centuries, and that is the state of well-­nigh constant warfare into which the Levant was plunged at this time. The chronicles make dispiriting reading in this respect. Events are reduced to a sequence of campaigns and battles. Below the conflict between major dynasties for control of Greater Syria, that is first between the Fatimids and the Seljuqs, and then between the Crusaders and the Turkish military barons such as Zengi and Nur al-­Din, there raged endless strife at a lower and local level. The same pattern characterised the Jazira. All this was

‘the v ie w f ro m a bo ve ’  | 341 exemplified by the violent rivalry between Seljuq princes (such as Ridwan and Duqaq) or between them and their governors (for example, between Tutush and Atsiz) or between the various Turkish atabegs (the so-­called tutors of Seljuq princes), generals and amirs (such as Tughtegin and Aq Sunqur), each ensconced in his own patch of territory and ambitious to seize that of his nearest rival. A further complicating factor was the regular irruption of nomadic Turcoman groups, who continued to follow their traditional way of life which depended on seeking winter and summer pastures, horse-­breeding and plunder. Such groups formed the core armies of smaller dynasties such as the Artuqids of Mayyafariqin. So the entire region was in a chronic state of political and social disorder throughout most of the eleventh and twelfth centuries, and the Turks were in the eye of the storm. As the historian of Damascus, Ibn al-­Qalanisi (d. 1160) expressed it: The people of Khurasan, Iraq and Syria were in a state of constant bickering and hatred, wars and disorder, and fear of one another, because their rulers neglected them and were distracted from the task of governing them by their dissensions and mutual warfare.8

This context helps to explain why the Turks figure as they do in the Arabic texts of this period. Power within the tribe lay with certain families. On the death of a senior male member of the ruling clan, his patrimony would be shared out amongst his male relatives. The sharing out of power and of territory, the system of providing a powerful atabeg, the endless disputes over ­succession – ­all of this departed significantly from Muslim norms and underlined the alien nature of Turkish rule. This paper is deliberately entitled ‘The View from Above’, in order to focus attention on a basic feature of the Arabic sources that inflects the attitudes they reveal about the warriors of Turkish stock who irrupted onto the scene in Syria and the Jazira from the mid-­eleventh century onwards. These ­foreigners – ­perhaps ‘aliens’ might be a better ­term – ­were of undeniable military and political importance. But as a daily presence in everyday affairs, they scarcely figure. They seem to have aroused no curiosity, no sustained interest, in the Arabic sources that deal with Syria and the Jazira in the eleventh and twelfth centuries. As a direct result of this indifference, it is difficult to discover exactly how Arab Muslims and Turks interacted in this region during

342  |  the me d ie va l tur k s this period. Suddenly a casual aside, a brief authorial comment, a revealing anecdote, brings unexpected enlightenment. But even then, the view from above is somewhat patronising. Medieval Arabic chroniclers are not usually closet sociologists or anthropologists. And it is very telling indeed that no sources in Turkish covering this time and place are known. So effectively the Turks themselves have no voice. This raises another and fundamental difficulty that underlies the history of this period, namely that of communication between Arab and Turk. Obviously, the Turks, who comprised several distinct subgroups, spoke some form of Turkish. The sources are silent on the question of how many of them spoke any kind of Arabic, whether that was classical or colloquial. The sources are equally silent about how many Arabs spoke some form of Turkish. A highly educated Arab aristocrat such as Usama b. Munqidh remarks unashamedly that he does not understand Turkish.9 Of course it is likely that, as time passed, more and more people on both sides of the linguistic divide would acquire the other language with some degree of fluency. But again the sources reveal no detailed information on this. Thus, it seems reasonable to assume that, at least up to the accession of the famous opponent of the Crusaders, Nur al-­Din, in 1146, the rank and file of the warriors of Turkish stock remained solidly monolingual and were thus unable to communicate properly with the Arabic- or Aramaic-­speaking population of Syria and the Jazira. It is plain that most Turkish warriors did not live in the towns and had no wish to do so. When they were forcibly billeted with the local population in an urban setting, as the Artuqid ruler Najm al-­Din İl-­Ghazi tried to do in the Jazira, the result was a spectacular failure. Their entire background and culture were nomadic. So, living together in isolation as a Turkish-­speaking community, they would have had little chance to learn Arabic. And conversely, how would the local population have been in a position to learn a Turkish dialect? Moreover, the sources do not discuss the presence of the women and children of the Turkish troops and they also make no mention of intermarriage. The Nomenclature of the Turks By the twelfth century there was at least some attempt by Muslim chroniclers to use special nomenclature to denote the different groups of Turks. Although

‘the v ie w f ro m a bo ve ’  | 343 a variety of names for the Turks are found in the sources, it is still not always easy to ascertain what the differences are between them. The chronicler Sibt b. al-­Jawzi (d. 1256) mentions the presence of ‘al-Turcoman wa’l-atrak’ (Turcomans and Turks) in his account of the events of the 1060s.10 These are identified as two distinct groups. He is critical of the Turcomans, the majority of whom were keen on pillaging; according to him, they did all kinds of foul and forbidden things and they had taken complete possession of Syria. But he does not mention any involvement by the ‘atrak’ (Turks) in these misdeeds; this term probably refers to the Turkish slave soldiers in the sultans’ armies. Another chronicler, Ibn al-­Qalanisi, seems to be writing about another category of Turk whom he calls kafir Turk (infidel Turk). But he does not explain who they were. He writes in his account of the year 532/1137–8, when Byzantine cavalry were in the environs of Aleppo, ‘some “kafir” Turks took refuge in Aleppo to escape them and to warn the people of Aleppo that the Greeks (were there)’.11 Inevitably biased but still valuable evidence about the Turks is provided by Eastern Christian chroniclers. Michael the Syrian (d. 1199) writes that the people of Edessa said of the Turks: ‘Behold a new and barbarous nation pours forth . . . (they) have the faces of men and the hearts of dogs. O Christians, take heed!’12 Later on in his chronicle, he has a

Figure 23.2  ‘Cavalry pursuit’, Rashid al-Din’s Jami‘ al-tawarikh (Edinburgh University Library Or. Ms. fol. 129b)

344  |  the me d ie va l tur k s more positive or pragmatic approach: ‘Those towards the west, the middle of the inhabitable earth, their journey was in the kingdom of the Arabs. They united (with the Arabs), adhered to their faith and embraced (it).’13 How Did Muslim Sources Portray the Nomadic Turks? The nomadic Turks (Turcomans) had endured unremittingly hard lives for millennia. Their most striking characteristic was mobility. This was the key to their survival. It was by their mobility that they could attack the enemy, plunder and then depart at great speed. Their mobility was not confined to warriors alone; it entailed movement of men, women, livestock, tents and baggage wagons. Their military strength lay in their mounted archers, who discharged their arrows at a distance and avoided direct encounters with the enemy. Already in the ninth century the Arabic writer al-­Jahiz praised the Turkish horseman, saying: The Turk has four eyes, two at the front and two at the back of his head. They have taught their horsemen to carry two or three bows and strings to

Figure 23.3  Parthian shot from a fifteenth-century Persian manuscript (Istanbul, Topkapı Sarayı Müzesi H.2153)

‘the v ie w f ro m a bo ve ’  | 345 match them. And the Turk has with him on his raid all that he needs for himself, his armour, his beast, and the harness of his beast.14

The nomads’ dress contrasted forcibly with that of the settled Muslims. The Turcomans wore furs or shapeless sack-­like cloth garments. Both sexes wore trousers. Their lack of hygiene must have shocked the Muslim Arabs. The men commonly wore their hair down to their waists and greased it with rancid butter. Indeed, one Eastern Christian source, Matthew of Edessa, mentions that the Seljuq army could be smelt at a distance of three days’ travel. The basis of the nomadic economy was a combination of pastoralism and raiding. The Turcomans’ major food supply came from their herds. Their wealth and

Figure 23.4  Two nomad women in conversation (after İpşiroğlu, Bozkir Rüzgâri Siyah Kalem (Istanbul, 1985), pl. 34)

346  |  the me d ie va l tur k s status were usually assessed by how many horses they owned. The sedentary Muslim populations were disgusted when they saw what the Turcomans ate: wolves, foxes, dogs, mice, rats and snakes. The eleventh-­century Arab writer Ibn Hassul says: ‘They consume only meat, even if it is dripping blood or filthy, and they do not wish for anything else.’15 During the pre-­Seljuq period Arabic and Persian geographers had written about the Turkish nomads in salacious detail, providing accounts that mixed fact and fantasy. The anonymous author of the late tenth-­century Persian work entitled Hudud al-‘alam, for example, wrote of the nomadic Turks: ‘These people have the nature of wild beasts and have rough faces and scanty hair. They are lawless and merciless.’16 In the tenth century Ibn al-­ Faqih mentions the type of details about the nomadic Turks which probably helped to formulate stereotypical views of the Turks in general in the minds of later Muslims. He writes that the Turks are the most skilled bowmen and are very courageous and strong. But he also mentions many lurid details about Turcoman customs: ‘They have sexual intercourse with every creature they meet, be it a woman, a boy or an animal.’17 The Infiltration of the Seljuq Turks into Syria and the Jazira The coming of nomadic Turkish tribes into Syria in the eleventh and twelfth centuries was an extension of their movements from Central Asia into Iran and Iraq and especially the invasions led by the Seljuq family. As already mentioned, by the early twelfth century the Levant was in disarray. The Seljuq dynasty that had ruled the eastern Islamic world since 1055 was no longer as strong or as centralised as it had been. Weaknesses inherent in the Seljuq system of g­ overnment – w ­ hich was based on the divisive principle of family ­rule – ­had become more obvious and debilitating since the deaths in quick succession of the mighty Seljuq vizier Nizam al-­Mulk and the Seljuq sultan Malikshah in 1092. Petty scions of the Seljuq family, whose power was concentrated in such key cities as Damascus and Aleppo, were being replaced or controlled by ambitious Turkish military barons who served as their atabegs and then took power themselves.

‘the v ie w f ro m a bo ve ’  | 347 How the Turkish Military Barons Ruled Ibn al-­Azraq, a twelfth-­century chronicler from the Jazira, describes the plight of the citizens of Mayyafariqin before its conquest by the Turcoman military leader İl-­Ghazi in 512/1118. He writes in some detail about the state of devastation in the city: The people were in great distress because of billeting in their homes. Most of the city was ruined because of constant changes of régime and overlord. Those who conquered them treated them unjustly, tyrannised and mulcted them, because they knew that they themselves would not endure and their rule would not ­last . . . ­The soldiers who had no homes began living in it (the city) and setting up tents in the ruins of the city because most of the city was devastated and the roads were terrorised by robbers and highwaymen.18

However, Ibn al-­Azraq is at pains to praise the nature of İl-­Ghazi’s rule in Mayyafariqin: ‘He showed justice and kindness to the i­ nhabitants . . . F ­ rom the moment that he assumed power, the roads and the countryside became safe. The robbers fled and the villages flourished. Mayyafariqin began to prosper and he ruled the people very well.’19 But this lavish praise should be treated with caution. Ibn al-­Azraq worked as a scribe for the Artuqid family. He knew on which side his bread was buttered. Two Key Turkish Rulers in Syria: How Islamised Were They? a. İl-­Ghazi (d. 1122), the Ruler of Mardin, Mayyafariqin and Aleppo (for a short while) How important was jihad for a Turkish military baron, such as İl-­Ghazi, in Syria and the Jazira in the early Crusading period? Certainly this concept had been well known in tenth-­century Syria where the Hamdanid Shi‘ite dynasty waged jihad on the Byzantine/Muslim frontier. Grandiose jihad sermons were preached by Ibn Nubata (d. 374/984–5) in the Jazira20 and the Arabic poet al-­Mutanabbi praised the Hamdanid ruler, Sayf al-­Dawla, for his promotion of jihad.21

348  |  the me d ie va l tur k s But generally amongst the sedentary Muslim populations of the eleventh and early twelfth centuries, they probably remembered the words of al-­Jahiz: ‘The Turk does not fight for religion nor for the interpretation of ­Scripture . . . ­nor for defence of the home nor for wealth, but only for plunder.’22 Ibn al-­Qalanisi (d.  1160) the chronicler of Damascus, likened İl-­Ghazi’s Turcomans to birds and beasts of prey: ‘lions seeking their prey and gerfalcons hovering over their victims’.23 In his view, therefore, although undoubtedly skilled fighters, they were motivated only by lust for booty. However, İl-­Ghazi is mentioned as deliberately exploiting jihad propaganda with his Turcoman forces in 513/1119–20 before the Battle of the Field of Blood. While in Aleppo, which was being increasingly pressurised by the Crusader leader Roger of Antioch, Il-­Ghazi found himself immersed in a feverish atmosphere of jihad and he came to realise that he could use this to boost the morale of his troops in the forthcoming battle. So, according to the account of Ibn al-‘Adim, a qadi called Ibn Khashshab (d. 528/1133–4) preached jihad to the assembled Turcoman troops and İl-­Ghazi made his amirs swear an oath ‘to perform their duty courageously, to fight heroically, and not to retreat, even if they had to shed all their blood in the jihad’. Initially the Turcomans displayed indifference and even contempt; one of them cried out: ‘Was it to obey this man with a turban that we left our country?’ Later on, Ibn al-‘Adim continues, Ibn al-­Khashshab brought ‘tears of ecstasy’ to the eyes of the Turcomans who then went into battle and gained a glorious victory.24 The caliph in Baghdad sent İl-­Ghazi a robe of honour. The narrative about Ibn al-­Khashshab does, however, remain problematic. How could the assembled Turcoman troops of İl-­Ghazi understand a jihad speech proclaimed in Arabic? It is much more likely that this grandiose speech was composed before or after the event rather than delivered on the spot. İl-­Ghazi apparently never again utilised the weapon of jihad propaganda in his later campaigns against the Crusaders. Nor did he follow up this military triumph; according to Usama b. Munqidh, this was due to İl-­Ghazi’s habit of excessive drinking of alcohol. After his victory over the Crusaders at the Battle of the Field of Blood in 1119, Usama writes that İl-­Ghazi remained in a drunken state for twenty days.25

‘the v ie w f ro m a bo ve ’  | 349 b. Tughtegin, the Ruler of Damascus (1104–28) The most important source for the life of Tughtegin is the town chronicle of Damascus written by Ibn al-­Qalanisi (died 1160).26 Despite his obvious bias towards his Turkish overlords and his convenient omissions and obfuscations of negative i­nformation – h ­ e would have been foolish to do o­ therwise – I­ bn al-­Qalanisi is a keen observer of events. The name Tughtegin means Warrior Falcon. His full names were Abu Mansur Tughtegin b. ‘Abdallah. His honorific titles, acquired whilst he was ruler of Damascus, were Zahir al-­Din and Sayf al-­Islam. Most often in the Islamic sources which recount his career he is simply mentioned as the Atabeg or Atabeg Tughtegin. Perhaps it is worth noting that Tughtegin’s father is here called ‘Abdallah. This could be an indication that he was a Muslim, thus showing that Tughtegin’s family had converted to Islam at least one generation back, or that the name of his father was unknown. What’s in a name? It is dangerous to speculate too much on the question of Tughtegin’s personal name, but it is perhaps not too fanciful to suggest that this major Turkish leader in Syria deliberately chose to keep his Turkish name rather than using his Arabic titles; and he was generally known by it. The use of Tughtegin, a totemistic Turkish name, must have enhanced his reputation as an alien ruler, a steppe warrior whose not-­so-­distant ancestors had hailed from distant Central Asia. In due course, naturally, the inhabitants of Syria must have got used to the fact that their military warlords bore Turkish names. Tughtegin had been a mamluk belonging to Tutush, son of the Seljuq sultan Alp Arslan. Tughtegin served as a military commander for his master who manumitted him, married him to Safwat al-­Mulk, the mother of Duqaq, son of Tutush, and made him Duqaq’s atabeg. After the death of Tutush, his son Duqaq took over Damascus, where Tughtegin joined him and became his atabeg. In 497/1104, Duqaq died and Tughtegin became the openly independent ruler of Damascus. Tughtegin’s Muslim piety and beneficent rule are emphasised by Ibn al-­Qalanisi in the following terms: ‘He modelled his conduct on the precepts of the Faith; he inspired fear in the lawless and the evildoers, and was bountiful in excess in rewarding the loyal and well-­doing; all hearts were attached to his cause by his liberality.’27 Thus, admittedly from the pen of the panegyrist of his dynasty, Ibn al-­Qalanisi, we see Tughtegin

350  |  the me d ie va l tur k s using a traditional combination of ingratiation and coercion to consolidate his takeover of power. His conduct is, however, also described as that of the pious Muslim ruler who exercises hisba – the task of ‘commanding the good and prohibiting the bad’. Legitimacy was a key preoccupation of the Turkish military commanders of Syria. Such men seized lands, but the mere possession of territory was not sufficient for them; they needed Islamic legitimisation of their power by the caliph in Baghdad. Tughtegin pursued, whenever he could, an aggressively expansionist policy not only against the Franks but also against other petty Muslim rulers left in the wake of Seljuq decline. When he died, medieval Arabic sources present his achievements in the usual clichéd obituary notices. Ibn al-­Athir writes of Tughtegin: ‘He was wise and generous. He raided and waged jihad against the Franks frequently and ruled his subjects well, eager to be just to them.’28 In turbulent times Tughtegin managed to rule Damascus with an iron hand for twenty-­four years and was a much respected and feared opponent of the Crusaders. He can be seen as the major early Turkish defender of Islam against the Crusaders. This judgement is not based on his personal virtues or piety but on his sheer longevity, persistence, tenacity and ability to stay in power. He outlived all the Turkish military ­barons – ­Balak, İl-­Ghazi and ­others – ­in the first phase of Crusading activity in the Levant. Moreover, he managed to die not at the hand of an assassin or on the battlefield but in his own bed. Much of his success sprang from his firm commitment to Syria. Seljuq offensives from further east were destined to fail since they had no local Syrian support. Other early Muslim leaders such as İl-­Ghazi and, later on, Zengi, dabbled in politico-­military struggles in the Jazira or in Baghdad and dissipated their energies, much to the advantage of the Crusaders. Tughtegin was a worthy precursor of Nur al-­Din in his concentration on Syria, and especially Damascus, as his sphere of operation. The Survival of Pre-­Islamic Shamanistic Practices amongst the Turks in Eleventh- and Twelfth-­century Syria Turcoman society had long been shamanistic, and even after the conversion of their tribal leaders to Sunni Islam, as dutifully proclaimed in the Arabic and Persian sources, it is probable that the Turcomans retained many of

‘the v ie w f ro m a bo ve ’  | 351

Figure 23.5  Shaman (A Cultural Atlas of the Turkish World: The Pre-Islamic Period (Istanbul, 1997), p. 200)

352  |  the me d ie va l tur k s their basic shamanistic practices for many years to come. For example, the Syriac Christian chronicler Bar Hebraeus (d. 1286) writes that the first Seljuq sultan, Tughril, had the habit of sitting on an elevated throne opposite a splendid bow made of gold ‘and he held in his hand two arrows which he played with’.29 This detail has shamanistic resonances, since the bow and arrow were used for divination. Moreover, there is clear evidence of Turkish totemistic names in the sources; these names often refer to animals and birds. A truly notorious event in the career of Tughtegin deserves special attention in this context: the death of Gervase de Basoches, the Crusader lord of Tiberias, in 501/1107–8. Versions of the story are given in a number of sources, both Muslim and Crusader. The account of Gervase’s killing given by Albert of Aachen (d. 1120) is detailed and shocking. To his audience back home in Europe it might even have seemed amazing and almost unbelievable: Gervase was brought out in the middle of the town of Damascus, and after much mockery he was shot through by the Turks’ arrows and gave up the ghost. Gervase, eminent knight, having died in this way, one of the most powerful of the Turks ordered his head to be cut off, and the skin of his ­head . . . ­to be pulled off and dried.30

Similar details can be found in the account of another Crusader author, Guibert of Nogent (d. 1124), who relates that Gervase was killed by arrows, and thereafter the crown of his head was removed and a cup was made from his skull.31 What of the Muslim accounts of this episode? Most of these go so far as to mention that Gervase was killed whilst he was Tughtegin’s prisoner, but they generally gloss over certain macabre details. Clearly, stories similar to those told by Albert of Aachen and Guibert of Nogent must have been circulating in the Levant. They are eventually mentioned, in rather laconic but frank fashion, by the much later Muslim historian, Ibn al-­Furat (d. 807/1405). He provides the gruesome detail that Tughtegin hollowed out Gervase’s skull while he was still alive and drank from it: Tiberias remained in their (Crusader) hands until Tughtegin attacked the Count, the Lord of Tiberias, in Shawwal in the year 501. He hollowed out the skull of his head while he was still alive, and drank wine in it, while he was looking at him [sic]. He lived for an hour, then he died.32

‘the v ie w f ro m a bo ve ’  | 353

Figure 23.6  Minbar of Nur al-Din, the Aqsa Mosque, Jerusalem (courtesy of Alistair Duncan)

Clearly this particular version of the story is sensationalist rather than accurate, though it does succeed in conjuring up, in vivid pictorial terms, an abominable spectacle. But what was behind this savage killing? It is worth remembering that the earliest known Turkic people, the Hsiung-­Nu, as well as the non-­Turkic Scythians, made drinking bowls from the skulls of prestigious fallen enemies. Wine would be poured into such a ‘bowl’ and then mixed with the blood of a sacrificial victim and stirred with a sword.

354  |  the me d ie va l tur k s The Turkish Ruler of Damascus, Nur al-­Din Mahmud (d. 1176) The accession of Nur al-­Din as the ruler of Damascus in 1146 marked a turning point.33 His personal name was Mahmud, a solidly Islamic name. His predecessors often bore Islamic titles, but their personal names told a very different s­tory – ­Tughril, Chaghri, Alp Arslan, Tutush, Atsiz, Duqaq, Buri, Barkyaruq, Tapan, Balak, İl-­Ghazi, Tughtegin, Aq Sunqur, Zengi. Even the supreme Seljuq monarch, who reigned from 1065 to 1085, bore the name Sultan Malikshah, conjoining the royal titles of Turk, Arab and Persian in his cognomen, and making no reference to the Islamic faith. These names signal a marked loyalty to Turkic tradition at the expense of Islam. In contrast, Nur al-­Din quickly established an enviable reputation for Islamic piety based on the celebrated asceticism of his private life, his honouring of the Islamic ‘ulama’, his ambitious programme of erecting buildings such as mosques and madrasas to serve the Islamic faith, and his deliberate fostering of jihad ideology and propaganda. It was he who exemplified the transition from outsider to insider, from foreigner to native, from infidel to Muslim. His personal name says it all. And the sources make it clear that he was fluent in Arabic. He knew the Qur’an well and allegedly often engaged in theological conversations with Muslim scholars. This sea-­change in the behaviour of a Turkish warlord brought a concomitant change in his way of life. Nur al-­Din made Damascus his capital. Those words sound unremarkable, but much of consequence flowed from that decision. It meant that he abandoned the aimless, peripatetic lifestyle that characterised, say, İl-­Ghazi or his own father Zengi, with its backdrop of endless warfare. It meant that he put down roots. It meant that he gave up on nomadism. It meant that he threw in his lot with the settled population, almost all of whom lived in cities, towns and villages. And that meant that he adopted the norms and forms of their culture. There was no turning back. His commitment to Damascus expressed itself in the costly building programme he funded but also in his decision to strengthen its walls and gates. Nur al-­Din, then, exemplified the increasing accommodation that characterised both sides of the linguistic and cultural divide in the later twelfth century, and with it an increasing mutual dependency on the part of Arab and Turk alike. Once again architecture provides a yardstick for social

‘the v ie w f ro m a bo ve ’  | 355 change, for one Turkish warlord after another followed the lead of Nur al-­ Din in erecting buildings that proclaimed their piety and memorialised their names while also serving the Muslim community to which they increasingly saw themselves as belonging. Concluding Remarks The presence of the Turcomans was viewed as a necessary evil by the sedentary Muslim populations. The Seljuq leadership exercised control, sometimes precariously, over them. The Seljuqs also came to an understanding with the urban elites; the Seljuqs would provide military support and the cities would provide taxes and the bureaucratic expertise necessary to administer the Seljuq empire. Despite the disruption to agriculture caused by the Turcomans, and their tendency to damage cities and monuments, Muslim writers valiantly made the best of a bad situation, for they saw that these nomads were here to stay. There were, however, new aspects of Seljuq government that came from the shamanistic steppe heritage. The first two Seljuq sultans retained their Turkish totemistic n ­ames – ­Tughril, meaning ‘falcon’, and Alp Arslan, meaning ‘hero lion’. Thereafter, most, but not all subsequent, Seljuq sultans were known by Muslim names. Even so, the title ‘Sultan of east and west’ bestowed by the ‘Abbasid caliph in Baghdad on the first Seljuq sultan Tughril in 1055 reflects the nomadic steppe concept of world ­dominion – a­ symbol of Turkish sovereignty, extending from where the sun rises to where it sets. The Turkish symbol of a bow and arrow denotes sovereignty. Legend has it that the first Seljuq sultan Tughril is said to have placed a bow and arrow in the mosque of Byzantine Constantinople. The Seljuq Turkish institution of atabeg, usually denoting a powerful Turkish military commander whose duty it was to supervise the upbringing of young Seljuq princes, also came from steppe tradition. But when it was introduced into the Seljuq government it had dire consequences. The institution proved to be a serious centrifugal force, working against the centralised government model beloved of Nizam al-­Mulk. After 1092 atabegs seized power in their own right and this contributed to the disintegration of the Seljuq state. Traditionally, the Seljuq Turkish élite are presented in the medieval Arabic and Persian sources as pious Sunni Muslims who had converted to

356  |  the me d ie va l tur k s Islam before their major invasions began. This portrayal of the Turks, who went on to govern the heartlands of the Islamic world for many centuries, is not surprising. Their Arab and Persian bureaucrats, court chroniclers and poets needed to put a brave face on political realities and to help to forge the necessary alliance between the military force of the Turks and the long-­ established prestige of the Sunni Arab and Persian religious classes. No doubt, over time, some of the Turkish sultans did indeed grow into the role accorded them by their chroniclers and court p ­ oets – ­namely, that of defenders of Sunni Islam and fighters of jihad against infidels and heretics. But that could not be said of the Turkish rulers of Syria and the Jazira up to 1146. As an alien, invading military force, the Seljuq Turkish sultans generally adopted a publicly supportive and deferential stance towards the Sunni caliph in Baghdad, from whom they sought the key religious credentials that they craved. And the caliphs gave them grand robes and honorific titles. A major religious development was the establishment of a network of Sunni madrasas by high-­ranking Seljuq administrators across the empire. Nizam al-­Mulk set the example by funding and supervising at least ten such madrasas or religio-­ legal colleges; the principal one of these was in Baghdad. Thus, a reliable, well-­trained class of Sunni religious scholars could promote orthodox Islam across the empire under the umbrella of the Seljuq state. Stories about the Turks, such as the tragic end to the life of Gervase de Basoches, analysed above, were narrated to titillate and mock and were saturated with feelings of religious or cultural superiority on the part of Muslim and Crusader sources alike. To the Muslims, the Turks were both their overlords and also a barbarian ‘Other’. To the Crusaders the Turks were another ‘Other’, even more alien than the infidel Arabs (or Saracens, as they were often called). Both Muslim and Crusader narratives, when analysed, suggest, unwittingly but persuasively, a subtext that contradicts the main narrative. Through the tales they tell of Turkish customs, dating back to the rituals of the Inner Asian nomads and to their pastoral hunter lifestyle, they suggest that the process of sedentarisation and Islamicisation amongst the Turks must have been much slower and more protracted than the Sunni Muslim sources would have us believe. Even if some or many of the Turkish leaders, and even the nomadic tribes, had converted to Islam either before or after coming to Syria, this

‘the v ie w f ro m a bo ve ’  | 357 conversion was probably still superficial and some at least of their ancient shamanistic traditions remained alive and active. Certainly the story of the killing of Gervase, featuring as it does the enactment of pre-­Islamic Turkish rituals in the very heart of ­Damascus – ­a Muslim city since the seventh ­century – ­is thought-­provoking. It highlights the difficulties that the sedentary Muslim Arabs (and the Persians further east) must have experienced in coming to terms with these new and powerful overlords. In the case of Gervase’s death, it is the precise account of a Crusader author, Albert of Aachen (whatever his motivation may have been), which is the most valuable testimony. Empires built by usurping nomadic warlords are notoriously ephemeral, and nomadic traditions sit lightly on the civilisations they take over. As with the nomadic Mongols in Iran and China, it was to be expected that indigenous Perso-­Islamic statecraft would eventually prevail over the incoming Seljuq Turks. But the military power of the Turks was an essential ingredient too, notably in Syria where there was urgent need for the Muslims to withstand the Crusaders and expel them. What came into being in this period was a symbiosis of Iranian statecraft, the Sunni Islamic norms followed by the Arab ‘ulama’, and Turkish steppe traditions. Notes   1. For a general historical overview, see P. Golden, Turks and Khazars (Farnham and Burlington, 2010); A. C. S. Peacock, The Great Seljuk Empire (Edinburgh, 2015).   2. M. G. S. Hodgson, The Venture of Islam. Conscience and History in a World Civilisation, Vol. 2: The Expansion of Islam in the Middle Periods (Chicago and London, 1974), 399.  3. For example, in his work entitled the Kitab al-Mustazhiri, addressed to the ‘Abbasid caliph during the short period 487–8/1094, al-­Ghazali argues persuasively that the Turks provide indispensable military support to the caliph in waging jihad against the infidels; see C. Hillenbrand, ‘Islamic orthodoxy or Realpolitik? Al-­Ghazali’s views on government’, Iran XXVI (1988), 81–94, and especially 84.   4. Two important works which deal with Crusader–Muslim warfare, using the evidence of Western medieval sources, are R. C. Smail, Crusader Warfare 1097–1193 (Cambridge, 1967) and C. Marshall, Warfare in the Latin East, 1192–1291 (Cambridge, 1992).

358  |  the me d ie va l tur k s   5. Albert of Aachen, Historia Ierosolimitana. History of the Journey to Jerusalem, ed. and tr. S. B. Edgington (Oxford, 2007), 770–1.   6. Y. Frenkel, The Turkic Peoples in Medieval Arabic Writings (London and New York, 2015), 41–53 and especially 51–2.   7. A. C. S. Peacock, Early Seljuq History: A New Interpretation (London and New York, 2010) 47.   8. Ibn al-­Qalanisi, Dhayl Ta’rikh Dimishq, tr. H. A. R. Gibb as The Damascus Chronicle of the Crusades (London, 1932; repr. Mineola, NY, 2002), 54.   9. Usama b. Munqidh, Kitab al-i‘tibar, ed. P. K. Hitti (Princeton, 1930), 130. 10. Sibt b. al-­Jawzi, Mir’at al-zaman fi tarikh al-a‘yan, ed. A. Sevim (Ankara, 1968), 33, 77 and 142. 11. Ibn al-­Qalanisi, Dhayl Ta’rikh Dimishq, 249. 12. Michel le Syrien, Chronique de Michel le Syrien, Patriarche Jacobite d’Antioche (1166–1199), ed. and tr. J.-B. Chabot (Paris, 1899–1910), 154. 13. Michel le Syrien, Chronique, 156. 14. Al-­Jahiz, Manaqib al-atrak, tr. C. T. Harley Walker, ‘Jahiz of Basra to al-­Fath Ibn Khaqan on the “Exploits of the Turks and the Army of the Khalifate in General”’, JRAS (1915), 666–7. 15. Ibn Hassul, The Superiority of the Turks over Other Regiments, tr. Frenkel, in The Turkic Peoples, 77. 16. Anon., Hudud al-‘alam: Regions of the World. A Persian Geography 372 a.h.– 982 a.d., tr. V. Minorsky (London, 1970), 96. 17. Ibn al-­Faqih al-­Hamadhani, Kitab al-buldan; excerpts tr. Frenkel, in The Turkic Peoples, 50. 18. Ibn al-­Azraq, Ta’rikh Mayyafariqin wa-Amid, partial ed. and tr. C. Hillenbrand as A Muslim Principality in Crusader Times: The Early Artuqid State (Istanbul, 1990), 35. 19. Ibn al-­Azraq, Ta’rikh Mayyafariqin, 35–6. 20. Ibn Khallikan, Kitab wafayat al-a‘yan, tr. B. W. M. de Slane as Kitab wafayat al-a‘yan: Ibn Khallikan’s Biographical Dictionary (Paris, 1843), vol. 2, 10–11. 21. Al-­Mutanabbi, Mutanabbi carmina cum commentario Wahidii, ed. F. Dieterici (Berlin, 1856–61), 481–4. 22. Al-­Jahiz, Manaqib al-atrak, 670. 23. Ibn al-­Qalanisi, Dhayl Ta’rikh Dimishq, 159. 24. Ibn al-‘Adim, Zubda, 188–9. 25. Usama b. Munqidh, Kitab al-i‘tibar, 149. 26. Ibn al-­Qalanisi, Dhayl Ta’rikh Dimishq, 62–186.

‘the v ie w f ro m a bo ve ’  | 359 27. Ibn al-­Qalanisi, Dhayl Ta’rikh Dimishq, 63. 28. Ibn al-­Athir, The Chronicle of Ibn al-Athir for the Crusading Period from al-Kamil fi’l-tarikh, Part 1. The Years 491–541/1097–1146. The Coming of the Franks and the Muslim Response, tr. D. S. Richards (Aldershot and Burlington, 2006), 274. 29. Bar Hebraeus, The Chronology of Gregory Abu’l-Faraj 1225–1286, tr. E. A. Wallis Budge (London, 1932), vol. 1, 201. 30. Albert of Aachen, Historia Ierosolimitana, 770–1. 31. Guibert of Nogent, cited by Edgington, Albert of Aachen, Historia Ierosolimitana, 771. 32. Ibn al-­Furat, Ta’rikh al-duwal wa’l-muluk, partial eds and trs U. and M. C. Lyons (Cambridge, 1971), vol. 2, 55. 33. For a detailed account of the career of Nur al-­Din, see N. Elisséeff, Nur al-Din: un grand prince musulman de Syrie au temps des Croisades (Damascus, 1967).

Bibliography Primary Sources Albert of Aachen, Historia Ierosolimitana. History of the Journey to Jerusalem, ed. and tr. S. B. Edgington (Oxford, 2007). Anon., Hudud al-‘alam: Regions of the World. A Persian Geography 372 a.h.–982 a.d., tr. V. Minorsky (London, 1970). Bar Hebraeus, The Chronology of Gregory Abu’l-Faraj 1225–1286, tr. E. A. Wallis Budge (London, 1932). Ibn al-‘Adim, Zubdat al-halab min tarikh Halab, ed. S. Dahan (Damascus, 1954). Ibn al-­Athir, The Chronicle of Ibn al-Athir for the Crusading Period from al-Kamil fi’l-tarikh, Part 1. The Years 491–541/1097–1146. The Coming of the Franks and the Muslim Response, tr. D. S. Richards (Aldershot and Burlington 2006). Ibn al-­Azraq al-­Fariqi, Ta’rikh Mayyafariqin wa-Amid, partial ed. and tr. C. Hillenbrand as A Muslim Principality in Crusader Times: The Early Artuqid State (Istanbul, 1990). Ibn al-­Furat, Ta’rikh al-duwal wa’l-muluk, Vol. 2, eds and trs (partial) U. and M. C. Lyons (Cambridge, 1971). Ibn Khallikan, Kitab wafayat al-a‘yan, tr. B. W. M. de Slane as Kitab wafayat ala‘yan: Ibn Khallikan’s Biographical Dictionary, 4 vols (Paris 1843–71). Ibn al-­Qalanisi, Dhayl Ta’rikh Dimishq, tr. H. A. R. Gibb as The Damascus Chronicle of the Crusades (London, 1932; repr. Mineola, NY, 2002). al-­Jahiz, Manaqib al-atrak, tr. C. T. Harley Walker, ‘Jahiz of Basra to al-­Fath Ibn

360  |  the me d ie va l tur k s Khaqan on the “Exploits of the Turks and the Army of the Khalifate in General”’, Journal of the Royal Asiatic Society (1915), 666–7. Michel le Syrien, Chronique de Michel le Syrien, Patriarche Jacobite d’Antioche (1166– 1199), ed. and tr. J.-B. Chabot (Paris, 1899–1910). al-­Mutanabbi, Mutanabbi carmina cum commentario Wahidii, ed. F. Dieterici (Berlin, 1856–61). Sibt b. al-­Jawzi, Mir’at al-zaman fi ta’rikh al-a‘yan, ed. A. Sevim (Ankara, 1968). Usama b. Munqidh, Kitab al-i‘tibar, ed. P. K. Hitti (Princeton, 1930). Secondary Sources Elisséeff, Nikita, Nur al-Din: un grand prince musulman de Syrie au temps des Croisades (Damascus, 1967). Frenkel, Yehoshua, The Turkic Peoples in Medieval Arabic Writings (London and New York, 2015). Golden, Peter, Turks and Khazars (Farnham and Burlington, 2010). Hillenbrand, Carole, ‘Islamic orthodoxy or Realpolitik? Al-­Ghazali’s views on government’, Iran, XXVI (1988), 81–94. Hodgson, Marshall G. S., The Venture of Islam. Conscience and History in a World Civilisation, Vol. 2: The Expansion of Islam in the Middle Periods (Chicago and London, 1974). Peacock, A. C. S., Early Seljuq History: A New Interpretation (London and New York, 2010). Peacock, A. C. S., The Great Seljuk Empire (Edinburgh, 2015).

24 A Cosmopolitan Frontier State: Relations among the Kurds, Arabs, Byzantines, Armenians, Persians and Turks under the Marwanids of Diyar Bakr, 990–1085

T

he dissolution and fragmentation of ‘Abbasid power in the tenth and eleventh centuries led to the appearance of a number of small dynasties across many areas of the Muslim world. Various Kurdish groups formed principalities which were in practice autonomous from the Baghdad caliphate. Such Kurdish dynasties included the Shaddadids (c. 951–1174) in western Azerbaijan,1 who waged jihad against Christian Georgians, Armenians and Byzantines, and the Hasanuyids (c. 960–1014) who flourished in the central Zagros area and supplied troops for the Buyid amirs of Persia and Iraq.2 In northern Syria, Diyar Bakr and Armenia, territories which lay near or on the eastern borders of the Byzantine empire or the fringes of the Fatimid empire, small states, ethnically diverse, whose peoples spoke Arabic, Armenian, Kurdish, Persian or Turkish, clustered together in close proximity, sometimes forming temporary and volatile alliances, and at other moments engaged in fierce hostilities with each other.3 Some of these small dynasties, such as the Hamdanids of Aleppo, the Mazyadids of Hilla, the ‘Uqaylids of Mosul and the Mirdasids of Aleppo, depended on Bedouin Arab tribal support, whilst others such as the Hasanuyids and the Marwanids relied on Kurdish nomadic groups.4 The historical sources dealing with the Marwanids include some information from well-­known medieval Arabic geographical works, such as that of al-­Muqaddasi,5 as well as more especially the Kamil al-ta’rikh of Ibn al-­Athir.6 However, the major source for the study of the Marwanids of Diyar Bakr is without doubt the long section about them in the still little-­used Arabic 361

362  |  the me d ie va l tur k s

Figure 24.1  Map of Marwanid Kurds (Thomas Ripper, Die Marwaniden von Diyar Bakr, Wiesbaden, 2000)

chronicle entitled Ta’rikh Mayyafariqin wa-Amid of Ibn al-­Azraq al-­Fariqi (d. after 1176–7).7 He worked as a scribe for the Artuqids of Mayyafariqin and wrote this detailed history of his home town from early Islamic times. In this work Ibn al-­Azraq provides detailed coverage of the Marwanid dynasty, 990–1085. This part of the chronicle, well edited by ‘Awad in 1959,8 still remains untranslated, but its contents were outlined in some detail in a long article by Amedroz in 1903.9 When recording the history of the Marwanids, Ibn al-­Azraq’s account is essentially focused on Mayyafariqin, which was their capital. It chronicles the violent events which preceded the establishment of a Kurdish state centred on the city. It then moves on to record the short periods in power of the first two Marwanid amirs, Badh the Kurd and Mumahhid al-­Dawla, before dealing in great detail with the apogee of Marwanid rule, the fifty-­one year-­long reign of Nasr al-­Dawla. The subsequent downfall of Marwanid power is also covered by Ibn al-­Azraq.10 It must be admitted that Ibn al-­Azraq’s chronicle is disordered and at times repetitive. Small snippets of information about local events in Diyar Bakr are more frequent than occasional longer narratives. This is especially

a cosm op ol ita n f ro ntie r s ta te  | 363 the case when dealing with Nasr al-­Dawla. Moreover, Ibn al-­Azraq does not produce many insights into the characteristics of the society he is describing. The impact of the nomads, who lived in the hinterlands of Diyar Bakr, on the settled urban populations is not mentioned very much, and he shows little interest in the lives of the predominantly Christian population in the cities of Mayyafariqin and neighbouring Amid. It seems that he took for granted the ethnic, linguistic and religious diversity of the society in which he lived. The Marwanids ruled from 380/990 until 478/1085. This dynasty established itself in the province of Diyar Bakr, seizing territory situated on the southern and western fringes of Armenia and Kurdistan. Their dynasty began with a Kurdish chief known as Badh the Kurd who came from the hills near Hizan, in the province of Bitlis. After the death of the principal Buyid amir, ‘Adud al-­Dawla, in 373/983, Badh took possession of Mayyafariqin and thereafter seized the cities of Amid, Nasibin and Akhlat, despite the efforts of Buyid and Hamdanid troops to unseat him.11 In 380/990, however, after he had tried unsuccessfully to take Mosul, Badh was killed in a battle against a coalition of Hamdanid and ‘Uqaylid troops. After Badh’s death, members of his family succeeded in establishing the Kurdish Marwanid dynasty (five rulers in all) which lasted almost a hundred years. Badh’s sister had married a mill owner called Marwan, from whom the name of the dynasty derived. Three of their ­sons – ­al-­Hasan, Sa‘id and ­Ahmad – a­re mentioned in the Arabic sources. The eldest son, al-­Hasan, known as Abu ‘Ali, had been with Badh when he died. He made his way to Mayyafariqin with his uncle’s widow, a woman from Daylam, whom he subsequently married. He took possession of Mayyafariqin and Amid and became the first ruler of the Marwanid dynasty in the province of Diyar Bakr. After his murder at Amid in 387/997, his brother Sa‘id, known as Mumahhid al-­Dawla, ruled until 401/1011. These two precarious reigns paved the way for the accession of the third brother, Ahmad, known by the honorific title of Nasr al-­Dawla. His reign of fifty years marked the crowning period of Marwanid power and prestige. After his death, his son and grandson ruled. The dynasty came to an end in 478/1085. In the thirteen-­year reign of the second Marwanid ruler, Mumahhid al-­Dawla (387/997–401/1010), the state gradually began to stabilise. The Byzantine historian Michael Psellus records that the emperor Basil ‘spent the

364  |  the me d ie va l tur k s greater part of his reign serving as a soldier on guard at our frontiers’.12 So it is not surprising that Ibn al-­Azraq records rather elliptically that in 390/999 the Byzantine emperor Basil II (976–125) went to Amid and Mayyafariqin. His behaviour on this occasion proved to be in keeping with his regular policy towards his empire’s eastern neighbours, which often involved the exercise of peaceful diplomacy rather than military means. Clearly, he well understood the value of face-­to-­face contact. Basil met Mumahhid al-­Dawla and they swore oaths to each other and reached an agreement.13 Ibn al-­Azraq does not mention any more details about the reasons for this visit but he records that Mumahhid al-­Dawla then wrote to various rulers and to the caliph in Baghdad. The caliph and the Buyid amirs, Baha’ al-­Dawla and his son Fakhr al-­Mulk, honoured him. He also received letters and presents from the Fatimid caliph, al-­Hakim, in Egypt.14 These high-­level contacts show the Marwanid ruler working hard to improve his status vis-­à-vis his neighbours. During his reign Mumahhid al-­Dawla built up the walls of Mayyafariqin, including the cylindrical tower (dated 1000–1 in a surviving inscription15) at the north-­east corner of the walls, where a Hamdanid attack might be expected. Ibn al-­Azraq records seeing his name on the outside of the wall in twenty-­two places and on a number of further places on the inside wall.16 This remarkable epigraphic overload offers further evidence of his marked tendency towards self-­promotion. Mumahhid al-­Dawla was able to rule for twice as long as his elder brother, and it was in his reign that the beginnings of a foreign policy involving his neighbours took shape. But much more was to come in the heyday of Marwanid power under Nasr al-­Dawla, who ruled long and very successfully from 401/1010 until 453/1061. The Marwanid state at this time stretched as far south as Akhlat, the borders of Lake Van and Jazirat ibn ‘Umar.17 The Marwanids controlled the key routes from the eastern Anatolian plateau to the plains of the Jazira, with Mayyafariqin in the centre, Arzan and Si‘ird to the east and Amid to the west. The Byzantines seemed to have accepted the Marwanids, like the Hamdanids and the Mirdasids further to the south-­west, as a buffer state between them and the wider Muslim world.18 It is worth investigating how Nasr al-­Dawla built up Mayyafariqin into a well-­defended capital city. The medieval Arab geographer al-­Muqaddasi describes Mayyafariqin as a fine city, with a sturdy stone wall, battlements, a

a cosm op ol ita n f ro ntie r s ta te  | 365 deep ditch and extensive suburbs.19 Early on in the reign of Nasr al-­Dawla, since his palace had been destroyed, he decided to renovate the existing castle, which had belonged to the Hamdanids, and to build next to it a burj al-mulk or centre of government adjoining it. He thus ensured the safety of his official residence.20 The castle was situated high up on a hill, thereby offering the ruler ample protection and allowing guards to warn of any approaching troops or caravans. The Marwanid principality was after all a buffer state, and buffer states were justifiably nervous of their neighbours and therefore needed to take precautions against potential hostilities. So, it is not surprising that the celebrated vizier of Nasr al-­Dawla, Abu’l-­Qasim al-­Maghribi, advised him to build this tower in a position dominating the whole city. Clearly this strategy would impress and intimidate the population both of the town and of the neighbouring settlements in the area, as well as any non-­local travellers, merchants, ambassadors and other visitors. But even more importantly these buildings would serve as prestige symbols emphasising his power and grandeur. So, an alternative translation for the Arabic name of this structure could well be ‘tower of power’. According to Ibn al-­Azraq, Nasr al-­Dawla ordered work on the castle and tower to begin in the year 403/1012 and he spent a lot of money on it. He even embellished the interior of the castle, placing gold on its walls and ceilings. He constructed an elaborate water system for the castle and placed pools and a bath there. What is extraordinary is that the work was begun and completed within a single year.21 That is a truly remarkable achievement. It testifies not only to the amir’s determination but also to his wealth and the efficiency of his administration. The building of the tower and the reconstruction of the castle alone would have required the quarrying, transport and dressing of a huge quantity of stone, which would have needed hundreds of skilled stonemasons as well as even larger numbers of unskilled labourers. The internal decoration of these buildings called for yet more specialised craftsmen. Many spectators came to admire these innovations, so it is clear that this was no run-­of-­the-­mill project. No detailed examination of the fortifications of Mayyafariqin has yet been carried out, and the dating of its component surfaces is uncertain, though it is very likely that what survives is largely of Marwanid and Artuqid construction, since those were the glory days of the city.22 Happily the observant and

366

|

the me d ie va l tur k s

Figure 24.2 Tower on the walls of Mayyafariqin (collection of the author)

well-travelled Persian Isma‘ili writer, Nasir-i Khusraw, visited Mayyafariqin in 1046, and his reactions are worth quoting: The place has an enormous fortification made of white stone . . . The top of the rampart is all crenellated and looks as though the master builder had just finished work on it. The city has one gate on the west side set in a large gateway with a masonry arch and an iron door with no wood in it. It has a Friday mosque that would take too long to describe.23

Clearly Nasir-i Khusraw was struck by the sheer scale of the fortifications, and that chimes with the tenor of Ibn al-Azraq’s comments. He notes the whiteness of the stone, a stark contrast to the black basalt of the walls of Amid, which he praises in hyperbolic terms. And he lays special stress on the iron gate with no wood in it. This obviously added much to the strength of the city’s defences; the norm in medieval cities of the Mashriq was to have gates of wood reinforced with iron spikes. But this was still the most vulnerable point in the defences. A gate made entirely of iron was a significant technical undertaking. There is now no trace of the ramparts admired by Nasir-i Khusraw, but the largest surviving tower of whitish stone, dressed in large blocks carefully fitted together, strengthened by molten lead and

a cosm op ol ita n f ro ntie r s ta te  | 367 furnished with projecting bosses, again represents a significant technical achievement. The news of this imposing building project would have spread far and wide and word of it would certainly have reached his neighbours. Small wonder, then, that ambassadors from all four of the larger neighbouring powers should hasten to test the veracity of these reports. And their visits were timed precisely to coincide with the end of the building programme. So, it is anything but coincidence that we read of the sumptuous reception planned by the amir for the end of the last month of the very year in which he had begun work on this grand project, a project which had clearly from the start been intended to mark the arrival of a new power in the land, a ruler to be reckoned with. And so it proved. The brand-­new buildings proclaimed his wealth, his power and his ambition, and the ambassadors of the Fatimids, of the caliph in Baghdad, of the Buyids who controlled him, and of the Byzantine emperor were there to appraise the newcomer in his seat of power. The amir must have had faith in the ability of the master builder and his men to complete the work on time, for the simultaneous arrival of ambassadors from three polities that were far away from M ­ ayyafariqin – ­Baghdad, Constantinople and ­Cairo – ­implies that they had received the invitation to attend some considerable time in advance. As Ibn al-­Azraq writes: On the fourth day of the ‘id, Nasr al-­Dawla sat on the throne to celebrate the ‘id. The emissaries of the caliph and the sultan sat on the right and the emissaries of Egypt and Byzantium sat on the left. Poets and reciters came. It was a great day and a blessed ‘id . . . Proclamations were read out to the people in the presence of the envoys and the commanders. The amir wore robes of honour and he gave robes of honour to the envoys, the like of which there was not.24

This account of Ibn al-­Azraq vividly evokes a new repurposing of an ancient tradition that stretched back to Sasanian times, namely the family of kings gathered around the throne of their overlord. That was by no means the political reality in 403/1012, but for those local people gathered together to participate in that grand reception, the sight of their amir enthroned amidst the representatives of all the great neighbouring powers, with the seating plan making explicit their lower status on this occasion, would have been

368  |  the me d ie va l tur k s unforgettable. The stage management of this occasion, timed to coincide with the first day of the ‘id, compels admiration as a piece of political theatre. It is interesting to speculate on the background to this grand occasion in Mayyafariqin in 403/1012. It was surely masterminded by Nasr al-­Dawla or his vizier, or by both of them in concert. Perhaps they had spread the news of the achievements of Nasr al-­Dawla across the borders into Iraq, Syria and Byzantium. An alternative possibility would be that the four emissaries had been sent by two caliphs, the Byzantine emperor and a Buyid amir to establish good relations with Nasr al-­Dawla, the new ruler of Diyar Bakr, an important state that bordered or was close to their own territories. But the precise timing of the occasion makes that somewhat unlikely. It is also significant that Nasr al-­Dawla was willing to welcome and honour envoys sent by both Sunni and Shi‘ite caliphs as well as the Christian emperor of Byzantium. This event was typical of the strategy of accommodation and self-­preservation consistently used by Nasr al-­Dawla vis-­à vis the three great powers of the time: Byzantium, Egypt and Baghdad. He made it his business to cultivate friendly relations with all of them, even the redoubtable Byzantine emperor. Within his own territory, as Ibn al-­Azraq writes, ‘his command was strong. There was nobody who defied him’.25 What of the government, the court and the public works of Nasr al-­Dawla during his extremely long reign? As early as the year 403/1012, according to Ibn al-­Azraq, Nasr al-­Dawla was busy ordering that parts of the walls of Mayyafariqin which had collapsed should be built up again, and he placed small dwelling places for the citizens on the walls.26 In the year 414/1023 he built and endowed a hospital and renovated the Friday mosque.27 Also in that year he constructed the minaret of the mosque in the suburb. He built an excellent bridge. So clearly he fulfilled one of the basic duties of a virtuous Muslim ruler, namely to build widely for the public good. Indeed, he increased the amenities of Mayyafariqin substantially. He restored the old observatory, put a clock in the Friday mosque, planted the citadel garden, repaired and added to the city walls,28 constructed and endowed several bridges, and erected public baths. He also made Mayyafariqin a hub of religious scholarship, and his court was a safe haven for many a political refugee. Happily Nasr al-­Dawla also had a lighter side. Ibn al-­Athir provides with obvious relish an account of his lifestyle:

a cosm op ol ita n f ro ntie r s ta te  | 369 He lived a life of ease and comfort unheard-­of by any other of his contemporaries. He possessed singing girls, some of whom he had purchased for five thousand dinars or more. He maintained five hundred concubines, and five hundred eunuch . . . He sent cooks to Egypt and spent a vast sum on their mission until they had learnt the local cuisine. He sent an enormous present to the Seljuq sultan Tughril Beg, part of which was the ‘Ruby Mountain’ which had belonged to the Buyids. In addition he sent with it one hundred thousand dinars.29

Thus this shrewd and experienced ruler of a small border state knew how to ingratiate himself with an intimidating and dangerous newcomer from the distant east. Nasr al-­Dawla placed a great deal of responsibility for the governing of his state on his viziers, two of whom were famous in their time. The first one was Abu’l-­Qasim al-­Husayn al-­Maghribi, who came from a family of famous bureaucrats. Ibn al-­Azraq describes him as ‘the administrator (mudir) of the state and the lord of its politics’.30 The other vizier of special note was the even more famous Fakhr al-­Dawla Ibn Jahir who remained in the position of vizier for the rest of the reign of Nasr al-­Dawla. Both men had served other masters of high rank. These men were brilliantly chosen. Al-­Maghribi is warmly praised by Ibn Khallikan, who mentions him in his biography of Nasr al-­Dawla, extolling the vizier’s erudition in both the arts and the sciences and mentioning his previous employment as secretary of state to the ‘Uqaylid amir and several other dynastic leaders in the area.31 Even more importance is accorded by Ibn Khallikan to Ibn Jahir (d. 1090), who went on to exercise great power in Baghdad and to whom he devotes an unusually lengthy biography.32 So under Nasr al-­Dawla the Marwanid state was in the safest of hands. Nasr al-­Dawla also enjoyed a harmonious relationship with the citizens of Mayyafariqin, distributing charity on a regular basis, and his rule ushered in a period of great prosperity. Only his death put an end to this idyll. Prestigious marriages were a valuable tool in the political strategy of the Marwanids. A typical example was a marriage arranged between a member of the Marwanid family and an Armenian princess from the Sinasina tribe who owned strong fortresses near Akhlat. Furthermore, Nasr al-­Dawla married a

370  |  the me d ie va l tur k s lady called al-­Fuduliyya, the daughter of Fadlun b. Minuchihr, the lord of the province of Arran and upper Armenia. This wife bore him two sons.33 The Marwanids took suitable measures to guard against internal rivals and external threats. In gaining control of the city of Mayyafariqin and their other territorial possessions in Diyar Bakr, the Marwanid rulers were violent and ruthless, not hesitating to banish from the city undesirable and hostile elements; Ibn al-­Azraq records that Nasr al-­Dawla banished from Mayyafariqin ‘those who were sinful and harmful and evildoers’.34 It happened quite frequently that opponents were murdered. Speaking of a potentially dangerous rival of Nasr al-­Dawla, a man called Sharwa, Ibn al-­Azraq records laconically that Nasr al-­Dawla ‘strangled him and crucified his corpse’.35 On the matter of the defence of the Marwanid state and its capital, Mayyafariqin, the city proved to be a most impressive stronghold against possible enemy attacks. The Marwanids would have heard about Turkish nomads streaming into south-­eastern Anatolia and threatening Byzantine territories and they were also aware of rumours of even more dangerous threats from the east after the Seljuq leaders Tughril and Chagri had taken possession of Khurasan, and Tughril and his nomadic followers were moving westwards across Iran. A taste of what was to come occurred in the year 434/1042 and it directly involved the Marwanids. Ibn al-­Azraq writes a long account of the arrival in Diyar Bakr of Turkish troops from the east. Tughril sent two amirs, accompanied by 10,000 horsemen, to Diyar Bakr, having given the area to them as iqta‘s. They arrived, raided the lands, plundered and then encamped outside the gates of Mayyafariqin, which remained closed for several days. Conversation between those outside the walls and those inside continued for some time. The account of Ibn al-­Azraq continues as follows: They (the Turcomans) were offered about fifty thousand dinars to retreat but they did not agree to that. It happened one night that they drank and got ­drunk . . . ­and they quarrelled and came to blows. Each one of them struck his neighbour with a knife and they both fell dead . . . The amir (Nasr al-­Dawla) and his troops went out, plundered what they had, killed many of them and took a great number of prisoners . . . This was the first appearance of the Turcomans in this country.36

a cosm op ol ita n f ro ntie r s ta te  | 371 This recalls the words of Matthew of Edessa when the Turks invaded Armenia in 467/1018–19: it was the ‘first irruption of fierce bloodthirsty beasts, the savage nation of infidels called Turks’.37 It is time to attempt an overview of the reign of Nasr al-­Dawla. He emerges as a flamboyant ruler with plenty of political acumen and extravagant tastes. His religious stance appears to have been a pragmatic one, suitable for the ruler of a vulnerable buffer state which was surrounded by powers of the most divergent ethnic and confessional loyalties. It seems likely that he ruled a predominantly Christian population in the towns of the Diyar Bakr province and that he enjoyed a good relationship with Byzantium. The Marwanid capital attracted prominent Muslim religious figures. It is noteworthy that in the reign of Nasr al-­Dawla, the famous religious scholar ‘Abd Allah al-­Kazaruni (d. 455/1063) went to Mayyafariqin and spread the Shafi‘i madhhab throughout Diyar Bakr.38 Shaykh Abu Nasr al-­Manazi, a high official of Nasr al-­Dawla, collected books and established waqfs for libraries in the mosques of Mayyafariqin and Amid.39 Poets, among them al-­Tihami, sought out Nasr al-­Dawla and they were lavish in their praise of him.40 Nasr al-­Dawla also took good care to remain on friendly terms with the local Christians. Thus, when he undertook building operations on the hill where the convent and Church of the Virgin had formerly stood, he had their relics transferred to the Melkite church.41 Nasr al-­Dawla died in 453/1061. Ibn al-­Athir writes warmly about his rule: He was 80 and more years of age, and had been emir for 52 years. He had absolute control over affairs in his lands and kept the frontier provinces flourishing and in order. He lived a life of ease and comfort unheard of for any other of his contemporaries.42

The death of Nasr al-­Dawla in 453/1061 heralded the final stage of Marwanid rule; thereafter, the power and prestige of the dynasty declined markedly. His son, Nizam al-­Din Nasr, succeeded him, at first only in Mayyafariqin and then two years later in Amid too. On his death (472/1079) his son Nasir al-­Dawla Mansur, the last Marwanid ruler, came to power. The vizier Ibn Jahir, who had left Diyar Bakr for Baghdad, turned against the son of his former master, Nasr al-­Dawla, and used his influence with the Seljuq sultan

372  |  the me d ie va l tur k s Malikshah and his vizier Nizam al-­Mulk to persuade them to bring the Marwanid dynasty to an end and to seize their treasures. In 478/1085 Diyar Bakr fell to Ibn Jahir and direct Seljuq Turkish control was imposed.43 Ibn Jahir took their treasury for himself and the last Marwanid ruler, Mansur, was given Jazirat Ibn ‘Umar, where he lived on until 489/1096. And now to conclude. Hugh Kennedy has rightly called the period 950– 1050 ‘a Kurdish interlude’44 during which Nasr al-­Dawla was renowned for excellent government and for enjoying good relations with local civilian elites. The eastern frontier between the Byzantine empire and the Muslim states on the border remained relatively stable in this period, and much of the credit for this must be laid squarely on the shoulders of this astute and far-­sighted ruler who enjoyed the good things of life while keeping an eagle eye on both the big and the little picture. He was in many ways a worthy precursor of Saladin. Notes   1. Ripper 2000, 41–2, 323–6.   2. Ripper, 63–6, 68–70; Minorsky 1953, 1–59.   3. Beihammer 2017, 57–61; Holmes 2001, 41–56.   4. Kennedy 2016, 264.   5. Al-­Muqaddasi 2001, 119.   6. Ibn al-­Athir 1979, IX, 35–8, 71–3, 349, 362, 397 and X, 10, 17–18.   7. British Library Ms. Or. 5803.   8. Ibn al-­Azraq 1959.   9. Amedroz 1902 and 1903. 10. Hillenbrand 1991, VI, 626–7, 930–2. 11. Ibn al-­Athir, IX, 35–6: Ibn al-­Azraq, 49–52. 12. Michael Psellus (1966), 46. 13. Ibn al-­Azraq, 84. However, Bar Hebraeus mentions Basil’s successful campaign in the east a few years earlier: Bar Hebraeus (1927), 180. Ostrogorsky also refers to Basil’s return to the east, including Syria and the Caucasus region, ‘several years’ after 995: Ostrogorsky (1968), 308. 14. Ibn al-­Azraq, 86. 15. Lehmann-­Haupt 1910, I, 424, with plate. 16. Ibn Khallikan 1970, I, 157–8. 17. Lehmann-­Haupt 1910, I, 423. 18. Kennedy, 264.

a cosm op ol ita n f ro ntie r s ta te  | 373 19. Al-­Muqaddasi 2001, 140. 20. Ibn al-­Azraq, 107. 21. Ibn al-­Azraq, 108. 22. For the topography of the city, see Markwart 1930, 193–8. 23. Nasir-­i Khusrau 1986, 7. 24. Ibn al-­Azraq, 110. 25. Ibn al-­Azraq, 104. 26. Ibn al-­Azraq, 110. 27. Ibn al-­Azraq, 122. 28. Van Berchem 1906, 129–32. 29. Ibn al-­Athir, X, 17–18. 30. Ibn al-­Azraq, 103. 31. Ibn Khallikan, I, 158. 32. Ibn Khallikan, IV, 280–7. 33. Ibn al-­Azraq, 121. 34. Ibn al-­Azraq, 102. 35. Ibn al-­Azraq, 103. 36. Ibn al-­Azraq, 160–1. 37. Matthew of Edessa 1858, 40–1. 38. Matthew of Edessa, 52. 39. Ibn al-­Azraq, 131. 40. Ibn al-­Azraq, 144. 41. Ibn al-­Azraq, 107–8. However, Nasr al-­Dawla repeatedly fought the Byzantines in a bid to control of Edessa (Honigmann 1935, 134, 136–8). 42. Ibn al-­Athir, tr. Richards, 2002, 135. 43. Ibn al-­Athir, tr. Richards, 140. 44. Kennedy, 266.

Bibliography Primary Sources Bar Hebraeus (1927), Makhtebhanuth Zabhne, partial tr. E. A. Wallis Budge as The Chronography of Gregory Abu’l Faraj, Amsterdam. Ibn al-­Athir (1979), Al-Kamil fi’l-ta’rikh, ed. C. J. Tornberg, IX and X, reprint Beirut. Ibn al-­Athir (2002), Al-Kamil fi’l-ta’rikh, partial tr. D. S. Richards, as The Annals of the Saljuq Turks, London.

374  |  the me d ie va l tur k s Ibn al-­Azraq (1959), Ta’rikh Mayyafariqin wa-Amid, ed. B. A. L. ‘Awad, Cairo. Ibn Khallikan, Wafayat al-a‘yan, tr. M. de Slane (repr. 1970) as Ibn Khallikan’s Biographical Dictionary, Beirut. Matthew of Edessa (1858), Armenia and the Crusades in the Tenth to Twelfth Centuries: The Chronicle of Matthew of Edessa, tr. A. E. Dostorian, Lanham, NY. Michael Psellus (1966), The Chronographia, tr. E. R. A. Sewter, as Michael Psellus. Fourteen Byzantine Rulers, Harmondsworth. Michael the Syrian (1905–10), Chronique de Michel le Syrien, Patriarche Jacobite d’Antioche (1169–99), ed. and tr. J. B. Chabot, Paris. al-­Muqaddasi (2001), Ahsan al-taqasim fi ma‘rifat al-aqalim, tr. B. A. Collins, as The Best Divisions for Knowledge of the Region, Reading. Nasir-­i Khusraw (1986), Safarnama, tr. W. M. Thackston, Albany NY. Secondary Sources Amedroz, H. F. (1902), ‘Three Arabic manuscripts on the history of the city of Mayyafariqin’, in Journal of the Royal Asiatic Society, 785–812. Amedroz, H. F. (1903), ‘The Marwanid dynasty at Mayyafariqin in the tenth and eleventh centuries a.d.’, in Journal of the Royal Asiatic Society, 123–54. Beihammer, A. (2017), Byzantium and the Emergence of Muslim-Turkish Anatolia, c. 1040–1130, Abingdon and New York. Cahen, C. (1988), La Turquie Pré-Ottomane, Istanbul. Canard, M. (1953), Histoire de la dynastie des Hamdanides de Jazira et de Syrie, Algiers. Hillenbrand, C. (1991), Encyclopaedia of Islam, 2nd edition, VI, art: ‘Marwanids’, 626–7. Hillenbrand, C. (1991), Encyclopaedia of Islam, 2nd edition, VI, art: ‘Mayyafariqin’, 930–2. Holmes, C., ‘How the east was won’, in Eastern Approaches to Byzantium (2001), ed. A. Eastmond, Aldershot, Burlington USA, Singapore, Sydney, 41–56. Honigmann, E. (1935), Die Ostgrenze des Byzantinischen Reiches von 363 bis 1071 nach griechischen, arabischen, syrischen und armenischen Quellen, Brussels. Kennedy, H. (2016), The Prophet and the Age of the Caliphates: The Islamic Near East from the Sixth to the Eleventh Century, Abingdon. Lehmann-­Haupt, C. F. (1910), Armenien einst und jetzt, Berlin. Markwart, J. (1930), Südarmenien und die Tigrisquellen nach griechischen und arabischen Geographen, Vienna. Minorsky, V. (1953), Studies in Caucasian History, London.

a cosm op ol ita n f ro ntie r s ta te  | 375 Ostrogorsky, G. (1968), tr. J. M. Hussey as History of the Byzantine State, Oxford. Ripper, T. (2000), Die Marwaniden von Diyar Bakr, Würzburg. Van Berchem, M. (1906), ‘Arabische Inschriften’, in C. F. Lehmann-­ Haupt, Materialien zur älteren Geschichte Armeniens und Mesopotamiens (Abhandlungen der königlichen Gesellschaft der Wissenschaften zu Göttingen. Philologisch-­ Historische Klasse. Neue Folge Band IX, No. 3, Göttingen), 125–60.

Original Sources of the Items in this Volume

  1. ‘The career of Najm al-­Din İl-­Ghazi’, Der Islam 58/2 (1981), 250–91.  2. ‘The establishment of Artuqid power in Diyar Bakr in the twelfth century’, Studia Islamica LIV (1981), 129–53.   3. ‘The history of the Jazira: A short introduction’, in Julian Raby (ed.), The Art of Syria and the Jazira, 1100–1250 (Oxford: Oxford University Press, 1985), 9–19.  4. ‘Malazgird’, Encyclopaedia of Islam (2nd edition), Vol. VI, C. E. Bosworth, E. van Donzel, B. Lewis and Ch. Pellat (eds) (Leiden: Brill, 1991), columns 243a–244b.   5. ‘Marwanids’, ibid., columns 626a–627a.   6. ‘Mayyafariqin’, ibid., columns 930a–932b.  7. ‘Mu‘in al Din Parwana: The servant of two masters?’, in F. de Jong (ed.), Miscellanea Arabica et Islamica, Orientalia Lovaniensia Analecta 52 (Louvain: Peeters Press, 1993), 267–74.   8. ‘Mu‘in al-­Din Sulayman Parwana’, in The Encyclopaedia of Islam (2nd edition), Vol. VII, C. E. Bosworth, E. van Donzel, W. P. Heinrichs and Ch. Pellat (eds) (Leiden: Brill, 1993), columns 479a–480b.   9. ‘1092: A murderous year’, in A. Fodor (ed.), Proceedings of the 14th Congress of the Union Européene des Arabisants et Islamisants (Budapest: Eötvös Loránd University, 1995), 281–96. 10. ‘Ibn al-‘Adim’s biography of the Seljuq sultan, Alp Arslan’, in Concepción Vazquez de Benito and Miguel Ángel Manzano Rodríguez (eds), Actas XVI Congreso Union Européene des Arabisants et Islamisants (Salamanca: CSIC Press, 1995), 237–42. 11. ‘The power struggle between the Saljuqs and the Isma‘ilis of Alamut, 487–

376

o ri gi na l sou rce s  | 377 518/1094–1124: The Saljuq perspective’, in Farhad Daftary (ed.), Studies in Isma‘ili History and Thought (Cambridge: Cambridge University Press, 1995), 205–12. 12. ‘Some reflections on Seljuq historiography’, in Antony Eastmond (ed.), Eastern Approaches to Byzantium (Aldershot: Ashgate, 2000), 73–88. 13. ‘Women in the Seljuq period’, in Guity Nashat and Lois Beck (eds), Women in Iran from the Rise of Islam to 1800 (Illinois: University of Illinois Press, 2003), 103–20. 14. ‘Ravandi, the Seljuk court at Konya and the Persianisation of Anatolian cities’, Mésogeios (Mediterranean Studies) 25–6 (2005), 157–69. 15. ‘Artuqids’, in The Crusades: An Encyclopedia, Vol. 1, Alan V. Murray (ed.) (Santa Barbara, Denver and Oxford: ABC Clio, 2006), 110–12. 16. ‘What’s in a name? Tughtegin – ‘the Minister of the Antichrist’?’, in Omar Alí-­de-­Unzaga (ed.), Fortresses of the Intellect: Ismaili and Other Islamic Studies in Honour of Farhad Daftary (London: I. B. Tauris, 2008), 459–71. 17. ‘Aspects of the Seljuq court’, in Christian Lange and Songül Mecit (eds), The Seljuqs: Politics, Society and Culture (Edinburgh: Edinburgh University Press, 2011), 22–38. 18. ‘Nizam al-­Mulk: A maverick vizier?’, in Edmund Herzig and Sarah Stewart (eds), The Age of the Seljuqs (The Idea of Iran, Vol. 6) (London: I. B. Tauris, 2015), 29–40. 19. ‘The life and times of ‘Amid al-­Mulk al-­Kunduri’, in A. C. S. Peacock and D. G. Tor (eds), Medieval Central Asia and the Persianate World (London and New York: I. B. Tauris, 2015), 161–73. 20. ‘The Nizamiyya madrasas’, in I. K. Poonawala (ed.), Turks in the Indian Subcontinent, Central and West Asia: The Turkish Presence in the Islamic World (Delhi: Oxford University Press, 2016), 144–57. 21. ‘The life and times of the Artuqid ruler Najm al-­Din Alpı (ruled 548/1154– 572/1176)’, in Lorenz Korn and Martina Müller-­Wiener (eds), Central Periphery? Art, Culture and History of the Medieval Jazira (Northern Mesopotamia), 8th–15th centuries: Studien zur Islamischen Kunst und Archäologie (Wiesbaden, 2017), 111–24. 22. ‘What is special about Seljuq history?’, in Sheila Canby, Deniz Beyazit and Martina Rugiadi (eds), Court and Cosmos: The Great Age of the Seljuqs (Edinburgh: Edinburgh University Press, 2020), 6–16. 23. ‘The view from above’: Muslim perceptions of the Turks of Syria and the Jazira in the period 1070 to 1176’, in Robert Hillenbrand (ed.), The Making of

378  |  the me d ie va l tur k s Islamic Art: Studies in Honour of Sheila Blair and Jonathan Bloom (Edinburgh: Edinburgh University Press, 2021), 49–70. 24. ‘A cosmopolitan frontier state: Relations among the Kurds, Arabs, Byzantines, Armenians, Persians and Turks under the Marwanids of Diyar Bakr, 990–1085’, in Deborah Tor (ed.), The Islamic–Byzantine Border: From the Rise of Islam to the Fall of Constantinople (forthcoming).

Index

Note: italic indicates illustrations, n indicates notes Abaqa, 108, 110–14, 115n, 117, 118–19 Abarsam, 274–5 ‘Abbasids Baghdad architecture, 291 court, 234 al-Ghazali, 160n, 357n İl-Ghazi, 35 Iraq, 77–8, 337 linguistic diplomacy, 276 Mayyafariqin, 98, 103 men of letters, 245 Mongols, 155 poets, 247 and Seljuqs, 130 Tughril, 330, 355 viziers, 295 women, 196, 197, 200n ‘Abd Allah al-Kazaruni, 95, 100, 371 al-Abiwardi, 246 Abu ‘Abdallah al-Tabari, 291 Abu ‘Ali al-Hasan b. Marwan, 94, 363 Abu Bakr, 61 Abu Hamid, 297–8n Abu Ishaq al-Shirazi, 290, 292 Abu Mansur Tughtegin ‘Abd Allah, 230n Abu Muhammad ‘Abd al-Wahhad al-Shirazi, 291 Abu Nasr Ahmad al-Manazi, 305 Abu Nasr b. ‘Umar al-Isfahani, 149 Abu Sa’id al-Nishapuri, 291

Abu Taghlib al-Ghandanfar, 99 Abu Tahir, 141–2 Abu’l-Hasan al-Basrawi, 141–2 Abu’l-Kasim al-Husayn al-Maghribi, 95 Abu’l-Mahasin al-Dihistani, 40n Abu’l-Qasim al-Maghribi, 365, 369 Abu’l-Wafa’, 99 Achaemenids, 234, 235 adab (secular literature), 170, 184–5, 194 ‘Adud al-Dawla, 94, 99, 363 al-Afdal, 6, 39n, 215 Aflaki, 112, 119 Agadzhanov, 262 Ager Sanguinis (Field of Blood), 25, 216–17, 348 Ahmad al-Ghazali, 297–8n Ahmad b. Husayn al-Hisnkayfi, 306 Ahmad b. Nizam al-Mulk, 151 Ahmad b. Suqman, 303 Ahmad Ibn ‘Attash, 150 Ahmadil, 159n Ajay, 118 Akhbar al-dawla al-saljuqiyya, 169, 179–80n, 243–4, 249 Akhlat Badh, 94, 363 Da’ud, 60 Romanus IV Diogenes, 90–1 Suqman al-Qutbi, 14, 27, 41n, 43n Aksaray, 109, 117

379

380  |  the me d ie va l tur k s ‘Ala al-Din Qayqubad I, 188 Alamut, 123–4, 126 Albert of Aachen, 222, 225, 226, 228, 2 31n, 352, 357 Albistan, 109–10, 113–14, 116n, 118, 119 Aleppo Alp Arslan, 138 Alpı, Najm al-Din, 311 Artuqids, 219 Badr al-Dawla Sulayman b. ‘Abd alJabbar, 51 Balak, 217–18 Bursuq b. Bursuq, 11 Hamdanids and Mirdasids, 361 Ibn al-‘Adim, 137, 138–40 İl-Ghazi, 21–2, 25, 29–30, 37, 43n, 44n, 52 Isma‘ilis, 34 Nizam al-Mulk, 258, 278 Roger of Antioch, 23, 32, 216, 348 Sayf al-Dawla ‘Ali, 98 Seljuqs, 337 Sulayman, 53–4 taxation, 47n Temürtash, 55, 57–9, 63 Turcomans, 13–14 Turks, 343, 346 ‘Ali al-Rida’, mausoleum, Mashhad, 191, 192 ‘Ali b. Munqidh, 137, 139 Alp Arslan Byzantines, 170–2 drunkenness, 139–40, 142–3 al-Kunduri, 272–4, 276 Malazgird, 89–93 name, 330, 355 nawba, 243 Nizam al-Mulk, 122, 256–9, 262–3, 288, 294 Seljuqs, 235, 332, 326 speech, 171–2 succession problem, 47n women, 194 Alpı, Najm al-Din, 302–20 bridges, 74n, 101 buildings, 101

ceremonial robe of honour, 305 coins, 307, 308 translation, 219 al-Alusi, The Mosques and Monuments of Baghdad, 293 Amedroz, H. F., 51, 362 Amid Artuqids, 77, 218 Badh the Kurd, 363 Basil II, 364 Diyar Bakr, 103 Isma‘ilis, 36 libraries, 100, 371 Marwanids, 94–6 Shaybanids, 98 Temürtash, 63 Zengi, 60 Anatolia Alp Arslan, 326 architecture, 209–10 Artuqids, 36 bridges, 306 Byzantine, 334 cultural life, 206–8 Islamisation of, 207 monumental inscriptions, 209 mosques and madrasas, 208–10 Mu‘in al-Din Sulayman Parwana, 108–20, 260 Persianisation of, 206–10 Seljuqs, 89, 163–81 Seljuqs of Rum, 333 Turcomans, 77 al-Anbari, 193 Antioch, 17, 20, 24, 25, 215, 217 Anushirwan b. Khalid, 125, 168–9, 280–1 Anvari, 247 Aq Qoyunlu, 103–4 Aq Sunqur al-Bursuqi, 5, 8–9, 27, 44n, 47n, 58, 216 Arabic sources, 168–70, 176–7, 236–7, 339–42 Arabs Bedouin, 94–5, 361 language, 342 and Marwanids, 361–75 poets, 246–7

ind ex |  381 Sadaqa, 156 Turks, 220–1, 338, 344, 355–6 Ardashir, 275 Armenia, 7, 17, 45n, 90, 98, 103, 202–3, 361–75 Arslan, 195 Arslan Khatun, 188, 271–2, 282n Artaxerxes I, 284n Artuq, 5, 13, 14, 64, 215 Artuqids, 215–19 administration, 29 buildings, 83, 219 court, 84 Diyar Bakr, 50–75, 77 family tomb, Mardin, 65 genealogy, 68–9 Ibn al-Azraq, 362 Ibn Khallikan, 39n İl-Ghazi, 5–6, 13, 28, 36, 41n Mayyafariqin, 47n, 100–1, 341, 365–6 political orientation of, 78–9 succession problem, 51–4 Sunni Islam, 79–80 women, 188 Arzan, 103 Ashraf Musa, 102 Assassins, 121–4, 126, 130, 131–2, 132n, 194 Atabeg Amir Qutb al-Din Kül-sarigh, 262 atabegs, 237, 267n, 330, 341, 346, 349, 355 Attaliates, 89, 90, 91 ‘Awad, 51, 362 Awhad Najm al-Din Ayyub, 102 Ayaz, 8, 10, 216 Ayyubids Artuqids, 215, 218 atabegs, 263 court, 84 Diyar Bakr, 77 Mayyafariqin, 101–2 religion, 81 Sunni Islam, 79, 129 Takiya bint Ghayth, 193–4 ‘Ayyuqi, Varqa va-Ghulshah, 245 Azerbaijan, 7, 98, 361 al-‘Azimi, 25, 56

Bab Hitta, Jerusalem, 95 Badh, 94, 99 Badh the Kurd, 363 Badr al-Dawla Sulayman b. ‘Abd al-Jabbar, 51, 54, 58 Badr al-Din Lu’lu’, 81, 82 Badr al-Jamil, 260–1 Baghdad ‘Abbasids, 77 Artuqids, 80 Artuqids and Zengids, 78 Banafsha bint ‘Abdallah al-Rumiyya, 192 buildings, 367 Dubays b. Sadaqa, 35 Hanbalites, 79, 82, 86n Ibn al-Azraq, 316n Ibn Jahir, 96, 369 İl-Ghazi, 6–11, 14, 18–19, 40n, 41n, 348 Isma‘ilis, 160n Kunduri, 271–2 Malikshah, Sultan, 121–3, 128 men of letters, 245 Mirjaniyya madrasa, 293 Mongols, 155 Mumahhid al-Dawla, 364 murder of sailor, 43n Nasr al-Dawla Ahmad, 367–8 nawba, 241 Nizam al-Mulk, 124–5, 331, 356 Nizamiyya madrasas, 289–91, 293, 295 Temürtash, 66 Tughril, 276–7, 325–6, 330, 355 Tughtegin, 31, 350 Zengi, 81 Baha’ al-Dawla, 364 Bahr al-fava’id (Sea of Precious Virtues), 185 al-Bakharzi, 247, 274 Balak Artuqids, 217–19 Da’ud, 60 Franks, 215 İl-Ghazi, 28, 54 Jazira, 85 Sulayman, 55 Temürtash, 57, 58–9 al-Bal‘ami, 256 Balat, 19–20, 23–7, 30, 32–3, 37, 41n

382  |  the me d ie va l tur k s Baldwin I, 215 Baldwin II, 57–8, 217, 223 Banafsha bint Abdallah al-Rumiyya, 192 Banu Marwan, Mayyafariqin, 100 Banu Nubata, 29, 83, 304 Bar Hebraeus, 34, 303, 305, 352, 372n Barkyaruq assassination attempt, 147–8 drunkenness, 147, 332 İl-Ghazi, 7–8, 41n Isma’ilis, 147–9, 156, 157 khutba, 123 Nizam al-Mulk, 128–9 praise for, 160n reign of, 145–9 Sadaqa, 40n Terken Khatun, 125–6 Barmakids, 295 Basil II, Emperor, 363–4, 364, 372n Batinis, 148, 150, 151 Batiniyya, 149, 152–3, 155–6 Bayazid, 103 Baybars, 109–14, 116n, 118, 119 Bayju, 109, 117 Bell, Gertrude, 102, 316n Berk-Yaruq, 40n Bianquis, Thierry, 260–1 al-Biruni, 245 Bishop John of Mardin, 316n Bitlis, 363 Bitlisi, Idris, Hasht Behesht (Eight Paradises), 165 Book of Marriage, 183 Bosworth, C. E., 267n, 275–6 Bowen, H., 132n Boz-Qush, 148 bridges, 65, 74n, 101, 104, 306–7, 368 Bu Tahir Arrani, 124 Bulliet, R. W., 208, 209, 289, 294 al-Bundari, 73n, 125, 147, 168–9, 273, 276, 277, 280–1, 291 Burids, 226, 237 Bursuq b. Bursuq, 9, 11, 27, 31, 40–1n, 216 Buyids Badh, 363 Hasanuyids, 361 Marwanids, 94

Mayyafariqin, 98–9 Mumahhid al-Dawla, 364 Nasr al-Dawla Ahmad, 367 nawba, 241–3 Sunni Islam, 79 women, 197 Buza’a, 23 Byzantines Diyar Bakr, 13 Ibn Bibi, 181n Jazira, 99 Manzikert, battle of, 170–4, 326 Marwanids, 361–75 Nasr al-Dawla Ahmad, 95, 367, 368, 371, 373n Ravandi, 167 Seljuqs of Rum, 202, 211, 214n Cahen, Claude Alpı, Najm al-Din, 302, 314 Artuqids, 1, 41n, 50 Assassins, 132n atabegs, 263 Ibn al-Athir, 180n Ibn al-Azraq, 51 İl-Ghazi, 29 Manzikert, battle of, 89 mosques and madrasas, 210 Mu’in al-Din Sulayman Parwana, 116n Pre-Ottoman Turkey, 108 La Syrie du Nord à l’époque des Croisades, 137 Temürtash and Christians, 66 Chaghri, 138, 241, 262, 325, 370 Chavlı Saqao, 9–10, 41n Cheynet, J. C., 89, 90 Chökermish, 5, 19, 41n, 47n Christianity Alp Arslan and Byzantine emperor, 172 Alpı, Najm al-Din, 305, 309 and castration, 272–3 Diyar Bakr, 95, 363 fall of Constantinople, 173–4 fanaticism, 80 Manzikert, battle of, 173 Mayyafariqin, 104–5 Nasr al-Dawla Ahmad, 368, 371

ind ex |  383 Shaddadids, 361 in Syria, 340 Temürtash, 66 and Turcomans, 82–3 coins Alpı, Najm al-Din, 307–9, 307, 308 Artuqids, 219 Ayyubids, 102 copper, 74n Seljuqs, 330 Temürtash, 65, 101 Constantine X, 95 Constantinople, 173–4, 330, 334, 355, 367 copper mine, 65, 101, 307–8 Counter-Crusade, 37, 40–1n Crawford, R. W., 49n Crusaders Artuqids, 77, 78, 80 Barkyaruq, 146, 147 chroniclers, 339 fall of Constantinople, 173–4 Gervase of Basoches, 222–5, 228–9, 357 İl-Ghazi, 348 Manzikert, battle of, 173 Muhammad, Sultan, 158 Muslim historians, 340 Nur al-Din, 342 Romanus IV Diogenes, 91 Seljuqs, 159n Tughtegin, 350 Turks, 220–2, 337–8, 356 Zengi, 80–1 Damascus Alpı, Najm al-Din, 311 Gervase of Basoches, 225, 229, 357 Hanbalites, 82, 86n İl-Ghazi, 39n Nur al-Din, 354–5 Seljuqs, 337, 346 al-Suhrawardi, 82 Sulayman, 54 Takiya bint Ghayth, 193–4 Tughtegin, 11, 30, 42n, 226, 230–1n, 349–50 Danishmendids, 203, 333 Danith, 11, 12

Dar al-Islam, 112, 119, 220–1, 309, 338 Dara, 60, 61, 304, 313 Daryaee, T., 248 Da’ud, 10, 28, 55, 59–62, 63, 85 David II, 217 al-Dawla, Sultan, 94 Daylam, 155 Daylamis, 156 De Slane, W. M., 244 Dimitri, 17, 26 Dioscorides De Materia Medica, 219, 305 Diyar Bakr Artuqids, 50–75, 77, 78, 84 bridges, 307 Ibn Jahir, 96, 100 İl-Ghazi, 8–15, 31–2, 35 Isma‘ilis, 36 Mayyafariqin, 103 Nasr al-Dawla Ahmad, 94–5 Safavids, 104 Sayf al-Dawla ‘Ali, 98–9 Suqman, 27–8 Temürtash, 80, 101 drunkenness Alp Arslan, 139–40, 142–3 Barkyaruq, 147 İl-Ghazi, 2, 24, 35, 348 Seljuqs, 332 Turks, 221, 338, 339 Dubays Aleppo, 44n, 53, 58 İl-Ghazi, 23, 48n al-Mustarshid, 35 Temürtash, 57 Tiflis, 17, 25–6 Duqaq İl-Ghazi, 5, 39n Mayyafariqin, 100 royal women and politics, 195 Tughtegin, 349 Turks, 220, 338 Edessa, 64, 90, 95, 216, 218, 343, 373n Edgington, S. B., 231n Egypt Artuqids, 218

384  |  the me d ie va l tur k s Egypt (cont.) Fatimids, 294, 337 hunting, 249 Jalal al-Din Rumi, 115n Mamluks, 164, 337 Mu‘in al-Din Sulayman Parwana, 112, 119 Mumahhid al-Dawla, 364 Nasr al-Dawla Ahmad, 367, 368 Encyclopaedia of Islam, 2, 270, 278 Esin, E., 228 Ewliya Čelebi, 104 Fadlun b. Minuchihr, 257, 370 al-Fadluya Shabankara’i, 243 Fairbanks, S., 260, 266n Fakhr al-Dawla Ibn Jahir, 95, 369 Fakhr al-Din ‘Ali, 111, 118 Fakhr al-Din ‘Iraqi, 111, 119 Fakhr al-Mulk, 364 Farkhunda Khatun, 35 Fatima Khatun, 188, 194 Fatimids Badr al-Jamil, 260–1 Hanbalites, 79 Jerusalem, 5 khutba, 49n, 95 Mumahhid al-Dawla, 364 Nasr al-Dawla Ahmad, 94–5, 367 Nizam al-Mulk, 294 Romanus IV Diogenes, 91 Syria, 337, 340 Field of Blood (Ager Sanguinis), 25, 216–17, 348 Fink, H. S., 41–2n fiqh, 183, 204, 257, 293 Firdawsi, Shahnama, 274–5 First Crusade, 2, 215–16 Firuzabad, India, 306 Flury, S., 102 Fourth Crusade, 173–4 Franks Aleppo, 15–16, 47n, 58 Alpı, Najm al-Din, 311–12, 314 Artuqids, 215–19 Balak, 28, 54 Constantinople, 174

counter-crusades, 40–1n Ibn al-Qalanisi, 223–4 İl-Ghazi, 2–3, 19–25, 30–2, 37 Jerusalem, 39n Mawdud, 41–2n Muhammad, Sultan, 9, 12 prisoners, 30, 46n Tughtegin, 11, 181n, 350 Friendly, A., The Dreadful Day, 90 al-Fuduliyya, 370 Fürstenspiegel genre, 144n, 164–5 Gabriel, A., 101, 102 Galen, 238 gazanama genre, 172 Georgia atabegs, 267n Da’ud, 60 Greater Iran, 203 Ibn al-Azraq, 79 İl-Ghazi, 12, 17, 26, 33–4, 53, 80 Shaddadids, 361 Georgian Chronicle, 46n Gervase of Basoches, 221–9, 352–3, 356, 357 Ghars al-Ni‘ma, 137, 141, 180n ‘Uyun al-tawarikh, 144n al-Ghazali Assassins, 124 Ihya’ ‘ulum al-din (Revivifying of the Sciences of Religion), 183 Isma‘ilis, 160n Kimiya-yi sa‘adat (Alchemy of Happiness), 183 Kitab al-Mustazhiri, 357n men of letters, 245 ‘Mirrors for Princes’ genre, 236 Nasihat al-muluk (Counsel for Kings), 185–6 Nizamiyya madrasas, 291, 293 Turks, 340 Ghaznavids, 138, 234, 236, 245, 262, 276 Ghuzzoğlu, 156 Gibb, H. A. R., 2, 40–1n, 40n, 57, 178n Girdkuh, 146, 156, 272, 279 Goitein, S. D., 278 Golden, Peter, 262, 267n

ind ex |  385 Goliath’s spring, battle of, 207 Gordlevski, V., 210, 214n Great Seljuqs, 163–81 Artuqids and Zengids, 78 court, 234–53 government, 203, 333 Iran, 207 Malikshah, Sultan, 157 Muhammad, Sultan, 154 Greeks, 91, 343 science, 238 Grousset, R., 2, 40–1n Grube, Ernst J., 316n Guibert of Nogent, 222, 352 Gulpayagan, 153 Gümüshtegin, 7 Habashi, Amirdad, 62, 156 hadith, 164–5, 167, 184, 185 al-Hakim, 94, 364 al-Halabi, 137 Hamadan, 235, 243 Hamdallah Mustawfi, 124, 291, 315 Hamdani, 91 Hamdanids, 94, 98–9, 347, 361, 363, 364 Hammer-Purgstall, J., 298n Hanafis, 206, 289 Hanbalites, 79, 82, 86n, 103 Hani, 63, 303–4 Harim, battle of, 311–12, 313 Harran, 32 Harran, battle of, 216 Harun al-Rashid, 234 al-Hasan b. ‘Ali b. Ishaq, 289 Hasan-i Sabbah, 131, 151, 155–6, 159 Sarguzasht-i Sayyidna, 123–4 Hasanuyids, 361 Hattakh, 65 Haysa Baysa, 246, 247 Hayton, 110, 115n, 118 Herodotus, 225 Hilal al-Sabi’, 241–3 Hilla, 7, 361 Hillenbrand, C., Zubdat al-halab fi ta’rikh Halab, 138 Hisn Kayfa Artuqids, 77, 84, 218–19

bridges, 307 buildings, 306–7 Da’ud, 10, 28, 59 Qara Arslan, 63, 312 Suqman, 8, 13, 47n, 215 women, 190, 196 Hodgson, Marshall, 337 Houtsma, M. T., 132n Hsiung-Nu, 228, 353 Hudud al-‘alam, 346 Hülegü, 102, 117, 155 Hulwan, 6–7 hunting Alp Arslan, 141 Great Seljuqs, 237, 240, 248–9 Malikshah, Sultan, 121–3, 332 Nizam al-Mulk, 259 al-Husayni atabegs, 262 Barkyaruq, 147 Ibn al-‘Adim, 137 Isma‘ilis, 151, 156 Mahmud, 243–4 Nizam al-Mulk, 257 Seljuqs, 256 Terken Khatun, 194–5 Zubdat al-tawarikh, 179–80n Ibn al-‘Adim Alp Arslan, 136–44, 144n Alpı, Najm al-Din, 312 Bughyat al-talab fi ta’rikh Halab, 136–44 Frankish prisoners, 46n Ibn al-Khashshab, 348 İl-Ghazi at Aleppo, 15, 29–30, 44n Nizam al-Mulk, 264, 296n, 298n Romanus IV Diogenes, 90 Seljuqs, 143n Sulayman, 53 taxation, 47n Temürtash, 55–6, 56, 58 Turks, 221, 338 Ibn al-Athir Aleppo, 44n Alpı, Najm al-Din, 303, 312 Barkyaruq, 146, 147–8 Constantinople, 173–4

386  |  the me d ie va l tur k s Ibn al-Athir (cont.) Da’ud, 59 Gervase of Basoches, 223–4, 229 History of the Atabegs of Mosul, 56–7 Ibn al-Qalanisi, 40n İl-Ghazi, 6–7, 18–32, 41n, 43n, 48n Jerusalem, 39n Kamil al-ta’rikh, 41n, 71n, 361 Kashani, Abu’l-Qasim, 146, 159n Kunduri, 274, 279 Malikshah, Sultan, 132n, 248, 327 Maslama, 334n Mayyafariqin, 101 Muhammad, Sultan, 150–1, 155, 161n Nasr al-Dawla Ahmad, 368–9, 371 Nizam al-Mulk, 257, 262–3 Nizamiyya madrasas, 289, 291–2 Persian language, 180n Sanjar, 156 Sulayman, 53–4 Temürtash, 63–6 Terken Khatun, 127 Tughril, 330 Tughtegin, 350 Turcomans, 14, 18–32, 83 Universal History, 152–4, 169 Zengi, 73n Ibn al-Azraq Artuqids, 72–3n Assassins, 126 Da’ud, 59 Ibrahim, 47n İl-Ghazi, 2, 13–14, 26, 29, 43n, 44n, 48n Mahmud, 72n Malikshah, Sultan, 127 Marwanids, 95 Mayyafariqin, 46n, 98–101, 104, 347 Mumahhid al-Dawla, 364 Nasr al-Dawla Ahmad, 365–70 peripatetic officials, 79 Sulayman, 54–6 Ta’rikh Mayyafariqin wa-Amid, 1, 17, 50–75, 302–20, 362–3 Temürtash, 60–6 Turcomans, 83 ‘ulama’, 316n women, 190

Ibn Badi‘, 139 Ibn al-Banna’, 142 Ibn Battuta, 186–7 Ibn Bibi Baybars, 116n Konya, 208 marriage alliances, 188–9 Seljuqs of Rum, 109, 165, 181n, 214n Ibn Butlan, 187 Ibn al-Faqih, 103, 346 Ibn Funduq, 165 Ibn al-Furat, 34, 54–6, 224, 229, 352–3 Ibn al-Habbariyya, 247 Ibn Hassul, 276, 279, 346 Ibn Hubayra, 86n Ibn al-‘Imrani, 279, 282n Ibn Jahir, 13, 51–2, 96, 100, 369, 371–2 Ibn al-Jawzi al-Ghazali, 291 Isma‘ilis, 147–8, 152 Kitab ahkam al-nisa’ (Book of Rules for Women), 183–4 Malikshah, Sultan, 259 Manzikert, battle of, 171–2 women, 188, 194 Ibn Jubayr, 190–1, 192, 197, 292, 293 Ibn Khallikan Abu Nasr Ahmad al-Manazi, 305 Abu’l-Qasim al-Maghribi, 95 Artuqids, 39n atabegs, 263 Da’ud, 262 Ibn al-Qattan, 247 Kunduri, 274, 280 Malikshah, Sultan, 83, 248, 332 Mayyafariqin, 127 Muhammad, Sultan, 161n Nizam al-Mulk, 124 Nizamiyya madrasas, 294, 297n al-Suhrawardi, 82 Tughril, 81 viziers, 369 Wafayat al-a‘yan, 193–4 Ibn al-Khashshab, 21–2, 32–3, 43n, 44n, 348 Ibn Mukhtar, 66 Ibn Muyassar, 261

ind ex |  387 Ibn Nubata, 99, 304, 347 Ibn al-Qalanisi Amedroz, H. F., 51 Assassins, 126 Gervase of Basoches, 223, 228–9 Ghars al-Ni’ma, 180n Ibn al-Athir, 40n İl-Ghazi, 17–21, 41n, 44n Isma‘ilis, 42n, 149 Manzikert, battle of, 170 Seljuqs, 146 Temürtash, 55 Tiflis, 26 Tughtegin, 230–1n, 349–50 Turcomans, 23 Turks, 341, 343, 348 Ibn al-Qattan, 247 Ibn al-Sa‘i, 187, 189, 193, 195, 196, 197 Ibn Shaddad Alpı, Najm al-Din, 101, 306 Hisn Kayfa, 84 İl-Ghazi, 35–6 Mamluks, 116n Mayyafariqin, 98, 103–5 Saladin, 313 Ibrahim, 27–8 Ibrahim, Amir, 14 Ibrahim b. Suqman, 28, 47n Il-Aldi, 60 İl-Ghazi, 1–49 administration, 29–34 at Aleppo, 15–18 Alpı, Najm al-Din, 311, 314 Artuqids, 77, 80, 215–19 ceremonial robe of honour, 25–6, 32, 35, 348 coins, 309 Crusaders, 78, 350 death of, 195 drunkenness, 2, 24, 35, 348 Ibn al-Azraq, 304 Islam, 35–6 Isma‘ilis, 34, 82 jihad, 347–8 Mayyafariqin, 71n, 101, 347 military ability, 18–27 Nur al-Din, 354

personality, 36 political ability, 27–9 relations with the caliphate, 35 religion, 35–6 Savran, 302 succession problem, 51–4 Turcomans, 18–22 Turks, 342 Ilkhanids, 103, 110, 111 ‘Imad al-Din al-Isfahani, 125, 168–9, 280–1, 291, 293–4 Kharidat al-qasr, 247 Inanch Khatun, 195, 196 iqta‘, 5, 6, 48n, 52, 370 Iran architecture, 209 Barkyaruq, 145–6, 157 Chaghri and Tughril, 325, 370 Muhammad, Sultan, 6 Nizam al-Mulk, 254–69 Persian historiographical tradition, 165 Seljuqs, 78, 203, 234, 328 Turks, 337, 346 Iraq Barkyaruq, 145–6, 157 dynastic histories, 165 Hasanuyids, 361 Ibn al-Azraq, 51 İl-Ghazi, 6 İl-Ghazi and Zengi, 78 Malikshah, Sultan, 127 al-Nasir, 77 Tughril, 271 Turks, 238, 337, 341, 346 Isfahan Barkyaruq, 148–9 Friday mosque, 150 Hanafi madhhab, 192 İl-Ghazi, 8 Isma‘ilis, 150, 152, 158 Malikshah, Sultan, 127–8, 235, 327–8 Muhammad, Sultan, 153 Nizamiyya madrasas, 289 patronage, 192 Terken Khatun, 127 Tughril, 325 Isma’ Khatun bint Malikshah, 192

388  |  the me d ie va l tur k s Isma‘il I, 104, 228 Isma‘ilis Aleppo, 16 Assassins, 121, 126 Hanbalites, 79 Jazira, 82 Mawdud, 11 Nizamiyya madrasas, 294 Ridwan, 49n Seljuqs, 145–62, 331 taxation, 150 Temürtash, 65 al-Istakhri, 103 ‘Iyad b. Ghanm, 98 ‘Izz al-Din Kayka’us I, 189 Ja‘ar, 125, 129, 130, 131 al-Jahiz, 196, 324, 344–5, 348 Jalal al-Din Rumi, 333 Fihi ma fihi, 111–12, 119 Jamal al-Din, 303–4 al-Jamaliyya Baltun bint ‘Abdallah, 191 Al-Jazari, Kitab fi Ma‘rifat al-Hiyal alHandasiyya, 242 Jazira, 76–88 Alpı, Najm al-Din, 302, 313, 314 Artuqids, 50–1 bridges, 307 buildings, 305–6 Byzantines, 99 İl-Ghazi, 14, 28, 36–7, 41n Marwanids, 95 Mawdud, 9 Mayyafariqin, 98 Muhammad, Sultan, 12 Muslim learning, 305 taxation, 83 Turks, 340–2, 347–8, 350, 356 women, 190 Jazirat Ibn ‘Umar, 96, 307, 372 Jekermish, 216 Jerusalem Alpı, Najm al-Din, 311 Aqsa Mosque, 353 Artuqids, 5–6, 39n, 215 Ayyubids, 81 Crusaders, 337

Temürtash, 64 Zengids, 77 jihad Alpı, Najm al-Din, 312–14 Artuqids, 80 Balak, 28 gazanama genre, 172 Ibn Nubata, 99 İl-Ghazi, 11, 21–2, 32, 35, 348 Isma‘ilis, 82 Manzikert, battle of, 173 Muhammad, Sultan, 153–4 Nur al-Din, 354 Seljuqs, 175 Shaddadids, 361 Tughtegin, 350 Turks, 221, 331, 338, 347, 356, 357n Zengids, 80–1 John Tzimisces, 99 Joscelin I, 217 Joscelin II, 62, 63, 218 Joseph Trachaniotes, 90 al-Juwayni Alamut, 154–5, 157 Assassins, 123–4, 126 Isma‘ilis, 151 Nishapur, 289 Nizamiyya madrasas, 289–90 Kafesoğlu, Ibrahim, 145 al-Kalkali, 156 Kamal al-Din Muhammad Khazin, 260 Kashani, Abu’l-Qasim, 146–7, 159n Kaykhusraw I, 167, 179n, 204, 205 Kaykhusraw II, 109, 117, 209 Kaykhusraw III, 109, 117 Kaymaz, N., 108 Kayqubadh I, 209 Kayseri, 110, 113, 118, 192 Kennedy, Hugh, 245, 372 Khalidi, T., 136 Kharput, 215, 217 Khartabirt, 55, 60 khatun, 52, 54, 139–40 al-Khazi, 45n Khurasan Barkyaruq, 146

ind ex |  389 Ibn al-‘Adim, 137 Ibn al-Qalanisi, 341 Kunduri, 271, 276 Nishapuri, 166 Nizam al-Mulk, 236, 258, 278, 288–91 Nizamiyya madrasas, 293, 295 Persian historiographical tradition, 164 Sanjar, 6, 78 Tughril, 370 khutba Alpı, Najm al-Din, 303 Banu Nubata, 304 Barkyaruq, 7, 8, 123 Fatimids, 49n Mahmud, 123, 127 al-Mustansir, 95 Sulayman b. Chaghri, 272 Tughril, 271 King David the Restorer, 17, 26 Kitab al-Wuzara, 144n Klausner, C. L., 258 Kononov, A. N., 244 Konya, 110, 117, 166–7, 190–1, 202–14, 333, 334 Köse Dagh, battle of, 109, 117, 118, 206, 209 Kosh Hisar, battle of, 104 Kuhistan, 148, 151 al-Kunduri, 270–87 al-Bakharzi, 247 castration, 132n, 272–5 Nizam al-Mulk, 122 Ravandi, 167 Shafi’is, 289 Kurds Anatolia, 202 Marwanids, 94, 99, 361–75, 362 Mayyafariqin, 103–5 Sadaqa, 156 Lambton, A. K. S., 255, 267n, 278 Lane-Poole, Stanley, 311 Lehmann-Haupt, C. F., 100, 102 Letter of Tansar, 237 Lewis, Bernard, The Assassins, 121, 148 libraries, 82, 279, 298n, 371

Liparit, Prince, 95 Lyons, Malcolm, 224 Lyons, Ursula, 224 madhhab, 95, 100, 290, 295, 296, 371 Mah Malik Khatun, 125 Mahmud (son of Sulayman), 72n Mahmud (son of Terken Khatun), 123, 125–6, 127, 129, 194–5 Mahmud, Sultan Haysa Baysa, 246 hunting, 248 İl-Ghazi, 31–2, 35 Mayyafariqin, 52 Muhammad, Sultan, 42n Sanjar, 151, 156, 243–4 Temürtash, 11, 48n, 57 Tughril, 44–5n Mahmud b. Nasr b. Salih, 139–40 Mahmud of Ghazna, 245 Mahperi Khatun, 192 Makdisi, George, 270, 275–6, 278, 281, 294 Malazgird, 89–93 Maliknama, 137, 180n Malikshah, Sultan, 327 Artuq, 13 climbing walls, 265n daughter of, 189 death of, 4–5, 77, 121–35, 132n, 220, 338 Great Seljuqs, 157 hunting, 248, 259, 332 Ibn Jahir, 96, 372 Ibn Khallikan, 83 İl-Ghazi, 8, 36 Isfahan, 235 Mayyafariqin, 100 man of letters, 245–6 al-Muqtadi, 135n name, 354 Nizam al-Mulk, 256–9, 262–4, 288 Nizamiyya madrasas, 292 Qizil Sarigh, 151 Seljuqs, 326–8, 346 succession problem, 47n Terken Khatun, 71n, 194–5

390  |  the me d ie va l tur k s Mamluks historians, 168 Mu‘in al-Din Sulayman Parwana, 108–16, 116n, 118 relations with Mongols, 108–16 Seljuq historiography, 164 Turks, 337 mamluks, 259, 321, 332–3, 349 Manbij, siege of, 57, 217 Manzikert, battle of, 77, 89–93, 138, 167, 170–5, 180n, 235, 326 modern Turkish celebratory stamp, 328 Maqdisi, G., 79 Mardin Alpı, Najm al-Din, 302–20 Artuqids, 101, 215–19 Diyar Bakr, 103 Great Seljuqs, 78 İl-Ghazi, 9–16, 20, 27–9, 32, 34–7, 41n, 44n Jacobite church, 80 religion, 81 Sunni madrasa, 35–6 Suqman, 8, 47n Temürtash, 51, 55–61, 63, 65–6 Zengids, 77 Marwan, 363 Marwanids, 94–7 Arabs, 361–75 Armenians, 361–75 Artuq, 13 Byzantines, 361–75 Kurds, 361–75, 362 Mayyafariqin, 47n, 51–2, 99–100 Persians, 361–75 religion, 104 Temürtash, 65 Turks, 361–75 Maslama, 334n Mas‘ud, 169, 260 al-Mas’udi, 273 Matthew of Edessa, 54, 89, 298n, 345, 371 Mawdud death of, 10–12, 27, 31, 42n İl-Ghazi, 41n, 216 Mosul, 9–12, 37 religious idealism, 41–2n

Mayyafariqin, 98–107 Alpı, Najm al-Din, 302–20 Artuqids, 83, 215–18, 341 Artuqids and Zengids, 77 Badh the Kurd, 363 Basil II, Emperor, 364 buildings, 100, 364–8 Christians, 104–5 city walls, 99–100, 103, 364 fortifications, 365–7, 366 Friday mosque, 65, 98, 100, 101, 306, 309–10, 368 Ibn al-Azraq, 362 İl-Ghazi, 13–14, 27, 29, 32, 36, 48n, 71n, 195 Jacobites, 104–5 and Mardin, 43n, 94–6 Marwanids, 47n monasteries, 104–5 Nasr al-Dawla Ahmad, 369, 371 succession problem, 51–3 Suqman, 46n Temürtash, 58–66 Turks, 347, 370 Meisami, Julie, 165 Melkites, 104, 371 Merzifon, 110, 111, 119 Mevlevi order, 111–12 Michael the Syrian Alpı, Najm al-Din, 303 Qara Arslan, 80 Seljuqs, 89 Sulayman, 54 Temürtash, 53, 56, 66, 305 Turcomans, 18 Turks, 343–4 Michael VII Ducas, 89, 91 Michell, George, 306 al-Mihani, 297–8n Minorsky, V., 45n, 51, 102 Mirdasids, 95, 361 ‘Mirrors for Princes’ genre Nizam al-Mulk, 255, 257, 261, 334n, 340 Ravandi, 167, 205–6 Seljuqs, 236 women, 184–6, 194

ind ex |  391 Mongols Alamut, 146, 154–8 Anatolia, 206–7, 209, 333 Artuqids, 218 Assassins, 123, 129 death sentences and cannibalism, 115n Jazira, 77 Mamluks, 108–16 Mayyafariqin, 102, 103 Mu‘in al-Din Sulayman Parwana, 108–20, 260 nomads, 199n, 357 Seljuqs, 197 Morgan, D. O., 155 Mosul Abu Taghlib al-Ghandanfar, 99 Badh, 94, 363 counter-crusade, 40–1n İl-Ghazi, 37 Mawdud, 9–12, 27 Nur al-Din, 311–12 Sayf al-Din Ghazi, 63 al-Suhrawardi, 82 Suqman, 215–16 ‘Uqaylids, 361 Zengi, 60 Zengids, 77–8, 84, 311–12 Mouton, J.-M., 225 Mu’ayyad al-Mulk, 292 Muhadhdhab al-Din ‘Ali al-Daylami, 109, 117 Muhammad, Sultan, 149–55 Alamut, 157–9, 161n Barkyaruq, 146–9 daughter of, 188, 194 Franks, 216 Haysa Baysa, 246 İl-Ghazi, 6–12, 27–8, 35, 37, 40n, 41n nawba, 243 Sadaqa, 40n Sanjar, 31 Seljuqs, 77, 78, 237 Suqman al-Qutbi, 46n Tughtegin, 30–1, 42n Muhammad b. ‘Ali al-‘Azimi, 231n Muhammad b. Dushmanziyar, 148

al-Muhtadi, 98 Mu‘in al-Din Sulayman Parwana, 108–20, 115n, 260 Mu‘izzi, 246 Mumahhid al-Dawla, 94, 100, 363–4 al-Muqaddasi, 103, 104, 273, 361, 364–5 al-Muqtadi, 123, 124–5, 129, 130, 131, 135n al-Muqtafi, 77, 86n, 131, 188, 193, 194 Murad II, 168, 211 Musa, 47n al-Musta‘li, 39n al-Mustanjid, 86n, 305 al-Mustansir, 95, 193, 196, 261 Mustansiriyya madrasa, 296 al-Mustarshid, 35, 168 al-Musta‘sim, 193 al-Mustazhir, 7, 160n, 188–9, 191–2 al-Mu’tadid, 98 al-Mutanabbi, 99, 347 Muzaffar Shihab al-Din Ghazi, 102 Najm al-Din Qummi, Ta’rikh al-vuzara’, 260 Nasibin, 19, 32, 60, 61, 94, 95, 363 al-Nasir, 77, 189 Nasir al-Dawla Mansur, 96, 371–2 Nasir-i Khusraw, 95, 103, 366–7 Nasr al-Dawla Ahmad, 94–6, 100, 363–71, 372, 373n nawba, 240–5, 242, 249 Nicephorus Bryennius, 91 Nishapur, 141–2, 271, 276, 289, 291, 294–5, 297n, 328 Nishapuri, 165, 167, 243, 248, 271 Saljuqnama, 166 Nizam al-Din Nasr, 96, 371 Nizam al-Mulk, 121–5 Alp Arslan, 139–40 Alp Arslan and Malikshah, 326 atabeg, 261–4 death of, 123–6, 132n, 296n, 346 Ibn Jahir, 96 İl-Ghazi, 4–5 Kunduri, 270–81 Malikshah, Sultan, 248, 328 man of letters, 245–6, 255, 258

392  |  the me d ie va l tur k s Nizam al-Mulk (cont.) ‘Mirrors for Princes’ genre, 334n Nasr al-Dawla Ahmad, 372 Nizamiyya madrasas, 288–301, 297n, 331, 356 poetry about, 246–7 Seljuqs, 330, 355 Siyar al-nuluk, 236 Siyasatnama (Book of Government), 124, 184–5, 194, 255, 259, 275, 332–3, 340 viziers, 254–69 women, 194–5 Nizami-yi ‘Arudi, 246 Nizamiyya, 122–3, 147–8, 149 Nizamiyya madrasas, 288–301 in Baghdad, 297–8n comet, 298n professors, 297–8n, 297n nomads, 12–13, 321–5, 333–4, 336, 337, 344, 357, 361, 370 dress, 325 portrayal in Muslim sources, 344–6 women, 345 Nubata family, 101 Nur al-Din Alpı, Najm al-Din, 311–12, 314 Aqsa Mosque, Jerusalem, 353 buildings, 354–5 Crusaders, 77 Damascus, 354–5 jihad, 80–1, 219 Joscelin II, 63 Saladin, 218 al-Suhrawardi, 82 Syria, 340, 350 Turks, 342 Nura Khatun, 190 Odoric of Pordenone, 228 Oghuz tribes, 262, 325 Origen, 272–3 Ostrogorsky, G., 372n ‘Other’, 221, 229, 338, 356 Ottomans gazanama genre, 172 Mayyafariqin, 104

Persian historiographical tradition, 165, 211 Turks, 337 Palestine, 215, 220, 338 Peacock, Andrew, 267n, 340 Persia Great Seljuqs, 77, 78 Marwanids, 361–75 Turcomans, 76–7 Turks, 220–1, 338, 355–6 Persian chronicles, 236–7 historiographical tradition, 164–5, 169–70, 180n names, 208 sources, 176–7, 339–42 Persianisation, of Anatolian cities, 206–10 poets, 193–4, 245–7, 371 Psellus, Michael, 89, 363–4 Pseudo-Galen, Kitab al-Diryaq, 239 Pseudo-Nishapuri, 277 Pullyblank, E. G., 244 qadi, 21–2, 32–3, 101, 304, 348 al-Qa’im, 188, 241, 271, 276, 277, 279 Qal‘at Ja‘bar, 35, 53, 62 Qara Arslan Alpı, Najm al-Din, 312–13 bridges, 307 forbade new churches, 80 marriage alliances, 304 Temürtash, 63, 73–4n Zengi, 73n, 218 Qaraman bridge, 65, 74n, 101, 306–7 Qavurt, 47n, 138 Qilij Arslan I, 14, 27, 55, 100 Qilij Arslan II, 196 Qilij Arslan IV, 109, 111, 117 Qizil Sarigh, 151 Qudama, 103 Qummi, Najm al-Din, 280 Qur’anic quotations, 164–5, 167, 172, 185, 217, 340 Qutb al-Din al-Nishapuri, 297n Qutb al-Din İl-Ghazi, 313 Qutb al-Din Mawdud b. Zengi, 307, 313

ind ex |  393 Qutlugh Balka Sayyida Turkan, 192 Qutlumush, 138, 257, 272 Rashid al-Din, 35, 116n, 123–4, 126, 160n, 248 Jami‘ al-tawarikh, 226, 324, 326, 327, 343 World History, 109 Rásonyi, L., 225–6 Ravandi, 165, 169, 179n, 202–14 Rahat al-Sudur, 166–8, 204–6 Rawandi, 160n, 195, 275 Rayy, 156, 272 Ribat-i Sharaf, Khurasan, 191–2, 192 Richards, D. S., 180n Ridwan death of, 31 İl-Ghazi, 5, 6, 16, 19, 30, 39n Isma‘ilis, 34, 49n marriage alliances, 188 Turks, 220, 338 Tutush, 39n Robert (Frankish leader), 30 Roger of Antioch, 11, 19–20, 21, 23–4, 31–3, 46n, 216–17, 348 Röhricht, R., 2, 64 Romanus IV Diogenes, 89–93, 167, 170, 171–2 Roussel of Bailleul, 90 Rum see Anatolia; Seljuqs of Rum Runciman, S., 2, 24, 42n, 64 Sa‘d al-Mulk, 150 Sadaqa, 7, 9, 40n, 156 Sadr al-Din al-Khujandi, 293 Safi, Omid, 258, 263, 294 Safwat al-Mulk, 349 Sahib Diwan Shams al-Din, 114 Saladin Artuqids, 218 Castle of Sorrows, 316n Crusaders, 77 illness, 313 jihad, 80, 81, 219 Mayyafariqin, 101–2 women, 196 Saljuqa bint Mas‘ud, 190–1

Saljuqa Khatun, 189, 192, 196 Samanids, 164, 234, 245 Sanaullah, M. F., 145 Sanjar Barkyaruq, 147, 156 İl-Ghazi, 7, 31 Iran, 328 Isma‘ilis, 148, 151–2 Khurasan, 6, 78, 146 mausoleum, 329 Muhammad, Sultan, 149, 157–9 nawba, 243–4 poetry about, 247 wife, 192 Sasanians atabegs, 264 castration, 273–5 court, 234 Fürstenspiegel genre, 164 government, 203 hunting, 248 Nizam al-Mulk, 264, 296 rank, 237 Sauvaget, J., 101 Savran, 302–3 Sayf al-Dawla ‘Ali, 98–9, 347 Sayf al-Din, 102 Sayf al-Din Ghazi, 63 Sayyida Khatun, 188–9 Scythians, 228, 353 Seljuqs, 79–80, 321–35 Alp Arslan, 136, 138, 140, 142 Anatolia, 89 Arabic and Persian sources, 176–7 atabegs, 263 buildings, 83 ceremonial robe of honour, 356 chronology of deaths, 134n counter-crusade, 40–1n court, 331–3 deaths of Nizam al-Mulk and Malikshah, 121, 130–1 Diyar Bakr, 96 drunkenness, 332 fighting viziers, 260 genealogy, 134n government, 330

394  |  the me d ie va l tur k s Seljuqs (cont.) historiography, 163–81 Ibn al-‘Adim, 143n Ibn Jahir, 372 İl-Ghazi, 3–6, 12, 35, 36–7, 215–16 Isma‘ilis of Alamut, 145–62 Jazira, 28, 76–7 jihad, 80 map of empire, 322 Mayyafariqin, 100, 104–5 men of letters, 245–7 Muhammad, Sultan, 8 Nizam al-Mulk, 256–8 religion, 81, 330–1 Romanus IV Diogenes, 90–1 in Syria and Jazira, 346 Tiflis, 43n Tughril, 370–1 Tughtegin, 11 Turcomans, 20 women, 332–3 Seljuqs of Rum, 55, 77, 108–20, 189, 214n, 218, 333 Sevim, ‘Ali, 1, 29, 49n, 138 Shaddadids, 361 Shafi’is Kunduri, 371 Mayyafariqin, 95, 100, 103 Nishapur, 289–90, 297n Nizam al-Mulk, 257 Nizamiyya madrasas, 289–95, 297–8n Shahan, 193, 196 Shahdiz, 146, 149–50 Shah-i Arman, 77, 190, 303, 311 Shahnama, 208 al-Sham, 98 Shams al-Dawla Sulayman, 54–6 Shams al-Din, 303–4 al-Shanfara, Lamiyyat al-‘arab (The verses rhyming in lam of the Arabs), 246 Sharaf al-Din Khan Bidlisi, 104 al-Sharif, citadel of, 34 Shaybak Khan, 228 al-Shaybani, 137 Shaybanids, 98 al-Shaybaniyya, citadel of, 312–13 Shaykh Abu Nasr al-Manazi, 100, 371

shihna, 6–7, 9, 10, 14, 18–19, 28, 41n Shi‘ite Islam, 36, 79, 337, 347, 368 Shirgir, Anushtegin, 151, 153, 158 Shuhda al-Katiba, Fakhr al-Nisa’, 193 Sibt b. al-Jawzi Alpı, Najm al-Din, 312 comet, 298n Gervase of Basoches, 223 Kunduri, 279 Manzikert, battle of, 89–90 nawba, 243 Nizamiyya madrasas, 292 Tiflis campaign, 26 Tughtegin, 42n Turcomans, 343 Sinclair, T., 316n Sinope, 111, 112, 117, 119 Sinor, D., 244 Sitti Radawiyya, 316n Sivan, E., 21–2, 32–3, 41–2n, 44n, 80–1 Spuler, F. B., 115n Stevenson, W. B., 42n, 57, 64 al-Subki, 124, 129, 135n, 258, 278, 295 Sufism Anatolia, 208–10, 333 Artuqids and Zengids, 81 Assassins, 124 women, 191 al-Suhrawardi, 82 Sulayman, 16, 51–4, 58–9, 72n, 195, 217, 333 Sulayman b. ‘Abd al-Jabbar, 16, 25 Sulayman b. Chaghri, 272, 276 Sulayman b. Qilij Arslan, 173 Sulayman II, 179n Sunni Islam Alp Arslan, 140 Alpı, Najm al-Din, 306 Assassins, 121, 124, 126, 129 Baghdad, 325 Isma‘ilis, 82 Jazira, 79–80 Muhammad, Sultan, 35–6, 149–50, 153–4 Nasr al-Dawla Ahmad, 368 Nizamiyya madrasas, 290, 294, 296 Qadriyya creed, 86n

ind ex |  395 Seljuqs, 163, 330–1, 334, 337–8, 340, 355–7 Seljuqs of Rum, 167 Turcomans, 350–2 Turks, 221 Suqman, 5–6, 7, 14–15, 39n, 40n, 47n, 215–16 Suqman al-Qutbi, 10, 14, 27, 41n, 43n, 46n, 100, 216 al-Sur, citadel of, 60 Surkhan b. Kaykhusraw al-Daylami, 150 Süssheim, H., 40n Syria Alpı, Najm al-Din, 312 Arabic and Persian sources, 339–42 Artuqids, 218 Balak, 54 Baybars, 110 Crusaders, 77, 78 Franks, 25, 37 hunting, 249 Ibn al-Azraq, 51 Ibn al-Qalanisi, 229 İl-Ghazi, 5, 9–10, 23–4, 30–1, 52, 347–8 Isma‘ilis, 36, 82 jihad, 21, 80 Malikshah, Sultan, 128 Muhammad, Sultan, 12 Mu‘in al-Din Sulayman Parwana, 115n Nur al-Din, 311 Seljuqs, 146, 163–81 Suqman, 6, 215–16 Tughtegin, 349–50 Turkish shamanistic practices, 350–3 Syriac Chronicle, 23, 26, 28–9 al-Tabari, World History, 274–5 Tabas Masinan, 148 Tabriz, 109, 111, 113, 118 Taj al-Mulk, 122–3, 125, 128, 130, 132n, 259, 278 Takiya bint Ghayth, 193–4 Tall Danith, 31, 216 Taylor, John G., 102 Temürtash administration, 64–6

Aleppo, 15–16 Alpı, Najm al-Din, 311 Artuqids, 50, 80–2, 218–19 ceremonial robe of honour, 66, 81–2 Christians, 305 coins, 307–9 Da’ud, 28 Diyar Bakr, 77 Ibn al-Azraq, 13, 302, 314 İl-Ghazi, 36 Is’ird, 73–4n Mahmud, 11, 48n Mayyafariqin, 101, 306 succession problem, 51–4 Sulayman, 55–6 taxation, 65–6 women, 190 Tenreyro, 104, 105 Terken Khatun, 71n, 123, 125–6, 128, 130, 194–5, 332–3 Tiberias, 221–5 Tiflis Aleppo, 16 Dubays, 53 Ibn al-Azraq, 51 İl-Ghazi, 14, 17–18, 23, 25–6, 33 Tughril, 43n, 44n Turcomans, 13, 22 al-Tihami, 95, 371 Toghan Arslan, 14, 26 Tokat, 109–13, 117, 118 Transoxiana, 137–8, 258, 278 al-Tughra’i, Lamiyyat al-‘ajam (The verses rhyming in lam of the non-Arabs), 246 Tughril Chaghri, 241, 325–6 gifts, 332 Kunduri, 270–80, 282n lack of court, 235 Liparit, Prince, 95 Mahmud, 44–5n Marwanids, 370 mosques and madrasas, 81 name, 330, 355 Nasr al-Dawla Ahmad, 96 Qutlumush, 257 Ravandi, 204

396  |  the me d ie va l tur k s Tughril (cont.) shamanism, 352 Tiflis, 17, 26, 43n, 44n women, 188 Tughril II, 195 Tughril III, 328 tughs, 225–8, 227 Tughtegin, 220–33, 349–50 Franks, 181n Gervase of Basoches, 352–3 İl-Ghazi, 11, 23, 30–3, 37, 216–17 Mawdud, 27, 42n Muhammad, Sultan, 12 name, 349 Sulayman, 54 Suqman, 8 Turcomans, 20–1 women, 195 Tughtegin (Atabeg), 230n Tughtegin (brother of Saladin), 226 Turan, Osman, 1–2, 33–4, 41n, 42n Turcoman Qaraqoyunlu confederation, 215 Turcomans Aleppo, 44n, 58 Alpı, Najm al-Din, 303, 312, 314 Anatolia, 109 Artuqids, 215–19 buildings, 83, 305–6 and Christians, 82–3 Da’ud, 59–60 drunkenness, 140 Fatimids, 337 İl-Ghazi, 6–7, 10, 12–16, 23–37, 41n, 45n, 48n, 348 Jazira, 76–7, 85 madrasas, 81 Mayyafariqin, 100–1, 103, 370–1 Mu‘in al-Din Sulayman Parwana, 112–13, 118–19 murder of sailor, 43n nomads, 321–5, 340–1 Oghuz tribes, 325 portrayal in Muslim sources, 344–6 Seljuqs, 78, 146, 174–5, 328–9, 355 Seljuqs of Rum, 333 shamanism, 350–3 Sibt b. al-Jawzi, 343

succession problem, 47n Temürtash, 64 titles, 82 women, 187 Türk Ansiklopedesi, 2 Turkish language, 204–5 names, 355 Turks Arabic and Persian sources, 339–42 cavalry, 324 court, 238–40 drunkenness, 221, 338, 339 Ibn al-Qalanisi, 230–1n Islamic, 202 Islamisation of, 340 Marwanids, 361–75 military barons, 347 Muslim perceptions of, 336–60 names, 342–4 Orkhon inscriptions, 234 Seljuqs, 334 shamanism, 350–3, 351 Tughtegin, 220–33 Tutush Ibn Badi‘, 139 İl-Ghazi, 30 iqta‘, 5, 39n Mayyafariqin, 100 Ridwan, 39n Tughtegin, 226, 349 Turks, 220, 338 ‘ulama’, 65–6, 79, 81, 95, 197, 354, 357 ‘Umar b. Tabarzad, 137 ‘Umar Khayyam, 245 ‘Uqaylids, 94, 95, 361, 363, 369 Usama b. Munqidh Alp Arslan, 140 Artuqids, 219 court, 84 drunkenness of İl-Ghazi, 24, 348 hunting, 248–249 process of scalping, 231n Turkish language, 342 Turks, 221, 338 al-Usuli, 297–8n

ind ex |  397 Vale, M. G. A., 235 Van Berchem, M., 102 van Renterghem, V., 289–90 viziers, 254–69, 266n Von Moltke, H., 104 Vryonis, S., 89, 202 Walter the Chancellor, 225 waqf, 81, 82, 292, 298n, 305, 371 Weil, G., 64 Wilken, F., 64 William of Apulia, Gesta Roberti Wiscardi, 89 William of Tyre, 64, 91, 228 women, 182–201, 200n, 332–333, 345 Ya’qub Arslan, 312 Yaqut, 103 Yusuf Khass Hajib, 189–90 Wisdom of Royal Glory, 182

al-Zahid b. al-Tawil, 306 Zaynab bint al-Shari‘, 193 Zengi Artuqids, 218 cruelty, 73n death of, 63, 77 Ibn al-Athir, 56 Iraq, 78 Jazira, 350 lack of jihad, 81 peripatetic lifestyle, 354 Syria, 340 taxation, 66 Temürtash, 101, 311 Zengids, 77–81, 83, 84, 312, 314 Zubayda Khatun, 123, 125, 126, 129–30 Zubdat al-tawarikh, 137 Zumurrud Khatun, 191