Early Islamic Art, 650-1100: Constructing The Study Of Islamic Art (1) 0860789217, 9780860789215

Table of contents :
Cover
Half Title
Title Page
Copyright Page
Table of Contents
List of Illustrations
Preface
Acknowledgments
Introduction
Part One: Origins and Context
Chapter I: Islamic Art and Byzantium
Chapter II: Islam and Iconoclasm
Chapter III: Upon Reading Al-Azraqi
Chapter IV: The Iconography of Islamic Architecture
Chapter V: Art and Architecture and the Qur'an
Part Two: Secular Culture under the Umayyads
Chapter VI: Sondages à Khirbet El-Minyeh
Chapter VII: Umayyad “Palace” and the ‘Abbasid “Revolution”
Chapter VIII: Notes sur les Cérémonies Umayyades
Chapter IX: The Date and Meaning of Mshatta
Chapter X: La Place de Qusayr ‘Amrah Dans L’art Profane Du HautMoyen Age
Chapter XI: L’Art Omeyyade En Syrie, Source De L’Art Islamique
Chapter XII: Umayyad Palaces Reconsidered
Part Three: Fatimid Egypt and the Muslim West
Chapter XIII: Imperial and Urban Art in Islam: The Subject-Matter of Fatimid Art
Chapter XIV: Fatimid Art, Precursor or Culmination
Chapter XV: Notes Sur Le Mihrab De La Grande Mosquée De Cordoue
Chapter XVI: Two Paradoxes in the Islamic Art of the Spanish Peninsula
Chapter XVII: Qu’est-Ce Que L’Art Fatimide?
Part Four: The Muslim East
Chapter XVIII: Sarvistan: A Note on Sasanian Palaces
Chapter XIX: Notes on the Decorative Composition of a Bowl from Northeastern Iran
Chapter XX: A Tenth-Century Source for Architecture
Index

Citation preview

variorum collected studies series

early islamic art, 650–1100

Oleg Grabar 1970, at Qasr alHayr in central Syria, directing excavations

Oleg Grabar

Early Islamic Art, 650–1100

Constructing the Study of Islamic Art, Volume I

First published 2005 by Ashgate Publishing Published 2016 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN 605 Third Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business Copyright © 2005 Oleg Grabar The author has asserted his moral right under the Copyright, Designs and Patents Act, 1988, to be identified as the author of this work. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing in Publication Data Grabar, Oleg Early Islamic art, 650–1100. – (Constructing the study of Islamic art; V. 1) (Variorum collected studies series) 1. Art, Islamic 2. Architecture, Islamic 3. Islamic antiquities I. Title 709.1’767’09021 Library of Congress Cataloging-in-Publication Data Grabar, Oleg Early Islamic art, 650–1100 : constructing the study of Islamic art / Oleg Grabar. p. cm. – (Variorum collected studies series) Includes bibliographical references and index. ISBN 0–86078–921–7 (v. 1 : alk. paper) 1. Art, Islamic. I. Title. II. Series: Collected studies. N6260.G688 2005 709’.17’6709021–dc22

2004011462

ISBN 13: 978-0-86078-921-5 (hbk)

Typeset by Manton Typesetters, Louth, Lincolnshire, UK

VARIORUM COLLECTED STUDIES SERIES CS809

Contents

List of Illustrations

vii

Preface

xiii

Acknowledgments

xix

Introduction

xxi

Part One: Origins and Context I

Islamic Art and Byzantium

3

II

Islam and Iconoclasm

43

III

Upon Reading al-Azraqi

57

IV

The Iconography of Islamic Architecture

69

V

Art and Architecture and the Qur’an

87

Part Two: Secular Culture under the Umayyads VI

Sondages à Khirbet el-Minyeh

107

VII

Umayyad “Palace” and the ‘Abbasid “Revolution”

131

VIII

Notes sur les Cérémonies Umayyades

141

IX

The Date and Meaning of Mshatta

151

X

La Place de Qusayr ‘Amrah dans l’art profane du Haut Moyen Age

159

v

vi

contents

XI

L’Art Omeyyade en Syrie, Source de l’Art Islamique

173

XII

Umayyad Palaces Reconsidered

187

Part Three: Fatimid Egypt and the Muslim West XIII

Imperial and Urban Art in Islam: The Subject-Matter of Fatimid Art

215

XIV

Fatimid Art, Precursor or Culmination

243

XV

Notes sur le Mihrab de la Grande Mosquée de Cordoue

257

XVI

Two Paradoxes in the Islamic Art of the Spanish Peninsula

267

XVII

Qu’est-ce que l’Art Fatimide?

277

Part Four: The Muslim East XVIII

Sarvistan: A Note on Sasanian Palaces

291

XIX

Notes on the Decorative Composition of a Bowl from Northeastern Iran

299

XX

Index

A Tenth-Century Source for Architecture

311

321

List of Illustrations

Part One: Origins and Context I

Islamic Art and Byzantium 1 2 3 4 5 6 7 8

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Damascus, Great Mosque. Court (after Creswell, Early Muslim Architecture) Damascus, Great Mosque. Plan Damascus, Great Mosque. Minaret Khirbat al-Mafjar. Plan Khirbat al-Minyah. Plan Mshatta. Plan (after Creswell) Qusayr ‘Amrah. Plan (after Creswell) Qasr al-Hayr. Façade, as reconstructed in the Damascus Museum (after Creswell) Qasr al-Hayr. Sculpture on façade Khirbat al-Mafjar. Sculpture from Palace entrance Arab–Byzantine coin Arab–Byzantine coin Arab–Byzantine coin Arab–Byzantine coin Post-reform Umayyad gold coin Damascus, Mausoleum of Baybars. Mosaic fragment Khirbat al-Mafjar. Mosaic fragment Khirbat al-Minyah. Mosaic fragment Qusayr ‘Amrah. Fresco of the Six Kings (after Musil, Kusejr Amra) Qasr al-Hayr. Sculpture of prince Khirbat al-Mafjar. Sculpture of prince Qasr al-Hayr. Sculpture of prince Qusayr ‘Amrah. Fresco of prince Istanbul, Topkapi Saray. Ahmet III, no. 2147, frontispiece with portrait of Dioscorides Vienna, Nationalbibliothek, A.F.10, authors’ portraits Kitab al-Aghani, frontispiece Vienna, Nationalbibliothek, A.F.10, frontispiece vii

8 9 10 12 14 15 16 17 18 19 21 21 21 21 21 22 25 26 27 30 31 32 33 34 35 36 37

viii

list of illustrations

28 29

Ravenna, Museo Nazionale. Ivory diptych Damascus, Hospital of Nur al-Din, façade

IV

The Iconography of Islamic Architecture

1 2 3

4 5 6 7

The Dome of the Rock, Jerusalem Taj Mahal, Agra Tile mosaic mihrab, Isfahan, dated 1354; outermost inscription is Qur’an 9:18–22, including masajid Allah. New York, Metropolitan Museum of Art, Harris Brisbane Dick Fund, 1939. Courtesy The Metropolitan Museum of Art Ottoman imperial mosque: Suleymaniye, Istanbul Hypostyle mosque: Great Mosque of Qairowan (photo: Oleg Grabar) Qutb Minar, Delhi Muqarnas, Hall of the Two Sisters, Alhambra, Granada

38 39

70 71

75 77 78 81 83

Part Two: Secular Culture under the Umayyads VI

Sondages à Khirbet el-Minyeh

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Le site de Khirbet el-Minyeh en contexte Le site de Khirbet el-Minyeh Vue générale du site Plan du palais Plan du secteur fouillé et niveau supérieur Plan des salles ABC – niveaux inférieurs Coupe des salles ABC La voûte de la salle B effondrée sur place Le mur 101 contre le mur ouest du palais La cuve de sarcophage, loc. 228 Le mur 102, salle B, et le passage vers la salle A Mur 102 – salle B Mur 106 – salle B Blocage 106a dans le mur 106 Céramique

IX

The Date and Meaning of Mshatta

1 2 3

Mshatta, façade (after Creswell, Early Muslim Architecture) Mshatta, plan (after Creswell) Ukhaydir, plan (drawing M. al-Asad)

108 108 109 110 111 112 113 114 115 116 118 119 120 121 123

152 153 155

list of illustrations ix

X 1 2

La Place de Qusayr ‘Amrah dans l’art profane du Haut Moyen Age

12 13

Qusayr ‘Amrah, vue extérieure (cliché Claude Vibert) Qusayr ‘Amrah, première salle, vue d’ensemble (angle nordouest) Qusayr ‘Amrah, première salle, mur nord avec l’abside (ensemble) Qusayr ‘Amrah, première salle, mur est (ensemble) et arche centrale Qusayr ‘Amrah, première salle, mur est, détail: scène de battue Qusayr ‘Amrah, scène de battue, détail Qusayr ‘Amrah, scène de battue, détail Qusayr ‘Amrah, scène de battue, détail Qusayr ‘Amrah, scène de battue, détail Qusayr ‘Amrah, première salle, mur sud de la nef ouest, détail Qusayr ‘Amrah, première salle, mur nord de la nef ouest, détail Qusayr ‘Amrah, première salle, mur ouest, détail Qusayr ‘Amrah, deuxième salle, plafond

XI

L’Art Omeyyade en Syrie, Source de l’Art Islamique

3 4 5 6 7 8 9 10 11

1 2

Qusayr ‘Amrah: Vue extérieure (cliché Claude Vibert) Qusayr ‘Amrah: Vue intérieure: façade sud (cliché Claude Vibert) Qusayr ‘Amrah: Salle principale, ensemble du mur nord Qusayr ‘Amrah: Salle principale, détail: travaux des champs ou joutes Qusayr ‘Amrah: Salle principale, détail: femme à la piscine Qusayr ‘Amrah: Salle principale, mur est: moine chrétien (?) Qusayr ‘Amrah: Salle principale, voûte sud: détail des travaux de construction Qusayr ‘Amrah: Salle principale, écoinçon sud: musicien et danseuses

3 4 5 6 7 8

XII 1 2 3 4

160 161 162 163 164 164 165 166 167 168 169 170 171

175 176 177 178 181 182 183 184

Umayyad Palaces Reconsidered So-called Palmyrene sculpture from Qasr al-Hayr West Qusayr ‘Amrah. View toward right wall Qusayr ‘Amrah. Standing nude woman on right wall in main hall Qusayr ‘Amrah. “Lady Niké” on back wall of right vault

195 198 199 200

x

list of illustrations

5

Qusayr ‘Amrah. Standing women on either side of ruling figure Qusayr ‘Amrah. Standing women on either side of ruling figure Qusayr ‘Amrah. Dancing woman on arch in main hall Qusayr ‘Amrah. Dancer accompanied by guitar player on spandrel of central vault Qusayr ‘Amrah. Standing nude woman in central hall Qusayr ‘Amrah. Alighting (?) woman, on left wall of main hall Qusayr ‘Amrah. Personifications on the side of windows Qusayr ‘Amrah. Female figure in upper part of central nave Qusayr ‘Amrah. Meditating woman in small side room Qusayr ‘Amrah. Naked women and children in side room Qusayr ‘Amrah. Naked women and children in side room

6 7 8 9 10 11 12 13 14 15

201 202 203 204 205 206 207 208 209 210 211

Part Three: Fatimid Egypt and the Muslim West XV 1 2

Notes sur le Mihrab de la Grande Mosquée de Cordoue Cordoue, mihrab de la Grande Mosquée, 965 (Foto Mas, Barcelona) Cordoue, détail du mihrab de la Grande Mosquée (Foto Mas, Barcelona)

258 259

Part Four: The Muslim East XVIII 1 2

Sarvistan: A Note on Sasanian Palaces

Sarvistan: plan Sarvistan: elevation

XIX

Notes on the Decorative Composition of a Bowl from Northeastern Iran

1, 2

Bowl, painted and glazed earthenware. Persian, from Nishapur, ninth–tenth century. The Metropolitan Museum of Art, Harris Brisbane Dick Fund, 63.159.1 Deep bowl, painted and glazed earthenware. Persian, from Nishapur, tenth century. Freer Gallery of Art, 57.4. Courtesy of the Smithsonian Institution, Freer Gallery of Art

3

292 293

302

304

list of illustrations xi

4

Bowl, painted and glazed earthenware. Persian, from Nishapur, tenth century. The Metropolitan Museum of Art, Rogers Fund, 56.44

305

Preface

Beyond the usual objectives of prefaces to thank those who helped in the preparation of these books and to identify the technical idiosyncrasies of their appearance, this particular preface is also meant to explain and justify these four independent volumes given the general title of Constructing the Study of Islamic Art, 1954–2004. These volumes include eighty-three articles published during a period of half a century. These articles constitute about two thirds of the contributions I made over the years to periodical literature, encyclopedias and collective books of one sort or another (with some exceptions noted below). Almost all book reviews have been eliminated, as have articles which contain major mistakes or which lead to incorrect conclusions without the redeeming value of useful reasoning or of otherwise unavailable data. Chapters or sections in historical or art-historical surveys or in introductions to Islamic culture have been excluded for the most part. Most of these, like those written for volumes 4 and 5 of the Cambridge History of Iran, for The World of Islam (London, 1976, with many subsequent editions), for the Larousse Histoire de l’Art (Paris, 1985), the Encyclopedia Britannica (Chicago, 1974), or the Grove Dictionary of Art (New York, 1996), are reasonably valid summaries of the state of knowledge at the time of their appearance, sometimes a generation ago. But they are dated by now and make better sense in the context of the volumes in which they appear rather than as contributions to scholarship. And, in any event, nearly all of them are available in most reference libraries. Just as with any retrospective, there is an element of self-centered vanity for any author or artist to present anew his or her achievements. The usefulness of the task lies, primarily, in making accessible items which were often spread in many different and sometimes inaccessible places and, secondarily, in reflecting the evolution of a field and of a person during decades of many changes in the academic as well as political and cultural spheres. Even this large selection reflects only part of the energies and efforts of a life of learning and of teaching. Large numbers of files, photographs and hand-written notes have been preserved in the archives kept under the names of André and Oleg Grabar at the Getty Research Institute in Los Angeles. Some documents were passed on to former students and colleagues or given to a few institutions in places with restricted facilities for learning or to young scholars who could profit from them immediately. In providing xiii

xiv

preface

such gifts, I followed, more modestly, the example of Eric Schroeder (1904– 71), curator of Islamic art at the Fogg Museum of Art, Harvard University, who, when he knew that his days were numbered by a fatal disease, passed on to me and to a few other young colleagues some of his books and notes. What he gave me is now, duly inscribed by him, at the Getty Research Institute or in the possession of younger scholars. And there is something soothing in continuing in this manner to preserve the use of resources for scholarship. The first decision to be made, after selecting the articles to be included, was how to organize them. One way could have been according to the different methodological directions taken by these studies. Such an approach could have been justified by the two directions suggested in a couple of short articles written when I was in my early twenties, which are not included in this selection. One is a precise and detailed presentation of two unusual and until then unpublished bronze coins of the early thirteenth century minted by a minor ruler of the northern Jazirah, the upper Mesopotamian valley now in Turkey; their analysis led to comments on the meaning of the word sultan as a title.1 The other one is the hypothesis that a verse attributed to an Umayyad caliph can explain a very fragmentary painting in the bath of Qusayr ‘Amrah, even though there is no reason to believe that the verse or its author had anything to do with the painting.2 In the first instance, all references are to written or numismatic evidence from the time of the coins involved, in the second one none are (even the verse is only known from a later source), and much of the bibliography deals with arguments around the representation of royal power. The information in the first article has by now been superseded and the second one was incorporated in a later book, The Formation of Islamic Art. Alternately, the articles could have been put in the chronological sequence of their appearance, which would have illustrated the development of an individual’s scholarly thinking and interests and of the ways in which that thinking and these interests were affected by new information and by changing intellectual fashions. But we finally settled on a compromise: two volumes reflecting the history of the Islamic world and of its art, and two others with a thematic focus. There is, first, the early Islamic period, these first centuries which transformed an enormous area into a primarily Muslim one. Then there is the Islamic visual culture which overwhelmed these territories and which is still the dominant one from Senegal to the Philippines. But then, no one dealing with Islamic art can avoid explaining to himself or herself and to 1 2

“On Two Coins of Muzaffar Ghazi,” American Numismatic Society Museum Notes 5 (1953). “The Painting of the Six Kings at Qusayr Amrah,” Ars Orientalis, 1 (1954).

preface

xv

others what it is that characterizes that art in contrast or as a parallel to other traditions, especially, for the medievalist that I was as a student, to Christian art with many of the same sources. The search for verbal formulas to explain visual phenomena or for the ideological bases of the arts is an endless pursuit that often has to respond to new challenges of thought and of political and cultural events. Furthermore, the unique ways of Islamic art as it formed itself and as it developed lead to important issues of the history and criticism of art. In the late 1970s, I began a long and fruitful association with the Aga Khan Foundation and I was introduced to contemporary activities in art and architecture, as well as in the complex operation of cultural policies. Thus a third volume is devoted to general ideas on Islamic art up to our own time and to the theories derived from it or applied to it. And then, partly by accident, I began my acquaintance with the Islamic world and with the Near East in Jerusalem, and I have devoted much time and effort to understanding its monuments and their meaning over the centuries. A whole volume is devoted to that extraordinary city and it includes one totally new contribution, a lengthy response and reaction to the many works on Jerusalem which have appeared during the past fifteen years. This division is an interpretation of fifty years of scholarly activities. But I hope that it will be of better use for other scholars than a purely chronological one would have been or the artificial one of various poles of scholarly procedures. Yet it is not entirely possible to separate the shadows of one’s scholarly life from one’s written accomplishments. For this reason, short introductions to each volume seek to recall the atmosphere surrounding many of the works and especially the people and institutions who over the decades created a context for learning and for growing which is almost impossible to imagine in the academic world of today. For Volume I, I shall introduce the archaeologists and archaeological institutions which helped and inspired me, especially in the 1950s and the 1960s. In Volume II, I shall mention the teaching and research institutions that became my home for nearly forty years and the fascinating evolution that took place in the ways students and colleagues in the United States and elsewhere became involved in the study of Islamic art. For Volume III, I shall sketch out the festival of ideas that accompanied so much of my academic life and some of the nonacademic activities which, since the late 1970s, played an important role in the processes of my learning. Finally, when dealing with Jerusalem, I shall sketch the unique circumstances of working in the Holy City during the 1950s and 1960s. The initial division of the articles was proposed as early as the late 1990s by Professor Cynthia Robinson, who first assisted me in sorting them out. But I had too many other commitments to fulfill at that time and could not manage to concentrate on the project in suitable fashion. Then, in 2001, the Institute for Advanced Study agreed to support the project of a retired professor and the Mellon Foundation provided the funds needed for a full-

xvi

preface

time assistant. Mika Natif, a finishing graduate student at the Institute of Fine Arts of New York University, took on the job. She helped in making the final choice of publications to be included, devised and proposed the arrangement of articles found in these volumes, and undertook the tasks of scanning articles published in many different journals into a single format, of gathering illustrations, and, in general, of keeping the project going. Her sharp and critical mind was essential in transforming what could well have become a disorganized exercise into a reasonably coherent whole for future scholars and critics. Without her energy, dedication and commitment, these books could not have been completed and I owe her a deep debt of gratitude for having stuck with the life and works of an older generation than hers. Additional help was gracefully and intelligently provided by Elizabeth Teague, the copy editor, and I am most grateful to her. Thanks are also due to two institutions. One is the Aga Khan Trust for Culture, which contributed to the publication of these books through ArchNet, a branch of the Aga Khan Program in Architecture at Harvard University and the Massachusetts Institute of Technology. I was involved in the early creation of the program and am grateful to Mr Shiraz Allibhai, manager of the program, and to Dr Luis Monreal, the head of the Trust in Geneva for having continued to support my work so many years later. The second institution is the Institute for Advanced Study from whose School of Historical Studies I retired in 1998. Two successive directors, Dr Philip Griffith and Dr Peter Goddard, supported all aspects of the work involved in preparing these volumes and in making available to Mika Natif and to me the technical facilities of the Institute and the expertise of its staff. A special word of thanks is due to Rachel Gray, Associate Director of the Institute, through whom all needs and requests were channeled. A last expression of gratitude goes to John Smedley from Ashgate Publishing, who, I suspect, did not quite know what he was getting himself into when he agreed to consider the publication of the eighty-odd articles found in these volumes. His gracious help and patience was essential to the completion of the work. The following institutions gave permission to reproduce articles and pictures published under their copyright: Pennsylvania State University Press, Dumbarton Oaks, E. J. Brill, Freer Gallery of Art, Smithsonian Institution, New York University Press; Ukrainian Research Institute, Harvard University, Metropolitan Museum of Art, State University of New York Press, Israel Exploration Journal. A number of editorial decisions were made to ensure consistency across all four volumes, to simplify the task of publishing them, and to facilitate the use of the books. Diacritical marks and macrons were given up altogether. The hamza is shown as ’ and the ‘ayn as ‘. The date and place of the original publication of each article are indicated with an asterisk on the first page of each article. All notes are put at the bottom of pages. References to the original pagination are given in square brackets. Not all original illustrations

preface

xvii

have been included. Some prints or negatives could no longer be located and scanning or photographing anew a mediocre print seemed senseless. At times substitutions were found and in a few instances original illustrations were simply omitted. Typos were corrected whenever we noticed them and minor emendations were made to the original texts to ensure clarity of expression. Bibliographical notes were not brought up to date, except when works announced in the notes were actually published.

Acknowledgments

The chapters in this volume were first published as follows: I

II III IV

V

VI VII VIII IX

X XI

XII XIII

“Islamic Art and Byzantium,” Dumbarton Oaks Papers, 18 (Dumbarton Oaks, Trustees for Harvard University: Washington DC, 1964), pp. 69–88. “Islam and Iconoclasm,” Iconoclasm, A. Bryer and J. Herrin, eds (Birmingham, 1977), pp. 45–52. “Upon Reading al-Azraqi,” Muqarnas, Vol. 3 (E. J. Brill: Leiden, 1986), pp. 1–7. “The Iconography of Islamic Architecture,” in Priscilla P. Soucek, ed., Content and Context of Visual Arts in the Islamic World (Pennsylvania State University Press, University Park and London, 1988), pp. 51–65. “Art and Architecture in the Qur’an,” in Jane D. McAuliffe, ed., Encyclopaedia of the Qur’an, Vol. I (E. J. Brill: Leiden, 2001), pp. 161–75. “Sondages à Khirbat el-Minyeh” (with J. Perrot, B. Ravani and M. Rosen), The Israel Exploration Journal, 10 (1960), pp. 226–43. “Umayyad ‘Palace’ and the ‘Abbasid ‘Revolution,’” Studia Islamica, 18 (1963), pp. 5–18. “Notes sur les Cérémonies Umayyades,” in Myriam Rosen-Aylon, ed., Studies in Memory of Gaston Wiet (Jerusalem, 1977), pp. 51–60. “The Date and Meaning of Mshatta,” Dumbarton Oaks Papers, 41 (Dumbarton Oaks, Trustees for Harvard University: Washington DC, 1987), pp. 243–7. “La Place de Qusayr ‘Amrah dans l’art profane du Haut Moyen Age,” Cahiers Archéologiques, 36 (1988), pp. 75–84. “L’Art Omeyyade en Syrie, Source de l’Art Islamique,” in Actes du Colloque international Lyon – Maison de l’Orient Méditerranéen, ed. P. Canivet and J.-P. Rey-Coquais (Paris, 1992), pp. 188–93. “Umayyad Palaces Reconsidered,” Ars Orientalis, 23 (1993), pp. 93– 108. “Imperial and Urban Art in Islam: The Subject-Matter of Fatimid Art,” Colloque International sur l’Histoire du Caire (27 mars–5 avril 1969), Ministry of Culture of the Arab Republic of Egypt, General Egyptian Organization, pp. 173–90. xix

xx

acknowledgments

XIV XV

XVI

XVII XVIII XIX

XX

“Fatimid Art, Precursor or Culmination,” Isma‘ili Contributions to Islamic Culture, S. H. Nasr, ed. (Tehran, 1977), pp. 209–24. “Notes sur le mihrab de la Grande Mosquée de Cordoue,” A. Papadopoulo, ed., Le Mihrab dans l’Architecture et la Religion Musulmanes (E. J. Brill: Leiden, 1988), pp. 115–18. “Two Paradoxes in the Islamic Art of the Spanish Peninsula,” in The Legacy of Muslim Spain, ed. S. K. Jayyusi (E. J. Brill: Leiden, 1992), pp. 583–91. “Qu’est-ce que l’Art Fatimide?” M. Barrucand, ed., L’Egypte Fatimide, Son Art et son Histoire (Paris, 1999), pp. 11–18. “Sarvistan: A Note on Sasanian Palaces,” Forschungen zur Kunst Asiens (Istanbul, 1970), pp. 1–8. “Notes on the Decorative Composition of a Bowl from Northeastern Iran,” in Islamic Art in the Metropolitan Museum of Art, ed. R. Ettinghausen (Metropolitan Museum of Art: New York, 1972), pp. 91–8. Reprinted with changes by permission. “A Tenth-Century Source for Architecture” (with Renata Holod), Harvard Ukrainian Studies, Vol. III/IV (1979–80), pp. 310–19. Copyright © 1980, by the President and Fellows of Harvard College. Reprinted by permission.

We would like to thank all individuals, publishers and institutions for their permission to reproduce articles and illustrations published under their copyright. Every effort has been made to trace the copyright holders, but if any have been inadvertently overlooked the publisher will be pleased to make the necessary arrangement at the first opportunity.

Introduction

These first few pages are meant to set the academic and, to a degree, personal stage for twenty articles written between 1960 and 2001 and concentrating on the first four and a half centuries of Islamic history, roughly until 1100. Two very different impulses from my days as an undergraduate student at Harvard University, the University of Paris, and the École des Langues Orientales, and then as a graduate student at Princeton University may help to explain my involvement with early Islamic art, with archaeology and with the patronage of the Umayyad and Fatimid dynasties. One impulse was my formation as a medievalist dealing first with Western Europe and Byzantium before 1000, and then with classical Arabic and Islamic history. The second impulse was my year of study, in 1953–54, as the Annual Fellow at the American School of Oriental Research (now the Albright Institute) in Jerusalem. Original training as a medievalist provided me with a set of academic tools and intellectual role models which simply did not exist at that time for the history of Islamic art and culture. André Grabar, Marc Bloch, Henri Pirenne, Emile Mâle, Kurt Weitzmann, Hugo Buchthal, E. Baldwin Smith, Henri Stern, Ernst Kantarowicz, Meyer Schapiro, A. M. Friend Jr, J. Strzygowski, R. Krautheimer are only a few randomly selected names of scholars (all but three of whom I had known personally) whose contributions to the highest levels of scholarship are countless, but who also knew how to provide or inspire theoretical schemes for the understanding of the arts and who exuded, as individuals or in their writing, a mixture of charisma and authority made to attract budding enthusiasts for knowledge. Many of these medievalists of the Christian world showed much interest in Islamic art and culture and even made occasional contributions to its study. Some, like Kurt Weitzmann, felt very strongly that their approaches had a universal value, at least as a way to handle illustrated manuscripts first and then all images. Whether or not one agreed with their position, the impression on a student was clear: Islamic art was a medieval art with sources firmly entrenched in Classical Antiquity and its study needed the more developed and more elaborate methods used for Christian art. There was a Byzantine version of Christianity with carefully worked out rules, regulations and procedures, and with an established doctrine for the interpretation of images or the xxi

xxii

introduction

construction of religious buildings. And there was a Western European or Latin version, much more varied in its multifarious appearances and much more complicated in its development. Both versions had formulated paradigms which affected the study of all arts at all times. In those days, no one but collectors of Persian or Mughal painting or inveterate travelers to India or Iran considered the possibility of dealing with Islamic art after 1500 and it was natural enough that a young student would be attracted to the early Islamic period, the time when a new culture emerged out of its Late Antique roots. In a sense, classical Orientalism dominated the field. Its essentialist emphasis on a single Islamic culture with occasional aberrations from the norm sought explanations for whatever happened later in the formative decades of the seventh and eighth centuries. Scholars were fascinated by the ways in which a new faith was established, followed by a new state and then a new cultural system. The texts that had been edited, like Tabari, Baladhuri, or the early geographers, dealt mostly with the first centuries of Islam and Leone Caetani’s monumental Annali dell’Islam in eleven volumes and nearly 6000 pages only covered the first thirty-seven years of the hijrah. On a more popular level, the most easily available general book was Philip Hitti’s History of the Arabs. More than half of it deals with the first 200 years of Islamic history and the rest of the book is organized around “petty” dynasties in the west and in the east. In short, there was a methodological model, the history of Christian art, for the history of Islamic art and the historical or cultural ground had been prepared for the first two centuries of its existence. It was a lucky accident that, in these days before major fellowship programs for study in the Near East, I was awarded the annual student fellowship of the American Schools of Oriental Research (ASOR). That venerable institution with countless contributions to the study, mostly archaeological, of the Bible and of the Ancient Near East, sought, I believe, to show that its concerns were not limited to biblical times or to the New Testament. But there was another reason as well for my being chosen. The year was 1953, just five years after the division of Palestine into a new state of Israel and an Arab area taken over by what was then the Hashemite Kingdom of Transjordan. Some of the political and administrative complexities of the times dealt with the Palestine Archaeological Museum (PAM, now the Rockefeller Museum) in Jerusalem, then a privately endowed institution run by a consortium of representatives of various learned societies in Europe and the United States. One of these representatives was the president of the ASOR, at that time Karl Kraeling, an archaeologist and historian of early Christian times and of Judaism in Roman times. The museum was the repository of nearly all the finds from the fabulous site of Khirbat al-Mafjar, an Umayyad palatial ensemble with sculptures, paintings and mosaics that had become known through a number of articles published in the Quarterly of the Department of Antiquities of Palestine, but whose revolutionary character was only

introduction

xxiii

dimly perceived at the time. R. W. Hamilton, the last Director of Antiquities of Palestine, had embarked on a major publication of the site, eventually published in 1959 by Oxford University Press as Khirbat al-Mafjar: An Arabian Mansion in the Jordan Valley. The subtitle of the book reflected the Orientalist mood of the time and guaranteed that no one outside the field would notice its existence and importance. Kraeling, a wonderful exemplar of the political scholar–administrator, was anxious that the publication come out as rapidly as possible because of the investment into the site made by the PAM and also to show the importance of the archaeological work carried out in the now defunct Palestine. He proposed the services of a graduate student for whatever purposes Hamilton needed. This is how, after a delightful visit with Hamilton and his wife at their house in or near Oxford, I came to catalog and eventually published the paintings of Khirbat al-Mafjar. I was offered the opportunity to handle the mosaics as well, but turned this task down for fear that I could not complete my Ph.D. thesis in a reasonable time (and thus find a job), if I had to learn about a technique which was not then very familiar to me. Conscientious scholar that he was, Hamilton completed the work on the mosaics in a competent but perfunctory manner and I have at times regretted a decision made for practical rather than intellectual reasons. In any event, my entry into the world of archaeology took place in the store-rooms of the Palestine Archaeological Museum, where I acquired a life-long dislike for archaeological ceramics. It was combined, during the year of my fellowship, with the usual visits to Umayyad sites in what was then Transjordan, and with study trips to Syria, Lebanon, Egypt and Anatolia. In Damascus, I could see the masses of paintings and sculptures from Qasr al-Hayr West, which are still only partly available and whose parallels with and differences from Khirbat al-Mafjar are so striking. The archaeology and understanding of these early monuments of Islamic art was then dominated by two very remarkable individuals who were as different from each other as night and day. In Cairo, there was K. A. C. (“Archie”) Creswell (1879–1974), who was still called Captain because of his World War I activities, a small and energetic man always impeccably dressed, with an imperial feel for command on his face. He was passionate about early Islamic architecture and the first two of his monumental volumes on the subject had already been published, at least in their first version. He was a man of systematic precision in descriptions and details, an excellent draftsman and photographer, fanatically committed to a linear chronological progression of construction techniques and of shapes in design, with a minimum of concern for the cultural context of buildings, opinionated with prejudices on nearly everything. He could be very generous, as when entertaining a student who could be his grandson, but also dismissive in his criticisms. And he easily became angry and red in the face. His wonderful library (now at the American University in Cairo) was beautifully kept in his

xxiv

introduction

apartment near the Abdin palace, in a neighborhood which was no longer fashionable when I met him in 1953, but which he terrorized by his knowledge of municipal regulations and by his control over the possession of radios. The other figure was the much younger Jean Sauvaget (1901–50), who had died recently at the early age of forty-nine. Based in Paris during the last fifteen years of his life, he had spent a long time in Damascus at the École Française de Damas. The author of major books on many subjects, his most remarkable contribution consisted in masses of articles dealing with buildings, history, inscriptions, reviews of scholarly works and so on. They burst with energy, with the vitality of a mind constantly alert and fascinated by everything it encountered. I only met Sauvaget once when I was still a teenager and still recall the excited flamboyance with which he described the Arabic language as a “geometry.” He died as I was preparing to study with him in Paris rather than return to the United States where I had just spent two years. He was devoted to his students, who truly adored him and have remained amazingly faithful to his memory. Although he never, I believe, indulged in actual excavations, he was an archaeologist in the broadest and most fruitful sense of the word, seeking to explain everything that remained from the past and to relate these remains to history, even to contemporary practices. The connector, to him, between the written sources with which one writes history and the world of buildings or of things was epigraphy, the reading of inscriptions. With these tools – observation of the land, reading of appropriate written sources, epigraphy – Sauvaget did give new directions to a type of investigation which had been initiated by the great masters of the previous generation, Max van Berchem and Ernst Herzfeld (whom he frequently criticized). He did so with an astounding rhetorical panache, but also with an awareness of the land and the people surrounding monuments, thus forecasting better than his predecessors what will later be defined as “total history,”1 exemplified most successfully in an approach which welded together archaeology, geography and the knowledge of texts. Sauvaget and Creswell did not appreciate each other, although I do not know whether they ever met. Sauvaget was ironic about Creswell and Creswell had warned me to “beware of Sauvaget.” Both of them were, however, the beneficiaries of the last decades of European imperialism. And, even though by the time I arrived in the Near East, nearly all areas were independent countries, the old imperial system remained active through the staffing of archaeological services and the maintenance of national research institutes. These were French, American and British in Jerusalem, primarily French in Damascus and Beirut, British in Iraq, British and French in Iran, French in Afghanistan. Everybody was present in Egypt and Turkey had French and 1

The term was, I believe, first used by Maxime Rodinson in a review of George Tchalenko’s book Villages Antiques de la Syrie du Nord (Paris, 1953); see Maxime Rodinson, “De l’archéologie à la Sociologie Historique,” Syria, 38 (1961), pp. 170–200.

introduction xxv

British institutes in Istanbul and Ankara respectively. The aftermath of World War II was that there were no German institutes any place at that time, a situation soon to be remedied. These establishments, schools or institutes, were havens for students and, in fact, for all travelers. Cars could be made available for expeditions to remote areas or visits to major monuments. Introductions were provided for officials and ways explained to obtain permits of all kinds. If lucky, one could join for a week or a month an excavation somewhere, and a whole mythology had developed about the style of individual “digs.” In general, a certain camaraderie grew between young scholars with shared interests, some of whom became major figures in their respective fields, while others faded away altogether. With slow mail, few airplanes, no television, expensive and unreliable telephones, radios that still needed electric plugs in walls, these institutions provided bodily and spiritual sustenance, a rest from the exciting novelties of exploration, the feeling of an international, highly western, community in foreign lands. Those of us who had traveled in the arid lands of South Jordan, whose bodies ached from being tossed about over the large ruts of bare tracks, who had to push cars from muddy holes, who lived on hard-boiled eggs and drank theoretically boiled tea at Beduin camp sites, who interpreted with enthusiasm any crummy wall somewhere in the steppe as possessed with the spirit of seventh-century Arabs or eleventh-century Turks, did not feel jealous of our colleagues sleeping in their beds and peacefully looking at manuscripts in the British Library or in the Vatican. We felt, on the contrary, that we brought to the story of visual remains an essential component, the component of the spaces in which events took place, the man-made spaces of buildings, cities or villages, but also the natural spaces of rivers, seas, mountains, rocks and sand. Our thought was that the awareness and the experience of these spaces defined the people who transformed the area by creating in it a new culture. The archaeological record, we used to argue, that of extensive surveys or of actual excavations, was the core documentation for history; written sources were but the veneer of a time’s memory, the self-conscious image it left of itself rather than the objective, even if randomly preserved, remains of a material culture in its natural setting. Such were the thoughts and arguments that we developed and polished during long trips over bad roads. A particularly memorable one for me took place in 1956. Led by George Forsyth, then Chairman of the Art History department at the University of Michigan and driven by an extraordinary Lebanese driver who was a wrestler in his spare time, a marvelous new International Harvester four-wheel-drive station wagon took a group of us – George Forsyth, the biblical scholar George Mendenhall, Father Jean Fiey, a French Dominican monk from Mosul with a unique knowledge of Near Eastern Christianity over the ages, George Tchalenko, the Franco-Russian explorer of the so-called “dead cities” of northern Syria, the marvelous Swiss-American companion and photographer Fred Anderegg who was

xxvi

introduction

eventually to become the photographer of my excavations at Qasr al-Hayr East and for the recording of paintings and mosaics in Jerusalem, Bethlehem and Qusayr ‘Amrah, and finally myself as the junior member of the team (I was only twenty-seven) – from Beirut to Palmyra, Raqqah, Dura-Europos, Baghdad, Mosul, Nisibin, Diyarbakr, Midyat, Urfa, Harran, Aleppo, and back to Beirut during a period of five to six weeks. Even now, almost fifty years later, the memories of the landscapes we crossed like the deep wadis just south of the Syro-Iraqi frontier, the discovery of so many different ways of life such as the marsh Arabs of southern Iraq with their huts covered with vaulted reeds or the sugar-cone dwellings of north-central Syria, the incongruous encounter with the spirit of Marilyn Monroe among the devilworshipers or Yazidis of the Jabal Sinjar in northwestern Iraq, the sad remains of Christianity in the Tur Abdin, negotiations needed occasionally to cross frontiers or to please policemen, all these memories are still fresh in my mind as the only member of the group to be still alive in 2005. It is, of course, easy and fun to recall anecdotes, especially when there is no one to contradict them. What really matters is that this trip, like many others, provided a visual and physical contact with the spaces in which history took place and with the people who now occupy these spaces. One of the most powerful lessons of field archaeology in a land that is not one’s own is the establishment of a contact with other people than yours and other social classes than the ones with which one usually mingled. It is an enriching and a humbling experience, but it is also a difficult one to transfer into scholarship, especially art-historical. It is rare indeed that some extraordinary object or some revolutionary remains are found. What does appear and can be reconstructed is the life of the common man. Even in excavated palaces, kings are present no doubt, but not as much as the artisans who built the palaces, the designers and artists who decorated them, the servants who worked in them, or the visitors who came from neighboring places or from afar. It is the archaeological base that is essential for the historian of art, as it sets up the mass of things against which he can measure the value of his masterpieces. In the early 1950s and 1960s and until the completion, in 1972, of the excavation of Qasr al-Hayr Sharqi in Syria, my own experience was mostly in Egypt and the Near East, with the exception of a month-long investigation of west-central Iran. For the most part, Andalusia, the Maghrib and the Mashriq from Iraq eastward were areas in which others worked, like R. McC. Adams, whose Land Beyond Baghdad is a model of an archaeological overview of a region, and especially the Russian expeditions in Turkmenistan, Khorezm and Tajikistan, which truly revolutionized our knowledge of Central Asia before Islam and helped in suggesting explanations for the ways of its Islamization. My experiences in the Levant and a few exploratory trips to Iran, Afghanistan and then Soviet Central Asia led me, just a few years before the Iranian revolution of 1979, to draw up a grandiose and absurdly

introduction xxvii

optimistic plan for the full exploration of Khorasan, the large province of northeastern Iran. It would have involved collective work carried out in Iran, Afghanistan, Turkmenistan, Uzbekistan and Tajikistan. The idea was great on an intellectual level as well as on a human one, as it would have involved scholars from many countries trained in many different ways and included the investigation of literary sources as well as archaeological ones. In a manner befitting the style of my generation, the project would have followed a brilliant survey sketched by Herzfeld in Der Islam XI of 1921.2 I lost or maybe purposefully destroyed the many pages which described this project and it is a sad commentary on the decay of the world that this sort of dream is even more impossible now, thirty years later, than it was then. Retrospectively, it is easy enough to be critical of the surveys we used to practice then without the fancy photographic equipment available now (my brief encounter with satellite photography and the Office of Naval Research is part of the more hilarious episodes of that time) as well as of the largescale excavations of the 1950s, 1960s, and early 1970s. A season lasted more than two months then; it included more than 150 workers and moved enormous quantities of dirt. But we did not use the new technologies being developed at that time with their cohort of specialized staff members. These novelties made the excavation of large sites prohibitive and the exploration of large urban sites so necessary for the history of Islamic culture, especially during the first centuries of its existence, has become unlikely unless new techniques of investigation are developed. The more serious defect of archaeological work concerned with the Islamic world is that most excavations and surveys are never fully published, if at all. The reason is partly social and economic in the sense that few archaeologists are provided with time for research or teaching jobs. Therefore, they have to run from one “dig” to the next one without time to breathe. But another and perhaps more important reason is that no one has figured out how to translate the mass of very concrete information provided by surveys and excavations into conclusions that can be used by others. In other areas, ancient Palestinian, classical or ancient Egyptian archaeology, for instance, there is a critical mass of intermediary scholarship on architecture, society or ceramics which allows one to fit and explain whatever is found. This is missing in Islamic archaeology and probably will never be developed, unless some stunning transformation occurs in the informatization of archaeological knowledge. The problem becomes intriguing when one turns to what may be called an archaeology of written materials, of words. This is the case with the Geniza fragments from Cairo, thousands of documents illustrating and depicting a Jewish life in and around the Mediterranean which can easily be extended to Muslims and 2

Herzfeld’s subtitle “Denkmalersgeographische Studien zur Kulturgeschichte des Islams in Iran (Studies in the monumental geography for the cultural history of Islamic Iran)” is an ambitious blueprint for one man to accomplish multidisciplinary research.

xxviii

introduction

Christians. But an archaeology of the written word also exists around some early texts, like Azraqi’s description of Mecca or geographers and travelers like Muqaddasi and Nasir-i Khosro. Several of the articles in this volume are not illustrated, precisely because they introduce conclusions or hypotheses from written sources alone. Here again, as with dirt archaeology, it is difficult to relate the information provided by these fragments to the high levels of art, but it is possible to reconstruct the human context of that art. This is particularly true of the Fatimid period (909–1173), for which literary and material documents provide a unique archaeological basis for research into cultural history. And it is this cultural history which became the ultimate aim of investigations initiated by collections of written documents, architectural monuments or ruins, and scores of artefacts. Such investigations are never ended, as each generation or each individual scholar brings something new to what has been gathered before. But research in these areas is an unending trek toward the elusive aim of knowing what happened, who created what, and why. The definitive answer will never be found, but, during the search, the technical and intellectual powers of scholars are sharpened to even greater competence. One last remark. As I complete this introduction, I realize the contrast between the excitement of recalling excavations and archaeological expeditions on the one hand and the rather ponderous scholarship so often inspired by them. Perhaps like many other things in life, the essential experience and at times joys of a broadly conceived archaeology cannot be fully reflected in academic writing.

Part One Origins and Context

Chapter I Islamic Art and Byzantium*

According to Eutychius, a Melkite priest from Alexandria, who died in ad940, the following event is said to have taken place in Qinnasrin, in northern Syria, sometime in the latter part of the thirties of the seventh century, during the Muslim conquest of Syria. An Arab force, under the celebrated general Abu ‘Ubaydah, had signed a truce of one year with the Christians of Qinnasrin in order to allow those Christians who so desired to leave Syria and to follow Heraclius into Anatolia. A line was established between Christian and Muslim possessions and the line was marked by a column (‘amud) beyond which the Muslims were not to go. With the agreement of Abu ‘Ubaydah, the Christians painted on this column a portrait of Heraclius seated in majesty (jalis fi mulkihi). One day, however, an Arab Muslim rider, who had been practicing horsemanship, accidentally defaced the representation of the Byzantine Emperor by planting the point of his spear in its eye. The head of the local Christian community immediately accused the Muslims of having broken the truce. Abu ‘Ubaydah agreed that a wrong had been done and asked what reparations could be offered. The Christian answered: “We will not be satisfied until the eyes of your king are put out.” Abu ‘Ubaydah suggested having an image of himself so mutilated, but to no avail, since the Christians insisted that the likeness should be of the Muslims’ highest authority (malikukum al-akbar). Finally Abu ‘Ubaydah agreed, and the Christians then made an image of ‘Umar, the caliph of the time, whose eye was then duly put out by one of their riders. The Christian patricius concluded the whole affair by saying to the Muslim general: “Indeed, you have treated us equitably.”1 Like most good stories, this one is probably apocryphal, especially in the wealth of its details, since other accounts of the conquest of Qinnasrin do not mention events which could have made it possible and since it comes from the ecclesiastical milieu of Christian Arabs within the Muslim Empire, among whom a whole body of stories developed tending to minimize the * First published in Dumbarton Oaks Papers, 18 (1964), pp. 69–88. 1 Eutychius, Annales, ed. by L. Cheikho and others, in Corpus Script. Christ. Orient., Scriptores Arabici, 3rd Ser., 7 (Beirut, 1909), p. 19.

3

4

early islamic art, 650–1100

tragedy of the Muslim conquest. Yet the story certainly reflects the spirit of the time and may serve to define our subject by focusing our attention on two central points: the period with which we are primarily going to deal and the type of problem posed by that period. First, then, it may serve to define the chronological framework of this paper. The event it describes is supposed to have taken place at the very beginning of the Muslim conquest; it identifies a crucial moment in the political and cultural contacts between Byzantium and the Arabs, when the buffer world of pre-Islamic Arabs, who basked at a distance in the glow of high Byzantine culture, was about to become the Islamic Empire, the strongest power of the Near East and the Mediterranean since the days of ancient Rome. [70] Nothing will be said of the period which preceded the Muslim conquest. The monuments which are definitely Ghassanid or Lakhmid are few and do not seem to have developed original styles, techniques, or functions, although further research may some day modify this picture; nor were the artistic contacts which the Arabs of Arabia had with Byzantium through trade and through the Christian Church more than secondhand. In greater part, they were fleeting impressions of mediocre objects. Occasional texts do refer to a more profound impact of Christian art, but their usefulness for archaeological purposes is often questionable.2 Altogether, the relations between Arabs and Byzantium until the formation of the Muslim Empire were not relations of cultural equality. After the seventh century the two empires were to become powerful bastions of two independent cultures confronting each other. But if the peculiar contact at Qinnasrin may serve to indicate the upper limit of a study of Byzantine–Arab contacts in the arts, can one also define a lower limit that is equally valid historically? This problem is more complex and requires some elaboration. A crucial phase during which the art of the Arabs and that of Byzantium were closely related to each other was the era of the creation of the first Islamic classicism, that is, the first syntheses between the Near East and the Mediterranean, on the one hand, and, on the other, the new political, social and religious entities elaborated by the Prophet and his immediate successors. This era is usually associated with the Umayyad and early ‘Abbasid periods (661 to about 800), and it has long been recognized that in monuments of early Islamic art as diverse as the Dome of the Rock, the desert baths of Qusayr ‘Amrah, or the wooden beams of the Aqsa Mosque, considerable direct or indirect Byzantine influences can be detected. In this early period, especially in Syria and Palestine, Byzantium played the part of one of the many parents who brought a new Islamic art to life. 2

There has not yet been any attempt to put together systematically the information which exists in pre-Islamic poetry concerning Arab knowledge of Christian art, but such knowledge existed (cf. R. Ettinghausen in N. Faris, ed., The Arab Heritage [Princeton, 1946], p. 252; C. J. Lyall, The Mufaddaliyat [Oxford, 1918], p. 92 ff.).

islamic art and byzantium

5

In later centuries a different and more complex series of relationships may be defined. In Cordoba, the Muslim Arab caliphs apparently repeated an early Islamic practice and, in the tenth century, called on Byzantine artists to decorate parts of their great mosque.3 But, in a more general sense, a constant stream of influences flowed in both directions. The Byzantines acquired a taste for Islamic objects and an Orientalized aspect was given both to the court of the Constantinopolitan emperors and to many a church treasure.4 At the same time, the Arab world continued to seek or to feel the impact of the art of Byzantium. Thus the Fatimid caliphs in Cairo, whose ceremonial was closely related to that of the Byzantine emperors, borrowed or imitated Byzantine [71] artistic techniques, such as cloisonné enamel.5 Even later, in the twelfth and thirteenth centuries, when a new art of book illustration developed among the bourgeoisie of the Fertile Crescent and Mesopotamia, models directly or indirectly derived from one of Byzantium’s most developed media of artistic expression were commonly used. The dependence of Arabic Dioscorides manuscripts or of the Automata of alJazari on Byzantine models has often been shown, especially by Professor Weitzmann,6 and it has often been pointed out that many stylistic devices commonly used in the illustration of so purely Arabic a book as the Maqamat of Hariri were derived from Byzantine or, at least Christian, models.7 After the thirteenth century and, at the high level of the princes, even much earlier the rule of Turks or Kurds and the taste of Iran introduced a component into Islamic culture and art which is no longer Arab and is, therefore, outside our specific concern, even though it is evident that in the monuments of Seljuq or Ottoman Anatolia or in the Persian miniatures of the fourteenth century there are numerous features clearly related to Byzantium. These few examples are sufficient to show that artistic contacts existed between the world of Byzantium and the world of Arab Islam, and that these contacts continued over many centuries, or at least throughout the 3

4 5

6 7

Several versions of this event exist, and there is a need for a thorough study of the various texts referring to it. The main account is in Ibn Idhari, Bayan, ed. by G. S. Colin et E. Lévi-Provençal (Leiden, 1951), 2, pp. 237–8. It might be added that columns are supposed to have been brought from Byzantium for the palace of Madinah al-Zahra (ibid., pp. 231–2), but this statement is not very credible. A. Grabar, “Le succès des arts orientaux à la cour byzantine sous les Macédoniens,” Münchener Jahrbuch der bildenden Kunst, 2 (1951). M. Canard, “Le cérémonial fatimite et le cérémonial byzantin,” Byzantion, 21 (1951). On the question of cloisonné enamel, see P. Kahle, “Die Schätze der Fatimiden,” Zeitschrift d. deutschen morgenlandischen Gesellschaft, N.F., 14 (1935), p. 345 and passim. Concerning specific works of art, it is difficult, in the present state of research, to do more than hypothesize that much in the development of Fatimid imagery reflected direct or indirect Byzantine influences. K. Weitzmann, “The Greek Sources of Islamic Scientific Illustrations,” Archaeologica Orientalia in Memoriam Ernst Herzfeld (Locust Valley, 1952). H. Buchthal, “‘Hellenistic’ Miniatures in Early Islamic Manuscripts,” Ars Islamica, 7 (1940).

6

early islamic art, 650–1100

major phases of artistic creativity in the Arab world. After the middle of the thirteenth century, in the Mamluk art of Egypt or in the late medieval art of Morocco or Spain, Byzantine elements, even though not completely absent, are quite rare and contribute little to the definition of these artistic traditions. Similarly, in Byzantine art, after the conquest of Anatolia by the Turks, Oriental themes, when they exist, are no longer those of the Arab world. Having thus defined the period within which Byzantine–Arab relations were meaningful, one could draw up a roster of such mutual influences, study their frequency at certain times, explain this frequency, and discuss the stylistic or thematic changes and modifications which one or the other of the two cultures introduced into the elements they borrowed or used, after that day in the fourth decade of the seventh century when a first contact was established between the new Muslim state and Byzantine art. The interest of this type of investigation is self-evident and coincides with well-established practices in the history of art. This is not, however, the aspect of the subject of Islamic art and Byzantium which I propose to examine. The story of Abu ‘Ubaydah at Qinnasrin has more far-reaching implications than merely those of symbolizing a contact. Its more striking feature is its suggestion that, from the very first moment of meeting, the two cultures did not quite understand each other. It is quite obvious that Abu ‘Ubaydah, in proposing first to have his own picture [72] painted and in agreeing then to having instead a likeness of ‘Umar, did not take very seriously what seemed to him a peculiarity of the Christians with whom he dealt; conversely, it is also clear that to the Christians the image of Heraclius on a column separating two armies had a meaning far greater than that of a landmark. But here a problem emerges, for, on the one hand, a continuity of contacts and influences existed, and, on the other, there appears at the very beginning a misunderstanding in regard to the significance of images and objects. Should we assume that this misunderstanding was peculiar to the early period and was later to be repaired? Or should we conclude that, through centuries of demonstrable formal or iconographic relationships, there was a more profound lack of understanding between the two artistic traditions? If there was, can one discover its reasons or the significance of the contacts? Considered from this point of view, our problem becomes less that of identifying specific themes than that of suggesting the ways in which artistic traditions grew in relationship to each other. The problem could be examined either by looking at what Islamic art meant to Byzantium or by studying the significance of Byzantine art to the Arabs. It is to the latter that I should like to direct my attention. I propose to analyze in some detail two questions which bear on the problem with particular force: first, the transformation of a Byzantine–Christian material culture into an Islamic one and the impact of this transformation on art, and, second, the iconography of power in early Islamic art. These two topics differ considerably in kind; while the second concerns the deliberate creation of a set of visual

islamic art and byzantium 7

symbols with precise meanings, the first derives from the automatic inheritance by the Muslims of several provinces of the Byzantine Empire. Together they may serve to answer the fundamental questions posed by our story and by the unique historical circumstances in which Islamic art and culture were born: what did the Muslim Arab world know and understand of Byzantine art? How did the Muslims use what they knew? What effect did their use of it have on Islamic art? As we deal with Islamic art and with Islamic civilization in general, we find that the man-made setting within which the culture grew and from which its art developed is of particular significance in determining the relationship between Byzantium and the Arab world. In Syria and Palestine the Arabs inherited a complete and complex entity with well-known physical, human, economic and artistic characteristics. To a degree, the same is true of Egypt and of North Africa, but our archaeological information there is much less complete. In northern Mesopotamia also the Muslims took over a Christian province previously ruled by Byzantine emperors, but a considerable difference exists between the medieval Jazirah and the other provinces, for its incorporation into the Muslim world changed its character: what had been exclusively a frontier area was transformed into an important agricultural and commercial center. Only Syria and Palestine became Muslim without immediate alteration in character, with practically no destruction brought about by the [73] conquest itself, and with a considerable archaeological documentation. An analysis, then, of the character of the art and culture of the Muslims in Syria and Palestine may serve to illustrate a central aspect of our problem: it may show us how the new civilization used those elements of the Christian–Byzantine world which fell into its hands unscathed. As far as the main cities are concerned, little was changed. A new type of building was introduced, the mosque, which in almost all instances known in Syria was located on the site of some older sanctuary. The most celebrated example is at Damascus, where the church of John the Baptist was destroyed and the Great Mosque erected in its stead.8 It is well known that practically all the elements of construction of this mosque (Fig. 1) are characteristic of the architecture prevalent in Syria under Byzantine rule; also that there was a major innovation in the composition of the plan (Fig. 2): the peculiar relationship between court, portico and deeply recessed sanctuary is new and is probably derived from the earlier House of the Prophet in Medina. There is, however, another point which I should like to stress and to which I shall return several times: as one compares this new building with what preceded it, the striking feature is that the new Islamic composition 8

Basic bibliography in K. A. C. Creswell, Early Muslim Architecture, I (London, 1932), p. 97 ff.; important interpretations by J. Sauvaget, La mosquée omeyyade de Médine (Paris, 1947), p. 95.

8

1 Damascus, Great Mosque. Court

early islamic art, 650–1100

re-established the unity of the classical Roman architectural ensemble which had been abandoned by the intervening Christian church, for the Umayyads used as the foundation of their mosque the shape and the dimensions of the Roman temenos and developed their religious structure within the mold created by Classical Antiquity. On the other hand, the Christian building was of much smaller dimensions and could not use the frame provided by the classical construction. In Jerusalem the celebrated Dome of the Rock and the Aqsa Mosque also exhibit techniques of construction and partially of decoration characteristic of Christian art, but what is ultimately the most remarkable feature of the new Muslim creation – the Haram al-Sharif – is again the fact that the Muslims, for political and historical, but especially for ideological, reasons, gave a new holiness to the most ancient sacred spot in the Holy City.9 In other instances, as at Hamah, Christian churches were converted into mosques.10 9 10

O. Grabar, “The Umayyad Dome of the Rock,” Ars Orientalis, 3 (1959). The matter has been debated, but K. A. C. Creswell’s latest discussion seems to have secured the point (“The Great Mosque of Hama,” Aus der Welt der Islamischen Kunst, Festschrift für Ernst Kühnel [Berlin, 1959]).

islamic art and byzantium 9

2 Damascus, Great Mosque. Plan

Except for mosques, however, very little is known of the physical changes introduced by the new culture into Christian cities that had been taken by treaties, usually with considerable limitations on the freedom of action of the Muslim settlers. One example exists of a new city of the early Muslim period in Syria and Palestine – the city of Ramlah, about whose early state unfortunately almost nothing is known.11 In all likelihood little change was forced upon these cities. Yet the very fact that there was a new culture altered the character of some of their forms of life. The new masters influenced the sounds [74] of the city in that most early treaties prohibited the use of the semantron and of religious processions. They affected the ceremonies of the city in that banners, crosses and religious symbols had to be kept inside buildings. And, as for monuments, it is worth noting that minarets, according to the most plausible traditions, appeared first in the conquered cities of Syria.12 Their shape (Fig. 3) derived from the Roman temenos towers of Damascus or from church towers, but a new Islamic meaning was given to these high square towers: not only were they used for calling the faithful to prayer – a function presumably of equal significance in new and purely Muslim cities – but they also served the more important purpose of symbolizing the presence of the new faith in the midst of a predominantly non-Muslim population. In the cities, then, one may assume a continuation of older patterns of life and construction, with the addition of only a few new architectural 11 12

Cf. G. LeStrange, Palestine under the Moslems (Cambridge, 1890), pp. 303–8. There is no recent study of the minaret and its origins; the older bibliography is summarized in Creswell, Early Muslim Architecture, I, pp. 38–40.

10

early islamic art, 650–1100

3 Damascus, Great Mosque. Minaret

compositions serving precise new needs but without as yet the use of many new forms or techniques; it is probably only the mood of the cities that changed, but this mood cannot be reconstructed with the evidence in our possession. In the countryside, on the other hand, things seem to have been much more complicated. What evidence can be reconstructed may be summarized as follows.13 During late Hellenistic and Roman times a striking development took place in the whole of Syria, Palestine and Transjordan, as enormous investments in hydraulic equipment (cisterns, canals, dams, aqueducts, waterworks of all types) brought about a tremendous growth of the agricultural infrastructure of the area. The whole region which extends from the Euphrates at Rusafah through Palmyra and the Hawran and then straight south to the gulf of ‘Aqabah became as busy agriculturally as it had always been 13

For a more detailed statement of the problem, see O. Grabar, “Umayyad ‘Palace’ and ‘Abbasid ‘Revolution,’” Studia Islamica, 18 (1963). [Reproduced in this volume as Chapter VII.]

islamic art and byzantium

11

commercially. Under the Christian emperors the development of the area did not stop; perhaps it even increased, as cults of saints, cenobitic practices and other phenomena which have frequently been analyzed created a sort of fashion for inaccessible places of retreat, most of which required a minimal supply of water.14 The financial bases and economic justification of these numerous settlements are not well known, except in the case of northern Syria, where G. Tchalenko has shown that the cultivation of olive trees and the manufacture of olive oil, principally for export, were the major occupations of these settlements.15 An additional cause for the growth of agriculture was the necessity of feeding and caring for the large numbers of local and foreign pilgrims attracted from the whole of Christendom to the Holy Places of Syria and Palestine. The architectural typology of this period has not been studied [75] in all of its details, except in the instance of churches and in the small area studied by Tchalenko, but even a cursory examination of travelers’ reports indicates a tremendous variety of buildings, from churches to simple farmhouses. It is striking to note, however, that there are almost no extant examples of true palaces in this whole agricultural area, the single exception being Qasr ibn Wardan, whose strictly Constantinopolitan inspiration has been noted more than once.16 Such palaces as existed were in the cities, where the larger landowners or the representatives of the government lived. As the Muslim conquest took place, this whole area fell into the hands of the Arabs, and, with the creation of the first Muslim Empire, almost all the identifiable economic reasons for the continued existence of the area began to disappear. Trade with the outside world dwindled considerably, as far as cheap commodities were concerned, in spite of several attempts to revive it; holy places were no longer visited by Christians as commonly as before; the wealthy investors from the cities emigrated, as did Byzantine officials. Yet no destruction took place and, according to the practice of the conquest and of Islamic law, most of the land thus abandoned by the owners fell into the category of booty to be distributed by the new state among members of the ruling family and their allies.17 As the new owners took over, a most remarkable change began to affect the whole area: in a few decades it became covered with palaces or at least very large and very rich private houses. Some thirty to forty early Islamic castles remain which were built on land the hydraulic infrastructure of 14

15 16 17

Christian Palestine was discussed in some detail at a symposium held at Dumbarton Oaks in 1955. The extraordinary development of Christian Syria and Palestine is made abundantly clear by the well-known older explorations of de Vogue, Butler and others, and, more recently, by the work of N. Glück (summarized in Rivers in the Desert [New York, 1959]) and the explorations of the Franciscan fathers in Jerusalem (B. Bagatti and S. J. Saller, The Town of Nebo [Jerusalem, 1959]). G. Tchalenko, Villages antiques de la Syrie du Nord, 3 vols (Paris, 1953–58). Butler on Qasr ibn Wardan; K. Swoboda, Römische und Romanische Paläste (Vienna, 1924), p. 156 ff. For references, see Grabar, “Umayyad ‘Palace’ and ‘Abbasid ‘Revolution’”.

12

4 Khirbat alMafjar. Plan

early islamic art, 650–1100

which was pre-Islamic; the neighboring farmhouses and often the very stones of the palace itself were also pre-Islamic, but the main buildings were new.18 What these buildings were can easily be seen in the six examples which can be analyzed in some detail: Jabal Says and the two Qasr al-Hayr in Syria, Khirbat al-Mafjar and Khirbat al-Minyah in Palestine, and Mshatta in Transjordan. All these are square, fort-like structures with strongly emphasized gateways (Fig. 4), small mosques, luxurious baths, and an extraordinary wealth of mosaics, paintings, stuccoes, stone sculpture and other symbols of rich life. Agricultural and ecclesiastical settlements were transformed into manorial enterprises. At times, the old agricultural exploitation probably continued; at other times, what had been a site of extensive farming was transformed into parks, game preserves and other characteristic features of high princely living. Altogether the area acquired an aristocratic residential character developed by and for the major princes of the dynasty. Amenities of high urban living, such as baths, were introduced into a land which fell, so to speak, by default into the hands of the Arabs and whose earlier functions could not be continued in the same fashion. 18

A complete list of these settlements has not yet been made; the fullest is that compiled by J. Sauvaget (“Remarques sur les monuments omeyyades,” Journal asiatique, 231 [1939]); cf. also Grabar, “Umayyad ‘Palace’ and ‘Abbasid ‘Revolution’”.

islamic art and byzantium

13

This phenomenon explains, among other [76] things, the constant reference in texts to peregrinations of Umayyad princes.19 These texts should be interpreted not as expressions of an unproved traditional nomadism, but as references to visits to agricultural enterprises or to estates. So far our analysis has suggested that the phenomenon of the Umayyad castles in Syria should be explained by the existence of the infrastructure of an agricultural organization which had been carried on, if not always created, by the Byzantine world. It remains to be seen whether the formal characteristics of the castles were as original as their existence. In a general way they were, like the mosques, a new combination of old forms. The square with round towers and a central porticoed courtyard belongs to a characteristic type of fortress and palace found in the Mediterranean area and in Mesopotamia.20 This central enclosure was the main residential area in which the princes lived and received. The throne room was often on the second floor, over the entrance, but in most instances has not been preserved. Throne rooms extant on the ground floor exhibit a ubiquitous “basilican” plan, as at Minyah (Fig. 5) and Mshatta (Fig. 6), although in the latter case the plan was modified by introducing an Oriental relation of covered hall to court and by the addition of a triconch.21 Baths also became a standard feature of these palaces; the small rooms actually used for bathing (Fig. 7) were taken directly from a type of small bath which had developed in Syria and has already been found in Dura-Europos. But the significant feature of the Umayyad bath is the extraordinary transformation of the apodyterium; at Qusayr ‘Amrah it was turned into a throne-room, at Khirbat al-Mafjar into a spectacular centrally planned hall (Fig. 4). The architectural origin of the latter is still unclear but, by analogy with Qusayr ‘Amrah and for reasons to be suggested presently, it can be assumed also to have been an official secular hall. Mosaics, usually without any figurative elements, have been found on the floors of these palaces; they range from the superb geometric designs of Mafjar and Minyah to the badly preserved and still unpublished ones at Qusayr ‘Amrah.22 Rich stucco sculptures and mural paintings not only served decorative purposes, but also depicted various 19 20

21 22

The texts have been gathered repeatedly, for instance, by H. Lammens, “La Badia et la Hira sous les Omeyyades,” Mélanges de la Faculté Orientale, Univ. St. Joseph, 4 (1910). H. C. Butler, Princeton University Expedition, II, Architecture, 2 vols (Leyden, 1919–20), A, p. 145 ff.; B, pp. 47 ff., 63, etc; A. Poidebard, La trace de Rome dans le désert de Syrie (Paris, 1934), passim. Latest discussion in I. Lavin, “The House of the Lord,” The Art Bulletin, 44 (1962), p. 11, where the appropriate bibliography will be found. For Mafjar, see R. W. Hamilton, Khirbat al-Mafjar (Oxford, 1959), p. 327 ff.; for Minyah, the only publication is that of O. Puttrich-Reignard, “Die Palästanlage von Chirbet el Minje,” Palästina-Hefte des deutschen Vereins vom Heiligen Lande, Heft 17–20 (1939), with a drawing in color; the Qusayr ‘Amrah mosaics are to be published by K. Kessler in the forthcoming volume in honor of K. A. C. Creswell.

14

early islamic art, 650–1100

5 Khirbat alMinyah. Plan

themes of a royal iconography of pleasure and power.23 This use of the bath as a ceremonially significant part of the palace is not entirely new, since it does find parallels in Roman imperial art and in some of its provincial manifestations, but, within the Umayyad [77] culture, it should be considered as an architectural expression of the majlis al-lahwa, that is, a place which was used for a ceremonial entertainment; traditional Arab customs, such as the recitation of poetry, were expanded to include official drinking, dancing and music in the Sasanian manner, as well as various other activities, such as banqueting, which relate the Umayyad practice to those associated also with Roman triclinia.24 23 24

Cf. below and, for the interpretation of a mosaic at Khirbat al-Mafjar, R. Ettinghausen, Arab Painting (Geneva, 1963), p. 38. The demonstration of this point is beyond our immediate subject. It has been treated in my doctoral thesis at Princeton University, Ceremonies and Art at the Umayyad Court (1955), and I hope to return soon to some of these problems.

islamic art and byzantium

15

6

While most of the architectural components of these palaces and many of the habits of life enjoyed in them find parallels in the Mediterranean tradition, the specific combination of functions illustrated by the palaces cannot be explained as a Byzantine Christian type modified by various characteristics of a new taste. For, except for a few depictions on mosaics in North Africa,25 there is no evidence that the combinations of forms and purposes which appear in Umayyad palaces were characteristic of secular architecture of the Near East before the Muslim conquest. Even the Tunisian mosaics are over a century earlier, and it is hardly likely that they exerted an impact on early Islamic palaces. Furthermore, the obviously makeshift arrangement of a plan like that of Khirbat al-Mafjar (Fig. 4) and the anomalously composite character of a façade like that of Qasr al-Hayr (Fig. 8) suggest that the Umayyads were in fact creating something new, that they were searching for some kind of formal entity that would tie together functions which had not, until then, been organized. It is in this fashion that one can explain the

25

Swoboda, Römische und Romanische Paläste, pl. V.

Mshatta. Plan

16

early islamic art, 650–1100

7 Qusayr ‘Amrah. Plan

progression from Khirbat al-Minyah (Fig. 5), with its mosque and official halls bursting the straitjacket of the fort-like castle, to Mshatta (Fig. 6), unfinished but superbly planned as a single entity. The other palaces fit between these two extremes as various steps in the direction of a complete composition. This very progression and the frequently inferior quality of the architecture contribute toward the argument that no real models existed for the Umayyads in Syria and Palestine or, for that matter, in Iraq. Indeed, the Umayyad palace – when seen in the context of the Byzantine architecture which preceded it in Syria and Palestine – appears as an original creation, made possible by the peculiar combination of four features: a highly developed agricultural infrastructure created several centuries earlier; the emigration of large landowners; the existence of an aristocratic ruling group; and the availability of themes, ideas, tastes and modes of behavior drawn from the entire breadth of the newly conquered world and amalgamated with older Arabian habits. But, and this point is crucial, what was thus created resembled in many aspects Roman and late antique palace architecture rather than Byzantine. The apparent fortification, the tremendous display of decoration, the comparative seclusion, the relationship to cultivated land – all are features which are close to Spalato, Piazza Armerina, Pliny’s villas, especially the Villa Urbana, or North African manors.26 [78] However, the significance of the Umayyad palace does not lie merely in the peculiar fashion in which it developed or in the manner in which its 26

This is a point with which I have dealt in papers presented at the Centro Italiano di Studi sull’ alto Medioevo in Spoleto in April 1964 and at the annual meeting of the College Art Association in Philadelphia in February 1964. The papers will be published in the journals of these two organizations, respectively.

islamic art and byzantium 17

8 Qasr al-Hayr. Façade, as reconstructed in the Damascus Museum

18

early islamic art, 650–1100

9 Qasr al-Hayr. Sculpture on façade

function and its architectural characteristics relate it to older types of building. It became also the crucible to which themes and ideas from the whole length and breadth of the newly conquered world were brought; there were borrowings from the past of that world, as in the case of Palmyrene sculpture imitated at Qasr al-Hayr (Fig. 9),27 or from its far-flung provinces, as in the case of paintings or sculptures inspired by Central Asian art.28 It is in this crucible that a new Islamic decorative art was created; elements from many areas were mixed together, at times incongruously, as in a fragment from Mafjar (Fig. 10) in which a flat frame of interlocking circles serves as a background for strangely projecting busts. In this development of decorative themes Byzantine art played an important part, but only as one of the contributors to a new series of syntheses. For the formation of early Islamic palace art, the conquest of a large world with an immense wealth of styles and objects was as important as – if not more important than – the location of the palaces in an area formerly ruled by Byzantium. 27 28

This sculpture was published by D. Schlumberger, “Les fouilles de Qasr el-Heir Gharbi,” Syria, 20 (1939), pl. XLVI, 2. Many sculptures from Khirbat al-Mafjar (Hamilton, Khirbat al-Mafjar, pls XLIV, 4 and LVI) and paintings both from Mafjar and from Qasr al-Hayr (the latter for the most part unpublished) seem clearly to have been inspired by Central Asian works of art.

islamic art and byzantium 19

The palaces were doomed before their art had developed to its fullest extent. The fall of the Umayyad dynasty in 750 led to the almost immediate abandonment of the whole agricultural area in which palaces have been found. The cities of Syria and Palestine also declined, but the impact of the Umayyad creations remained for many centuries and in many regions. The persistence of mosaic decoration in Cairo as late as the fourteenth century; the square minarets of Morocco; the fact that so much of the architecture in the Arab world continued for so long to depend on the arch and the column rather than on vaults; the persistence of ornamentation with a clear organization of vegetal forms, such as can be seen in Fatimid Egypt or in Spain; the plan imposed on the mosque of Damascus by the old sanctuary of the city; all these and many others are features which remained Muslim because they were naturalized in Umayyad Syria and Palestine. At the same time, what emerged from the palaces or the city mosques to have a lasting effect on Islamic art was not Byzantine art but a number of techniques and a large number of motifs. However, the brilliance and wealth of Byzantine Syria and Palestine and the peculiar ecology inherited by the Arabs were the principal factors that made possible the brilliance and wealth of the Umayyad art of Syria and Palestine. These are all the more remarkable when compared to the primitive simplicity of Umayyad art in Iraq. Umayyad art was not a province of Byzantine art, but the extraordinary fact that there was an identifiable art sponsored by the Muslim Arabs a few decades after their emergence from Arabia is to [79] be attributed to the setting which they inherited from Byzantium and which they transformed in ways and for reasons peculiar to them. From a survey of the setting in which early Islamic art was created, I turn now to my second topic: the consideration of an iconographic theme. It is only in the past two decades that iconographic studies have begun to revolutionize our knowledge of Islamic art and to replace the romantic

10 Khirbat alMafjar. Sculpture from Palace entrance

20

early islamic art, 650–1100

conception of a purely decorative art. The task of so changing our vision has been the almost single-handed work of R. Ettinghausen, who, beginning with an article published in 1943,29 has been continuously pointing out that the most characteristically “decorative” objects and motifs possess also a level of social, intellectual, or even religious meaning hitherto rarely seen. Without denying the considerable ornamental values developed by Muslim artists, as early as for the mosaics of the Dome of the Rock, it remains true that these artists and their patrons on numerous occasions used works of art to express certain ideas. For our purposes here the interest to be found in an analysis of such a theme is quite evident, since the creation of an iconography imposes on the patron or on the artist a deliberate choice of subject-matter. Furthermore, if we recall that from its Arabian past the new Muslim art could draw almost nothing, it follows that it was from the rich heritage of the Mediterranean and of the Near East that its main themes were borrowed; and thus the question of the conscious uses of Byzantine art is raised. I have chosen the theme of power as expressed in monuments because it was one of the first to be developed, because, due to the high positions of its sponsors, it is one of the better-documented themes, and also because it was a theme hardly peculiar to Islamic art, but characteristic of all imperial arts, and it was brought to particular intensity in the art of Late Antiquity and of Byzantium. There are two main periods and two groups of monuments around which I should like to develop the theme: the Umayyad period and its palaces and mosques, and the twelfth and thirteenth centuries and their illustrated books. In discussing these monuments, however, I shall stress only their specific relation to Byzantine art, it being assumed that there are other coordinates through which many of them can and should be understood. The first group of monuments occurs in the Umayyad period, when, in the later part of the seventh century and in the first two decades of the eighth – more particularly under the rules of ‘Abd al-Malik and of al-Walid I – the Muslim princes established the first elements of a characteristically Muslim imperial organization. Together with various administrative and other practices, there appeared also the first elements of a visual symbolic system which will serve to identify the Islamic world. The main purpose of this system was to symbolize power, i.e., to emphasize the existence, the greatness and the individuality of the new world. Its first characteristic expression occurred on coins. There, as is well known, the Muslims had first used Byzantine and Sasanian types [80] (Fig. 11) with very few modifications;30 then changes began to occur: new inscriptions with the caliph’s name, new iconographies, such as 29 30

R. Ettinghausen, “The Bobrinski Kettle,” Gazette des beaux-arts, 24 (1943). The basic publications are by J. Walker, Arab–Sassanian Coins (London, 1941) and ArabByzantine and Post Reform Umaiyad Coins (London, 1956). See also G. C. Miles, “Mihrab and ‘Anazah,” Archaeologica Orientalia … (Locust Valley, 1952).

islamic art and byzantium 21

12 Arab– Byzantine coin

11

Arab–Byzantine coin

14 Arab– Byzantine coin

13

Arab–Byzantine coin

15 Post-reform Umayyad gold coin

a caliph’s portrait (Fig. 12), a peculiar orant (Fig. 13), or, in a rare instance, a search for a specifically Muslim iconography. In a celebrated coin (Fig. 14) published by G. Miles the main themes were symbols of Islamic power, a mihrab (or merely a niche of honor) and the ‘anazah, the lance of the Prophet. The significant points about this coin, however, are that it remained an extreme rarity and that, for reasons to be suggested presently, its very Islamic symbolism was not continued. In fact, sometime thereafter, probably in 695, a new Muslim coinage (Fig. 15) was established in which words alone, the profession of faith and the qur’anic verse of the mission of the Prophet to the

22

16 Damascus, Mausoleum of Baybars. Mosaic fragment

early islamic art, 650–1100

world, expressed the new culture. This coinage remained standard, with only minor exceptions, almost until today. At the very same time, in the Dome of the Rock and in the mosque of Damascus, respectively, the imagery created was of crowns suspended around a sanctuary and of an idyllic landscape. This imagery is, again, an expression of power, power of the victory of Islam over its Christian and Sasanian antagonists and power of the idealized complete world ruled over by the Muslims. The proposed justifications for these interpretations have been given elsewhere by R. Ettinghausen and myself,31 but the point which I should like to stress here is that in none of these instances do we find human or animal representations, even though practically all symbols and images of power in Antiquity tended to center around human or animal symbols. The exact iconographic themes of Damascus and Jerusalem do not seem to have been used again in the following centuries, and it is not very clear whether their apparent revival in a few instances by early ‘Abbasid caliphs,32 and especially under the Mamluks in the late thirteenth century (Fig. 16),33 should be given the same precise meaning or whether these examples of architectural themes in mosaics were mere ornaments, probably imitating Damascus mosaics. Although in the intervening centuries the theme of power did not disappear, it tended to find expression not so much through organized imagery as through epigraphy and architectural compositions.34 31 32

33 34

Ettinghausen, Arab Painting, p. 20 ff.; O. Grabar in Ars Orientalis, 3. The case in point is that of the crowns and royal insignia of conquered princes of Afghanistan which were hung in Mecca, al-Azraqi, K. Akhbar Makkah, in F. Wüstenfeld, Die Chroniken der Stadt Mekka, I (Leipzig, 1858), p. 155 ff.; Grabar, in Ars Orientalis, 3, pp. 50–51. These mosaics in the mausoleum of Baybars have often been mentioned, but have not yet been properly published. This point cannot be demonstrated in the context of this paper, but the highly organized plan of Baghdad – city and palace at one and the same time (cf. O. Grabar, “Mshatta, Wasit, and Baghdad,” The World of Islam, ed. by R. B. Winder and J. Kritzeck [London, 1959]; and the forthcoming studies by D. J. Lassner) – the very remarkable inscriptions

islamic art and byzantium

23

[81] The conclusion, therefore, would be that, after a few essays in coinage, the Muslim world abandoned, at least for a while, an official imagery of power and replaced it with certain monumental symbols and especially with the written word. This change, which has long been recognized, though perhaps not quite in these terms, was usually attributed to Islamic iconoclasm, but the difficulty with this hypothesis is that the evidence, derived from works of art, of opposition to representations of living beings precedes any theoretical statement of an Islamic opposition to images, the latter not being evident before the second half of the eighth century.35 It can hardly be doubted that the Muslim interpretation of the Qur’an and perhaps a generally iconoclastic tendency among the Semitic population of the Near East theologically justified the later doctrine and brought about its general acceptance with respect to religious art. It remains true, however, that a conscious refusal to use representations for official purposes and their partial replacement with words had occurred before any systematic formulation of any doctrine took place. An explanation for this phenomenon can be provided if we consider the setting within which the Muslim search for a symbolism of power took place. However one is to interpret the specific changes in coinage brought about by Justinian II, it is clear that a complex iconography of Christ was used on the new coins, and many recent studies have pointed out in general the intellectual and emotional intricacies of religious art in the sixth and seventh centuries which led to the great iconoclastic quarrel.36 Furthermore, even though specific styles varied, the basis repertory of iconographic themes and motifs used in the Christian and imperial arts of early Byzantium was one issued from Antiquity and basically common, or at least understandable, to the whole Christian world. The choice confronting ‘Abd al-Malik, al-Walid and, even more urgently, al-Hajjaj, was either to invent a new iconography (thus the mihrab and lance of the American Numismatic Society coin, Fig.

35

36

on the mosque of Ibn Tulun (E. Combe, J. Sauvaget, G. Wiet, Répertoire chronologique d’épigraphie arabe [Cairo, 1931 and ff.], no. 682), or on the so-called Juyushi mosque (ibid., no. 2753 and O. Grabar, “The Earliest Islamic Mausoleums,” Ars Orientalis, 6, forthcoming), or, finally, the bejeweled axial nave and series of domes built by alHakam II in Cordoba – all of these examples would indicate that very precise architectural practices and official epigraphy served to show the sovereign character of buildings. The problem of the sources of an Islamic opposition to images is still incompletely solved. The latest formal statement on the strictly Islamic side of the subject is by K. A. C. Creswell, “Lawfulness of Painting in Early Islam,” Ars Islamica, 11–12 (1946); see also R. Paret, “Textbelege zum islamischen Bilderverbot,” Das Werk des Künstlers, Hubert Schrade zum 60. Geburtstag (Stuttgart, 1961). The problem, however, has to be considered in the light of contemporary Byzantine ideas, on which see E. Kitzinger, “The Cult of Images in the Age before Iconoclasm,” Dumbarton Oaks Papers, 8 (1954); and A. Grabar, L’iconoclasme byzantin (Paris, 1957). There is also the question of the so-called edict of Yazid II, A. A. Vasiliev, “The Iconoclastic Edict of the Caliph Yazid II, A.D. 721,” Dumbarton Oaks Papers, 9–10 (1956). On this, see below, note 40. See the works quoted in the preceding note, and J. Breckenridge, The Numismatic Iconography of Justinian II (New York, 1959).

24

early islamic art, 650–1100

14) whose meaning would miscarry since it would not fit the existing formal vocabulary of the Mediterranean and of the Near East, or to use the formal vocabulary (as on early coins) which, while familiar, would inadequately identify the novelty and uniqueness of the new empire and the new faith, for the Umayyads could hardly in one generation acquire the sophisticated practice of imagery which characterized Byzantium. Faced with this dilemma, the Muslims tried both alternatives, but soon discarded imagery, and, as we have seen, adopted the techniques of Byzantium without its formulas. Thus, to avoid the pitfalls [82] inherent in the use of a highly developed iconography, the Muslims changed the rules of the game. There are two conclusions to be drawn from this analysis. The first is that Muslim theoretical iconoclasm followed a practical refusal to use representations for highly official purposes, and perhaps was influenced by this refusal, although it is possible that philosophical and intellectual iconoclasm developed, at least in part, independently of political and imperial considerations. The second is that Byzantine art was responsible for this early Muslim attitude; yet it is not a particular style or iconographic motif of Byzantine art that is involved here, but rather the fact of that art and of the ideological depth it had achieved. It is in this context that I should like to remark upon a very curious phenomenon, about which much has been written: that of the use of Byzantine workers for the erection of Islamic monuments. I do not refer to the probably very numerous workers from everywhere – including Byzantium – who were attracted by the money and employment available in Umayyad times. I refer, rather, to the specific incident of al-Walid’s request that workers be sent to him by the Byzantine emperor. The evidence gathered recently by Professor Gibb37 has clearly shown that Byzantine workers were brought by al-Walid I to help decorate the great mosques of Damascus, Madinah, and perhaps Jerusalem. Professor Gibb also noted that for the Byzantines this action had one meaning and for al-Walid quite another. To the Byzantines it was an imperial act granting to “barbarians” the privileged use of highly technical training which, by its very quality, served to enhance the prestige of the Byzantine emperor and, presumably too, to bring the “barbarian” into the imperial fold. To al-Walid it was partly a way of “learning the ropes,” as Professor Gibb has put it, for the Muslim caliph was anxious to possess all the characteristics of an emperor (one of these being the sponsorship of superb monuments for his own followers), and partly a way to impress the Christians of the empire. Thus, as in the story of Abu ‘Ubaydah at Qinnasrin, opposite meanings were given to a single event. But perhaps we may be able to go a step further. It was not technically necessary for mosaicists and decorators to come officially from Byzantium, for the examples of mosaics in the palaces and especially in the Dome of the 37

H. A. R. Gibb, “Arab–Byzantine Relations under the Umayyad Caliphate,” Dumbarton Oaks Papers, 11 (1958).

islamic art and byzantium

25

Rock are of as fine quality as those of Damascus, yet there is no evidence that they were done by Byzantine artists sent from Byzantium as part of a high-level treaty agreement. In fact, with regard to the Jerusalem sanctuary the circumstances of the times and the specific meaning of the building itself make it highly unlikely that any such artists could have been there. Furthermore, the iconographic program of the mosque of Damascus refers to specifically Islamic ideas and does not merely copy Byzantine models; and the palace mosaics, with their highly developed geometric designs (Figs 17 and 18), are the conscious choice, with modifications, of late antique models rather than [83] of the latest Byzantine styles. If workers were available in the Muslim Empire and if iconographic ideas had been developed by the Muslims, why did al-Walid send for artists from Byzantium? Two explanations for his action can be suggested. The first would derive from an element in the early Muslim state which has not yet been sufficiently examined in discussions of Umayyad culture. It may be defined as an almost messianic feeling that the Empire of Rome would surely fall to the Muslims.38 Military and psychological 38

M. Canard, “Les expéditions des Arabes contre Constantinople dans l’histoire et dans la légende,” Journal asiatique, 208 (1926).

17 Khirbat alMafjar. Mosaic fragment

26

18 Khirbat alMinyah. Mosaic fragment

early islamic art, 650–1100

hopes for this event collapsed after the failure of 717, but the call for artists and artisans up to that date can be understood as having symbolized in fact the subservience to the Muslims of the Byzantine emperor, who, like a vassal, must provide his overlord with artisans. This explanation would be in keeping with the interpretation of the iconography of the mosaics of the Mosque of Damascus suggested by R. Ettinghausen, and with at least some of the texts describing al-Walid’s call for Byzantine artists, texts which emphasize that it was the Muslim prince who gave the “orders.” The second explanation is simpler: al-Walid’s request derived from snobbism, the conviction that first-rate works could come only from Byzantium. Whether or not the latter explanation is fully valid for the caliph may be open to some doubt. It hardly coincides with his truly imperial ideas and construction program. I would rather, therefore, suggest that, like the new Muslim coinage and like the iconography of the mosque of Damascus and of the Dome of the Rock, al-Walid’s call for workers from Byzantium was meant to be a sign of the Muslim prince’s accession to universal power, the very theme of the celebrated fresco of the Six Kings at Qusayr ‘Amrah (Fig. 19).39 39

O. Grabar, “The Painting of the Six Kings at Qusayr ‘Amrah,” Ars Orientalis, 1 (1954); Ettinghausen, Arab Painting, p. 30.

islamic art and byzantium

27

19 Qusayr ‘Amrah. Fresco of the Six Kings

But, as the defeat of Byzantium became impossible and as the Muslim world turned eastward, the events of al-Walid’s reign became myths and the Rumi, the Byzantine, became the artist par excellence, later to be joined by the Chinese. The myth survived because the buildings of al-Walid in Jerusalem, Damascus and Madinah remained central sanctuaries in the succeeding history of Islam and because the early Muslim dream lingered, however hazily, in later historiography and legend. Thus, once again an important characteristic of the Islamic attitude toward the arts, i.e., the evaluation of the Rumi artist, derived from the peculiar situation of Byzantine and Arab relations in the early eighth century; in this instance, it was the result of an expression of power by al-Walid; the Muslim vision of the defeat of Byzantium led eventually to the assumption of the superiority of the Byzantine artist.40 40

It could be noted that much in our analysis of early Islamic art finds parallels in early Christian art, where a similar dichotomy of opposition and attraction existed between the new faith and Classical Antiquity. The main difference is that Islamic culture did

28

early islamic art, 650–1100

[84] While coins and the decoration of mosques show an ultimate refusal to use more than Byzantine techniques, and while an attempt at imperial power by al-Walid led in part to the mythical eminence of the Byzantine artist, palace art reflects a quite different story. There also an art was developed whose aim was to glorify the prince and to emphasize his power. We could not have a clearer illustration of this than the celebrated fresco of the Six Kings at Qusayr ‘Amrah (Fig. 19). Yet an important difference existed between the art of early Islamic palaces and the imagery or symbols on coins and in mosques, for palace art was designed primarily for private audiences. It was no less an art of power than the other, but its uses were more intimate and less accessible to the general public. Basically it revolved around a ceremonial theme much less Byzantine than Oriental, the theme of the royal pastime, in which the life of the prince was expressed through its association with hunting, games, banqueting, dancing, music, scantily clad females, etc.,41 and it is these very activities that form the major subject-matter of the paintings of Qusayr ‘Amrah and of the sculptures of Qasr al-Hayr and Khirbat al-Mafjar. The styles in which these themes were executed vary considerably from pseudo-classical to Central Asian and Indian, and each style poses separate problems of influences which cannot concern us here. In one respect, however, they do bear directly on our subject. In every one of the three buildings I have mentioned there are representations of princes and attendants. In one palace, Qasr al-Hayr, there are two representations of princes, making a total of four royal images.42 Two of

41 42

not have the four centuries of incubation of Christianity before emerging as a unique political and cultural entity. It is in the light of these conclusions that I should like to place the edict of Yazid II, although a full treatment of the subject is beyond the scope of this paper. There is no doubt that political persecution of Christians took place at that time, as is evident from the fact that Christian sources (whether in Arabic, Greek or Syriac) are primarily responsible for the preservation of the edict. It also stands to reason that anti-iconoclastic milieux, in fighting their own battles in Constantinople, would have tended to assimilate their own enemies to the one caliph who clearly had persecuted them. That Yazid had become a “scapegoat” can be shown by the passage in Dionysius of Tell-Mahre which follows the text of the edict (trans. by J. B. Chabot [Paris, 1895], pp. 17–18); there Yazid is also accused of having created a fearful stench throughout the Muslim world by having ordered the killing of all white dogs, pigeons and roosters. The absurdity of the accusation suggests a “smear campaign” rather than a factual account. I would, therefore, prefer to consider the edict of Yazid as reflecting an anti-Christian program, which, because of contemporary Christian problems and later developments in both Byzantium and Islam, took an iconoclastic turn in the sources. The precise demonstration of this point in Umayyad palaces, the textual evidence for its uses in ceremonies, and its Iranian origins have not so far been put together in print. All these images have been published. Mafjar: Hamilton, Khirbat al-Mafjar, pl. LV, 1 and 5; Qasr al-Hayr: Schlumberger, “Les fouilles,” pl. XLV, and S. ‘Abd al-Haqq, “I‘adah tashiyid jinah Qasr al-Hayr,” Annales archéologiques de Syrie, 1 (1951), pl. 92; Qusayr ‘Amrah: A. Musil, Kusejr ‘Amra (Vienna, 1909), pl. XV. The problem is whether these four figures are correctly interpreted as princes or caliphs. In the case of ‘Amrah, there is absolutely no doubt about it, since the inscription around it refers to an amir (J. Sauvaget in Journal asiatique [1939], p. 14); in the instances of the Sasanian image at

islamic art and byzantium

29

these princes, one at Qasr al-Hayr (Fig. 20) the other one at Mafjar (Fig. 21), are clearly of Sasanian inspiration, and the considerable differences which exist between them suggest that they derived from representations rather than from actual vestimentary practices. The other two princes, from Qasr al-Hayr (Fig. 22) [85] and from ‘Amrah (Fig. 23), just as obviously reflect Byzantine, or at least late classical, art, as do some of the attendants. For us the important points are that, in the more intimate atmosphere of the palaces, an iconography for princes did fully develop, and that its themes were adopted by the Muslims from conquered lands. But it was images of emperors, rather than of just any human figures, that were chosen as models, and this indicates the Muslims’ awareness of the meaning of the imagery in the original works. Furthermore, the vestimentary variations of these images serve to illustrate yet another point, to which we have been led before: it is that in the early eighth century a prince could be represented only in Sasanian or Byzantine garb, for the Muslim princely image had not yet acquired proper iconographic values in any other system of clothing. However, in those instances where Muslim identification was essential, devices were found, either through a new kind of headgear as at Qasr al-Hayr43 or, most significantly, through an inscription as at Qusayr ‘Amrah, where the written word gives concrete meaning to an abstract iconographic cliché of a prince. In any case, in these instances the role of Byzantine art is clear: it was one of the sources from which the new Muslim art chose elements which served to illustrate its own needs and practices. It is particularly unfortunate that, after a considerable documentation for the first half of the eighth century, there is almost no evidence from succeeding centuries which would allow us to trace the growth of an Islamic iconography of power, let alone the precise Byzantine sources of such an iconography. The shift of power toward the east certainly led to an increase in Iranian influences.44 It is only in the twelfth and early thirteenth centuries that new documents once again come to light. At that time, in the midst of a sudden and

43 44

Qasr al-Hayr and at Mafjar, there is no certainty, but considerable likelihood, because of the position of the statues on major entrances and because of other symbols of power (such as the crown at Qasr al-Hayr and the lions at Mafjar). The doubtful example is the second one at Qasr al-Hayr, which, according to the reconstruction in the Damascus Museum (Annales archéologiques de Syrie, pl. 8), was on the inner façade of the main part of the building and not quite in the center of the composition. Yet I fail to see what other meaning could be given to this fragment, inasmuch as the ‘Amrah example makes it quite clear that Byzantine stylistic origins were indeed possible for princely images of Umayyad times. See Schlumberger’s remarks in Syria, 20 (1939), pp. 353–4. Unfortunately, we are very badly informed concerning ‘Abbasid or Fatimid visual symbols of power. It is possible that more thorough searches through texts may uncover interesting instances of the development of an idea which is not likely to have disappeared or to have been entirely sublimated into epigraphy. In the meantime, see M. Canard in Byzantion (1951) and D. Sourdel, “Questions de cérémonial ‘Abbaside,” Revue des études islamiques (1960).

30

20 Qasr alHayr. Sculpture of prince

early islamic art, 650–1100

islamic art and byzantium

31

21 Khirbat alMafjar. Sculpture of prince

32

early islamic art, 650–1100

22 Qasr alHayr. Sculpture of prince

widespread emergence of new techniques, new types of objects, new purposes, and new centers of artistic activity and production, the art of book illustration was created or revived. Two peculiarities of this new art concern us here. First, from what is known about it, by far the greater bulk was in Arabic, and some of its most typical examples consisted of characteristically Arab texts. Second, these books were intended mainly for a sophisticated urban community, a milieu quite different from that of the princes. The quality of the books was often expressed by their frontispieces, and these can be divided into two broad groups. One consists of a series of variations on the author portrait, in which Byzantine models were used almost systematically, as in an example from a Dioscorides manuscript (Fig. 24) and in a well-

islamic art and byzantium

33

known instance from the Vienna pseudo-Galen (Fig. 25);45 incidentally, some of these instances are related more [86] closely to early Byzantine models than to later ones. The second group consists of a well-developed princely cycle, in which a tradition much more specifically Islamic, with considerable Turkish and Iranian elements, was depicted. Yet in the composition of frontispieces from an Aghani manuscript (Fig. 26) and from the Vienna Galen (Fig. 27)46 there is a striking resemblance to earlier Byzantine compositions, as we know them, for instance, in ivories (Fig. 28), in which the upper and lower as well as the side borders – the latter arranged in two rows of figures – framed and enlarged upon the central theme.47 45

46

47

The Dioscorides frontispiece was first published by A. Süheyl Unver, Istanbulda Dioscorides Eserleri (Istanbul, 1944), figs 1–2; for the Galen, see Ettinghausen, Arab Painting, p. 92. See also R. Ettinghausen, “Interaction and Integration in Islamic Art,” in Unity and Variety in Muslim Civilization, ed. by G. von Grünebaum (Chicago, 1955), p. 119 ff., where some relatable examples are discussed. The Aghani group was published by B. Farès, Une miniature religieuse de l’Ecole de Baghdad (Cairo, 1948); a considerable controversy developed around B. Farès’ thesis, which does not concern our subject, but further literature included an important article by D. S. Rice, “The Aghani Miniatures,” The Burlington Magazine, 95 (1953), and a second volume by B. Farès, Vision chrétienne et signes Musulmans (Cairo, 1961). It should be pointed out that this particular arrangement was not peculiar to Byzantine compositions, but existed also in Manichean and Buddhist compositions of Central

23 Qusayr ‘Amrah. Fresco of prince

34

early islamic art, 650–1100

24 Istanbul, Topkapi Saray. Ahmet III, no. 2147, frontispiece with portrait of Dioscorides

It seems apparent, then, that even in the thirteenth century, when the artists of the Arab world renewed the search for iconographic modes to add distinction to the newly created art of book illustration, they once again turned to Byzantine models which, by their presence, seemed almost magically to exalt the quality of the book. The practice could be pursued in other areas as well: in architectural decoration particularly a fascinating return to the use of classical themes on façades is apparent. The late D. S. Rice once showed me an extraordinary stone found at Harran in which a perfect fifthcentury molding was dated by an inscription in the thirteenth century, and Asia as well as in those of the Carolingian West, all probably having common roots in Roman art. For examples and relation to a frontispiece of the thirteenth to fourteenth centuries in Istanbul, see now E. Esin, “Two Miniatures from the Collections of Topkapi,” Ars Orientalis, 5 (1963); for Carolingian examples, see, for instance, Homburger, “L’art carolingien de Metz et l’Ecole de Winchester,” Gazette des beaux-arts, 62 (1963). A case can indeed be made to the effect that it was in Central Asia that the princes of the thirteenth century acquired a taste for such compositions, and the subject deserves a fuller investigation than it has received so far; however, the very Arab character of the illustrated texts leads me, at least for the time being, to prefer a western Byzantine or Roman background for this type of composition.

islamic art and byzantium 35

25 Vienna, Nationalbibliothek, A.F.10, authors’ portraits

36

early islamic art, 650–1100

26 Kitab alAghani, frontispiece

on the façade of one of the monuments of Nur al-Din in Damascus, a completely classical lintel is surmounted by a very medieval muqarnas (Fig. 29). The same idea was applied to some of the coins minted in Northern Mesopotamia and in Anatolia,48 on which images of Constantine and other Byzantine or Roman emperors appear. Obviously the analyses and examples I have gathered for this study do not tell the whole story of the artistic relations between Byzantium and the Arab world; nor can I pretend to have mentioned all the elements in Islamic art that derived from or were fashioned in Byzantium. I have tried, rather, to present an interpretation of the systems of association between forms and functions and between images and needs which Byzantine art, for various historical and geographical reasons, imposed on the new culture. I have focused my attention on the Umayyad period in which documents are particularly plentiful and relations between the cultures especially strong, and in [87] the use of one illustrative instance I have tried to show that the practice was continued at later periods. The problem of Byzantine– Arab artistic relationships could be extended to deal with such subjects as the palaces of Theophilus in Constantinople and of the ‘Abbasids in Baghdad, or with the uses of architecture evident under the Fatimids, or, even more strikingly, with certain very archaic features of the secular architecture of North Africa and Umayyad Spain. Even in a monument as 48

R. Lane Poole, Catalogue of Oriental Coins in the British Museum, pl. V.

IV

(London, 1879),

islamic art and byzantium

37

27 Vienna, Nationalbibliothek, A.F.10, frontispiece

38

28 Ravenna, Museo Nazionale. Ivory diptych

early islamic art, 650–1100

Islamic as the Alhambra there are antecedents which extend to Byzantium and beyond.49 Three main conclusions emerge from this analysis. First, by the very nature of the history of the seventh century, certain clearly identifiable ecological practices in Byzantium and Syria created the extraordinary phenomenon of Umayyad palaces. Similarly, in later times the social contacts 49

The bibliography on all these subjects, though fairly extensive, is not very enlightening from our point of view. Exceptions would be: A. Grabar, L’iconoclasme byzantin, pp. 144–5; G. Marçais, L’architecture musulmane d’Occident (Paris, 1954), passim; and F. Bargebuhr, “The Alhambra Palace of the Eleventh Century,” Journal of the Warburg and Courtauld Institutes, 19 (1956), p. 217 ff.

islamic art and byzantium

39

29 Damascus, Hospital of Nur al-Din, façade

40

early islamic art, 650–1100

with Christian populations, the reliance on Greek scientific books, and, in a general sense, the Mediterranean orientation of most Arabic-speaking areas in the Middle Ages made inevitable the Arabs’ knowledge of, use of, and reliance upon Byzantine themes. Second, there are almost no instances of the Muslims having borrowed from Byzantium without there being an identifiable need within their culture, or of their having continued with old traditions without making modifications demanded by the new world. This statement may require some correction as one deals with certain details of ornamental themes, but it can be maintained with regard to official art. Islamic art used Byzantine art when it needed iconographic expressions. Byzantine art thus became an essential ingredient in the formation of Islamic art. However, if we examine the nature of the impact of the former upon the latter, we note that, despite the reputation of Byzantine emperors as patrons of art and possessors of artists and treasure, it was not Byzantine art but the themes of Byzantine art which were used by the Muslims. In the one instance of religious and imperial symbolism where the Byzantines had developed a complex iconographic and stylistic mode of representation, the Muslims declined to adopt any of it precisely because it was highly developed. What we must conclude, then, is that Byzantine art provided the new culture with a vocabulary and with the rudiments of a grammar, but that the language developed therefrom was a new one. During its development, as the need for new themes and modes occurred, the Muslims turned again and again to the wellspring of Byzantium, much as Renaissance word-makers, partly out of snobbery and partly out of a genuine need for new words, turned to Greek for a vocabulary, from which certain words have since become popular, while others have faded away as artificial and meaningless concoctions. The idea that for the Muslims Byzantine ways were a means, not an end, an essential element in the creation of a symbolic system, not the system itself, is clearly illustrated in a celebrated story concerning Mu‘awiyah.[88] At one time he was upbraided by the Caliph ‘Umar for having adopted the foreign ways of the Caesars and the Khosros. Mu‘awiyah answered that Damascus was full of Greeks and that none of them would believe in his power if he did not behave and look like an emperor.50 To answer, then, the question posed by my original story: the early Muslims never fully understood Byzantine art, but circumstances having forced it on them, they could not but be impressed by its existence. The evidence I have presented leads us, however, to a third conclusion. One of the striking characteristics of the themes used by Islamic art, brought to it by Byzantine hands, is that so many of them are quite classical. The illusionist style of fragments of the Damascus mosaic decoration, the arch on column, the agricultural structure of Syria, the palace-villas, the 50

The story has been related often. One of its earliest versions is in Tabari, Annales, ed. by M. de Goeje and others (Leiden, 1879 ff.), 2, p. 207.

islamic art and byzantium

41

compositions of frontispieces – and to these can be added other features not treated here, such as the shapes of mausoleums and vegetal decorative designs – all reflect the art of Antiquity, and almost never did the newer and emotionally deeper Byzantine mode make its appearance. Should we understand from this that, at the time of the formation of Islamic art in Syria and throughout the period of later Byzantine art, the classical mode was so lively that it was more easily adopted? Should Umayyad art be used, then, as a major document in the assessment of Antiquity in Byzantium? Or should we rather feel that the artistic language of Antiquity was wider in spirit, more abstract and more adaptable to new needs than the engaged art of Christian Byzantium? Answers to these questions are beyond the scope of this paper and beyond my competence, but, whatever the answers, the questions illustrate the deeper meaning of Byzantium to Islamic art. Far more than from any other artistic tradition which created Islamic art, it was from the Byzantine that the new culture most consciously – if not necessarily most often – and with due acknowledgement took its vocabulary of forms and images. This was so, in part, because the Byzantine world, more than any other, carefully nurtured the great inheritance it had assumed from Antiquity. It was so also because Byzantium was the one world Early Islam most wanted, and failed, to conquer. But, above all, it was so because, to the Islamic and especially to the Arab Middle Ages – as well as to the Christian and especially to the Carolingian West – Byzantium, even at its lowest and weakest moments, partook of that mysterious aura which at given periods of history has endowed certain cultures and countries with a prestige of artistic genius which, deservedly or not, they alone at that period possessed. Therein, more than in any precise artistic motif, lies what the late Louis Massignon, in an eloquent article,51 has called the Byzantine mirage in the Arab mirror.

51

L. Massignon, “Le mirage byzantin dans le miroir baghdadien d’il y a mille ans,” Annuaire de l’Institut de Philologie et d’Histoire Orientale, 10 (1950).

Chapter II Islam and Iconoclasm*

The most obvious difference between Byzantine and Islamic iconoclasm is that the former is usually spelled with a capital “I” and the latter with a small “i”. This secondary typographical distinction illustrates first of all the difference between a historical moment (these are presumably capitalized) and an attitude or mode of behavior, the latter being apparently too common to deserve capitalization. In Byzantium the historicity of the events which are called the Iconoclastic period is defined through very specific dates: a crisis involving the relationship of worshippers and religious images in the seventh century; a succession of edicts and debates starting around 726 and remarkable for their verbal wealth, intellectual content and occasional violence; and finally in 843 a final Restoration of Images, once again capitalized by becoming a Feast of the Church. There is nothing comparable in the Muslim world. There are no internally decided edicts; the Qur’an is totally silent on images except insofar as they were used as idols, which are most forcefully condemned; to the extent that there was a debate, it was hardly a significant one, and our evidence for it is far more tangible in the twelfth and thirteenth centuries or, even more so, in the twentieth, than in the seventh or eighth. Forceful destruction of images is usually quite late, the action of brutal conquerors like the Ghaznavids in India or Nadir Shah attacking the Buddhas of Bamiyan with his artillery, and almost always directed against nonMuslim monuments, with a few exceptions of late defacing of miniatures by pious librarians. While there are instances of technical destruction of representations, as when Mahmud of Ghazna asked his son to remove erotic * First published in Iconoclasm, A. Bryer and J. Herrin, eds (Birmingham, 1977), pp. 45–52. This lecture was left essentially in the shape in which it was delivered at the symposium with two modifications other than the elimination of most obvious features of oral presentation. I have simplified the second half of the paper, since many of its points have been reworked for a meeting of German Orientalists in Freiburg (September 1975) and it was published in the proceedings of the meeting. “Das Ornament in der islamischen Kunst,” Zeitschrift der Deutschen Morgenlandischen Gesellschaft, suppl. III, I (1977), XLI–LIV. I have also eliminated all illustrations as they consisted almost exclusively of well-known and often-illustrated monuments; for the same reason bibliographical notes have been limited to references which are immediately pertinent to the text.

43

44

early islamic art, 650–1100

paintings from his palace,1 such instances are few, and on the whole, literary references to the use of representations tend to take the form of titillated disapproval, not unlike our own cultural, if not individual, reaction to pornography. There does not seem to be in fact any real parallelism between the Byzantine phenomenon and a Muslim attitude, even though it is possible that the adjective “iconoclastic” can justifiably be attributed to both. Why then concern oneself with the Muslim world when dealing with Byzantine Iconoclasm? Beyond the theoretical possibility that a general phenomenon can be of use in explaining a specific one, the most important reason is the coincidence in time between Byzantine Iconoclasm and the rise of Islam. Several decades ago many a scholar had seen the Christian phenomenon as directly influenced by the new faith and culture, and a number of rather troubling arguments do exist to support such a contention. There is the celebrated edict of Yazid in 721 which is supposed to have led to the destruction of Christian images.2 It is the most clearly focused of a group of alleged instances of forceful removals of Christian images from religious buildings, and several examples of replacements of representations by vegetal or geometric ornament seem prima facie to demonstrate the actual implementation of such edicts. The Ma’in mosaic is the most celebrated instance but there are others, curiously concentrated in Jordan and still unpublished.3 Then it is possible to relate the well-known changes in the coinage of Justinian II, both iconographic and epigraphical ones, to the cold and at times hot war between the first Umayyad caliphs and their Byzantine counterparts.4 And finally, the mosaics of the Dome of the Rock, of Bethlehem’s Church of the Nativity, of the Aqsa mosque (redone in Fatimid times but certainly reflecting Umayyad models), perhaps even those of the mosque of Damascus – all following each other over twenty years at most – can be construed to reflect, in their unusual sequence, a series of statements about Christian or Muslim, imperial or caliphal, power, for which it is possible to propose the conscious will of formal assertions directed to the competitor or the enemy rather than to one’s own culture.5 However these 1 2 3 4 5

Quoted by T. A. Arnold, Painting in Islam, repr. edn (New York, 1965), p. 86. Most texts gathered by Vasiliev, “The Edict of Yazid,” Dumbarton Oaks Papers, 9–10 (1956). R. de Vaux, “Une mosaïque byzantine à Ma’in,” Revue Biblique, 47 (1938). Grabar, Iconoclasme, p. 67 ff.; J. Breckenridge, The Numismatic Iconography of Justinian II (New York, 1959). For the Dome of the Rock see K. A. C. Creswell, Early Muslim Architecture, 2nd edn (Oxford, 1969) and O. Grabar, “The Umayyad Dome of the Rock,” Ars Orientalis, 3 (1957); for the mosque of Damascus, Creswell and now B. Finster, “Die Mosaiken der Umayyadenmoschee,” Kunst des Orients, 7 (1970–71), where the arguments of R. Ettinghausen, K. Otto-Dorn and others are mentioned; for the basilica of the Nativity at Bethlehem I am following H. Stern, “Les Représentations des conciles,” Byzantion, 11 and 13 (1936 and 1938), and Grabar, Iconoclasme, p. 50 ff. rather than R. W. Hamilton,

islam and iconoclasm

45

examples of wall mosaics must in fact be interpreted, it seems hardly [46] accidental that the Muslim statements reject all representations of anything living and that the Christian answer in Bethlehem is equally devoid of the usual personages of Byzantine imagery, just a few years before the formal opening of the Iconoclastic period. It seems altogether possible to hypothesize for the Byzantine phenomenon an Islamic influence or impact and it could be imagined that a deeper investigation into the few documents we possess would clinch the argument. But, even if one could delve more profoundly into the matter – and I for one am far from sure that it is really possible – it is hardly likely that the argument would be clinched, for there is no real need for Islam to explain Byzantium. The edict of Yazid is a very suspicious document, curiously absent from early Muslim sources and appearing in the Christian ones as an anti-Muslim and anti-Jewish document of doubtful historical importance, and, in my judgement, even validity. Floor mosaics may have been redone but is it necessary to make Muslim power responsible for the changes rather than local Christian practices, perhaps the impact of Monophysites or changes in ecclesiastical allegiances? The changes in coinage, be they iconographic like the sudden emphasis on Christ, or verbal like the servus Christi so close to the ‘Abd Allah of the Muslims, can easily be explained in purely internal Christian terms.6 And the mosaics in official buildings all make sense internally, the Dome of the Rock, Aqsa and Damascus as statements of some significance to Muslims, and the Nativity mosaics for Christians. At best, one could argue that the lack of representations in the latter was a concession to a prevailing mood, an expression of taste rather than a conscious and formal rejection of representations. It is not going to be my purpose to pursue the investigation of these wellworn questions, for in the absence of new evidence or of a new perspective, arguments become circular and interpretations more persuasive by the eloquence of their expression than by their intrinsic value. It is even possible that stylistic changes and iconographic purposes within the Muslim world are not pertinent to the Christian phenomenon, and in a wider sense, that the two empires were at that time irrelevant to each other artistically, if not politically. At best the Muslim world served as a political haven for actually persecuted or ideologically threatened Christians, but it is curious to note that from John of Damascus to Theodore and Theophanes, two monks from Palestine who were punished under Theophilos and one of whom became bishop of Nicaea after the restoration of Orthodoxy, it is Christian Iconodules rather than Iconoclasts who resided within the Muslim world. The examples of John of Damascus or of the saintly brothers from Palestine a century later illustrate a medieval version of a well-documented

6

The Church of the Nativity (Jerusalem, 1947), pp. 59–69; for the Aqsa mosque see H. Stern, “Recherches sur la mosque al-Aqsa,” Ars Orientalis, 5 (1963). O. Grabar, The Formation of Islamic Art (New Haven, 1973), pp. 89, 97.

46

early islamic art, 650–1100

phenomenon of the nineteenth century or of our own time: the possibility of living in peace within one culture while fighting battles, mostly verbal ones, within another. There is no evidence known to me during the Iconoclastic century (as opposed to the end of the seventh century) that either empire utilized such groups for political purposes, although one can wonder whether the striking Muslim military successes between 813 and 840 may not have been helped by Iconoclastic populations. A more interesting point may be that the Iconoclastic period, and most particularly its second phase in the ninth century, was a time of considerable cultural contacts between Greek and Arab Muslim worlds. It was the time of the largest group of translations of science and philosophy into Arabic and it has long been recognized that the Emperor Theophilos was strongly influenced by the themes and possibly styles of the ‘Abbasids.7 His pavilions bearing names like Love, Pearl or Harmony, as well as his gardens filled with spectacular automatic machines, all find parallels in the art of the caliphs in Baghdad. The ruins of Küçükyalı, identified with some likelihood as those of Theophilos’ Bryas, bear some resemblance to earlier or contemporary Muslim monuments.8 Regardless of its other aspects, the later Iconoclastic period, with its secular concerns, seems to be a reversal of the phenemenon of the late seventh and early eighth centuries. Instead of Muslim patrons struggling with Christian Byzantine art, we find the emperors of Constantinople fascinated with the forms developed in Baghdad and laying the ground for the partial Orientalization of Macedonian taste.9 The actual nature of these forms transmitted from one culture to the other is difficult to illustrate in the ninth century, for both in Islam and in Byzantium, a number of key documents are missing and will probably never be recovered. There are also problems of dating individual monuments, which complicate matters even further. Regardless of the solutions to be given to these, a possible conclusion may be that, instead of thinking about Iconoclasm and Islam in terms of relatable and comparable or irrelevant and separate ideologies and attitudes, we should rather investigate questions of taste within a period of time in which each entity developed its own modes of behavior and forms; emphasize, in other words, a system of cultural contacts and of mutual or contradictory [47] concerns rather than seek out a single explanation for diverse phenomena. One should concentrate on the Muslim search for Greek books to be translated into Arabic and on the Orientalization of Christian taste, not only in Constantinople but also in the provinces, most particularly in Armenia and northern Mesopotamia, those extraordinary regions which still await coherent cultural and artistic 7 8 9

Grabar, Iconoclasme, p. 169 ff. S. Eyice, “Contributions à l’histoire de l’art byzantin,” Cahiers Archéologiques, 10 (1951). A. Grabar, “Le Succès des arts orientaux,” Münchner Jahrbuch der bildenden Kunst, 2 (1951).

islam and iconoclasm

47

investigation. Within this perspective, the religious and intellectual quarrel recedes somewhat in importance, except in the relatively narrow area of specifically Byzantine religious art. What emerges instead as particularly significant is a period of shifting cultural currents, and the church of Aghtamar datable between 915 and 921 as well as Gagik’s palace described by the Armenian historian Artzrumi are the most remarkable results of the formal and ideological currents of the Iconoclastic period.10 These preliminary remarks can be summed up in the following manner. Just before the Iconoclastic controversy in Byzantium, a group of Islamic monuments utilized primarily Byzantine forms but avoided representations of living beings, even though there is no intellectual or doctrinal argument for such an abandonment of living things in what we know of Islamic thought at the time, and the roughly contemporary Byzantine monuments and symbolic images like coins do not require the explanation of an external impact, at least not on their “iconoclastic” level. A long century later, Byzantium established a reasonably rational and formalized definition of images, but in the meantime there had occurred a shift in the relationship between the new Muslim culture and the old Byzantine one and it has been argued that the changes introduced by Iconoclasm – but not necessarily Iconoclasm itself – were at least in part responsible for the new Byzantine receptivity to eastern things. In both instances, the early eighth century or the middle of the ninth, two concurrent questions are raised. One is historical and requires a precise reconstruction of the kinds of events and social, political or cultural attitudes which would make it reasonable, if not always necessary, to explain an impact of Byzantium on Islam or vice versa. The second question is one of artistic ideology: what made certain forms and themes rather than others pass from one culture to the other? Is it an aspect of the “giving” culture or of the “receiving” one? Is it both? Can such aspects be defined? Or should one tend to interpret similarities in terms of some vague common Zeitgeist? Since a full investigation of all four of these questions – Byzantine and Muslim history and ideologies c. 680–730 and c. 820–50 – is clearly beyond the possibilities of a single paper as well as way beyond my competence, I should like to limit my remarks to one major question, which is the nature and meaning of what has been called Islamic iconoclasm, thus trying to answer the subsidiary one of whether or not the nature of a Muslim attitude to the arts is of any real pertinence to the Byzantine phenomenon. The historical setting in the Muslim world of the years 680–720 seems reasonably clear, and I only want to summarize conclusions reached elsewhere on fuller evidence.11 My argument is essentially that the early Muslims were indifferent to images, that in a group of coins they sought to create an 10 11

S. Der Nersessian, Aght’amar (Cambridge, 1965). Grabar, Formation, p. 75 ff.

48

early islamic art, 650–1100

imagery and a symbolism of their own,12 and that this effort failed for many reasons, but primarily because they could not create a meaningful and effective imagery without becoming like the Christians. The result appears in a succession of major creations – the Dome of the Rock, Abd al-Malik’s new coinage, the mosque of Damascus – which rejected representations in official art, emphasized writing, and sublimated symbolic meanings of whatever sort into what appears to us now as being primarily ornament. This rejection was not carried down to private art, as is amply demonstrated by the so-called “desert palaces” of the Umayyads or by the more hidden parts of Samarra in the ninth century. But in the public art of the mosque or of coinage, it is only exceptionally that representations occur, at least until the eleventh or twelfth centuries. A rational and intellectual justification for these changes did not occur on a doctrinal level until later. Its earliest traces seem to be in polemical writing of the middle of the eighth century, although I have often wondered whether the celebrated text of Theodore Abu Qurra which has so frequently been used to indicate this particular date really does have all the implications attributed to it.13 It is possible that a careful study of early traditions and of legal literature will bring out some additional information on the elaboration of the doctrine within Islam itself, but these are notoriously difficult sources to use, especially when one seeks to date them properly, and their serious investigation has barely begun.14 But, even if more information is found, I rather doubt that it will alter in any [48] significant way the conclusion that there is nothing within early Islamic thought comparable to the late antique and especially medieval Christian concern with formulating a precise relationship between representation and the represented. The paradox seems then to be that, for immanent cultural, psychological and political considerations of a very specific time, official Islam succeeded in doing, apparently painlessly, bloodlessly, and without rational justification, what required, at almost the same time in Byzantium, endless discussions, often of a very high level of sophistication, and a violence hardly seen before or after, around what we would consider as works of art. Furthermore, Byzantine Iconoclasm was eventually rescinded, while an absence of representations in all aesthetic or symbolic creativity pertaining to the faith and a reluctance to images in other arts became accepted characteristics of Islamic art. Without a capitalized Iconoclasm, iconoclasm with a small “i” became apparently the norm for an enormous culture extending from Spain to India and the inheritor of practically all the forms which had grown in 12 13 14

The latest statement on the evidence, with full bibliography, is found in G. C. Miles, “The Earliest Arab Gold Coinage,” ANSMN, 13 (1967). Creswell, Early Muslim Architecture, pp. 411–12. R. Paret, “Textbelege zum islamischen Bilderverbot,” Studien H. Schrade dargebracht (Stuttgart, 1960). A statement on the evidence, with full bibliography, is found in Miles, “The Earliest Arab Gold Coinage,” ANSMN, 13 (1967).

islam and iconoclasm

49

the Mediterranean and the Middle East. How was it possible? And, if an explanation can be provided, does it in any way help in understanding Byzantine Iconoclasm? Let me begin with a parenthetical remark dealing with an important problem of formal interpretation of monuments and affecting the hypothesis I will propose later. While it is true that calligraphy and geometric or vegetal ornament predominated in the visual tradition created by Islamic culture, these two modes did not exclude representations of living beings whose analysis brings out frequently enough complex meanings. Umayyad palaces like Qusayr ‘Amrah and Khirbat al-Mafjar are obvious examples which are reasonably well known, but the ‘Abbasids, the Fatimids, and almost every secondary dynasty as well as the non-dynastic substructure of Islamic civilization sponsored and utilized figural art, with the exception so far of North African Berber dynasties.15 After the middle of the twelfth century, there occurred a true explosion of such images which continued in India, Iran and the Ottoman empires, although a taste for and interest in representations disappeared almost entirely in the Arab world c. 1350 onward.16 Under these circumstances, is it really correct to talk of an Islamic iconoclasm or even of an avoidance of representations of living things? And should one not limit investigations to those aspects of the arts which are directly and exclusively related to the faith? Maybe so, but it is important to note that, in scale and in formal character, Islamic representations are, with a few exceptions, of a different order than figures in classical, Byzantine, or western medieval arts after the early Romanesque period. Three examples may suffice to make my point. In the Cairo Museum there is a celebrated group of wooden beams which presumably came from the Fatimid palace and which are usually dated in the latter part of the eleventh century.17 They are elaborately carved and contain a simple geometric progression of cartouches with a standard imagery of royal pastime. In the twelfth-century ceiling of the Cappella Palatina in Palermo, almost every facet of an elaborate muqarnas ceiling contains images with an unusually wide variety of subjects.18 The point, however, is that in neither case could the figures be easily seen. Obviously their point as decoration, as a manner of transforming a high ceiling or wooden beams far above the eyes of the users of whatever room contained them, overshadowed, not to say obliterated,

15

16 17 18

There is no coherent history of Islamic iconography. A general introduction may be found in R. Ettinghausen, Arab Painting (Geneva, 1962) and specific examples are discussed by many studies by Ettinghausen, the latest one being From Byzantium to Sassanian Iran (Leiden, 1973). See also various studies by E. Baer. O. Grabar, “Les Arts Mineurs de l’Orient Musulman,” Cahiers de Civilisation Médiévale, 11 (1968). G. Marçais, “Les figures d’hommes et de bêtes,” Mélanges Maspéro, III (Cairo, 1940). U. Monneret de Villard, Le Pitture Musulmane al Soffitto della Capella Palatina (Rome, 1950).

50

early islamic art, 650–1100

their immediate visual impact as figures belonging to a variety of otherwise definable iconographic cycles. My third example is that of one of the celebrated group of Spanish ivories of the last decades of the tenth century. Their quality has long been recognized and, although some of the subjects of their decoration are not entirely clear, there is a general agreement that it included principally typical scenes of princely life.19 These ivories differ from the Cairo woodwork or the Palermo paintings in that they were meant to be handled directly and their representations could not remain unseen or barely imagined. Yet one of their stylistic characteristics is that there is no hierarchical distinction between medallion frames, vegetal ornament and personages. They are all shown on the same plane or planes, and subjects with the same iconographic meaning are found inside and outside medallions. It is as though they were meant to be equal in the visual effect they produced. This rather curious egalitarianism in the use of forms and subjects lends itself to several interpretations, one of which is ambiguity or uncertainty about the meaning to be given to images. Similar ambiguities occur in the eighthcentury sculpture at Khirbat al-Mafjar,20 in Nishapur pottery of the ninth or tenth centuries,21 and, in a different way, in later Persian miniatures. [49] The first point of these examples is that even if it is incorrect to imagine Islamic art as devoid of representations, it probably is correct to say that they played in it a different role, perhaps best defined at this stage as undifferentiated from other motifs, as though part of a “total” system of designs in which not a single category, except writing, was unduly emphasized. But there is a second point to these examples which is more important to my subject. It is that a culture and an art which picked up its visual vocabulary in the richest repertoire of representations known in history managed to use these forms in a manner least suited to the models and kept up for centuries, with only occasional disruptions after the twelfth century, a use of representations which seemed at first glance to be defeating its subjects. In other words, the central tradition of Islamic art always downplayed representations of living things and, even though the contrast between the right side of the Mshatta façade which is in front of the mosque and contains no animals and the left side which does have animals and human heads makes the point in a rather crude manner,22 the visual contrast between the two sides is one of degree rather than of kind, and there is something disingenuous in doing what so many of us have done: to excerpt a representation of man, however deeply hidden it may be, and to give it the importance only of a design. 19 20 21 22

See now E. Kühnel, Die Elfenbeiskulpturen (Berlin, 1969), esp. pls XVIII and XXII. R. W. Hamilton, Khirbat al-Mafjar (Oxford, 1957), pls XXXV.1 or LIII.2. E. Atil, Ceramics from the Islamic World (Washington, 1973), pl. 12. Creswell, Early Muslim Architecture, pls 119 and ff.

islam and iconoclasm

51

To propose an explanation for this phenomenon we must turn to what may be called considerations of social ethics. For it is probably neither artists and artisans nor legal scholars and theologians nor even caliphs and lesser princes who determined a Muslim artistic attitude, but rather that peculiarly unique social phenomenon which has still evaded coherent definition by social scientists, the corporate ummah. Centered primarily in urban settings, its components were related to each other by a complicated set of bonds, at times formalized as through the workings of the shari’ah, the legal system, at other times more informal, as the complex of ties which has been called the asabiyah.23 It is this body which created the norms around which the visual creativity of the culture was developed and it is rather remarkable that only one scholar, the late Marshall Hodgson, has tried in a brilliant article published over ten years ago to disentangle the components of this social ethic as it applied to the arts.24 Starting with his rather abstract analysis, I shall try to develop a slightly different interpretation of a Muslim attitude toward the arts and suggest a way in which it may be, at least conceptually, of some value in understanding the Byzantine phenomenon. I will only deal with monuments and ideas which seem to me valid for a period earlier than the middle of the twelfth century, for, in spite of a generally accepted notion of a monolithic Islamic position, so many things changed in the latter part of the twelfth century that generalizations applicable to earlier times simply do not work, as mysticism together with many ethnic and regional modifications require entirely different explanations. Influenced in many ways by the formally expressed or subjacent views of various forms of Protestantism, most specifically early Calvinism and the Quakers, Hodgson identifies four key aspects of a Muslim social ethos: populism in the sense that a working urban bourgeoisie established itself, at least until the thirteenth century, as the main taste-maker in opposition to the court of princes; moralism in the sense that an ethical justification, usually in legal terms, must be given to any action or form of behavior; factualism in the sense that there is no real mystery in the universe, and since God alone is Power and Reality, there is no need for the obvious substitution for physical reality which characterizes any representation; and, finally, historicalism, a term which strikes me as less clear than the previous one but which I interpret to mean that no event or no life (with only a very partial exception for the Prophet) could acquire the kind of extra-temporal or parachronic value which has led to the Christian or Buddhist attempts to represent the intelligible. Although an Islamicist may quibble with some aspects of this rough definition of a culture, it seems to me at least that, on the whole, it succeeds quite adequately in identifying the key features of early Sunni Islam. It implies a distrust if not an outright rejection of two essential 23 24

See Marshall Hodgson, The Venture of Islam, esp. vol. 1 (Chicago, 1974). M. Hodgson, “Islam and Image,” History of Religions, 3 (1964).

52

early islamic art, 650–1100

aspects of artistic creativity. It denies the possibility or the value of escaping time through the freezing of a moment or of an idea; this objection is particularly valid for the visual arts, for poetry or music were precisely arts which are intimately bound to the time of recitation or of performance. And, secondly, Hodgson’s model implies a rejection or at the very least a disapproval of expensive possessions, and, in our fascination for the public art of churches, it is essential to remember that the arts most commonly known or seen were privately owned or used objects, ranging from simple ceramics to fancy textiles and gold cups. The rejection is not of art per se, nor even of representations, but, as Ibn Miskawayh, a typical honnête homme of the tenth century, put it, of becoming attached to something beautiful [50] and expensive.25 The cardinal sin is not one of artistic creativity but of greed and of temptation away from divine truth. Up to this point, Hodgson’s analysis, which I have only modified, I trust, on minor details, seems reasonable enough, for it does provide an explanation for the striking practicality of so much of Islamic art of early centuries: its emphasis on public architecture with social functions, its development of an elaborate art of textiles and of ceramics. It also explains why so much of the fancier arts is identified with royal or princely luxury: the palaces of the Umayyads with their exuberant sculptures and paintings, the silver objects continuing Sasanian practices, so many themes of the Cappella Palatina. This was precisely the stuff which was morally and socially wrong and it is curious indeed, especially when compared to Christian historiography, that so little information about all these royal orgies of forms is given in the literature written by and for the learned bourgeoisie. But Hodgson and many other writers may have overstated their point when they deduce from their analyses of a Muslim ethos that there was no Islamic visual symbolism and that there was no concern for the intellectual problems of images and of representations. I should like to suggest, on the contrary, that Islamic culture did in fact develop both a visual symbolic system and a rather complex, if limited, argument about representation. The key point about symbolism and the reason why it is not usually detected is that it rarely took the form expected in the classical tradition of a one-to-one relationship between a visual motif and an external subject. The celebrated third style of Samarra stucco decoration from the ninth century is usually interpreted as a fascinating ornamental achievement of utilizing the full surface of a wall in the creation of an endlessly repetitive pattern which can only be defined in abstract terms: that is to say, not in terms of separate motifs which can be excerpted from the design but as a set of relationships of symmetry, light and shade, and so on.26 The Kalayan minaret in Bukhara is a striking composition of brick combinations in which each stage is a 25 26

Ibn Miskawayh, Traits d’Ethique, tr. M. Arkoun (Damascus, 1969), pp. 301–3. Grabar, Formation, p. 198 ff.

islam and iconoclasm 53

different aspect of the same basic diamond motif.27 And the surface of a celebrated plate from Nishapur in the Freer Gallery contains an inscription with a proverb and five floral designs in the center.28 All these examples can be interpreted as visual interpretations of medieval Islamic intellectual currents. The abstraction of the Samarra design can be seen as a visual translation of the philosophical, legalistic and ethical concerns of the ninth century which attempted to discover the basic, abstract principles of thought in such a fashion that no aspect of reality would be immune from them, just as no part of the decorated wall escapes from a meaning in the design. The Nishapur plate or the Kalayan minaret are not really examples of a horror vacui but of the fact that everything has meaning and value. And in the instance of the minaret, as in several other examples more particular to Iran, it may even be suggested that the expression of the same motif in several different appearances is a translation through elaborate geometry of a profoundly Islamic notion of one God with many equivalent names. What gives some credence to these hypothetical explanations is, first of all, that symbolic meanings through inscriptions pervaded so much of Islamic art,29 and secondly, that there is such striking correspondence in time between the visual phenomena and scientific, legal and philosophical interests, just as the mystical developments of later times will, together with other reasons, lead to a reappearance of images and of concrete symbols. And then there is a logical argument which is that the absence of a developed visual symbolism is simply an unlikely phenomenon for a huge world over several centuries. Now in all these examples there are no representations of living beings, men or animals. The reason, I submit, is only secondarily the negative one of rejecting actual or potential idols. It is rather that the Muslim world had not quite managed to come to grips with the key problem of visual reality. Or rather it gave to this problem an unusual and contradictory solution. Two examples will serve to make my point. One is the celebrated text of Maqrizi describing the competition around 1058 between two painters, Qusayr and Ibn Aziz. One claimed to be able to represent a personage so that it can be seen in relief, while the other one said that his personage will seem to be imbedded in the wall. Both succeeded by an artful combination of colors (in one case white on black, in the other red on yellow), according to the account. But preceding this story is another one, less frequently quoted, which relates that in a mosque of Cairo there was a representation of a fountain with steps leading to it and that the combination of colors was such as to give, from one particular spot, the typical trompe-l’œil feeling of an actual staircase.30 Both stories suggest [51] an illusionist perfection which 27 28 29 30

D. Hill and O. Grabar, Islamic Architecture and its Decoration (London, 1967), figs 10–11. Quoted in Atil, Ceramics from the Islamic World. E. Dodd, “The Image of the Word,” Berytus, 18 (1969). Maqrizi, Khitat (Bulaq, 1869), II, 318.

54

early islamic art, 650–1100

hardly accords with the works known from that time. The explanation is that the eye of the medieval viewer in the Muslim world interpreted automatically the simplest outlines of what it saw as an illusion of reality, because there is no way of interpreting something one knows otherwise except as potentially real. My second example comes from Ghazali’s Alchemy of Happiness written around 1106 and discovered in a very rarely mentioned article by Ettinghausen.31 His main emphasis is on the pleasure of that which is beautiful and he writes: the beauty of a thing lies in the appearance of that perfection which is realizable and in accord with its nature. When all possible traits of perfection appear in an object, it represents the highest stage of beauty; when only part of them occur, it has that measure of beauty which appears in the realized degree of perfection.

But then, in ways which recall Plotinus, Ghazali adds that there is a difference of degree existing between those things which belong to the outer world and to the inner world: “there is a great difference between him who loves the painted picture on the wall on account of the beauty of its outer form and him who loves a prophet on account of the beauty of his inner form.” The point of these passages is that, however one enjoys it, one can only represent that which one knows, a doctrine strikingly akin to that of socialist realism. The further point, however, is that the truth or reality of anything alive does not lie in its shape, its physical character, which is but an accident, but in its inner worthiness, which is only hidden by the accidental shape. What these two examples suggest is not at all iconoclasm, but beyond a truly sensuous feeling for pictorial beauty, the ultimate impossibility of representations of living things. It is of course true that out of these two examples of the eleventh century it is probably not proper to define a doctrine, and clearly a search for additional texts, not in juridical and religious treatises where scholarship has tended to look for them but in philosophical and ethical ones, is a needed task for research. Yet, when related to my earlier example from Ibn Miskawayh and to the monuments I have mentioned they suggest a pattern for the Muslim attitude to images. First of all, it is not proper to refer to it as iconoclasm. At best aniconism is perhaps acceptable, for its key point is that it saw images not as evil per se (only man can do evil with them) but as irrelevant since unable to capture reality, and at worst temptations away from the requirements of a good life. This position was not achieved immediately and is not pertinent to the monuments of c. 700. They are the result of a specific inner growth of the Muslim world and not necessarily influenced by other cultures and 31

R. Ettinghausen, “Al-Ghazzali on Beauty,” Islamic Art and Archaeology: collected papers, ed. Myriam Rosen-Ayalon, with an introduction by Oleg Grabar (Berlin: Gebr. Mann Verlag, 1984), pp. 16–21.

islam and iconoclasm

55

developments. But aniconism did not mean absence of symbols or a negative rejection of representations, at least not in early times. It rather meant the elevation of other visual forms, writing, vegetal ornament, geometry, abstract pattern, possibly color, to the level of meaningful forms, around which the culture developed its own systems of association whose study is still in its infancy. From these remarks it could be concluded that an understanding of Islamic attitudes has nothing to do with Byzantine Iconoclasm, except perhaps to the extent that the phenomenon of Islam is pertinent to explain the Byzantine crisis. Such has been the conclusion of many scholars in recent years, for instance that of Grünebaum, among others.32 At best one can simply argue, as many have done, that Islamic monuments serve as examples of styles and motifs which would have disappeared from Byzantium because of Iconoclastic destructions. But there are perhaps two areas in which an awareness of the Muslim phenomenon is quite pertinent to the Byzantine one. One is technical. The fact that the Muslim attitude developed as it did was one of the reasons why Islamic art could have an impact on other arts, especially in Christendom, whereas the reverse was no longer possible. For the very qualities of its abstraction and purely interiorized symbolism made it adaptable to other tastes and other purposes, and even its writing could be copied with impunity if one did not become aware of its meaning.33 It transformed Islamic forms into a uniquely intercultural system of forms in the Middle Ages and even in the Renaissance. And, when its princely cycles were copied, they did not represent so much Islamic art as the Near Eastern version of the art of all princes. The uniqueness of the Muslim ideological development may therefore explain its impact on Byzantium in the ninth century. But the second aspect of the Muslim attitude is perhaps more important, although more difficult to focus properly. One can argue that Islam elaborated a mode of creativity which illustrates far more than itself. It is in fact a manner of “seeing” and of “showing” which tended to refuse the hierarchy of subjects imposed by Antiquity in which the representation of man [52] occupied the first place. Thereby, like early Buddhist art from which the Buddha is absent, it avoided the dilemma of Christian art which was that God can only be represented in human form but how can one make a human form to be understood as God? The choice existed in Byzantium as well, just as it existed later in the West, and the problem becomes less why Iconoclasm with a capital “I” began than why it did not succeed, for the questions it posed were universal ones rather than localized ones in time and 32 33

G. von Grünebaum, “Byzantine Iconoclasm and the influence of the Islamic Environment,” History of Religions, 2 (1962). Much remains to be done in this area: see R. Ettinghausen’s contribution to S. Schacht and J. Bosworth, eds, The Legacy of Islam (London, 1972).

56

early islamic art, 650–1100

space. To this question also the Muslim phenomenon may provide an answer, or at least the elements of an answer. It is that the starting point for an explanation of a phenomenon like the destruction or avoidance of images does not lie in immanent formal changes or in specific philosophical positions, but rather in the coexistence, at all times and in all places, of two equally potent attitudes toward the visual world: one which seeks the relationship between the thing made and its subjects or functions, and the other one which seeks to emphasize the relationship between works of art and their viewers and users. The Muslim world, for historical reasons of its own, chose to develop primarily the latter, while being occasionally tempted by the former. Christianity adopted the former and was compelled to struggle with the complex of Pygmalion, of images which become alive. Islam, by leaving the choice of interpreting visual creativity to men rather than by creating a doctrine of visual interpretation, may have succeeded in setting the nature of the arts in a totally different dimension.

Chapter III Upon Reading al-Azraqi*

It is altogether curious that so little scholarship has been devoted to the physical features of Mecca in early Islamic times. Except for the informative and detailed entries in the two editions of the Encyclopaedia of Islam,1 most writing on Mecca has been concerned either with its socioeconomic and religious–cultural setting at the time of the Revelation to Muhammad or, even more often, with the pilgrimage, its complex liturgical practices, its concomitant economic and other practical problems, and, especially in more recent times, the powerfully moving emotional and spiritual experiences of the faithful on this holiest of journeys.2 Yet, however fascinating and emotionally charged the practices and symbolic associations of the pilgrimage might be, they are only one part of the impact the holy city has on the pilgrim. The Ka‘ba, the Masjid alHaram, or Sacred Mosque, surrounding it, and the whole city of Mecca are today, as they were in the past, part of a common visual memory of the Muslim community, even if colored by the emotional make-up and sensitivities of each particular individual. Every Muslim has in common an awareness of its forms and spatial compositions. Since the beginning of the twentieth century, photographs and films have been available to serve as reminders of the holy city and of the events taking place in it. In earlier centuries people relied on printed, stenciled, drawn and painted pictures rendered on paper, cloth, tile, stone, or any other available material for the images of Mecca that became the souvenirs or mementos that were the permanent signs of a believer’s association with the city of the qibla.3 * First published in Muqarnas, 3 (1986), pp. 1–7. 1 In the second edition the key article is “Ka‘ba” by J. Jomier and A. Wensinck, but the article “Mekka” in the first edition contains considerable additional information. 2 The fundamental study is still that of M. Gaudefroy-Demombynes, Le Pèlerinage à la Mekke (Paris, 1923); classic accounts of pre-contemporary times are C. Snouck Hurgronje, Het Mekkaansche Fest (Leiden, 1880), Muhammad L. al-Batanuni, al-Rihla al-Hijaziya (Cairo, 1329H), and Rif‘at Pasha, Mir’at al-Haramayn (Cairo, 1925). A particularly striking contemporary Muslim account is Abdel Magid Turki and Hadj Rabah Souami, Récits de pèlerinage à la Mekke (Paris, 1979); Malcolm X’s experience of brotherhood during the pilgrimage is one of the most moving passages of his autobiography. It would be most desirable to have a good historical anthology of pilgrimage accounts.

57

58

early islamic art, 650–1100

These representations are, however, for the most part very conventional and stereotyped and, pending detailed investigation, contribute very little to an understanding of the holy city’s physical character and evolution. They are pious images, not historical documents, and reflect a standard, toponymically accurate but visually simplified vision of a rectangle with the places of commemorative or liturgical importance clearly marked. The contrast between them and the complicated and frequently difficult to interpret sixth-century Byzantine phials from Jerusalem is striking.4 Furthermore, the Muslim images all belong to a time when the sanctuary had acquired more or less the shape it would keep until the momentous and irreversible transformations of recent decades, a shape fixed in its major features by the end of the ninth century. Even the principal monuments punctuating the holy place had been built by the fifteenth century, and such stylistic variations as occurred during reconstructions were rarely, if ever, recorded on images. They were not meant to be descriptions of places, but evocations of holiness, and they do not provide any sense of the range of emotion or reaction the faithful experienced as they reached the sanctuary, nor do they express the complex memories carried away by pilgrims afterward. There is nothing in these depictions that is comparable in range to Ibn Jubayr’s rapturous but very precise description studded with qur’anic quotations,5 Ibn Battuta’s chatty but equally concrete account full of stories and minor human events,6 Ibn Khaldun’s perfunctory statement with a long reference to the letters he received and the important people he met in Mecca,7 or Ibn al-Arabi’s transformation of the holy place and of the pilgrimage into a stunning cosmological vision.8 But even these literary examples are relatively late (the earliest author, Ibn Jubayr, was born in 1145); they belong to essentially post-Fatimid centuries, when the Muslim world had fully developed a material culture of piety around Mecca and probably other religious sanctuaries as well.9

3

4 5 6 7 8 9

There has never been, to my knowledge, any attempt to gather together a corpus of representations of Mecca in Islamic art. The most common examples are on Ottoman tiles, K. Erdmann, “Ka‘bah-Fliesen,” Ars Orientalis, 3 (1959), pp. 192–7, with bibliography to that year. The earliest example known to me, dated 498 (1098), is a small basalt plaque in the Baghdad museum, which had the accession number ‘ayn 1149 when I saw it in 1956. To my knowledge it is not published, and my notes of over a quarter of a century ago indicate that the dating inscription may not be contemporary with the object. Vincenza Strita, “A Ka‘bah Picture in the Iraq Museum,” Sumer Journal, 32 (1976), pp. 195–201. André Grabar, Ampoules de Terre-Sainte (Paris, 1958); D. Barag and J. Wilkinson, “The Monza-Bobbio Flasks and the Holy Sepulchre,” Levant, 6 (1974). Ibn Jobair, Voyages, trans. M. Gaudefroy-Demombynes (Paris, 1949), pp. 93 ff. Ibn Battuta, Voyages, ed. C. Defremery and B. R. Sanguinetti (Paris, 1969), 1, pp. 303 ff. Ibn Khaldun, Le Voyage d’Occident et d’Orient, trans. A. Cheddadi (Paris, 1980), pp. 159 ff. See the lengthy commentary on the appropriate passages in the Futuhat by C. A. Gilis, La Doctrine initiatique du pèlerinage (Paris, 1982). Needless to say, this material culture of piety has not been investigated in any systematic way. See Janine Sourdel-Thomine, “Clefs et serrures de la Ka‘ba,” Revue des Etudes

upon reading al-azraqi

59

What happened to the Meccan sanctuary between the Prophet’s glorious and official return to it in 631 and the, properly speaking, medieval restructuring of the Muslim world from the eleventh century onward? Is it possible to imagine the attitudes of people during the formative centuries of Islam toward the shape, the physical form, of their holiest sanctuary? The investigation I am proposing, of which this essay is only a very preliminary step, has as its long-range objective an understanding of the interplay between specific building activities – the erection of a colonnade or of a [2] portico, the repaving of a court, the addition of some decoration – which can easily be documented through an inscription or a chronicle’s reference, and the practical, ideological, pietistic and symbolic motivations and explanations attached to these activities. This type of investigation may allow us little by little to develop a profile of the synchronic and diachronic mental attitudes of Muslims and of the relationship between those attitudes and architecture. Because of its overwhelming importance to Muslims, the Haram in Mecca can serve as an exemplar for this sort of investigation, and whatever hypotheses or conclusions can be reached for Mecca should apply to other holy buildings and places as well. There is no available archaeological record for the Haram, and none is likely to be forthcoming. We are therefore restricted to incidental references in chronicles, to the factual but, with a partial exception in the case of Maqdisi, remarkably sober descriptions in tenth-century geographies,10 and to the lengthy volume Kitab Akhbar Makka (“Book of Information on Mecca”) by Abu al-Walid Muhammad b. ‘Abdallah b. ‘Ali al-Azraqi.11 The book was put together before 865 by a native of Mecca claiming descent

10

11

Islamiques, 39 (1971), pp. 29 ff.; D. Sourdel and J. Sourdel-Thomine, “Nouveaux documents sur l’histoire religieuse et sociale de Damas,” ibid., 32 (1962), pp. 1–25. For other sanctuaries, also late, there is the extraordinary luster plaque in the Musée de la Céramique at Sèvres published by Chahriyar Adle, “Un diptyque de fondation en céramique lustrée,” Art et société dans le monde iranien, C. Adle, ed. (Paris, 1982), pp. 199–218. I am sure that a reexamination of ceramics, metal objects and possibly textiles will yield interesting results for a corpus of pious objects. On these remarkable writers in general, see André Miquel, La Géographie humaine du monde musulman, 3 vols (Paris, 1967–80). Specific passages are: Ibn Rusteh, Les Atours précieux (Cairo, 1955), pp. 21–62, with historical, topographical and pietistic notations, the latter taken from adab literature; Maqdisi, Ahsan al-Taqasim fi Ma‘rifat al-Aqalim (Descriptio Imperii Moslemici), ed. M. J. de Goeje (Leiden, 1906), pp. 71 ff.; Ibn Hauqal, Configuration de la terre, trans. J. H. Kramers and G. Wiet (Paris, 1964), 1, pp. 26 ff. I am leaving aside the issue of the relationship between geographers and litterateurs like al Jahiz, Ibn Qutayba, Ibn ‘Abd Rabbih, and others, who do occasionally mention the Arabian sanctuary, because they set it within a frame of reference which is less immediately involved with the specifics of the sanctuary’s form, but much more with its meaning in the consciousness of a man of culture. For the long-range objective of understanding the ethos of a period, however, these authors will also need investigation. Edited by F. Wüstenfeld as vol. 1 of the Geschichte der Stadt Mekka (Leipzig, 1858; photo reps, Beirut, n.d.).

60

early islamic art, 650–1100

from a Byzantine soldier who was taken prisoner during the Persian wars of the seventh century. It is primarily a text completed by a member of his family before 837, but, as preserved and edited, also includes references from as late as 922–3. It was probably revised by a pupil of al-Azraqi called alKhuza‘i.12 Although there is some evidence that works on Mecca and on the Ka‘ba were written earlier, including one by Wahb ibn Munabbih in the early eighth century,13 none has survived. Azraqi’s mid-ninth-century work is therefore not only the earliest extant book on Mecca, but the earliest preserved example of a book devoted to a single city. An added peculiarity is that it does not deal at all with the city’s notables, as most Arabic literature on cities does, but with its buildings and their history. By its very nature, therefore, it establishes that Mecca’s physical character and evolution had primacy over the people who lived in it. Several recent studies on early Muslim writers have begun to formulate methods for investigating written sources that will define the attitudes of their authors and explain the experiences and thoughts that lie behind a book’s content and structure.14 Such investigations can be of great value to the historian and interpreter of visual forms, but they also require complex and painstaking philological, linguistic, critical and historical inquiries that are far removed from the art historian’s or archaeologist’s concerns and competence. A brief description of al-Azraqi’s work will suggest the type of information that could be extracted from a literary and structural analysis of this kind. Structurally the book can be divided into four unequal parts: the first covers the Ka‘ba from the Creation to the Yemenis’ attempt to destroy the Ka‘ba late in the sixth century (pp. 1–84); the second, the “historical” Ka‘ba and the immediately surrounding holy spots (Maqam Ibrahim, Zemzem well) from the time of their reconstruction by the Quraysh before the Revelation to al-Azraqi’s time, with sections on the chronology, characteristics and liturgical or daily uses of the holy places (pp. 84–301); the third, the Masjid al-Haram, i.e., the open space which surrounds the Ka‘ba and which is entirely a Muslim creation (pp. 301–445); and the fourth, the living quarters of the city and a few miscellaneous items (pp. 445–505). Except for the last section, which is fairly straightforward and enumerative, each part consists of a large number of chapters, some as long as ten or fifteen pages, some as short as a paragraph or a few lines. Some are purely descriptive, either of a building or of a fragment of a building (e.g., the nails and the gutter 12 13 14

See Wüstenfeld’s introduction and Fuat Sezgin, Geschichte des arabischen Schrifttums, 1 (Leiden, 1967), pp. 339 ff. Azraqi, Akhbar, p. 9. Mas‘udi has been studied by Ahmad M. H. Shboul, Al-Mas‘udi and His World (London, 1979), and by T. Khalidi, Islamic Historiography (Binghampton, 1975). For Bayhaqi, see Marilyn R. Waldman, Toward a Theory of Historical Narrative (Columbus, 1980). For alJurjani, see Kemal Abu Deeb, Al-Jurjani’s Theory of Poetic Imagery (Warminster, 1979).

upon reading al-azraqi

61

of the Ka‘ba); some deal with an event (e.g., the Yemeni invasion) or with a sequence of events (e.g., the various pieces of cloth put on the Ka‘ba and the scents sprayed on it over several centuries), or with good or bad practices and obligations (e.g., on p. 316, the virtue of circumambulating the Ka‘ba in the rain). In other words, the book is neither an account in chronological sequence nor is it an orderly description of space. There is a constant interplay between specific moments, usually established quite precisely with names and dates when known, and equally specific places in the sanctuary. It is as though the understanding of something seen requires its connection with a historical or a mythical event, often drawn from the lives of Abraham or Hagar, which were connected with so many places in Mecca. The same events are repeated several times, and while a coherent chronology can be derived from al-Azraqi’s account, establishing a sequence of events does not seem to be its main point. Only a careful structural analysis of many passages15 in al-Azraqi’s book and their collation with other historical or religious sources would reveal whether he was using events to explain anomalous as well as regular features and practices, or whether he was seeking to connect sacred and, later, [3] human and dynastic history with the physical features of the sanctuary. Each of his chapters seems to answer a question: why are the nails of the Ka‘ba gilt? What is, or was, inside the Ka‘ba? Why is there a small dome over the Maqam Ibrahim? Can one catch one of the pigeons nestled in the courtyard? If this colonnade was built by order of the caliph al-Ma’mun, what was there earlier? Who is asking the questions, why he is doing so, and especially whether al-Azraqi was reflecting one, two, or more systematic interpretations of the holiest place in Islam are the questions we should ask, however. Once answers, even hypothetical ones, to them can be given, we will be closer than we now are to understanding early medieval Mecca through the eyes of those who lived there. Perhaps we can then better understand the architectural decisions made for its sanctuary. Because its subject is so closely tied to a particular architectural ensemble, al-Azraqi’s text is replete with architectural terms pertaining to a building’s construction as well as its appearance. It is, therefore, a prime source for the architectural vocabulary of the ninth century, at least the one prevalent in the Arabian peninsula. It probably reflects the high Arabic of the central lands of the Fertile Crescent, but not necessarily the technical language being developed, also in Arabic, in eastern Iran.16 Three passages can be used 15

16

I have in mind the type of analysis carried out by Kemal Abu Deeb, “Toward a Structural Analysis,” International Journal of Middle Eastern Studies, 6 (1975), pp. 148– 84, and sketched recently by A. Hamori, “The Comic Romance from the Thousand and One Nights,” Arabica, 30 (1983), pp. 38–56. Oleg Grabar and Renata Holod, “A Tenth-Century Source for Architecture,” Harvard Ukrainian Studies (Eucharisterion, Essays Presented to Omeljan Pritsak), 3–4 (1979–80), pp. 310–19.

62

early islamic art, 650–1100

to illustrate both the quality and the limitations of the information on vocabulary that al-Azraqi can provide. The first passage is an account of the church allegedly built in San‘a by Abraha, a more or less legendary Christian, probably Ethiopian, king who in the second half of the sixth century sought to conquer Mecca.17 The account is fictional, as are the main accounts of the castle of Ghumdan, also in San‘a.18 Neither al-Azraqi nor his immediate sources had ever seen the church, or even its ruins – assuming it ever existed. Nevertheless, for reasons which are not entirely clear but which may well have involved the then prevalent mythology about pre-Islamic architecture in Yemen, al-Azraqi’s description is so precise that R. B. Serjeant was able to sketch out a reconstruction from it.19 The building consisted of a large columnar hall (bayt), followed by a more formal hall (iwan)20 and then by a domed one (qubba). The sequence suggests a long church of a sort that would not have been impossible within the typology of early churches removed from the main urban centers of the Mediterranean. But a more interesting point lies in the three terms – bayt, iwan, and qubba – used to define the parts of the elaborate building. One of them refers to a form, the other two to functions, but all three are in standard use in early Arabic texts for defining (usually secular) built and covered spaces, and must correspond to some way of perceiving and organizing one’s perception of architectural ensembles. On a more technical level, this same passage provides a vocabulary for elements of construction – kibs (platform?), sur and ha’it (two types of walls, probably to be differentiated as outer wall and partition), and daraj (steps), and for materials and ways of building, including a very precise description of a closely fitted stone masonry (mutabaqah). It also makes several references to decoration – either in general, as with the word manqush, or more specifically, as in the description of crosses decorated with mosaics and in a technique known as balaq, which channels the light of the sun and of the moon inside the dome. Although such descriptions of long-gone monuments must have been couched in terms that were understandable to a ninth-century reader, the same terms are used in literary accounts of ancient masterpieces, and only a comparative study of several such texts will permit the clear separation between literary clichés and contemporary facts. No such problem exists 17 18 19 20

Azraqi, Akhbar, pp. 88–91. Barbara Finster and Jürgen Schmidt, “Die Kirche des Abraha im San‘a,” Norbert Nebes, ed. Arabia Felix (Wiesbaden, 1994), pp. 67–85. Al-Hamdani, Iklil, trans. N. A. Faris, The Antiquities of South Arabia (Princeton, 1938). R. B. Serjeant and R. Lewcock, San‘a (London, 1983), pp. 45–47. I am less sure than Serjeant is that iwan should be translated as “arched space” and maintain my earlier position (s.v. “Iwan” in Encyclopaedia of Islam, 2nd edn) that in classical times the word referred to a function rather than to a form. Serjeant’s emendations of al-Azraqi’s text are all eminently reasonable, even if at times hypothetical.

upon reading al-azraqi

63

with those passages in al-Azraqi which deal with roughly contemporary constructions.21 There we can cull a wonderful example of a process of construction. When the caliph al-Mahdi decides to enlarge the mosque, he calls surveyors (muhandisun) for advice; spears (rimah) are used as markers; private properties are expropriated; a budget is fixed; materials such as columns are brought by boat from Syria; a flood during construction compels a modification of levels, for which surveyors are brought in again; some supports (asatin) are built with marble, and arcades are roofed with gilded and decorated wood; other supports are of stone.22 Al-Azraqi’s description of the gates of the Haram is so precise as to have made it possible for Jonathan Bloom to propose a reconstruction for them,23 but such reconstructions are equally possible for colonnades, minarets, ceilings, cornices, inscriptions, ornaments and even light fixtures. The whole sanctuary is broken down into its constituent elements, and these elements are then enumerated, measured and described. But in order to make these reconstructions plausible in all details, the Meccan information must be related to monuments and descriptions from other places of contemporary [4] building activities, and its vocabulary compared with whatever is known to have been in use elsewhere. In addition to legendary accounts of the past and descriptions of contemporary practices and activities, al-Azraqi’s account also contributes to our knowledge of architectural terminology and architectural and religious history when it deals with particularly complex features of the sanctuary in which old pre-Islamic practices had not been entirely transformed into a set Muslim liturgy. Such is the case of the Zemzem well and of the area around it, which so puzzled Gaudefroy-Demombynes24 and in which the following curious sequence of constructions took place.25 Adjoining the well with its two basins was the majlis (usually meaning at this time a reception hall) of Ibn ‘Abbas, whose family had some sort of control over the well. During the rule of the first ‘Abbasid caliph, al-Saffah (749–54), the governor of Mecca put a dome (qubba) over the majlis. In the reign of al-Mansur (754–75), two shubbak (literally openings or windows, but usually interpreted to mean wooden screens) were built around the well, and under al-Mahdi (775–81), a kanisa (literally “church,” interpreted as “petite loge” by GaudefroyDemombynes) was built in the domed room. As a curiosity, al-Azraqi adds that it was built by a carpenter (whose name he gives) brought from Iraq by the ‘Abbasid governor, and that this carpenter also made a roof and a door for the governor’s house.

21 22 23 24 25

Azraqi, Akhbar, pp. 312 ff., among several places. Ibid., pp. 317–19; for another example of process, see p. 344. Jonathan Bloom, “The Mosque of al-Hakim in Cairo,” Muqarnas, 1 (1983), pp. 25–6. Gaudefroy-Demombynes, Pèlerinage, pp. 71 ff. Azraqi, Akhbar, pp. 299–300.

64

early islamic art, 650–1100

This passage has four terms with architectural connotations, of which only one – qubba – can be understood in its obvious and common meaning of dome. For two others, some alternative or secondary meaning must be found, and for the fourth, kanisa, a meaning has to be invented. It is likely that in this account of structures that had been built a hundred years before al-Azraqi’s time, our chronicler was quoting or repeating a terminology which no longer made sense, for, as we shall see, the whole setting of the sanctuary was changing rapidly in the second half of the eighth century. His expressions, the functions which would have been associated with them, and the forms they imply must therefore be replaced, if at all possible, in a very precise context of early ‘Abbasid history. The full elucidation of the issues raised by all the examples I have so far given would require considerable linguistic, philological and historical investigation in many different sources. Alongside the broad structural problems raised earlier, whose solution would situate the perspective from which al-Azraqi wrote, such investigations would do more than identify a technical vocabulary of forms and functions; they would also provide what might be called the “analytical process” of a ninth-century observer, a unique process no doubt, but one whose perception and understanding can only be formulated within the verbal and conceptual competence of a time. A final series of remarks derived from reading a section of al-Azraqi’s text will clarify a further methodological point I am trying to make. There is a fascinating interplay in the Akhbar between events recalled by the chronicler and the visual perception of the sanctuary in his time, and it is this interplay which becomes the recorded history of the sanctuary. The section deals with the development of the mosque proper, that is, the area that eventually became the large open space around the initial and mostly pre-Islamic core of holy places.26 For it the following sequence can be reconstructed: 1.

2.

3.

26

Under the righteous caliphs ‘Umar (634–44) and ‘Uthman (644–56), a few houses adjoining the Ka‘ba and the few holy spaces in its immediate vicinity were bought and razed, and a low wall or fence (jidar qasir) was marked out (ahatta); there was no covered (musaqqaf) area anywhere. When Ibn Zubayr ruled over Mecca (680–92), he acquired a few more houses and even parts of houses (including one belonging to an ancestor of al-Azraqi) and enlarged the sacred space, but without altering its simple character. Most of the account actually deals with the location of the old houses in the mosque of al-Azraqi’s time; it also includes an obscure explanation of how to walk backwards from the corner with the Black Stone. The Umayyads did not modify the size of the mosque, but they did transform its character. Both ‘Abd al-Malik (685–705) and al-Walid Ibid., pp. 306 ff.

upon reading al-azraqi 65

4.

5.

27 28 29 30

(705–15) are credited with beautification (husna), and it is probable that we are dealing with a single activity which lasted many years. The outer walls were raised and a covered area was built, consisting probably of a portico with a wooden ceiling; the capitals or upper parts (ru’us) of the supports (asatin, piers or columns) were gilt. Al-Walid is remembered for having covered supports with marble, and soffits or spandrels (wajh al-tayqan) with mosaics; he also built its crenellations and moldings, if this is the correct way of interpreting the word shurrafat.27 It is interesting to note that the verb used to mean “cover,” as with marble, is azzara, which means to “veil” or to “cover with a piece of clothing.” Elsewhere in the text, [5] the most common verb for “decorating a surface” is albasa, “to clothe.” Between 754 and 758, a major program of construction was ordered by the caliph al-Ma’mun. It was commemorated by a triumphal inscription on a newly built gate in black mosaic cubes on a gold background, and one of its two qur’anic quotations (3:96) proclaims that “the first house appointed to Me was the one at Bakka” (i.e., the Ka‘ba).28 It consisted of increases in area in whatever directions were available through the acquisition of houses; in extensive decoration; and in the building of formal gates and of a manara (minaret), whatever was meant by that term at the time. The caliph al-Mahdi (775–85) went on a pilgrimage in 776–7 and immediately undertook a major program of repairs and modifications of the usual sort. It is described in such detail by al-Azraqi that it should be possible to sketch out a full reconstruction, if not of the whole building, at least of the main units (gates, vaulted arcades, open spaces) which were its constituent parts. Four years later, in 780–81, al-Mahdi returned to Mecca and, having inspected the work done, saw that “the Ka‘ba was on one side (fi shiqq) of the mosque; he did not like that and wished that it be in the middle (mutawassata) of the mosque.”29 He called in surveyors and architects, bought houses, fixed the wadi going through the mosque so that it would not flood as often as it was wont to do, had columns brought from Egypt and Syria, and altogether created a complete monument with the Ka‘ba in the center, arcades around an open space, a minaret, and gates leading into various parts of the city.30 Al-Azraqi then proceeds to detailed descriptions of the architectural elements of the masjid, eventually providing a list of such changes or repairs as occurred between 781 and his own time, and Cornice is also possible. E. Combe et al., Répertoire chronologique d’épigraphie arabe (henceforth RCEA), 1 (Cairo, 1931), no. 40; Azraqi, Akhbar, pp. 311–12. Azraqi, Akhbar, p. 317. Some of these building activities have been recorded in inscriptions listed by Ibn Jubayr, RCEA, nos 49 ff.

66

early islamic art, 650–1100

discusses the rules, practices and regulations affecting behavior in what is both the congregational mosque of a city and the sacred space around that unique holy place. This sequence of five changes is of considerable interest for many reasons. One is that nearly every other feature of Mecca, as remembered by al-Azraqi, was affected by roughly the same set of interventions: those made by the Prophet and his immediate successors; by Ibn Zubayr, the Meccan aristocrat who sought to establish his city as the center of Islam; by the Umayyads, primarily al-Walid; and by al-Mansur, and al-Mahdi, the second and third ‘Abbasid caliphs (one of whom was also the founder of Baghdad). But this raises a question: why were so many major modifications made to the mosque over a period of not more than five generations? One explanation may simply be that the growth of Islam and the presumed increase in the number of pilgrims required an ever-larger space. That this was so can hardly be questioned. But, in addition to that, each of these interventions in the mosque and elsewhere also seems to have had a different motive behind it. The first caliphs merely wanted to maintain the status quo established by the Prophet and to accommodate the still barely elaborated rites of the pilgrimage.31 Ibn Zubayr had more complex ambitions, and he alone rebuilt the Ka‘ba as it allegedly had been before the Quraysh rule over the city.32 A full investigation of his ideological and other motives is sorely needed. The Umayyads were remarkably inactive in Mecca. They did, of course, rebuild the Ka‘ba as it had been in the Prophet’s time, but the remainder of their work either involved ornamentation and is known only through statements smacking of literary clichés, which occur in many accounts of their buildings, or else it was downright blasphemous, as when one of their governors downgraded the Zemzem well in order to dig a well with sweeter water.33 It was the early ‘Abbasids who first, under al-Mansur, officially proclaimed the unique holiness of the Ka‘ba and built a full-fledged mosque around it, and then, under al-Mahdi, reached the aesthetic decision to make of it a true monument. At this stage of scholarship, we can only speculate as to why it was alMansur and al-Mahdi who so deliberately formalized the sacred mosque and gave it a visual and compositional definition. Yet it is not an accident that 31

32 33

Azraqi, Akhbar, p. 234, for a curiously skeptical statement by Caliph ‘Umar about the real holiness of the Black Stone. Had he not seen the Prophet’s veneration for it, says ‘Umar, he would not have believed in it. For a model analysis of the formation of Muslim practices, see Klaus Lech, Geschichte des islamischen Kultus: I. Das Ramadan Fasten (Wiesbaden, 1979). M. Ibrahim, “Social and Economic Conditions in Pre-Islamic Mecca,” International Journal of Middle Eastern Studies, 24 (1982), p. 343, for an interesting approach. Azraqi, Akhbar, pp. 339–40; Gaudefroy-Demombynes, Pèlerinage, pp. 74–5 and p. 98, n. 3; for an interesting commentary of another Umayyad attempt at innovation based on Azraqi, p. 461.

upon reading al-azraqi

67

the patron of Baghdad, with its palace complex in the center of an urban ring, also transformed the sanctuary of Mecca into a large space around a holy place. But it is curious that both al-Mansur and al-Mahdi had traveled to Jerusalem, had seen there a vast and only partially rebuilt esplanade with a stunning Umayyad monument in its psychological, if not actual, center, and made major contributions to its monuments. The building up of Mecca’s sacred mosque into the shape it finally acquired can perhaps best be seen as the result of a new taste for centralized planning developed by the early ‘Abbasids and of the very precise memory of the Dome of the Rock in Jerusalem. This suggestion introduces yet another element in the visual as well as ideological contest or parallelism between the two holy cities in the seventh and eighth centuries.34 Beyond this admittedly speculative possibility, the ‘Abbasid formulation of Mecca as a shrine, almost a dynastic one in part, explains the official use of the Ka‘ba by Harun al-Rashid when he put his political [6] testament there and by al-Ma’mun when he filled it with treasures from conquered Eastern potentates. It also explains the succession of repairs and additions made to the mosque throughout the ninth century, and, on a more practically pious level, the development of the Darb Zubayda, the great road for pilgrims from Iraq to Mecca. While describing what he saw, al-Azraqi reflected the formal ideology of ‘Abbasid power and the incorporation of Mecca within it. Thus from practical issues of architectural practice and vocabulary to speculation about the relationship between holy places and the growth of an ‘Abbasid aesthetic, al-Azraqi’s text offers a mine of information which has hardly been tapped and whose full exploitation requires a variety of investigations barely sketched out in these remarks. The uniqueness of Mecca led to a unique source about it, but, perhaps more important, an analysis of alAzraqi’s text requires a modification of Gaudefroy-Demombynes’s early judgment that the Meccan sanctuary was a “monument built without any method.”35 Its “method” can be understood once the several discrete moments of its history can be both visually and ideologically isolated.

34

35

I expect soon to return to the topic in a far more systematic way, but I have discussed some of the religious associations common to both places in “The Umayyad Dome of the Rock,” Ars Orientalis, 3 (1957); there are many more. Gaudefroy-Demombynes, Pèlerinage, p. 116.

Chapter IV The Iconography of Islamic Architecture*

Over 950 years elapsed between the construction of the earliest fully documented monument of classical Islamic architecture (the Dome of the Rock in Jerusalem, completed in ad691) (Fig. 1) and its latest celebrated masterpiece (the Taj Mahal in Agra, completed in 1654) (Fig. 2). During the millennium that separates them, tens of thousands of monuments were built from Spain to China, from Siberia to sub-Saharan Africa. Archaeologists, art historians and architectural lexicographers have made enormous progress in classifying buildings according to function (mosques, mausolea, palaces, houses, baths); geographical region. (Egypt, Iran, Turkey, Central Asia, India; period (early Islamic, Seljuq or middle Islamic, Ottoman, Mughal); or any combination of these essentially taxonomic categories – that is to say, categories of definition and ordering. They are taxonomic because they are valid, incontrovertible and, once established, definitive. Anyone among us can give examples of discoveries – a dated inscription, an excavation or sounding, the subject-matter of a decorative program – that have permanently altered the history of a monument. An obvious recent example occurs in Galdieri’s excavations in Isfahan.1 These led to postulation of the architectural evolution of the parts of the Great Mosque that is clear up to the construction of a large dome in front of the mihrab, even though what happened later is still uncertain, and even though the rather incongruous visual impression provided by Galdieri’s pictorial reconstruction still raises doubts about the architectural competence of the time. In a more general way, Creswell’s enormous achievement is the demonstration of a noble concern for the establishment of “facts” and for their reasonable classification in sequences according to formal and temporal characteristics and relationships.2

* First published in Priscilla P. Soucek, ed., Content and Context of Visual Arts in the Islamic World (Pennsylvania State University Press, University Park and London, 1988), pp. 51–65. 1 E. Galdieri, Isfahan: Masgid-i Gum’a, II (Rome, 1973), and “Précision sur le Gunbad-e Nizam al-Mulk,” Revue des Etudes Islamiques, XVIII (1975). 2 This is not to say that Creswell’s volumes are free from deep-seated prejudices and preconceptions, but his system of identifying characteristic features and of seeking their prototypes bears all the external appearance of straightforward “scientific” rationality.

69

70

1 The Dome of the Rock, Jerusalem

early islamic art, 650–1100

Why not remain satisfied with the immense progress over the past fifty years in our knowledge [52] of facts about Islamic architecture and in our organization of this information into reasonably accepted categories?3 Two reasons not to, I believe, have emerged over the past decade. One derives from my involvement over the past several years with the activities of the contemporary architects and planners, Muslim and non-Muslim, who are reshaping the face of the whole Muslim world. The questions they ask are never: who built something? and why?, but nearly always: what is Islamic in this? and how can I, a modern builder (frequently alien to Islamic culture), use the traditional past and its monuments to create something today? From the point of view of real contemporary interests, this is necessary information – like the knowledge of anatomy for a doctor. In other words, out of the taxonomic order of knowledge, something is expected that is not normally required. I shall return to this expectation in my conclusion – but let me

3

Two recent publications illustrated two ways of categorizing Islamic architecture. John Hoag’s Islamic Architecture (New York, 1976), provides a historical survey with clearly indicated and, for the most part, acceptable temporal and regional subdivisions. The volume edited by George Michell, The Architecture of the Islamic World (London, 1978), catalogs monuments according to areas and provides broad thematic essays.

the iconography of islamic architecture 71

2 Taj Mahal, Agra

now turn to the second reason for dissatisfaction with the taxonomic progress of the field. It lies in the monuments themselves. Let me return to the Dome of the Rock and the Taj Mahal. In both instances we are dealing with masterpieces remarkable for the fact that their continuing importance within Islamic culture has very little to do with the reasons for their actual construction. The Dome of the Rock has become a commemorative monument for the Prophet’s mystical journey into the heavens, but it was built in 691–2 for the very ideological local purposes of sanctifying the old Jewish Temple according to the new Revelation and of demonstrating to the Christian population of the city that Islam was the victorious faith.4 The Taj Mahal has always been considered the most romantic monument to a dead spouse, but a brilliant recent investigation has demonstrated that it was actually an extraordinary attempt to show God’s throne on earth as it will appear at the time of the Resurrection.5 In a fascinating contrast to what happened with the Dome of the Rock, the romantic Western vision of the Taj Mahal was accepted by the Muslim world as a convenient explanation for a monument with an unorthodox purpose. 4

5

O. Grabar, “The Umayyad Dome of the Rock,” Ars Orientalis, 3 (1957). For different views, see Creswell, Early Muslim Architecture, 2, 226 ff., and W. Caskell, Der Felsendorn und die Wallfahrt nach Jerusalem (Cologne, 1963). On the whole, I have not been swayed by most of these arguments, although the historical and cultural context of the last decades of the seventh century can now be explored in a sharper fashion than I had done. Wayne Begley, “The Myth of the Taj Mahal and a New Theory of Its Symbolic Meaning,” The Art Bulletin, 61 (1979).

72

early islamic art, 650–1100

The reinterpretation of both monuments was based on two elements. One was historical logic demonstrating that the later view of the buildings could not have been imagined at the time of their creation. The other was that each monument contains a document neglected by previous investigators: qur’anic inscriptions chosen to provide the immanent, concrete, specific meaning of the monument. Both historical logic and qur’anic inscriptions are not architectural features but extrinsic sources of information and understanding. Why, then, is it legitimate to consider the Dome of the Rock and the Taj Mahal as masterpieces of Islamic architecture? Is it because they simply happen to have been sponsored by Muslim patrons, the great ‘Abd al-Malik (founder of the first coherent state outside of Arabia), and Shah Jahan (the complex figure who ruled one of the last great Muslim empires)? Or is it because they both express, at a distance of almost a thousand years and in very different lands and political situations, a common idea, a shared thread, something which reflected the cultural needs and uniqueness of the Muslim world? By raising the question in this fashion, I am immediately raising two subsidiary but fundamental questions of the history of art and of cultural history. Does something become Islamic because a Muslim builds it or uses it? In the case of the Dome of the Rock, for instance, all scholars agree that the shape of the building, its technique of construction, its decoration, and nearly all its physical attributes were not created by Islam but were part of the traditional – Christian – vocabulary of the eastern Mediterranean. To say this, however, is to indulge in academic pedantry. The origins of the forms used in the Dome of the Rock bear on the meaning given to it over the centuries only if one can demonstrate that a consciousness of these origins remained with the culture, or if one accepts a Jungian notion that every culture requires the same basic forms to [53] express its religious or social needs and that formal alterations are merely secondary.6 The other question, or, rather, the other way of posing the problem, is this: is there anything in the forms of these monuments – as opposed to their use – that makes them Islamic? And, if there is, what is it? This second series of questions is the subject of this essay. My aims are to develop an intellectual strategy for further research on Islamic architecture, and to meditate on a key issue of contemporary thought: whether it is valid to apply the same investigative methods to the art of all cultures, or whether the very nature of artistic experience requires methods created by the culture itself. One last introductory remark is in order. Much of what follows here is preliminary, and not all of it is my own work. It is the result of research, 6

This approach permeates parts of N. Ardalan and L. Bakhtiar, The Sense of Unity (Chicago, 1972), as well as studies – less directly involved with the Muslim world – by scholars such as M. Eliade.

the iconography of islamic architecture 73

often still unpublished, by a half-dozen students at Harvard, MIT and elsewhere, and it reflects what I can only call a collective laboratory effort at scholarship – to my mind the only way for research in the humanities to become true research. It is, first of all, easy enough to demonstrate that Muslims quite consistently used certain forms, that courtyards with porticoes or with iwans, domes and towers became part of the setting in which Muslims live – but there is nothing intrinsically Muslim about a courtyard, an iwan, a dome, or a tower. Each one of these forms has a pre-Islamic history and non-Islamic functions. In these instances, the important issue is simply to discover the nature of the “charging” of forms that makes the towers of San Gimignano in Italy clearly non-Muslim but the towers of Fez or Cairo the minarets of a Muslim setting. There are two kinds of more or less traditional methods of dealing with such issues. One approach I would like to call symbolic: its assumption is that there are features, perceptible visually, which, whatever their origin, possess or have possessed an immediately accepted cultural association. A most obvious example in Islamic architecture is the minaret, whose meaning as the place for the Muslim call to prayer is accepted by all, even though, as I shall suggest later, this was not always so. A less obvious but more important example is the muqarnas, that fascinating composition of three-dimensional units often called a stalactite or honeycomb. The muqarnas has two features lacking in the minaret: it is an entirely Muslim invention, almost never copied in a non-Muslim context except by Armenians in the thirteenth and seventeenth centuries, and it is a form used in nearly all kinds of Islamic monuments, not only in mosques. The symbolic approach would then be to say that the minaret or the muqarnas must have been uniquely meaningful to Islamic culture, and meaningful only in the Muslim world. They must symbolize something deep within the culture; they must stand for something essential to the purposes and existence of the ummah or community. Our task then is to study what within the culture was symbolized, and to explain the complex of meanings and references involved in a Muslim’s reaction to a minaret or muqarnas. As in most symbolism, the proof of meaning lies less in the form itself than in the conscious or unconscious make-up of the viewer or user. For instance, purely optical observation, whatever the physiological or psychological reasons, tells us about the “projecting activism” in red or yellow and the “receding passivity” in blue or green, but only literary or ethnographic sources identify green as the color of the Prophet or as the color of the naked heart.7 One cannot explain the symbolism of the colors of a Persian dome as a reflection of the mystical or even archetypical unity of creation, or the whiteness of a North African town as a reflection of 7

Ardalan and Bakhtiar, Sense of Unity, p. 54. See also H. Corbin, L’Homme de Lumière dans le soufisme iranien (Paris, 1971), pp. 164 ff.

74

early islamic art, 650–1100

the purity of the Prophet’s message, without concrete textual or ethnographic evidence, that is, without the agreement of the culture itself. The second approach is iconographic, meaning that in Islamic architecture certain forms denote or describe a Muslim idea or concept. A simple example would be the mihrab; its location in a religious building, its decoration, and its most common inscriptions (Qur’an 9:18 on the masajid Allah (Fig. 3) or 24:35–8, the verses on Light) indicate that an architectural form of no particular [54] significance – the niche found in thousands of buildings in classical architecture – has been transformed into a sign denoting very precise Muslim purposes: the direction of prayer, the commemoration of the Prophet’s presence, and other even more complex meanings.8 If, in any covered space, archaeologists find a niche directed to the qiblah, they decide that the building is a mosque. And the presence of a niche supported by columns (and sometimes including a lamp or a vegetal motif ) on tiles, tombstones, rugs and other media indicates that the mihrab form became an iconographic sign with some constant meanings and a number of variables. The extension of the iconography of the mihrab, however, is less an architectural phenomenon than a decorative one, as it appears in twodimensional rather than three-dimensional space. A related mode of iconographic interpretation can be demonstrated at Qusayr ‘Amrah, at Khirbat al-Mafjar, possibly at the Aqsa mosque or the Cordoba maqsurah, the Cappella Palatina and, I suppose, later monuments with which I am less familiar. These include the Chihil Sutun or the Hasht Bihisht, where paintings, sculpture and other techniques of decoration provide the charge to architectural forms. It is, however, usually difficult to demonstrate that such meanings as are provided are an intrinsic part of the architectural forms themselves. The wealth and ubiquity of this type of iconographic charging through decoration has been amply demonstrated in Karl Lehman’s great study of the Dome of the Heaven.9 But, with a few exceptions such as the Pantheon,10 Hagia Sophia, or the church described in a celebrated Syriac hymn,11 it is not in the architectural forms that the complexities of meanings were found. Architecture here is iconophoric, not iconographic. More complex instances of architectural iconography occur when one monument becomes a model for successive copies, imitations and transformations or when a certain type of monument denotes something special in the culture. To my knowledge, there are few examples of the first type in classical Islamic architecture, but they do exist, as, for instance, the visual 8 9 10 11

The history of the mihrab still needs to be written, as more energy has been spent investigating its origins than on the 1400 years of its development and use. Karl Lehman, “The Dome of Heaven,” The Art Bulletin, 27 (1945). W. Macdonald, The Pantheon: Design, Meaning and Progeny (Cambridge, Mass., 1976). C. Mango, The Art of the Byzantine Empire, 312–1453: Sources and Documents (Englewood Cliffs, 1972), pp. 72–102, 57–60.

the iconography of islamic architecture 75

3 Tile mosaic mihrab, Isfahan, dated 1354; outermost inscription is Qur’an 9:18–22, including masajid Allah

imitation of the Dome of the Rock in Qala’un’s mausoleum in Cairo, and I have no doubt that further studies will uncover iconographic derivations of the type that Richard Krautheimer described around the Holy Sepulchre in Christian architecture.12 For instance, an iconographic sequence of the hazirah 12

R. Krautheimer, “Introduction to an Iconography of Medieval Architecture,” Studies in Early Christian Medieval and Renaissance Art (New York, 1969).

76

early islamic art, 650–1100

in funerary complexes goes back to the Prophet’s mosque in Medina in ways only barely sketched out up to now.13 And I have tried elsewhere to construct a similar iconographic history for the Court of the Lions in the Alhambra.14 The point in all these cases, as in all such instances of architectural iconography, is that some mechanism of cultural perception makes a genetic association between forms that may be quite different in detail. This kind of iconographic investigation can probably only be made through texts, for instance, by comparing Ibn Jubayr and Ibn Battuta’s accounts of buildings. More interesting are the meanings attached to specific architectural types. Let me give two examples. One is a fairly simple one. Sometime in the sixteenth century the Ottoman mosque acquired its classical characteristics (Fig. 4): a large central dome supported by half-domes, a porticoed courtyard, slender minarets, and a unique compositional logic based on the diameter of the cupola, with size, light and decoration as variables. This type, almost certainly a creation of the Ottoman capitals, is best expressed in the great mosques of Istanbul, but it occurs in Algiers, throughout the Balkans, in Syria, and even in Muhammad ‘Ali’s Cairo. It does not occur in Morocco, Iran, India, or Central Asia because this type is tied to Ottoman supremacy. It serves an Islamic function, but its architectural forms signify a specific empire. The second example is more complicated. It seems clear that in the seventh and eighth centuries, the central lands of Islam (primarily Iraq, if my conclusions are correct) developed a type of mosque based on a multiplicity of single supports known as the hypostyle mosque (Fig. 5). This [55] type both served and reflected the characteristics of the early Muslim community and acquired a more elaborate regional variant with the mosque of Damascus. Nearly every early mosque in the new cities of the Muslim world – Cordoba, Qairawan, Isfahan, Siraf, Nishapur – was hypostyle. What is interesting is what happened later. The hypostyle mosque with a single minaret and an elaborate mihrab area became characteristic of the entire Arab world until today, from Morocco to Iraq; it appears in all sizes, from huge buildings, as in Rabat, to small ones, neatly fitted within their urban setting, such as the Aqmar mosque in Cairo and any number of masjids in Syria. The hypostyle mosque was also often the first type of mosque built when a new area was conquered or converted. In Konya, the new congregational mosque of the thirteenth century is a hypostyle, and so are some of the earliest mosques in India. Most African mosques of any size tend to the hypostyle. It is as though at those moments and places when the important cultural objective was the strengthening of Islam, and not the extension of a 13 14

L. Golombek, The Timurid Shrine at Gazur Gah, ROM Art and Archaeology Occasional Paper 15 (Toronto, 1969). O. Grabar, The Alhambra (London, 1978), especially chapter 3.

the iconography of islamic architecture 77

4 Ottoman imperial mosque: Suleymaniye, Istanbul

78

5 Hypostyle mosque: Great Mosque of Qairowan

early islamic art, 650–1100

state, the hypostyle mosque provided the architectural form through which the presence of the faith could most easily be expressed. Why? There may be practical reasons, for instance, the possibility of creating a space tailored to any size of community (the hypostyle is an unusually flexible form) or the absence of a hierarchy of parts reflecting the equality of the Faithful. But a more profound explanation is that the hypostyle form remained in the collective memory of Muslims and was associated with an early, unadulterated Islam, and that it expressed that view of itself that the Muslim world was particularly anxious to project. Such revivals of early Islam are probably more numerous than we recognize. For instance, in the early fourteenth century, in the north-central Iranian city of Bistam, two small cells were built in the sanctuary of the great mystic Abu Yazid ai-Bistami.15 They were probably meant for private meditation, but it is interesting that in an inscription they are called sawma‘ah,16 an old word found in the Qur’an (22:41) but with unclear significance. It means “minaret” in North Africa for reasons elucidated by Creswell, but in some 15 16

A definitive treatment of the sanctuary has not yet appeared. Pope, SPA, p. 1080; and D. N. Wilber, The Architecture of Islamic Iran (Princeton, 1955), pp. 127–8. RCEA, 8 (1944), no. 5155.

the iconography of islamic architecture 79

historical sources it also refers to the small rooms for hermits found in the towers of the mosque of Damascus at the time of the conquest. The term is rarely used in medieval literature, but its reappearance in the early fourteenth century to describe the setting of a pious function so reminiscent of the stories spun around seventh-century mosques in Syria seems to me to demonstrate the persistence of early Islamic concepts and their reappearance when an association was made, for whatever reason, with the first century of Islam. These examples indicate the existence within the evolution of Islamic architecture of an order of meaning which is inherent neither to forms nor to functions, nor even to the vocabulary used for forms or functions, but rather to a relationship among all three. This relationship had a history, a development, almost certainly a number of constants and variants. That history, these constants and variants, still requires an extraordinary amount of research, not only in the monuments themselves but in the huge literature that deals with them or refers to them. But that there is (or was) an iconography of Islamic architecture seems clear to me. This is not surprising, for there is every reason to assume that Islamic architecture contains the same complex meanings as does classical, Christian and Hindu architecture. It is simply that so little effort has been spent on the meanings of Islamic architecture that their depth has been overlooked. How does one deal with this underappreciated area? The problem is that Islam does not possess the two vehicles through which Christian or Hindu architecture can be understood. One is a complex, codified liturgy that would affect architecture. (There are cases, known to me especially in the Mediterranean area, in Cordoba or in Fatimid Egypt, where complex ceremonies did accompany Friday prayer and affected the shape and possibly the decoration of minbars and mihrabs,17 but as a rule the absence of a liturgy and of a clergy makes matters more difficult.) A [56] second traditional aid to the understanding of architectural meaning is the creation of decorative programs. Such programs actually existed, but the fact that they were not based on images makes them extremely difficult to approach, because the West-centered universal culture of today finds it difficult to understand anything without a system of representations. At this stage of research, all that can be done is to indicate some of the techniques that I believe will help to deal with the question of how to look for meaning in Islamic architecture. I will use the two examples I mentioned at the beginning of this essay as apparently symbolic of something within Islamic culture: the minaret and the muqarnas. Almost everyone agrees that the minaret derived from a specifically Islamic requirement, the idhan or call to prayer. There is little doubt that, from 1500 17

Caroline Williams, “The Cult of Alid Saints in Cairo,” Muqarnas, 1 (1983), pp. 37–52; and Jonathan M. Bloom, “The Mosque of al-Hakim in Cairo,” Muqarnas, 1 (1983), pp. 15–36.

80

early islamic art, 650–1100

onward, and perhaps even as early as the middle of the fourteenth century, the call to prayer was the main official purpose of the minaret and most minarets were used for that purpose, as they are today. (Iran is the exception; some complexities were introduced there through the existence of a guldastah.) When minarets were provided with inscriptions, the most common one was from Qur’an 62:9–10: O you who believe, when the call is made for prayer on Friday, then hasten to the remembrance of God and leave off trade; that is better for you, if you know. But when the prayer is ended, then disperse abroad in the land and seek of God’s grace, and remember God much, that you may be successful.

There is a perfect coincidence between the purposes of this structure and the verbal sign of God’s revelation inscribed on it. When we turn to the beginning of Islam, matters are confused. Whereas the call to prayer is as early as Islam, the precise time, place and manner of its association with the tower, a form known since time immemorial, are extremely unclear. (I have presented my own explanation of what happened elsewhere, but would be perfectly happy to be proved wrong.)18 The earliest consistent and authentic evidence we possess is of the eleventh to thirteenth centuries, a period from which nearly two hundred minarets have survived. A large number of them are provided with inscriptions, and at least twenty of these are qur’anic passages. Only one, the Qutb-minar in Delhi (Fig. 6), uses Qur’an 62:9–10, and this passage is only one-thirtieth of the qur’anic inscriptions on the minaret. All the other minarets, from Hakim’s in Cairo to the minaret of Jam, in Afghanistan, are not merely inscribed with other qur’anic passages but also with passages that differ from each other: at Jam the whole of the Surah 19 (“Maryam”) appears, whereas the inscription on Aleppo’s minaret is a passage (2:121–22 and 60:60) dealing with those who have erred from the straight path, and Delhi’s minaret has not only the Ayat al-Kursi but the five ayats that follow, which have little to do with prayer.19 In Sangbast, the inscription is 41:33: Who is better in speech than one who calls (men) to God, works righteousness, and says I am of those who bow in Islam.20

If we add that an unusually large number of minarets are not even located near mosques (especially in Iran and the eastern part of the Muslim world),21 we must conclude that the towers we call minarets fulfilled a broad range of 18 19 20 21

O. Grabar, The Formation of Islamic Art (New Haven, 1973), pp. 118 ff. A. Maricq and G. Wiet, Le Minaret de Djam (Paris, 1959); all other inscriptions are taken from the RCEA. D. Sourdel and J. Sourdel-Thomine, “A propos des minarets de Sangbast,” Iran, 17 (1979). See a study of these minarets by Prof. R. Hillenbrand in his unpublished doctoral dissertation. In the meantime, see J. Sourdel-Thomine, “Deux Minarets,” Syria, 30 (1953).

the iconography of islamic architecture 81

6 Qutb Minar, Delhi

82

early islamic art, 650–1100

functions within the architectural sign system of the time. My own explanation, very tentative at this stage, is that minarets at that time were still primarily expressions of power and wealth, of victory, and of Islamic presence in non-Muslim settings, or else were signals and signposts leading to major holy sites or to other significant mosques, rather than identifying them. [57] A fascinating example of the use of the minaret occurs in Jerusalem, where, until the Crusades, a physical balance had been established between a western Christian sector centered on the Holy Sepulcher and an eastern Muslim area around the Haram al-Sharif, with a small Jewish quarter probably to the north. No Muslim building, with one small and very temporary exception, was found in the western city and no Christian one remained in the east. After the defeat of the Crusader state in the thirteenth century and the development of a Jewish quarter to the south, two minarets were built in the Christian quarter equidistant from and framing the Holy Sepulcher.22 Minuscule sanctuaries, which hardly fulfilled an important social role, are attached to them. But the minarets serve to emphasize the victory of Islam, just as the later towers of the Lutherans and of the Franciscans identified both the return of Western Christianity to Jerusalem and the competition there between Protestantism and Catholicism, whereas a Russian tower on the Mount of Olives showed the presence of Orthodoxy. Several very different conclusions derive from these observations. One is that the single, collectively accepted source of Truth in Islamic culture, the Qur’an, is used in so many different fashions on minarets that we must assume that its message takes precedence over architectural forms; whatever use or explanation may have been given to these towers later on, they initially had a practical, specific, time-bound purpose, and it is only within the limits of the time that created them that their iconographic meaning can be securely established: unless otherwise demonstrated, iconographic time is short. The second point is that even though different local circumstances led to the creation of each of these minarets, when considered as a group they belong to two subsets. One is the subset of the tower, the strong, high unit, visible from afar and dominating its social setting; at this level, the minaret is today no longer “Islamic” (nor was it ever), since television towers, towers of silence, or even office buildings fulfill a universal human need for a vertical architectural focus. The other subset is stylistic, for instance, the treatment of brick on Iranian minarets can be related to the treatment of brick on Iranian mausolea of the time. For us as historians of forms, the universal value of the tower, its more limited relationship to a period style, and whatever local need led to its creation are all essential categories, but they should not be confused in

22

A. Walls, “Two Minarets Flanking the Church of the Holy Sepulchre,” Levant, 8 (1976).

the iconography of islamic architecture 83

explaining monuments. One point, however, is clear: the tower form we call the minaret was not originally used in the whole Islamic world with the sole purpose of calling for prayer; this function emerged gradually. Now it has been superseded by the loudspeaker and tape recorder in its function, and by the office building in its form. It has lost its iconographic value both as a universal form and as a concrete expression of very varied functions, but it has retained the symbolic function of indicating the presence of Islam. As a sign in the past, its strongest meaning, its greatest change, lay not in its form but in the confluence between its form and decisions by several layers of the community that endowed the form with whatever needs the community had at any one time. The implications of this point will emerge shortly. Before dealing with these, let me turn to the muqarnas (Fig. 7), this ubiquitous combination of three-dimensional or curved shapes that can be used on anything from a flat wall, where it becomes a frieze, to a whole cupola.23 The origins of the motif are not altogether clear, but it seems to have developed first in eastern Iran and then, perhaps independently and perhaps not, in Egypt and North Africa some time in the tenth century. Why did it develop? What does it or 23

The most useful recent works in this also insufficiently studied form are D. Jones and G. Michell, “Squinches and Pendentives,” AARP, 1 (1972), and U. Harb, Ilkhanidische Stalaktitengewölbe (Berlin, 1978), with an excellent bibliography.

7 Muqarnas, Hall of the Two Sisters, Alhambra, Granada

84

early islamic art, 650–1100

did it mean? For how can one possibly even begin to discuss the meaning of Islamic architecture when no explanation exists for its most uniquely original form? External sources are, to my knowledge, of little help. A fourteenth-century manual by al-Kashi24 and a sketch discovered in the excavations of Takht-i Sulayman25 indicate that at least the plane geometry of the muqarnas had been worked out and was available in simple manuals. It is [58] also probably not an accident that the muqarnas appears at the time of al-Farabi and the first major school of mathematicians in the Muslim world. But the exact relationship between the muqarnas and scientific development is difficult to establish, because no source exists, at least to my knowledge, that would explain why a theory of numbers or advanced geometry should have found an application in the muqarnas. I am not aware of anything comparable to the celebrated Syriac hymn previously mentioned or a number of Greek texts that explain the symbolic and iconographic meaning of the domed church.26 If we turn to the muqarnas itself, some tentative answers may be suggested. First of all, there are instances when an inscription does provide a specific meaning to a muqarnas. The most remarkable instance is at the Alhambra, where Ibn Zamraq’s poetry makes it legitimate to understand the cupola as a dome of heaven, in this instance even a rotating one. But this obviously does not mean that every muqarnas dome is a dome of heaven. In other words, the form itself may be considered as neutral, as simply a technical device of construction or decoration, unless a vector charges it with some meaning. The problem with this explanation is that it weakens, in fact even cheapens, the social effort necessary to make, for instance, a muqarnas-covered portal in stone; is it likely that the stupendous muqarnas of the Sultan Hasan Madrasah in Cairo was nothing but an ornament? Hence a second explanation may be provided that is broader than the first one and in fact does not exclude it. One peculiarity of the muqarnas, wherever it is found, is its ability to suggest almost infinite subdivision and, by modulating whatever surface or shape it occupies, to create the illusion that an architectural form – a wall, a ceiling, a doorway, a hall – is different from (usually larger than) what it is. This peculiarity can be understood as a game, which the muqarnas certainly was at times, but it can also be seen as an illustration of a profoundly Islamic notion of the immateriality of human creation, a notion often expressed through inscriptions like al-mulk lillah, la ghalib illa Allah, or albaqi huwa Allah. An architectural form serving to deny the materiality of

24

25 26

L. Bretanitsky and B. Rosenfeld, “Arhitekturnaia Glava,” Iskussivo Azerbaidjana, 5 (1956), pp. 125 ff. See also M. S. Bulatov, Geometricheskaia Garmonizatziia v Architekture Srednei Asii (Moscow, 1978). It is the subject of Harb’s booklet, note 23 above. Note 11 above.

the iconography of islamic architecture 85

forms would in the deepest sense be an affirmation of divine truth. And then, in monuments of secular architecture like the Alhambra, an inscription can charge the muqarnas in different fashion, but only superficially so, for whatever meaning is temporarily given to it, the fragility of human creation is always conveyed, for the other – endlessly repeated – Alhambra inscriptions refer to the eternal permanence of the divine alone. It is tempting to understand the muqarnas as a visual metaphor for a certain traditional Muslim view of reality, as the abstract carrier of a message that also found verbal equivalents and often learned ones in the more elaborate qur’anic and hadith inscriptions on some monuments. Nevertheless, I must conclude that, whereas we know that some Islamic meaning is associated with the muqarnas, we do not yet know what it is; we may be compelled to follow the example of the minaret and argue that each muqarnas is an independent form, whose discrete meanings must be understood before pan-Islamic meanings are proposed. These remarks and observations are in many ways inconclusive, for they clearly indicate the insufficiency of the information we possess and the absence of intelligent thinking about whatever information we do possess. But perhaps a few directions for work and thought can be suggested. The first and most important one is that, regardless of how much pressure is put on us singly or collectively to generalize about Islamic architecture, we cannot do so without a clearer understanding of the meaning of any one monument in its time, in the fullness of its historical circumstances. This is essential if we are to deal usefully and meaningfully with the traditions of Islamic architecture. It is not simply a scholar’s professional interest that is involved here but a contemporary’s judgment that today’s world can only be true to itself if it is aware of the immense complexity of its time past. [59] The second conclusion is that the forms of Islamic architecture, like those of any architecture, carry meanings that build systems of communication and of social relations. The carriers of these meanings, however, are, in architecture, less the forms themselves than signs added to these forms. The uniquely Muslim one is writing, qur’anic inscriptions in particular, but also hadith, poetry, and simple aphorisms or formulas, which served to define the precise aims of a monument, at least at the time of its creation. Other carriers are still incompletely understood – like geometry, for instance, the muqarnas, color, and perhaps certain formal orders. There is still some uncertainty as to whether these carriers were culturally distinctive or merely universal forms of honor or focus in architectural compositions. And we are only beginning to grapple with the infinitely more complex question of the time of art, that is, today, the duration of acceptance of synchronic meanings for any one monument. However we resolve these issues, the point remains that in order for the meanings of Islamic architecture to be understood, both a high level of literacy and an unusual power of abstract thinking were required.

86

early islamic art, 650–1100

Is it likely that these requirements were actually met in classical time? Here again, I cannot answer with a demonstration, for lack of preliminary studies, but I can argue that Islamic civilization with its many variants in time and space could not have existed and maintained itself if its message had not reached the kind of abstract sophistication that made it accessible to the Hinduist culture of Java, to the pagan Berbers, or to the Christians of Spain. What we do not know is how it worked, and this is what we have to find out. If this were 1880, or preferably 1850, the way to answer these questions would be fairly easy, as Max van Berchem outlined in an extraordinarily perceptive paragraph in his modestly entitled “Notes d’archéologie arabe.”27 We would learn languages slowly and well, read what our predecessors had done, put it all down on cards and in our heads, travel without haste by boats and on horseback, take thousands of notes and a few intelligently chosen photographs, rework them in the evening, share them with each other in long letters full of references and comments, return to comfortable and properly endowed homes and institutions, and publish it all within six months of completing the writing. But these are the 1980s and other ways must prevail. Beyond the obvious taxonomic task, I see three clear directions. One is the investigation of the historical vocabulary of architecture in all pertinent languages. Words like dar, iwan, maqsurah, or buq‘ah have acquired such a range of meanings over the centuries that only a careful chronological investigation could bring out an approximation of their meaning at any one time; it is highly unlikely that they always had the confused sense they have now. The second direction is the more complicated one of identifying the mental processes and expectations of the environment of these structures as they developed in any one time or place. Of course, we shall never be able to reach the precision that texts and monuments allow for the nineteenth century in Europe, but we should be able to achieve the kind of precision found in Western medieval, Early Christian or Byzantine art. Finally, we must be able to show that, by understanding and explaining Islamic architecture, we are doing more than explaining a specific culture and its inheritance; we are also observing a unique way of creating an architecture that, because of its discrete and unique cultural setting, focuses on the relationship between men and buildings, not between buildings and buildings, not even between architects and buildings. This is a strikingly contemporary effort, and thus I return to my early remark about our role in the contemporary scene. We, as historians, can indeed bring something to the new world created in the Middle East and elsewhere, but only if we are allowed to do it with the secure knowledge of the past. This, as yet, we do not possess. 27

Max van Berchem, “Notes d’archéologie arabe,” Opera Minora I, (Geneva, 1978), pp. 78–9.

Chapter V Art and Architecture and the Qur’an*

The relationship between the revealed scripture of Islam and attitudes towards art and architecture and the practice thereof will be discussed under three headings: 1. Qur’anic references or allusions to art and architecture, including passages later [162] cited with respect to artistic creativity, even if they were not initially so intended; 2. The uses of the Qur’an as a source for citations in the making and decorating of works of art; and 3. The enhancement of the Qur’an itself through art.

Art and architecture in the Qur’an It must be stated at the outset that, with the partial exception of Q27:44, which will be discussed later, the Qur’an does not contain any statement which may be construed as a description of manufactured things or as a doctrinal guide for making or evaluating visually perceptible forms. The world in which the revelation of the Qur’an was made was not one which knew or particularly prized works of art and later hadith – the reports recording the Prophet’s words and deeds – only briefly mention a few fancy textiles owned by the members of the entourage of the Prophet. Furthermore, although hadith do attribute to the Prophet theoretical positions or practical opinions on the making of works of art, none is directly asserted in the Qur’an itself, but only deduced from various passages. Finally, while the Qur’an is quite explicit about such practices as prayer or pilgrimage being specifically restricted to Muslims, it provides no direct or implied definition or even a requirement for a particular locale for the accomplishment of these practices. For all these reasons, the consideration of art and architecture in the Qur’an does not lead to a coherent whole, but to a series of disjointed observations which may be divided into two groups: the direct references to things made or to spaces built; and the indirect implications for the making of things and the design of spaces. *

First published in Jane D. McAuliffe, ed., Encyclopedia of the Qur’an, vol. I (Leiden, 2001), pp. 161–75.

87

88 early islamic art, 650–1100

Direct references There are, first of all, references to categories of manufacture and especially of construction. One rather striking set of examples involves concrete items which are mentioned only once. All of them are described as being in the possession of Solomon, the prophet–king whose patronage for works of art was legendary and whose artisans were usually the no less legendary jinn. In Q34:12 he ordered the making of a fountain of molten brass, a Muslim adaptation of the celebrated brazen sea in Solomon’s temple in Jerusalem as it is described in 2 Kings 25:13 and 1 Chron. 18:8. Then in Q34:13, the jinn manufacture for him maharib, tamathil, jifan of enormous size and qudur, which were anchored down so that they could not easily be removed. The meaning of the word mihrab (sing. of maharib), which will be discussed later, appears in other contexts as well. Jifan – meaning some sort of receptacle, usually translated as “porringer,” a term of sufficiently vague significance to hide our uncertainty as to what was really involved – and qudur, “cookingpots,” are only mentioned in this particular passage. The exact meaning and function of these two items are somewhat mysterious. Timthal, also in the plural, appears again in Q21:52, where it clearly refers to the idols worshiped by the father of Abraham. These idols would have been sculptures of humans or of animals and it is probably sculptures in general rather than idols in particular that must be understood in Q34:13. The association of Solomon with unusual buildings is confirmed by Q27:44, where, in order to test the Queen of Sheba and ultimately to demonstrate his superiority to her, Solomon orders the construction of a sarh covered or paved with slabs of glass (mumarrad min qawarir). Usually translated as “pavilion” or “palace,” the word sarh occurs also in Q28:38 and Q40:36. Both times it is modified by the adjective “high” and refers to a construction ordered by Pharaoh [163]. Since all these passages deal with mythical buildings and because the root of the word implies purity and clarity, the term may reflect the attribute of transparency in a building rather than its form. It would then be a pavilion comparable to the elaborate construction alleged to have existed on top of pre-Islamic Yemeni palaces. Generally speaking, it seems preferable to understand the term as a “constructed space of considerable merit and attractiveness,” without being more specific, though the matter remains open to debate. What is of import here is not the exact meaning of the term but the presence within the qur’anic images of works of art that have not been seen, but only imagined. Further on it will be seen that the story of Solomon and the Queen of Sheba, as told in the Qur’an, has many additional implications for the arts. A second category of qur’anic terms dealing with or applicable to the arts consists of much more ordinary words. There is a series of terms for settlements, such as qarya (Q25:51), usually the term for a city as well as for smaller settlements; madina, a word used only twice (Q28:18, 20), possessing very

art and architecture and the qur’an 89

broad connotations; masakin “[ruined] dwellings” (Q29:38), which often occurs in poetry; and a more abstract term like balad in al-balad al-amin, “the place of security” (Q95:3), which is probably a reference to Mecca. Bayt is the common word for a house and it is supposed to be a place of privacy (Q3:49; 4:100; 24:27–9), a quality which has been sought until today by architects building in what they assume to be an Islamic tradition. The word was used for the dwellings of the wives of the Prophet (Q33:33–4), for whom privacy was an essential criterion, and also for the presumably fancy abode of Zuleika, the wife of Potiphar (Q12:23). When it is mentioned as adorned with gold (Q17:93), it is meant pejoratively as an expression of vainglorious wealth. Dar occurs occasionally (e.g. Q17:5; 59:21) with no clear distinction from bayt except insofar as it implies some broader function as in al-dar al-akhira in Q28:83, indicating “the space of thereafter.” The rather common word qasr (castle, palace) occurs only four times, twice metaphorically, once in a wellknown cliché referring to the destroyed “palaces” of old and once with reference to paradise in a passage which will be examined later. Other terms for something built or at least identified spatially are rarer, like mathwa (dwelling, Q47:19) or masani‘ (buildings, Q26:129). There are a few instances when techniques of construction are indicated, often in a metaphorical way as in Q13:2, where the heavens are depicted as a miraculous, divine creation built without columns. A third category of terms consists of words which, whatever their original meaning, acquired a specifically Muslim connotation at the time of the Prophet or later. The two most important ones are masjid and mihrab. Masjid (place of prostration) occurs twenty-eight times in the Qur’an. In fifteen instances it is modified by al-haram, a reference to the Meccan sanctuary whose pre-Islamic holiness was preserved and transformed by the Muslim revelation, i.e. the Ka‘ba, the holy house (al-bayt al-haram in Q53:97) which Abraham and Ishmael built (Q2:125). It is mentioned as the qibla or direction of prayer (Q2:142–7) and as the aim of the pilgrimage (Q5:96–7). However, nothing is said about its form or about the space around it and there is only a vague reference to the importance of its proper maintenance (Q9:19). Even this action is not as important as professing the faith in all of its truth. In Q17:1, the word is once used for the Meccan sanctuary while in Q17:7 it refers to the Jewish temple in Jerusalem. The word is used a second [164] time in Q17:1 in the expression “the farthest mosque” (al-masjid al-aqsa), the exact identification of which has been the subject of much debate. There is no doubt that, at some point in history and possibly as early as the mid-second/eighth century, it became generally understood as a reference to Jerusalem. This, however, was not the case during the first century after Muhammad’s emigration to Medina (hijra), when it was identified by many as a place in the neighborhood of Mecca or as a symbolic space in a miraculous event. The remaining ten occurrences of masjid do not form a coherent whole except insofar as they all mention a place where God is worshiped (Q7:29). It literally belongs to God (Q72:18, a passage often used in mosque

90

early islamic art, 650–1100

inscriptions, see below) and unbelievers are banned from it (Q9:17). “Those who believe in God and his last day, practice regular prayer and give to charity, and fear none but God must maintain and frequent [the verb ‘amara has a complex range of meanings] the mosques of God” (Q9:18, another passage frequently used in inscriptions). In recounting the story of the seven sleepers of Ephesus, the Qur’an asserts that God built a masjid over them (Q18:21). A most curious and somewhat obscure passage is Q22:40, which contains a list of sanctuaries that would have been destroyed had God not interfered to save them. The list includes sawami‘, biya‘, salawat and masajid, usually – but there are variants – translated as “monasteries [or cloisters], churches, synagogues and mosques.” The first two words are never again used in the Qur’an. The third term, salawat, is the plural of salat, the word commonly used for the Muslim ritual prayer. Here it seems to mean a place rather than the act of prayer. But the sequence itself suggests four different kinds of sacred spaces, probably representing four different religious traditions. If there are four religious groups implied, Islam, Judaism and Christianity are easy to propose – even if one does not quite know which term goes with which system of faith – but what is the fourth religion? It is, in fact, with some skepticism that the word masajid is translated as “mosques” since nowhere else in the Qur’an is the word masjid used alone to be understood correctly as a place of prayer restricted to Muslims. It always means a generally holy space which could be used by Muslims. This verse must, therefore, be connected to some particular event or story whose specific connotations are unknown. In short, the proper conclusion to draw from the evidence is that, while the Qur’an clearly demonstrates the notion of a sacred or sanctified space, it does not identify a specifically Muslim space as a masjid. The only specifically Muslim space mentioned in the Qur’an is the masjid of Mecca and its sacred enclosure. The vagueness of nearly all references to it may explain some of the later problems in actually defining the exact direction of prayer (qibla). Was it toward the city of Mecca, a large enclosure, the Ka‘ba, one of its sides or the black stone in its corner? In short, the word masjid – destined for a long and rich history in Arabic and in many other languages – soon after the death of the Prophet in 11/632 came to mean a special type of building restricted to Muslims. In the Qur’an it appears to have a very broad significance with a very uncertain relationship to exclusively Islamic worship. Matters are almost as complicated with the word mihrab, which also possesses a range of practical and symbolic meanings. It too was destined for a long and distinguished history as the name for the niche indicating the direction of prayer on the wall of all Muslim sanctuaries. The term [165] mihrab also refers to a type of decorative recess found on tombstones, faience panels and rugs. As has been shown in a recent article (N. Khoury, “The mihrab”), the word originally designated elevated structures which had acquired some sort of honorary significance, although the element of height

art and architecture and the qur’an 91

is only clearly present in one qur’anic verse. In Q38:21 the disputants go up to the mihrab where David is. The honorific quality applies to this particular place by inference, as it does in the three instances (Q3:37, 39; 18:11) where the term is used for Zechariah, the servant of God and the father of John the Baptist. When used in the plural maharib (Q34:13) it has usually been interpreted as “places of worship,” but, even if consecrated by tradition, this interpretation does not seem necessary since the other terms listed in this passage – the maharib, tamathil, jifan and qudur (see above) that the jinn manufactured for Solomon – are mostly exemplars of power and wealth rather than of religious, though pagan, needs. Altogether, the exact meaning of this word in the Qur’an seems to be more secular than pious and bears no direct relationship to the word’s later uses in mosques and as a theme of design. While masjid and mihrab became terms to define major elements of Islamic architecture and while other terms dealing with created forms remained consistent and relatively clear (bayt or dar) or rare and fairly obscure (sarh), there is a category of qur’anic references to visually perceived matters which have not been seen, but which nonetheless are held to exist. The numerous accounts of paradise include a great number of references which fall into the category of architecture and planning. These accounts may have had an impact on the design of gardens, most particularly in Mughal India as with the tomb of Akbar in Sikandara near Agra and with the Taj Mahal in Agra itself (see W. Begley and Z. A. Desai, Taj Mahal, although their arguments are not universally accepted). It has also been argued that these qur’anic passages were literally illustrated in the decoration of mosques, most specifically in the early second/eighth-century mosaics of the Great Mosque of Damascus, also known as the Umayyad Mosque (B. Finster, “Die Mosaiken”; C. Brisch, “Observations”) although others (O. Grabar, The formation) have remained more skeptical. Whatever turns out to be appropriate to explain later developments in decoration and in design, an architectural and decorative imagery pervades most of the Qur’an’s vision of paradise and even, at times, of hell. Both paradise and hell are entered through fancy gates, green being the color of the ones for paradise (Q39:72). Rivers and formal – as opposed to natural – gardens abound (Q43:70–3; 44:51; 47:15; 76:12, among many places) in paradise. There are also fountains (Q76:6). In a celebrated passage (Q61:12) gardens are described above underground rivers and beautiful dwellings (masakin in Q61:12 or qusur in Q25:10) are erected in the gardens. In five passages (Q25:75; 29:58; 34:37; 39:20–21), these dwellings are called ghuraf (sing. ghurfa), in all cases but one modified by the adjective “lofty” with apparently the same equation between height and importance as in the instance of the word mihrab. It is difficult to know what was meant or imagined by the term in its singular occurrence in a strange passage (Q25:75), which seems to state that there is only one ghurfa in paradise. Were these meant to be whole architectural establishments

92

early islamic art, 650–1100

or simple pavilions? Inasmuch as we have no means to enter the imaginary world of qur’anic sensitivity, the question cannot be answered in historical terms, although it possibly, as will be seen, may be entered in the fiction of later art. [166] The same difficulty appears when we try to imagine the khiyam, “tents or pavilions” (Q55:72) in which houris (hur) are found, the surur (sing. sarir, one of the several words for “throne,”) with perpetually youthful companions (Q56:15) and especially the throne of God himself. The word for God’s throne is ‘arsh, as in Q40:7, only one of its twenty-nine occurrences in the Qur’an. Most of the time the word is used in the singular and refers to the throne as the place of divine presence. The word ‘arsh is also used once in the story of Solomon and the Queen of Sheba (Q27:41–2). When used in the plural (‘urush, Q2:259; 18:42; 22:45), it refers to some part of a larger architectural composition. Here it is usually translated as “turrets” or “trellis,” which reflects the uncertainty of the translators and commentators about a feature which is always shown as destroyed by divine wrath. One last visually significant qur’anic reference dealing with paradise and with visually perceived matters is that the elect are beautifully dressed (Q35:33; 76:21) and the companions they find there (Q76:15–17) carry vessels (aniya), cups (akwab) and goblets (ka’s) polished to look like crystal or silver (this seems to be the correct interpretation of qawarira min fidda, Q76:16). Their clothes are of silk, the most precious metals are silver and crystal, and polished glass is the model for the expected visual effect. These images are important in suggesting the materials and objects which were considered luxurious in early first/seventh-century Arabia and also serve as inspiration for later Persian painting, where the association between paradise and luxury through expensive clothes and other objects was fully exploited.

Implications for art Quite early passages from the Qur’an came to be used to justify and explain Muslim attitudes toward the arts in general and the representation of living beings in particular. This last topic has been – and will continue to be – the subject of much debate and discussion because it reflects the ever-changing needs and concerns of the prevailing culture and society as much as the actual positions apparent in the Qur’an. The latter is, on the whole, quite clear. Unlike the second commandment of the Old Testament, there is no opposition to art or to representation, just as there is no call for the creation of works of art or of a material culture that would be distinctly Muslim. Thus terms like “iconoclasm” (a call for the destruction of images) or even the German Bilderverbot (forbidding the making of images) are inappropriate to define any part of the message of the Qur’an. The term “aniconism,” meaning simply “the absence of a doctrine or even of much thought about

art and architecture and the qur’an 93

representational imagery,” has found favor among some scholars and is more accurate in reflecting the attitude of the Qur’an. On the other hand, once a broad Muslim culture had been established over vast territories, it was compelled to deal with the rich and varied artistic traditions of the alien cultures it encountered and it sought in the Qur’an either direct answers to its own questions about the validity of artistic activities or, at the very least, references that could lead to such answers. In the absence of direct statements, three kinds of arguments could be, and were, derived from the Qur’an. One is based on a few passages which may be construed as dealing with representations. The “statues” made for Solomon (Q34:12–13) have already been mentioned. A more frequently used passage to uphold a prohibition of images is Q6:74, where Abraham, a far more saintly figure than Solomon in the Islamic tradition, says to his father Azar: “Do you take idols [167] (asnam) as gods? Indeed, I see that you and your people are in manifest error.” This passage must be connected with Q5:90, where idols (ansab) are also mentioned, together with wine and games of chance, as “abominations of Satan’s handiwork.” Both words mean “idols,” which usually have the shape of men or animals, or “statues” of figures that could be used as idols. The two passages are usually seen as expressing an objection to images, but they are more appropriately construed as being in opposition to idols regardless of their shape. A third passage is more specific and, therefore, more pertinent. In Q3:47–9, God says to Mary: “God creates what he wills. When he decrees something, he only says to it ‘be’ and it is.” An example is the case of Jesus, who comes with the following message: “I have come to you with a sign from your lord. I will make for you out of clay the figure of a bird. I will breathe into it and it will become a [real] bird by God’s leave.” Here it is clear that the making of a representation is only meaningful if life is given to that representation. Since the giving of life is reserved to God alone, it is only with his permission that the creation of a three-dimensional and lifelike bird can occur. These few specific passages dealing with representations are not conclusive in themselves, but they served as important points of reference in the later development of the opposition to the making of images. They acquired their particular importance when put next to a second type of argument based less on specific passages than on two themes which pervade the Qur’an: the absolute opposition to idolatry and God’s uniqueness as creator. These two Islamic doctrines were used as arguments against the legitimacy of images as long as images were indeed worshiped and the belief existed that they partook of the spirit of what was represented. It may also be argued that they lost their pertinence once the old equation no longer held. Over the years, much has been written arguing that abstraction, visual distortion and ornamentation occur with such frequency in Islamic art because mainstream Muslim patrons and artists sought to conform to a doctrine that always

94

early islamic art, 650–1100

aimed at the equation of the representation and the represented. According to this view, alternate modes of expression had to be found in order to avoid criticism or even condemnation for vying with God, as a result of such an alleged doctrine. Another doctrine alleged to have been derived from the Qur’an has been that of opposition to luxury – what may be called an ideal of reasonable asceticism in private and public life. Its premise is that art is a luxury, a point which has certainly been argued forcefully by fundamentalist groups and more moderately by moralists down through the centuries. Although common enough in any religious movement with a populist base, as Islam was certainly at the beginning, such a doctrine is difficult to represent as one which has maintained itself on a significant scale throughout time, and even its qur’anic basis is somewhat uncertain. In spite of a number of contrary arguments, on the whole it is difficult to explain the development of an Islamic art through doctrines derived from the Qur’an. This view may only appear to be correct because too many problems have not received the proper attention. Instead, it would seem to have its roots in the complex contingencies of a new ethic encountering the well-developed cultures of the world with their rich visual heritage. There is a need for a careful investigation of the terminology dealing, directly or potentially, with the arts. Words like asnam (idols), ansab (idols), tamathil (statues), sura (shape, Q82:8), hay’a (form, Q3:49; 5:110) are all terms [168] which actually refer to or imply a likeness or copy and suggest some sort of relationship to a previously existing original. The full investigation of the occurrences of these terms in the Qur’an and in early Arabic poetry, as well as later usage both among littérateurs and in technical philosophical thought, may well provide a sketch of the conceptual framework implied by the revelation and give some idea of what the arts may have meant at the time. An interesting beginning in that direction occurred in a recent article by Muhammad Qlaaji published in a Saudi Arabian legal journal which argues, on the basis of a set of qur’anic citations, for the canonical value of ornament in Islamic art. A much more imaginative work by the young French esthetician Valérie Gonzalez (Piège de crystal) demonstrates the deep philosophical problems behind the qur’anic passage mentioned earlier (Q27:44) in which Solomon creates an object, the mysterious sarh, which is supposed to appear real and to be understood as such, without in fact being what it appears to be. The implications are striking not only for Islamic art, but for the very nature of art in general. Comparable statements have been made by twentiethcentury surrealists like René Magritte. Yet such efforts at an interpretation adapted to the needs, tastes and paradigms of our own century are rare. Also they may well go against an interpretative current which asserts that only in its historical truth can the divine message be accepted. Altogether, there is no doubt that the Qur’an will continue to be mined for answers to the aesthetic and social needs of Muslim societies and cultures

art and architecture and the qur’an

95

as they evolve with time. It is also fairly clear, however, that the arts were not a significant concern of the revelation nor did they play a large role in the modes of life prevalent in the Arabian peninsula during the first decades of the first/seventh century. Fancy and elaborate objects were largely absent in the surroundings of Mecca and Medina, and the vision of architecture was limited to the simple Ka‘ba. There was a vision of art and architecture based on the legends of Solomon and memories of the ancient Arabian kingdoms. Ruins in the desert or in the steppe could then, as they do now, be transfigured into mirages of a lost man-made world of awesome proportions. It does not, however, seem that the milieu in which the Qur’an appeared was truly aware of the great artistic traditions of the Mediterranean, Mesopotamia, Iran, India, or even of the Yemen and Ethiopia. Furthermore, the Qur’an contains no trace of the Neoplatonic debates about the nature of art. The emerging universal Muslim culture had to seek in the Qur’an answers to questions which were only later formulated.

Uses of the Qur’an in later art It is well known that script played an important part in the arts of all Muslim lands, regardless of whether that art was primarily secular or religious. Large inscriptions are a common part of the decoration of buildings and many objects have long bands or short cartouches with writing, at times even with imitations of writing. These inscriptions often used to be identified in older catalogs and descriptions as “Koranic” without proper concern for what they really say. It is, of course, true that there is an ornamental or aesthetic value to these inscriptions which is independent of whatever meaning they convey. In order to organize a subject, which heretofore has received little attention, it has been broken into two headings: iconographic uses of the Qur’an and formal uses of qur’anic scripts. [169] iconographic uses The founder of the systematic study of Arabic epigraphy, Max van Berchem, was the first scholar to establish that most formal inscriptions in monumental architecture consist of citations from the Qur’an which bear or may bear a relationship to the function of the buildings on which they were found. He initiated the systematic publication with commentary of all Arabic inscriptions. Beginning in 1931, these were published under the title Matériaux pour un corpus inscriptionum arabicarum as part of the Mémoires of the Institut Français d’Archéologie Orientale in Cairo. M. van Berchem himself published the volumes on Cairo (with a supplementary volume by G. Wiet), Jerusalem and Anatolia, while E. Herzfeld produced the volumes on Aleppo. A similar survey, although less elaborate in its commentaries, was made by Muhammad

96

early islamic art, 650–1100

Husain for the Archaeological Survey of India. In recent years, G. Wiet and M. Hawary, using almost exclusively secondary sources, produced collections of the inscriptions of Mecca. In addition, S. Blair recently collected the inscriptions of pre-Mongol Iran and M. Sharon published those of Palestine. Unfortunately, M. van Berchem adopted the practice of providing only the sura and verse numbers of the qur’anic quotations, usually according to the verse division of the G. Flügel edition, which does not always agree with the now standard Egyptian edition. Therefore, there are problems whenever one tries to identify the exact wording of an inscription. Although most recent publications have abandoned this practice, it is still found in the most important corpus of Arabic epigraphy, the eighteen volumes published so far of the Matériaux pour un corpus inscriptionum arabicarum. A particularly important tool has been derived from all these efforts. Erica Dodd and Shereen Khairallah produced the work The image of the word, the first volume of which contains a list of all of the qur’anic passages cited in inscriptions and where they have been used, thus allowing one to study the frequency of use of certain passages and the temporal or geographical variations in their use. The second volume comprises a series of essays on individual monuments and on questions which grow out of these catalogs, for example why certain inscriptions were placed in different places on different monuments. All of the essays show the influence of a major article written by E. Dodd in 1969 entitled “The image of the word,” outlining the historical and psychological premises behind the existence of an iconography of the Qur’an. She argues that in trying to avoid or even reject the religious imagery of Christianity and paganism, the mainstream of Islamic culture replaced images with words whenever it wished to make some pious, ideological or other point. Within this scheme, the Qur’an was preeminent both because of its sacredness and because most Muslims were familiar with it. Therefore, the viewer appreciates the significance of the selection of the particular passages from the Qur’an and interprets them in accordance with the expectations of the patron. It may be noted that Buddhism and Hinduism do not appear to have been pertinent to the formation of Islamic culture, even though this assertion may be modified by future research. Though never established as a formal doctrine, this “iconography” of the divine word developed quite early in Islamic times under Umayyad rule (r. 41/661–132/750). It might even be proper to associate its appearance with the caliph ‘Abd al-Malik (r. 65/685–86/705), who made the [170] language of administration Arabic, and introduced Arabic inscriptions on the coinage. For the latter, the so-called “mission verse,” “It is [God] who sent his messenger with guidance and the religion of truth to proclaim it over all religion, even though the pagans may detest it” (Q9:33) became the standard formula for thousands and thousands of coins. It is, in fact, rather remarkable how rarely alternate passages were used. Even if there are sixty-one qur’anic citations identified in North African coinage (H. W. Hazard, Numismatic

art and architecture and the qur’an 97

history), many are only pious statements rather than fuller citations and should not be considered as iconographically or semantically significant quotations. The ideological and political assertion of truth made by the passage chosen for coins is easy to explain for a coinage that was used all over the world and which, quite specifically, competed in its inscriptions with Byzantine gold and silver. It is also quite early that the glass weights and stamps used for internal consumption received as decoration “Give just measure and be not among the defrauders” (Q26:181; G. C. Miles, Early Arabic glass weights). This selection demonstrates a considerable and very early sophistication in the manipulation of qur’anic passages for pious as well as practical purposes. The most spectacular early use of qur’anic quotations on a building occurs in the Dome of the Rock (dated 71/691) in Jerusalem, where 240 meters of Umayyad inscriptions running below on either side of the dome octagon are divided into seven unequal sections, each of which begins with the phrase known as the basmala, “In the name of God, the merciful and the compassionate.” The first five sections contain standard proclamations of the Muslim faith. “There is no god but God, one, without associate” is the most common of these. There is also a series of short passages which are probably excerpted from the Qur’an (Q112; 35:36; 17:111; 64:1 combined with 57:2), but which might also be merely pious statements not taken from the Qur’an. The sixth section contains historical data while the seventh, occupying half of the space, repeats a few of the formulas or citations from the first half and then creates a composite of Q4:171–2; 19:33–6 and 3:18–19 with only one minor addition in the middle. This statement exposes the main lines of the Christology of the Qur’an, which makes sense in a city which was at that time a major ecclesiastical and devotional center of Christianity. Other inscriptions in the Dome of the Rock use various combinations of Q2:255 and 2:112 (or 3:1 and 6:106); 3:26; 6:12 and 7:156, 9:33, 2:139 or 3:78 (slightly modified) in order to make clear the eschatological and missionary purpose of the building. Although the matter is still under much discussion, it is possible that the transmission of the qur’anic text used for the decoration of the Dome of the Rock was done orally rather than through written copies of the text. This would seem to account for the fact that many of the inscriptions do not exactly agree in wording with the most common version of the Qur’an in circulation. While the use of the qur’anic passage Q9:33 on coins remained a standard procedure throughout Islamic history and the selection of verses made for the Dome of the Rock remained unique, other citations appear in several early Islamic inscriptions and deserve to be studied in detail. Such is the case with the series, known from later texts, of inscriptions from Mecca and Medina (see RCEA, nos 38, 40, 46–52; G. Wiet and H. Harawy, Matériaux) with a primarily religious content. A curious painted graffito in Medina

98

early islamic art, 650–1100

dated 117/735 contains a long citation dealing with faith (RCEA, no. [171] 30), but its context is unclear and slightly troubling. There are no such concerns about the fragment of an inscription found on a floor mosaic in a private house, probably from the Umayyad period, excavated in Ramallah in Palestine. It contains a fragment of Q7:205, “Do not be among the unheedful,” next to the representation of an arch which may or may not be a mihrab (Rosen-Ayalon, “The first mosaic”). The actual point of the inscription and the reason this particular citation was chosen are still difficult to explain. These early examples all suggest a considerable amount of experimentation in the use of qur’anic citations during the first two centuries of Muslim rule. A certain norm became established from the third/ninth century onward. Epitaphs will almost always contain the Throne Verse Q2:255, sura 112 in its entirety, or both. These verses proclaim the overwhelming and unique power of God. Often these passages are accompanied by Q9:33, with its missionary universality. Mosques will have the throne verse and Q9:18 beginning with “the masajid of God will be visited and maintained by such as believe in God and the last day.” Mihrabs have their own qur’anic iconography with the beautiful Q24:3: “God is the light of the heavens and of the earth, the parable of his light is as if there was a niche [mishkat, another mysterious architectural term] and within it a lamp, the lamp enclosed in glass, the glass like a brilliant star, lit from a blessed tree, an olive neither of the East nor of the West, whose oil is luminous, although fire hardly touches it. Light upon light. God guides whom he wills to his light.” There is little wonder that the decoration of mihrabs and of tombstones often included lamps hanging in a niche and tree-like vegetal ornaments. The history of this iconography is still in its infancy but almost every major monument of Islamic architecture bears, in addition to the common and frequently repeated passages, citations expressing some special function or purpose or references to events which have been mostly forgotten. Examples include the great mosque of Isfahan (O. Grabar, The great mosque); the minarets of Iran (J. Sourdel-Thomine, Deux minarets and S. Blair, Monumental inscriptions); the striking minaret at Jam in Afghanistan (A. Maricq and G. Wiet, Le minaret and J. Moline, “The minaret”); the inscriptions of the small al-Aqmar mosque in Cairo, which expresses Shi‘i aspirations through qur’anic citations (C. Williams, “The cult”); the Ghaznavid palace of Lashkari Bazar in Afghanistan, which is the only building known so far to have used the Solomonic reference of Q27:44 (J. Sourdel-Thomine, Lashkar-i bazar); the Firdaws law school (madrasa) in Aleppo, where a relatively unusual qur’anic passage (Q43:68–72) is found together with an extraordinary mystical text made to look like a qur’anic inscription (Y. Tabbaa, Constructions of power). In the great mausoleums of the Mughal emperors of India (r. 932/ 1526–1274/1858), a wealth of qur’anic inscriptions have allowed some scholars (W. Begley and Z. A. Desai, Taj Mahal) to interpret the buildings themselves

art and architecture and the qur’an 99

in an unusual way as slightly blasphemous attempts to create on earth God’s own paradise. These interpretations have not convinced all historians, but the point still remains that the choice of inscriptions and of qur’anic citations is not accidental and reflects precise concerns on the part of patrons and constitutes a powerful message to the outside world. In general, it is proper to conclude that qur’anic citations were important signifying components of Islamic art, especially of architecture. They became part of the monument and served as guarantors or witnesses of its function and of the reasons [172] for its creation. They could be highly personalized, as in the epitaphs filling graveyards, where endlessly repeated statements are attached to individuals or more general proclamations of power, glory or good deeds projected to the whole of humankind and especially to the faithful. What is, however, less clear is the extent to which these messages were actually understood and absorbed. It is, in part, a matter of evaluating the level of literacy which existed over the centuries or at the time of a building’s construction. It is also a matter of seeking in the chronicles and other sources describing cities and buildings actual discussions of the choice of inscriptions made. These descriptions, however, are surprisingly rare. Often it seems as though this powerful visual instrument, from which modern scholars have derived so many interpretations, was hardly noticed in its own time. Much remains to be done, therefore, in studying the response of a culture to its own practice, if one is to accept the position that the use of the qur’anic word can be equated with the use of images in other religious systems. It is just possible that modern, primarily Western, scholarship misunderstood the meaning of these citations by arbitrarily establishing such an equation. In a fascinating way, the contemporary scene has witnessed rather interesting transformations of this iconographic practice. A recently erected mosque in Tehran, the al-Ghadir Mosque designed by the architect Jahangir Mazlum and completed in 1987, is covered with large written statements, for the most part in glazed or unglazed bricks. Some of these calligraphic panels are indeed placed like icons or images in a church and contain qur’anic passages. Others are pious statements or prayers, for example the ninetynine names of God on the ceiling and the endlessly repeated profession of faith. While the aesthetic success of the structure is debatable, the building itself is impressive for its use of writing so well blended into the fabric of the wall that its legibility is diminished and its value as a written statement difficult to perceive. It is almost as if the difficulty of reading the words contributes to their aesthetic and pious values (M. Falamaki, “al-Ghadir mosque”). Many other contemporary mosques, especially the monumental ones, provide examples of the same difficulties (R. Holod and H. Khan, The mosque). A particularly spectacular use of the Qur’an has been proposed by the architect Basil al-Bayati for the city of Riyadh in Saudi Arabia. He envisioned

100

early islamic art, 650–1100

huge arches in the shape of open books of the Qur’an along the main highway leading into the city as a sort of processional alley greeting the visitor. The project, however, has not been executed. Yet an open book appears as the façade of a mosque designed by the same architect in Aleppo and the Pakistani sculptor Gulgee created a stunning free-standing mihrab in the shape of two leaves from an open Qur’an for the King Faisal Mosque in Islamabad. The effect is striking, if unsettling for those who are used to traditional forms, but it demonstrates the contemporary extension of an iconography taken from the Qur’an to one that uses the book itself as a model. Whether successful or not as works of art, these recent developments clearly indicate that the future will witness further experiments in the use of the Qur’an as a book or as a source of citations, to enhance architecture, especially that of mosques, and to send religious and ideological messages. Thus, shortly after the end of the Cultural Revolution in the primarily Muslim Chinese province of Sinkiang, a modest plaque at the entrance of a refurbished mosque in the small town of Turfan (Tufu in Chinese) on the edge of the Tarim Basin quoted in [173] Arabic script, which presumably was inaccessible to the secret police, Q9:17: “It is not for idolaters to inhabit God’s places of worship (masajid), witnessing unbelief against themselves. Their work has failed and in fire they will forever dwell.” Thus the Qur’an continues to reflect the passions, needs, and aspirations of Muslims everywhere.

The forms of the Qur’an Thanks to important recent studies in the paleography of early Arabic (F. Déroche, Les manuscrits du Coran; Y. Tabbaa, “The transformation”; E. Whelan, “Writing the word”) and to the stunning discovery of some forty thousand parchment pages of early Islamic manuscripts of the Qur’an in the Yemen, we are beginning to understand the evolution of the Arabic script used in manuscripts of the Qur’an in spite of the total absence of properly dated examples before the third/ninth century: the variety of early scripts was already recognized by the bibliographer Ibn al-Nadim (d. c. 385/995) and modern collectors have transformed early pages of what is known in the trade as “Kufic” writing into works of art which frequently fetch high prices on the market. It is much more difficult to decide whether these early manuscripts were indeed meant to have a formal aesthetic value independent of their sacred content. Some of them acquired many forms of ornamental detail, which will be examined in the following section of this chapter. It is also difficult to evaluate whether they or the many styles of angular writing discovered in the San‘a’ trove or elsewhere were meant primarily for the pleasure of the beholder. Matters changed considerably after the introduction of a proportioned script

art and architecture and the qur’an 101

(al-khatt al-mansub) by the ‘Abbasid vizier Ibn Muqla (d. 328/940) in the fourth/tenth century. The establishment of a modular system of writing made it possible to create canons for scripts and variations of these scripts around well-defined norms. As a result, from the time of the small Qur’an of Ibn al-Bawwab (d. 413/1022) in the Chester Beatty Library dated to 391/ 1001 (D. S. Rice, The unique Ibn al-Bawwab) until today, thousands of professional scribes and artists have sought to create variations on the conventional scripts which would attract and please the eyes of buyers. These scripts were not restricted to the text of the Qur’an but, with the major exception of manuscripts of Persian poetry, the holy book was the text on which the most effort was lavished. This is demonstrated by the magnificent Qur’ans of the Mamluks (r. 648/1250–822/1517) in Egypt, Syria and Palestine and those of the Ilkhanids (r. 654/1256–754/1353) in Persia (D. James, Qur’ans and After Timur). It is also for the accurate reading of the qur’anic text that diacritical marks and other identifying signs were carefully integrated into the composition of words and of letters without detracting from the availability of the text. Already with the celebrated “Qarmatian” Qur’an of the fifth/eleventh or sixth/twelfth centuries, the leaves of which are spread all over the world (B. St Laurent, “The identification”), each page became a composed entity to be seen and appreciated in its own right and in which writing and ornament are set in an even balance. A potential conflict between form and content has begun, with the former of greater importance to the ordinary faithful and the latter more important to the collectors of artistic writing or calligraphy.

Enhancement of the Qur’an through art Two aspects of the enhancement of the Qur’an have already been mentioned: the varieties of styles of writing and the addition of small, ornamental, usually abstract or floral, features in the midst of the text itself or in the margins. At some point, large [174] headings were introduced between suras and some of these acquired decorative designs. A group of pages, presumably in the Egyptian National Library in Cairo but not seen since their publication by B. Moritz almost a century ago, uses arcades and other architectural features, perhaps representing or symbolizing places of prayer, as well as geometric and floral designs. Large floral compositions project into the margins and the design of these headings has been compared to the tabulae ansatae of Classical Antiquity. In Mamluk, Ilkhanid or later manuscripts, the cartouches with the titles of each sura are often dramatically separated from the text proper, while in earlier manuscripts they are more closely imbricated with each other. Enhancement could also be provided by variations in size. There are minuscule copies of the Qur’an and gigantic ones, like the Timurid one which requires a special stand to be used and whose pages

102

early islamic art, 650–1100

cannot be read and turned simultaneously. Accounts of calligraphers, especially in Iran, often boast of such feats of marvelous transformations of the holy book, thereby illustrating the major traditional aesthetic value of being “astonishing” (‘ajib). Qur’ans were also honored with fancy and expensive bindings. Especially valued copies were even kept in special boxes. When the Almohad ruler ‘Abd al-Mu’min (r. 524/1130–558/1163) received from the people of Cordoba the copy of the Qur’an which had allegedly belonged to the caliph ‘Uthman (r. 23/644–35/656) and preserved traces of his blood, he hired jewelers, metalworkers, painters and leather-workers to embellish it. In Ottoman times (r. 680/1281–1342/1924) particularly beautiful cabinets were made for keeping pages and manuscripts of the holy book. It is, on the whole, clear and not particularly surprising that many techniques were used to honor manuscripts of the Qur’an by making them more attractive and more exciting than other books and by treating them like precious items, if not literally like works of art. What is more difficult to decide is whether certain styles of writing, certain techniques of binding, certain ways of ornamenting pages and certain motifs were, generally and exclusively, restricted to the Qur’an. The argument may be made for the composition of pages after the fifth/eleventh century and for scripts which, angular or cursive, were written with particular care when used for the holy text. More tentatively, it may be argued that certain types of decorative feature like the marginal ornaments – which also served to signal divisions within the text – were exclusively restricted to the Qur’an. All these hypotheses, however, still await investigation and discussion. The difficulty they present is well illustrated by two hitherto unique pages from the trove in Yemen which were published by H. C. von Bothmer (Architekturbilder) and discussed by O. Grabar (The mediation). They illustrate large architectural ensembles, which have been interpreted as mosques shown in a curious but not unique mix of plans and elevations. Are they really images of mosques? If so, are they representations of specific buildings or evocations of generic types? Could they be illustrations of passages in the Qur’an describing buildings in paradise? There are as yet no firm answers to these questions, but it may be suggested that there was a complex vocabulary of forms more or less restricted to the enhancement of the Qur’an. These forms did indeed create an art.

Bibliography M. Aga-Oglu, “Remarks on the character of Islamic art,” The Art Bulletin, 36 (1954), pp. 175–202. [175] W. Begley, “The myth of the Taj Mahal,” The Art Bulletin, 61 (1978), pp. 7–37. W. Begley and Z. A. Desai, Taj Mahal (Seattle, 1989). M. van Berchem, Matériaux pour un corpus inscriptionem arabicarum, Jérusalem (Cairo, 1922–27).

art and architecture and the qur’an

103

S. Blair, The monumental inscriptions from early Islamic Iran and Transoxiana (Leiden, 1992). H. C. von Bothmer, “Architekturbilder im Koran. Eine Prachthandschrift der Umayyadenzeit aus dem Yemen,” Pantheon, 45 (1987), pp. 4–20. C. Brisch, “Observations on the iconography of the mosaics in the Great Mosque at Damascus,” in P. Soucek, ed., Content and context of visual arts in the Islamic world (University Park, Penn., 1988). Dar al-Athar al-Islamiyya, Masahif San‘a’ (Kuwait, 1985). F. Déroche, The Abbasid tradition. The Nasser D. Khalili collection of Islamic art (Oxford, 1992), i. F. Déroche, “Les écritures coraniques anciennes,” REI, 48 (1980), pp. 207–24. F. Déroche, Les manuscrits du Coran. Catalogue des manuscrits arabes de la Bibliothèque Nationale, 2 vols (Paris, 1983–85). E. Dodd, “The image of the word,” Berytus, 18 (1969), pp. 35–62. E. Dodd and S. Khairallah, The image of the word (Beirut, 1981). R. Ettinghausen, “The character of Islamic art,” in Nabih A. Faris, ed., The Arab heritage (Princeton, 1948), pp. 251–67. R. Ettinghausen, “Arabic epigraphy,” in D. Kouymijian, ed., Near Eastern numismatics, iconography: epigraphy and history in honor of George C. Miles (Beirut, 1974), pp. 297–317. M. Falamaki, “al-Ghadir mosque, Tehran,” Mimar, 29 (1988), pp. 24–9. B. Farès, Essai sur l’esprit de la décoration islamique (Cairo, 1952). B. Finster, “Die Mosaiken der Umayyadenmoschee von Damaskus,” Kunst des Orients, 7 (1970–71), pp. 83–141. V. Gonzalez, Piège de crystal au palais de Solomon (Paris, forthcoming). O. Grabar, The formation of Islamic art (New Haven, 1983). O. Grabar, “The mihrab in the mosque of Cordova,” in A. Papadopoulo, ed., Le mihrab dans l’architecture et la religion musulmanes. Actes du colloque international tenu à Paris en mai 1980 (Leiden, 1988). O. Grabar, The great mosque of Isfahan (New York, 1990). O. Grabar, The mediation of ornament (Princeton, 1993). O. Grabar, The shape of the holy (Princeton, 1996). H. W. Hazard, The numismatic history of North Africa (New York, 1952). E. Herzfeld, Matériaux pour un corpus inscriptionem arabicarum, Alep (Cairo, 1954– 56). R. Holod and H. Khan, The mosque and the modern world (London, 1997). M. A. Husain, Qu’ranic epigraphy in Delhi province (Calcutta, 1936). D. James, Qur’ans of the Mamluks (London, 1988). D. James, After Timur. Qur’ans of the 15th and 16th centuries (New York, 1992). N. Khoury, “The mihrab. From text to form,” IJMES, 30 (1998) with a good bibliography. A. Maricq and G. Wiet, Le minaret de Djam (Paris, 1959). G. C. Miles, Early Arabic glass weights and stamps (New York, 1946). J. Moline, “The minaret of Jam,” Kunst des Orients, 9 (1973–74), pp. 131–48. B. Moritz, Arabic palaeography (Cairo, 1904; repr. Cairo, 1974). I. A. Muhammad, “Muslims and taswir,” MW, 45 (1955), pp. 250–68. M. R. Qlaaji, “Decoration and the position of Islam,” Majallat al-buhuth alfiqhiyya al-mu’asira (Contemporary jurisprudence research journal), 8 (1991).

104

early islamic art, 650–1100

D. van Reenen, “The Bilderverbot, a new survey,” Der Islam, 67 (1990), pp. 27–77 (the latest survey of the subject). D. S. Rice, The unique Ibn al-Bawwab manuscript in the Chester Beatty Library (Dublin, 1955). M. Rosen-Ayalon, “The first mosaic discovered in Ramleh,” Israel Exploration Journal, 26 (1976). M. Sharon, Corpus inscriptionum arabicarum palaestinae (Leiden, 1997f.). J. Sourdel-Thomine, “Deux minarets d’époque seljoukide,” Syria, 30 (1953), pp. 108–36. J. Sourdel-Thomine, Lashkari Bazar IB. Le décor non-figuratif et les inscriptions, vol. 18 of the Mémoires de la délégation archéologique française en Afghanistan (Paris, 1978). B. St Laurent, “The identification of a magnificent Koran manuscript,” in E. Déroche, ed., Les manuscrits du Moyen-Orient (Istanbul, 1989); the whole volume contains many pertinent contributions. Y. Tabbaa, “The transformation of Arabic writing. Qur’anic calligraphy,” Ars Orientalis, 21 (1991). Y. Tabbaa, Constructions of power and piety in medieval Aleppo (University Park, Penn., 1997), ii. E. Whelan, “Writing the word of God,” in Ars Orientalis, 20 (1990), pp. 113–47 with 22 pls. G. Wiet and H. al-Hawary, Matériaux pour un CIA. IV Arabie, ed. N. Elisséeff (Cairo, 1985, rev. edn). C. Williams, “The cult of Alid saints,” Muqarnas, 3 (1985), pp. 39–60.

Part Two Secular Culture under the Umayyads

Chapter VI Sondages à Khirbet el-Minyeh* O. Grabar, J. Perrot, B. Ravani et Myriam Rosen†

Introduction Le site de Khirbet el-Minyeh, à 14 km au nord de Tibériade et à 200 m environ de la rive du lac (Figs 1 et 2), avait été l’objet de nombreuses identifications historiques depuis les premiers rapports de voyageurs et de savants au milieu du XIXème siècle. Une expédition allemande en a commencé la fouille en 1932, mais ses travaux ont été interrompus par la guerre en 1939.1 Ces fouilles ont permis de reconnaître qu’il s’agissait, à Khirbet el-Minyeh, d’un bâtiment approchant le carré (66,4 ¥ 73 ¥ 67 ¥ 72,3 m.) orienté dans la direction des points cardinaux (Fig. 3). L’existence d’une mosquée et d’une inscription remployée au nom d’el-Walid ibn ‘Abd el-Malik, qui régna de 705 à 715, ainsi que certains détails architecturaux et décoratifs, prouvèrent que le bâtiment était un palais omeyyade, comme on en connaît d’autres en Syrie, Palestine et Transjordanie. Les côtés est, sud et nord du palais ont été entièrement fouillés, et les archéologues allemands ont mis au jour sur le côté sud une succession de pièces officielles (mosquée, salle du trône publique, salles privées) décorées de panneaux de marbre sur

* Premièrement publié dans Israel Exploration Journal, 10 (1960), pp. 226–43. † Bezalel Ravani, Inspecteur du Service des Antiquités pour la région de Tibériade, est mort dans des conditions dramatiques au cours de cette campagne de fouilles à laquelle il participait. Portant un intérêt particulier à la période arabe, il rêvait depuis de longues années de voir reprendre la fouille du palais de Khirbet el-Minyeh auprès duquel il vivait et où il avait lui-même effectué des travaux. Son enthousiasme, qui contribua à notre décision de reprendre l’exploration, le dévouement avec lequel il s’employa à faciliter l’organisation matérielle de l’expédition, sa cordialité et sa modestie, le choc que fut pour nous tous sa disparition, attachent son nom et son souvenir à une œuvre à laquelle il s’était consacré avec un si total dévouement. 1 Le dernier rapport sommaire a été publié par O. Puttrich-Reignard, Die fünfte Grabungskampagne auf Chirbet et Minje (Köln, 1939; Palästinahefte des Deutschen Vereins vom Heiligen Lande, Heft 17/20); résumé général des fouilles par A. M. Schneider, “Hirbet el-Minye am See Genesareth,” Annales archéologiques de Syrie, 2 (1952), pp. 23– 45.

107

108

early islamic art, 650–1100

1 Le site de Khirbet elMinyeh en contexte

2 Le site de Khirbet elMinyeh

les murs et de marbre et de belles mosaïques sur le sol.2 L’énorme cour centrale et la partie occidentale du palais n’ont pas été touchées – sauf par quelques tranchées rapides dont le but semblait être l’établissement du plan 2

L’hypothèse de Sauvaget que l’on trouverait un bain dans le coin sud-ouest du palais n’est, pour l’instant, pas à retenir; J. Sauvaget, “Remarques sur les monuments omeyyades,” Journal Asiatique, 231 (1939), p. 37.

sondages à khirbet el-minyeh 109

général de l’édifice – et restaient ensevelies sous près de 6 m. de débris (Fig. 4). [227] C’est afin de pouvoir décider si la partie occidentale du palais justifierait une fouille systématique que, grâce à une subvention du Horace H. Rackham Fund for Research, University of Michigan, une série de sondages ont pu être faits et une partie des murs ont pu être dégagés. Les rapports qui suivent, par Jean Perrot sur la fouille et par Myriam Rosen sur la céramique, décrivent en détail les résultats que nous avons obtenus et permettent de poser les problèmes qu’une fouille ultérieure aura à résoudre.

La fouille Les travaux se sont déroulés du 19 juillet au 10 août avec la collaboration de Mlle Myriam Rosen, assistante au Service des Antiquités, pour l’étude de la poterie, et de Bezalel Ravani, inspecteur des Antiquités pour la région de Tibériade, Jean Liger, architecte, Arieh Volk, photographe, Yehezkiel Menashé, catalogues, Shulamit Havlin et Avraham Yosef, étudiants. Les locaux de l’ancien hospice allemand à Qarey Deshé (Tabgha) ont été gracieusement mis à notre disposition par le ministère de l’Agriculture. Nous avons employé en moyenne une trentaine d’ouvriers venant de Tibériade. Le Service des

3 Vue générale du site

110

4

early islamic art, 650–1100

Plan du palais

Antiquités a contribué pour les deux-tiers au payement de cette maind’œuvre. [228] Le but principal de la fouille était de reconnaître par un sondage la stratigraphie des débris accumulés sur 6 à 7 m. d’épaisseur à l’intérieur et audessus des ruines du palais par les occupations successives et d’essayer de dater ces occupations. Pour ce sondage, nous avons choisi, de préférence à la cour centrale, la moitié nord de l’aile ouest du palais, dans l’espoir qu’il nous serait possible de vérifier en même temps la fonction de cette partie de l’édifice incomplètement [229] fouillée par les archéologues allemands. Leurs tranchées, en effet, encore visibles, n’avaient guère été poussées à plus de deux mètres sous le sommet des murs dans les chambres A et B (v. plan, Fig. 5). Le secteur C non seulement n’avait pas été fouillé mais avait reçu, à l’est, une partie des déblais provenant de la tranchée de dégagement de la façade sur cour de l’aile occidentale. Sous ces déblais qui contenaient des tessons de

sondages à khirbet el-minyeh 111

5 Plan du secteur fouillé et niveau supérieur

poterie vernissée, nous avons mis au jour deux niveaux arabes récents; nous avons dégagé aussi la salle B jusqu’au niveau de sa voûte effondrée, tandis que plusieurs sondages poussés jusqu’à la base des murs (Fig. 6) en A (225), B (222–228–229) et C (227), permettaient d’établir la coupe AB (Fig. 7) et de reconnaître en B, dans la moitié ouest de la salle, une belle mosaïque à dessin géométrique dans un très bon état de conservation. Nous décrirons successivement: (a) (b) (c)

les niveaux supérieurs (secteur C) la coupe ABC la chambre B

[230] (a) les niveaux supérieurs (secteur c) (fig. 5) Presque en surface sont apparus quelques murs, formés le plus souvent d’une seule rangée de pierres; ces murs paraissent une pauvre réfection des habitations du niveau sous-jacent, dont les murs à double parement, mais également mal construits et mal conservés, reposent le plus souvent sur le sommet des murs du palais, hauts de 4 m 50, qu’ils utilisent comme fondations et dont ils suivent approximativement le tracé (109–112–119). Le mobilier ne comprend, avec un rouleau de basalte et un mortier à oreillettes, également en basalte, que des tessons de poterie commune, de types qui ne se sont guère différenciés du XVIème siècle jusqu’à nos jours.

112

6 Plan des salles ABC – niveaux inférieurs

early islamic art, 650–1100

sondages à khirbet el-minyeh 113

7 Coupe des salles ABC

114

early islamic art, 650–1100

8 La voûte de la salle B effondrée sur place

(b) la coupe des salles abc (figs 6 et 7) Elle a été établie par les sondages 222–225 et 227. Sondages 222 Apres avoir enlevé au centre de cette chambre ce qui avait été laissé par les fouilleurs allemands, nous avons dégagé des pierres qui apparurent être les claveaux de la voûte, effondrée en bloc (Fig. 8); au dessous de ce niveau, en 222, la fouille a rencontré successivement (v. coupe, Fig. 7): 7 – une mince couche de terre cendreuse. 8 – un sol de terre battue, avec de nombreux tessons, surmontant une terre grise cendreuse, compacte, d’aspect stratifié; puis une mince couche de sable jaune. 9 – un sol, couvert de cendre noire avec quelques pierres et des morceaux de plaques de marbre, sur une épaisse couche de terre argileuse, brunrouge, compacte; au-dessous, une couche de terre sableuse, meuble, recouvrait 10 – un sol de mosaïque. Mêmes observations en A (tr. 225) pour les couches 7 à 9; mais ici pas de sol de mosaïque; à 0 m 30 au-dessus de la plinthe de basalte sur laquelle s’élève le mur ouest du palais (100), nous avons rencontré quelques dalles

sondages à khirbet el-minyeh 115

formant semelle sur le sol vierge. La base du mur ouest du palais (100) est à quelques centimètres au-dessous de la base du mur 102 qui vient s’appuyer contre lui. Le mur 101 s’appuie lui aussi contre le mur 100 (Fig. 9), mais il correspond à une réfection tardive du palais; il n’appartient pas au plan original comme l’indiquent les plans allemands.3 Son appareil est d’ailleurs bien different de 3

Le plan publié sous le titre “Letzter Plan des Palastes” par Schneider, “Hirbet el-Minye,” 1, Pl. 3; en fait le dernier plan publié par Puttrich-Reignard est plus correct, quoique lui

9 Le mur 101 contre le mur ouest du palais

116

10 La cuve de sarcophage, loc. 228

early islamic art, 650–1100

[231] celui des murs du palais; il est formé de blocs grossièrement équarris de calcaire et de basalte, calés par de petites pierres; sa largeur est de 1 m 40. Les fondations de ce mur s’enfoncent jusqu’au sol 8 mais il n’a pas été possible de distinguer si [233] elles ont été posées sur le sol 8 ou si elles l’ont été dans une tranchée de fondation ouverte à partir du sol 6 ou du sol supérieur, tous deux couverts d’un grand nombre de tessons (Fig. 7). Toutefois, comme le sol 6 est postérieur a l’effondrement de la voûte de la chambre B et comme le découpage du haut de la porte A–B ne peut guère correspondre qu’au sol 8, il est probable que la construction du mur 101 (et la réfection des murs 103–5) correspondent au sol 8. La poterie trouvée sur le sol 8 n’est pas antérieure aux XIIIème–XIVème siècles (v. p. 125); c’est donc probablement aux Mameluks que l’on peut attribuer la réfection du palais. Cette conclusion trouverait encore un appui dans la découverte en 228 (B), sous la voûte effondrée de la cour, d’un grand sarcophage (Fig. 10) évidemment réutilisé (peut-être comme abreuvoir? – la nature du remplissage aussi indique le début du mur comme étant de la même époque que le mur extérieur du palais.

sondages à khirbet el-minyeh

117

[couches 8–9] suggère l’accumulation rapide et régulière de débris organiques; il est possible que la salle ait servi d’écurie). Le sarcophage correspond au sol 8 dans lequel il a pu être à demi enfoncé; nous l’avons trouvé plein de tessons de poterie du même type que ceux trouvés sur le sol 8 en A (Fig. 13). Sondage 227 A 1 m 90 sous le mur 108, nous avons rencontré le sommet d’un mur (123) perpendiculaire au mur 106; ce mur 123, conservé sur 1 m de haut, repose sur une mince couche de cendres; par la hauteur de sa base et par son appareil (gros blocs mal équarris, calés par de petites pierres; Fig. 7), il semble correspondre au mur 101 ou tout au moins à son mur de fondation; mais, en 227, les couches sous-jacentes sont de nature différente de celle des chambres A et B; le remplissage est fait, sur 2 m d’épaisseur, de blocs et de terre meuble; ces débris recouvrent un sol de galets, au même niveau que le sol de mosaïque de la chambre B. Le passage, qui à l’origine mettait les deux chambres en communication, semble avoir eu un seuil assez élevé (0 m 50) si l’on considère que la dalle en forte saillie, à la base du blocage 106a, faisait partie de ce seuil; une marche aurait été nécessaire de part d’autre. Entre B et A le seuil ne mesure que 0 m 38 de hauteur. (c) la chambre b Cette salle mesure à l’origine 12 m 50 de long sur 3 m 50 de large; elle est divisée en deux parties (6 m 50 et 5 m 50) par des pilastres de 0 m 70 dont il ne reste que les arrachements formant, une assise sur deux, une saillie de 0 m 25; ces pilastres devaient supporter un arc doubleau soutenant la voûte en berceau. Cette division transversale de la pièce est marquée sur le sol par un seuil qui limite, a l’est, le sol de mosaïque. [234] Cette mosaïque à décor géométrique (entre-lacs et motifs végétaux, semble-t-il) est du même style que celles dégagées dans l’aile sud du palais; elle fait emploi de pierres de sept couleurs différentes. Elle a été dégagée à la base du sondage 222 sur 1 m2 environ; également en 228, mais son état de conservation était ici moins bon. En 229, après avoir observé que la première assise du mur 102 présente un assez fort empattement et que la plinthe de basalte n’apparaît pas au-dessous, nous avons rencontré une vaste cavité remplie de terre cendreuse dont, pour des raisons de sécurité et de temps, nous n’avons pu achever la fouille. Le mur 102 (Figs 11 et 12) Large de 1 m 05 et haut de 4 m 50, ce mur est formé, sur une plinthe de basalte de 0 m 35, de sept assises, [235] hautes de 0 m 56 (0 m 58 pour l’avant-dernière) avec un couronnement de dalles (0 m 26) formant le sommier de la voûte. L’appareil est celui des plus anciens murs du palais; chaque assise est formée de blocs de longueur variable posés de champ,

118

11 Le mur 102, salle B, et le passage vers la salle A

early islamic art, 650–1100

contenant un bourrage cimenté de pierres et de galets, et alternant avec les deux faces des dalles posées en parpaing liant. La hauteur de la porte vers A était à l’origine de 1 m 70; son découpage vers le haut est postérieur, datant vraisemblablement de la reconstruction du palais. L’extrémité du mur 102, a l’est, a été refaite au même moment. Le mur 106 (Fig. 13) Ce mur n’est conservé que sur six assises; il a beaucoup souffert à l’est où il ne reste plus que l’assise inférieure. On passe en C par une porte monumentale en plein cintre, large de 1 m 65 et haute de 4 m. L’arc du linteau s’appuie sur la 5e assise du mur et sa hauteur correspond aux 6e et 7e assises. Cet arc était encore en place, quoique menaçant ruine [236] (une des pierres du jambage E de la porte est tombée), au moment ou le passage a été muré (106a) à l’aide de blocs calés par de petites pierres (Fig. 14). Ces blocs sont de la même pierre que celle du palais mais de dimensions variables et mal équarris. La fermeture du passage 106a est antérieure a l’effondrement de la voûte; elle pourrait donc dater de la reconstruction mameluk; mais il n’est pas exclu

sondages à khirbet el-minyeh

119

12 Mur 102 – salle B

120

13 Mur 106 – salle B

early islamic art, 650–1100

sondages à khirbet el-minyeh

121

qu’elle soit plus ancienne; seule la fouille et la relation du mur 106a avec les sols 8 et 9 pourra donner la réponse. Lors de la reconstruction du palais, au XIVème siècle, la salle B sera fermée à l’est par un mur de pierres liées par un ciment (115) avec entrée en 118; lorsque la voûte s’effondrera, la partie haute du mur s’abattra à l’intérieur de la chambre, la partie basse se renversant vers l’extérieur. Au-dessus du pan de mur effondré et des débris qui le recouvrent (couches 4 et 5) nous avons observé (3) les restes d’un mur NNO–ESE, qui, par sa hauteur, pourrait correspondre à un niveau intermédiaire entre la reconstruction du XIVème siècle et les niveaux supérieurs déjà décrits; il reste à éclaircir ce point. résumé Les travaux que nous avons mené à Kh. el-Minyeh ont permis de reconnaître: ●

en surface, au-dessus des murs du palais, les restes de deux niveaux d’occupation arabe récents (XVIIème–XIXème siècles);

14 Blocage 106a dans le mur 106

122

●

●

●

early islamic art, 650–1100

au-dessous (salle B) les traces (niveau 3) d’un niveau arabe plus ancien, postérieur à un effondrement de la voûte de la salle. Cet effondrement suit lui-même une réfection générale de cette partie du palais (murs 101, 103, etc.) qui pourrait dater, selon la poterie et la documentation historique (v. plus loin), du XIVème siècle. La porte de communication entre la salle B et la salle C a pu être murée plus anciennement encore lors du premier achèvement du palais. Le fait que le sol de la salle B porte une mosaïque qui ne le cède en rien à celles de la mosquée dans l’aile sud du palais laisse supposer que cette partie de l’aile ouest a, elle aussi, une fonction importante: elle pourrait avoir abrité les appartements; la fosse circulaire remplie de cendres, dans la moitié est de la chambre B, est peut-être en relation avec une installation de bains; mais seule la poursuite de la fouille permettra d’apporter une réponse à ces questions.

[237] La céramique La céramique recueillie a été trouvée en densité inégale, avec des accumulations particulièrement abondantes en 228 et en 225. La poterie vernissée est beaucoup moins fréquente que la poterie non décorée. Aucune pièce entière n’a été retrouvée et ce ne sont par conséquent que des tessons qui seront étudiés. céramique à décor gravé sous couverte vitreuse du type ‘sgraffiti’ Sur la pâte rouge le dessin est gravé au trait sur engobe et peint d’un vert éclaboussé d’autres couleurs, le plus souvent jaune et marron, à effets de marbrures, le tout recouvert d’un vernis. Cette céramique a été utilisée longtemps4 et en différents endroits.5 Nous l’avons retrouvée à Kh. elMinyeh dans les déblais rejetés en surface et provenant du fond d’une tranchée de fouille allemande et parfois aussi dans les quatre couches supérieures6 (v. Fig. 15:1), où leur présence est accidentelle. 4

5

6

En Palestine, elle est signalée par B. Bagatti, I Monumenti di Emmaus el-Qubeibeh e dei Dintorni, Resultato degli scavi e sopralluoghi negli anni 1873, 1887–90, 1900–02, 1940–44 (Gerusalemme, 1947), Photo 57 – No. 6, la datant de la deuxième moitié du XIIème siècle au milieu du XIIIème siècle; à ‘Atlit, v. QDAP, 3 (1933), p. 137, et à Kh. el-Minyeh, citée par Puttrich-Reignard, les deux ayant pour date XIIIème–XIVème siècles, laquelle date est reprise par R. de Vaux et A. M. Stève, Fouilles à Qaryet el-‘Enab-Abu Gôsh, Palestine (Paris, 1950), p. 138. En Iran aux VIIIème–IXème siècles, v. A. U. Pope, A Survey of Persian Art, V (London, 1938), Pl. 569A et B; en Égypte, aux XIéme–XIIéme siècles; en Syrie aux XIIIème– XIVème siècles, v. H. Ingholt, Rapport préliminaire sur la première campagne de fouilles de Hama (Copenhague, 1934), p. 36. Ce tesson provient de la couche 2, dans le secteur C.

sondages à khirbet el-minyeh 123

15

Céramique

124 early islamic art, 650–1100

céramique peinte vernissée Sur une terre rouge les sujets sont presque toujours géométriques ou floraux, peints à grands coups de pinceau assez épais, et le plus souvent blancs. La couverte incolore donne après la cuisson un fond marron sur lequel se détache le décor en ocre jaune ou blanc coquille d’œuf. Parfois le fond est noir et le décor vert. Or, on retrouve cette céramique courante à la période mameluk, sur de nombreux sites en Palestine, notamment à Affoula,7 Emmaus,8 Abou Gosh9 et jusqu’à Hama10 en Syrie. Certains fragments ont permis la reconstitution de formes donnant toute la gamme, allant du bol et des coupes aux plats, avec des bases en disque assez mince, parfois concaves et annulaires, et des bords droits ou retroussés. Cette céramique apparaît en profondeur jusqu’à la couche 6; on pourrait la dater des XIIIème–XIVème siècles. [238] céramique vernissée sans décor Probablement dans la tradition de la céramique parthe-sassanide,11 elle est connue sur tous les sites du Moyen Age au Moyen Orient. De couleur vert uni, elle est, ainsi qu’à ‘Atlit,12 la plus abondante; on la retrouve dans toutes les couches. Quelques fragments d’argile rouge et fine, recouverte uniquement de glaçure incolore, semblent appartenir à des marmites de la période mameluk.13 céramique peinte a décor géométrique C’est une céramique modelée à la main et certains des tessons trouvés à Kh. el-Minyeh portaient à l’intérieur des traces de toile imprimée sur l’argile fraîche. La terre est assez friable, souvent mal cuite, recouverte d’engobe chamois, blanc ou quelquefois rose, avec un décor de méandres concentriques, treillis, zigzags, damiers, spirales. Le décor est peint en rouge-marron, marron ou noir. Ce type de céramique est fréquent en Palestine à Beth-Shean,14 Emmaus,15 Abou Gosh,16 Affoula,17 et hors de 7 8 9 10 11 12 13 14 15 16 17

M. Dothan, “The Excavations at ‘Afula,” ‘Atiqot, 1 (1955), pp. 25ss. Bagatti, I Monumenti, Pl. 26, Photo 56 – Nos 6–10. de Vaux et Stève, Fouilles, Pls F et XVII. P. J. Riis et Vagn Pouisen, Les Verreries et poteries médiévales (Copenhague, 1957), p. 238, Figs 824 et 825. Raymond Koechlin, Les Céramiques musulmanes de Suse au Musée du Louvre (Paris, 1928). QDAP, 3 (1933), p. 139. Ingholt, Rapport préliminaire, p. 241, Figs 840 et 841. G. M. Fitzgerald, Beth Shan Excavations III (Philadelphia, 1931), Pl. XXV, 2. Bagatti, I Monumenti, Photo 58 – No. 4, Photo 59 – Nos 3–5. de Vaux et Stève, Fouilles, p. 133, Pl. F. Dothan, “The Excavations,” pp. 25ss., et Figs 5 et 6.

sondages à khirbet el-minyeh 125

Palestine à Hama.18 Les tessons recueillis (v. Fig. 15:2–11) l’ont été dans la chambre B jusque dans la couche 4. On peut les dater des XIIIème– XIVème siècles. Un grand vase avec ce même décor géométrique, provenant de Kh. el-Minyeh, se trouve au Musée du Service des Antiquités à Jérusalem. céramique non vernissée à décor peigné (v. fig. 15:12–14). Ces tessons, aux parois épaisses, font partie d’un petit groupe trouvé dans la couche inférieure du locus 227. En dehors de la Palestine on peut signaler des analogies avec des vases trouvés à Hira19 en Iraq, datés des VIIIème– IXème siècles et en Afrique, à la Qal‘a des Beni Hammad,20 du XIème siècle. En Palestine, cette céramique a été trouvée à Emmaus21 où elle est datée du milieu du XIIème siècle au milieu du XIIIème siècle. Si l’on accepte le XIIème siècle pour ces fragments, ils nous donnent la date la plus ancienne pour la céramique de Khirbet el-Minyeh. céramique commune sans décor Cette céramique, la plus abondante, est aussi celle qui pose le plus de problèmes. Des fragments de cruches, vases et autres [239] récipients en argile grise ou rouge, décorés de cannelures du type courant, font place à partir de la couche 6 et au dessus à deux catégories bien distinctes. Le premier groupe est caractérisé par des tessons en terre rose, assez fine, côtelés, le bord est retroussé et légèrement montant, le col descend obliquement, rejoignant une panse bulbeuse (v. Fig. 15:15), les bases sont rondes et légèrement concaves, avec un épaississement au centre. Les parois sont de 12 mm d’épaisseur. Les fragments sont nombreux en A, dans la couche 6, mais il a été impossible de reconstituer une pièce complète. Il semble toutefois qu’il s’agit de vases ovoïdes, sans anses, ayant une ouverture d’une dizaine de cm de diamètre. Le deuxième groupe, caractérisé par ses grandes dimensions, est le plus énigmatique. Le bord est arrondi (v. Fig. 15:16–19), présentant parfois un léger épaississement a l’intérieur. L’ouverture de ces vases a base étroite, dépourvus d’anses et ressemblant à de grands pots à fleurs, atteint parfois 0 m 40 de diamètre. La hauteur des parois atteint jusqu’à 0 m 35, leur épaisseur de 2,50 à 3 cm. Les fonds sont tous percés d’un trou au centre. Cette poterie apparaît en A dans la couche 6 et on la suit jusque dans la couche 8. En 228, elle remplissait le sarcophage. Mais la poterie musulmane non décorée n’est pas suffisamment connue pour les

18 19 20 21

Ingholt, Rapport préliminaire, Figs 1020 et 1026. Ars Islamica, 1 (1934), p. 66, Fig. 20. G. Marçais, Les Poteries et faïences de la Qal‘a des Beni Hammad (Constantine, 1913), pp. 5–7, Pl. VII. Bagatti, I Monumenti, Photo 54 – No. 9.

126 early islamic art, 650–1100

périodes tardives. C’est la poursuite de la fouille qui permettra probablement de compléter ces formes et de les dater, sans doute vers les XIIIème–XIVème siècles. les lampes Nous avons trouvé quelques lampes de deux types, mais apparentés, dans la couche 4 (en B) et dans la tranchée 227 (en C). Du premier type, en céramique non vernissée (Fig. 15:21), est une lampe ronde, avec une plaque de couverture bombée percée d’un trou circulaire central et d’un trou luminaire plus petit pour la mèche. Cette lampe serait du XIVème siècle, ainsi qu’une lampe semblable trouvée a Tyr.22 Le deuxième type est du groupe de la céramique mameluk vernissée, semblable au précédent, tout en ayant un récipient central plus réduit. Une protubérance appliquée sur la plaque de prolongement dans l’axe des deux trous, sert d’anse; le trou central est entouré d’un sillon a rebord légèrement saillant (Fig. 15:20). Des lampes de ce type ont été retrouvées en Palestine près d’Ascalon; d’autres sont signalées à Hama.23 En conclusion, on peut constater que les périodes hautes de la céramique musulmane ne sont pas représentées a Khirbet el-Minyeh. La période des Croisades est faiblement représentée et par des fragments non stratifiés. La [240] plus ancienne poterie trouvée dans les couches de remplissage du palais ne parait pas remonter plus haut que le XIIème siècle. C’est des XIIIème–XIVème siècles que date l’ensemble de la céramique trouvée dans le sarcophage et sur le sol 8 en A. Dans les niveaux supérieurs (secteur C) n’a été trouvée qu’une poterie commune de type courant jusqu’à nos jours.

Conclusions et problèmes Quoique incomplets, ces résultats peuvent nous permettre, d’une part, d’esquisser l’histoire du site et, d’autre part, de poser les problèmes qu’une fouille systématique aura, dans un proche avenir, à résoudre. El-Ghuwayr, ou plaine de Gennesar, où se trouve Khirbet el-Minyeh, est une étroite et fertile plaine alluviale dans laquelle se déversent les eaux de trois bassins fluviaux. Depuis la plus haute antiquité la plaine était habitée et servait aussi de voie de passage pour les caravanes et les voyageurs. Mais la plaine de Gennesar était vulnérable aux incursions venant du sud et surtout du nord. Aux époques de faiblesse du pouvoir central, rien n’empêchait les nomades d’outre-Jourdain de venir ravager la plaine sur la rive occidentale du lac. A l’époque romaine et surtout à l’époque de l’empire chrétien, 22 23

Syria, 3 (1922), pp. 6–7, Fig. 3. Ingholt, Rapport préliminaire, p. 53.

sondages à khirbet el-minyeh 127

d’importants centres urbains des deux côtés de la vallée du Jourdain protégeaient la plaine et toute la région était non seulement un centre agricole important, mais servait aussi de voie de passage pour les pèlerins; des églises et des martyria y commémoraient les miracles du Christ. La conquête arabe, comme on sait, au début surtout, changea peu le caractère des lieux qui furent occupés. Sans doute, les grands centres religieux perdirent peu à peu de leur importance, mais les Arabes, surtout sous les Omeyyades (660–750), essayèrent autant qu’il était possible de maintenir la richesse économique des pays conquis. C’est ainsi que Sauvaget déjà24 cherchait à expliquer le grand nombre de palais et de châteaux omeyyades; ils symbolisaient la nouvelle autorité, les nouveaux ‘châtelains’ des grands domaines agricoles des époques romaine et chrétienne. C’est dans ce sens, je crois, qu’il faut interpréter la construction de Khirbet el-Minyeh. Les indications, malheureusement fort limitées, données par des photographies aériennes25 et par une tranchée poussée par les fouilleurs allemands vers le sud-est du château, permettent de suggérer l’existence d’une vaste exploitation agricole autour du palais. Un bain byzantin [241] refait à plusieurs reprises26 et peut-être un khan plus ancien au-dessous ou aux environs du khan ottoman actuel (Fig. 2), nous donneraient pour Minyeh les éléments d’une exploitation pré-islamique reprise ou continuée par les Omeyyades, telles que nous les connaissons à Qasr el-Hayr el-Gharbi, Qasr el-Abyad, et ailleurs. Le palais même est de construction omeyyade, mais, en plusieurs endroits, on peut retrouver des indications qui semblent montrer que dès le début un certain nombre de réfections y furent faites. Dans le cas des sondages effectués en 1959, on peut signaler les curieuses inégalités de niveau entre les chambres B et C. Or, justement, les sources épigraphiques et littéraires nous permettent de montrer qu’au moins deux princes régnants omeyyades ont habité à Minyeh ou, tout au moins, participé à sa construction. D’une part l’inscription trouvée en remploi à la porte d’entrée parle d’el-Walid; de l’autre, el-Qazwini, dans sa Cosmographie appelle notre site ‘ayn Minyah Hisham.27 A une époque où les princes avaient tendance à se construire chacun leur propre palais, une telle continuité s’expliquerait d’autant plus facilement si l’on admet qu’il s’agissait d’un centre d’exploitation aussi bien que d’un château de plaisance. Mais le résultat de la fouille le plus important pour la période omeyyade se trouve dans la découverte d’une mosaïque dans une belle salle voûtée. Dans la tradition des palais omeyyades à un étage, on s’attendrait en effet à trouver les grandes salles d’apparat sur le côté de la cour centrale qui est opposé à l’entrée. Or à Minyeh ces salles avaient été identifiées, sans contredit 24 25 26 27

Sauvaget, “Remarques,” p. 53. V. Puttrich-Reignard, Die fünfte Grabungskampagne, Abb. 1. Ce bain, aujourd’hui disparu, avait été fouillé par le regretté B. Ravani; nous espérons pouvoir publier ses résultats dans une étude plus complète du site. Zakariyah ibn Muhammad el-Qazwini, Kosmographie, ed. Wüstenfeld (Göttingen, 1849), p. 195.

128 early islamic art, 650–1100

possible, sur le côté sud et c’est à leur existence qu’est due la curieuse asymétrie du palais par rapport à l’axe est-ouest. Nous pouvons être assurés maintenant du fait que la partie occidentale du palais possédait, elle aussi, des pièces d’apparat. Qu’étaient-elles? Un bain pourrait être indiqué par certains détails et par l’exemple de Khirbet el-Mefjer, où l’on voit un bain souterrain avec mosaïques en face de l’entrée. Une salle de réception privée ou une salle de banquet seraient justifiées par les cérémonies omeyyades que l’on connaît. Ou bien encore – mais ce n’est là qu’une hypothèse – les groupes de chambres officielles sur le côté sud et sur le côté ouest pourraient appartenir à des décades différentes de l’époque omeyyade, car on connaît bien la passion des princes du Moyen Age de ne pas habiter dans les demeures de leurs pères ou prédécesseurs. La solution de ces problèmes [242] devra attendre le dégagement total de la partie occidentale du château, ce qui permettra aussi de reconstituer avec certitude la façade de cette partie sur la cour. Les siècles qui suivirent la chute des Omeyyades en 750 ne furent pas favorables au développement de la Palestine, ce qui est amplement confirmé par l’épaisseur des niveaux au-dessus de la mosaïque. Il est possible que des occupations temporaires y aient eu lieu, mais ni la documentation archéologique ni les textes ne permettent de les identifier.28 A l’époque des Croisades de nombreux casals se trouvaient dans la région de Tibériade, un des grands fiefs du royaume de Jérusalem, et il est fort vraisemblable qu’un de ces casals ait été dans la plaine de Gennesar. Mais aucun n’y a été identifié et l’échec du système fortifié des Croisés au nord du lac aurait rendu la vie d’une exploitation agricole franque bien aléatoire. Le régime mameluk, après les Croisades, recommence la mise en valeur de la Palestine. Mais la région de Tibériade ne fut pas une des premières à être relevée. Abu-l-Fida (1321) et Ibn Battutah (1355) la décrivent ruinée, et la grande route du Caire à Damas passait ou bien plus à l’est ou bien le long de la côte. Ce n’est que vers la fin du XIVème siècle et au XVème que l’ancienne via maris reprend son importance. Le célèbre voyage de Kaitbay en Syrie (1477) indique Minyeh comme étant une station de grande importance sur la route de Safed au Caire.29 Tibériade, en fait, semble n’avoir pas regagné son importance ancienne et c’est directement de Minyeh à Nazareth que s’acheminent les courriers. C’est à cette deuxième période mameluk que nous pouvons donc attribuer la réutilisation du palais omeyyade. Une partie de la porte d’entrée, notre mur 101, et nombre d’autres indices, dont une fouille complète pourra peut-être donner l’ensemble, doivent donc être attribués à cette époque. Pour la céramique, qui justifie d’ailleurs une date 28 29

Seul el-Maqdisi mentionne de nombreux villages autour du lac, v. G. Le Strange, Palestine under the Moslems (London, 1890), pp. 334 ff. V. Textes et bibliographies dans R. Hartmann, “Die Strasse von Damaskus nach Kairo,” Zeitschrift der Deutschen Morgenländischen Gesellschaft, 64 (1916), pp. 665–702.

sondages à khirbet el-minyeh 129

mameluk, on peut donc assumer qu’elle est postérieure à la première moitié du XIVème siècle. Un certain nombre de problèmes se posent cependant. Il semblerait (mais seul un dégagement de la cour centrale le prouverait) que la reconstruction mameluk se fit sur le plan du palais omeyyade, c’est-à-dire en utilisant les murs anciens, en divisant parfois les pièces anciennes, et en maintenant la cour centrale. Mais en même temps, il apparaît que la destination des constructions fut différente [243]. Aucune publication n’existe des niveaux supérieurs trouvés par les fouilleurs allemands,30 et nos hypothèses devront donc être basées sur les découvertes de l’an passé. Celles-ci nous permettent de suggérer que les grandes pièces omeyyades furent divisées en pièces plus petites (ainsi mur 101), que les murs omeyyades furent altérés ou réparés pour les besoins nouveaux (murs 102 et 106) et que l’une des pièces (B) servit d’abreuvoir pour les chevaux (hypothèse la plus vraisemblable pour la cuve de “sarcophage”). Ne pourrait-on donc conclure que le khan mameluk était bien l’ancien palais omeyyade et que le khan ottoman, dont les ruines subsistent toujours au nord-ouest du palais, aurait été bâti plus tard, peutêtre par la suite d’un changement du tracé de la route? A une époque que l’on ne peut encore déterminer, l’établissement tomba en ruines (chute de la haute voûte omeyyade trouvée dans la chambre B). Plus tard le site fut réoccupé, mais les deux réoccupations postérieures que l’on peut noter deviennent plus ou moins indépendantes de la configuration des constructions anciennes. De khan, le palais omeyyade devint un tell. Telles sont les grandes lignes du développement du site que nos travaux de quelques semaines permettent de proposer. On y trouve sans doute certaines hypothèses que la fouille future confirmera ou modifiera, mais surtout on y trouve posés trois problèmes essentiels que la fouille devra résoudre. Tout d’abord, il s’agirait de dégager le plan exact de la partie occidentale du palais et ainsi d’expliquer l’existence apparente de pièces d’apparat sur deux côtés d’un même palais. En même temps un dégagement de la cour centrale s’impose, afin de mettre en valeur ce qui reste du palais et de reconstituer sa façade sur la cour. Ensuite une série de sondages autour du palais, dans les environs du khan ottoman et peut-être dans d’autres endroits que l’on pourrait détecter à l’aide de photographies aériennes, permettraient de donner une idée de la nature et du développement des installations qui se trouvaient dans la région de Khirbet el-Minyeh. Enfin, ces mêmes fouilles pourraient amener à écrire d’une manière plus précise et plus certaine l’histoire du site à l’époque musulmane et, ainsi, contribueraient à notre connaissance de la civilisation musulmane aussi bien que de la riche histoire de la terre palestinienne. 30

Il nous a été malheureusement impossible jusqu’à présent de pouvoir aller examiner les documents laissés par les fouilleurs allemands, mais le Professeur Ernst Kühnel m’a aimablement assuré de sa coopération dans l’étude des notes et rapports qui sont toujours à Berlin.

Chapter VII Umayyad “Palace” and the ‘Abbasid “Revolution”*

Since the time of Wellhausen and Lammens, almost everything that has been written on the dynastic change from Umayyads to ‘Abbasids has been primarily based on the written sources already available, for the most part, to early twentieth-century scholarship. It is true, of course, that much in our understanding of the events and of their religious, ethnic and social components has been refined to a far greater degree than is apparent in earlier works. One of the consequences of the refinement – as it appears either in B. Lewis’s assessment of the ‘Abbasids in the second edition of the Encyclopedia of Islam or in H. A. R. Gibb’s essay on “The evolution of Government in Early Islam”1 – has been to underplay the “revolutionary” character of the change in dynasty in the sense that it involved massively the whole Iranian world and to consider it instead as the “Third Civil War” of early Islam, at the time of no greater or lesser significance than the ‘AliMu‘awiyah conflict three generations earlier, although portentous of considerable later changes. All the arguments and discussions so far have centered on written sources, whose partiality has often been pointed out and whose specific reliability in precise instances of fact as well as interpretation is hence often difficult to assess. It is only comparatively rarely that the considerable and impartial archaeological documentation has been brought into the [6] discussion. The two monuments from the Umayyad period which have been known for many decades – Qusayr ‘Amrah, the single bath with extraordinary paintings in the wilderness of Transjordan, and Mshatta, the unfinished but grandly planned palace a few miles south of ‘Amman – have usually been taken to be symbols of an ill-explained Umayyad megalomania and romantic attachment to the desert; the recently discovered palaces of Khirbat al-Mafjar in the valley of the Jordan near Jericho, Qasr al-Hayr al-Gharbi along the road from Damascus to Palmyra, Khirbat al-Minyeh, north of Tiberias along the shores of the sea of Galilee, Qasr al-Hayr al-Sharqi, some sixty miles northeast of Palmyra, in an area hardly visited any longer by caravans, to name but * First published in Studia Islamica, 18 (1963), pp. 5–18. 1 Studia Islamica, 4 (1955).

131

132 early islamic art, 650–1100

those few with extraordinary architectural characteristics or in which particularly remarkable discoveries of sculpture, paintings, or mosaics were made, have not yet fully penetrated, as historical documents rather than artistic oddities, into the consciousness of most historians who have dealt with the early Islamic period. As far as the early ‘Abbasid period is concerned, the Round City of Baghdad, the mysterious palace at Ukhaydir, and illdefined remains at Raqqah have often been described, but rarely brought to bear on the wider historical and cultural problems of the period. To this general rule there is one major exception. In a series of articles2 J. Sauvaget challenged many of the traditional views on Umayyad secular constructions, but his untimely death prevented the completion of the promised Châteaux [7] Omeyyades de Syrie, in which presumably a new interpretation of the archaeological evidence as well as a more flexible explanation of the historical development which led from Umayyads to ‘Abbasids would have been proposed. Regardless, however, of what Sauvaget’s final conclusions would have been, the hypotheses and explanations he proposed in the late 1930s can for the most part be considered as established and it is only in comparative details – such as the precise identification of Hisham’s Rusafah with Qasr alHayr al-Sharqi or the location of the bath of Khirbat al-Minyeh – that they have been modified by subsequent research. If we take then Sauvaget’s remarks on Umayyad “palaces” and relate them to several other investigations carried out either in earlier periods of the archaeology of Syria and Palestine or in entirely different areas of the Middle East, a series of conclusions – or at least working hypotheses – suggest themselves, not only for our understanding of the monuments themselves, but, as we shall try to show, also for the wider question of the ‘Abbasid “revolution”. The final validity of these hypotheses will have to be tested through more thorough archaeological surveys than have been made so far and through new studies of the information provided by texts; but the hypotheses themselves may be valuable in showing ways in which archaeological materials supplement literary evidence and at times replace it altogether. This has, of course, been demonstrated already by S. P. Tolstov in Khorezm, where the whole social and political development of an area in the early Middle Ages can be suggested almost exclusively on 2

The most important one is “Observations sur les monuments omeyyades, 1: Châteaux de Syrie,” Journal Asiatique, 231 (1939), although two other articles published in the same year contain important notes for our purposes, “Les ruines omeyyades du Djebel Seis,” Syria, 20 (1939), and “Les ruines omeyyades de ’Andjar,” Bulletin du Musée de Beyrouth, 3 (1939). The essential bibliography on Umayyad palaces will be found there, but one should add D. Schlumberger, “Les fouilles de Qasr el-Heir Gharbi,” Syria, 20 (1939); R. W. Hamilton, Khirbat al-Mafjar (Oxford, 1959); K. Otto-Dorn, “Grabung in Umayyadischen Rusafeah,” Ars Orientalis, 2 (1957); M. Chéhab, “Excavations at Anjarr,” Ars Orientalis, 5 (1963). The author is preparing an annotated list of Umayyad settlements known archaeologically and through texts, which should be completed in the near future. The work was never completed, but see J. Sauvaget, “Châteaux umayyades de Syrie,” Revue des Etudes Islamiques, 35 (1967).

umayyad “palace” and the ‘abbasid “revolution” 133

archaeological grounds, or in Turkmenistan, where recent surveys permit a reconstruction of the social and physical setting of pre-Seljuq Khorasan. The single most important point made by Sauvaget was that, in almost all instances (Qusayr ‘Amrah being the only possible but by no means certain exception), clearly identifiable Umayyad sites, as we may recognize them through the existence of a mosque or through analogy with undoubted Umayyad buildings, were not simply places for high princely living and entertainment, but also centers for agricultural exploitation. The extensive irrigation system, the canals, cisterns and aqueducts [8] which surrounded many of them did not merely serve to supply the palace with water, but also to provide for fields and meadows. As agricultural exploitations established most of the time in areas not easily provided with water, these centers of early Islamic occupancy are related to a large series of farms, manors, at times even small villages with a few houses, a cistern, enclosures for animals and equipment, and the ubiquitous and fairly diversified water systems, which literally cover the whole area between desert and sown extending northwest or west of a line drawn from the Euphrates around Raqqah to Damascus through Palmyra and then through the mountains of the Hawran down to the gulf of ‘Aqabah in an ill-defined fashion approximated by a straight line from the present oasis of Azraq to the modern town of ‘Aqabah. The precise date at which these settlements were developed is not known. Some probably have historical peculiarities of their own, but the investigations of N. Glueck and Fr. Franck in Transjordan and the wadi ’Arabah, as well as the air surveys of A. Poidebard in Syria,3 have clearly shown that it is generally in Hellenistic and especially Roman times that, in spite of considerable natural difficulties, the agricultural growth of this area took place. The development was hardly slowed down by the creation of the Christian empire; and, from the fourth century on not only do most of the earlier settlements acquire churches, but many new settlements appear either in the areas of the previous settlements or in entirely new places such as certain parts of the Negeb and the superb wilderness of Khilwah in the southeastern part of the present Kingdom of Transjordan. Some of these new settlements were purely religious centers in which anchorites and cenobites lived. These must have had a limited economic significance, but in all at least a rudimentary program of conservation of water and subsistence agriculture has to be assumed. [9] The second point made by Sauvaget is that in the large majority of cases clearly identifiable Umayyad buildings or settlements were intimately related to existing older buildings or settlements. This relation was especially 3

N. Glueck, Explorations in Eastern Palestine, in Annual of the American Schools of Oriental Research, 14 (1933–4) and following. Fr. Frank, “Aus der Araba”, Zeitschrift der Deutsche Palästinavereins, 57 (1934). A. Poidebard, La trace de Rome dans le désert de Syrie (Paris, 1934).

134

early islamic art, 650–1100

close in so far as water was concerned, as at Qasr al-Hayr al-Gharbi, where a superb dam and system of canalization from Palmyrene times served the Umayyad palace, at Khirbat al-Mafjar, where a Roman aqueduct brought water to the palace, or in the very important but still unmapped site of Humaymah in southern Transjordan, where the ‘Abbasid plot was hatched and where all the presently visible means of water supply are of pre-Islamic origin. The point can be generalized to say that there are practically no Umayyad “palaces” which fail to show stones or foundations from an earlier origin and that the physical or ecological infrastructure which made the Muslim settlement possible existed before Islam had conquered the area. In fact in three of the four instances of excavations carried out on Umayyad buildings – Khirbat al-Minyeh, Qasr al-Hayr al-Gharbi, Khirbat al-Mafjar (the exception is Rusafah, where the excavator actually sought Hisham’s palace) – the excavators were not looking for and did not at first expect to find Umayyad buildings, and Mshatta was considered for a long time to be a pre-Islamic palace. From these remarks two conclusions emerge: that Umayyad settlements were directly related to the economic organization of the preIslamic world, which the Muslims had to maintain in working order lest their establishments perish; and that apparently the choice made in. so many instances by the Muslims to settle in the difficult area between desert and sown, which requires extensive waterworks, did not derive from a presumed atavistic craving for the desert but from reasons which involved a consciousness of the value of land ownership and land use. In order to elucidate these reasons, it is essential to understand in some detail the exact functions of these lands in pre-Islamic times and to explain why it is that, in so many instances, the land on which the Umayyads built required a highly artificial system of water supplies and was on the edge of the desert. It is true, of course, that not all Umayyad settlements and palaces [10] are in these areas. Khirbat al-Mafjar, Khirbat al-Minyeh and ‘Anjarr are all in fertile valleys with sufficient water, requiring usually only systems for distribution and not for storage. But there is no doubt that at least the first two of these palaces (little is known so far about the third one)4 were surrounded by agricultural exploitations in the past just as they are today and that the reasons which will be given below for the existence of Umayyad settlements in less hospitable lands actually apply in fertile areas as well, except for the fact that in fertile areas the circumstances of the conquest did not permit the systematization of a practice which was probably common in more inclement regions. A partial answer to the question of the exact functions of these lands in pre-Islamic times is now provided by G. Tchalenko’s admirable volumes, Villages Antiques de la Syrie du Nord.5 In describing with painstaking details 4 5

The excavations began only two years ago. Three vols, Paris, 1953–8.

umayyad “palace” and the ‘abbasid “revolution” 135

the remaining buildings and the evolution of settlements in a small area of northwestern Syria which has been since then almost totally in disuse, Tchalenko has reached the following conclusions of significance to Umayyad problems. First, the villages of northern Syria during the period of prosperity – i.e. before the seventh century – were predominantly concerned with the monoculture of olive trees and the accessory industry of oil-pressing; second, these agricultural and semi-industrial enterprises were almost entirely sponsored from the great Hellenized cities where enough capital existed to finance the undertakings and to amortize their losses; third, much of the final product – oil – was used for export beyond the metropolis of Antioch into the great cities of Greece and Anatolia; finally, the decadence of the region in the seventh century is entirely due to human rather than natural changes. These human factors were partly the destructions which accompanied the Persian wars and the obliteration of Antioch, but especially the loss of outside markets after the elimination of the great city, and after the Muslim conquest. [11] The region between Aleppo and Antioch which has been studied by Tchalenko has a number of peculiarities, some of a geographical character, others human, and its position in Umayyad times practically on the frontier with Byzantium certainly explains the almost total lack of any early Islamic settlements there. But as one begins to move southwards, in the areas of Palmyra, Damascus, the Hawran, and especially the long depression of the Ghawr and the plateau east of the depression – the very area of most identified Umayyad settlements on or near pre-Islamic sites – the question is raised whether the pre-Islamic sites of this area, whose number is astounding (in a comparatively limited area of Transjordan Fathers Saller and Bagatti have counted 141 Christian sites)6 can be and should be explained in the same fashion as the villages of northern Syria. No final answer can be given as long as these sites have not been subjected to the type of careful and complete analysis made by Tchalenko in the north; however, there are two features in which the north Syrian villages are comparable to the villages of Syria and Transjordan. On the one hand, the extensive techniques of water use and conservation – canals, cisterns, aqueducts – which are still visible in most places can only be explained through external sponsorship, either by government or private capital, because the investments involved were considerable and losses in bad years would have to be covered. On the other hand, in spite of the evident Umayyad concern in maintaining and utilizing the infrastructure provided in these locations, the sites did not survive the fall of the first Muslim dynasty. This decay cannot be attributed to ‘Abbasid destructions, which mostly involved centers, such as Rusafah, that had an administrative and political meaning, and it is interesting to note that the

6

S. J. Saller and B. Bagatti, The Town of Nebo (Jerusalem, 1949).

136

early islamic art, 650–1100

completely waterless city of Rusafah survived for a considerable time after it had stopped being an Umayyad capital, because it had a commercial significance. The implication of these points is that, just as in the north, the prosperity of these centers was [12] related to some aspect of Christian Syria and Palestine which could not be permanently carried over into Islamic times. The possibility of a monoculture of wine rather than oil is not excluded, although it could only be proved by extensive archaeological surveys (see however below for a possible argument from a literary source); but an explanation that these establishments bear some relation to the tremendous growth of Christian sanctuaries and pilgrimage routes can perhaps be suggested. The numerous and large establishments needed to house and feed the multitude of travelers and settlers in the Holy Land certainly required a considerable infrastructure of agricultural and animal husbandry, even though differences existed between such dissimilar areas as, for example, the Hawran and the region east and south of ‘Amman. The suggested character of these agricultural settlements and their likely relationship to social, economic, political and religious institutions of preIslamic Syria and Palestine may serve to explain also why it is that so many Muslims settled there in Umayyad times. Two major sources – the treaties signed by the Muslim conquerors with the Syrian cities and the fiscal rescript of ‘Umar II – have in recent years been studied afresh7 and, together with works on the economy of the first Muslim century by Løkkegaard and Salih al-‘Ali, have established several conclusions of significance to our purposes. It appears, first of all, that the large and rich landowners usually emigrated before or shortly after the conquest; therefore if our conclusions with respect to the ownership of the irrigated lands in the semi-desert and mountainous areas are acceptable, it follows that such lands would, as abandoned property, fall into the category of booty. Land falling into this category has been shown by Dennett to have been taken by the Umayyad princes for their own use or for redistribution among their clients or allies; in some of the earliest occurrences, as at Humaymah, even ‘Alid sympathizers could end up by [13] settling on these lands. Since these properties possessed an infrastructure for successful operation and presumably a labor force, it is quite likely that the new owners sought to maintain them in working condition and to use their revenues. Some of the owners may have used the revenues to build estates in the Hijaz, while others settled there and built their châteaux or manors and thereby changed the character of some of the settlements by introducing into them not only mosques, but, more interestingly, baths, richly decorated palaces and houses, in short, an element of “gracious” living, imitated from urban centers, which had hitherto been lacking. This hypothesis may also explain the well-documented phenomenon 7

H. A. R. Gibb, “The fiscal rescript of ‘Umar II,” Arabica, 2 (1955); D. Dennett, Conversion and Poll-tax in early Islam (Cambridge, 1950).

umayyad “palace” and the ‘abbasid “revolution” 137

of the peregrinations of caliphs from one place to another; rather than an expression of peripatetic restlessness of princes of nomadic origin, these were visits to productive enterprises. Furthermore, this hypothesis permits us to solve some of the more puzzling features of Umayyad secular architecture: the extraordinary number of buildings erected in a short period and the rather considerable variations in quality which exist between the presumed splendor of a Mshatta and the rusticity of other sites, such as Qasr alHallabat where an old Roman fort was rebuilt by the Umayyads; or even the apparent lack of major Muslim buildings in a town like Humaymah, which was demonstrably a major early Muslim center. The reason for these divergences is that the land appropriated by booty was not kept by caliphs alone but distributed among many individuals whose taste and use of their possessions varied considerably. One last point may be suggested, although more tentatively: could it be that the inclusion in ‘Umar II’s rescript of a whole paragraph on prohibition of wine in the midst of complex problems of taxes and land tenure derived from the considerable production of wine in the very type of settlement we have tried to define? These settlements could not survive, even though the Umayyad princes and their clients had at their disposal the capital needed for their exploitation. For, as Tchalenko has shown in the instance of north Syria and as has been suggested for the rest of Syria and Transjordan, the economic usefulness of these settlements depended on their relationship to markets [14] outside the Muslim world, or on wine, which lost much of its significance, or on Christian religious establishments which weakened considerably. The Umayyads may have maintained the system for a while on a more or less artificial basis, but as it became meaningless economically, it declined rapidly; and, after the fall of the dynasty, only a few places survived in which supporters of the Umayyad regime barely managed to live, and often revolted against ‘Abbasid governors.8 Eventually the whole system was obliterated, to be revived briefly in the Ayyubid period and then in the middle of the twentieth century, when a new influx of population created anew a need for the development of Syria and Transjordan as agriculturally productive areas. Three conclusions are suggested by this explanation of the evidence provided by Umayyad “palaces”. First, the location of most Umayyad manors clearly shows that the economic system of pre-Islamic Syria and Palestine was, on the whole, maintained until the fall of the dynasty. Second, the palaces themselves appear as a highly original type of architecture; in function they are comparable to Roman villas as they were described by Pliny or to the manorial houses – at times also imitating fortified structures – which grew in the nineteenth century in much of agricultural Europe or in the American South. As an intrusion of sophisticated living in the agrarian 8

An incident of that order will be described and discussed in an article written for the memorial volume being prepared for L. A. Mayer.

138

early islamic art, 650–1100

world of the older tradition, they actually illustrate a considerable change in at least one aspect of the use of the land: in some instances, as at Qasr alHayr al-Sharqi, agriculture was replaced by gardens and game preserves, a change of sufficient notoriety to have impressed the Byzantine chronicler Theophanes, who, in an often quoted passage, speaks of Umayyad “paradises”. But spectacular though the latter may have been, they did not constitute the majority of cases. The explorations of N. Glueck have shown that in a large proportion of the sites with “Nabatean” or Christian settlements, [15] remains of early Islamic times are also found although without monumental architecture. For a few estates transformed into true palaces, the majority remained under the new order what they had been before the conquest. The third conclusion requires some elaboration. The pattern of early Muslim settlement in Syria and Palestine was in strong contrast to what occurred in the Jazirah and in Iraq. The Jazirah had been, for the most part, a frontier area between Byzantium and Iran, covered with forts along major routes and strategic points rather than with agricultural settlements. Iraq’s economic development had declined considerably under the late Sasanian princes. This character of the two provinces in pre-Islamic times is confirmed by the tremendous economic endeavors of the Umayyad governors of Iraq and of the Umayyads themselves in the Jazirah, at least along the Middle Euphrates. In Iraq canals were rebuilt, the land was divided, and the systematic agricultural exploitation of the area, which had such momentous results in the ninth century, began with all the abuses known for later times.9 In the Jazirah Hisham undertook considerable work along the Euphrates.10 The tribal warfare between Qays and Taghlib constantly involved presumably new settlements along the Balikh and the Khabur.11 Maslamah b. ‘Abd alMalik built canals in Iraq and his name is to be connected with the Hisn Maslamah mentioned by Ibn Khurdadhbih near Raqqah, one of the many properties identified by Gabrieli as owned by the great Umayyad hero.12 One may even wonder whether the choice of [16] Rusafah as a capital by Hisham should not be explained by the growing consciousness – often discussed – in the mind of that prince of the economic significance of the Mesopotamian valley in the new Muslim world. The classical explanation of the change from Damascus as the result of the plague and of fear of big 9

10

11 12

H. Lammens, “Études sur le règne de Mo‘awia,” Mémoires de la Faculté Orientale, Univ. St. Joseph, vol. IV (1907), pp. 123 ff.; L. Caetani, Annali dell’Islam (Milan, 1905 ff.), vol. 1, pp. 353 ff., based mostly on Baladhuri. Hisham’s work is tied up with the complicated problem of Zaytunah, on which most of the texts have been gathered in the last instance by Sauvaget, “Remarques”, pp. 1 ff., although his conclusions are not necessarily definitive. See now O. Grabar, “Le nom ancien de Qasr al-Hayr Sharqi,” Revue des Etudes Islamiques, 38 (1970). Baladhuri, Ansab al-Ashraf, vol. V (Jerusalem, 1936), pp. 316, 326, 314, 321. There is no doubt that the eventual publication of this work will add a great deal of new information. F. Gabrieli, “L’eroe omayyade Maslamah”, Rendic. Accad. Lincei, ser. VIII, vol. 5 (1950), p. 33. See works by D. Whitcome, T. Leisten and K. Haase.

umayyad “palace” and the ‘abbasid “revolution” 139

cities13 should be relegated to a romantic view of the Umayyads, to which Tabari as well as twentieth-century historians have succumbed, even though for different reasons. Altogether, then, literary sources suggest that a considerable transformation of the land took place, giving a great impetus to possible agricultural activities. But the instances derived from textual sources have not been verified on the spot. The actually visible remains in the Syrian Jazirah are very numerous but have usually been considered from the point of view of a more ancient history. They have never been subjected to any sort of systematic archaeological investigation since the famous trip taken before World War I by F. Sarre and E. Herzfeld. Their account sheds little light on our problem, and it is only Miss Bell who has pointed out the existence on the eastern bank of the Euphrates south of the Khabur of a series of ruins which can typologically be related to Umayyad constructions in Syria.14 For Iraq the important survey carried out by the University of Chicago’s Oriental Institute is still unpublished. It is clear that archaeological investigations in all these areas must be undertaken, but the evidence of the texts is sufficient to lead to the general conclusion that, whereas in Syria and Palestine the Umayyads took over an existing form of land use, in the western Jazirah and in Iraq they created – or, more correctly, recreated – a new system, related, no doubt, to the older practices of these areas. Because the old system had decayed or been abandoned, the new Umayyad creations were more adequately suited to their own actual needs and possibilities [17]. It is the Umayyads, then, who in reality fostered the growth of the area which became the center of ‘Abbasid rule. That the principal emotional attachment of most Umayyad princes was to Syria remains true; it is there that the members of the ruling family built their palaces, while the known Iraqi palaces of Wasit and Kufah were urban centers of authority rather than places for personal enjoyment. This point may explain, for instance, why the fully excavated palace at Kufah was so poor in ornament and decoration, when Syrian palaces are so rich. But on this score our information is still very uncertain, as neither archaeological nor literary sources have yet been properly searched and coordinated for a detailed analysis of Iraqi topography in the first Islamic century. While it is doubtful that there were many castles and manors built there by Umayyad princes, it would be interesting to know for whom exactly the many qusur of Iraq mentioned by Tabari were built and what they looked like. The very fact that this information is more difficult to obtain in Iraq is a sign of the difference between Umayyad Iraq and Syria, for in the former the Umayyad foundations, by becoming the nucleus around and over which the later growth of the province took place, were much more often destroyed and 13 14

Tabari, Annales, ed. M. de Goeje and others (Leiden, 1879 and ff.), part II, pp. 1737–8. G. Bell, From Amurath to Amurath (London, 1924), pp. 77 ff. Of particular significance are also the explorations of A. Musil.

140

early islamic art, 650–1100

built over than in Syria, where they symbolize a basically pre-Islamic use of the land by man. These remarks, in spite of their tentative character as far as Iraq is concerned, confirm the suggestion that the traditional conception of a complete revolution in Islamic history as power shifted from the Umayyads to the ‘Abbasids needs to be revised, if not truly abandoned. Both the growing wealth of Iraq and the coming decadence of Syria are phenomena of the Umayyad period, and it is one of the paradoxes of the Umayyad century that the center of political and imperial power and of a rather exuberant princely life artificially remained in an area which continued a by then meaningless economy, while the region to whose development these same princes and their powerful viceroys devoted so much money and energy, and from which they derived much of their wealth, became both a center of political opposition to them and, at the same time, of greater contributions to the elaboration of an Islamic culture than were Syria and Palestine. [18] In suggesting the addition of this paradox to the paradoxes of Umayyad government discussed by H. A. R. Gibb15 in an earlier issue of this journal, I do not want to imply that economic factors – as they are suggested by an analysis of archaeological remains – were the only or even the main causes of the ‘Abbasid takeover. It is rather that, as one tries to understand the peculiarities of Umayyad secular monuments and the purpose they fulfilled, their meaning and their implications only become clear if instead of considering them merely as creations of the late seventh and early eighth centuries, one contrasts them with what they replaced and with what followed them. As one proceeds along these lines, however, it becomes apparent that Umayyad culture – best known though it may be through its Syrian forms – involved the whole of the Fertile Crescent and that, whereas it appropriated and used, but misunderstood, the features of life, artistic or economic, available in the western part of the Crescent, it created a more viable and newer basis for further developments in the eastern part. As seen from the point of view of the Umayyad palaces in deserted parts of Syria and Transjordan, the ‘Abbasid revolution appears as the natural result of the conscious efforts of the masters of the palaces.

15

H. A. R. Gibb, “The evolution of government in early Islam,” Studia Islamica, 4 (1955), pp. 1–17.

Chapter VIII Notes sur les Cérémonies Umayyades*

Deux conclusions historiques dominent l’article fondamental de Dominique Sourdel sur les cérémonies ‘abbasides: le premier siècle de la nouvelle dynastie ne modifia que dans le détail les pratiques umayyades et ce n’est qu’à partir de la fondation de Samarra et surtout de l’avènement des Buyides que l’on découvre une spécialisation accrue des locaux cérémoniels et les effets de l’iranisation du califat.1 Quelles que soient les données nouvelles qui aient été découvertes depuis 1960 – et il n’y en a guère beaucoup – , ni l’une ni l’autre de ces conclusions ne prête à controverse, tout au plus à des précisions ou discussions de détail. Par contre les cérémonies umayyades qui auraient servi de modèle à celles des premiers ‘Abbasides n’ont jusqu’à présent fait l’objet que de remarques passagères, à l’exception de quelques pages de Sauvaget autour de l’explication qu’il proposait de la mosquée umayyade.2 Les remarques qui suivent ont pour but d’apporter quelques précisions sur les cérémonies umayyades. D’une part je voudrais attirer l’attention sur une série de textes qui n’ont pas été étudiés du point de vue des cérémonies. Et d’autre part il n’est peut-être pas inutile de revoir ces cérémonies dans le cadre plus général de ce que les études beaucoup plus poussées des cours romaine, byzantine, ou baroque ont appelé le “jeu” et l’“expression” de la vie aulique.3 Deux séries de documents existent pour reconstituer les cérémonies umayyades. L’archéologie a mis à jour un groupe d’établissements princiers – Khirbat al-Mafjar, Qasr al-Hayr Gharbi, Mshatta en particulier [52]4 – qui fournissent un cadre architectural pour les cérémonies umayyades. Malheureusement ces monuments, sauf peut-être Mshatta, sont des * Premièrement publié dans Studies in Memory of Gaston Wiet (Jerusalem, 1977), pp. 51– 60. 1 D. Sourdel, “Questions de Cérémonial ‘Abbaside,” Revue des Etudes Islamiques (1960). 2 J. Sauvaget, La Mosquée Omeyyade de Médine (Paris, 1947), pp. 129–34. 3 A. Grabar, “Pseudo-Codinos et les Cérémonies de la cour byzantine an XIVème siècle,” Art et Société à Byzance sous les Paléologues (Venise, 1971); O. Treitinger, Die ostromische Kaiser- und Reichsidee (Jena, 1938); divers ouvrages d’Ernst Kantorowicz, en particulier The King’s Two Bodies (Princeton, 1957). 4 K. A. C. Creswell, Early Muslim Architecture (Oxford, 1969); O. Grabar, The Formation of Islamic Art (New Haven, 1973), pp. 141 et suiv.

141

142 early islamic art, 650–1100

monuments privés dont le patronage n’était pas nécessairement celui des princes régnants. Leur valeur documentaire est donc diminuée en partie. Même s’il est vraisemblable que tel motif ou ensemble – grandes entrées, salles d’apparat – se trouvait également dans les palais officiels de Damas ou de Rusafah, nous ne possédons pas pour les Umayyades d’équivalent du Grand Palais de Constantinople, du Palatin ou même de Samarra. Il suffit de penser au contraste entre la villa d’Hadrien à Tivoli et les palais de Rome ou bien entre Spalato et Piazza Armerina pour apercevoir les limites du document architectural umayyade pour reconstituer le cadre de leurs cérémonies. Il n’est de même pas facile d’utiliser les représentations peintes ou sculptées pour imaginer les costumes ou les gestes officiels. Rien ne prouve en effet que les décorateurs des palais utilisaient comme modèle la vie qui les entourait plutôt que des images d’origines byzantines ou iraniennes.5 A Qusayr ‘Amrah un prince est habillé à la byzantine, tandis qu’à Khirbat al-Mafjar un modèle persan fut utilisé.6 Le contraste est frappant avec les monnaies qui, en tant que documents officiels, cherchèrent parfois à donner un caractère umayyade aux effigies du prince.7 Ces efforts furent cependant abandonnés, à quelques exceptions près, après la réforme d’Abd al-Malik. Nous verrons plus loin que la documentation archéologique n’est pas sans apporter des renseignements importants, mais a priori il ne s’agit pas d’équivalents immédiats des palais impériaux de Rome ou de Byzance ni d’images précieuses comme les sculptures de certaines colonnes de la fin de l’Antiquité ou bien les mosaïques et ivoires byzantins. Les textes sont également incomplets. Il n’y a pas d’équivalent d’un Livre des Cérémonies à la manière de Constantin Porphyrogènete ni même d’ouvrages comme le Rusum dar al-khilafah d’al-Sabi, le Kitab al-Taj du pseudo-Jahiz ou les extraits de Maqrizi qui ont permis de reconstituer les cérémonies fatimides.8 Il s’agit donc d’extraire des renseignements [53] fournis par les historiens, Tabari et surtout Baladhuri, et les littérateurs comme l’auteur du Kitab alAghani des incidents qui permettraient d’imaginer les cérémonies umayyades. Ces textes, comme les monuments, sont limités par des objectifs qui leur sont propres. Le Kitab al-Aghani parle de poètes et il n’est pas nécessaire qu’une réception fréquentée par un poète ait été du même type que celle d’un ambassadeur. Comme al-Walid b. Yazid fut poète, trop de documents parlent de ce personnage à bien des égards exceptionnels. En général toute étude des

5 6 7 8

Le problème est en fait beaucoup plus compliqué. Je compte y revenir à l’occasion d’un étude en préparation sur les fresques de Qusayr ‘Amrah. Les exemples sont décrits et discutés dans O. Grabar, “Islamic Art and Byzantium,” Dumbarton Oaks Papers, 18 (1964). Grabar, Formation of Islamic Art, figs 15–19, et les travaux de G. C. Miles cités, pp. 222– 3. L’ouvrage d’al-Sabi a été édité par M. Awad (Baghdad, 1964); M. Canard, “Le cérémonial fatimite et le cérémonial byzantin,” Byzantion, 21 (1951); le K. al-Taj a été traduit par C. Pellat, Le Livre de la Couronne (Paris, 1954).

notes sur les cérémonies umayyades

143

cérémonies umayyades se trouve dans l’impossibilité de prouver que les exemples qu’elle utilise aient vraiment été typiques et non pas accidentels. Dans le cas des Umayyades on pourrait, il est vrai, faire un recensement complet de tous les événements qui ont eu lieu à la cour et qui ont été préservés dans les textes et en déduire d’une manière bien plus précise la vie de la cour,9 mais c’est là une entreprise de grande haleine qui n’a pas encore été commencée. Ceci dit, il n’en reste pas moins que même une sélection de textes permet de définir sinon un système précis de cérémonies, du moins une ambiance cérémonielle que l’on peut décrire à partir de deux types d’événement caractéristique, l’audience et les divertissements.

1. Audience L’audience princière est l’événement le plus répétitif de la vie de cour. Peu de textes en décrivent le cadre architectural, à l’encontre de ce que l’on sait des ‘Abbasides.10 Les expressions les plus courantes sont majlis, sans qu’il soit souvent clair s’il s’agit d’un lieu spécifique ou bien de l’activité elle-même, ou bien bayt et dar. Bayt apparaît le plus fréquemment et semble signifier un ensemble architectural incomplet, une partie seulement d’un monument plus grand, comme dans le cas du “grand bayt” d’al-Hajjaj11 ou du petit dans la Rusafah de Hisham.12 Dar est un terme plus général et dans l’expression laysa laku bidar13 il s’agit en fait d’un banissement de la cour. Le prince est assis sur un tronc, sarir. Il est parfois tentant d’imaginer [54] le sarir comme on le voit sur des peintures et sculptures à Qusayr ‘Amrah ou à Qasr al-Hayr Gharbi,14 c’est à dire un siège pour une seule personne, décoré de sculptures ou d’incrustations, et peut-être surmonté d’un baldaquin. Un texte qui parle de colonnes (‘amud) sur un sarir pourrait confirmer l’existence de tels supports, mais l’occasion est funéraire et l’objet en question semble plutôt être un cénotaphe.15 En général, le sarir umayyade était plutôt une espèce de lit et le prince invitait fréquemment un hôte de marque ou un favori à s’installer auprès de lui sur le sarir.16 Au début de l’époque umayyade le kursi, siège plus simple et sans dos, pouvait aussi être réservé au prince, 9 10 11 12 13 14 15 16

Un exemple intéressant de ce genre d’études systématiques est D. R. Hill, The Termination of Hostilities (London, 1971). Sourdel, “Questions,” pp. 129 et suiv. Abu al-Faraj al-Isfahani, Kitab al-Aghani (Bulaq, 1868), XVIII, p. 130. Aghani XX, p. 80. Tabari, Akhbar II (Leyden, 1876–1901), p. 1468. Grabar, Formation of Islamic Art, fig. 80; et Dumbarton Oaks Papers, 18 (1964), figs 22–3. Aghani I, p. 19. Mas‘udi, Muruj al-Dhahab (Paris, 1861–77), V, pp. 40, 151 et suiv.; Tabari II, p. 190; Baladhuri, Ansab al-Ashraf, ed. W. Ahlwardt, Anonyme Arabische Chronik (Greisswald, 1883), p. 310; Aghani VII, p. 76; le calife apparaît seul dans Tabari II, pp. 1636, 1819–20; Aghani XV, p. 78.

144

early islamic art, 650–1100

mais il devint rapidement un siège moins restrictif, parfois, comme dans le cas d’un kursi appartenant à une fille d’Abd al-Malik, de grande beauté.17 Sa caractéristique est d’être mobile et le cas d’un kursi-litière utilisé par Mukhtar à des fins religieuses semble indiquer une origine pré-islamique pour ce meuble.18 Cependant, même s’il n’était pas réservé au prince, le kursi n’était pas un objet très commun au premier siècle de l’Islam. Il est beaucoup plus difficile d’établir si les Umayyades avaient des vêtements de cérémonie, comme en auront les ‘Abbasides et les Fatimides. Le texte célèbre décrivant l’inauguration de la mosquée de Médine mentionne le fait que le calife portait une tunique en laine (darra‘ah) et avait une qalansuwah sur la tête.19 Mais il s’agit là d’un événement exceptionnel. Abd al-Malik, lors d’une réception, portait une tunique à longues manches (jubbah) et une cape (rida).20 Al-Walid II change de costume suivant les occasions21 et Sulayman était connu pour ses fastes vestimentaires.22 Mais il n’est pas vraiment possible de déduire de ces quelques textes l’existence d’un costume de cérémonie constant et uniforme. On peut être un peu plus précis au suject du couvre-chef. Je n’ai pas trouvé de texte prouvant l’existence d’un taj umayyade et l’hypothèse de [55] Schlumberger qui voyait une couronne califale dans une représentation de prince a Qasr al-Hayr Gharbi23 ne me semble pas acceptable. Il est par contre clair que le turban et la qalansuwah étaient des couvre-chefs officiels. Le turban semble bien avoir été le symbole du prince en tant qu’imam, car il fut surtout porté a la mosquée.24 La qalansuwah semble avoir été une nouveauté sous les Umayyades. Peut-être représentée sous sa forme courte sur une monnaie célèbre,25 la qalansuwah avait des dimensions variables qui semblent avoir eu un sens restrictif. Al-Hajjaj ne permettait pas qu’une qalansuwah semblable à la sienne fut portée en sa présence26 et Richard Ettinghausen a récemment proposé d’expliquer une chaîne de pierre se terminant par un long objet conique trouvée à Khirbat al-Mafjar comme étant la qalansuwah longue du prince, suspendue à la manière des couronnes sassanides.27 La qalansuwah était vraisemblablement le symbole du prince en tant qu’amir.

17 18 19 20 21 22 23 24 25 26 27

Muruj V, pp. 73 et suiv.; Aghani VII, pp. 189–90. Tabari II, p. 703; Muruj V, p. 185. Tabari II, 1232–4; Sauvaget, Mosquée Omeyyade de Médine, p. 152. Ansab (Ahlwardt), p. 195. Aghani II, pp. 65–6; III, pp. 98–9; VI, pp. 65 et 141. Muruj V, pp. 400–403. D. Schlumberger, “Qasr al-Hayr,” Syria, 20 (1939), p. 353; cf. J. Sauvaget, “Remarques sur les Monuments Omeyyades II,” Journal Asiatique, 232 (1944), p. 48. Aghani XIX, p. 60; Sauvaget, Mosquée Omeyyade de Médine, p. 143. G. C. Miles, “Mihrab and Anazah,” Archaeologica Orientalia in Memoriam E. Herzfeld (Locust Valley, 1952), fig. 3. K. al-Taj, tr. Pellat, p. 76. R. Ettinghausen, From Byzantium to Sasanian Iran and the Islamic World (Leyden, 1972), pp. 17 et suiv.

notes sur les cérémonies umayyades

145

Le prince se tenait derrière un rideau et al-Walid II avait un préposé au rideau, un sahib al-sitr.28 Le chambellan, hajib, existait depuis l’époque de Mu‘awiyah et ses fonctions furent si nombreuses qu’Abd al-Malik le décrit à son frère comme étant “ton visage et ta langue.”29 En dehors du chambellan, cependant, les textes apportent peu de renseignements sur les fonctionnaires de la cour. Les eunuques ont vraisemblablement été introduits par Mu‘awiyah,30 mais, sauf dans quelques cas exceptionnels comme celui du poète Turayh banni de la cour par les intrigues d’un eunuque,31 leur rôle semble avoir été limité à la vie familiale du prince.32 Il y eut aussi quelques restrictions de droits et de devoirs parmi les compagnons immédiats du prince. Ainsi seuls les nobles (ashraf) purent s’asseoir avec Abd al-Malik33 et de nombreuses audiences furent réservées [56] aux “membres de la famille du calife, ses mawali, ses poètes, et ceux qui avaient à traiter une affaire particulière.”34 On chercha à interdire l’entrée des gens après le début de l’audience,35 mais dans presque tous les cas il s’agissait d’éviter le désordre plutôt que de former une hiérarchie officielle de courtisans et de visiteurs. En fait les princes aimaient inviter les gens à s’asseoir près d’eux36 et une atmosphère détendue régnait souvent pendant les audiences. Pourtant, lorsque Khalid al-Qasri, infirme, demanda à être porté dans sa kursi à l’audience d’al-Walid II, il lui fut répondu que “personne ne peut se faire porter vers la présence du calife”37 et lors du même événement il est fait mention de gardes autour du prince et de tables, vraisemblablement couvertes de nourriture, comme le suggèrent des textes parlant de Mu‘awiyah et d’Abd al-Malik.38 Dès l’époque de ce dernier, il y avait une classe de julasa’, de courtisans, attachés d’une manière permanente à l’entourage du prince.39 Nous trouvons dans ces références presque tous les thèmes caractéristiques des cérémonies ‘abbasides: fonctionnaires spécialisés, arrangement précis des locaux, vêtements limités au prince, courtisans, restrictions dans l’admission des personnes, un vague sens de l’étiquette. Mais ces thèmes ne s’y retrouvent qu’à l’état embryonnaire et deux éléments des cérémonies postérieures manquent entièrement. D’une part aucun aspect des audiences umayyades n’est constant et l’on trouve une exception pour toutes les occasions. D’autre part l’ordre de 28 29 30 31 32 33 34 35 36 37 38 39

Aghani XX, p. 72; Muruj VI, p. 8. Ansab (Ahlwardt), pp. 171–5. Ibn al-Faqih, dans M. de Goeje, Bibliotheca Geographorum Arabicorum, V (Leyden, 1885), p. 109. Aghani IV, p. 78. Muruj V, p. 274. Ansab (Ahlwardt), p. 175. Aghani IV, p. 81. Aghani VI, p. 131; IX, p. 169; X, p. 80. Aghani IX, p. 169. Tabari II, pp. 1819–20. Ansah (Ahlwardt), p. 177; Aghani VIII, p. 107. Ansah (Ahlwardt), p. 259.

146

early islamic art, 650–1100

grandeur et de précision des gestes et des symboles n’est nullement comparable. Les textes umayyades parlent peu de symboles concrets et même le sceau et le sceptre ne sont mentionnés que lors de l’investiture d’un prince; les prosternations sont rares et l’appareil militaire est presqu’entièrement absent. Il semble clair que les cérémonies de l’audience, telles que nous les connaissons et avec les restrictions imposées par la nature des sources, n’étaient pas sous les Umayyades l’exercice et l’expression de majesté qu’elles deviendront sous les ‘Abbasides. Et pourtant sur un point capital les Umayyades donnèrent aux cérémonies de l’audience un ton qui passera chez leurs successeurs. Il s’agit de la manière par laquelle le prince se rend visible et accessible à ses sujets et à ses hôtes. Deux traditions pouvaient [57] servir de modèle aux Umayyades. L’une, la tradition iranienne, avait un prince immobile qui, à un moment cérémonial précis, se montrait au peuple.40 Cette manière existait certes à Byzance mais les empereurs chrétiens avaient également développé un rituel complexe de la procession de l’empereur qui allait d’un endroit à l’autre et se faisait acclamer. Les Fatimides à un certain moment adopteront aussi cette deuxième manière. Les Umayyades dès le début n’adoptèrent que la première qui prendra sous les ‘Abbasides un éclat particulièrement brillant. S’agit-il simplement d’une élaboration des habitudes tribales de la péninsule arabique ou bien les débuts d’une adoption par les princes musulmans du système cohérent des cérémonies iraniennes? L’examen des divertissements princiers permet en partie de répondre à cette question, ou tout au moins de formuler une hypothèse.

2. Divertissement D’après le Kitab al-Aghani et Baladhuri, Yazid b. Mu‘awiyah fut le premier des califes à avoir introduit des instruments de musique dans l’Islam, à recevoir des chanteurs, à se conduire d’une manière désordonnée, à boire du vin, à chasser, à fréquenter les femmes; et à assister à des combats de chiens et de coqs.41 Sans nous attacher à l’invraisemblance d’une accusation de cet ordre qui appartient au genre littéraire spécial des awa’il ni aux critiques souvent discutables du Père Lammens au sujet des préjugés des chroniqueurs du Moyen Age,42 il n’en reste pas moins que toutes sortes d’activités qui entrent dans la catégorie générale de divertissements firent partie de la vie des princes umayyades et que ces divertissements ont pu apparaître plus tard comme étant des innovations déplorables. Il est vrai que dans certains cas comme celui d’alWalid II et du chanteur Ma‘bad,43 les divertissements avaient un caractère

40 41 42 43

A. Christensen, L’Iran sous les Sassanides (Copenhague, 1944), p. 397 et suiv. Aghani XVI, pp. 71–2; Ansab IVB, ed. M. Schloessinger (Jerusalem, 1938), p. 1. H. Lammens, “Le Califat de Yazid I,” Mélanges de la Faculté Orientale, Université St Joseph, 5 (1911), pp. 446 et suiv. Aghani I, pp. 26–7; voir aussi III, pp. 98–9.

notes sur les cérémonies umayyades 147

d’orgie mais, même dans l’histoire du bassin rempli de vin autour duquel s’étaient installés le prince et le chanteur, on retrouve la présence de motifs cérémoniaux typiques: rideau séparant le prince des autres, chambellan introduisant le poète, changements de vêtements, parfums, présence d’un auditoire de courtisans. Il est en fait assez facile de multiplier les exemples concernant al-Walid II qui montreraient que la plupart de ses rencontres avec les poètes ou les chanteurs des deux sexes avaient [58] un caractère officiel.44 Mais il ne fut pas le seul. Après un repas, Abd al-Malik s’installe dans son majlis avec ses courtisans devant lui et d’autres personnes en deux rangées; un poète apparaît qui demande la permission de réciter ses oeuvres, ce qui lui fut accordé.45 Des poètes et des musiciens furent attachés à la cour de Hisham.46 Dans le cas des poètes et des musiciens on pourrait simplement expliquer les choses comme étant une transformation du poète de tribu en poète de cour. Mais l’hypothèse d’une innovation radicale n’est nullement exclue d’autant plus qu’elle seule semble expliquer la pratique de la boisson à la cour des Umayyades. Un messager envoyé au Khorasan raconte au gouverneur de la province qu’il vit al-Walid II en train de boire du vin sur son trône.47 A une autre occasion, dès qu’il fut installé sur son trône, al-Walid fit venir des boissons et des compagnes.48 Un jour il s’enivra et commença à servir du vin à l’audience, ce que le chroniqueur considéra comme un manquement à l’étiquette.49 Abd al-Malik boit et offre à boire à un poète qui le loua lors d’une audience officielle.50 Le même Abd al-Malik reçoit les gens avec une esclave à ses pieds et une autre debout devant lui.51 A l’annonce de son élection au califat, alors qu’il était félicité par les gens présents, al-Walid fait venir des musiciens et boit du vin devant tout le monde.52 Un texte tardif signale que les Umayyades s’enivraient d’une manière régulière et officielle.53 Il est peu vraisemblable qu’ils aient adopté d’une façon systématique cette pratique sassanide, mais le caractère officiel de l’acte de boire semble indiqué par plusieurs textes. Un chambellan à qui al-Walid offre à boire refuse en alléguant que sa fonction ne lui en donne pas le droit.54 Un rideau sépare parfois le prince buvant des autres.55 Lorsqu’Abd al-Malik reçoit al-Hajjaj, il tient dans ses mains une coupe en or.56 44 45 46 47 48 49 50 51 52 53 54 55 56

Aghani I, p. 119; II, pp. 65–6. Aghani XII, p. 36. Aghani III, p. 125; Tabari II, p. 1733. Tabari II, p. 1766. Aghani IV, p. 116. Aghani IV, pp. 122–3; voir aussi VI, p. 130. Ansab (Ahlwardt), p. 154. Ibid., p. 185. Aghani VI, p. 109. A. von Kremer, Kulturgeschichte der Chalifen I (Vienne, 1875–7), p. 149. Aghani VI, p. 130. Tai, tr. p. 59. Aghani VII, p. 67.

148

early islamic art, 650–1100

Présence de poètes et de musiciens aux réceptions officielles, princes [59] buvant pendant les audiences, ce sont là des motifs qui deviendront les clichés de la représentation du prince dans l’art du Moyen Age musulman. Or, si les textes sur les Umayyades confirment bien que telle était la pratique de l’époque, les images umayyades de princes ne les montrent pas ainsi, sauf si nous considérons comme umayyades un petit groupe de plats en argent dits post-sassanides où presque tous ces motifs se retrouvent.57 C’est que la représentation de divertissements royaux de ce type n’était pas dans la lignée principale du langage artistique utilisé par les Umayyades. Les cérémonies par contre ont été directement adoptées de la pratique sassanide, où les “maîtres de plaisir” faisaient partie de l’entourage royal et l’enivrement du prince était une manière d’exprimer son caractère surhumain.58 L’auteur anonyme du Kitab al-Taj a correctement compris cette filiation directe entre les Sassanides et les Umayyades lorsqu’il explique que les cérémonies de Bahram Gur ne se retrouvèrent pas dans les cérémonies des rois avant le règne de Yazid b. Abd al-Malik.59 Le nom des princes importe peu, quoiqu’il y eut beaucoup à dire sur le fait que Bahram Gur ait symbolisé la dynastie iranienne. L’essentiel est que l’iranisation des mœurs princières commença dès l’époque umayyade et s’explique en partie par l’énorme prestige de tout ce qui venait de l’Orient dès l’époque d’al-Walid et en partie par l’effort fait par les Umayyades à se différencier des empereurs byzantins.60 Deux conclusions ou tout au moins hypothèses dérivent de cette rapide et nécessairement incomplète série d’exemples. La première est que l’iranisation des mœurs princières a commencé dès l’époque umayyade. Il n’est certes pas toujours aisé de distinguer dans les incidents rapportés par le Kitab alAghani ou par Baladhuri l’influence directe et consciente des pratiques sassanides d’un luxe de nouveau riche se permettant toutes sortes de fantaisies. Dans bien des cas il ne s’agissait que de modifications d’habitudes arabes plus anciennes. Et pourtant l’impression principale qui se dégage de la lecture de ces textes est celle d’un système en voie de création; dans la formation de ce système, en particulier dans l’utilisation de divertissements en tant qu’activités exclusivement princières, l’Orient iranien plutôt que la Méditerranée servit de modèle. La deuxième conclusion est que ces cérémonies n’avaient pas de caractère systématique. Il ne semble pas que l’on puisse parler de vêtements [60] exclusivement réservés au prince, comme l’indique Tabari pour les ‘Abbasides.61 Ni les audiences ni les divertissements n’avaient encore de caractère rituel, comme l’auront les processions fatimides. Le contexte politique des 57 58 59 60 61

Ces plats ont été souvent reproduits; voir A. U. Pope et Ph. Ackerman, A Survey of Persian Art (Oxford, 1939), pls 208A, 230B. Christensen, L’Iran, pp. 394–6. Taj, tr. p. 58. Grabar, Formation of Islamic Art, p. 96 et suiv. et passim. Tabari III, p. 916.

notes sur les cérémonies umayyades 149

Umayyades, leur passé et leurs associations personnelles avec les chefs des grandes tribus rendaient certains types de cérémonies impensables. Mais, si leurs cérémonies n’étaient pas une expression de majesté, du calife en tant qu’“ombre de Dieu,” elles formaient déjà un jeu officiel, un ensemble de signes qui servaient à séparer le prince des autres. Ces signes étaient parfois vulgaires et l’auteur du Kitab al-Taj se moque des Umayyades lorsqu’il raconte qu’ils aimaient se mettre à nu et en général se conduire d’une manière désordonnée derrière le rideau qui les séparait de l’audience.62 Néanmoins un épisode assez bizarre des activités d’al-Walid II complique un peu les choses. Apres avoir écouté des chansons, le calife fit un signe à un serviteur qui se tenait, dit le texte, comme un soleil. Le serviteur tira un rideau et quarante esclaves des deux sexes firent leur apparition, comme si, poursuit le texte, ils étaient des perles éparpillées; tous tenaient entre les mains des aiguières et des serviettes.63 En partie nous avons là une espèce de tableau vivant digne d’un style cinématographique d’une autre époque, comme dans le cas des musiciens déguisés en singes qui apparaissent à une autre occasion.64 Il est possible en fait que certaines sculptures de Khirbat alMafjar et peintures de Qusayr ‘Amrah illustrent précisément ce genre d’activités. Mais il n’est nullement exclu qu’il s’agisse aussi d’un jeu symbolique avec le prince se tenant dans un firmament, motif de la plus haute antiquité que l’on retrouve constamment dans les palais musulmans postérieurs.65 Le monde des cérémonies umayyades se présente ainsi avec une ambiguïté constante et peut-être voulue. On y trouve à la fois un formalisme qui tend à créer des symboles et des rites et un enjouement qui ne prend pas entièrement au sérieux la valeur profonde des actes accomplis. C’est cette ambiguïté qui explique peut-être un palais privé comme Khirbat al-Mafjar, dont la salle la plus décorée et, si l’on suit Ettinghausen,66 imprégnée de symboles iraniens anciens n’est qu’une immense salle de déshabillage pour un bain.

62 63 64 65 66

Taj, tr. p. 59. Aghani VI, p. 133. Aghani VI, p. 123–4. O. Grabar, The Alhambra, à paraître. R. Ettinghausen, From Sassanian Iran, surtout le dernier chapitre.

Chapter IX The Date and Meaning of Mshatta*

The unfinished palace of Mshatta, in the steppe of Jordan some thirty miles southeast of Amman, is probably the best-known secular monument of early Islamic times. Its notoriety rests primarily on the unique monumental decorative band of sculpted stone on its façade (Fig. 1), which was transferred in the early part of the twentieth century to the Berlin Museum. But it is also the result of two debates of more than half a century ago. One dealt with the character and significance of this particular decorative ensemble and of ornament in general. It was a debate that extended beyond universities and museums, as it spread to artists, critics and industrial designers. But within the academy it focused a great deal on a set of antinomies (Orient/Rome, representation/ornament, pure art/ decorative art, among others) which required a precise date for Mshatta in order to identify the meaning of its forms. The second debate involved a curious group of academics and para-academics: usually wealthy travelers, archaeologists and historians who, in the first two decades of the twentieth century, became, in very different ways and with very different points of view, fascinated by material and visual evidence for the formation of Islamic culture. Ernst Herzfeld, the youngest, most learned and most imaginative scholar of that group, returned to questions about Mshatta more than once, always from a different angle, but ended up by being more concerned with the archaeology of Mshatta and with the confirmation of its Umayyad date than with its decoration.1 More recent discussions and summaries by Leo Trümpelmann2 and K. A. C. Creswell have tended to by-pass the decoration or to deal with it exclusively as the sum of more or less localized motifs. Once the date of the building was securely established, all that was needed, or so it seems, was to identify the sources of its * First published in Dumbarton Oaks Papers, 41 (Dumbarton Oaks, Trustees for Harvard University: Washington DC, 1987), pp. 243–7. 1 Complete bibliography in K. A. C. Creswell, Early Muslim Architecture, I, 2nd edn (Oxford, 1969), pp. 604–6. Relatively little has been written about Mshatta since then, except in O. Grabar, The Formation of Islamic Art (New Haven, 1973), pp. 197–200, or else in general manuals. 2 L. Trümpelmann, Mshatta (Tübingen, 1962).

151

152

1 Mshatta, façade

early islamic art, 650–1100

designs. This procedure is a wonderful instance of the common art-historical practice of defining a monument through its genetic pool. The point of these rapid introductory remarks is, first, to note that the study of Mshatta seems from the very beginning to have led to broad concerns about the history of forms or the methodology of visual analysis. The second point is that there does not seem too much doubt about the date and significance of Mshatta. In reality, however, there still is a “problem” of Mshatta, not only because its decoration has not really been explained, but also because even its date, for which a sort of scholarly consensus seemed to have been established, may be far from assured. A summary of that consensus would run as follows: Mshatta is a palace built toward the end of the Umayyad dynasty, most likely by order of al-Walid ibn Yazid, caliph for a few months of 743–4 (hence its unfinished character), but a longtime resident of this part of the badiyah or steppic intermediary zone between the real desert and permanently cultivated areas. This conclusion rests on the following set of arguments. First, the late tenth-century Egyptian Christian chronicle of Severus ibn al-Muqaffa contains a passage referring to the building of a “city” by alWalid in a waterless area which required many workers from Egypt and which was left unfinished. H. Lammens argued that it was a reference to Mshatta; in spite of J. Sauvaget’s rather cryptic doubts concerning the text’s appropriateness for Mshatta, Creswell and nearly everyone else accepted it.3 [244] Second, the alleged combination in Mshatta of Mediterranean themes and techniques (square plan with round towers [Fig. 2], throne room, in the shape of a triconch preceded by a basilical hall, size based on a seventy meters to the side module, technique of relief sculpture, most of the motifs in the decoration) with eastern, Iraqi or Iranian, ones (combination of stone 3

H. Lammens, “La Badia et la Hira,” Etudes sur le siècle des Omayyades (Beirut, 1930), p. 348; J. Sauvaget, “Remarques sur les monuments omeyyades,” Journal Asiatique, 231 (1939), pp. 31–5.

the date and meaning of mshatta

153

2

and brick construction, iwan-like opening of main hall on court, a relatively small but clearly identifiable number of details in ornament) seemed appropriate to an evaluation of Umayyad art as one of juxtaposing disparate elements available to Umayyad patrons through their conquests.4 Third, the combination of a formal core centered on a large reception hall and of living quarters (mostly never built but traceable through bonds in standing walls) on the sides (Fig. 2) seemed to correspond to a typology of functions known in other Umayyad palaces in the badiyah or elsewhere and to illustrate an alleged Umayyad style of living which required the coexistence of official rooms and private apartments, even when the building is far from urban centers or other significantly populated areas. Fourth, the whole Syrian and Transjordanian badiyah is covered with structures of very varying dimensions and importance which can be shown, on archaeological, comparative or literary grounds, to have been built or 4

Such is essentially the thesis developed by Herzfeld. As an additional item not in Creswell’s bibliography, see I. Lavin, “The House of the Lord,” Art Bulletin, 44 (1962), pp. 3–27.

Mshatta, plan

154

early islamic art, 650–1100

heavily rebuilt in Umayyad times;5 hence any major construction from Islamic times6 in that area is assumed to be Umayyad, unless the contrary can be proved. Each of these arguments is open to question. The statement in a Christian chronicle dealing with the ecclesiastical history of Egypt reflected no doubt some conscription of Egyptian labor still remembered over two centuries later, but it can at best be used as evidence to supplement other data, certainly not to identify a specific place in a different province. Umayyad building activities were numerous, and “cities” were built on the Cilician frontier and in the Jazirah. Many of them were left unfinished or, rather, an initial large-scale investment in funds and labor from some central authority was usually followed by slower and more localized construction. The pseudo-stylistic argument of the juxtaposition of forms of different origins is valid to a certain extent. But one need only compare Mshatta to Khirbat al-Mafjar, Khirbat al-Minyah, Qusayr ‘Amrah, or Qasr al-Hayr West, four definitely Umayyad establishments, two of which are less than a day’s horse ride from Mshatta,7 to realize how different Mshatta is from other Umayyad ensembles. The almost geometric rigor of its plan with the diagonals of the square generating the dimensions of the formal part of the building (Fig. 2) contrasts with the seemingly haphazard, additive arrangement of Khirbat al-Mafjar, the contrived composition of Khirbat al-Minyah around its court, the typological simplicity of Qusayr ‘Amrah, or the constructional incongruencies of Qasr al-Hayr West. In Mshatta’s dominant decorative band, as well as in many ornamental fragments still in place, the mass of details is subordinated to a very rigid order – triangles on the façade, simple and sober acanthus leaves or vine tendrils elsewhere. The four other buildings, even the relatively sober Khirbat al-Minyah, have more decoration and, especially, a much greater range of themes and techniques. In construction, design, or decoration Mshatta exhibits none of the exuberance of the other major Umayyad palaces, nor does its façade show the wild festival of motifs that could be reconstructed for the façade of Qasr al-Hayr West, now in the court of the Damascus National Museum. What characterizes Mshatta is, first of all, the logical and technical coherence of its design and, second, that its architectural forms and the organization of decoration on its façade are innovations without clear antecedent in Umayyad art, with one exception to

5

6

7

There is no complete list of such sites. See Sauvaget, “Remarques” and “Châteaux Umayyades de Sync,” REI, 35 (1967); see also O. Grabar, R. Holod, J. Knudstad and W. Trousdale, City in the Desert (Cambridge, Mass., 1978), pp. 150 ff. There is, in my opinion, no point in reopening the debate on a pre-Islamic date for Mshatta, as the presence of a mihrab built into the outer wall makes clear. This is not to say that there was no earlier settlement in or around the palace. All these monuments are described by Creswell and discussed by O. Grabar. I am leaving out such places as Muwaqqar, Ziza, or Qastal, which are even closer to Mshatta, but about which too little is known.

the date and meaning of mshatta

155

3 Ukhaydir, plan

which I shall return later. The fact that Mshatta is so different from other Umayyad buildings does not necessarily make it non-Umayyad. Stylistic coherence hardly characterizes Islamic art in the first half of the eighth century, and, within the confines of an evolutionary theory of formal growth, Mshatta could indeed be seen as the culmination of an Umayyad stylistic progression. These are possible [245] but not necessary conclusions, and my point at this stage is simply that such stylistic characteristics of Mshatta as can be defined do not make the traditional conclusion about its date necessary. A similar reasoning weakens the third argument for an Umayyad date. The presence of living and formal zones within a single ensemble is not peculiar to Umayyad establishments in the badiyah, as witnesses Ukhaydir in Iraq (Fig. 3), usually dated to the late eighth century,8 with the same characteristics. More important and, once again, unique within the series of Umayyad palaces is that the official zone in Mshatta is so axial, so clearly the main part of the complex, with presumed living quarters relegated to the 8

Creswell, Early Muslim Architecture, II (Oxford, 1940), pp. 35–46; in addition, see W. Caskel, “Al-Uhaidir,” Der Islam, 39 (1964) and reports on recent discoveries in Sumer, 22 (1966), pp. 79–110 and Mesopotamia, 2 (1967), pp. 195–218.

156

early islamic art, 650–1100

side. Within Islamic architecture this feature is first known in the urban governor’s palace and administrative seat in Kufah. It is as though the message of Mshatta is its visual assertion of power, first through a wide and symmetrical horizontal band across the façade and then through a centrally placed reception hall or throne room. The message is quite different from that of the Umayyad châteaux, where functions of life and of pleasure predominate and reception halls are to the side of a main axis (Khirbat alMinyah), have to be hypothesized on the second floor (Jabal Says, Khirbat al-Mafjar, Qasr al-Hayr West), or are absent altogether (Qasr al-Tuba).9 Thus Mshatta has relatively little in common with other Umayyad buildings in Syria and Palestine. But it is remarkably close to early ‘Abbasid buildings from Iraq in composition and in many constructional details. At Ukhaydir, also, a strongly axial formal area ending with a reception hall or throne room opening on a court is separated from living areas relegated to side courts. Ukhaydir did not have a large decorative statement on its façade. But the two-storied domed entrances to the city of Baghdad illustrate at least an ‘Abbasid concern for the external broadcasting of its message of power and authority. Only the Dome of the Rock, in Umayyad times, was built and decorated with the intention of affecting those outside the building, and in this case even quite far away. Since it is not possible to explain Mshatta through the Dome of the Rock, the conclusion is that Mshatta is more likely to be an early ‘Abbasid building than an Umayyad one.10 And even the stylistic details of the ornament on the façade are easier to see as the forerunners of the first styles of Samarra than as the more rigidly controlled and more tightly coherent continuations in stone of the stucco ornament found in Qasr al-Hayr West or Khirbat al-Mafjar. In terms of style, composition, or technique, nothing, it seems to me, makes an ‘Abbasid date impossible and much makes it likely. But such a conclusion runs against the fourth argument for an Umayyad date which is that large-scale building in the Syrian and Jordanian badiyah, in fact throughout most of Syria and Palestine, was, in early Islamic times, restricted to the Umayyads. This argument, however, cannot be maintained in such absolute terms, even though the maintenance and development of these areas as viable agricultural or commercial enterprises with attendant private villas had often been sponsored by Umayyad princes or by other Arab aristocrats under Umayyad rule. In fact it is precisely the success of these Umayyad or Umayyadaided activities that attracted the early ‘Abbasids. The excavations of Qasr al-Hayr East demonstrated a continuity between the time of the Umayyad foundation and a decline in the late ninth century.11 An Umayyad settlement 9 10 11

The argument for second-floor reception rooms was first made by J. Sauvaget. It is rather curious that not one of the 116 opinions as to the date of Mshatta (Creswell, Early Muslim Architecture, I, pp. 633–4) lists the ‘Abbasid possibility. O. Grabar et al., City, pp. 156–9.

the date and meaning of mshatta 157

in Transjordan was taken over by a son of the caliph al-Mahdi and eventually acquired by a consortium of oil merchants from Kufah.12 Ramlah was maintained and developed by the ‘Abbasids, and, until al-Mutawakkil in the ninth century, nearly every ‘Abbasid caliph traveled to Palestine and Transjordan, at times to visit Jerusalem, at others on their way to or from Mecca. Although their exact number is still a matter of debate, there is no doubt about the considerable investments made by the ‘Abbasids in restoring the sanctuaries of Jerusalem.13 Finally, large estates and whole cities like Humaymah in southern Transjordan were owned by members of the ‘Abbasid family even in Umayyad times, and there is no reason to believe that they were given up after 750. In short, an important ‘Abbasid presence can [247] easily be demonstrated in this area at least during the first century of ‘Abbasid rule, after which the whole badiyah began to decline, especially on its Syrian and Palestinian borders. Historical circumstances, therefore, do not make an early ‘Abbasid date for Mshatta impossible, any more than formal comparisons did. For the history and stylistic interpretation of the most celebrated feature of Mshatta, the ornament on its façade, this change of dates by circa half a century is not significant, as it is ultimately not very important or interesting to know whether Mshatta represents the last stage of one style or the first stage of a new one. It does matter, however, in suggesting a different purpose for and ideology behind the building and in explaining perhaps the oddity of the shape of its decoration. Within a radius of some thirty miles from Mshatta are located no fewer than eight major or minor Umayyad foundations, some like Muwaqqar provided with an elaborate irrigation system for agriculture, and therefore particularly likely to have remained in use after the ‘Abbasid takeover. In this environment I propose to see Mshatta as the center of the new authority, as a formal building of power rather than as a château. And, like the elongated rectangular tiraz bands on official ‘Abbasid vestments, a large rectangular band of ornament on the facade was meant to serve as a sign of the presence of a new rule. In later times such ideas will be expressed through inscriptions or through towers and green or, much later, gilt domes.14 At this time of experimentation with visual signs, an ornamental band of grandiose dimension was deemed sufficient.15 12 13 14

15

Baladhuri, Futuh, trans. as The Origins of the Muslim State by R. K. Hitti (New York, 1916), p. 197. For other examples see Grabar et al., City, p. 155. See, among other places, R. W. Hamilton, The Structural History of the Aqsa Mosque (Jerusalem, 1951) and Guy Le Strange, Palestine under the Moslems (Boston, 1890). The symbolism of domes as authority already existed; see O. Grabar, “Al-Mshatta, Baghdad, and Wasit,” The World of Islam (Studies in Honor of P. K. Hitti) (New York, 1960), pp. 99–108. A more sophisticated explanation of these domes will soon be proposed by Jonathan Bloom. Another such example of experimentation may well be the citadel at Amman. In spite of recent work (A. Almagro, El Palacio omeyyade de Amman [Madrid, 1984] and the unpublished thesis of A. Northedge), I still have lingering doubts about an Umayyad date for the celebrated large square pavilion.

Chapter X La Place de Qusayr ‘Amrah dans l’art profane du Haut Moyen Age*

La découverte, à la fin du siècle dernier, de peintures murales décorant le petit bain omeyyade de Qusayr ‘Amrah dans la steppe alors arabique et maintenant jordanienne eut parmi les historiens de l’art un retentissement bien plus grand que ne le méritaient la qualité de ces peintures ni le contexte archéologique ou technique de leur découverte. Les islamisants furent surpris par la présence de représentations de tout genre dans l’art d’un monde jugé iconophobe et s’empressèrent de faire sortir une série de travaux sur le prétendu iconoclasme de l’Islam. La grande majorité de ces travaux sont encore essentiels à l’entendement d’un sujet qui ne sera jamais résolu. Les historiens de l’art antique ou chrétien pensèrent avoir trouvé à Qusayr ‘Amrah un exemple de ce que l’art profane aurait été à une époque, le début du huitième siècle, dont il est presqu’absent. Par ailleurs, dans des cas précis comme les personnifications ou bien les représentations pseudo-scientifiques, Qusayr ‘Amrah aurait conservé des motifs antiques disparus des centres plus importants et plus connus. C’est ainsi que les images de Qusayr ‘Amrah entrèrent dans le corpus iconographique de Salomon Reinach et dans les premiers manuels d’art byzantin, ceux de Dalton et de Diehl.1 Les princes omeyyades qui commandèrent ces peintures auraient simplement adopté le vocabulaire visuel d’un art profane prévalant en Méditerranée en y ajoutant peut-être un iranisme de service ou même quelques éléments d’origine plus lointaine, comme l’avait jadis proposé Ettinghausen.2 Et, d’un point de vue formel plus général, il y avait le jugement de Herzfeld qui avait vu dans les peintures de Qusayr ‘Amrah un jalon important dans la déhellénisation de l’art islamique, un processus qui se serait achevé avec les peintures de Samarra au neuvième siècle, du

* Premièrement publié dans Cahiers Archéologiques, 36 (1988), pp. 75–84. 1 Une bibliographie pour ainsi dire complète jusqu’à 1965 se trouve dans K. A. C. Creswell, Early Muslim Architecture (Oxford, 1969), pp. 414–15. Pour ces derniers vingt ans, il faut consulter Creswell et autres, A Bibliography of Architecture, Arts and Crafts of Islam, Supplements I et II (Le Caire, 1972 et 1984). 2 R. Ettinghausen, La Peinture Arabe (Genève, 1962), p. 30.

159

160

early islamic art, 650–1100

1 Qusayr ‘Amrah, vue extérieure

moins jusqu’aux renouveaux fatimides et ayyoubides des onzième et douzième siècles.3 Ces interprétations dérivent d’une position intellectuelle et méthodologique qu’il est essentiel de dégager. Chaque monument s’y trouve placé sur une courbe définissant le degré de ce qu’on pourrait appeler son “isomorphisme” avec l’art si riche et si varié de la basse antiquité. Ces courbes permettent des “renaissances” ou des “renouveaux,” mieux étudiés dans les arts byzantins et occidentals qu’islamiques, et éventuellement, du moins en Occident, ces courbes deviennent annonciatrices de la vraie et grande Renaissance. A un niveau très général d’une histoire universelle des arts, une interprétation de Qusayr ‘Amrah du point de vue de ses relations avec l’Antique est peut-être valable mais elle est peu utile lorsqu’on cherche à comprendre pourquoi un minuscule bain isolé dans un coin perdu de la steppe (Fig. 1) aurait été transformé en festival d’images. Un historien au positivisme fanatique comme Creswell et quelques observateurs locaux avaient bien vu que le contexte chronologique et spatial purement local du petit bain omeyyade méritait d’être compris, mais ils hésitèrent à en tirer les conséquences appropriées.4 Toutes ces considérations resteraient purement théoriques sans les restaurations entreprises il y a maintenant plus de dix ans par une mission espagnole dirigée par M. Martin Almagro. Les résultats des travaux de la mission n’ont pas été complètement publiés, quoique le beau livre polyglotte 3 4

E. Herzfeld, “Die Genesis der islamischen Kunst,” Der Islam, 1 (1910), pp. 106–8; Die Malereien von Samarra (Berlin, 1927), pp. 4–7. Creswell, Early Muslim Architecture, pp. 402–4; F. Zayadine, “The Umayyad Frescoes,” Archaeology, 31 (1978).

qusayr ‘amrah dans l’art profane 161

(texte espagnol, résumés en anglais, français, allemand et arabe) qui a paru en 1975 en donne l’impression. 5 La construction d’une route et la permanence d’un gardien ont rendu le site facilement accessible. Mais ce qui est bien plus important, c’est que l’on y voit tout ce que l’on y verra. Car, à l’exception possible de renseignements recueillis pendant les travaux de restauration et restés inédits, les peintures telles qu’elles apparaissent sont devenues la totalité de la documentation visuelle à la disposition des savants. Ses lacunes resteront des lacunes, car, à l’encontre des arts religieux, les arts profanes ne [77] se répètent pas automatiquement et surtout pas d’une manière qui permette d’imaginer facilement les détails manquants. Le champ sémantique ou l’index archéologique, selon que l’on emprunte une expression d’André Grabar ou de Max van Berchem, des peintures de Qusayr ‘Amrah en devient d’autant plus important à définir pour pouvoir répondre aux questions posées par l’existence même de Qusayr ‘Amrah. Quelle place les peintures doivent-elles occuper dans l’histoire des arts profanes? Est-il juste d’y voir un exemple d’arts par ailleurs disparus? Ou 5

M. Almagro et autres, Qusayr ‘Amrah (Madrid, 1975); voir aussi le travail de J. M. Blázquez, “La Pintura helenística de Qusayr ‘Amrah”, Archivo Español de Arqueología, 56 (1983), dont le titre est révélateur.

2 Qusayr ‘Amrah, première salle, vue d’ensemble (angle nord-ouest)

162 early islamic art, 650–1100

3 Qusayr ‘Amrah, première salle, mur nord avec l’abside (ensemble)

qusayr ‘amrah dans l’art profane 163

bien est-ce un cas tout à fait particulier qui n’a d’autre valeur et d’intérêt que lui-même? Une réponse définitive à ces questions demanderait l’analyse détaillée de l’ensemble des peintures du bain. Mais une ébauche de réponse est possible autour d’images pour lesquelles le répertoire des premiers siècles de notre ère est particulièrement riche. Tel est le cas des scènes de chasse comprises d’une manière assez large pour inclure la vaste majorité des scènes avec animaux. Quatres scènes plus ou moins directement associées à la chasse ont été préservées. Toutes se trouvent dans la salle principale du bain – un mélange entre la salle du trône médiévale et l’apodyterium romain6 – entièrement recouverte de peintures dont presque toutes étaient invisibles depuis le début du siècle et ou les restaurations ont été particulièrement réussies (Figs 2, 3). Sur le mur est, au-dessus de trois images séparées dont la célèbre représentation des Six Rois (Fig. 4), une longue bande de plus de sept mètres sur une hauteur qui a au moins un mètre et demi est consacrée à la représentation d’une battue d’onagres ou de chevaux sauvages (Figs 5 à 9). La scène est facile a reconstituer. A côté d’un campement de bédouins dont on aperçoit les tentes à l’extrême droite (Fig. 6) et à l’extrême gauche,

6

O. Grabar, La Formation de l’Art islamique (Paris, 1987), pp. 219–20.

4 Qusayr ‘Amrah, première salle, mur est (ensemble) et arche centrale

164

early islamic art, 650–1100

5 Qusayr ‘Amrah, première salle, mur est, détail: scène de battue

6 Qusayr ‘Amrah, scène de battue, détail

un filet a été tendu entre une série de grands pieux et recouvert lui aussi d’une teinture noire, comme celle des tentes. Le filet est rattaché à une corde posée à terre, mais qui, à un signal donné sera relevée par des personnages cachés derrière des morceaux de tissu attachés à des piquets qui avaient vraisemblablement été plantés en terre (Fig. 7). Les personnages ne sont représentés que par une tête et un bras tenant un instrument

qusayr ‘amrah dans l’art profane

165

bizarre que je m’explique comme servant à faire passer la corde une fois que l’enclos [78] a été monté. Les animaux (Figs 8 et 9) sont poussés vers l’enclos par des cavaliers dont trois sont en partie visibles. Deux galoppent parmi les onagres, le troisième est tombé à terre à l’extrême gauche de la scène. La scène est vivante et animée, schématique dans la composition et la représentation des gens et des choses, mais d’une certaine vigueur dans l’observation des animaux en mouvement. Les deux scènes suivantes se font face sur les murs sud et nord de la nef ouest. Elles sont beaucoup moins bien préservées, quoique leur sens général semble clair. Sur l’une (Fig. 10), dans un enclos temporaire semblable à celui de la scène précédente, six personnages sont occupés à tuer des onagres et peut-être un chameau. Il est possible qu’il ne s’agisse que de les maîtriser pour les marquer d’un fer, mais l’hypothèse moins agréable d’une boucherie semble être confirmée par la troisième scène (Fig. 11), qui [80] montre clairement deux personnages aux jambes nues en train d’éviscérer et d’écorcher des animaux déjà tués. Le cadre semble être une clairière naturelle ou, plus vraisemblablement, une étable recouverte de grandes feuilles. Le quatrième et dernier exemple, d’un style plus rude et nettement moins réussi, ne montre aucun personnage, mais seulement des chiens attaquant des bovins aux têtes plutôt paisibles (Fig. 12). Il s’agirait ou bien de l’attaque d’un troupeau par des chiens sauvages ou bien d’une attaque de bovins sauvages. Cette deuxième hypothèse est la plus vraisemblable, car des références dans

7 Qusayr ‘Amrah, scène de battue, détail

166

8 Qusayr ‘Amrah, scène de battue, détail

early islamic art, 650–1100

la poésie arabe de l’époque implique l’existence de vaches sauvages.7 On pourrait certes ajouter à ces quatre scènes les animaux paisibles arrangés en quinconce sur le plafond de la deuxième pièce du bain (Fig. 13) ou bien le combat mal préservé d’un lion avec un taureau dans la première salle. Les animaux en quinconce sont en général comparés, et à juste titre, avec une célèbre mosaïque d’Antioche, tandis que le combat du lion et du taureau fait partie des grands thèmes que l’art iranien propagea dans le reste du monde. Il est facile d’expliquer ces deux images comme étant des copies, peut-être modifiées dans tel ou tel détail, de modèles courants. Autrement dit, rien de nouveau n’est proposé ni pour le modèle, ni pour Qusayr ‘Amrah. Au mieux, pourrions-nous ergoter sur la nature du modèle; était-ce un tissu ou un livre? Mais le principe même de la transmission d’images à partir d’images [81] n’est plus à démontrer lorsqu’une composition ou un motif apparaissent aussi fréquemment que les animaux en quinconce et le lion-ettaureau. Or, mes quatre premiers exemples de chasse sont remarquables par le fait que je ne leur ai trouvé aucun modèle dans l’immense répertoire des scènes de chasse qui décorèrent les palais, villas, et autres habitations de Piazza Armerina, le Palais impérial de Constantinople et les palais allant 7

Il s’agit des odes de Labid et d’Imrulqays. Voir K. Abu-Deeb, “Toward a structural analysis,” International Journal of Middle Eastern Studies, 6 (1975).

qusayr ‘amrah dans l’art profane 167

d’Espagne à la Syrie.8 Certains détails comme les vêtements des bouchers ou les enclos temporaires se retrouvent certes, mais ces détails sont, jusqu’à preuve du contraire, trop minimes et trop rares pour rendre l’existence de modèles vraisemblables. Plusieurs autres détails me semblent confirmer cette conclusion. Par exemple, chacune de ces scènes est très facile à expliquer dans le contexte immédiat de la région. Il s’agissait d’une culture de nomades dont les tentes noires étaient déjà du type que l’on retrouve aujourd’hui. Les animaux sauvages étaient encore monnaie courante dans la steppe et on les attrappait pour les apprivoiser ou bien pour s’en nourrir. La spécificité de l’image et la possibilité quasi-automatique de situer l’événement figuré dans le contexte immédiat de la représentation rend tout modèle imagé externe superflu. On peut en fait aller plus loin et suggérer que dans chaque cas c’est un événement concret et local qui est représenté. C’est ainsi que j’explique le cavalier tombé de la première scène: il s’agissait [82] de quelque chose qui était arrivé récemment. Il suffisait de donner quelques indications minimes pour que l’événement soit rappelé à la mémoire et fasse partie d’un souvenir collectif ou individuel qu’un commanditaire inconnu, vraisemblablement un prince omeyyade (l’inscription parle bien d’un amir), voulut perpétuer. 8

I. Lavin, “Antioch Hunting Mosaics,” Dumbarton Oaks Papers, 17 (1963).

9 Qusayr ‘Amrah, scène de battue, détail

168

10 Qusayr ‘Amrah, première salle, mur sud de la nef ouest, détail

early islamic art, 650–1100

qusayr ‘amrah dans l’art profane 169

11 Qusayr ‘Amrah, première salle, mur nord de la nef ouest, détail

170 early islamic art, 650–1100

12 Qusayr ‘Amrah, première salle, mur ouest, détail

L’exemple des scènes de chasse n’est qu’un cas particulier d’observations sur des séries d’images provenant de Qusayr ‘Amrah. Dans chaque série on trouverait un contraste curieux entre des sujets pour lesquels on trouve tout de suite un modèle approprié, par exemple la représentation d’un prince dans l’alcôve principale du bain, et d’autres qui sont imaginaires et fantaisistes ou bien s’expliquent comme référence à des événements locaux. Il y a ainsi dans ces peintures un premier niveau de tension dans la typologie des images, une tension entre des clichés visuels d’origines diverses et des images locales et donc uniques. Un deuxième niveau de tension est iconographique; des motifs officiels et formels (prince trônant, les Six Rois, princesse sous une tente,9 danseuses chastes, musiciens) côtoient des scènes de nudité, des détails pornographiques, des copies d’objets de provenance diverses, même peut-être la traduction en images de thèmes littéraires, en somme des motifs entièrement privés. Et puis il y a une tension d’expression entre les sujets que le ou les peintres, vraisemblablement des Syriens ou des Palestiniens formés à la tradition antique et byzantine, savaient traduire en images parce qu’ils entraient dans le potentiel de leurs compétences, et les motifs qui étaient au-delà de leurs moyens, comme par exemple les vaches et les chiens de notre Fig. 12. Ces tensions reflètent la personnalité de celui qui 9

J’en donnerai un exemple dans une étude à paraître dans les Mélanges Dominique Sourdel.

qusayr ‘amrah dans l’art profane

171

13 Qusayr ‘Amrah, deuxième salle, plafond

transforma un minuscule bain en une espèce d’album photo-[83]graphique de sa vie personnelle, en témoignant de ses connaissances, de ses souvenirs, et peut-être même d’une certaine idéologie princière, quoique le côté unique du personnage l’emporte sur son idéologie.10 Pour répondre à la question de la valeur et de l’importance de Qusayr ‘Amrah pour l’historien de l’art du Moyen Age, il faudrait dire tout d’abord que chaque image risque d’avoir une signification différente. De ce point de vue, Qusayr ‘Amrah et ses peintures n’ont d’intérêt que dans la mesure où nous pouvons retrouver ou reconstituer le système visuel ou le langage auxquels telle image aurait appartenu. Tel est le cas, par exemple, du plafond astronomique. Mais le fait qu’une telle généralisation est parfois possible n’implique en aucun cas le reste des images. Car, d’un autre point de vue, Qusayr ‘Amrah est un objet unique avec un sens synchronique précis qui s’élabore à partir d’observations de détail comme celles que nous avons faites sur les scènes de chasse. Dans cette perspective, Qusayr ‘Amrah est moins une œuvre d’art profane qu’une œuvre d’art privée où la présence du commanditaire est plus visible et plus importante à définir que celle de l’artiste.

10

Pour cette idéologie, voir Grabar, Formation, surtout les chapitres 2 et 5.

Chapter XI L’Art Omeyyade en Syrie, Source de l’Art Islamique*

L’expression “art omeyyade” est généralement utilisée pour recouvrir les monuments fort nombreux datés ou datables entre 680 et 750 de notre ère et localisés en Syrie, au Liban, en Palestine et en Transjordanie. C’est la région que la tradition arabe et islamique a appelé le bilad al-Sha’m, la moitié occidentale du Croissant Fertile. Il s’agissait en fait d’un état d’être culturel traditionnel plutôt que d’une terre aux frontières précises et claires. Pour des raisons sur lesquelles je reviendrai en conclusion, les monuments de la même période qui proviennent d’Iraq ou d’Égypte ne sont généralement pas inclus parmi les œuvres d’art omeyyade. Tout comme les monuments syriens, cependant, ceux d’Iraq, d’Égypte, ou le petit nombre datable au premier siècle de l’hégire qui proviennent d’Iran, de la Mésopotamie du Nord ou du Maghrib font tous partie de ce que l’on a appellé l’art islamique. Cette distinction entre un art omeyyade qui aurait été limité aux monuments de Syrie et un art islamique plus inclusif est-elle justifiée, ou bien est-elle simplement le résultat d’une de ces nombreuses confusions taxonomiques dont l’histoire des arts et des cultures est pleine, confusions rarement créées par les monuments eux-mêmes mais plutôt par les idéologies et préjugés de ceux qui étudient la culture ou les arts, et, encore plus, de ceux qui interprètent la science et les connaissances archéologiques ou autres pour les besoins des sociétés qui les entourent? Il n’appartient vraisemblablement pas à un colloque scientifique de débattre ce genre de questions d’historiographie ou de politique culturelle. Mais il est bon de les garder en tête pour mieux situer les problèmes plus précis qui vont m’occuper. Il est en fait tout à fait normal d’associer les monuments de cette époque en Syrie avec la dynastie régnante des Omeyyades, la seule dynastie, dans toute l’histoire de l’Islam, à avoir plus ou moins maintenu son emprise sur le monde immense, des Pyrénées aux frontières de l’Inde et de la Chine, tombé * Premièrement publié dans Actes du Colloque international Lyon – Maison de l’Orient Méditerranéen, ed. P. Canivet and J.-P. Rey-Coquais (Paris, 1992), pp. 188–93. Le texte écrit est, à quelques détails près, celui qui fut lu à l’Institut du Monde Arabe.

173

174

early islamic art, 650–1100

aux mains des Arabes musulmans. Il est, cependant, moins certain que ce qui semblerait normal soit aussi légitime et justifié. Car les monuments de l’art dit omeyyade, les contextes sociaux et culturels dans lesquels ils apparaissent, et surtout les interprétations qui en ont été proposées posent toute une série de problèmes et de difficultés que je me propose de passer en revue sous trois rubriques, chacune basée sur un paradoxe ou, [189] tout au moins, une contradiction dans l’évidence même ou bien dans l’entendement que nous en avons: 1.

2.

3.

Pourquoi y a-t-il tant de monuments omeyyades en Syrie, lorsque les monuments du siècle précédent et surtout des siècles suivants se mesurent au compte-goutte? Pourquoi y a-t-il tant de monuments profanes, alors que l’histoire des arts avant le septième siècle est presqu’entièrement celle des arts chrétiens et du Christianisme et celle des siècles suivants l’histoire des mosquées? (et ce sera là le sujet principal des remarques qui suivent) Est-il vraiment possible d’interpréter ces monuments à la fois comme l’expression ou une des expressions des arts du Haut Moyen Age ou de l’Antiquité tardive autour de la Méditerranée et également comme les premiers pas d’un art nouveau et d’une attitude nouvelle envers les arts?

Avant de passer à ces questions, je voudrais établir le point de vue auquel je me place, en partie pour les besoins de ces remarques et en partie par choix professionnel, pour réfléchir aux monuments du premier siècle de l’hégire. Ce ne sera pas l’optique de l’archéologue concerné par tout ce qui reste de la culture matérielle d’un lieu ou d’une époque. Ce n’est pas non plus l’option de l’historien qui cherche à établir et puis à élucider dans le temps des événements concrets ou généraux, les hommes qui y participèrent et les facteurs, immédiats ou profonds, qui les expliquent. Mon point de vue sera celui de l’historien de l’art, ce qui implique deux choses. D’une part il y a une vie, une ontologie et une langue des formes qui sont en partie indépendantes des circonstances qui les ont créées. D’autre part, chaque instant culturel, chaque chronotope identifié par la place et le temps d’un monument préservé ou même comme abstraction, fait des choix esthétiques (tel objet ou telle manière de décorer sont beaux ou plaisent davantage) ou bien des choix de luxe plutôt que de goût (l’or fait plus “riche” que le bronze et les mosaïques sont plus chères que la peinture). L’explication de ces choix est la trame du discours de l’historien de l’art. Ces distinctions et ces définitions auront, vous le verrez, leur importance dans les conclusions que je vous proposerai. Voyons tout d’abord pourquoi y a-t-il tant de monuments en Syrie entre les dernières décennies du septième siècle et le milieu du huitième. Une seule nouvelle ville fut créée dans les vieilles terres du bilad al-Sha’ m, Ramlah en Palestine, mais la frontière avec l’Anatolie chrétienne et surtout la vallée

l’art omeyyade en syrie 175

1 Qusayr ‘Amrah: Vue extérieure

de l’Euphrate moyen se couvrirent d’établissements, urbains ou autres, nouveaux, dont l’exploration ne fait que commencer. Des changements urbains importants, quoique peu étudiés, sont certains pour Damas, Jérusalem, Hama, et Bosra et vraisemblables pour beaucoup d’autres centres anciens. Une vingtaine de mosquées sont plus ou moins connues, dont les deux plus célèbres sont celles de Damas et de Jérusalem. Il y a l’extraordinaire et unique sanctuaire de la Coupole du Rocher admirablement posée sur un octogone aux faces parfaitement identiques et avec ses 240 mètres carrés de mosaïques restaurées certes, mais en gros omeyyades par leur programme décoratif et par leur style original. Et puis il y a ce que l’on a appelé les “châteaux” omeyyades, une quarantaine (leur nombre exact est sujet à caution et dérive de jugements et de décisions archéologiques qui ne sont pas de mon ressort aujourd’hui) de résidences installées à l’intérieur ou aux alentours de grandes propriétés foncières, de grands axes routiers, de centres d’élevage, et peut-être d’entreprises semi-industrielles. La grande majorité de ces “châteaux” n’étaient que de grosses fermes ou de simples résidences, mais dans cinq cas certains (Khirbat al-Mafjar, Khirbat Minyah, Qusayr ‘Amrah, Mshatta, Qasr al-Hayr Ouest) et trois cas problématiques (Jabal Says, Qasr al-Hayr Sharqi, Anjarr) une superstructure de fonctions artisanales (pressoirs), commerciales (caravansérail), ou de plaisir (bains) aussi bien que de décoration (peintures, mosaïques, sculptures) fut ajoutée au simple habitat. Le décompte exact de la présence omeyyade en Syrie n’a jamais été fait et se complique lorsqu’on prend en considération les églises et autres établissements fondés par des non-Musulmans après 680 et les objets, moins rares qu’on ne le pense, attribuables à l’époque omeyyade. Par ailleurs les références écrites sont également nombreuses dans les textes arabes, syriaques, grecs, arméniens

176 early islamic art, 650–1100

2 Qusayr ‘Amrah: Vue intérieure: façade sud

et géorgiens et méritent une attention qui n’a été accordée jusqu’à présent qu’aux textes arabes en prose. Si l’on pense simplement en termes économiques de chantiers en marche et d’objets vendus ou transités par le bilad al-Sha’m, nous avons certainement affaire à des centaines de cas démontrables, dont une vingtaine sont ou avaient été des œuvres d’art exceptionelles.1 Il y a, me semble-t-il, deux explications principales et complémentaires pour ce nombre si considérable de fondations ou d’utilisations significatives sous les Omeyyades. La plus importante est ce que l’on appellerait aujourd’hui le phénomène pétrolier, c’est-à-dire la présence soudaine d’un nouveau mécénat nanti de ressources financières énormes. Au huitième siècle, comme à toutes les époques depuis les Achéménides, sinon les Assyriens, les “faiseurs de choses” (architectes, artisans, orfèvres, peintres, décorateurs de tout genre) ou bien les revendeurs de choses, les marchands d’objets d’art ou de souvenirs historiques, se précipitent là où se trouvent les bailleurs de fonds. L’autre explication est l’existence d’une idéologie associée à la fortune. Je reviendrai sur certains aspects précis de cette idéologie. Pour l’instant il suffit d’en saisir un élément essentiel: le besoin simultané de s’identifier comme “autre” et 1

On trouvera des descriptions sommaires de tous ces monuments ainsi que les références bibliographiques nécessaires dans K. A. C. Creswell, A Short Account of Early Islamic Architecture, rev. James W. Allen (Scolar Press, Aldershot, 1989). Pour des interprétations voir Oleg Grabar, La Formation de l’Art Islamique (Paris, 1987).

l’art omeyyade en syrie

177

nouveau, donc différent, et, en même temps d’être accepté par le monde tel qu’il existe. Les princes de la dynastie régnante et la pléiade des grands gouverneurs et des grandes familles qui les entouraient avaient les moyens matériels et le besoin psychologique de faire valoir leur richesse et leur puissance pour toutes sortes de raisons allant de la politique générale aux particularités individuelles de chacun. Mais, étant à la tête d’une communauté qui se définissait comme nouvelle et comme différente de tout ce qui la précédait, l’expression de la richesse et du pouvoir se devait de trouver des formes originales. C’est ainsi, dans l’hypothèse que je propose, la richesse des princes omeyyades et de leur entourage qui expliquerait la quantité d’œuvres d’art omeyyades et de témoins archéologiques que l’on peut leur attribuer. Et c’est le contexte idéologique de la richesse et de l’identification islamique qui expliquerait le nombre de monuments profanes, d’habitations, de bains, de palais, plutôt que de monuments religieux ou sectaires réservés, par définition, à un groupe déjà unifié par le mode de vie et les règles de conduite sinon par la politique. Dans toutes les autres régions des premiers siècles de l’Islam (en Iraq, en Ifriqiah, en Égypte, au Khorasan) les monuments profanes qui existent comme le dar al-‘imarah de Kufah sont pauvres en art et il faut attendre le neuvième siècle pour trouver de vrais équivalents aux palais omeyyades de Syrie, et encore en nombres restreints et sans l’exubérance des grands monuments de la première dynastie de

3 Qusayr ‘Amrah: Salle principale, ensemble du mur nord

178

4 Qusayr ‘Amrah: Salle principale, détail: travaux des champs ou joutes

early islamic art, 650–1100

l’Islam. Dans la Syrie du huitième siècle, une région encore chrétienne et à la fidélité incertaine, la visibilité du pouvoir était essentielle. A l’opposé de l’Iraq où dominaient les mouvements religieux et où l’islamisation fut [190] relativement rapide, la Syrie des Omeyyades exigea une présence séculière plutôt que pieuse, car cette présence devait être comprise par les non-Musulmans encore plus que par les fidèles et c’est ainsi un art profane qui domina. La formule suivante résumerait mon argument pendant près d’un siècle (de 661 à 750, et certainement à partir de 690 lorsqu’Abd alMalik organise le pouvoir de la dynastie): la Syrie, le bilad al-Sha’m, pour la première (et dernière) fois de son histoire, était capitale d’empire et en épousait les conséquences. Dans cette perspective, comment devons-nous comprendre l’art des Omeyyades? II est facile d’y voir simplement un exemple de l’art méditerranéen avec quelques bizarreries locales. Les mosaïques de la Grande Mosquée de Damas appartiennent directement à une tradition de représentations architecturales que l’on trouve à Salonique au siècle précédent ou bien à Antioche et qui descend en filiation directe et relativement bien établie de la peinture romaine du premier siècle. Du point de vue des formes, il n’y a rien d’islamique dans ces mosaïques, rien qui n’aurait pu être fait pour une église chrétienne ou un palais byzantin. Il y a certes l’absence de représentations d’êtres animés, mais ces représentations ne sont pas plus obligatoires dans l’art chrétien qu’elles ne sont exclues de l’art islamique.

l’art omeyyade en syrie

179

Dans la Coupole du Rocher, les mêmes arguments sont valables. On y trouve certes des combinaisons inusitées de motifs – joyaux et plantes, sujets iraniens et méditerranéens – mais ces combinaisons ne sont pas exclues du langage de l’art byzantin ni incompatibles avec ses habitudes et ses traditions. A Qusayr ‘Amrah, il y a certes des sujets – les Six Rois par exemple – qui ne peuvent être qu’omeyyades, mais cette conclusion n’est vraie que pour la précision iconographique des détails, car ce type d’image officielle n’est nullement invraisemblable dans plusieurs des traditions pré-islamiques. La majorité des sujets dépeints dans ce bain si original ainsi que leur style un peu grossier sont plus faciles à définir comme des exemples d’art provincial dont les images refléteraient d’une manière plus ou moins réussie les grands thèmes d’un “centre” historique ou culturel, tout en y ajoutant des particularités inspirées par la situation locale.2 On se trouverait, dans le cas de Qusayr ‘Amrah, en présence d’un phénomène semblable à la relation entre les mosaïques de pavement tunisiennes et celles, par exemple, d’Antioche, c’est à dire un exemple de monument, dont l’ambition n’était pas à la mesure de la compétence de ses créateurs. On peut se demander s’il y avait en fait un vrai “centre”, dans le sens sociologique ou géographique du terme, dont Qusayr ‘Amrah aurait été le reflet. Le seul candidat possible serait Byzance, une hypothèse peu vraisemblable. Peut-être pourrait-on démontrer à partir de Qusayr ‘Amrah le phénomène beaucoup plus intéressant de l’existence pour les arts princiers du Haut Moyen Age de centres mythiques plutôt que réels, d’un monde rumi, “chrétien” (ou plus tard dans l’imagination islamique sini, “chinois”) avec des Césars-qaysars ou des Kisras-Chosroès que l’on inventait à partir d’objets et de tissus, comme on inventait Constantinople, Rome, ou Ctésiphon comme villes imaginaires. Le style des peintures de Qusayr ‘Amrah s’expliquerait, comme en Occident à partir des Carolingiens et surtout de l’art roman, par la translation en termes de peinture monumentale des motifs des tissus coptes ou bien des argenteries iraniennes. Ainsi, même si l’existence des peintures de Qusayr ‘Amrah à l’endroit précis ou elles se trouvent dans la steppe jordanienne, ne s’explique que par le fait des princes omeyyades, la typologie de Qusayr ‘Amrah – peintures provinciales évoquant [191] l’imaginaire d’un monde royal éloigné et probablement mythique – n’a rien de particulièrement islamique ou même omeyyade. Qasr al-Hayr Ouest et Khirbat al-Mafjar, les deux derniers exemples dont je voudrais parler, sont beaucoup plus compliqués, mais là aussi il est relativement facile de montrer que presque tous les thèmes des arts représentationnels ou 2

Les peintures restaurées n’ont pas encore été publiées. Voir M. Almagro et autres, Qusayr Amra (Madrid, 1975); Oleg Grabar, “La Place de Qusayr ‘Amrah dans l’art profane,” Cahiers Archéologiques, 36 (1988), et “Une inscription grecque,” Mélanges Dominique Sourdel (à paraître).

180

early islamic art, 650–1100

bien ornementaux que l’on y trouve ont leurs ancêtres ou parallèles immédiats dans les diverses traditions de la Méditerranée post-Antique, et de l’Iran, surtout dans ses formes soghdiennes d’Asie Centrale. Les mêmes filiations et parallélismes pourraient se faire autour des techniques employées et des styles les plus courants dans ces deux palais. La coexistence de ces motifs dans un seul bâtiment est certes nouvelle et ne s’explique que par l’existence d’un monde omeyyade et d’un nouvel alignement des grandes régions de l’Asie occidentale et de l’Afrique du Nord. Mais si le fait historique de l’Islam explique ces nouvelles configurations visuelles, l’Islam en soi ne semble pas en avoir été la cause. Au contraire. Bien peu de thèmes, styles, et motifs de ces palais si riches semblent avoir été retenus par la tradition islamique postérieure. La seule exception à cette règle semble être la géométrie des mosaïques de pavement et surtout des panneaux en plâtre qui recouvraient les murs en pierre, une manière de décorer nouvelle pour la Syrie. On peut trouver des antécédents antiques à la géométrie de l’ornement en général comme à beaucoup de motifs individuels, mais ce qui semble original c’est un système de thèmes et variations à partir de rotations et de compressions qui semblent indiquer une réflexion sur la géométrie plane. Or ce genre de réflexion n’est, à ma connaissance, pas attesté dans la science avant le dixième siècle, le siècle de son apparition dans les arts d’Iraq et du Khorasan. Je ne connais pas d’exemples de cette géométrie en Méditerranée juste avant la fin du septième siècle. Des études de détail sur des exemples comme le fragment d’origine palmyréniennne trouvé à Qasr al-Hayr ou comme les princes habillés en byzantin ou en sassanide montreraient facilement que les artisans qui ont copié tout cela pour les princes omeyyades introduisirent des détails dans le drapé ou bien dans les proportions des corps qui sont absents des modèles originaux. Aucune de ces modifications ne peut être expliquée par des caractéristiques ou bien des doctrines exprimées par ou associées avec l’Islam. La sensualité un peu vulgaire des représentations peintes et sculptées dériverait, à mon sens, des préférences d’un mécénat sans grande tradition artistique. Même la présence d’ “orientalismes” iraniens ne requiert nullement le fait historique des Omeyyades, car le prestige de snobisme et l’apport visuel, réel ou mythique, de l’Iran ou même de la Chine existaient déjà dans l’Antiquité tardive. Il est bon de rappeler que le grand savant qu’était René Dussaud pensait que la mosquée de Damas était une église, et le seul aspect “islamique” de la Coupole du Rocher est son extérieur multicolore, vraisemblablement à l’imitation des soieries suspendues sur la Ka’abah de La Mecque.3 On pourrait ainsi conclure que ces juxtapositions déjà identifiées par Herzfeld et Schlumberger, entre autres, comme caractéristiques de l’art omeyyade ne sont que des exemples de l’art de la fin de l’Antiquité méditerranéenne rendus possibles par la richesse et les aspirations du nouveau 3

Oleg Grabar, “The Meaning of the Dome of the Rock in Jerusalem,” Derek Hopwood, ed., Studies in Arab History (Oxford, 1990).

l’art omeyyade en syrie

181

5 Qusayr ‘Amrah: Salle principale, détail: femme à la piscine

182 early islamic art, 650–1100

6 Qusayr ‘Amrah: Salle principale, mur est: moine chrétien (?)

mécénat omeyyade, de même que certains monuments contemporains de Jeddah ou d’Abu Dhabi serviront d’exemples pour l’histoire universelle de l’architecture de la fin du vingtième siècle. En fait, les manuels de Salomon Reinach, de Diehi, de Dalton, comme les grandes hypothèses de Strzygowski, de Riegl, d’André Grabar et de Kurt Weitzmann utilisaient, de manières plus ou moins réussies pour l’historiographie de la fin du vingtième siècle, l’art des Omeyyades comme illustration de la grande koiné visuelle [192] de la fin de l’Antiquité dont les exemples profanes ont presqu’entièrement disparu dans le monde chrétien.4 4

Carl Nordenfalk, Die Spätantike Zierbuchstaben (Stockholm, 1970), pp. 16–17 a écrit cette phrase curieuse: “Vielleicht lassen sich die Jahrhunderte zwischen dem Aufstieg

l’art omeyyade en syrie 183

Nous pourrions simplement en rester là et satisfaire deux critères d’interprétation. Comme toujours, diraient certains, la science occidentale rejette l’identité culturelle des arts autres que les siens et accapare un phénomène discret dans les maillons d’une manière universelle des valeurs et des compétences artistiques. Le critère n’en reste pas moins valable. Dans le passé comme aujourd’hui, à bien des égards les arts n’ont pas de patrie et le seul objectif de l’historien des arts, par exemple ceux de la Syrie, est de persuader ses collègues occidentalistes ou byzantinologues du fait que les distinctions et les oppositions si profondes à un niveau politique et religieux ne jouent pas dans les arts qui suivent un tout autre rythme de développement. Les révolutions culturelles prennent beaucoup plus de temps que les changements économiques ou même politiques. Des groupes profondément antagonistes peuvent parler les mêmes langues et porter les mêmes vêtements. Cela est possible justement parce qu’il y a un langage des arts visuels indépendant de celui des croyants ou des poètes. Et puis ce critère accepte lui aussi la notion de l’inévitabilité de l’Antique, surtout avec ses prolongements iraniens. L’autre critère d’interprétation qui serait satisfait par mes arguments est un critère développé à partir de l’historiographie islamique du Moyen Age. Pour cette dernière, les princes omeyyades n’étaient pas de vrais et bons Islams zur Weltmacht und der Ende der Ikonoklasmus am ehesten eine Übergangzeit zwischen Spätantike und Mittelalter.”

7 Qusayr ‘Amrah: Salle principale, voûte sud: détail des travaux de construction

184

8 Qusayr ‘Amrah: Salle principale, écoinçon sud: musicien et danseuses

early islamic art, 650–1100

musulmans, mais les héritiers de cette sirah qaysar wa kisra, “mode de vie de César et de Chosroès”, que la tradition pieuse rejette, ou bien encore les habitués du mulk, de la royauté honnie par la foi. Leurs monuments, si pleins d’images et de richesse visuelle, sont, presque par définition, des actes de blasphème. En fin de compte, cependant, cette manière de voir les arts des Omeyyades simplement comme une interprétation d’une Antiquité universelle n’est pas entièrement valable. Elle est juste dans la mesure où elle permet d’expliquer pourquoi et comment telle ou telle forme apparaît dans une mosquée ou dans un palais. On peut l’approfondir et trouver des explications uniques et individuelles pour des masses de petits (ou grands) détails que nous ne comprenons pas encore mais dont l’élucidation augmentera l’aspect unique et discret de chaque monument. Cette élucidation privatisera les arts et la connaissance des arts, encore un phénomène où l’idéologie économique contemporaine rejoint notre manière de connaître le passé. Dans cette perspective, les grandes explications “universelles” deviennent, à l’étude, des masses d’événements concrets. Le choix entre l’universel et le concret est fait par celui qui regarde et, à la manière de la critique contemporaine des arts, il est peut-être juste que chacun trouve, même dans les arts du passé, ce qu’il veut y chercher. Deux aspects de l’art des Omeyyades et de la réflexion musulmane contemporaine sur les arts ne seront pas, toutefois, expliqués par cette

l’art omeyyade en syrie

185

manière de comprendre les arts. L’un de ces aspects est symbolisé par la réforme monétaire d’Abd al-Malik qui crée des monnaies sans iconographie mais avec un texte. On peut ne pas considérer ces monnaies comme des œuvres d’art et je ne suis pas de ceux qui y voient les premiers pas d’un art calligraphique pour l’écriture arabe, mais elles entrent dans la langue visuelle de l’époque. Elles encadrent la vie courante tout comme les premières inscriptions des tissus. Ce qu’elles font en fait c’est d’éliminer entièrement la possibilité d’un choix d’interprétation – tel qu’il est possible pour les mosaïques de la Mosquée de Damas – et être immédiatement reconnaissables comme différentes [193] des monnaies byzantines ou sassanides; elles sont donc clairement arabes et musulmanes. Sans être des oeuvres à fins esthétiques, elles modifient le contexte culturel par lequel on regarde et on comprend. L’autre aspect nouveau est le début d’un débat sur la représentation d’êtres animés, pas seulement d’images. Le contenu de ces débats n’est pas important pour mon propos. Ce qui importe c’est que le caractère du mécénat des Omeyyades et les opportunités qui leur furent ouvertes par leur richesse ont obligé la culture islamique naissante à trouver son attitude envers les arts visuels, car cette culture rencontra sous les Omeyyades le monde des arts – le monde de choix esthétiques et de besoins visuels – qui lui était jusqu’alors presqu’inconnu, qui était absent en Arabie et assez peu développé en Iraq. A ces deux niveaux, celui d’une langue visuelle et d’une réflexion sur la nature de la représentation et de l’art, l’époque des Omeyyades a créé les premiers éléments d’une culture artistique et d’un discours sur les arts qui se devait d’être différent de ce qui l’entourait et de ce qui l’avait précédé. L’importance et l’intérêt de l’époque omeyyade pour l’histoire des arts islamiques résident ainsi moins dans ses monuments d’art, si nombreux et si passionants qu’ils soient, que dans les questions plus profondes qu’elle souleva sur les fonctions mêmes de l’art dans la vie des hommes.

Chapter XII Umayyad Palaces Reconsidered*

The latest, most complete, and most authoritative survey of early Islamic architecture – Allan’s reworking of Creswell’s standard history – contains seventeen partly documented buildings datable between 690 and 750, the main decades of Umayyad rule, and fitting into the general category of palaces. It is a loose category including nearly all foundations with living accommodations for which a reasonable assumption can be made of sponsorship or use by the state (a dar al-‘imara for instance), by the ruling dynasty, or by members of the Arabian aristocracy associated with the Umayyads.1 * First published in Ars Orientalis, 23 (1993), pp. 93–108. This paper is neither a coherent reevaluation of Umayyad palaces nor quite a repetition of remarks begun some forty years ago for my doctoral dissertation at Princeton University. It is rather a moment in a continuing meditation on the first monuments of Islamic art, perhaps with a different set of questions and interpretations than I and others developed in the 1950s. 1 K. A. C. Creswell, A Short Account of Early Islamic Architecture, revised by James W. Allan (Aldershot, 1989). The buildings are in the order in which they appear in Creswell, and in what is also supposed to be their chronological order: Kufa, Khirbat al-Minyah, Jerusalem, Qasr Kharana, Qusayr ‘Amrah, Jabal Says, ‘Anjar, al-Muwaqqar, Qasr alHayr West, Rusafa, Qasr al-Hayr East, Amman, Qastal, Khirbat al-Mafjar, Mshatta, Qasr al-Tuba, Tulul al-Sha‘iba. I have not included in this list places like Qasr Burqu‘ or Bayir, where there were Umayyad foundations but which are too damaged to allow for significant architectural rather than archaeological or geographical conclusions. Nor does this list include buildings like the Khadra in Damascus which are known to have existed but about which too little can be reconstructed, even though some of these constructions, like the Damascus palace of Mu‘awiya or Wasit’s palace of al-Hajjaj, in their time or eventually acquired a mythical value. Both of these buildings are mentioned in Creswell–Allan, but there are many places like Fustat, Basra, Merv, and probably other cities for which written sources yield some information. In the absence of a reasonably complete survey of these sources, occasional accidental references would have been misleading. I have also not counted in the list of seventeen “palaces,” buildings like Qasr Hallabat in Jordan or Raqqa in Syria, where recent excavations have found significant Umayyad uses and modifications of older buildings or Umayyad secular buildings like the ones being excavated in Aqaba or Humayma in Jordan which may have been palaces, but whose investigation is still incomplete. Finally, I have left aside the full investigation of the long list of possible Umayyad princely establishments made, partly as a heuristic exercise, by Jean Sauvaget, “Observations sur les monuments omeyyades,” Journal Asiatique 231 (1939). This list contains many places which are

187

188

early islamic art, 650–1100

Out of these seventeen buildings, five – al-Muwaqqar, Rusafa, Qastal, Qasr al-Tuba, and Tulul al-Sha‘iba – are too poorly known or too poorly preserved to allow for significant conclusions,2 two – Anjar and Amman – pose a very complicated set of problems in the interpretation of archaeological evidence, which, in my opinion, does not necessarily lead to the Umayyad dates proposed by the investigators of these sites, and four – Jerusalem, Qasr Kharana, Jabal Says and Qasr al-Hayr East – are not palaces or even royal or dynastic residences, at least not in my judgment, although this judgment is not necessarily shared by all scholars.3 This process of elimination leaves six relatively well-known buildings – Kufa, Khirbat al-Minyah, Qusayr ‘Amrah, Qasr al-Hayr West, Khirbat alMafjar and Mshatta – which can be reasonably assumed to be dynastic foundations of the Umayyad dynasty. For Kufa the evidence is essentially historical in the sense that written accounts of events from the last decades of the seventh century and the first ones of the eighth justify the interpretation of an excavated building as the dar al-‘imara of the Umayyad governors.4 At Khirbat al-Minyah, Khirbat al-Mafjar and Qasr al-Hayr West (all of which were first thought by archaeologists to be pre-Islamic Christian foundations), inscriptions were discovered which suggest that some aspect of the construction of these buildings was sponsored or ordered by a ruling Umayyad

2

3

4

merely settlements without palatial functions or whose exact dates are still under discussion. Actually quite a bit is known about Rusafa and about Qastal, but much of the available information is not easily accessible. For Qastal, for instance, the activities of Patricia Carlier are summarized in her thesis for a Doctorat de troisième cycle from the University of Aix-en-Provence, Qastal, château umayyade (Aix, 1984). For Rusafa, the pertinent results of the renewed German excavations have not yet, to my knowledge, been published. I am sure that departments of antiquities and private notes of travelers and archaeologists contain large quantities of useful information. For Qasr Kharana, see the rather complete study and original interpretation by Stephen K. Urice, Qasr Kharana in the Transjordan (Durham, NC: American Schools of Oriental Research, 1987). For the other three sites, see the comments provided in Oleg Grabar et al., City in the Desert (Cambridge, 1978), pp. 151–2. On Jerusalem the only easily accessible source consists in the very impressionistic and personal accounts of Meir BenDov, In the Shadow of the Temple (New York, 1982), esp. pp. 273–321. More detailed views on these buildings will appear in my forthcoming book on the early medieval city of Jerusalem. In my view, the date and functions of the Amman citadel have not yet been resolved, in spite of the assiduous and meticulous works by A. Almagro, El palacio omeyyade de Amman (Madrid, 1984), and by Alistair Northedge in his unpublished dissertation on the topic completed in 1982 at the University of London. ‘Anjar’s problem is quite different. Too little has been published of what was discovered there and its size as well as apparently commercial functions do not fit with what is known of Umayyad economic policies in this area as opposed, for instance, to northern Mesopotamia where new trade routes can be assumed. Furthermore, its alleged palace does not look like known Umayyad ones. Most of the immediate evidence on Kufa is found in Creswell’s larger volumes, pt 1, pp. 46 ff. But, since Kufa remained as a major and very lively city for several centuries, it should be possible to analyze and interpret the continuing use of this building, a possibility which cannot be pursued with the evidence which has been published.

umayyad palaces reconsidered 189

caliph or else that a caliph was alive while the building was being built. But not one of these inscriptions can be considered as a foundation statement or as decorative writing in the manner of later Islamic foundations or ornamental inscriptions.5 For Qusayr ‘Amrah, as we shall see shortly, there is a valid presumption for an Umayyad dynastic patronage on the basis of the paintings decorating the monument. As to Mshatta, it is primarily its inordinate size and its peculiar decoration that make it reasonable to assume that only princes, and in all likelihood ruling princes, had access to the funds and personnel necessary for its planning and construction, and for a completion which never took place.6 In short, we have no direct knowledge of these buildings as palaces built for ruling Umayyad princes, and I shall return in my conclusion to the hypotheses which can be derived from this apparent absence of clearly cut, written or archaeological, external labels. Nor can we really argue that their shape, the functions they imply, or their decoration make it necessary for all of them to be Umayyad creations. In fact, as I reread some of the papers, articles and books I wrote on them over the years, I realize that there may have been a major methodological flaw in parts of my reasoning. Following very respectable masters like Sauvaget, Herzfeld, Stern, and in his own way Creswell, I consistently tried to fit all these buildings into a single pattern by assuming that they all reflected the same architectural and functional type: a villa with living and formal or official quarters arranged around a porticoed court in standard square buildings with a single, usually massive, entrance; a bath, most of the time an independent building; a mosque; a variety of service areas; and possibly formal gardens or orchards. The type would have

5

6

At Qasr al-Hayr West and at Khirbat al-Mafjar fragments of statements written casually on marble plaques mention the name of the caliph Hisham, once as the sender of a message, the other time as the recipient of one. Both statements contain the names of otherwise unknown and unidentified individuals, and it is extremely difficult to imagine the circumstances which would have led to the writing of these messages on marble plaques presumably destined to be used in the buildings. They are correctly interpreted as graffiti, not as inscriptions. See Daniel Schlumberger, “Kasr el-Heir,” Syria, 20 (1939), p. 373, and R. W. Hamilton, Khirbat al-Mafjar (Oxford, 1957), p. 43. The written fragment found at Khirbat al-Minyah is possibly a genuine inscription since it begins with the traditional invocation of God’s name and with the words mimma amara bihi, “what has ordered,” the typical beginning of thousands of medieval inscriptions on objects or buildings. The problem is that the stone with the inscription is reused in the palace and there is no way of knowing to what it referred. See O. Puttrich-Reignard, “Die Palästanlage von Chirbet el-Minje,” Palästina-Hefte des Deutschen Vereins vom Heilige Lande 17–20 (1939). The three documents are quite different from clear-cut foundation inscriptions like the one from Qasr al-Hayr East. Elsewhere I have argued for an early ‘Abbasid date for Mshatta (Oleg Grabar, “The Date and Meaning of Mshatta,” Dumbarton Oaks Papers, 41 [1987]), but it is only slightly later than the one usually given to the palace and does not affect the reasoning of this paper. See now the excellent summary by Volkmar Enderlein and Michael Meinecke, “Mshatta-Fassade,” Jahrbuch der Berliner Museen, 24 (1992), with a discussion of epigraphic information I did not know.

190 early islamic art, 650–1100

developed within the rich tradition of secular architecture in the late-antique Mediterranean area. And, following primarily Sauvaget but frequently using documentation gathered much earlier by Lammens and Herzfeld,7 I further assumed that this particular type could be fitted into a reconstruction or a model of Umayyad aristocratic behavior. But that reconstruction of behavior was based on very different sources: written texts in which an originally orally transmitted poetry predominates; archaeological sources involving the interpretation of excavated spaces and other remains; and finally a more complicated source which may be [94] called “ecological logic” and which consists in apparent changes within the lands ruled by the Umayyads which require significant investments of money and labor.8 The resulting hypothesis was a simple one. Umayyad patrons, a nouveau riche class of aristocrats from central Arabia, invested in the land they inherited or conquered and in a life of varying but usually considerable luxury. The palaces illustrate the setting for that life. Variations and peculiarities in the model were then seen as aberrations reflecting individual needs or preferences in taste or behavior for which no explanation was likely to be found. A case in point is that of the twin quarters of Qasr al-Tuba for which some romantic or family explanation could always be invented and proposed but no proof can be found.9 To the historian of architecture, that is to say, of techniques which require considerable investments of funds and differentiated competencies from specifically trained practitioners, the assumption of types adapting rapidly or slowly to new patrons was then and remains now a reasonable one, for the hitherto unknown sponsorship of architecture by rich aristocrats from northwestern Arabian cities did not necessarily demand new functions or seek to satisfy new needs. This patronage could easily accommodate itself to the prevailing habits of building and of designing within the Mediterranean world, and one or two technical innovations from elsewhere – as, for instance, the covering of stone walls with stucco – easily found a place in an essentially traditional system. A typical problem of design arose from the new requirement for an oriented place of prayer. The mosque became a compositional problem with several early and more or less successful solutions 7

8 9

There is no one statement of Sauvaget’s definitive opinions on the subject of Umayyad palaces. The most important ones are “Remarques sur les monuments omeyyades,” Journal Asiatique, 231 (1939); and “Châteaux omeyyades de Syrie,” Revue des études islamiques, 39 (1967), published posthumously. But he made remarks on these “palaces” on many other occasions. H. Lammens, “‘La Badia’ et la ‘Hira’ sous les Omayyades,” Mélanges de la Faculté Orientale, Univ. St Joseph, Beirut, 4 (1910); E. Herzfeld, “Mshatta, Hira und Badiya,” Jahrbuch der preussischen Kunstsammlungen, 42 (1921). One example of ecological studies with implications for Umayyad palaces is A. M. Watson, Agricultural Innovation in the Early Islamic World (Cambridge, 1983). Such is essentially my position; see Oleg Grabar, The Formation of Islamic Art, 2nd edn (New Haven, 1989), and Richard Ettinghausen and Oleg Grabar, Islamic Art 650–1250 (London, 1985).

umayyad palaces reconsidered

191

until buildings apparently began to be oriented in order to meet the necessity of fitting a mosque into them. Oratories appear, most commonly to the right of the entrance (Mshatta, Ukhaydir).10 The more original aspect of these Umayyad secular buildings is how they transformed, or at least modified, the architectural profile of an area by introducing the amenities of a wealthier and restricted private life of “masters” into a simple landscape of farms and villages with their churches and occasional civic buildings such as baths or into old Roman military cities.11 The Umayyads, this hypothesis implied, did not really invent much that is new; they simply put whatever they found in spaces which had not previously been used for these purposes. There are several other, more specific issues of architectural history, like the technology of building, for instance, which come up whenever there is a combination of a set of buildings from a restricted period of time, in a relatively limited space, and with more or less the same patronage.12 Beyond elementary measurements and the establishment of the most basic modular principle of designing and with the exception of the question of brick in Syria and Palestine about which a few things have been written, these technical issues have hardly been touched, however important they are for establishing the technological potential of a time. A more specific approach to the establishments attributed to the Umayyad dynasty can be derived from a number of puzzles in the architecture and decoration of remaining buildings. These puzzles led historians of art, of visual forms, to seek an explanation in the history provided by texts. Such features as the elaborate reception rooms of Mshatta and the rich decoration of Khirbat al-Mafjar did not quite coincide with the simple-minded image created by earlier historians such as Lammens and Creswell of an easy-going, fun-loving, but not very serious life attributed to freely moving and seminomadic Umayyad princes and assorted aristocrats. Nor is fancy decoration necessary for landlords in their manorial and private estates. As a result, several studies, whether or not they acknowledge the original inspiration of 10

11

12

A thorough study of early solutions to the problem of finding a place for mosques in a language of planning that did not have them lies beyond my purposes here, but I do want to note that many early Islamic buildings (Ziza, Qastal, Umm al-Walid, Qasr Hallabat, all in Transjordan) have small mosques outside the main residential complex. The landscape of Christian times can best be seen in books such as G. Tchalenko, Villages antiques de la Syrie du Nord (Paris, 1959). Much has been accomplished since that time, and numerous excavations and surveys, especially in Jordan, have modified some of the views developed a generation ago, especially around the extent and character of the cultural continuity of the area. Consult the reports found in various archaeological journals. For a recent publication which touches on many of the pertinent problems, see P. Canivet and J.-P. Rey-Coquais, La Syrie de Byzance à l’Islam (Damascus, 1992). See, as an example, the work of J. Ward-Perkins on vaults in University of St Andrews, The Great Palace of the Byzantine Emperors, 2nd report (Edinburgh, 1958). On a much smaller scale similar analyses have been carried out on Palestinian churches and synagogues.

192

early islamic art, 650–1100

Sauvaget for the method involved, sought to describe an Umayyad ceremonial life and to fit that life within the ruins or reconstructions of whatever remained from alleged palaces.13 The best among these studies, notably Hamilton’s meditation on al-Walid and Khirbat al-Mafjar, have provided a very tempting and coherent picture, based to a large degree on poetry, of an unstable equilibrium between formality and fun, between rigidity and freedom, all of it bathed in virile sensuality. It is a coherent picture in the sense that the poetry of the time and especially the ways in which this poetry was recited in male gatherings seem to fit with the spaces provided by the ruins. But there are problems. For instance, too many of the examples of almost anything in these studies deal with al-Walid II, who was in many ways an eccentric hardly typical of his rank and of the office he occupied. And there are always dangers [95] in excerpting examples from literary contexts whose own rules have not been ascertained. In general, there are many theoretical and concrete pitfalls in explaining architecture through poetry, unless, as for instance with much later Persian lyrical poetry, the poetry itself deals with architecture or, as in the Alhambra, ad hoc poetry adorns architecture.14 The danger exists because the sphere of life and of behavior in which poetry appears is the immediate and the evanescent; much of it was not written down at the moment of recitation, but at the moment when the event in which it participated was recorded. Architecture, on the other hand, is more permanent, unless, of course, we consider decoration separately from building on the grounds that some of its techniques – especially painting and carved or molded stucco – were themselves changeable or could be covered with other materials such as textiles.15 The skin of architecture, like that of snakes, can be changed, at least in part. 13

14

15

J. Sauvaget, La mosquée omeyyade de Médine (Paris, 1947), is the first book to have used this approach of blending written and visual documentation. For later studies, see Oleg Grabar, “Notes sur les cérémonies omeyyades,” in Studies in Memory of Gaston Wiet, ed. Myriam Rosen-Ayalon (Jerusalem, 1977); Robert Hillenbrand, “La Dolce Vita in Early Islamic Syria,” Art History, 5 (1982); and especially Robert Hamilton’s admirable AlWalid and His Friends (Oxford, 1988). In reality the historiography of this particular approach is older than Sauvaget, as Lammens in particular and Herzfeld were inspired by buildings to search for texts, but they did not look for ceremonies and for behavior, something which has been forced on a scholarship affected by the decoration which came out of excavations for the most part. For the Alhambra, see Oleg Grabar, The Alhambra (London, 1978). To my knowledge no study has been devoted specifically to the references to architecture in Persian poetry or in Ottoman panegyrics, whether descriptions of buildings or metaphors and images using architecture. I owe to Professor Necipoglu the reference to A. S. Levend, Türk Edebiyattnda S¸ehrengizlerve S¸ehrengizlerde Istanbul (Istanbul, 1957). A typical early Islamic example of changing textiles is that of the Ka‘ba in Mecca, which was covered anew every year with fancy embroideries; al-Azraqi, Akhbar Makka, in Geschichte der Stadt Mekka, ed. G. Wüstenfeld (Leipzig, 1858), 1, 104 ff. Another kind of change using textiles was the temporary modifications made for special occasions, as

umayyad palaces reconsidered 193

A remarkable feature of Umayyad architecture is the quantity as well as the quality of the decoration it exhibits, still in situ, as in the Dome of the Rock and the mosque of Damascus, or through masses of fragments, some in place, others reconstructed into their original compositions, many still hiding in unopened crates in the basements of museums or exhibited as bulk remains in their drawers and galleries.16 The curious point is that the decoration of the main Umayyad palatial establishments has been less well studied than architecture or planning in that period, probably because Creswell spent relatively little time on the decoration, much of which was discovered after the first edition of his work. The only partial exception is that of classical art-historical issues dealing with genetic sources which have occupied a few scholars, especially in the early decades of the twentieth century.17 This absence of studies is all the more peculiar since the originality and the variety of decoration in Umayyad palaces are quite astounding. Yet, precisely because of its variety and originality, the study of this decoration leads to a number of methodological problems. Among approaches which have been tried is one I used for the first time some fifty years ago, and which was probably even more dubious than the procedure I used for architectural planning and design. Following the model of Christian and Buddhist art, I sought to find in the decoration of walls, floors and ceilings a set of types, that is to say, of common standards with local variations. The assumption, valid for religious arts with an ecclesiastical patronage, was that models (actual ones on monuments and objects or written ones in manuals) must exist for any visual program to be executed and to be understood. In reality, of course, the secular arts operate quite differently from religious art. It is therefore more fruitful and, in fact, necessary to attempt to understand the individual combinations of themes and forms located in one monument.18

16

17

18

when the Baghdad palaces were festively transformed through textiles for the arrival of a Byzantine embassy; see, among many places, Grabar, Formation, 168 ff. For a more general theory of textile aesthetics, see Lisa Golombek, “The Draped Universe of Islam,” in Content and Context of Visual Arts in the Islamic World, ed. P. Soucek (University Park, Penn., 1988). The principal structures with decoration are Khirbat al-Minyah (practically only mosaics), Qasr al-Hayr West (paintings and sculpted stucco), Khirbat al-Mafjar (mosaics, stucco, and sculpture in stone and in stucco), Qusayr ‘Amrah (paintings), Qastal (mosaics), and Mshatta (stone sculpture, a few stucco fragments). The basic bibliography on all of these except Qastal can be found in Ettinghausen and Grabar, Islamic Art and Architecture, pp. 26–70. For Qastal, see above. Genetic discussion of the origin of motifs was central to much of what has been written on the façade of Mshatta and on the mosaics of the Dome of the Rock. Most of it derived from the complex ways of analyzing the decorative arts developed out of grammars of ornament on the one hand and the theoretical studies of A. Riegl on the other. I first approached the subject of a general system in Umayyad art in my dissertation, “Umayyad Art and Ceremonies,” Princeton University, 1955. It is outdated in many ways and in error in some, but its impact can be found in most of the works quoted in n. 13.

194

early islamic art, 650–1100

The establishment of vocabularies and grammars of forms and the definition of typologies are proper endeavors, but because in secular art almost every statement through a given object or monument is a different combination of terms and means something different, typologies must follow rather than precede individual analyses. In a nutshell, system follows syntagm.19 In the absence of formal manuals and of a doctrine about the arts, the only way one can eventually reach a sense of the language used in the art of Umayyad palaces is first to provide reasonable hypotheses about every monument by itself. Such hypotheses are all the more important since every one of the monuments involved provides a range of information which is peculiarly its own. I shall illustrate the point with two examples. The first one is that of Qasr al-Hayr West, for which the last written and drawn statements by the excavators, all of whom are now deceased, have recently been published.20 The second one is Qusayr ‘Amrah, the first of the Umayyad dynastic establishments to have come to light, and the only one which is not in ruins. It underwent a major overhaul some twenty years ago, but its complete actual appearance today is, rather shamefully, not yet available in print.21 I shall be brief about Qasr al-Hayr. In addition to the sculptures and paintings which are well known, have been visible for over a generation in the Damascus Museum, have been mentioned in general books and manuals, and have even been subjected to elementary scholarly analyses, we are presented with painted fragments which were exhibited occasionally but never published and with stone and stucco sculptures of personages known only to visitors to the storerooms of the Damascus Museum.22 It is nearly impossible to reconstruct how these paintings and sculptures fitted anywhere, except for a few fragments on the façade of the palace rebuilt in the Damascus Museum and for two large paintings found on the [96] floor of the staircases.23 It is, therefore, nearly impossible to propose a program for these images, or even sequences of visual impressions with some iconographic or expressive meaning. This is so even when the sources of a motif are clear. For instance, 19

20 21

22

23

The major exception to the rule lies in coinage, where both written sources and existing series of coins emphasize the necessity for a system of forms in order for any one item to be understood. Daniel Schlumberger, Qasr al-Heir al-Gharbi (Paris, 1986). See M. Almagro and others, Qusayr ‘Amrah (Madrid, 1984), which unfortunately does not give a complete survey of the restored frescoes. It is also unfortunate that there is no account of the degree of restoration which has been provided in some instances, nor is anything really known about the technical details of the paintings. Bibliographic references will be found in Creswell–Allan, Early Muslim Architecture and in Ettinghausen and Grabar, Islamic Art. Some pieces have been shown in various traveling exhibitions organized by the Syrian Department of Antiquities, especially the one held at the Musée du Monde Arabe in 1990–91 which focused on Umayyad palaces. Their catalogs provide a few additional pictures and minimal comments. Daniel Schlumberger, “Deux fresques omeyyades,” Syria, 25 (1946–48); Richard Ettinghausen, Arab Painting (Geneva, 1962), pp. 35, 37.

umayyad palaces reconsidered

195

1 So-called Palmyrene sculpture from Qasr al-Hayr West

two sculptures found at Qasr al-Hayr resemble representations of rulers in Byzantine and Sasanian art.24 But if we argue, as I and others have done, that, because these had been forms for the representation of pre-Islamic rulers, they must now be representations of Umayyad princes, then we also have to argue that a sculpture with obvious Palmyrene funerary associations (Fig. 1) should be connected to some funeral theme in Umayyad times.25 What then is it doing on the façade of the building? A similar position of iconographic skepticism can be argued for almost every fragment found at Qasr al-Hayr West. But the same skepticism need not apply if one tries to identify, even tentatively, not so much the subject-matter as the sources of inspiration for the images. Within the limits and restrictions of existing information and investigations, the genetic pool from which these images came may be the only kind of meaning or semantic range available for the decorative motifs from Qasr al-Hayr West which can be traced with some

24 25

Oleg Grabar, “Islamic Art and Byzantium,” Dumbarton Oaks Papers, 18 (1964). It is, of course, true that as an exercise in pure connoisseurship the Qasr al-Hayr sculpture differs in many details from actual Palmyrene works. Does it matter? How different must an imitation be in order not to be meaningful as an imitation? Conversely, are meanings carried together with forms?

196 early islamic art, 650–1100

certainty. The most interesting and most unexpected point about this genetic pool seems to be the close relationship between the paintings and sculptures from Qasr al-Hayr and the art of Soghd in Central Asia, and even the Buddhist and Manichean arts found farther east in the Tarim basin.26 The question is how did themes and styles from the northeastern frontier of the Muslim empire (and even beyond) reach Syria, when none of the patrons of Syrian and Palestinian buildings had ever set foot there and the movement of artisans, while not impossible, seems technically unlikely at this time. The answer lies almost certainly in transmission through objects – silver objects sent as tribute or as gifts, textiles, perhaps rugs. Books are not likely to have been important at this time, a few decades before the appearance of paper through the same route, but painted bark and bone could have been among the marvels brought from the east.27 It is also through textiles and works in metal that Mediterranean motifs can be imagined to have reached a remote site like Qasr al-Hayr and through silver objects that Iranian ones would have become available.28 Thus the evidence from Qasr al-Hayr West does not lead one to interpret the decoration on the walls within a social, official, or personal setting, but it does explain how certain combinations of forms ended up in a striking location of the Syrian steppe. We become informed on the movement of works of art and of artistic motifs across western Asia rather than on their meaning in the setting they eventually adorned. This is so because the heterogeneity of the vocabulary of the palace seems to me to be its only characteristic which lends itself to a positive conclusion. Let me turn now to the bath house of Qusayr ‘Amrah, which is only one of a larger group of constructions including now-ruined living quarters, a watch-tower and a hydraulic complex. Its discovery a century ago affected nearly all historians of Islamic culture because of the paintings which adorn 26

27

28

There is no convenient access to the immensely rich treasures of early medieval art from Central and Inner Asia. A very general survey is Tamara Talbot Rice, Ancient Arts of Central Asia (New York, 1965). Rich and accurate, but geographically more restricted, information is found in the published catalog of an exhibition that never took place, Culture and Art of Ancient Uzbekistan, 2 vols, in Russian and English (Moscow, 1991). A little-known and unexpected source of major importance for Qasr al-Hayr as well as Khirbat al-Mafjar is Varakhsha in Uzbekistan, which has friezes of animals strikingly similar to the Umayyad ones; see V. A. Shishkin, “Arhitekturnaia Dekoratsia Dvortsa v Varahshe,” Hermitage Museum, Trudy Otdela Istorii Kultury i Iskusstva Vostoka, vol. 4 (Leningrad, 1947). For further examples see my comments on the paintings of Mafjar in Hamilton’s publication of the palace and A. von Le Coq, Die Buddhistische Spätantike (Berlin, 1922–33), 1: pl. 20 and ff.; 5: pls 12 and 33, comparable to fragments from Mafjar and Qasr al-Hayr. It is well ascertained that considerable booty came from the conquest of Central Asia and that taxes were at times paid in objects rather than cash. A complete gathering of all appropriate texts would be a very worthy enterprise. The arguments justifying this position are much beyond the range and purposes of this paper. From a methodological point of view I am trying to distinguish artistic transfers which took place through objects from the ones made by traveling artisans or artists.

umayyad palaces reconsidered

197

the bath and whose subjects seem to reflect the life of princely leisure. Now we have the advantage of possessing all the information we are ever going to get and of having it in situ.29 It is possible to reconstruct the probable models for many of the paintings. Just as at Qasr al-Hayr, objects, mostly from the Mediterranean, but also from the Sasanian world, had a clear impact on the paintings. Moreover, as I have already pointed out elsewhere, there were also many images inspired by a direct perception of the actual setting of the building and of the events which took place in it. This was not done in order to present the illusion of a physical world, but in order to commemorate events which had either taken place or been recalled there.30 In fact, those images which are new, without a clear model, seem often to be more caricatured and more awkward than the conventional representations in the bath. It is as though someone, a patron no doubt, had requested topics that may have been a bit outside the competence of the painters, for the whole place has the appearance of a lively amateur exhibition. The more important point, however, is that even here, where about five-sixths of what had once been there is preserved, no clear sense of what is shown emerges. This may well be an example of the art-historical principle that it takes two to tango, or that a unique building is almost impossible to understand, especially when additional contextual evidence is missing. It is tempting to give up, just as it is tempting to excerpt something immediately understood like the Six Kings and to elaborate from there on some grand iconographic message.31 So far, to my [97] knowledge, none of these approaches has worked, although admittedly so little has been written about this fascinating little place that even methodological judgments are dubious. What can be proposed is something akin to chemical tests used to determine the composition of some product, that is, try possibilities until something works. Here are two examples. The first one is difficult to illustrate because, when the photographs I used were made,32 I was concerned with recording two-dimensional painted

29

30 31 32

The restoration work was accomplished by a team sent by the Spanish government. The results of its work are rather impressive, as can be seen in the book by M. Almagro and others quoted earlier. There are, however, several places where the paintings visible now differ from the fragments seen by Musil a century ago and photographed by the Reverend Fathers Jaussen and Savignac, Mission archéologique en Arabie, vol. 3 (Paris, 1922). While important for the overall interpretation of the bath’s paintings, these queries are not particularly pertinent to the point of this essay. Oleg Grabar, “La place de Qusayr ‘Amrah dans l’art profane,” Cahiers archéologiques, 36 (1988). Oleg Grabar, “The Painting of the Six Kings at Qusayr ‘Amrah,” Ars Orientalis, 1 (1954). This is an appropriate occasion to recall the help of Fred Anderegg, for some time Head of Photographic Services at the University of Michigan, who came with me on so many Syrian and Jordanian trips and who took and often developed his pictures in most unusual places.

198

2 Qusayr ‘Amrah. View toward right wall

early islamic art, 650–1100

panels on walls. I did not try to record photographically or even to determine on the spot the impressions made on users and visitors from various points of view logical and unique to this particular place: the entrance, doorways, the floors on which one sat, the apse or throne niche in the center of the building, and so on. Pending verification at Qusayr ‘Amrah itself, I propose two visual impressions as correct assessments of the building’s impact. One concerns the wealth of pictures hanging on the wall: there are so many of them that one can hardly make sense of them as a whole, and one moves almost automatically to consider them individually. Just as in a museum, one is overwhelmed by the quantity of pictures, and one can only handle them one at a time. I am unclear about the psychological implications of this cascade of ever-present images. The second impression is that the first and richest hall of the bath, usually interpreted as a throne hall, is dominated by the image of an enthroned ruler in the apse and by a series of large panels in the right vault ending at the vault’s back wall (Fig. 2). The comparable area in the left vault is cut up into smaller sections and visually less effective from a distance. The question is whether this visual judgment, if accurate, justifies arguing that the woman by the pool (Fig. 3), the Six Kings, and the “Lady Niké”

umayyad palaces reconsidered

199

3 Qusayr ‘Amrah. Standing nude woman on right wall in main hall

200

4 Qusayr ‘Amrah. “Lady Niké” on back wall of right vault

early islamic art, 650–1100

umayyad palaces reconsidered

201

5 Qusayr ‘Amrah. Standing women on either side of ruling figure

202

6 Qusayr ‘Amrah. Standing women on either side of ruling figure

early islamic art, 650–1100

umayyad palaces reconsidered

203

7 Qusayr ‘Amrah. Dancing woman on arch in main hall

204

8 Qusayr ‘Amrah. Dancer accompanied by guitar player on spandrel of central vault

early islamic art, 650–1100

(Fig. 4)33 are, with the prince in the back of the apse, more significant than other images. One could propose several other similar readings of the walls from a variety of points of view corresponding to the likely uses of the building. The result of this kind of investigation would be to establish hierarchies according to the way the paintings are perceived, but I am not sure that such hierarchies would not themselves become arbitrary constructs rather than useful explanations. My second approach is even more subjective, at least in its premise. A striking feature of Qusayr ‘Amrah is the extraordinary number and variety of women represented on its walls in a truly minuscule space: half-dressed 33

The representation on this wall is of a well-dressed woman in a tent with various figures around her. To the right of the tent, in the upper part of the panel, there is a Greek inscription reading quite clearly NIKE, “Victory.” In an article published a few years ago, I connected NIKE known for a hundred years with a number of Greek letters discovered by the Spanish restorers to have been painted on the left of the tent at the same level as NIKE. In “Une inscription grecque à Qusayr Amrah,” Revue des études islamiques (Mélanges Dominique Sourdel), 54 (1986), I argued that there was a single name and proposed that it be read ARIS [TO] NIKE. Unfortunately, I no longer believe that this reading is possible for a whole series of reasons. A suggestive alternative will be proposed by Dr Garth Fowden in a forthcoming book which, if acceptable, will be of extraordinary importance for the understanding of Qusayr ‘Amrah. I am most grateful to Dr Fowden for having shared his interpretation with me and for many wonderful discussions of Qusayr ‘Amrah. See Garth Fowden, Empire to Commonwealth (Princeton, 1993), pp. 143–9.

umayyad palaces reconsidered 205

9 Qusayr ‘Amrah. Standing nude woman in central hall

206

10 Qusayr ‘Amrah. Alighting (?) woman, on left wall of main hall

early islamic art, 650–1100

ladies around the prince (Figs 5 and 6), Lady Niké (Fig. 4), the tall lady by the pool (Fig. 3), nude dancers (Fig. 7), a heavily dressed dancer (Fig. 8), another standing naked woman (Fig. 9; now in the Berlin Museum), a woman apparently alighting in front of a man (Fig. 10), meditating personifications identified by Greek inscriptions (Fig. 11), busts of welldressed women in the central nave of the main hall as though looking at the goings-on below (Fig. 12), a cogitating nude woman in one of the side rooms for bathing (Fig. 13), and partly damaged paintings of totally naked women and children also found in one of the small bathing rooms that seem to depict domestic activities that would have taken place in precisely such a room (Figs 14 and 15). Almost none of these representations of women finds adequate parallels in comparable or reasonably comparable artistic traditions, except occasionally in details like the stubby and awkward legs so typical of Coptic textiles, poses and gestures clearly belonging to the ways of representation in lateantique art, and occasionally the memory of a classical personification or of a Sasanian symbol. The point I would like to make, however, is a different one. While the presence of so many women predominates at Qusayr ‘Amrah, it is striking that they are so different from each other. Sensuous performances cohabit with highly proper ones, possible narratives coexist with apparent scenes of leisure and domestic life, and trite personifications are found alongside

umayyad palaces reconsidered

207

11 Qusayr ‘Amrah. Personifications on the side of windows

concrete references. Should one try to argue that what is shown or reflected in Qusayr ‘Amrah is the world of the women whose power, both political and cultural, was, in early Islamic times, considerable? Or should one interpret all of this as a men’s locker-room view of the world? A preliminary argument can be made for either possibility, and maybe others exist as well. The full appreciation of this strange and fascinating museum of rather mediocre paintings most probably requires the elaboration of a series of categories of understanding, one of which may even be the traditional art-historical one of specific iconographic meanings and of an equilibrium of sources from many origins.

208

12 Qusayr ‘Amrah. Female figure in upper part of central nave

early islamic art, 650–1100

What conclusions can one draw from these remarks? The most important one is that Umayyad [98] palaces, or at least a small number among them, contain a considerable amount of data which should at this time be considered syntagmatically, that is, in terms of the unit in which they are found, rather than systemically, in terms of the separate histories of each example. Before returning to types and to broad vistas, it may be worthwhile to plunge into the depth of each palatial establishment, to study all of its details, to imagine how it was built and what went on in it. This sort of analysis should not stop at information provided by the monument or by written sources about its time, nor even by the sources which can be identified for any part of it. It is perfectly appropriate to bring to bear everything from gender studies to technological analyses which can develop a grip on the monument. Such a monographic approach, like a proper trench in an archaeological investigation, would at least begin to show that each one of some six buildings can bristle with specific references and interpretations, acquiring a full personality of its own. The acquisition of such precise information together with a range of possible interpretations would allow for the first time three additional discourses. One is the art-historical one of witnessing, in the first half of the eighth century, a series of transfers of forms and of meanings, as new sponsors and new resources contribute to the recomposition of older forms,

umayyad palaces reconsidered

209

13 Qusayr ‘Amrah. Meditating woman in small side room

210

14 Qusayr ‘Amrah. Naked women and children in side room

early islamic art, 650–1100

on the one hand, and to the development of new meanings, on the other. The most obvious example of the latter is the ceremonial pastime, which acquires new dimensions with the luxurious baths of Khirbat al-Mafjar or the grandiose composition of Mshatta. A second discourse concerns the Umayyad phenomenon. It can explain how a new aristocracy from the outside established itself in ancient lands, but the more tantalizing question is whether the specific phenomenon of Umayyad palaces played any significant role in the development of a new art of princes eventually associated with Muslim rulers. One would also be able to return with better information and fuller perception to standard problems of early Islamic art like the formation of a princely cycle and the myths that developed around the Umayyads. At this time, I am doubtful that the Umayyads were as important and innovative in secular art as they were in the art of the mosque, but the matter will be discussed over the years to come, I hope. Finally, there is a thematic discourse. The Umayyad establishments belong somewhere in one or more schemes of royal and aristocratic living and behaving. Most of them are private palaces which constantly use public artistic triggers – like a decorated façade, for example – to audiences who are impossible to identify. But although their most brilliant features are inside, to be treasured and enjoyed and not to be shared, they lack clear spaces for private and family living. They seem different from anything else and yet

umayyad palaces reconsidered

211

can be related to French Renaissance palaces or to nineteenth-century villas in the West perhaps more easily than to their Roman ancestors. In short they have a significant contribution to make to the broader understanding of what rulers sought to do with their architecture and its decoration: to proclaim and display their glory or perhaps to create places in order to hide and enjoy their wealth.

15 Qusayr ‘Amrah. Naked women and children in side room

Part Three Fatimid Egypt and the Muslim West

Chapter XIII Imperial and Urban Art in Islam: The SubjectMatter of Fatimid Art*

The Fatimid period in Egypt has long been a puzzling one not only to the cultural historian but also to the art historian, and it may be worthwhile at the outset to define some of the apparent features of the puzzle. Three of these seem to me of primary significance. The first concerns the time of the dynasty. If we disregard its early decades in Ifriqiya, it was in 969 that the occupation of Egypt and the ceremonial foundation of Cairo symbolized the transformation of a comparatively remote heterodox dynasty with farflung “underground” connections into a major empire. The time of this transformation followed the political and social decadence of ‘Abbasid Iraq and preceded by a few decades the fall of the Umayyad dynasty of Spain as well as the slow and still very poorly known rise of the new Turkishcentered Muslim world in northeastern Iran. For a century or so, at least until the Seljuq successes of the middle of the eleventh century, only Fatimid Egypt existed as a strong and purposeful force in the whole of Islam and one can appropriately wonder whether Egypt did not become for a century the place to which artists and artisans came from all over and therefore a creative center with an impact throughout the dar al-Islam. With the great crisis which shook Egypt between 1052 and 1072 (and about which more will be said later), with the Crusades, and with the establishment of Seljuq power in Iran and Iraq, the second century of Fatimid rule was certainly not as successful politically, but artistic growth does not always coincide with political strength and quite frequently the arts, especially the industrial arts so characteristic of medieval Islam, followed and were influenced by ideas and patterns developed during periods of strength. It is furthermore remarkable that the decadence and eventual fall of the Fatimid dynasty in the second half of the twelfth century coincided with the striking and apparently sudden growth of a series of styles in ceramics, metalwork and manuscript illustration which have been called * First published in Colloque International sur l’Histoire du Caire (27 mars–5 avril 1969), Ministry of Culture of the Arab Republic of Egypt, General Egyptian Organization, pp. 173–90.

215

216 early islamic art, 650–1100

“Seljuq,”1 but which in reality affected all Muslim regions from Central Asia to the Mediterranean. Thus the question raised by the time of the Fatimid dynasty in Egypt is whether it may not have inspired or possibly even created the artistic changes which have characterized the rest of the Muslim world in the twelfth century, for during a crucial century of Islamic history it seems to have been the only major political and cultural force of the Muslim world. The second feature I would like to identify is a geographical one and concerns the space of [174] Fatimid power. There are several facets to what the term “space” means in our present context. One is purely technical. The Fatimids began their rule in North Africa, established their power in Egypt, and ruled over various other provinces such as Sicily, Syria, Palestine and Arabia, with variable degrees of success. But the center of their power was primarily Egypt and, since for the first time since the Ptolemies a major power was actually centered in Egypt, the question arises whether Fatimid art and culture did not develop characteristics which were primarily regional. Or to put it in another way, within the characteristically Islamic tension between regional traditions and universal ideals, how important was the Egyptian contribution to Fatimid art? The question is of some significance for even when under foreign rule Egypt has had for centuries an identifiable artistic personality. Since most Fatimid ventures outside of Egypt failed in spite of a considerable investment in propaganda and missionaries, should not Fatimid art be considered as the expression of a local tradition with deep roots in the past which would have been magnified to a particularly high, almost pan-Islamic importance because the dynasty itself filled a major power vacuum within the contemporary Muslim world and achieved in its time a unique level of sophistication and wealth? An answer, even a tentative one, would be of considerable interest both for the history of Egypt and as a document in the wider problem of the importance of local traditions within Islamic art. Yet there is also another facet to the question of Fatimid space. It is that most of the history of the Fatimids took place within the Mediterranean world. Only the strong Byzantium of the Macedonian and Comnenian dynasties could and did compete with the dynasty from Egypt, especially in Syria and on the seas. Otherwise it has been possible to show that, for most of the Fatimid period and for any number of reasons which need not concern us at this stage, the Fatimid centers of North Africa and of Egypt were the major magnets and the dominant forces of Mediterranean trade.2 1

2

O. Grabar, “La Révolution dans les arts mineurs de l’Islam au XIIème siècle,” Cahiers de Civilisation médiévale (1968); also O. Grabar, “The Visual Arts, 1050–1350,” Cambridge History of Iran (1968), vol. V, pp. 626–58. All these problems are being revolutionized by the research of S. D. Goitein, whose latest pertinent publication is A Mediterranean Society, vol. I (Berkeley and Los Angeles, 1967), especially pp. 29 and ff.

imperial and urban art in islam

217

There was indeed a Fatimid “miracle,” as Professor Goitein has called it, and, while one of the reasons for the miracle was certainly the weakness of the contemporary Christian Mediterranean, the more positive reasons were the remarkable ethnic and religious liberalism of the dynasty’s politics, the administrative if not always political stability of the state, and the freedom of trade. The question, then, is whether the part played by the non-Muslim Mediterranean in the economic and social life of Fatimid Egypt did not contribute another piece to the Fatimid puzzle. If it was indeed so, the importance of this period transcends the Muslim world and involves in an understanding of its art and of its culture an awareness of other artistic and cultural traditions than those of Egypt and the Muslim world. The last feature of the Fatimid period I should like to emphasize at this stage is that there is considerable information about the art of the dynasty. This information is of two kinds. First there are the monuments themselves. Architectural remains are best known, since they have been magnificently published by K. A. C. Creswell,3 and no discussion of Fatimid art as a whole can avoid being a tribute to his painstaking activity over six decades. Outside of architecture, three techniques are sufficiently well represented with monuments which, if not always dated, are clearly definable as Fatimid and have been accepted as such for many decades. One such series consists of textiles, many of which are inscribed and precisely dated.4 Another such series consists of woodwork, for which the catalogs of the Arab Museum in Cairo are the main source.5 The third series with large numbers of examples is ceramics, about which there is no comprehensive study but whose body of known pieces is constantly being increased by the publications of the staff of the Islamic Museum in Cairo,6 for it is in Cairo that we find the overwhelmingly largest collections of actual objects and of sherds. Numerous fragments of glass and of mural paintings exist as well,7 3 4

5 6

7

K. A. C. Creswell, The Muslim Architecture of Egypt, vol. I: Ikhshids and Fatimids (Oxford, 1952). E. Kühnel and L. Bellinger, Catalogue of Dated Tiraz Fabrics: Umayyad, ‘Abbasid, Fatimid (Washington, DC, 1952); also E. Kühnel, “Four Remarkable Tiraz Textiles,” Archaeologia Orientalis in Memoriam Ernst Herzfeld (New York, 1952), pp. 144–50; H. Schmidt, “Islamische Seidenstoffe der Fatimidenzeit,” Zeitschrift für bildende Kunst, 64 (1930–31), pp. 185 ff. Edmond Pauty, Les Bois sculptés jusqu’à l’époque Ayyoubide, Catalogue général du Musée Arabe du Caire (Cairo, 1931). See A. Bahgat and F. Massoul, La Céramique musulmane de l’Egypte (Cairo, 1930); G. M. Mehrei, “Al-Khazaf al-Fatimi,” University of Cairo, Bulletin of the Faculty of Arts (hereafter BFA), 7 (1944); M. Z. Hasan, “Tuhaf djadida,” ibid., 13 (1951); Hasan al-Basha and ‘Abd al-Ra‘uf ‘Ali Yusuf, “Tabaq Ghaban,” ibid., 18 (1956); ‘Abd al-Ra‘uf ‘Ali Yusuf, “Khazzafun min al-’asr al-Fatimi,” ibid., 20 (1962); there are numerous other articles on specific topics which will be mentioned in due course, but the above are of primary significance. C. J. Lamm, Mittelälterliche Gläser, 2 vols (Berlin, 1929); for latest published paintings see E. J. Grube, “Three Miniatures from Fustat in the Metropolitan Museum in New York,” Ars Orientalis, 5 (1963), pp. 89–97; but the fundamental study is that of R. Ettinghausen, “Painting in the Fatimid Period: A reconstruction,” Ars Islamica, 9 (1942).

218 early islamic art, 650–1100

in addition to a more limited number of crystals,8 ivories, and metalwork.9 One should add that works of Fatimid art have also been discovered archaeologically or else are found in situ and, aside from the still continuing Fustat excavations10 or the many monuments of Cairene architecture described by Creswell, one should mention the very important Fatimid remains from Syria and Palestine such as the mosaics of the Aqsa mosque in Jerusalem,11 or the various investigations carried on in the early Fatimid cities of North Africa.12 Altogether this information is quite remarkable by its quantity, by its quality, and by comparatively secure general dates and provenance. It should, therefore, be possible to use it to answer some of the questions we have raised, inasmuch as a whole category of monuments exists outside of the Muslim world itself for which a Fatimid background has been posited or assumed, such as the paintings of the [175] Cappella Palatina in Palermo, as well as certain Romanesque bronzes and Italian ivories.13 But in addition to this archaeological documentation, the Fatimid period is provided with literary sources of a kind which is unfortunately very rare for most other periods of Islamic art. There is a particularly precise description of Cairo by Nasir-i Khusraw,14 who lived there for seven years during the city’s greatest wealth. And then there is the wealth of information found in Maqrizi’s Khitat Misr to be supplemented now by the recent discovery of one of Maqrizi’s original sources, the Kitab al-Dhaka‘ir wa’ l-Tuhaf.15 As a result of the existence of this literary information, much of which goes back to Fatimid times, it is possible in dealing with Fatimid art to relate actual 8

9

10

11 12 13

14 15

D. S. Rice, “A Datable Islamic Rock Crystal,” Oriental Art, 2 (1956), with summaries and bibliographies; K. Erdmann, “Neue Islamische Bergkristalle,” Ars Orientalis, 3 (1959). No systematic publication of objects in these categories exists and one must refer either to very old surveys or to general works on individual techniques or on Islamic art; the simplest introduction is probably that of M. Z. Hasan, Atlas Funün al-Islamiyya (Cairo, 1956), who also wrote a suggestive introductory essay in his Kunuz al-Fatimiyyin (Cairo, 1937). In addition to the older publications summarized by Creswell, Muslim Architecture of Egypt, pp. 119–30, see the articles by G. Scanlon in the Journal of the American Research Center in Egypt, 4 and 5 (1965 and 1966). H. Stern, “Recherches sur la Mosquée al-Aqsa et ses Mosaïques,” Ars Orientalis, 5 (1963), pp. 28–48. A. Lézine, Mahdiya (Paris, 1965); S. M. Zbiss, “Mahdia et Sabra-Mansouriya,” Journal Asiatique, 144 (1956). U. Monneret de Villard, Le Pitture musulmane al soffito della Cappella Palatina in Palermo (Rome, 1950); for bronzes, see E. Meyer, “Romanische Bronzen und ihre islamischen Vorbilder,” Aus der Welt der Islamischen Kunst (Berlin, 1959); for ivories see P. B. Cott, Siculo-Arabic Ivories (Princeton, 1939). Nasir-i Khusraw, Voyages, edited and translated by C. Schefer (Paris, 1888), pp. 124 ff. Al-Rashid b. al-Zubayr, Kitab al-Dhakha‘ir wa’ l-Tuhaf, edited by M. Hamidullah (Kuwayt, 1959); M. Hamidullah, “Nouveaux documents sur les rapports de l’Europe avec l’Orient musulman,” Arabica, 7 (1960). For Maqrizi, we have used the old (1853) Bulaq edition in two volumes.

imperial and urban art in islam

219

objects with contemporary literary sources and, what is perhaps far more important, to suggest some sort of historical, cultural, or aesthetic setting for those monuments of art or of material culture which have remained.16 At this methodological level, of course, the study of Fatimid art transcends both the time and the space of the dynasty, for it may become possible to put together a sort of model of the relationship between man and his creation within Muslim culture. To be sure, the ultimate validity of this model for all regions and times is likely to vary, but the construction of the model itself may be of eventual use. There is thus a historical question of setting the art developed by Egypt in the tenth, eleventh and twelfth centuries within several possible cultural periods (early Islamic, later Islamic, Mediterranean). There is an interpretative question of deciding whether pan-Islamic, local Egyptian, North African, or Mediterranean features predominated in that art or else what sort of balance was achieved between possible sources for Fatimid art. There is finally a methodological question of how to relate to each other a considerable mass of visual and literary information in order to develop some general hypothesis of the styles, functions and artistic ideas which prevailed in Fatimid times. It is obvious that even a paper of some length cannot presume to answer all these questions and I should like to attempt to focus on one problem only and to suggest a historical and interpretative solution to it for our collective discussions. The simplest way to begin may be to discuss briefly the implications of one of the latest appreciations of Fatimid art by a leading scholar. In a recent general introduction to Islamic art, Professor Otto-Dorn has emphasized the following main characteristic of Fatimid art other than architecture: its surprising (überraschend) delight in the representation of living things (Figurenfreude) which appears in woodwork from the palaces, in painting, and in the minor arts.17 This Fatimid feature is stylistically related by the author to the art of the early ‘Abbasids as it is known primarily through paintings found in Samarra, but then the author concludes: “At the same time the parallels with Seljuq art are astonishing, a relationship which still needs investigation.” It would be difficult to quarrel with this appreciation. The most rapid survey of works of Fatimid art when compared to earlier art in Egypt or anywhere else in the Muslim world makes it very clear that a subject-matter with living beings makes its appearance in all media. But in a more general way one might say that there occurred a rather extraordinary extension of themes of decoration in all techniques and that these included a much higher proportion of personages and animals than in any of the previously known series, such as Samarra or northeastern Iranian ceramics, 16 17

Only M. Z. Hasan in his Kunuz al-Fatimiyyin has attempted to do something of this kind in dealing with a precise text to which we shall return. K. Otto-Dorn, Kunst des Islam (Baden-Baden, 1964), p. 109.

220

early islamic art, 650–1100

Tulunid woodwork, and ‘Abbasid or Spanish Umayyad stucco.18 The relationship between this new Fatimid art and ‘Abbasid art of the ninth century in Iraq is perhaps not quite as extensive as has been believed, but its existence can be demonstrated both through the undeniable impact of Iraq on late ninth- and tenth-century woodwork and ceramics in Egypt19 and through a number of details such as the celebrated sidelocks in the representation of personages, especially women. But, even if these details were not available, the impact on the rest of the Muslim world of Baghdad and Iraq in the latter part of the eighth century, during the whole ninth century and at least during the first half of the tenth, could be assumed from political and cultural history alone. The third point made by Professor Otto-Dorn, that of the apparent relationship of Fatimid themes with those associated with the so-called “Seljuq” art, is far more significant. This relationship is, however, not of the same kind as the apparent relationship between ‘Abbasid and Fatimid artistic traditions, for there are very few instances, if any, of direct stylistic or iconographic parallels between the two arts. It is rather a structural relationship in which two cultural moments chose to transform their industrial and decorative arts by introducing a large number of new subjects, including figurative and animal themes hardly used until then. In such a relationship subjects vary according to the [176] visual vocabulary particular to individual regions and it is by the fact of the new subjects that regional traditions are related to each other. In the following discussion we shall not be primarily concerned with the further exploration of these three points and to a certain extent we shall take them as axiomatic, although our comments may eventually lead to some refinement in the ways in which they might be put. Our objective will rather be to concentrate on the question of why, how, and if possible when, within the Fatimid period, the introduction of a new subject-matter took place. For on these topics unfortunately both the archaeological and literary sources are silent. Changes in taste are not usually recorded in chronicles and the vast majority of the objects which illustrate them are undated. It is a curious fact for instance that, whereas a great deal is known about the names of Fatimid potters, only two signed pieces exhibit any of the new subjects for decoration.20 Or, in the instance of woodwork, the celebrated frieze 18

19 20

A partial exception may have to be made for a group of Iraqi and Iranian ceramics dated usually to the tenth century, A. Lane, Early Islamic Pottery (New York, 1948), p. 16. But this particular group, like most Iraqi-Iranian ceramics outside of the northeast of Iran and of Samarra, still requires a more thorough investigation, with if at all possible, some archaeological confirmation of the dates given. See the very important study by R. Schnyder, “Tulunidische Lüsterfayence,” Ars Orientalis, 5 (1963). BFA, 18 (1958); M. Jenkins, “Muslim: an early Fatimid Ceramist,” Bulletin of the Metropolitan Museum of Art, May 1968. See also BFA, 20 (1962), for a rather thorough investigation of signed sherds. It is noteworthy that, while animal, vegetal, epigraphic

imperial and urban art in islam 221

found in Qala’un’s hospital is usually considered to be from the Fatimid palaces and from the middle of the eleventh century21 but there is no definite evidence for this and we are not dealing with more than a probability. On the other hand, the very remarkable changes in the composition of large wooden objects with their geometrically conceived star patterns and multiplicity of levels of decoration can be dated fairly precisely to the early twelfth century thanks to a group of inscribed mihrabs. Examples could be multiplied to show that, as a general rule, the objects in ivory, woodwork, ceramics or glass as well as the paintings which best illustrate what is supposed to be the most original novelty of Fatimid art are anonymous and undated, whereas monuments with more traditional purely ornamental designs are more often provided with specific historical information. This state of affairs could be purely accidental but it does complicate the task of the historian in understanding the reasons for the timing of a precise change. The method I should like to utilize will therefore be somewhat different from the usual art-historical one in the sense that instead of working from a small number of objects with a maximum number of known coordinates and then expanding whatever conclusions may be reached to include the mass of known objects, I shall consider primarily the mass itself and then fit individual objects within resulting schemes or models. Beginning with ceramics, with the largest number of fragments or complete pieces, an attempt at a classification of subjects and styles will be made which will then be briefly correlated with whatever evidence may be known from other techniques. Out of this survey a sort of profile will, it is hoped, emerge of the representational themes of Fatimid art. Then in the following part an explanation will be suggested for the appearances of these themes, while in our conclusion we shall return to some of the questions posed at the beginning of our paper. It is, of course, superfluous to add that what is being proposed here is not more than a hypothesis, or rather a series of hypotheses, for the study of Islamic art is still too young and too insecure to provide more than hypotheses for further research. Yet one may perhaps suggest that at times the wrong hypothesis can be more fruitful for scientific knowledge than the correct fact. One preliminary remark about Fatimid ceramics in general is of some pertinence. Although no thorough study of the techniques is available and although the exact dimensions of each individual object are not always published, it seems that a rather remarkable consistency was maintained

21

and geometric designs predominate on these fragments, human ones are rare. See below for an evaluation of this point. Often published and discussed; Pauty, Les bois sculptés, pls. XLVI and ff.; best comments by G. Marçais, reprinted in vol. I of his Mélanges d’Histoire et d’Archéologie de l’Occident musulman (Alger, 1957), pp. 81 ff.

222

early islamic art, 650–1100

throughout the Fatimid period in the quality of the technique used and in the shape and size of the individual objects. On both of these subjects, it is expected that the excavations now being carried out at Fustat will bring more definitive results. In the meantime the technique with which we are primarily concerned is a ware of rather coarse and sandy red or buff bodies, covered with a slip and a tin glaze, and then luster painted. Most lusters have a brown to yellow tinge but the color range can reach a reddish tone often found in Iraq. No significant change seems to have occurred in this technique, which had been fully developed by the time of the dynasty’s appearance in Egypt. Shapes of objects appear to have been equally consistent and changeless. A small number of fairly large jars is found, but the vast majority of Fatimid luster-painted objects are plates with rims of varying size or more or less deep bowls.22 It seems uncertain at the moment whether any evolution occurred in the proportions of these shapes, but it may be pointed out that the predominance of “open” shapes over “closed” ones can be explained by the fact that the former lent themselves much more easily than the latter to being decorated with a single design or image, for which it is therefore proper enough to propose an iconographic significance rather than a purely ornamental one. That this single design was in fact the main purpose of the ceramic is further suggested [177] by the rather rough and usually meaningless decoration of circles and hatched lines found on the back of most plates. The only important information occurring there is that of the names of potters, but in the two instances of inscriptions with precise information23 the inscription is set on the main, obverse side of the plate.24 We are 22

23 24

For a first attempt at systematizing the terminology to be used for shapes see J. Sauvaget’s posthumous article “Introduction à l’étude de la céramique musulmane,” Revue des Etudes Islamiques (1965). It is probable that eventually an Arabic terminology should be adopted for most of these shapes, and ours would be a sahn and a zubdiyya or a sultaniyya. BFA, XVI, pl. 5 in Hasan al-Basha’s article and G. Wiet, “Un céramiste de l’époque fatimide,” Journal Asiatique, 241 (1953). It seems to me in general that the whole question of the presumed names of potters on Fatimid objects needs considerable review; cf. BFA, 20 for ‘Abd al-Ra‘uf Yusuf ’s summary of the information available and Sauvaget’s article quoted above, pp. 44 ff. The matter can perhaps be best summarized in the following manner. There is little doubt that most of the inscriptions involved are in fact signatures of individual potters or of ateliers, although more than one potter by the same name may have existed. To derive from this information stylistic distinctions seems most unlikely to me, since there is little evidence in medieval Islamic art that stylistic differences were prized and there is considerable evidence to show that the appreciated artisan was one who was able to imitate many different ways imposed by the taste of the patrons. It is therefore elsewhere that we must search for the meaning of the information provided by these names. It could reflect a particularly organized system of guilds, but then why is Egypt so unique? Or it could reflect a snobbish taste which gave special importance to the maker without necessarily involving a style; such an interpretation would strengthen our point (to be discussed later) of the importance of the patron or the buyer of the object. Other interpretations can be found as well and the matter deserves further investigation.

imperial and urban art in islam

223

therefore, it seems to me, justified in considering that the main purpose of the ceramic object was the artful and successful presentation of a single design on one primary surface. In this, of course, Fatimid ceramics are not unique; they agree with most series of Islamic ceramics and are differentiated for instance from Chinese pottery with its emphasis on the purity of the body and of the shape and from direct imitations of Chinese types. Greater originality occurs in the general composition of the main design. If we except the jars whose two types of composition – decoration of the upper two-thirds of the object in concentric bands of varying width or vertical division of the object’s surface into panels25 – are quite common with this particular shape, three main types of arrangement are found on the plates and bowls. The first and by far the most common one exhibits a comparatively narrow border made up of a variety of themes and then a single subject occupying in more or less successful fashion the center of the plate. In general comparatively little effort was made to fit the design to the circular shape of the object, or rather the painter’s art was so consummate that he rarely needed the distortions of form found in some Iraqi-Iranian pottery of the time or the convention of dividing the circle into bands or by means of an exergue which are so common on earlier metalwork of the same shape or on Iranian pottery of the following centuries. Great though the merits of the decorator may have been, the essential point for our present purposes is that in this particular type of composition a single image was the subject of the decoration and compositional concerns were subordinated to it. The second type of composition is identified primarily by the great width of its border and the corresponding narrowness of the central area. It is a particularly common composition for the circular shape of plates and bowls all over the Near East and its peculiarity is that a developed design occurs both on the wide border (often inscriptions) and in the central area, thereby diluting somewhat any certainty one may have about the ornamental or iconographic meaning of a given motif. Finally the third composition is a radial one, with a narrow border and a division of the circular space from one or more centers. Although comparatively rare, it can reach at times considerable complexity26 and it is a type of composition which will become comparatively common in later Iranian ceramics, while in earlier times it is only in northeastern Iran that it was often used with stunning success. Like the preceding type of composition, the radial one is a device which takes into primary consideration the circular shape of the object and therefore here also some uncertainty may exist about the meaning to be attributed to 25 26

Compare Lane, Early Islamic Pottery, pl. 24 to Hasan, Atlas, fig. 40 or BFA, 13, pl. 9. Lane, Early Islamic Pottery, pl. 29 A; Hasan, Atlas, fig. 53; BFA, 13, figs 31 and 32; also R. Pinder-Wilson, “An Early Fatimid Lustre Bowl,” Aus der Welt der Islamischen Kunst (Berlin, 1959).

224

early islamic art, 650–1100

whatever themes of decoration can be identified. It should be noted that three types of composition common to pre-Fatimid pottery in Egypt, Iran and Iraq and not unknown in later times seem to be totally absent from Fatimid ceramics: the all-over composition found for instance in Samarra ceramics, a composition based on the creation of an exergue so common in the twelfth and thirteenth centuries, and a composition with limited subjects on a primarily empty background so typical of northeastern Iranian ninthand tenth-century series as well as of rarer Seljuq ones. The point of this comparison is that the two compositional types in which ornamental values take precedence over iconographic ones seem mostly absent from Fatimid series and we would have here a further confirmation for the hypothesis that the majority of known ceramic objects from Fatimid Egypt were not merely vehicles for their decoration but that the general composition of the decoration tended to emphasize a single subject rather than a complex design. It is therefore possible to attribute an iconographic rather than a simply ornamental meaning to the designs and, while the ambiguity between ornamental and iconographic so typical of medieval Islamic industrial arts was never completely removed, it seems less pronounced in many Fatimid ceramics. From this particular point of view, it is difficult to compare Fatimid pottery with other Islamic series without discussing the latter in detail, but it seems that, while hardly exhibiting the wealth of compositional and thematic inventiveness of twelfth- and thirteenth-century works or the simplicity and effectiveness of the earlier Iranian groups, Fatimid ceramic decoration stands out as being most consistently clear and definable. If we turn now to the themes of decoration, we may eliminate for the purposes of our discussion such features as borders (usually half-circles, but exhibiting also chevrons, braids, bands, waves, vegetal rinceaux, many of which recall traditional [178] Mediterranean ornamental borders of Hellenistic and Roman origin) or “fills” between main subjects which, when they occur, consist either of modifications of the “peacock’s eye” motif from Iran or of a miniaturized vegetal rinceau which almost merges visually with the “peacock’s eye.” Important though they may be in our eventual “phonetic” definition of Fatimid ceramics or in determining some of the influences which may have been at work in the creation of their designs, these elements are by definition secondary features of the decoration. Similarly one may conveniently eliminate a discussion of purely vegetal themes. Most of them are used in conjunction with some other subjects; they are supports for animals or backgrounds for larger scenes or else they appear to play a specific iconographic part in a number of fragments representing palm-trees.27 Only a small number of objects use vegetal motifs as a main subject.28 In most of these instances we meet 27 28

BFA, 18, fig. 50. BFA, 13, figs 27–30; Pinder-Wilson, “Early Fatimid Lustre Bowl.”

imperial and urban art in islam 225

either with a translation into painted design and dark–light contrast of the generally sculpted total ornamental surfaces developed in Samarra’s third style or else with a rather clumsy composition based on a single large leaf with sprouting tendrils. One would need lengthy analytical studies like Professor Shafi‘i’s on the calyx29 in order to classify and interpret such vegetal ornaments properly. Outside of borders, fills and vegetal themes, three main categories of potentially meaningful subjects can be defined. The first one consists of writing. As a theme it is not as common in Fatimid ceramics as in earlier northeastern Iranian series or in later Seljuq ones, but it is almost always meaningful and one encounters less commonly the pseudo-writing so typical of other Islamic series.30 Outside of the two historically significant fragments mentioned above, all inscriptions known to me consist of anonymous good wishes and no evidence exists either for the proverbs and aphorisms of northeastern Iranian ceramics or for the poems of Seljuq ceramics.31 There is probably a development in the style of writing which could be worked out in some detail by comparison with inscriptions on other media, a subject on which there exists a considerable bibliography.32 The important point for our purposes is, however, the limited use which has been made of epigraphical themes by Fatimid potters. A secondary point may be that anonymous inscriptions of good wishes are particularly characteristic of metalwork but the possible significance of this relationship will only be developed after we have discussed other motifs. The second category of subject-matter is much more common and far more important. It consists of animals. Zoologically the following animals are found: hares or rabbits in very large numbers; small birds which could be interpreted as anything from sparrows to partridges, magpies or even ducks; roosters; peacocks, eagles or other birds of prey; gazelles or other longhorned animals; fish; gryphons; elephants,33 and finally harpies. All these animals (except fish) occur at least once singly as the sole subject-matter of a given object, but they could also be shown in groups, symmetrical compositions of two similar animals or sets of three to five arranged according to some geometric or other pattern. It is interesting to note that no example is known to me of two different animals together on the same plate, except 29 30

31 32 33

F. Shafi‘i, Simple Calyx Ornament in Islamic Art (Cairo, 1957). Exception in BFA, 13, fig. 28. During the discussion of the paper at the Colloquium this point was challenged by a discussant on the basis of the many fragments found in the Cairo Museum. Many of these fragments are unpublished and the point, whose validity I cannot question, may serve to emphasize once again the need for complete publications of documents. On this point also a discussant, M. ‘Abd al-Ra‘uf Yusuf, pointed out the existence of one fragment with a proverb. The works of G. Flury, A. Grohman and G. Wiet are particularly numerous and one would wish for someone to put them all in a usable introductory form. BFA, 18, fig. 30.

226

early islamic art, 650–1100

on a jar in the Cairo Museum showing a wild animal attacking a hare.34 In this unique example an extraneous source – probably the very ancient theme of a peaceful animal attacked by a wild one known already in Achaemenid art – transformed a typical Fatimid animal and the image should be considered as exceptional. A last point to note about these animals is that the quality of their representation varies considerably and the larger the number of preserved examples (as in the case of the rabbit), the larger the variations. In some instances of unique objects there is even a grotesque quality to the animal,35 which suggests the misunderstanding of some model, but it is only with rabbits and birds that some sort of range of qualities of execution could eventually be developed. The point, as we shall see later, is of some importance in showing that a wide variety of technical abilities and tastes was involved in the making of these objects. Can one suggest any interpretation of these animals? A secure answer can only be provided by individual monographs on each animal and by appropriate textual studies on bestiaries and on the whole literary genre of the Manafi’ al-Hayawan. Yet recent studies by R. Ettinghausen and E. Baer36 may make a hypothesis possible. There are two sides to ours. First, while it is true that northeastern Iranian ceramics also exhibit from the tenth century on a rather remarkable bestiary, it is in the Fatimid ceramics that this bestiary acquires its most precise character and, except for the birds which are common elsewhere, its fullest extent. It is in Fatimid Egypt, furthermore, that Mrs Baer has seen the first development of the harpy which was to have such an elaborate later history in Islamic art. Thus we may suggest that whatever artistic or literary models may have [179] existed for these animals, it is in Fatimid Egypt that they were first transferred in such systematic manner on ceramics. And even if it may be proved some time that the northeastern Iranian group preceded the Egyptian one in time, it is likely that these were two quite independent phenomena since there are so few instances in which precise stylistic parallels can be made. The second side of our hypothesis about the animals is that, since in the instances of the harpy, of sphinxes, of rabbits, of fish and of birds, it has been possible for R. Ettinghausen and E. Baer to suggest, if not to prove, that the animal was connected with cosmic (especially solar) symbolism and, like the senmurv of Sasanian Iran, was meant to bring happiness, health, joy and prosperity, the same interpretation should perhaps be given to all these animals. They would be the visual equivalents of the inscriptions found on the objects. The problem to solve then becomes not only that of discovering

34 35 36

BFA, 13, fig. 9. BFA, 13, fig. 17. R. Ettinghausen, The Unicorn (Washington, DC, 1950) is the first model for this kind of monograph; see also his “The Wade Cup,” Ars Orientalis, 2 (1957); E. Baer, Sphinxes and Harpies (Jerusalem, 1965).

imperial and urban art in islam 227

the artistic origin of each type, a subject which is beyond the aims of this paper, but also to find out why it was at that time that this translation of a very common feeling of good wishes into images of animals was made. To this last question we shall return later. The third group of subjects found on ceramics involves representations of people. Since many of the objects with such representations are either fragmentary or unique, and since, in rather puzzling fashion, contemporary scholarship has been particularly ill at ease with the identification of human figures in early Islamic art, it is only tentatively that I would like to propose the following iconographic typology. One large group appears to be derived from what is generally agreed to be the princely cycle. The use of the cycle is peculiar in Fatimid ceramics because there is not a single object or fragment known to me which would clearly and obviously show a prince or a royal figure in some kind of formal pose. The personages illustrating the princely cycles are hunters (most often with falcons), musicians (a particularly common theme usually showing male or female rubab-players,37 or, more rarely, drummers), drinkers and pourers of wine (comparatively common and with a variety of poses and actions), and finally what has usually been interpreted as dancers, almost throughout female. These images can be interpreted in either one of two ways. One can argue that they were in some ways symbols or representations of individual personages identified here by their courtly functions and the image together with the object on which it is found should be considered as some sort of private memorialization. Alternately, and in our judgment preferably, these images are actually representations of functions, of hunting, music-making, drinking and dancing. These functions which identified princely life had acquired the more general meaning of symbolizing a “good life,” a life of pleasure, and the images served thereby a purpose similar to the purpose of animals and of inscriptions, that of wishing well to the owner of the object. While a fairly large number of complete objects and of fragments can be interpreted as part of a royal cycle, a number of others pose special problems. Thus a well-known fragment in the Cairo Museum38 showing a woman halflying on a bed, surrounded by attendants, and about to pick up a musical instrument (or giving it back after use), could be understood simply as a more developed and more precise illustration of a courtly life of pleasure. Yet it could also be the depiction of some event or the representation of a privately meaningful image. Or else is the fragment with an apparent representation of a nude dancer39 merely an erotic version of the princely dance theme or else part of some more complex subject? There is, in other 37 38 39

H. Bessier and M. Schneider, Musikgeschichte in Bildern, Vol. III Islam (Leipzig, 1966), figs 19, 28–30, etc. Lane, Early Islamic Pottery, pl. 27 B; Hasan, Kunuz, pl. 27. Hasan, Atlas, fig. 61.

228

early islamic art, 650–1100

words, a fairly large number of fragments which are only uncertainly relatable to the princely cycle but which cannot automatically be put into our second group. The latter may be called narrative, for its most immediately perceivable characteristic is that it appears to deal with a precise scene or event. Such are fragments with two personages wrestling,40 with an old man carrying a bucket on his back,41 with a bearded personage kneeling in front of a cheetah,42 with a boat being rowed alongside a strange building,43 and so forth.44 It is in this narrative group that I would like to put the well-known series of Christian images, either actual Christian scenes45 or representations of priests. Finally a celebrated and much discussed fragment with personages is provided with the names of the individuals involved, Abu Talib and Mansur, although the context of the story is unknown.46 The exact reference in the visually meaningful vocabulary of the time of all these images is, for the time being, almost impossible to find and it is possible that some of them – the wrestlers, for instance, or certain representations of personages with animals – were in fact part of a courtly cycle. On this score the art historian is in constant need of help from social and literary historians, since they alone can discover the contemporary practices [180] or literary references which could be illustrated with such images. It is equally difficult to suggest a general interpretation for these subjects similar to the interpretations we have proposed for the princely ones. It is true, of course, that the general theme of good wishes and happiness is always possible for some of them. Yet this explanation is only partially valid, for it seems difficult to associate whatever Abu Talib and Mansur may have been doing with pleasure alone. It is perhaps more appropriate to hypothesize that, for reasons as yet unknown, the object in ceramics became a vehicle for a far wider number of images than before and that these images corresponded to a much more developed system of visual symbols and of social needs than the simple one of good wishes. This, it may be noted, is exactly the conclusion which can be reached by surveying the ceramics of the twelfth and thirteenth centuries in Iran,47 even though the breadth and complexity of the subjects found in the latter is far greater. More important is the point that this development is completely unknown anywhere else in the Muslim world 40 41 42 43 44 45 46 47

Ettinghausen, Arab Painting, p. 55. Hasan, Atlas, fig. 63. J. and D. Sourdel, La Civilisation de l’Islam classique (Paris, 1968), fig. 143. Fragment in the Berlin Museum, to my knowledge unpublished, whose photograph I owe to Dr K. Brisch. Hasan, Atlas, fig. 44. Often reproduced, for instance, M. Mostafa, “Darstellungen des täglichen Lebens,” Bustan, 2 (1960), fig. 26. R. Ettinghausen, “The Mesopotamian Style in Lustre Painting,” Artibus Asiae, 18 (1958); Bahgat–Massoul, La Céramique, pl. XXXII. Cf. O. Grabar in the articles quoted in note 1.

imperial and urban art in islam 229

before Fatimid times or even at the same time as the flowering of the Egyptian dynasty. Two further remarks of slightly lesser significance may be added to our discussion of subject-matter. One is that there may also have been an astrological cycle, since some fragments show images which could be interpreted as suns.48 The other point is that most of these images with human scenes and especially the courtly ones exhibit a great variety in quality of execution, ranging from very finely drawn and composed images to strangely crude ones. There does not seem to be any correlation between type of subject-matter and quality, thus suggesting that the subjects themselves were quite widespread in their usage. We must turn now to the last aspect of our analysis of Fatimid ceramics, their style, i.e. the set of visually perceptible formulae through which individual subjects and themes were made understandable. If we consider the whole body of objects and especially the variations in quality found in them, it is clear that there is no single unified stylistic definition valid for them all regardless of subject-matter. At the same time, the lack of clearly dated objects does not allow us, it seems to me, to define a stylistic evolution, since no theoretical system has been formulated by which an evolution of ceramic designs took place and since, in a more general sense, the dynamics of stylistic change in the Muslim world have never been investigated. I would prefer therefore to propose that we consider the existence of two formal tendencies, which may have followed each other or may have been contemporary and corresponding to different modes. The first of these tendencies may be called two-dimensional, although there are many ways in which this expression is not satisfactory. It would have the following major characteristics: almost all forms and shapes are suggested by contrasts between two tones, one light, and the other dark; outlines predominate over shapes; repetitive patterns, at times unrelated to the subject-matter, are often used to cover identifiable units of design; great care is given to composition; space tends to be filled with motifs of all kinds or else an uncertain balance is created between main subject and background. This tendency appears to have had two main sources. One is the ninth-century style usually associated with Samarra; in Fatimid ceramics it tends to flatten out the curved shapes of stucco ornament49 and reaches at times considerable sophistication of formal arrangement.50 It is to this source that belong the objects datable to Hakim’s reign, but it is impossible to say how late these designs might have continued. The other source for this two-dimensional tendency was probably folk art, and it is to traditional patterns of folk art that I would like to attribute many of the simply drawn faces and elementary bodies found on a 48 49 50

Bahgat–Massoul, La Céramique, pl. 30; Sauvaget, Introduction, p. 48. Lane, Early Islamic Pottery, pl. 24. Pinder-Wilson, “Early Fatimid Lustre Bowl.”

230

early islamic art, 650–1100

large number of fragments. How much this latter feature may have also derived from textile patterns or from a popular Coptic art is still a very moot question which deserves further investigation. In any event this first tendency appears to have existed on two planes, one probably purely local and the other one relatable to the high ‘Abbasid tradition which continued in Iraqi and Iranian art during the tenth and eleventh centuries. The second tendency is far more original and it was first identified in stylistic terms by Dr Richard Ettinghausen.51 Its main characteristics are three. One is that in its depiction of the human or animal body it does not limit itself to a contour filled with a single color or with more or less arbitrary patterns, but emphasizes the volume of the body by a fairly artful use of dark and light tones. In most instances this has meant representing certain parts of the body – breasts and belly of female dancers, belly of animals, knees, thighs and arms of almost all personages – with both light and dark areas arranged no longer in clearly separated parts [181] but closely intermeshed with each other. This device is almost never used for faces and it never becomes transformed into the tonal variations of chiaroscuro. A second characteristic of this stylistic tendency lies in its interest in space and movement. On a number of objects such as the plate with stick fighters52 or one of a man leading a giraffe or on any number of dancing-girl plates, space and movement are identified by emphasizing and at times suggesting the pose of that part of the body which is primarily involved in the subject, thereby actually making the image an arrested movement from a sequence of movements rather than a pose. In a smaller number of examples this emphasis is strengthened by the disappearance of extraneous fills and the image becomes a lively or static (depending on the needs of the subject-matter) design creating its own space on a dark background by contrasting the background with a thin and precise line rather than with a heavy contour. This new interest is also remarkable for being used in details of texture, such as on fragments showing a liquid being poured into a glass.53 The last characteristic of this tendency is that it does not identify only complete objects but also occurs on details in images which would normally be put in our first tendency. The significance of this point is that this tendency cannot therefore be considered simply as a feature identifying a given school or a moment in the evolution of ceramic styles. Instead it has to be understood as having become part and parcel of the ways of representation available to Fatimid ceramicists, at least from a certain moment on. The more difficult question is that of giving a name to this tendency. Professor Ettinghausen has called it “realism” and, like Dr Mostafa, he has 51 52 53

R. Ettinghausen, “Early Realism in Islamic Art,” Studi Orientalistici in onore di Giorgio Levi della Vida, Vol. I (Rome, 1956). Hasan, Atlas, pl. 44. Ettinghausen, “Early Realism,” fig. 11.

imperial and urban art in islam

231

connected it with representations of daily life or of visually observed details which do indeed occur on many objects. Dr Ettinghausen has also related it to the celebrated painting competition between Ibn ‘Aziz and al-Qasir organized by the vizier Yazuri (before 1058). The story of the representation of dancing-girls coming in and out of walls through color contrasts actually follows in Maqrizi’s account the equally interesting but almost unnoticed description of a painting in a mosque. The latter consisted of a fountain with steps which were so colored that they appeared as a trompe l’œil when seen from one specific place while from other places the image was flat and the trompe l’œil is identified in the text as an illusion (wahm). This kind of achievement, writes Maqrizi, is “the ideal creation for painters” (fakhr alsina’i ‘ind al-muzawwaqin).54 In spite of this text with its clear statement of an illusionist ideal, I find it difficult to accept either the term “realism” or even the preferable one for stylistic purposes of “illusionism” for the images on ceramics or for the closest parallel known to me for the mosque paintings described by Maqrizi, namely the mosaics in the drum of the main dome of the Aqsa mosque in Jerusalem datable around AD 1035.55 It is always possible, of course, that the truly illusionist monuments have disappeared but it seems to me that the Aqsa mosaics which were certainly works of an imperial Fatimid art56 would have illustrated the tradition at its best. Yet, as has been amply demonstrated by H. Stern, they show instead a conscious return to Umayyad mosaic types and thus reflect only secondarily illusionist Antiquity. For these reasons I would prefer to consider this second stylistic tendency neither as a realist one nor as an illusionist one, but in contrast with the first one, as a spatial one, that is as a tendency to use a selected number of conventions in order to compel the viewer to notice details of texture, movement, action, or space. The tendency is always a selective one and never implies the conception of a total physical reality translated into image which is implicit in terms like realism or illusionism. While these considerations on terminology may be mostly theoretical, the important point about this second tendency is that it is quite unique in Islamic art, both at this time and later. Its uniqueness lies also in the artistic sources which can be given to it. For most of these have to be put outside of the Muslim world, in the Byzantine tradition. Thus it is in the manuscripts of the Macedonian Renaissance – almost contemporary with the Fatimids – that occurs the convention of creating volume by contrasting dark and light areas.57 And it has long been recognized that the lightly contoured personages set against a dark background as well as certain poses and profiles relate this

54 55 56 57

Maqrizi, II, p. 318; translation by Wiet in Syria, 13 (1932). H. Stern, “Recherches sur la Mosquée al-Aqsa’,” Ars Orientalis, 5 (1963). The importance of the Fatimids in the development of the Haram area in Jerusalem is a hitherto little-studied problem, to which I hope to return to the near future. A. Grabar, Byzantine Painting (Geneva, 1953), pp. 166–9.

232

early islamic art, 650–1100

tendency to the pre-Islamic classical art of the Mediterranean.58 How this relationship was possible has never been made clear and we will attempt to do so presently. Before proceeding to this topic, however, it may be useful to sum up our description of Fatimid ceramics and to compare them to other contemporary techniques. We are dealing with a large body of objects or fragments related to each other by shape and by a comparatively standard and consistent technique closely tied to earlier Egyptian and Iraqi techniques. Both technique [182] and shape tended to emphasize a single surface for decoration, again a feature common to much Near Eastern ceramics before and after the Fatimid period. In subject-matter and style, two modes coexisted. One, twodimensional, with primarily epigraphic, vegetal and animal themes, can easily be related to other Islamic traditions. The other one, spatial with an extraordinary variety of subjects many of which included human figures, is quite new and relatable structurally, but neither iconographically nor stylistically, to post-twelfth-century ceramic series in Iran. The two modes interacted a great deal with each other and with a possible third, folk source. Thus there are representations of figures in two-dimensional style and of animals and even of vegetal themes59 in spatial style. As far as meanings are concerned, the predominant one is that of happiness and well-being, but it is clear that other possible meanings were developed for ceramic images. In this respect, the Fatimid phenomenon seems to anticipate the later Seljuq one much more than follow the earlier Iraqi or Iranian ones. Finally, a Mediterranean non-Muslim source must be given to some of the stylistic conventions of Fatimid ceramics. At the same time, none of these characteristics can be dated; there is no internal indication of how a development took place within this large body of objects. Yet somehow, the historian has the responsibility of providing an explanation for changes in taste and form. Before suggesting such an explanation, we should turn briefly to techniques other than ceramics, for, even though ceramics are most numerous and therefore lend themselves most easily to systems of classification, it is possible that some pertinent information may be derived from the smaller numbers of works of art in other media. The closest parallel to ceramics occurs in woodwork with many fragments and complete pieces. Their importance is only tempered by the fact that a vast percentage of remaining woodwork was made for mosques and thereby limited in decorative subjects. Even with this limitation the evidence of carved wood agrees with that of ceramics. If we

58 59

Lane, Early Islamic Pottery, pp. 20–21. This point appears in objects like the one in Hasan, Kunuz, pl. 24, whose vegetal arabesque acquires a freedom of movement in space which is quite different from the total coverage of objects like the ones discussed by Pinder-Wilson (“Early Fatimid Lustre Bowl”).

imperial and urban art in islam 233

except the geometrically planned compositions of the early part of the twelfth century, two main groups of woodwork can be determined. The first one is an almost indistinguishable continuation of what had been done in Egypt since the middle of the ninth century and has usually been interpreted as a translation into wood of Samarra stucco themes. Its remaining masterpiece is the celebrated door from the Azhar mosque dated in 1010,60 where already a hitherto less common precision in the representation of individual leaves and stems is found. The second group exhibits characteristics which, just as in ceramics, can be called spatial. A greater contrast between subject-matter and background and a liveliness of individual motifs coincides with a muchexpanded vocabulary of themes. While the most important monuments of this group are the frieze from Qala’un’s hospital with its many subjects from princely and daily life and a comparable fragment from the Coptic Museum,61 the animation of wood carving with the help of a large variety of subjects and a more dynamic and natural treatment even of traditional vegetal designs are novelties of the Fatimid period. These novelties are, however, difficult to date with any precision, but it seems very likely that they do not antedate the middle of the eleventh century when al-Mustansir renovated a part of the Western palace.62 The important point is that it is in the imperial art of the caliphs that these themes are first seen, a conclusion which would be confirmed by Nasir-i Khusraw’s amazement at the sight of the hunting and feasting subjects painted on the imperial throne.63 Most of what is known of Fatimid metalwork is by inference only and the only more or less certain group consists of a series of animal-shaped vessels and sculptures in bronze. The most celebrated of these is the Pisa gryphon with its rather extraordinary surface decoration.64 None of these objects is clearly dated and therefore the evidence to be derived from them is limited. Far more important is the evidence provided by crystal, glass and ivories. Two points here seem to me of primary significance. One is that, whereas the early Fatimid period seems to have produced a number of superb imperial objects in crystal with precise epigraphical identification of the personages for whom they were made, the later Fatimid production tended to be anonymous and consisted primarily of smaller objects. Even though definite examples from Egypt itself are lacking, Spanish objects make it possible for us to reach the same conclusion about ivories.65 A group of precisely dated and personalized objects from the late tenth century and the first decade of the eleventh with the earliest and most explicit statement of the princely 60 61 62 63 64 65

Pauty, Bois sculptés, pl. 23. Ibid., pls XLVI and ff.; Ettinghausen, “Early Realism,” pp. 259 and ff. Creswell, Muslim Architecture in Egypt, Vol. 1, p. 129. Nasir-i Khusraw, Voyages, pp. 157–8. E. Kühnel, Islamische Kleinkunst (Braunschweig, 1963), pp. 164 ff. The subject is well summarized by J. Beckwith, Caskets from Cordoba (London, 1960), with bibliography.

234

early islamic art, 650–1100

cycle of images is followed by the same subjects or purely decorative ones represented on a large number of single objects or of plaques for inlays, almost all of which are anonymous. A [183] large number of the latter come from Egypt66 and almost all of these exhibit the very same spatial concerns as are found in woodwork and ceramics. Finally, even though many of their problems have not yet been solved, the paintings from the Cappella Palatina in Palermo demonstrate the very same type of interest and, whatever other influences may have been at work in them, they appear as a convenient summary both of the variety of subjects available to the Fatimid world and of the variety of ways of representation.67 As to the numerous painted remains from Egypt itself, most of them are much too fragmentary to lead to any other conclusion than the fairly obvious ones that there was an art of painting in Fatimid Egypt, that it occurred both as miniature and as wall painting, and that it involved a variety of subjects and an even greater variety of quality ranging from rather elegant personages all the way to grotesque or folk drawings.68 Brief though it had to be, this survey of arts other than ceramics leads to two conclusions. One is that the iconographic and stylistic concerns found in ceramics are not unique but typical of all the arts of the time and therefore that an explanation of the more numerous ceramics may be considered as valid for Fatimid art as a whole. The other conclusion is more hypothetical and deserves some elaboration. If we consider the two most expensive techniques which have been discussed, ivories and crystals, a curious point emerges. Both are known primarily through dated and individualized objects up to the early eleventh century and by anonymous objects thereafter, suggesting thereby some change in the social setting of the objects and of their use. Moreover, whereas all techniques exhibit novel interest in figural representation but generally alongside of other, more traditional, vegetal or animal themes, it is on royal ivories that this interest appeared first between 950 and 1050.69 Our hypothesis then would be that there occurred in the middle of the eleventh century a change in taste or in some other aspect of life likely to affect the arts which led to major modifications in the ways in which expensive materials were used and also to the spread to all media of figurative themes hitherto primarily limited to the more expensive ones. Furthermore, whatever it is that occurred must 66

67

68 69

Other illustrated panels in the Berlin Museum or in the Bargello; Kühnel, Islamische Kleinkunst, fig. 794; Sourdel, Civilisation, figs 148–9; Ettinghausen, “Early Realism,” figs 5–6. In addition to Monneret de Villard’s book mentioned in note 13, see R. Ettinghausen, Arab Painting, pp. 44 ff. and A. Grabar, “Image d’une église chrétienne,” Aus der Welt der Islamischen Kunst (Berlin, 1959). Most of the bibliography will be found in E. Grube’s article in Ars Orientalis, 5 (1963). While it is true that the objects involved are all from Spain, it can easily be shown that they belong to a generalized princely repertoire and their evidence can certainly be used for Fatimid Egypt.

imperial and urban art in islam 235

have taken place in Egypt, for no other part of the Muslim world exhibits the same changes at that time, even though a century later Iran was to undergo a similar revolution in the arts. It so happens that there is an event which took place in the middle of the eleventh century and which could be proposed as the event which triggered the artistic changes of Fatimid art. I am referring to the liquidation of the Fatimid imperial treasures either through looting, through cheap sale in order to raise cash, or through distribution in lieu of money. Although probably spread over several years, the main event took place in 1067 and has been remarkably well recorded.70 The operation was primarily a financial and economic one, but it is our contention that this dumping of a huge mass of expensive objects and works of art on to the public market was a revolutionary one in affecting the taste of large numbers of people in Fatimid Egypt. In order to justify this position we have to be able to show two things: that what is known of these treasures can indeed explain a certain number of the new features of Fatimid art and that there was a “market” for such novelties, i.e., that there was a probable “receiving” entity which could utilize, interpret and modify for its purposes whatever was suddenly made available. The first point to make is that the Fatimid treasure must be distinguished from the large number of private treasures gathered together by all sorts of individuals throughout the Muslim world. Most of the latter were primarily treasures for hoarding purposes, investments against bad days, and included only expensive materials, mostly gold and silver which could be melted down if there was a need for it.71 The Fatimid treasure belonged instead to a rarer category of imperial treasures in which were kept not only expensive objects but also rare and symbolic items of all sorts. It was divided into sections, one of which was even called the section of strange curiosities (tara’if). In each room there was a throne for the prince when he formally visited his treasure and large groups of servants were employed at keeping everything clean and in order. This official and ceremonial character of the Fatimid treasure relates it to the very similar Byzantine imperial treasures on which much information exists72 and probably to a similar ‘Abbasid treasure, 70

71

72

Maqrizi, I, pp. 408–25; Kitab al-Dhakha’ir, pp. 249 and ff.; P. Kahle, “Die Schätze der Fatimiden,” ZDMG, 89 (1935); much of the textual comparison between Maqrizi’s work and the Kitab al-Dhakha’ir was done by my student, Mrs Renata Holod-Tretiak, working at that time with Miss Lisa Volow, now Dr Golombek. The political events which led to the sale and distribution of the treasures are best recounted in G. Wiet, L’Egypte arabe, vol. IV of G. Hanotaux, Histoire de la Nation Egyptienne (Paris, 1937), pp. 239 ff., and in the article “Fatimids” by M. Canarad in Encyclopaedia of Islam, second edition. There is no study of such hoards, although the chronicles of Ibn al-Djawzi or of Ibn alAthir contain much interesting and important information for the arts and probably also for economic history. M. Canarad, “Le cérémonial Fatimite et le cérémonial Byzantin,” Byzantion, 21 (1951), p. 365.

236

early islamic art, 650–1100

although on this score our information is very scanty.73 It seems also unclear whether the Umayyads of Spain had developed a similar official treasure. The function of an imperial treasure was not merely that of a sort of dynastic museum. Its objects were made visible on certain occasions, such as the great Fatimid processions which [184] differentiate so remarkably Fatimid ceremonies from ‘Abbasid ones and relate them so closely to Byzantine ones or in official receptions of foreign ambassadors whose descriptions have been preserved for both Byzantium and the ‘Abbasids.74 The huge storerooms of clothes were certainly primarily for such official occasions, as were the numerous tents which are quite interesting for the history of secular architecture. In most cases, however, this visibility of the treasures was limited to foreign visitors only, and until the time of the looting, the treasures themselves were better known on a sort of mythic level than in reality. Another function of these treasures seems to be more peculiar to the Fatimids; at least I do not know of any evidence for this function in other imperial treasures. There were in one of the treasure’s sections toy-like sculptures representing gardens with the earth done in nielloed silver, trees of silver with fruits of amber, Nile boats with baldachins as they were used for princely outings, and pavilions with expensive decorations. I am unclear as to the use of such small objects illustrating imperial life, but it may be noted that Nasir-i Khusraw saw many of the same themes done in sugar for the caliph’s feast. In all probability, like some of the culinary achievements of our own time, the Fatimid ones and the sculptures of royal boats, gardens and pavilions served as symbol – souvenirs of the complex life of the palace. It is possibly in the same fashion that one should understand the 22,000 figurines (tamathil) in amber found in the treasure. A symbolic meaning should also be given to a huge silk hanging with maps of all the lands of the earth and representations of their kings with the names of each properly embroidered. A further point of importance about the treasures is that they were made up of things from three different sources. One was local manufacture, especially in the case of the clothes which were manufactured in a particularly elaborate and well-known way,75 but probably also in the case of many other items about which we are less informed. A second source was the rest of the Muslim world and an extraordinary traffic in historical souvenirs must have taken place if the Fatimids gathered in their treasures the rug on which Buran first appeared to Ma’mun or 100 jasper cups with Harun al-Rashid’s name on each one. Certain crystals had also been ordered 73 74

75

D. Sourdel, “Questions de Cérémonial ‘Abbaside,” Revue des Etudes Islamiques (1960). The texts describing the spectacular 917 embassy reception in Baghdad have often been discussed; G. Le Strange, “A Greek Embassy,” Journal of the Royal Asiatic Society (1897); M. Canard in A. A. Vasiliev, Byzance et les Arabes (Bruxelles, 1950), vol. II, pp. 72 ff. The best introduction to all these problems is found in R. D. Serjeant’s series of articles, “Materials for a History of Islamic Textiles,” Ars Islamica, 9 and ff. (1942–5).

imperial and urban art in islam

237

in Syria for the caliph al-‘Aziz and kept by his successors. But a third source is perhaps even more important and consisted of gifts from foreign rulers and foreign countries. This is where the Kitab al-Dhakha’ir is so important, for it provides us with long lists of what Byzantine emperors sent to ‘Abbasid and Fatimid caliphs. For instance in 937–8 the list included the following items: three golden arrows with jewels, two crystal bottles with silver incrustation and a crystal lion on top, a palm-stalk of gilded silver with engraved sparrows, an octagonal silver box with floral designs, a gilded palm-stalk with handles made in the shape of peacocks, brocades with lions and eagles, silk garments with the likeness of a Persian king holding a banner in his hand.76 In 1045–6 and 1052–3 similar groups of objects were sent and included even a saddle for the son of Mustansir which was supposed to have belonged to Alexander the Great. When one reads in the accounts of the looting that the looters found all sorts of crystals and cut glass, many enamelled gold plates, peacocks incrusted with precious stones, gold roosters and gazelles, or gold palm trees with gold leaves and imitations of dates, the possibility must be kept in mind that many of these items were of Byzantine manufacture, just as the 1204 looting of Constantinople was to bring to the West a large number of Muslim objects together with Byzantine ones. The evidence for Muslim gifts to the Byzantines is less numerous in the Kitab alDhakha’ir, probably because they were less noteworthy from a Muslim point of view, but in 1056 silver candlesticks and “apricot” colored porcelain were sent over by a minor prince. From this evidence we may conclude that ceramics from the Muslim world were particularly prized and I would suggest a local Muslim manufacture for the ceramic objects described by Maqrizi which included large urns set on tripods shaped in the form of lions and other wild animals. But what is probably more important to note is that, regardless of where a given object may have been made, the art which is illustrated through these treasures was an international or intra-cultural art common to the great empires, to the Family of Princes whose themes and ideals were all inherited from Roman and Sasanian Antiquity.77 It was an art in which a Fatimid prince would own a saddle supposed to have belonged to Alexander the Great and Byzantine craftsmen made textiles representing Persian kings. This is not to say that there was no enmity or competition between the princes. On the contrary, political competition was fierce and [185] acute, even if it took at times the form of artistic embellishments, as in the case of the Fatimid reshaping of the Haram area in Jerusalem shortly after the Byzantine rebuilding of the Holy Sepulcher. But, in spite of wars and strife, there was an imperial taste and an imperial tradition which were shared by the great empires, Byzantine and Fatimid, only after the decadence of the 76 77

Kitab al-Dhakha’ir, pp. 60 ff. For this theme see O. Grabar, “The Six Kings,” Ars Orientalis, 1 (1954).

238

early islamic art, 650–1100

‘Abbasid caliphate, in the middle of the tenth century.78 Thus what fell into the hands of looters or what was sold by the caliph did not consist only of works of contemporary Islamic art, but also of works from many different lands and from many different times. Most of them illustrated the practical or symbolic needs of an ancient imperial tradition which had been inherited by Fatimid princes and which, as the Sirat Qaysar wa Kisra shows, was already apparent in Umayyad times. To the mass of the Muslim population this tradition had only until then been known as a myth and its sudden availability and visibility were obviously major novelties. One last aspect of the looting of the Fatimid palace may be mentioned. It is that, as crowds went through the palaces, their interior decoration was made visible and known to all. Thus the whole Muslim tradition going back to Umayyad art of a private palace decoration of human figures and scenes involving people suddenly appeared to all. Little beyond the woodwork mentioned above remains from this decoration, but textual information, some of which was already given, makes it quite clear that the old Muslim princely traditions were fully preserved and the themes known from Umayyad, ‘Abbasid and Tulunid palaces can be assumed in Fatimid times as well. Even though much work is still needed to put together all the information which is thus available from these literary sources, I would like to suggest that a large number of features we have determined in ceramic and in other arts as being new in the Fatimid period can be explained though the impact of the art made for and gathered by the Fatimid caliphs. The rather curious spatial concerns with their Byzantinizing relations are easy to explain if so many of the treasures were actually Byzantine made or if much more ancient objects were available. The development of large cycles of images, with single personages and even whole scenes, can similarly be explained by the sudden impact of actual paintings and sculptures and of objects with all sorts of representations of personages and animals. Even certain details of the decoration like the circles and grooves on the Pisa gryphon or the Kassel lion could perhaps be explained as translations into cheaper bronze of the inlaid or enameled gold and silver objects found in the treasures, most of which were probably of Byzantine origin, perhaps made for the Muslim princes, as can be suggested for the celebrated Innsbruck plate.79 And a possible impact of metalwork with its inscriptions of good wishes on ceramics can also be suggested. This sudden influx of a mass of new objects and of new artistic themes did not simply lead to their translation into more common techniques and media 78

79

A. Grabar, “Le succès des arts orientaux,” Münchener Jahrbuch der bildenden Kunst, 2 (1951); O. and A. Grabar, “L’essor des Arts inspirés par les cours princières,” L’Occidente e l’Islam nell’ alto Medioevo (Spoleto, 1965). M. Van Berchem and J. Strygowski, Amida (Heidelberg, 1910).

imperial and urban art in islam

239

where they acquired a social meaning which was no longer internationally imperial but more locally popular. It also created an impetus for the creation of new images and of new ideas and it is in this fashion – as transformations into images of views and ideas of new groups of patrons – that I would like to interpret most of the new subjects of ceramic decoration. The impetus was successful because Egypt at this time was receptive to such an impetus. As the Geniza fragments show quite conclusively, Egypt had a remarkably developed middle class of Muslim, Christian and Jewish tradesmen and artisans who profited from what has been called the “free-trade community” of the Mediterranean at that time80 and who had the economic means to acquire or to order for themselves objects with new and sophisticated themes. It is this urban bourgeoisie which transformed the international art of princes into a locally Egyptian version of Islamic art. And this transformation was made possible in the middle of the eleventh century by the sudden availability of masses of works of art from an older tradition to a growing new social class. It is thus not an accident that Christian themes appear among these works, for Christians played a significant part in this social development. It is also true that Christian themes are easy to recognize and that our knowledge of the non-religious imagery other than that of princes of the medieval Near East is still far too elementary to allow for precise identifications of meanings to be given to visual forms. There lies one of the most important and most immediate tasks of scholarship in Islamic art. It would not be proper to conclude our paper without returning to some of the questions raised at the beginning. Before doing so, it is also necessary to point out that there are two areas immediately pertinent to our subject which have not been discussed but which certainly deserve [186] further work. One, perhaps the most important one, is the area of textiles. The point there is that texts give rather elaborate descriptions of textiles, most of which do not correspond to the fragments which have actually remained, however numerous the latter are. How is one to explain this anomaly? An answer to this question must eventually be provided, for the Fatimid period is uniquely rich in literary information. Furthermore, documents exist which suggest that it was through textiles that much of the contemporary world acquired its aesthetic judgment, as when Nasir-i Khusraw compares luster ceramics to buqallamun.81 The other subject, which has already occupied the attention of knowledgeable Arabists like P. Kahle, M. Mostafa and M. Hamidullah, is that of the exact meanings to be given to the terms found in literary texts.82 Yet there is still much to be done, especially in the study of verbs and nouns, for there is some intellectual presumption in claiming any 80 81 82

Goitein, A Mediterranean Society, p. 61. Nasir-i Khusraw, Voyages, p. 151. In addition to works quoted above, P. Kahle, “Bergkristall, Glas and Glasflüsse,” ZDMG, 90 (1936).

240

early islamic art, 650–1100

sort of historical competence in the arts when one does not fully understand contemporary references. These questions should be left for other occasions. Our aim at this stage is to try to draw some conclusions – or at least hypotheses – from our observations on Fatimid ceramics and the representational themes found on them. There are three such hypotheses I would like to propose. The first one is that the transformation of a “restricted” artistic tradition into a source of inspiration and an impetus for a more popular art is a characteristically Romanesque phenomenon. Since the correlation in dates is quite striking, I would suggest that it is not accidental that, at the time when the Fatimids utilized and modified the international art of empires, local entities all over Italy and the West translated into a sculpture visible and accessible to all the private, monastic or royal, art of manuscripts, ivories and metalwork of older generations. It is probably not an accident as well that antique themes reappeared in the West as they did in Egypt and the rather remarkable transformation in the second half of the eleventh century of Italian episcopal thrones into princely ones with many themes closely related to those of Fatimid art83 cannot be a coincidence. Yet, even though the Fatimid phenomenon can justifiably be called a Romanesque one, its very Romanesque nature of a sort of democratization of the visual arts world led to its separation from other Romanesque traditions, for control over taste and patronage was no longer in the hands of a small group of princes partly withdrawn from the area in which they lived but in those of a locally centered urban order. Thus, to recall a question raised at the beginning, Fatimid art belonged indeed to a Mediterranean, partly non-Muslim tradition and was influenced by it, but it is in Fatimid times that the first original art of Muslim Egypt acquired, so to speak, its aesthetic and thematic independence from the rest of the Mediterranean, as well one might add, as from the ‘Abbasid traditions of Iraq. It is far more difficult to assess the impact of Fatimid art on later Islamic art, outside of Egypt itself where it obviously maintained a continuing influence.84 On this score our hypothesis, the second one, would be that on the technical level of artistic forms the Fatimid impact was minimal. Even though it can be argued that the transformation of imperial objects into automatic toys for minor princelings, as appear in al-Djazari’s Automata,85 may have been influenced by the looting of the Fatimid palace (actually 83 84

85

A. Grabar, “Trônes épiscopaux du XIème et XIIème siècles en Italie Méridionale,” Wallraf-Richartz-Jahrbuch, 16 (1954). The point has been made for architecture by K. A. C. Creswell on a number of occasions (as in the cases of the mosque of Baybars or of the muqarnas). It could also be made for many other arts. O. Grabar in L’Occidente e l’Islam, pp. 870–71. For the manuscripts themselves, which have never been properly studied, see Ettinghausen, Painting in the Fatimid Period, pp. 194–5 with a good bibliography.

imperial and urban art in islam

241

other sources existed as well), on the whole Fatimid art remained a peculiarly Egyptian phenomenon which drew on and synthesized sources available to the specific land of Egypt. The Seljuq arts of the twelfth century with which Fatimid art shares so many structural features do not seem to have been influenced by what happened a century earlier in Egypt. The reason for this may be provided by our third hypothesis. It is that the arts – and perhaps the whole culture – of the classical Islamic world86 grew and developed through a complex interaction between princely and urban tastes and needs. While the princely ones tended to be interchangeable from one region to another, the urban ones tended to be more localized and to reflect regional characteristics. If our interpretations are acceptable, what created the originality of Fatimid art was the predominance taken by an urban Egyptian, Cairene taste in the second half of the eleventh century. Different developments and different regional traditions would have led to the impact of an urban taste in Iran and, for reasons as yet unknown, the phenomenon took place at a different time. Such seem to me to be some of the conclusions and hypotheses which may be derived from a consideration of some of the monuments of Fatimid art. Whether they are justified or not is for others to judge. It seems to me, however, that by compelling us at the same time to raise questions about the dynamics of all Islamic arts [187] and to define the uniqueness of medieval Egypt, the art of the Fatimids more than fulfills the expressions of those who founded the city of Cairo a thousand years ago.

86

I am using the term in the sense suggested by D. and J. Sourdel as meaning the centuries before the Mongol onslaught and the growth of the Ottoman empire.

Chapter XIV Fatimid Art, Precursor or Culmination*

There were several occasions over the past decade for scholars with different specialities to interpret rather than describe the historical and cultural phenomenon of the Fatimids. At the time of the great colloquium celebrating Cairo’s first millennium, the late Gustav von Grünebaum, Professor Bernard Lewis, and I in a more limited sense, sought to identify the reasons for Fatimid successes, failures, or simply cultural or historical peculiarities.1 The late Marshall Hodgson’s controversial synthesis of Islamic history also provides novel interpretations of the Fatimids within what he called the Middle Islamic Period.2 And for several decades now Professor Goitein’s studies on the Geniza fragments have led him to a particularly wide range of conclusions and ideas about the Fatimid period.3 One of the reasons for this unusual interpretative concern is that few periods in Islamic history are as well and accessibly documented as the Fatimid; preserved chronicles are for the most part published, even if not to everyone’s satisfaction; the monuments of Cairo and of Egypt are known; works of art and artisanal creations can be found in books and articles, even if at times in arcane reviews; and a considerable amount of ancillary documents – coins, official letters, sales contracts, dowries, legal texts – is available, if not always incorporated into general history books. But a more important reason may be that there is something unusual, puzzling about the Fatimids, as though they were an anomaly in the development of Islamic culture. At a time of localized, ethnocentric dynastic movements from Spain to Transoxiana, it was a pan-Islamic, Arab-centered but universal caliphate. It claimed universal religious and political power, but its most obvious legacy was the transformation of Egypt, at the time hardly

* First published in Isma‘ili Contributions to Islamic Culture, S. H. Nasr, ed. (Tehran, 1977), pp. 209–24. 1 G. von Grünebaum, “The Nature of the Fatimid Achievement”; B. Lewis, “An Interpretation of Fatimid History”; O. Grabar, “Imperial and Urban Art in Islam: the Subject Matter of Fatimid Art”; all in Colloque International sur l’Histoire du Caire (Cairo, 1972). 2 M. Hodgson, The Venture of Islam, vol. II (Chicago, 1974), esp. pp. 21 ff. 3 Most particularly S. D. Goitein, A Mediterranean Society, I (Berkeley, 1967).

243

244

early islamic art, 650–1100

Islamicized and economically exploited by others, [210] into a major center of Islamic culture. Instead of promoting a minor town like Ghazni, Tinmal or Rabat into the temporary capital of an empire, it coalesced Cairo into one of the truly great metropolises of the Muslim world for more than a millennium and gave it an Egyptian rather than an Islamic character. It lived and prospered off an international, Indian and Mediterranean trade, but its political ambitions lay in Baghdad. Its literary and intellectual creativity was not comparable to contemporary activities in Spain, Iraq or Khurasan, but its art was original and different from much of what is known elsewhere at the same time. And, as a last example of Fatimid peculiarities, the intensity of the dynasty’s religious and missionary zeal is at its most effective through the rule of al-Hakim (died in 1021) and decreases thereafter, as is amply demonstrated by al-Basasiri’s failure in Baghdad in 1057–9; but, weakened and at times battered, the Fatimid dynasty stays in power for another century and more. It is these paradoxes which have puzzled historians, and my purpose in this essay is to identify and explain the ways in which a study of the arts can contribute to their solution. I shall not try to be complete and thorough, nor shall I discuss all possible ways of examining and interpreting the available material. My purpose instead is to develop general observations about the monuments and the texts dealing with them, especially Maqrizi’s invaluable Khitat4 in order to assess the originality of the Fatimids within Islamic art and culture. The best-known monuments of Fatimid architecture are all in Egypt: large mosques like the Azhar and Hakim mosques, small ones like al-Aqmar, or now-vanished ones like the Qarafa mosque which can be reconstructed with some degree of certainty; mausoleums in Cairo, Qus and Aswan; palaces, either the great imperial ones within the walls of Cairo itself or pavilions all over the city, mostly known from literary sources and descriptions; fortifications, and ceremonial or functional gates.5 Several broad considerations are involved in understanding these buildings. One is the balance between indigenous Egyptian elements (some of which had been brought from Iraq during the ninth century) and imported ones. The latter are generally assumed to be North African for the early period of Fatimid history6 and northern Mesopotamian after the middle of the eleventh century.7 A second issue is the degree of [211] originality of the forms and 4

5 6 7

Maqrizi, Kitab al-mawa‘iz wa’l-i‘tibar bi-dhikr al-khitat wa’l-athar (Bulaq, 1853), 2 vols. Few Arabic texts need a new edition as much as this one. Additional important literary sources for the arts are Maqrizi, Itti‘az al-hunafa’, ed. Jamal al-Din al-Shayyal (Cairo, 1967), and al-Qad al-Rashid ibn al Zubayr, Kitab al-dhakha‘ir wa al-tuhaf, ed. M. Hamidullah (Kuwait, 1959). Practically all these monuments are in K. A. C. Creswell, Muslim Architecture in Egypt, vol. I (Oxford, 1952). To the references found in Creswell, one should add A. Lezine, Mahdiyah (Paris, 1965). These are particularly evident in military architecture and in the growth of stone construction.

fatimid art, precursor or culmination 245

functions of Fatimid architecture. This is far more complicated to establish than the drawing up of a list of influences, and the whole question requires some elaboration. Only one architectural function is truly new in Fatimid times, even though its roots are older than the dynasty. It is the mausoleum, whose development certainly owed something to Shi’ism but whose real sources are both secular and religious.8 Mosques, palaces and even city walls existed before and, at least at first glance, the Fatimids continued to use the hypostyle mosque developed several centuries earlier. No evidence exists that their palaces were significantly different in shape from what can be assumed to have been the imperial urban palace type found in Baghdad, Samarra, or even Constantinople. Walls and city gates are fairly rare before the tenth century and Fatimid originality is more difficult to establish, but then city walls have a sameness of purpose and form which limits their value for architectural history. Yet these rapid impressions of continuity in function are misleading, for changes were introduced which were quite far-reaching. Small mosques – alAqmar, Salih Tala’i, and several others listed by Maqrizi9 – were not simply miniaturized versions of large hypostyle ones but expressions of a shift in the patronage of public pious buildings from the central authority of the state to private sources. This became an irreversible change and, for several centuries thereafter, nearly all Muslim cities, and most particularly from Iraq to Morocco, acquired small private mosques and other pietistic monuments adapted to the physical reality of cities rather than imposed on them, and reflecting a plurality of pious ways rather than a single conception of the proper setting for the Muslim community.10 Functional changes in large congregational mosques are less immediately apparent, but Maqrizi shows that these monuments for the whole Muslim community became settings for major royal ceremonies with formal entries, official changes of vestments, incense burning, disappearance of caliphs behind curtains, chanting and processions.11 Whether these ceremonies survived the crisis of the middle of the eleventh century is not certain, but the concern for formality and ceremonial processions in most aspects of court life certainly went on and its unusual character within the Muslim tradition has been frequently discussed.12 If we turn to architectural forms, only one entirely new form [212] appears in Fatimid times, the muqarnas. There is as yet no truly satisfactory 8

9 10 11 12

O. Grabar, “Earliest commemorative monuments,” Ars Orientalis, 4 (1966); some of my conclusions have been challenged by Y. Raghib, “Les premiers monuments funéraires de l’Islam,” Annales Islamologiques, 9 (1970), and “Sur un groupe de mausolées du cimetière du Caire,” Revue des Etudes Islamiques, 40 (1972). For the most part, I will stand by my earlier interpretation. See especially Khitab, II, pp. 282 ff., 445 ff. O. Grabar, “The Architecture of the City,” in I. M. Lapidus, ed., Middle Eastern Cities, Berkeley, 1969. Khitab, II, pp. 227 ff. M. Canard, “Le Cérémonial Fatimite et le Cérémonial Byzantin,” Byzantion, 21 (1951).

246

early islamic art, 650–1100

explanation for its origins and there is some debate as to whether it is an eastern Iranian form imported into Egypt, a local creation, or, as is much less likely, a North African invention.13 The importance of the muqarnas lies in that, like the mausoleum and other small religious buildings, it becomes common in Fatimid times and remains a consistent feature of Islamic architecture for centuries to come. Another formal development with a similar although less well-defined history is the formal façade with a central, usually heavily decorated gate and with corner minarets. Decorated gates already existed in ninth-century Western Islamic architecture, but it is in Tunisian Fatimid mosques that they became the main element of façade compositions. Roughly at the same time the Buyids of Iran appear to have initiated the same type of composition, if the Jurjir façade in Isfahan has been correctly dated and identified.14 The later development of façades with or without minarets needs no demonstration. Two less well-known instances of Fatimid activities can refine and sharpen an interpretation of their architectural concerns. The first one deals with the Haram al-Sharif in Jerusalem. The archaeological and cultural history of this extraordinary sanctuary has not yet been written, in spite of the huge available documentation.15 From inscriptions, texts, and from Nasir-i Khusraw’s superb description of the monuments, it is clear that, next to the Umayyads, the Fatimids were responsible for the most consistent building up of the Haram. It is true, of course, that the shortening of the city walls during the time of al-Zahir and devastating earthquakes early in the eleventh century necessitated major reconstruction. But what is striking is their quality, as in the instances of the celebrated mosaics of the Aqsa mosque16 and of the western gate described by the Persian traveler. The mosaic technique of both monuments and the themes of decoration in the mosque are also striking for their traditional and archaizing character, even though the pendentives of the Aqsa mosque are architecturally quite original. Nothing remains of the western gate, but, as it became the main entrance into the sanctuary, it stands to reason that it was a particularly majestic expression of Fatimid power. If we recall Hakim’s destruction of the Holy Sepulcher and its subsequent rebuilding, for which the Byzantine emperor sent precious objects duly recorded by Fatimid chanceries,17 it can [213] be concluded that, 13 14 15

16 17

Creswell, Muslim Architecture, pp. 251 ff.; O. Grabar in R. N. Frye, ed., The Cambridge History of Iran, vol. IV (Cambridge, 1975), pp. 343 ff. A. Godard, “The Jurjir mosque,” A. U. Pope and Ph. Ackerman, eds, A Survey of Persian Art, vol. XIV (Oxford, 1967), pp. 3100–103. See S. D. Goitein and O. Grabar, al-Kuds, E12 (forthcoming). In the meantime, the only complete study of Islamic Jerusalem is Max van Berchem, Matériaux pour un Corpus Inscriptionum Arabicorum (Cairo, 1920–22). Out of date in a few details, it is still one of the true masterpieces of Islamic archaeology. Nasir-i Khusraw, Sefer-nameh, ed. and tr. Ch. Schefer (Paris, 1881), repr. 1970, pp. 66–99. H. Stern, “Recherches sur la Mosquée al-Aqsa,” Ars Orientalis, 5 (1963). Ibn al-Zubayr, Kitab, pp. 72–8.

fatimid art, precursor or culmination

247

with a different script, the Fatimids replayed in Jerusalem the confrontation between Christian Byzantium and Islam which characterized Umayyad times and that, like their predecessors, they expressed it in monumental form. The second example is less well documented. It involves Fatimid building activities in Mecca and Medina. As is obvious from the chronicles, Fatimid control over the two holiest sanctuaries of Islam was complicated by ‘Abbasid activities from Iraq, by local disturbances, and by Carmathian incursions. Pilgrimages often had to be given up and were not usually very safe. But the formal assertion of Fatimid presence was emphasized by a large inscription on the Ka‘ba,18 and it is interesting to note that the ‘Abbasid inscription of al-Muktafi which replaced the Fatimid one in 1155 contains a very Shi’ite reference to the ‘Abbasid caliph’s “pure ancestors” and descendants.19 All these examples lead to the following model. Fatimid architecture is basically conservative and traditional. Most of it consists of established compositional and constructional elements and the dynasty’s concern for Mecca, Medina, or Jerusalem is a way to express their legitimacy, perhaps in the case of Jerusalem because it was more easily controlled than the Arabian holy places and had been a major centre of Umayyad architecture.20 Whatever the reasons, acceptance by the Muslim community could only be expected through the maintenance and occasional overhaul of old and known forms. Changes were of two kinds. One was the ceremonial use of mosques and other sanctuaries21 by the caliphs, a practice for which some evidence exists in Umayyad, ‘Abbasid and Tulunid times, but never in the systematic manner of the Fatimids. This innovation did not last, but it is interesting to note that the ceremonial presence of the ruler in the mosque will reappear with the Ottomans, the last great Mediterranean Islamic dynasty. There were also formal changes. The mausoleum, the muqarnas, composed façades on religious buildings: all seem like minor innovations and none were invented under the Fatimids. Yet all were to become nearly universal fixtures of Islamic architecture and the question can be raised whether their adoption by the Fatimids did not compel their acceptance by the culture as a whole. Because of its automatic social connections, architecture lends itself more easily than other arts to the formulation of [214] hypotheses about the historical and cultural meaning of a dynasty or of a time. With these hypotheses in mind, we can more easily turn to the other arts. 18 19 20 21

Nasir-i Khusraw, Sefer-nameh, pp. 202–3. Ibn Jubayr, Voyages, tr. Gaudefroy-Demombynes, Paris (1949–53), p. 109. It is of course true that a renewed pious involvement in Jerusalem began under the Ikhshidids, who were buried there. An interesting instance is that of the monumental activities around the finding of the head of Husayn in Asculon; its main remaining monument is the minbar now in Hebron. See M. van Berchem, “La chaire de la mosquée d’Hébron,” Festschrift E. Sachau (Berlin, 1915); and G. Wiet, “Notes d’épigraphie syro-musulmane,” Syria, 5 (1924).

248

early islamic art, 650–1100

Much has been written about Fatimid painting, since Ettinghausen’s seminal article of more than thirty-five years ago22 and many recently discovered fragments have been, rightly or wrongly, attributed to the Fatimid period.23 Their meaning is rarely very clear and their quality hardly consistent. The task of sorting out sketches from actual paintings, doodles from illustrations, folk drawings from courtly ones, early from late examples is, for the time being, well-nigh impossible. If one adds, as well one should, to the art of painting the many examples of luster ceramics with very varied subjects,24 woodwork like the panels supposed to have come from the Fatimid palace,25 or ivories, metalwork and sculpture26 with occasional representations of animals and personages, matters become even more complicated, as stylistic or iconographic judgments are affected by different techniques and different purposes. Furthermore, there is hardly a single dated painting or object. In fact only the paintings from the Cappella Palatina in Palermo are datable to c. 1140, a time when Fatimid culture was but the shadow of its former self.27 An art of painting can be assumed for earlier times, as several texts mention mural paintings in private houses, palaces and mosques from the middle of the eleventh century, and the huge libraries of the caliphs almost certainly contained illustrated manuscripts, probably of far greater quality than whatever remains.28 None of these library references can be directly related to otherwise known paintings. Thus, except for the still incompletely interpreted Palermo paintings executed in a very special historical and cultural context, we lack for this huge body of documents some obvious interpretative key, some motif or idea from which one could imagine, at least in theory and hypothetically, that the rest can be understood. Granting, then, a certain amount of methodological despair in dealing with this mass of uneven and unenlightening information, a number of observations can be made which are related to the hypotheses formulated

22 23 24

25 26

27

28

R. Ettinghausen, “Painting in the Fatimid Period. A Reconstruction,” Ars Islamica, 9 (1942). Most up-to-date bibliography by E. J. Grube in B. W. Robinson and others, Islamic Painting and the Arts of the Book (London, 1976). Latest study with bibliography by E. J. Grube, Islamic Pottery … in the Keir Collection (London, 1976), pp. 126 ff.; for an interpretation see Grabar in Colloque. See also D. Jones, “Notes on a tattooed Musician,” AARP, 7 (1975). E. Pauty, Les Bois sculptés (Cairo, 1931); G. Marçais, Mélanges d’Histoire et d’Archéologie (Alger, 1957), vol. 1. See bibliography in Grabar, Colloque. For ivories E. Kühnel, Die Islamischen Elfenbeinskulpturen (Berlin, 1971). For other techniques, it is still necessary to consult museum and exhibition catalogs. U. Monneret de Villard, La Pitture Musulmane al soffitto della Cappella Palatina in Palermo (Rome, 1950). These paintings are about to be reinterpreted by Ms Annabelle Cahn in a doctoral thesis at Columbia University. The main texts are Maqrizi, Kitab, II, pp. 318, 486–7; for libraries, see G. Wiet, “Recherches sur les bibliothèques égyptiennes,” Cahiers de Civilisation Médiévale, 6 (1963).

fatimid art, precursor or culmination

249

around architecture. First of all the number of representations in all media attributed to the Fatimids is unusually high. The climate of Egypt and the relative security of a land free of destructive invasions for many centuries contributed, of course, to a better preservation of its [215] possessions than anywhere else in the Muslim world. In spite of this, it is reasonable to conclude that the Fatimid wealth of motifs differs from that of other provinces of the Muslim world during the latter part of the tenth century and during the eleventh. Neither the few Iraqi or eastern Iranian ceramics with representations, nor the rather pitiful fragments of paintings from Nishapur or Lashkari Bazar, nor the stuccoes from Afrasiyab and Tirmidh, exhibit a comparable variety of subjects. And it is not possible to argue insufficient evidence, because the Egyptian phenomenon clearly demonstrates that, at a certain level of saturation, representations spread to almost all levels of artistic activity and would have left more traces than are actually found. In the twelfth century, on the other hand, all these regions except the Muslim West experience an explosion of representations of all types and in all media.29 From numerical evidence alone it is not possible to argue an impact of the Fatimids on the later arts of Iran or Anatolia, but it is curious that, once again, a Fatimid Egyptian novelty remains a characteristic of several subsequent centuries. A second characteristic of the Egyptian remains is that their preservation is indiscriminate in the sense that the creativity of all classes of society has been preserved according to almost archaeological randomness. In fact, because of frequent lootings of caliphal establishments and the absence of coordinated excavations,30 the remains are probably more representative of the full range of Fatimid society than the finds of Samarra in Iraq or Madinat al-Zahra in Spain. Unfortunately, in spite of Professor Goitein’s pioneering studies, we cannot as yet provide a reasonable description of that society. We can, however, assume first of all that Fatimid society was pyramidal, with many more poor and simple people than rich and educated ones, and second that some correlation exists between an object’s technique and the social level which owned it or used it. But even this second assumption is debatable, as the ownership of an expensive item was then and is still now frequently the prerogative of the poor. The point is rather that the proper exploitation of the accidental preservation of large numbers of objects found in Fatimid art requires the identification and assessment of two variables: the social cost of a technique and the imaginative taste of a social group.31 29 30

31

O. Grabar, “Les Arts Mineurs,” Cahiers de Civilisation Médiévale, 11 (1968); R. Ettinghausen, “The Flowering of Seljuq Art,” Metropolitan Museum Journal, 3 (1971). The recent Fustat excavations have only been partly published and bear more on earlier times. See the various reports by G. Scanlon in the Journal of the American Research Center in Egypt. By social cost of a technique, I mean whatever an individual was willing to invest in order to own an object; beyond purely financial considerations, he certainly did consider

250

early islamic art, 650–1100

Since neither variable has been investigated, we can only formulate the hypothesis that under Fatimid rule wider [216] segments of society were able to express something of their taste, visual needs, or aesthetic requirements than ever before or than in any other part of the Muslim world. It is therefore all the more unfortunate that so few of the remaining monuments are dated or datable, for we cannot demonstrate whether the broadening of the social base of forms was a slow process, willed choice, or, as I have suggested in an earlier article, the more or less accidental result of the lootings in the middle of the eleventh century.32 A third characteristic of Fatimid representations, their style, has been discussed more frequently. Two aspects of that style have attracted particular attention. One is its sources. It has amply been demonstrated that Iraqi ‘Abbasid elements, local Egyptian folk ways, Hellenistic and ancient Egyptian memories, and important Byzantine artistic devices are all present. They are all easy to explain, as ninth-century Iraq was indeed the hub of the Muslim world and its styles had already penetrated into Egypt a century before the Fatimids, as Fatimid rule certainly fostered native self-awareness, as the growth of Egypt led to accidental or wilful rediscoveries of the country’s past monuments, and as Macedonian Byzantium was the most powerful rival as well as model and trading partner for the caliphs in Cairo. Most curious is the absence of a western Islamic or western Mediterranean influence in representational art, inasmuch as it is so prominent in architecture. This absence is perhaps more apparent than real, for in two limited instances a relationship between the Fatimids and the Muslim West is not excluded. One deals with sculpted ivories, where presumably Fatimid plaques like the ones in the Berlin museum may well have acquired their three-dimensionality and their misleading monumentality from the sculpted ivories of tenthcentury Cordoba.33 The other example is that of metalwork, where Fatimid works are almost impossible to distinguish from western Islamic ones.34 Whether this difficulty should be explained quite simply by a commonality of forms between two Mediterranean centers or through the impact of one on the other is still a moot question. Altogether, the point about the sources of Fatimid representational art is their variety. It may simply reflect the formation of a new and wealthy center of patronage, to which models and artisans come, on their own or invited, in expectation of commissions and successes. It may also mean that Fatimid culture never developed a single

32 33 34

such features as status, occasion, resale, and so forth. As to the imaginative taste of a social group, it consists of the shared memories which can be translated into visual terms and of the formal expectation of the translation. Grabar in Colloque. Kühnel, pl. XCVII. There is no easily accessible corpus of metal objects pertinent to our discussion. See Exhibition, Islamic Art (Cairo, 1969); and E. Dodd, “On the Origins of Medieval Dinanderie,” Art Bulletin, 51 (1969).

fatimid art, precursor or culmination 251

[217] dominant style of its own but reflected, passively and accidentally or wilfully and consciously, traditions developed elsewhere. The same difficulty of definition appears in the other frequently discussed aspect of Fatimid style, the character of its forms. It has been argued that two manners of representation coexisted. One, two-dimensional and ornamental, emphasizes patterns and is removed from physical reality. The other one is more unusual; whether to be called realist, illusionist, or spatial, it utilizes techniques of representation, largely of classical origin, which sought to suggest space and volume.35 It is tempting, as has been done more than once, to relate this second mode to several preserved texts which describe the naturalism of the works depicted by Fatimid painters. Whether these texts reflect actual reality or literary clichés, it seems reasonable to assume that the illusionism of some of the paintings corresponded to a genuine impulse for more natural forms than had existed before. To define the historical importance of these stylistic characteristics is more difficult than to identify or to describe them. If it is correct to date them to the second rather than to the first Fatimid century, then it may be legitimate to see them as forerunners of the realism which will appear in the latter part of the twelfth century and in the thirteenth, especially in the Arab art of the Fertile Crescent.36 But if they are indeed forerunners, they are not so much formal ones as structural ones. They identify a related attitude toward representation rather than a similar manner of representation. It is under the Fatimids, it seems, that one discovers for the first time within the medieval Islamic tradition, a concern for the physical reality of the surrounding world. This judgment must remain a qualified one, as it is just possible that what appears today as a kind of realism was nothing more than the impact on Fatimid Egypt of Hellenistic forms with illusionist characteristics. There is yet another sense in which Fatimid styles can be considered as forerunners. Like the changes which occurred in the twelfth-century art of Iran but one century earlier, Fatimid representations combined an unusual variety of artistic sources with a considerable range of quality and effectiveness. Although it has been argued that Fatimid artisans moved to Iran after the fall of the dynasty,37 their impact, if it existed at all, was probably limited to certain techniques like luster and is [218] difficult to argue for forms. Once again the Fatimid phenomenon is structurally rather than formally related to the later Iranian one. If we turn to the subject-matter of Fatimid art, somewhat similar conclusions emerge. Its purely royal level is hardly known except through texts. From Nasir-i Khusraw’s description of the paintings in the main 35 36 37

R. Ettinghausen, “Early Realism in Islamic Art,” Studi Orientalistici in onore di Giorgio Levi della Vida, I (Rome, 1956); Grabar, Colloque, p. 181. R. Ettinghausen, Arab Painting (Geneva, 1962). Grube, Pottery, p. 210.

252

early islamic art, 650–1100

throne-room of the palace or from the lengthy lists of objects found in their treasures,38 the impression is that of masses of expensive “things” and curios like whole gardens of gold and silver, but not one of great ideological originality. The art of Muslim princes tended to emphasize the same subjects, and to my knowledge no evidence exists to show that the Fatimids tried to introduce into their art dynastic, religious, or personal symbols.39 Few Muslim princes did so. Matters are quite different when we turn to other social levels. Since the variety of themes found in Fatimid paintings, woodwork, ivories, or ceramics is in the process of being investigated by several scholars, I shall limit myself to a few general and probably only temporarily valid remarks. Examples exist of illustrations of literary stories, common events, erotic topics, astrological symbols, princely activities, animals, possibly even Christian religious symbols. The range is comparable to the Iranian phenomenon of the following century, but, just as for styles, it is difficult to propose a direct filiation from Egypt to Iran. One explanation possibly valid for both areas is essentially a social one. The growth of a strong bourgeoisie, so clear from Professor Goitein’s research, would have created a more varied patronage than existed before and this patronage sought to express many more social, personal, ethnic, literary, sectarian, or other needs than the fairly consistent patronage of princes. Another explanation is that Egypt became, before Iran, a crossroads of international contacts and that the example of cultures with a rich imagery like India or the Christian world led to the development of Islamic representations, at a time when the historical justifications for a strict aniconism were no longer valid. Before attempting to conclude this essay on the art of the Fatimids, two additional points need to be made. One deals with artisanal techniques. Without taking sides in the problem of Egyptian and Iraqi priority in developing luster-painted ceramics, it is the Fatimids who first transformed this uniquely [219] Islamic luxury technique into a vehicle for an unusual variety of styles and subjects. A more complicated problem is posed with rock crystal and its apparent derivative, the so-called Hedwig glass.40 Because of a small group of magnificent crystal objects with an early Fatimid date, scholarship has tended quite naturally to assign most remaining examples to the patronage of the Fatimid court or to direct imitations of its taste. This may indeed be correct, but literary sources clearly indicate that rock-crystal objects were made for many courts and it is perhaps more appropriate to consider the technique as a generally courtly one with accidentally preserved 38 39 40

Nasir-i Khusraw, Sefer-nameh, p. 158; Maqrizi, Kitab, I, pp. 408–25; P. Kahle, “Die Schätze der Fatimiden,” ZDMG, 89 (1935). A possible exception occurs in the decoration of the Hakim mosque, Creswell, Muslim Architecture, p. 104, but the evidence is not very clear. D. S. Rice, “A datable Islamic Rock Crystal,” Oriental Art, 2 (1956); B. Gray, “Thoughts on … ‘Hedwig’ Glass,” Colloque.

fatimid art, precursor or culmination 253

Egyptian examples. A very similar difficulty in identifying technical originality occurs in textiles. Textual information is plentiful and long lists exist of the elaborately decorated silks used at the Fatimid court and housed in its treasures.41 But at this stage of research it is impossible to identify these textiles from among the hundreds of remaining ones and, what is far more important, there is some uncertainty whether major luxury techniques were tied to specific centers of manufacture or were available at every important court. These observations on the techniques of Fatimid art have one important corollary for our purposes. It is the near impossibility of defining a Fatimid technical originality. Perhaps there was none, as the eleventh-century Muslim world had already developed in various places and for various reasons the major means for its artistic expression. In this respect, Fatimid art is at the end of a period of artistic growth. The twelfth century with its new centers in Iran, Iraq and Anatolia will be quite different. The last point concerns Fatimid ornament. It is almost impossible to generalize on this little-studied topic. Its greatest curiosity seems to be that it moved away from the total abstraction of Iraqi-influenced Tulunid designs and preferred luxurious vegetal patterns or often naturalistic, even if fragmentary, animals, before becoming affected, in the twelfth century, like the rest of the Muslim world, with a fascination for complex geometric shapes.42 Liveliness and vivacity characterized the central period of Fatimid ornament. Partial and incomplete though they are, the preceding remarks lead to conclusions on Fatimid art at three different levels. A first one is primarily local. Even though the North African and more generally western Islamic connections of [220] Fatimid art become stronger and more interesting with every new excavation in Tunisia or investigation of a Maghribi technique,43 the most lasting impact of the Fatimids was on Egypt, whose social, economic and ecological character was irretrievably changed by the Fatimids. This is not to say that the peculiarly open symbiosis of different religious and ethnic groups so typical of most of Fatimid times was maintained over subsequent centuries. It is rather that this relatively open, mercantile system under the aegis of a spiritually committed dynasty consecrated the triumph of an Islamic taste in Egypt. However much we may know about the “things” owned by Jews, none of them seems to possess exclusively Jewish features 41 42 43

R. B. Serjeant, “Materials for a History of Islamic Textiles,” Ars Islamica, 9 ff. (1942 ff.); also the texts quoted in note 38. This ornament can best be studied in woodwork (see Pauty, Bois sculptés), or in architectural decoration (Creswell). It is particularly unfortunate that so many Tunisian excavations have not been properly published; S. M. Zbiss, “Mahdia et Sabra-Mansaouriga,” Journal Asiatique, 244 (1956); Lézine, Mahdiya. For ceramics, M. Jenkins, “Western Islamic Influences on Fatimid Egyptian Iconography,” Kunst des Orients, 10 (1975).

254

early islamic art, 650–1100

and, even if there is some debate on the matter, it is apparently under the Fatimids that the art of Christian Egypt – what used to be called quite erroneously Coptic art – becomes a folk art. From this point of view the Fatimid period can be seen as the culmination of a long process, in which the Tulunid half-century served as a major forerunner. But Tulunid art relied a great deal on Iraqi models. The Fatimid period did borrow from elsewhere but also rediscovered local sources, as Hellenistic or even ancient Egyptian motifs reappeared.44 Yet, even on a local level, it cannot be seen merely as the end of a tradition; for, with the creation of the city of Cairo and the artisanal as well as commercial development of the whole country, it established the physical foundations of centuries of Egyptian history. This was made possible, I submit, by the imperial and universally Islamic vocation of the Fatimids. The early Muslims, the ‘Abbasids and the Tulunids all created new urban establishments in a propitious part of the Nile valley, but it is the caliphal ambition of the Fatimids which transformed their own royal city into the catalyst which made the present Cairo possible. A second level of Fatimid art is interesting and important for its time, but perhaps less significant for later developments. It may be called its Mediterranean level. From Byzantine models in forms and ceremonies to Fatimid-inspired paintings in Norman Sicily or to romanesque objects, a Mediterranean “connection” is a constant aspect of Fatimid art, just as its society has been called Mediterranean. In part, of course, this is the practical result of Egyptian trade and of Fatimid sectarian liberalism. But it is curious to note the degree of consciousness [221] of Mediterranean objects which permeates texts like the Kitab al-dhakha’ir or Maqrizi’s description of imperial treasures. Byzantine ambassadors made a major point of bringing fancy gifts to the court of Cairo and some of the greatest masterpieces of Fatimid art now in Venice came from Byzantium.45 The contrast is striking with the comparative paucity of objects from the Muslim East. It is difficult to decide whether this Mediterranean taste of the Fatimids was a conscious choice or simply the result of political failure in the East. The Crusades and the eventual victory of an eastern Islam altered the character of Egyptian taste, although traces of its earlier tendencies appear occasionally throughout Mamluk architecture. Finally there is an Islamic, or Islamicate as Marshall Hodgson would put it, level of Fatimid art. It is the most fascinating and most puzzling one of all. As I tried to show earlier, Fatimid art is traditional in many ways and 44

45

Jones, AARP, 7; E. J. Grube, “Studies in the Survival and Continuity of pre-Muslim traditions,” Journal of the American Research Center in Egypt, 1 (1962). This whole problem permeates the existing controversies around the dating of the so-called “Coptic” textiles. With several students, I am preparing an annotated study of the texts about treasures. For the treasures of Venice, see K. Erdmann, A. Grabar and others, Il tesoro di San Marco (Venice, 1971).

fatimid art, precursor or culmination 255

hardly inventive in techniques, types of objects, or architectural functions. Yet its innovations – the muqarnas, the mausoleum, representational art – are almost always forerunners of the great changes which swept most of the Muslim world in the twelfth century. One could argue, as has been suggested for mausoleums,46 that Fatimid novelties were picked up by the eventually dominant Sunni world and transformed into vehicles for their views and ideas. Since almost every one of these innovations was less a true invention of the Fatimids than a minor motif in the earlier art of the tenth century, one could further argue that they were from the very beginning expressions of a Shi’ite taste, a point which may find some confirmation in some parallels between the Fatimids and Buyids. Yet, on balance, I doubt that Fatimid art should be interpreted on sectarian grounds. A more plausible explanation of their originality may lie in the coincidence under their aegis of imperial ambitions, of a mercantile society, and of cultural self-confidence. The forms and techniques created over the previous centuries were sufficient for their aims and there no longer was any need to control or limit the endless range of individual self-expression. Fatimid art was both the culmination of the past and the forerunner of future developments, because Fatimid culture was a novel one in a newly developed Islamic land but its legitimacy lay in the past.

46

O. Grabar, Ars Orientalis, 6, p. 39; J. Sourdel-Thomine, “Les anciens lieux de pèlerinage,” Bulletin d’Etudes Orientales, 14 (1951–4).

Chapter XV Notes sur le Mihrab de la Grande Mosquée de Cordoue*

Peu de monuments musulmans sont aussi connus que la Grande Mosquée de Cordoue. Décrite et étudiée depuis le début du 19ème siècle, elle a fait l’objet de toute une série de travaux de détail récents et nous devons à Henri Stern, Christian Ewert, Klaus Brisch et M. Ocaña Jiménez d’avoir mis à la disposition des savants une documentation complète et des interprétations nouvelles sur les mosaïques, les inscriptions, la décoration et les techniques de construction.1 Par ailleurs, les renseignements épigraphiques et historiques qui nous sont parvenus sur cette mosquée sont d’une richesse rare pour une aussi haute époque. En fait, ce n’est qu’à partir du 14ème siècle que les chroniqueurs ou autres lettrés arabes et iraniens semblent s’être préoccupés, rarement d’ailleurs, d’une manière systématique à la Maqrizi, des monuments qui les entouraient ou qui se construisaient autour d’eux. L’Espagne omeyyade est une exception qui a l’avantage de nous fournir des renseignements précieux, quoique problématiques et parfois contradictoires, sur ses monuments et mon but est simplement d’attirer l’attention sur quatre particularités du mihrab de la célèbre mosquée (Figs 1 et 2).

1.

La forme du mihrab et son effet

C’est là l’aspect du mihrab qui a été étudié le plus fréquemment et je me borne à résumer des observations qui ne sont pas nouvelles. Je voudrais seulement faire remarquer que la recherche si commune des origines de telle ou telle forme me semble avoir exagéré l’importance diachronique du mihrab de Cordoue – à savoir sa place dans une séquence chronologique allant de Damas ou Kairouan vers l’Aljaferia ou les mosquées Almohades – sans avoir

* Premièrement publié dans A. Papadopoulo, ed., Le Mihrab dans l’Architecture et la Religion Musulmanes (Leiden, 1988), pp. 115–18. 1 Henri Stern, Les Mosaïques de la Grande Mosquée de Cordoue (Berlin, 1976); Christian Ewert, Spanisch-islamische Systeme (Berlin, 1968); Klaus Brisch, Die Fensterglitter der Hauptmoschee von Cordoba (Berlin, 1966).

257

258

1 Cordoue, mihrab de la Grande Mosquée, 965

early islamic art, 650–1100

le mihrab de la grande mosquée de cordoue

259

2 Cordoue, détail du mihrab de la Grande Mosquée

260

early islamic art, 650–1100

résolu le problème synchronique du mihrab: pourquoi fut-il construit et décoré tel qu’il est entre 965 et 971? Il s’agit, comme on le sait, d’une petite chambre octogonale de 3,50 mètres de diamètre recouverte d’une coupole transformée en élégante coquille. Trois baies carrées recouvertes de coupoles et admirablement décorées forment la trame architecturale de la célèbre maqsurah devant le mihrab. De chaque côté du mihrab se trouvaient cinq petites pièces; la série occidentale servait de passage vers le palais, tandis que le groupe oriental était le trésor de la mosquée. A l’étage il y a onze petites pièces dont la fonction est inconnue. En dehors des qualités exceptionnelles de l’ornementation du mihrab et de la maqsurah, deux caractéristiques formelles le rendent unique parmi les mihrabs des premiers siècles de l’Islam: (1) le mihrab fait partie d’un ensemble architectural comprenant une variété de formes (par exemple les coupoles de la maqsurah, la hiérarchie des arcs entrecroisés analysée par Ewert, ou bien l’équilibre entre décor sculpté et décor en mosaïques) et plusieurs fonctions (prière, passage, trésor). Le mihrab est le centre de cet ensemble, mais la complexité de l’ensemble finit par dominer le mihrab; (2) de loin le mihrab apparaît non pas comme une niche faisant partie du mur de la qibla mais comme une porte ouverte vers l’invisible ou l’inconnu. Il est possible qu’on y suspendait une lampe allumée, mais il est peu vraisemblable qu’elle ait été allumée à toute heure. Nous nous trouvons donc en présence d’un signe architectural dont le caractère ou l’effet ne sont pas constants mais varient avec les heures ou les jours. L’interprétation du mihrab comme porte est confirmée par le fait que ses formes (arc en fer à cheval encadré et surmonté d’une arcade) sont les formes même des portes de l’architecture omeyyade du 10ème siècle à Cordoue. [116] Les mosquées de Kairouan et de Tunis avaient déjà été pourvues de grands ensembles architecturaux devant le mihrab et autour de lui. Mais le mihrab de Cordoue est le plus ancien des mihrabs préservés à avoir servi de module pour la composition de l’ensemble qui l’entoure et à avoir cet aspect de porte ouverte vers le néant ou bien vers la lumière.

2.

Les inscriptions

A l’intérieur du mihrab, une inscription de 965 relate le revêtement du mihrab avec du marbre, donne le nom de ceux qui s’en occupèrent et cite deux passages du Coran.2 Au début se trouve Coran 2:238: “Soyez assidus aux Prières ainsi qu’à la Prière Médiane et acquittez-vous [du Culte] envers Dieu, faisant oraison”.3 A la fin de l’inscription 31:22: “Quiconque se soumet 2 3

Répertoire Chronologique d’Épigraphie Arabe, 4 (Le Caire, 1933), no. 1578; E. LéviProvençal, Inscriptions Arabes d’Espagne (Leyde–Paris, 1931), no. 10. Traduction de R. Blachère, Le Koran (Paris, 1947).

le mihrab de la grande mosquée de cordoue

261

à Dieu (lit. “livre sa face à Dieu”), étant bienfaisant, s’est saisi de l’anse la plus solide. A Dieu revient la fin des choses.” Le choix de citations coraniques est inusité et curieux, d’autant plus que le premier exemple provient d’un verset au sens longuement débattu par les exégètes traditionnels. Sur les impostes des claveaux inférieurs de l’arc en fer à cheval se trouve une inscription, également de 965, qui se réfère, comme l’a bien montré Lévi-Provençal,4 à la mise en place dans le nouveau mihrab d’al-Hakam des colonnes provenant du mihrab d’Abdul-Rahman II. Laissant de côté le problème de vocabulaire architectural posé par cette inscription, notons simplement qu’elle est précédée de Coran 7:43: “Louange à Dieu qui nous a dirigés ici. Nous n’aurions pas été à même de nous diriger si Dieu ne nous avait pas dirigés. Certes les Apôtres de notre Seigneur sont venus avec la Vérité.” Le choix de citations est également curieux, car il s’agit d’une partie seulement d’un long verset donnant les paroles des Elus au Paradis. Il s’agirait de savoir si une citation provenant d’un contexte aussi précis implique l’ensemble du texte dont elle sort et suggère dans ce cas une référence au Paradis ou bien si son sens plus limité d’une direction divine (le verbe est en fait hada, indiquant à la fois direction et cadeau) donnée à une succession d’apôtres ou d’envoyés (rusul) a été extrait à dessein du contexte coranique. La réponse à cette question ne pourrait être fournie que par les historiens de l’exégèse du Coran. La grande inscription qui entoure l’arc du mihrab n’est malheureusement originale que dans sa partie centrale. Lévi-Provençal a démontré qu’il s’agissait de deux inscriptions commémorant séparément l’élargissement de la mosquée en général et la construction des arcs entrelacés de la maqsurah.5 Les deux citations coraniques du début sont 32:6 et 40:65 liées l’une à l’autre par la proclamation de la grandeur de Dieu et l’ensemble peut être reconstruit de la manière suivante: “[(Dieu) élabore l’Ordre (envoyé) du ciel sur la terre, puis l’Ordre remonte vers Lui en un jour dont la durée est de mille ans de votre comput;] Dieu qui connaît l’Inconnaissable et le Témoignage, est le Puissant, le Miséricordieux. Il est le Vivant. Nulle divinité excepté Lui! Priez-Le Lui vouant le Culte. Louange à Dieu, Seigneur des Mondes.” Au centre se trouve une des grandes proclamations de l’Unité divine, 59:23: “Il est Dieu, aucune divinité sauf Lui, le Roi, le Très-Saint, le Salut, le Pacificateur, le Préservateur, le Violent, le Superbe. Combien Dieu est plus glorieux que ce qu’ils [Lui] associent.”6 L’inscription de la coupole est entièrement coranique et contient les trois derniers versets de la 22ème surah, la proclamation des commandements divins: “O vous qui croyez, inclinez-vous, prosternez-vous, adorez votre 4 5 6

Lévi-Provençal, Inscriptions, pp. 13 et suiv. Ibid., no. 12. M. Ocaña Jiménez in H. Stern, Mosaïques, p. 49.

262

early islamic art, 650–1100

Seigneur, faites le bien. Peut-être serez-vous bienheureux. Menez le combat pour Dieu comme Il le mérite. Il vous a choisis et sur vous Il n’a placé nulle gêne en la Religion (din) et la millah de votre père Abraham. Il vous a nommés les Soumis, auparavant et en cette Révélation, afin que l’Apôtre soit témoin à votre encontre et que vous soyez témoins à l’encontre des Hommes. Accomplissez donc la Prière! Donnez l’Aumône! Mettez-vous hors de péril en vous attachant à Dieu! Il est votre Maître. Combien excellent est cet Auxiliaire.” La dernière série d’inscriptions que je voudrais mentionner est en mosaïques autour de la porte qui menait vers le palais. On y trouve Coran 41:30–32; 2:286; 3:8. Les trois fragments sont des prières indépendantes du texte qui les entoure et deux d’entre eux dénoncent les associations à la Divinité. [117] Les thèmes du choix de citations coraniques fait pour le mihrab me paraissent clairs: prières et exhortations à l’entrée ou sortie princière; proclamation de la Gloire Divine autour du mihrab et dans la coupole avec emphase sur l’Ineffable Vérité de Dieu et l’impossibilité de lui associer quoi que cela soit; l’obligation de la prière dans l’ensemble des devoirs du croyant; et enfin des références plus précises sur le succès de l’achèvement du monument agrandi. Il s’agit, en somme, d’un programme iconographique dans lequel se mêlent, comme dans tout programme iconographique, des généralités bien connues des croyants et des références localisées. Il n’y a, à ma connaissance, que deux exemples préservés de mosquées (j’exclus un monument à caractère unique comme le Dôme du Rocher) antérieures à celle de Cordoue qui auraient utilisé des inscriptions coraniques en nombre suffisant pour en justifier une interprétation iconographique. Il s’agit des mosquées d’Ibn Tulun et de Kairouan dont l’étude de ce point de vue est encore à faire. Mais, dès la fin du 10ème siècle, les grandes mosquées fatimides du Caire sauront employer le Coran d’une manière qui semble toute nouvelle, s’il n’y avait pas l’exemple antérieur des Haramayn, car c’est bien à La Mecque, à Médine, et d’une manière moins nette à Jérusalem, que les grandes inscriptions deviennent dès le 9ème siècle les signes chargés de sens qui transforment des formes souvent répétées et imitées en symboles concrets.7

3. Textes Deux textes permettent, me semble-t-il, de conclure à l’existence à Cordoue de ce que l’on pourrait appeler une liturgie de la prière. Le premier texte est certes tardif et il n’est pas absolument certain qu’il soit valable pour la mosquée omeyyade de Cordoue, mais il se réfère sans 7

Une série d’études sur ces monuments par Caroline Williams et Jonathan Bloom vont paraître incessamment.

le mihrab de la grande mosquée de cordoue 263

aucun doute à une pratique connue dans la mosquée avant 1157. Il s’agit d’un passage de Maqqari qui décrit le grand Coran attribué au caliphe Uthman qui se trouvait dans la mosquée et qu’y avait vu Idrisi.8 Chaque vendredi, deux serviteurs le tiraient du trésor où on le conservait et, précédés d’un troisième serviteur tenant un cierge, le portaient au kursi en bois d’aloes décoré de clous d’or qui lui était réservé a l’endroit d’où l’imam dirigeait la prière. Dès que la lecture du livre était terminée, on le rapportait dans le trésor. Le deuxième texte est dans le Bayan d’Ibn Idhari et relate que les muezzins, avant l’appel à la prière, venaient à l’intérieur de la mosquée, s’inclinaient devant le mihrab et puis montaient au minaret.9 Il s’agit d’un exemple rare mais non pas unique de la cérémonie de l’iqamah qui serait d’origine mecquoise.10 Nous connaissons bien mal le développement de la pratique religieuse à l’intérieur des mosquées,11 mais le premier de mes deux exemples est à tel point insolite que ce n’est que par l’influence de la liturgie chrétienne qu’il pourrait être expliqué.

4.

Événements symboliques contemporains

Les chroniqueurs de l’Espagne omeyyade ont beaucoup écrit sur la mosquée de Cordoue et ces textes ont été utilisés par les historiens du monument pour en reconstituer le développement. Mais, quelle que soit leur valeur purement documentaire, ces textes ont aussi une dimension symbolique précise: ils servent à rattacher la nouvelle mosquée aux traditions et habitudes antérieures, souvent même étrangères à Cordoue.12 Quelques exemples suffiront. Les textes insistent sur le remploi des colonnes de la mosquée précédente dans le nouveau mihrab ainsi que sur la reconstruction de l’ancienne maqsurah en bois. Comme à Damas au début du 8ème siècle, on fait appel à des mosaïstes de Constantinople pour la décoration du mihrab. Plus curieuse est l’histoire de l’orientation du mihrab d’al-Hakam. Les architectes, qui eurent gain de cause, voulaient maintenir la direction Sud des mosquées précédentes, tandis que les astronomes auraient préféré l’Est, les deux directions étant presque également inexactes. L’argument qui fit pencher en faveur du Sud fut que les Compagnons qui avaient créé les 8 9 10 11 12

A. Dessus-Lamare, “Le Mushaf de la mosquée de Cordoue et son mobilier mécanique,” Journal Asiatique, 230 (1938), p. 555. Ibn Idhari, Bayan, éd. G. S. Colôn et E. Lévi-Provençal (Leiden, 1951), vol. II. J. Pedersen, art. masdjid dans la première édition de l’Encyclopédie de l’Islam, section H 4. Ibid.; voir aussi des études comme Klaus Lech, Geschichte des islamischen Kultus, dont seul le premier volume a paru (Wiesbaden, 1979). G. Marçais, Architecture Musulmane d’Occident (Paris, 1954), pp. 139–40.

264

early islamic art, 650–1100

premières mosquées s’étaient tournés vers le Sud. Autrement dit, al-Hakam favorisa une opinion basée sur l’histoire et la tradition locales plutôt qu’un argument basé sur des [118] calculs qui n’avaient été établis que fort récemment13 et qui semblent avoir été mal faits à Cordoue. Quelle que soit l’interprétation à donner à tel ou tel événement associé au mihrab de la mosquée, leur rédaction dans les chroniques sert en fait à formuler ce que l’on pourrait appeler le pedigree historique et symbolique du mihrab. Que pouvons-nous conclure de ces remarques préliminaires, ces notes de lecture, sur les documents si complexes et si variés qui ont été préservés autour du mihrab de Cordoue? Tout d’abord, ce mihrab est bien plus qu’un signe indiquant la direction pour la prière, plus même qu’un endroit dans la mosquée commémorant la présence du Prophète, comme je l’avais proposé pour le mihrab en général.14 Ce n’est pas non plus un symbole royal, quoique placé dans une maqsurah royale et cérémoniellement rattaché à la porte d’entrée du prince. Le mihrab de Cordoue est un essai, peut-être unique pour l’époque,15 d’iconographie architecturale. Les formes uniques – une chambre flanquée de deux rangées de pièces à fonctions précises et précédées d’un avant-corps somptueux – l’organisation du décor épigraphique autour de grands thèmes religieux allant de la proclamation de la foi à des prières individuelles et au registre précis des événements ayant créé le mihrab, le détail des textes contemporains ou postérieurs décrivant le mihrab ou les cérémonies qui l’entouraient, ce sont là, me semble-t-il, des arguments démontrant une volonté sociale ou individuelle à exprimer quelque chose de précis et peut-être de nouveau. Qu’était-ce? Au risque de me lancer dans une hypothèse aventureuse, je voudrais suggérer qu’il s’agissait d’une réponse idéologique aux deux pressions constantes dans l’Andalousie omeyyade: l’Islam qui proclame ses sources locales, syriennes et hijaziennes et un monde chrétien à la fois soumis et turbulent. C’est en réponse aux cérémonies chrétiennes que l’on aurait transporté le Coran à travers la mosquée avec un acolyte tenant un cierge. C’est, par contre, de la Mecque que proviendrait la prière spéciale des muezzins. C’est à cause du précédent de Damas que l’on fit appel aux mosaïstes byzantins et c’est un débat intense de la communauté musulmane qui se reflète dans le problème de la qibla. Cette iconographie du mihrab de Cordoue s’explique par les contingences uniques de l’Andalousie omeyyade du milieu du 10ème siècle. Mais une deuxième conclusion s’impose peut-être. C’est que la dimension iconographique 13 14 15

D. King, “Kiblah,” Encyclopédie de l’Islam, 2ème ed. O. Grabar, The Formation of Islamic Art (New Haven, 1973), pp. 120 et suiv. C’est à dessein que j’hésite à le considérer comme unique, car des études détaillées sur d’autres monuments anciens risquent de démontrer que les créations musulmanes des premiers siècles avaient souvent des significations bien plus compliquées que l’esprit décoratif dont on les affuble si souvent.

le mihrab de la grande mosquée de cordoue 265

du mihrab de Cordoue n’existe que par la présence de “vecteurs” qui rechargent un système de formes ou de signes plus ou moins neutres en soi. Ces vecteurs sont les inscriptions et les événements qui restent associés au mihrab. Mais, dans le cas de Cordoue, la forme même du mihrab a changé. Et, s’il est juste de proposer une explication complexe du mihrab à partir d’inscriptions et d’événements contemporains, n’est-il pas vraisemblable qu’un même ordre de volonté iconographique préside à l’invention d’un mihrab-chambre au lieu du mihrab-niche traditionnel? Pour expliquer cette forme, je voudrais citer un texte plus ou moins contemporain de Mas‘udi qui explique qu’Adam fut créé par Dieu “comme un mihrab, une Ka‘ba, une porte sainte ou une qibla vers laquelle les purs esprits et les anges de lumière doivent se tourner pour prier”.16 L’équivalence entre mihrab, porte, et direction me semble indiquer que le monde musulman du 10ème siècle, comme d’ailleurs le monde omeyyade de Syrie autour de 700, était à la recherche de formes chargées de signification symbolique. Le mihrab de Cordoue serait un exemple de ces formes nouvelles, mais cette conclusion ou plutôt hypothèse demanderait à être vérifiée par des études sur les arts et la littérature de l’époque qui vont bien au delà des simples observations sur un seul monument que je voulais faire.

16

Mas‘udi, Les Prairies d’Or, éd. et tr. Barbier de Meynard et Pavet de Courteille (Paris, 1880–1905), I, pp. 57–8. Cité par P. Ravaisse, “Sur trois mihrabs,” Mémoires de l’Institut Egyptien, II (1889), p. 626.

Chapter XVI Two Paradoxes in the Islamic Art of the Spanish Peninsula*

Introduction Of all the lands of the earth which have preserved masterpieces of Islamic architecture, or from which unique monuments of craftsmanship attributable to Muslim artisans or to Muslim patronage have come, two are no longer ruled by Muslims. They are India, the home of the Taj Mahal and of Fatehpur Sikri, and then there is Spain. Of the numerous sub-cultures which shaped European Christian civilization in the Middle Ages and in pre-modern times, two were for several centuries in close connection with and at times even subjugated by the world of Islam. One is the Eastern and South-Eastern European world of, for the most part, orthodox Christians and the other is a major portion of the Iberian Peninsula, more specifically that part of the Peninsula which has been called al-Andalus, the southern section of which has become the contemporary province of Andalucia; for, in the Middle Ages, al-Andalus was to Arab Muslim writers every part of the Peninsula under Muslim rule and control. I shall not, in the context of this essay, pursue the parallels between the intercultural contacts of the Iberian Peninsula and those of other parts of the Eurasian and African worlds, though I shall refer to them toward the end of my observations, as they may well provide a useful interpretative framework within which to see and to explain the art of Muslim Spain. What I shall try to show is that the art of Islamic Spain can be seen in two ways. It can be part of a large body of monuments known as “Islamic,” that is to say as made by or for people who professed the Muslim faith; or else it can be seen as Spanish or Hispanic, that is as the creation of a land with traditions which would have been, in part at least, independent of the religious, ethnic or cultural allegiances of rulers of the moment. Good arguments can be made, and have been made in the past, in favor of either one of these positions or approaches toward the arts of Muslim *

First published in The Legacy of Muslim Spain, ed. S. K. Jayyusi (Leiden, 1992), pp. 583– 91.

267

268

early islamic art, 650–1100

Spain; for indeed each one of them is justified by some at least of the factual characteristics of the monuments involved, but especially by reference to two diametrically opposed ideological positions. I and others in this volume will deal with the monuments. The ideologies are less easy to define. On the one hand, there lies the achievement of a land remote from the centers of Muslim power and creativity; and that achievement can be interpreted as a demonstration of the divinely inspired power of a Muslim ethos or of the [584] brilliantly superior cultural bind that tied together, through a single faith with many variants, as diverse a crowd as Turkified Iranians from Central Asia and the descendants of Arabicized Berbers and of Hispanic women. But there is an alternative position to what may be called a panIslamic ideology explaining culture through the forceful mediation of the faith and of the ethic attached to it. From this other point of view, the qualities of a land’s art are explained through the permanent operation of a national spirit, of an indefinable attribute of a land and of its past, through the presence of the “earth” and of the “dead,” as theoreticians of nationalism defined the nation in the early years of the twentieth century. The debate between these ideologies is not one in which someone who is neither a Spaniard nor a Muslim should intervene, but it is proper to wonder why it is that apparently incompatible attitudes of interpretation have emerged around the art of Islamic Spain, as they had also grown around its culture and indeed its very existence. I shall explore this question by identifying two apparent paradoxes concerning the art of Islamic Spain and by weaving various thoughts and observations around these paradoxes. The first is the apparently unique character, both typologically and aesthetically, of so many works of Spanish Islamic art. The second one is the unusual fit between forms assumed to be Islamic and patrons of art or settings for art which are not. In conclusion I shall return to some of the broader issues brought up at the beginning.

I.

The monuments of Spain

The Great Mosque in Cordoba is acknowledged as a major masterpiece of Islamic architecture, and many scholars have used it as a prototypical exemplar of the hypostyle mosque which creates large spaces for the whole community by multiplying a single support, in this instance the column with arches, in a flexible manner adjusted to increases and decreases in the population of believers. And it is true that, at a very simple and elementary level, the mosque of Cordoba is planned and designed according to principles comparable to those which created the mosque of Qayrawan in Tunisia, the Azhar or the mosque of ‘Amr in Cairo, the Mosque of the Prophet in Medina, the Aqsa mosque in Jerusalem, and, in slightly different ways, the large brick mosques of Samarra in Iraq and of Ibn Tulun in Cairo. All these

two paradoxes in islamic art

269

are buildings earlier than the Cordoban one or roughly contemporary with its latest phase in the fourth/tenth century. After the fourth/tenth century thousands of mosques, especially the ones in the Muslim West, would continue this hypostyle tradition. But to see the mosque of Cordoba as “just” another example of a wellknown type is to misunderstand the peculiar qualities of the building. As several present-day architects and architectural critics have pointed out, it combines a number of unusual features: a subtle harmony of proportions between [585] elements like thin columns and horse-shoe arches which are not themselves original; a geometry of the arch which gives it a feeling of repose rather than the strain of being a carrier of thrusts; an equilibrium between single supports and mass ensembles like naves; occasionally the conscious breakdown of nuclear forms like arches into segments which can then be recomposed in alternate ways; and, finally, the stunning mihrab with the three domes in front of it, an ensemble glittering with rich mosaics for the representation of highly composed vegetal motifs and for the copying of long written messages, and yet mysterious in the deep niche of the mihrab itself, which is like an empty chamber, or else the gate toward another realm than that of man. Some of these features, like the mosaic technique or the expensive mihrab area, can be explained by specific local contingencies, namely the politicocultural relations with the Byzantine world which explain the mosaics themselves and the existence of more elaborate ceremonies than was usually the case around the daily prayers required of all Muslims. In Cordoba, perhaps in imitation of Christian practices, the muezzins came and prayed in front of the mihrab before calling for prayer. There was in the mosque a gigantic copy of the Qur’an which required two men to move it, and in which were included four leaves from a Qur’an attributed to the caliph Uthman, a hero of Umayyad tradition, who had allegedly been assassinated while reading the Holy Book; drops of blood were in fact found on these pages, which had obviously become symbols for something much greater than pages of text. This Qur’an was carried around at prayer time preceded by an acolyte with a candle, just as the Gospels are carried in a church. But, even beyond such specific details, which are original to the mosque of Cordoba but which are typologically not different from objects associated with other mosques, two features differentiate the mosque of Cordoba from nearly all other Muslim congregational buildings. One feature is that so much about it has been recorded and maintained even by historians and geographers who wrote much later, after the city had been taken by Christians. It is as though collective memory, Muslim, and probably also Christian since that particular mosque has been preserved, recognized something unique about the Cordoban monument. The second feature is the consistency of aesthetic purposes in the building, that is to say of creating visual effects which would affect the senses, which would give pleasure to the visitor or to the user. Few

270

early islamic art, 650–1100

other mosques (Ibn Tulun’s in Cairo is a major exception) are designed in such a way that everything in it, even later additions like the Christian chapels and the church, has to be done in the harmonic key of the constructions of the early third/ninth century. A concern for sensory effectiveness and for visual beauty is a hallmark of Cordoba’s mosque in ways that are more consistent, more fully anchored and more gripping than in most examples of congregational buildings within the medieval Muslim tradition. [586] An even stranger case is that of the ivory objects of the fourth/tenth and early fifth/eleventh centuries. Some twenty small boxes have remained, which were probably used for the storage of precious items, kerchiefs of different types or unguents. Many of them are dated and localized either in Cordoba or in the royal city of Madinat al-Zahra’ only a few miles from the urban center itself. Inscriptions often identify the owners of many of them as members of the ruling family or very high officials of the Umayyad state. In itself there is nothing unusual about expensive objects in a rare material being made for members of ruling classes in the Muslim world. Chronicles and other written sources are full of references to the fancy things and the fancy clothes which surrounded the princes and assorted aristocrats in Baghdad, Nishapur, Cairo, Herat, Rayy or Bukhara. But hardly anything has remained from these treasures, and one way of interpreting the Spanish ivories is to argue that they are an accidentally preserved set of princely artefacts of a type which would have existed elsewhere as well. In all likelihood it was the possibility of reusing these objects for church treasures which saved them from being destroyed or utilized, and then handled over the centuries to the point of becoming totally worn. To a certain extent this is probably the correct conclusion to draw. These ivories are indeed aristocratic household objects illustrating the wealth and the taste of the Umayyad court in al-Andalus. But there are several reasons for wondering whether we are not also dealing with a unique group of objects reflecting some uniquely local phenomena. I will mention just two peculiarities of these ivories which cannot be explained, at least within our present scholarly capabilities, in terms of a wider Muslim culture. One is that some among this group of objects – for instance the cylindrical casket of 357–8/968 in the Louvre, the 359–60/969–70 one in the Victoria and Albert Museum, and an undated one in the Museo Nazionale in Florence – are very deeply cut, so that the decoration on them appears in high relief, almost like the sculpture on antique and early Christian sarcophagi. This sculpted effect is, especially in the Louvre object, carried to the point where the personages, animals and plants of the design appear almost like freestanding sculptures in the round fixed on an object. Nothing like this is known in Islamic art elsewhere, nor, for that matter, is it known in early medieval Christian art. It is likely in fact that some antique model influenced the patron or artisans of these objects, but it is difficult to imagine how and why such an impression would have been sought.

two paradoxes in islamic art

271

The second peculiarity of some of these ivories is even more unsettling. The Louvre and Victoria and Albert examples, as well as several others in the Cathedral Treasury of Pamplona, the one in Burgos, and once again in the Victoria and Albert Museum, are decorated with personages and animals either arranged formally and symmetrically, as they often are on textiles, or else in what are clearly narrative or symbolic scenes: a prince enthroned, [587] wrestling, hunting, plucking eggs from a nest, riding elephants, picking dates, and so on. It is, first of all, remarkable that these scenes using personages in a narrative context occur in Spain nearly a century and a half before they become common in Egypt and the rest of the Islamic world. But even more remarkable is the fact that, while some of the representations would eventually become fairly common in Islamic art, most of them are unique. We are thus faced with the strange paradox of being unable to explain images which are easy to describe. At this stage we can only speculate about the reasons for these peculiarities of the Spanish Islamic ivories of the Umayyad period. They might have reflected, at the height of Umayyad power and wealth, the unusual cultural and artistic depth of the Umayyad court, in which new motifs are invented to give an old look, classical in mode, to the expensive materials brought from Central Africa. A hundred or more years later, under the rule of a Christian king, quintessentially Muslim motifs would adorn the ceiling of the royal chapel in the Norman palace of Palermo in Sicily. This later example suggests that in the Western Mediterranean a cultural mix was perhaps created different from that of places further east. A couple of minor points would confirm the sense of a difference in the art of Muslim Spain in its earliest and greatest time. The names of artists and craftsmen for objects and for architectural decoration have been preserved from Spain much earlier and much more frequently than from elsewhere in the Muslim world, as though the status of the artisan was higher there. And then it is interesting to note the visibility of the patronage of objects by women, again a phenomenon rare elsewhere at that time. The two earliest dated ivory objects were made for daughters of ‘Abd al-Rahman III, and one of the later ones was made for a Princess Subh. My third example is that of the most celebrated monument of Islamic art in Spain, the Alhambra. This is not the place to discuss either its archaeology or its stunning features which attract millions of tourists every year. What is important from the point of view I am developing in this essay is that it too is unique in Islamic architecture, even though everyone, from scholars who have written about it to Hollywood or rich Arab patrons from the Gulf who have copied it or imitated it or parts of it a thousand times, regards the Alhambra as so characteristic of Islamic culture that popular as well as sophisticated imagination has, since the early nineteenth century, woven its Orientalist fantasies around it. Yet it is curious that there is no other building, no other part of a known building, which resembles the Alhambra, some

272

early islamic art, 650–1100

later imitations in Morocco in particular notwithstanding. And it requires a considerable stretch of the imagination to see in the Top Kapi Seray in Istanbul, the palace of the Ottoman sultans or the later Safavid palaces of Isfahan and the Mughal palaces of India, more than occasional similarities with Granada’s masterpiece. We are less well informed about earlier and contemporary [588] palaces around the Mediterranean, but what is known for instance about the citadel of Cairo in the heyday of Mamluk rule bears very little relationship to the Alhambra. It is maybe just possible that a dying Muslim dynasty in al-Andalus did not create a “typical” palace belonging to a set which has disappeared elsewhere, but something adapted to its own specific history and to its own specific needs and expectations. The mosque of Cordoba, the fourth-/tenth-century Umayyad ivories and the eighth-/fourteenth-century Alhambra are all unique monuments which fit uneasily within the generic cultural types with which they have usually been associated. And yet all three – as well as several additional ones like a number of silks and bronzes, or the small mosque of Bib Mardum in Toledo – illustrate functions and tastes which were indeed part of the traditional and classical ethos of the Islamic world: the large congregational mosque, the princely household object of great value, the luxurious setting for the life of rulers. None of these needs, except to a degree the second one, was significant to the medieval Christian world. Their Spanish expression, however, seems to have obeyed other constraints, other forces than those which obtained elsewhere in the Islamic world. Why?

II.

Islamic forms and non-Islamic patrons

The second paradox I would like to develop is easier to define than the first, but equally difficult to explain. It has long been noted that the forms of Islamic art lingered on in Spain much longer than in Sicily or in the Balkans or Russia, where they had hardly affected the arts of the local population (except in clothes), even during Muslim domination. Examples abound. Pedro the Cruel’s Alcázar in Seville comprises architectural forms associated normally with Islamic art, and, in the decorative cartouches of plaster which appear everywhere, his name is clearly written out in Arabic letters. For several centuries the churches of Toledo and Saragossa utilized real or blind decorative arcades which come out of the façades and minarets of the earlier Islamic tradition. As profoundly Christian a building as the so-called “tempietto” in the monastery of Guadalupe bears unmistakable and obviously deliberately chosen traces of medieval Islamic themes. In Burgos, one of the main centers of Spanish life to escape Muslim rule, and which became one of the centers of the reconquista, the monastery of Las Huelgas, in the early thirteenth century, was designed in part as a commemorative monument for Alfonso VII, one of the main Crusaders

two paradoxes in islamic art

273

against Muslim power in the South. But not only is its stucco decoration entirely taken from Islamic models, but the textiles which had been kept there, often as shrouds, were for the most part either manufactured by Muslims or imitated Muslim types. Ceramic production remained for centuries under the influence of the high-luster techniques developed in the Muslim world and brought relatively late to Spain. And two remarkable synagogues built in [589] Toledo under Christian rule – one from the twelfth century, known today as the church of Santa María la Blanca, the other dated 1357, and transformed into a church under the name of El Tránsito – were decorated in the purest style of Islamic ornament. This is all well known, and for over a century now scholars have identified examples of what has been called Mudejar art, an art of Muslim forms within a non-Muslim context. Even its migration into Mexico and Peru has occasionally been noted. What is more puzzling is that this preservation of allegedly Muslim forms often took place while Islam itself and those who professed it were persecuted, often quite brutally, and eventually physically expelled from the Iberian Peninsula. Gothic art coming from the North appeared at times like an outright intruder within a formal system which would have been the accepted genuine local one; and it is only with the Italianate taste of the Renaissance that Islamic motifs began to fade away, as in the lovely House of Pilate in Seville. But, even then, Charles V built his grandiose palace of Granada next to the Alhambra, dominating it no doubt, as a victorious culture would, but recognizing something of its values by preserving it. And earlier Alfonso the Wise was deeply imbued with Muslim values and aware of all that went into the making of a cultivated Muslim Arab. How can one explain the contrast between policies that were leading to the destruction of Islam’s presence in the Peninsula and this fascination with forms issuing from Islamic art, which continued quite consciously for several centuries, and, according to some, has remained in the background ever since? What, especially, is it that made Spain so different from other lands? As most paradoxes do, mine about Islamic art in Spain end up with questions. Both questions imply that something happened in Spain which is different from what happened elsewhere. There does not seem, a priori, to be any reason why the Islamic monuments of Spain should be qualitatively and typologically unique within the huge spectrum of Islamic art, even though their functions were not. And it is strange that a land which had invested so much physical and psychic energy in reclaiming from an allegedly alien power what was presumed to be its own would, for several centuries, maintain and carefully nurture the artistic forms of the enemy. To be able to reach an answer or answers to these questions, we must be willing to explore two propositions which go against well-established assumptions of the history of art, and perhaps of cultural history in general.

274

early islamic art, 650–1100

The first of these assumptions is that which involves labeling forms with cultural or national identifications. What seem to us today to be valid or even accurate means for the classification of visual evidence from the past, and for the appreciation of that evidence within our own, present-day minds, may not have been the appropriate criterion at the time when the monuments through which this evidence appears were created. If we consider a motif or a type of design as first of all colorful, geometric or vegetal, rather than [590] Islamic or Gothic or Byzantine, an appreciation of forms emerges which may well correspond more closely to what actually happened than the national and ethnic constructs we have posited. Alternatively, one can consider a motif as “ours,” as belonging to a tradition within a land rather than to a system of belief in that land. Analyses freed of prejudices may indeed begin to argue for a complex growth, in medieval Spain, of a common heritage of forms which was, in part if not as a whole, differentiated by its presence in that particular land rather than by its association with religious or national groups in that land. Within that heritage, some specific feature may be charged with an Islamic, Arab, Christian, Castilian or Catalan connotation, but such distinctions will only be reached after the realization that there was a common language for the expression of different thoughts and of different tastes and purposes. Perhaps after all it was other factors than those of cultural identification that predominated in the arts of the Middle Ages in Spain, and even elsewhere. The second issue to be pursued springs not so much from a possibly wrong-headed assumption as from one’s awareness of the position of alAndalus within the huge body of Islamic culture. It was a frontier area, at the outer edges of the dar al-Islam, and like all frontier areas it was endowed with a peculiarly paradoxical ethos in which intense identification of differences between groups and allegiances, at times warped by hate and contempt, coexisted with open-minded cohabitation and creative inventiveness. Thirteenth-century Anatolia, twelfth-century Sicily, Central Asia until the sixteenth century, were all frontier areas between opposing and at times warring factions of many different kinds. They were also areas of intense visual (and perhaps other) creativity, in which the desire to show off one’s unique qualities went along with competition with others and understanding of various ways of achieving visual effectiveness. With the advent of the rational doctrines issuing from the Renaissance, such tolerance became more difficult to maintain. It is obvious that these hypotheses and assumptions need elaboration and reflection before they can be fully accepted as explanations for the Islamic arts of the Spanish Peninsula in the Middle Ages. That they can even be raised is a testimony to the extraordinary quality of the centuries which revolutionized a land and expressed some of the best ambitions of a universal religious and ethical system created far away.

two paradoxes in islamic art

275

Bibliography This essay is based on commonly known volumes of information such as Vols III and IV of Ars Hispaniae and the monumental E. Kühnel, Islamische Elfenbeinskulpturen (Berlin, 1971). I have much profited from the following more detailed or more recent studies: [591] Beckwith, John, Caskets from Cordova (London, 1960). Delgado Valero, Clara, Toledo islámico (Toledo, 1987). Gómez Moreno, Manuel, El Panteón Real de Las Huelgas de Burgos (Madrid, 1946). Instituto de Estudios Turolenses, II Simposio Internacional de Mudejarismo: Arte (Teruel, 1982). Moneo, Rafael, and others, “La Mezquita de Córdoba,” Arquitectura, 66 (1985). Pavón Maldonado, B., El arte hispano-musulmán en su decoración geométrica (Madrid, 1975). Sánchez Albórnoz, C., L’Espagne musulmane, French translation of earlier Spanish version (Paris, 1985). Stern, Henri, Les mosaïques de Cordoue (Berlin, 1976).

Chapter XVII Qu’est-ce que l’Art Fatimide?*

Les historiens, de l’art ou d’autre chose, organisent l’immense passé du genre humain qui se trouve à leur portée en distribuant des adjectifs modificatifs de tout ordre aux mots “histoire” ou “art”. Ces modificatifs peuvent être géographiques (l’art africain), religieux (art chrétien ou bouddhique), nationaux (art irlandais) typologiques (art profane ou ornemental), chronologiques (art moderne ou du 16ème siècle), et puis toutes sortes de combinaisons que l’on pourrait appeler culturelles (art roman, gothique ou baroque). Une de ces catégories est dynastique, avec, comme toute catégorie, toute une série de présupposés, sinon préjugés, conscients ou non, sur lesquels je reviendrai plus d’une fois dans le courant de cet exposé. Il est, par exemple, assez curieux qu’une distinction dynastique dans l’art de l’Europe occidentale est normale pour le Haut Moyen Age (arts mérovingien, carolingien ou ottonien), mais semble réservée pour les époques plus récentes aux arts dits mineurs et pratiques (mobilier Restauration ou chaise Louis XV). Les distinctions dynastiques sont plus fréquentes pour Byzance (la “renaissance” macédonienne ou la peinture des Paléologues) et la Chine (art des Tangs si différent de celui des Sungs), mais dans les deux cas on peut faire valoir une histoire culturelle continue dans un espace nettement délimité et gouverné, du moins dans le cas de l’instance byzantine, à partir d’un centre bien établi. Ce n’est certainement pas le cas pour le monde islamique après les Omeyyades avec leurs nombreux monuments préservés et avant la grande séparation en orbites dynastiques ou pseudodynastiques (je pense aux Mamlouks) qui suivit l’invasion mongole. Et pourtant on parle d’art tulunide, aghlabide, samanide, seljuq, même s’il n’y a parfois qu’un seul monument ou œuvre d’art associé à cette dynastie. C’est comme si le caractère du pouvoir en place, en général celui d’une famille, détermina la nature d’un art. Et la première question qui se pose ainsi est la suivante: est-il juste de définir un art par le rappel du nom d’une dynastie régnante? Ou peutêtre faudrait-il poser le problème d’une manière différente: y a-t-il un art qui formerait un ensemble cohérent de sujets, formes et expressions et dont la présence coïnciderait avec celle de l’autorité politique et culturelle des Fatimides? *

Premièrement publié dans M. Barrucand, ed., L’Egypte Fatimide, Son Art et son Histoire (Paris, 1999), pp. 11–18.

277

278

early islamic art, 650–1100

Et puis, si l’historien des arts peut en effet préparer un tel inventaire de choses clairement datées et localisées et en déduire une définition d’un art différent de ce qui aurait été créé à la même époque mais dans d’autres régions du monde musulman, est-il juste et légitime d’en attribuer la responsabilité sinon aux princes de la dynastie, tout au moins à un climat social et culturel que ces princes auraient, volontairement ou non, rendu possible? Ces questions paraissent simples et raisonnables. En réalité elles contiennent plusieurs pièges méthodologiques. J’en citerai trois dont l’élaboration me permettra de proposer un certain nombre d’idées ou d’hypothèses pour la discussion. Le sens du premier piège m’est venu en cherchant le mot “Fatimide” dans le Larousse. J’y ai trouvé la phrase suivante: “Leur nom (celui des Fatimides) reste attaché à une période de l’art [12] musulman en Égypte, marqué par le souvenir de l’Ifriqiyah allié aux traditions tulunides.” Comme il se doit pour le Larousse, les choses sont claires, tout au moins grammaticalement, car, sémantiquement, il est difficile d’imaginer une alliance de souvenirs avec des traditions pour faire un art. L’art fatimide est l’art d’un pays particulier, l’Égypte, pendant une période de deux siècles (973 à 1171, dit le Larousse); cet art aurait d’une part maintenu une tradition locale et, de l’autre, pris à son compte un certain nombre de nouveautés provenant de l’Ifriqiyah, plus ou moins la Tunisie actuelle. On trouve dans cette définition un dosage de responsabilités digne des grandes organisations internationales. Mais, au début de notre siècle, par contre, c’était d’Iran que seraient venus les thèmes caractéristiques de l’art fatimide, car, et je caricaturise un peu l’argument, l’Iran étant devenu shi’ite au 16ème siècle, il est normal que le shi’isme ait véhiculé des éléments iraniens au 10ème; et puis une certaine tradition historiographique voyait l’Iran à l’origine de beaucoup de motifs artistiques. Sous ces formes, l’argument n’a pas de sens, mais un article récent sur les peintures de la Cappella Palatina à Palerme leur attribue un caractère fortement influencé par l’Iran à cause de l’aspect soi-disant fatimide de ces peintures.1 Nous nageons dans l’absurde, un absurde qui s’explique d’ailleurs assez facilement par un siècle et demi d’orientalisme qui demande de voir toutes les expressions de l’art musulman unifiées dans un seul système plutôt que d’accepter que les arts représentent un ordre de créativité qui exige l’apparition des différences et non pas la transformation constante d’une artificielle unité. Avec la montée des nationalismes ou, tout au moins, de natiocentrismes, on en arrive à la définition du Larousse avec deux pays ou zones géographiques associées à l’art fatimide. Or, il est facile de démontrer que la Palestine, la Syrie et le Hijaz ont également été transformés à l’époque des Fatimides, et souvent sous leur impulsion. La Jérusalem prise par les Croisés en 1099 était une ville fatimide dans ses formes, sinon dans son gouvernement, et c’est à eux que l’on 1

G. M. d’Erme, “Contesto Architettonico e Aspetti Culturale di Dipinti de la Cappella Palatina di Palermo,” Bolletino d’Arte, 92 (1996).

qu’est-ce que l’art fatimide?

279

doit non seulement toutes sortes de réparations à la Coupole du Rocher, y compris la construction du mihrab plat sous le Rocher même, mais la transformation radicale de la mosquée al-Aqsa avec la construction d’un arc triomphal qui, pour la première fois à Jérusalem cite le passage célèbre du Coran (17:1) mentionnant la Mosquée Lointaine visitée par le Prophète, et puis l’achèvement du projet omeyyade pour le Haram avec une porte monumentale de la ville au sanctuaire, ornée de mosaïques et décrite par Nasir-i Khosrow. Le minbar fabriqué en 1091 pour le martyrium construit à Ashkalon pour sauvegarder la tête, mystérieusement découverte, d’Ali se trouve actuellement dans le sanctuaire d’Hébron mais, avec l’étonnante inscription découverte récemment à Ashkalon également,2 elle fait partie de tout un groupe d’œuvres syro-palestiniennes d’époque fatimide. S’agit-il d’un art égyptien importé? Ou faut-il ajouter encore une région à un monde fatimide pluraliste? Je voudrais appeler mon premier piège, le piège de l’espace historique. Il s’agit de savoir si un terme dynastique n’est qu’un subterfuge pour cacher les réalités qui seraient géographiques ou même religieuses, sectaires ou confessionnelles, une manière commode pour identifier une période dans la continuité d’un art local, en l’occurrence égyptien. Le deuxième piège pourrait s’appeler celui de l’archéologie tentatrice. Il a l’avantage d’être méthodologiquement beaucoup plus précis que le premier, car il part toujours de documents concrets. J’en donne un exemple inédit dont je dois la connaissance au fonds israélien pour la recherche scientifique. Les fouilles récentes et par ailleurs si importantes de Césarée, la capitale de la Palestine byzantine, ont mis à jour un trésor contenant une trentaine d’objets utilitaires en bronze qui, d’après son contexte archéologique, serait d’époque fatimide. Évidemment, rien, à ma connaissance, ne dit que ces objets ont tous été fabriqués en même temps et au même endroit, ni même qu’ils provenaient de la même région. Autrement dit, comme il est souvent vrai pour les [13] renseignements d’origine archéologique, il faut prendre garde à la manière dont on tire des conclusions d’une documentation qui semble claire. Tout ce qu’on peut dire, c’est que ce groupe d’objets avait été assemblé à une certaine époque (autour de l’an 1000, donc à l’époque fatimide) et qu’il reflète la culture matérielle et le goût d’un instant dans l’histoire. On peut imaginer plusieurs scénarios pour expliquer le rassemblement de ces objets dans un endroit connu et à une date précise et j’espère que les savants qui se préparent à les étudier sauront trouver le plus vraisemblable. Entre temps, on peut dire que cet ensemble existait à l’époque, mais on ne peut pas en déduire nécessairement qu’il s’agissait d’objets d’art fatimide, quelle que soit la tentation de le faire. Le petit nombre de fouilles systématiques et complètes empêche l’évaluation adéquate de celles qui ont été faites, car il manque une masse critique de 2

M. Sharon, “A New Fatimid Inscription from Ascalon,” Atiqot, 26 (1995).

280

early islamic art, 650–1100

documents nécessaires à toute conclusion qui irait au-delà d’un site spécifique. Pour cette raison, la documentation la plus authentique que nous puissions avoir nous échappe encore en grande partie. Tels ont été les cas de la céramique d’al-Mina, de plusieurs sites fouillés en Syrie et en Palestine, et surtout de l’énorme documentation archéologique provenant de fouilles officielles ou clandestines en Égypte depuis plus d’un siècle et pour lesquelles on voudrait bien avoir une liste annotée et commentée. Un troisième piège est celui de la tentation des sources. J’en citerai trois exemples. Le premier est celui, bien connu, des objets en cristaux de roche. Trois exemples bien connus contiennent des inscriptions dont deux mentionnent le nom d’un calife fatimide ou bien un titre associé à un calife. Le groupe tout entier de ces cristaux est généralement attribué aux Fatimides, avec parfois des dérogations nécessitées par des styles incompatibles. Comme il s’agit par ailleurs d’objets associés, grâce à des références littéraires en plus des inscriptions, aux califes eux-mêmes, ils sont facilement installés dans une série “art impérial ou royal, manière islamique” et entrent dans la grande lignée des arts profanes des cours du haut Moyen Age et éventuellement des trésors des églises. C’est ainsi que la belle carafe au nom d’al-‘Aziz, vraisemblablement donnée en cadeau à un empereur byzantin, s’est retrouvée au trésor de St Marc à Venise. Le cas des cristaux de roche est intéressant, car, si une date correspondante à l’époque fatimide pour beaucoup d’objets est tout à fait vraisemblable, il n’en est pas de même pour leur provenance; ces objets ne forment pas un ensemble stylistique clair et il semble parfois plus raisonnable d’y voir surtout des objets appartenant à un certain niveau social – les princes et les rois – plutôt qu’à une chronotopie, fatimide ou autre. Le destin éventuel de beaucoup de ces cristaux dans des milieux religieux ou profanes bien éloignés de l’Islam correspond en fait à la réalité de leur importance. Ils étaient faits pour une classe d’utilisateurs, pas pour une région ou pour un milieu précis. Mon argument serait que les objets de cour ont une autre valeur sémantique que celles de l’époque ou du lieu qui les a produits ou qui les utilise. Un autre exemple de tentation des sources est celui des peintures du plafond de la Chapelle Palatine à Palerme terminée autour de 1140 par ordre de Roger II, le grand roi normand de Sicile. Aucune documentation directe, à ma connaissance du moins, ne rattache les activités et intérêts culturels des rois de Sicile à ceux d’un monde fatimide plus ou moins restreint à cette époque à l’Égypte et à une petite partie de la côte palestinienne. Mais, dans un article célèbre publié il y a plus d’un demi-siècle, le très grand savant et connaisseur qu’était Richard Ettinghausen avait proposé, même avant la publication du grand livre de Monneret de Villard sur les peintures du plafond de Palerme, de voir dans ces dernières les traces d’une peinture fatimide plus ou moins disparue.3 Quoique modifiée parfois dans le détail, 3

R. Ettinghausen, “Painting in the Fatimid Period, a reconstruction,” Ars Islamica, 9 (1942).

qu’est-ce que l’art fatimide? 281

l’argumentation d’Ettinghausen a été maintenue depuis sa publication en 1943, car elle reflète deux procédures, je dirais même marottes, de l’histoire de l’art. L’une est d’établir une généalogie pour toute création artistique et, comme des modèles occidentaux ou byzantins manquent ou bien sont rares, c’est dans une Asie occidentale, des Sassanides en Iran aux ‘Abbasides de Samarra et de Baghdad, que l’on se doit de trouver les [14] origines des fresques siciliennes et le “missing link” darwinien nécessaire pour cette évolution se doit d’être en Égypte. Le contexte de ce colloque n’est pas le bon moment pour faire une critique de ce darwinisme historique qui domine encore l’histoire de l’art, mais il est évident que l’établissement de toute génétique des formes va à l’encontre d’une autre direction, également puissante et beaucoup plus à la mode d’aujourd’hui, celle de l’originalité indépendante de chaque région. L’autre courant de pensée qui se reflète dans l’hypothèse d’Ettinghausen est celui de l’importance des “grands centres” de création, ce qui donne automatiquement une importance particulière au Caire et à l’Égypte placés au centre des grands courants économiques et culturels, au détriment par exemple de l’Ifriqiyah et du Maghrib, voire de la Sicile ellemême, qui étaient bien plus riches en nouveautés que l’Égypte au milieu du 12ème siècle; les choses étaient bien différentes au 11ème, mais, hélas, c’est au 12ème que le plafond fut décoré. Pour notre propos d’expliquer et de comprendre l’art fatimide, l’essentiel est que les peintures de la Chapelle Palatine, le document central pour reconstituer la peinture fatimide, comme d’ailleurs le célèbre manteau de Roger II, sont en dehors de l’orbite fatimide, dans un monde féodal occidental très particulier. Il faut comprendre ce dernier avant de pouvoir en identifier les origines peutêtre fatimides. Mais il est intéressant que le dernier livre sur le sujet de cette chapelle de Palerme, paru il y un an à peine,4 explique encore – et, je dois avouer, d’une manière sophistiquée et convaincante – les pratiques cérémoniales de Roger II à partir de la description du Caire faite par Nasir-i Khosrow cent ans plus tôt. Je ne nie pas qu’il y ait eu une filiation entre le monde des princes fatimides d’Égypte et celui des Normands de Sicile, mais cette filiation ou plutôt relation fraternelle, comme dans l’antique tradition de la “famille des princes,” a, me semble-t-il, un tout autre aspect, bien plus intéressant d’ailleurs, que celui des formes qui auraient traversé une grande partie de la Méditerranée sans que l’on en connaisse les transporteurs. Pour l’instant, il me semble dangereux de baser la reconstitution de l’essentiel d’un art de la peinture fatimide à partir d’un seul monument, aussi unique et étonnant qu’il soit, à plus de 1000 kilomètres de l’espace où cette peinture se serait trouvée. C’est un piège très différent qui est tendu par les sources écrites, particulièrement riches pour l’époque des Fatimides, mais également 4

W. Tronzo, The Cultures of His Kingdom, Roger II and the Cappella Palatina in Palermo (Princeton, 1997).

282

early islamic art, 650–1100

problématiques, en partie par le fait même de leur richesse. Les nombreuses citations de la Geniza qui parlent d’objets ou de techniques de confection et de pratiques de vente, la célèbre description du Caire fournie par le voyageur persan Nasir-i Khosrow au milieu du 11ème siècle à l’apogée de la gloire et de la force de la dynastie, les dizaines de passages de Maqrizi qui décrivent les monuments du Caire et les activités de ses princes, y compris la commande de toutes sortes de peintures publiques et privées, et qui répètent des récits pittoresques sur les talents des artistes présents au Caire, et enfin les comptesrendus du Kitab al-Dhakha’ir wa’l Tuhaf, récemment traduits en anglais,5 où des témoignages pris sur le vif du sac du palais fatimide en 1068 côtoient des récits détaillés des cadeaux officiels envoyés du Caire à Byzance, de Byzance au Caire, entre autres exemples de contacts entre cultures par le truchement d’objets de valeur, ce ne sont là que les principaux exemples d’une littérature abondante qui donne facilement l’impression que, pour une fois, la corrélation pourrait être faite entre des œuvres d’art également nombreuses, leur contexte culturel ou liturgique et même les opinions des contemporains sur ces œuvres. Or ce n’est pas le cas, du moins pas encore, car je n’ai pas perdu l’espoir que ces corrélations immédiates pourront encore se faire. Mais, pour l’instant, à ma connaissance et avec l’exception, majeure certes, des monuments de l’architecture du Caire, aucune œuvre d’art précise n’a pu servir d’illustration à un seul texte et aucun texte n’a expliqué un monument. Et, ce que les historiens de l’art reconnaissent comme des nouveautés majeures de l’époque fatimide – les [15] mausolées et la création d’une architecture funéraire, le muqarnas qui apparaît simultanément (ou est-ce simultanément?) en Égypte, Sicile et Afrique du Nord, la richesse d’une peinture sur céramique à reflets métalliques – n’ont justement pas été identifiés dans les textes. S’agit-il de deux domaines ou de deux discours séparés, celui des choses et celui des idées ou des souvenirs sur les choses? Ou bien devons-nous conclure que même l’abondance de ce qui est, à juste titre, considéré comme fatimide et de ce qui parle de choses fatimides n’est en fait qu’une petite partie de ce qui existait, une quantité insuffisante pour que l’on puisse s’attendre à des corrélations faciles? Pour toutes sortes de raisons, j’hésite à accepter l’hypothèse de trop de choses disparues; toute comparaison du texte de Maqrizi avec ce qui reste de la ville fatimide du Caire et de ses monuments architecturaux, même souvent restaurés, montre que le nombre de constructions disparues est relativement petit et concerne surtout les palais et autres bâtiments profanes. On pourrait, devrait peut-être, en déduire que le langage des mots et celui des formes ne sont pas tout à fait sur le même diapason et qu’ils doivent être bien compris séparément avant d’essayer de les unir. Mais c’est là un domaine méthodologique qui demande bien plus de réflexion que je ne peux lui accorder dans cet essai.

5

Gh. Al-Qaddumi, Book of Gifts and Rarities, 9 (Cambridge, 1996).

qu’est-ce que l’art fatimide? 283

Ainsi les trois pièges que j’ai cherché à identifier – le piège de l’espace historique et culturel autour d’une Égypte dominante, le piège de l’archéologie insuffisante, le piège de la tentation des sources – permettent de formuler deux paradoxes autour de l’art fatimide. L’un serait que c’est un art régional centré sur l’Égypte qui se définit plus facilement par comparaison avec Byzance ou la Sicile, même les Umayyades d’Espagne, qu’avec les ‘Abbasides et l’Orient musulman, mais c’est un art issu d’un mouvement profondément lié à l’Islam classique et à ses structures. L’autre est que la vaste majorité de ce que nous savons de cet art, par les choses qui restent ou par les textes, est profane. Ainsi se pose ce qui me semble être la question fondamentale de l’art (mais non pas de la culture) fatimide. S’agit-il d’un art islamique, c’esta-dire d’un art qui refléterait les traditions et les raisonnements d’une culture visuelle formée, surtout autour de Baghdad mais en partie aussi au Khorassan, au cours des 9ème et 10ème siècles? Ou bien s’agit-il d’un art méditerranéen bien plus rapproché de l’Andalousie et de Byzance, tous les deux en pleine floraison artistique au 10ème siècle et dont les contacts nouveaux sont avec une Europe occidentale en effervescence après l’an mil? Ces questions sont toujours d’actualité dans l’entendement que nous avons de la Méditerranée orientale, de l’Afrique du Nord et de l’Asie occidentale. Ce sont aussi des questions qui resteront toujours sans réponses définitives pour un historien, car elles dépendent autant du climat idéologique et intellectuel dans lequel l’historien opère que des renseignements fournis par des sources, elles-mêmes teintes par leur temps. Mais dans deux domaines il est possible d’envisager une recherche fructueuse qui permettrait, sinon de résoudre, du moins de revoir les paradoxes que j’ai proposés. Ces deux domaines sont l’architecture avec tous les arts décoratifs qui l’accompagnent et puis la céramique et je pense en fait presque exclusivement à la céramique à reflets métalliques dont la vaste majorité d’objets connus sont plus ou moins uniques et qui semble avoir vraiment été une céramique d’art. En fait beaucoup a déjà été fait dans les deux domaines et je voudrais simplement signaler deux aspects de cette recherche qui me semblent particulièrement utiles pour revenir aux grands problèmes de l’art fatimide. Le premier aspect est la chronologie. Les monuments architecturaux sont, en général, bien datés et permettraient ainsi de définir une évolution dans les fonctions et dans les formes, continuant d’ailleurs dans des chemins relativement bien rodés depuis l’évolutionnisme positiviste de Creswell. Même si les chronologie et typologie exactes du muqarnas, surtout en Égypte, ne sont pas entièrement claires, le fait même de son apparition sous les Fatimides est indéniable et il s’agit simplement d’en évaluer la signification: était-ce une technique de construction de voûte, un signe indicatif de mode ou de prestige, ou bien un symbole religieux ou autre? L’extériorisation du décor des mosquées comme message public est un autre exemple d’une nouveauté (quoiqu’il faudrait encore en expliquer la relation, si tant est qu’il y en ait une, avec la mosquée des Trois Portes à [16] Kairouan) qui a été expliquée par la nécessité d’ostentation

284

early islamic art, 650–1100

visuelle pour une dynastie hétérodoxe, une nécessité idéologique qui serait devenue une forme visuelle incorporée dans plusieurs traditions locales et continuées par les siècles qui suivirent. Et puis l’apparition au 12ème siècle dans les boiseries fatimides d’une structure géométrique complexe remplaçant les variations de l’arabesque qui prévalaient auparavant pourrait être expliquée par un changement de goût ou bien par l’adaptation d’une iconographie de la géométrie comme reflet ou représentation de la création divine. Il est plus difficile d’établir une chronologie de la céramique d’art en l’absence de documents datés. Mais il semble juste de penser que la représentation d’êtres humains est plus fréquente à partir de la fin du 11ème siècle que pendant le premier siècle fatimide et, il y a une trentaine d’années, j’avais proposé d’y voir l’impact du sac du palais autour de 1068 qui aurait distribué pour un plus grand public des images accessibles seulement dans la vie privée des princes.6 Il s’agirait de savoir si cette évolution iconographique est valable et si, par ailleurs, les contrastes entre une manière hiérarchique et à deux dimensions et une autre plus souple et avec des velléités de représentation en profondeur des corps humains et animaliers ou des choses comme des gobelets remplis de liquide sont des contrastes de goût ou d’époque. Il s’agirait également de distinguer les œuvres, en général uniques, provenant du Caire ou trouvés au Caire et celles des ateliers syriens, tunisiens et autres sous l’emprise des Fatimides. Seul ce genre de travail pourrait définir l’originalité ou, au contraire, son absence, dans l’art de l’Égypte. L’autre aspect de ces recherches, dont beaucoup ont été faites depuis les listes mémorables de feuilles et de fleurs préparées par Shafi’i ou bien qui sont en cours, serait d’isoler le vocabulaire ornemental de l’art fatimide, c’est-à-dire les motifs, floraux par exemple ou épigraphiques, qui se retrouvent régulièrement et dont la présence servirait d’indice exclusif mais pas nécessaire pour une identification fatimide des choses. Les comparaisons auxquelles ce genre de travail mène sont parfois douteuses, surtout lorsque des techniques très différentes avec des contraintes particulières ne permettent pas toujours de conclure à une similarité entre motifs, mais la quantité des décors fleuris des mosquées et sanctuaires divers et la décoration des céramiques devraient permettre l’établissement d’un vocabulaire qui, en fait, n’a vraiment été fait que pour les inscriptions, grâce aux grands travaux précurseurs de Flury. On ne peut évidemment pas dire d’avance à quoi mèneraient ces travaux d’apparence taxonomique. Mais ce que mes remarques me permettent de signaler, en guise de conclusion, c’est que tous les paradoxes de l’existence même des Fatimides et les incertitudes qui planent sur les œuvres d’art attribuées à leur époque mènent à deux grandes manières de voir les choses et de les comprendre. L’une, traditionnelle dans l’histoire des arts, est celle de prendre chaque monument d’architecture, chaque groupement légitime d’objets, voire 6

O. Grabar, “Imperial and Urban Art in Islam: the Subject-Matter of Fatimid Art,” Colloque International sur l’Histoire du Caire (Cairo, 1972).

qu’est-ce que l’art fatimide?

285

chaque objet individuel comme des unités discrètes et accidentellement préservées provenant d’un passé lointain et d’en sortir d’une part une interprétation centripète, axée sur l’objet même, aussi exacte que possible et, de l’autre, les éléments d’un vocabulaire en cours à l’époque. II est difficile de critiquer cette manière de faire les choses, même si les résultats sont parfois longs à se faire connaître et s’il faut beaucoup de ce genre d’exercices pour en tirer des conclusions historiques et culturelles valables. L’autre manière est de proposer une série d’hypothèses qui pourraient aider à créer un cadre pour les explications à venir d’objets individuels, quitte à abandonner les hypothèses lorsqu’elles ont perdu leur valeur et à les remplacer par de nouvelles. Deux hypothèses, en particulier, me semblent ressortir des observations et remarques qui précèdent. L’une est que les arts que nous associons aux Fatimides et à leur époque se comprennent au mieux lorsqu’on identifie les classes (dans un sens différent de celui du marxisme classique) ou les niveaux sociaux pour lesquels ils furent faits. Il y a ainsi le niveau des princes, souvent plus proches les uns des autres sans distinction [17] de frontière ou religion que de leurs compatriotes ou de ceux dont ils partagent la foi; l’art de la peinture, par exemple, était réservé, sinon exclusif, à ce niveau-là ainsi que la fabrication d’objets en métaux et autres matériaux précieux et de certains tissus. Il est vraisemblable que les grands vizirs ou militaires appartenaient à cette classe de consommateurs d’art. Puis il y a, surtout à partir de la fin du 11ème siècle, la classe ou le niveau des notabilités, un terme plus générique que celui de bourgeois, par trop européen, et qui correspond peut-être aux ashraf dont parlent les textes, quoique le terme de patriciens ait aussi été utilisé, des villes à qui je continue à vouloir attribuer la céramique d’art que nous connaissons; les distinctions locales y sont plus grandes que chez les princes, les thèmes juifs ou chrétiens y sont possibles, parce qu’il s’agit d’un monde attiré par la qualité de l’artisanat plutôt que par celle du matériel utilisé; ou, peut-être, pour poursuivre une hypothèse déjà ancienne, c’est une classe qui utilise des substituts parce qu’elle ne peut pas se permettre les matières premières chères. Il s’agirait aussi d’une classe relativement nouvelle dans l’ordre du mécénat et ce serait le monde culturel des Fatimides qui aurait été le premier à promouvoir et à favoriser ce monde, je ne dirais pas, de parvenus, mais certainement de nouveaux riches. Un troisième niveau serait celui de la foi; il s’agirait moins d’individus que de fonctions associées à l’expression de la foi et auxquelles participaient des gens d’origines sociales diverses; les musulmans dominent dans ce groupe, mais il est juste d’y inclure juifs et chrétiens, chacun servant ses propres besoins. Il y aurait enfin un niveau populaire dont on peut vraisemblablement reconstituer les lieux d’habitation, certains tissus et des céramiques de types courants. J’ajoute que les Dhimmis devraient peut-être être considérés séparément, quoique, pour cette époque, je préfère, faute d’arguments contraires, les inclure dans des divisions valables pour tous.

286

early islamic art, 650–1100

Cette classification n’est qu’une hypothèse préliminaire pour comprendre la variété de goûts et de besoins qui expliquerait l’existence d’une créativité difficile à comprendre par hypothèse d’une quelconque communauté de formes acquises à tous et partagées par tous. Ce qui expliquerait les paradoxes des arts de l’époque fatimide serait la conjoncture de deux ordres, un ordre de fonctions diverses (du palais ou de la mosquée, aux habitations ou au bazar) et puis un ordre de goût et presque de snobisme (de l’universalisme géographique et historique des-princes dont l’un acheta des selles ayant appartenu à Alexandre le Grand aux habitudes visuelles de chaque lieu). Les deux ordres ont trouvé leur langage dans le vaste vocabulaire de l’époque ‘abbaside, dans le souvenir des Umayyades d’Espagne et de Syrie, et, en fait, dans l’énorme richesse de cette antiquité tardive qui est tellement à la mode dans notre monde de savants. Le paradoxe le plus étonnant des Fatimides dans ce sens est le contraste entre deux caractéristiques. On en aperçoit une dans la Jérusalem qu’ils parachèvent: ils reconstituent un espace idéal antique et umayyade, tout comme ils retrouvent un peu d’illusionnisme dans la représentation, un réalisme de formes aussi bien que d’intention noté jadis par Ettinghausen.7 De l’autre côté il y a l’éclatement des prescriptions antérieures et les premiers pas vers cet art dont la richesse et la variété embraseront le monde musulman à partir du milieu du 12ème siècle. Seule la première de ces caractéristiques peut vraiment être attribuée aux califes fatimides eux-mêmes et à leur entourage qui chercha à légitimer leur présence et à imposer leur droit à une tradition impériale et islamique à la fois. Mais c’est par accident plutôt que par dessein qu’une ambition idéologique trouva des formes antiquisantes par le truchement du rappel du passé omeyyade. En fait, ce n’est, du moins dans notre entendement actuel, que dans certaines inscriptions monumentales – à la mosquée d’al-Hakim au Caire, à celle d’Aqmar ou sur le minbar d’Ashkalon, maintenant à Hébron – que le monde idéologique ou la pratique de leur foi aient été rendues visibles. Il est bon de reprendre l’hypothèse que le maintien de l’écriture angulaire pour les beaux Corans fatimides était une manière de se rattacher au passé plutôt que d’adopter l’écriture nouvelle légiférée de Baghdad. [18] Ce n’est pas la dynastie elle-même qui fut responsable des énormes changements économiques, sociaux et politiques qui, à partir du milieu du 11ème siècle, modifient toute la carte de l’histoire de trois continents et surtout de la Méditerranée. D’autres forces entrèrent en jeu dont les Fatimides tirèrent profit. Et, si la majorité de ce que l’on connaît de l’art fatimide provient d’Égypte, peut-on dire que les styles et motifs égyptiens aient été imposés ailleurs? Ainsi, pour répondre aux questions posées au début de cet exposé, l’art sous les Fatimides installés au Caire ne se distingue ni par son caractère 7

R. Ettinghausen, “Early Realism in Islamic Art,” Studi Orientalistici in Onore di Giorgio Levi della Vida (Rome, 1956).

qu’est-ce que l’art fatimide? 287

sectaire ni par ses éléments régionaux, ni même par ses innovations lorsqu’on le place dans son contexte synchronique, mais par une énergie et une richesse que l’on retrouve en même temps dans le monde romain occidental, la renaissance macédonienne et comnène, l’Andalousie des Umayyades et des reyes y taifas. Dans cette perspective, l’art des Fatimides est un des volets d’une époque brillante d’un monde méditerranéen bien plus ouvert et tolérant, du moins dans les formes, qu’il ne le deviendra deux siècles plus tard. Et, dans un autre sens, c’est un volet particulièrement original d’un monde musulman en train d’échapper à l’emprise culturelle et religieuse de Baghdad grâce à l’établissement de centres artistiques nouveaux qui, comme Cordoue, Le Caire, Isfahan, Nishapur ou Balkh, donneront des directions plus variées aux arts des hommes et femmes de l’Islam.

Part Four The Muslim East

Chapter XVIII Sarvistan: a Note on Sasanian Palaces*

Even if we do not necessarily follow Sauvaget’s caustic and negative remarks about Sasanian art,1 it must be admitted that our knowledge of Iranian art in the centuries which preceded the Muslim conquest is still very fragmentary. Only one excavation dealing exclusively with Sasanian remains has been published so far, and even in this instance not entirely.2 Outside of Bishapur, we have to rely on descriptions by travelers and explorers of varying degrees of reliability and chance discoveries have been the major sources from which scholars have had to draw conclusions about what can be assumed a priori to have been one of the most important formative elements of medieval Near Eastern art. The recent revolution in our awareness of Central Asian art before the Muslim conquest has made the poverty of our understanding of Sasanian art all the more unfortunate. It is not the purpose of this paper to solve the problem, since only properly equipped and serious archaeological expeditions can do that. My aim is rather to illustrate the practical and methodological complexity of the problem by discussing one specific instance of a building which has been assumed to be one of the most important examples of Sasanian architecture. The instance is that of Sarvistan.3 There is general agreement on the facts that it is a Sasanian palace of the fifth century ad (perhaps even more specifically from the time of Bahram Gur, 420–38) and that its construction and decoration are quite remarkable. There is no argument about this last point, for the building is comparatively well preserved and does exhibit some [2] extraordinary uses of piers supporting vaults, of façade designs, and of stucco coverings. It represents, without doubt, an important step in the

* First published in Forschungen zur Kunst Asiens (Istanbul, 1970), pp. 1–8. 1 J. Sauvaget, “Remarques sur l’art sassanide,” Revue des Etudes Islamiques, 12 (1938); cf. also his notes in “Remarques sur les monuments omeyyades,” Journal Asiatique, 232 (1940–41). 2 R. Ghirshman, “Les Fouilles de Bishapur,” Revue des Arts Asiatiques (1936–42). 3 R. Ghirshman, Parthians and Sasanians (Paris, 1962), p. 181; L. Vanden Berghe, Archéologie de l’Iran Ancien (Leiden, 1959), p. 47, with bibliography; A. Stein, “An Archaeological Tour in the Ancient Persis,” Iraq, 3 (1936). The most complete study of the building is still by O. Reuther in A. U. Pope, A Survey of Persian Art (Oxford, 1939).

291

292

1

early islamic art, 650–1100

Sarvistan: plan

development of Iranian architecture, even though the exact position of that step is more difficult to establish because of the uncertainty of our knowledge of other buildings. On the other hand, the questions of the date and of the purpose of Sarvistan do not appear to be as clearly assured as has been assumed so far. Let us consider first the possible function of the building. Sarvistan (Fig. 1) is a small building, 36 by 42 meters. Its main façade opens to the west and consists of three sets of steps separated by a wall with engaged columns (Fig. 2); the central steps lead into a wide but shallow eyvan, the southern steps into a similar, but smaller, eyvan, the northern ones into a domed room. The central eyvan leads into a large domed hall – almost 13 meters to the side – which is clearly the main feature of Sarvistan, since it occupies nearly a third of the central part of the building. Behind the dome there appears a court followed by a small eyvan flanked by two side-rooms. On the north side the central dome opens into an eyvan leading to the outside; this eyvan communicates also with a long hall which has doors leading outside, into the courtyard, and into a small, almost square, room in the northeast corner; the latter, in turn, is provided with doors into the court and into the siderooms of the eastern eyvan. The southern side of Sarvistan [3] consists of a long and narrow vaulted hall followed by a domed square area; it is provided with three outer doors, one of which is on the axis of the main cupola, and with two passageways into the court.

sarvistan: a note on sasanian palaces 293

2 Sarvistan: elevation

Tedious though it may have been, this description allowed us to focus attention on a number of peculiarities in the plan of Sarvistan: its openness to the outside with four monumental gates and four minor ones, its complex system of internal communications arranged in such a fashion that from any one defined area one can penetrate into any one of the adjoining ones (the one exception is the southern side-room of the interior eyvan), and the central position of a huge dome which serves as the turning plate of the building to the point where the unity of the long southern hall has been broken up for a door centered on the domed room. The axial position of the dome does not appear only in the fact that openings permitted one to move to and from the building in every direction. The position of the passageways also suggests the possibility of movement around the domed hall, a sort of circumambulation of the central feature of the building. The most remarkable feature of Sarvistan’s plan lies, however, in the contrast between apparently clear functions of parts – central, axial dome and possibility of circulation around the dome – and the asymmetry of the forms used. No axis divides the internal arrangement of the building into symmetrical parts, which is all the more peculiar since the façade is quite symmetrically composed. At the same time it can be suggested that almost every one of the definable units of Sarvistan had in fact a precise purpose. This conclusion may be a bit conjectural [4] in the light of the uncertainties of our understanding of Sasanian architectural forms. And yet it is a peculiarity of Sarvistan that almost every one of its halls and rooms is architecturally differentiated from

294

early islamic art, 650–1100

all the others; there are no two similar units in the whole building. A priori and as long as we do not know more about the development of Sasanian architecture, we have to assume that some precise purpose was attached to each unit. This is also the only explanation for the unusual contrast between the symmetrical façade and the asymmetrical internal arrangement.4 Thus we may define Sarvistan as a small building, asymmetrically composed of several clearly defined units among which a large dome appears particularly important, with numerous ways in and out of the building, and with a complex system of communications inside the building. Now we may turn to other monuments and see whether they can suggest explanations for the actual purpose of Sarvistan. The building has been called a palace, but it differs from known palaces in several significant ways. First, its small size is unusual; in fact it could almost be fitted within the great eyvan of Ctesiphon, whose dimensions are 43 by 26 meters. The comparison with Ctesiphon is perhaps unfair, but it remains true that Firuzabad is 55 by 103 meters, Qasr-i Shirin over 370 meters long, the excavated part of the so-called Palace II at Kish 45 by 50 meters and the single great hall of Bishapur with its ambulatory occupies an area of some 50 by 50 meters.5 Hence, if it is to be related to palace architecture, Sarvistan must, on account of size alone, be explained in a manner which would differentiate it from other palaces.6 Yet it is not only size that is involved. The very facts of the numerous entrances and of the easy communications between parts inside the building sharply distinguish Sarvistan from those palaces – like Firuzabad and Qasr-i Shirin – whose plans are known and in which axial symmetry, a single entrance, and a clear compartmentalization of parts give a totally different impression. Furthermore, the very logic of a palace architecture with its presumed clear separation [5] between public and private parts makes it unlikely that Sarvistan belongs to the main typological series of Sasanian palaces.7 This point can be extended into several details. It has usually been agreed that the central feature of a Sasanian palace is the combination of an eyvan and of a dome. Some uncertainty exists as to the exact purpose of each of these units, but the existence in Ctesiphon of an eyvan without a dome on 4

5

6

7

An alternate explanation would involve the hypothesis of a series of major reconstructions. As long as the building is not excavated, this possibility must remain open, but, to my knowledge, the existing descriptions have not made it very likely. All these buildings are discussed in the Survey or in the works of Ghirshman and Vanden Berghe quoted in note 3. See also K. Erdmann, Die Kunst Irans zur Zeit der Sasaniden (Leipzig, 1943), esp. pp. 19 ff. This point has been recognized by E. Herzfeld in F. Sarre and E. Herzfeld, Archäologische Reise im Tigris – und Euphrates – Gebiet (Berlin, 1911–20), vol. II, pp. 332–3, but his further reasoning is not totally acceptable. This point is, I feel, quite important, and the easy identification of purposes which has been made for each room of Sarvistan by R. Ghirshman (Parthians and Sasanians, p. 281) or by Sir Aurel Stein (Iraq, 3, pp. 178 ff.) seems for the time being quite unwarranted.

sarvistan: a note on sasanian palaces 295

the same scale and various literary and archaeological documents8 do suggest that the main unit used for audiences was primarily the eyvan and not the dome. Yet the striking features of Sarvistan are the size of the dome and the shrinking of the eyvan. Another detail derives from the point made earlier that almost every one of the halls in Sarvistan has clearly differentiated architectural characteristics. None of the rooms has features which could be explained as part of a residential complex and it is difficult to imagine that the small rooms in the back of the court (3 meters by 4 meters) could be so interpreted. On the contrary, every unit of Sarvistan seems to have had some precise ceremonial or symbolic meaning. One way of solving this problem, while maintaining the aulic character of the building, would be to suppose that it belonged to a type of royal building which was not primarily residential. The pavilion in a garden could be such a building. The assumption of gardens around Sarvistan has been made by several writers,9 although it is by no means proved archaeologically; and the existence of pavilions is ascertained in medieval Islam and fairly likely in Sasanian times.10 We are [6] unfortunately totally uninformed about the likely size of such pavilions, but the openness of Sarvistan and the central position of the domed room make the interpretation possible. The rooms in the back could be considered as service-rooms, while the two long halls would fulfill secondary ceremonial functions, such as banqueting. The whole building would then be a prime example of what may be called an architecture of pastime and a major prototype, in function if not in form, of later Islamic monuments. There are, however, uncertainties and difficulties about this interpretation. The main uncertainty resides in the fact that no clear palace parallel can be brought to strengthen an explanation of Sarvistan as a garden pavilion. Another uncertainty lies in the problem of explaining the differentiation of 8

9 10

It is not possible as yet to write a history of Sasanian ceremonies, but a number of arguments do exist to suggest that the prince, when appearing officially, was sitting in the eyvan. The assumption of the domed room as the place for the enthroned king is, first of all, difficult to visualize architecturally, for these rooms have usually very thick walls with four small openings and, therefore, the king could hardly be visible to more than a minute fraction of the audience. Furthermore, the ceremony of the lifting of the curtain in front of the king does not make sense if there are four doors to the room in which he sits. Finally, it does seem from literary sources that the announcement of the appearance of the king was made to throngs gathered outside, in the open air. See Mas‘udi, Prairies d’Or, vol. VII, pp. 264–5; Ferdosi, Shah-nameh, tr. J. Mohl (Paris, 1876–8), vol. VI, pp. 290, 139–40, etc. Admittedly these texts are late and their usefulness limited, but they do fit with evidence derived from an observation of the architecture. It would follow that the dome fulfilled either some other royal purpose such as that of private audiences or some non-royal purpose; cf. below. Herzfeld, Reise, II, pp. 332–3; Erdmann, Kunst, pp. 31–2. The only clear archaeological evidence is that of the celebrated plate in the Berlin Museum (Survey, IV, pl. 237), for whose interpretation we accept Pope’s hypothesis as in “A Sasanian garden palace,” The Art Bulletin, 15 (1933), although other explanations also exist: Survey, pp. 554–7.

296

early islamic art, 650–1100

parts and the asymmetry of the building, for these features suggest a more complex purpose to Sarvistan than that of a simple pleasure pavilion. These uncertainties are compounded when we consider the further point that the characteristics of Sarvistan can, in part at least, be related to a totally different and comparatively well-documented series of buildings, the firetemples. The best-known type of Sasanian fire-temples is the Chehar-taq, the simple domical structure open on four sides.11 Its central feature is a cupola, just as in Sarvistan.12 What is more significant, however, is that simple domes were not the only type of Sasanian sanctuary. Already in Qasr-i Shirin such a dome is set inside a large (80 by 120 meters) enclosure which contains courts, apartments and halls, and which is provided, insofar as limited surface surveys allow us to judge, with a complex system of passageways.13 A more remarkable religious building was discovered by L. Vanden Berghe at Kunar Siah; its fire-temple is surrounded in very irregular fashion by a whole series of courts and halls and the asymmetry of the composition is striking.14 An equally asymmetrical composition involving many different parts was excavated at Takht-i Sulayman and its component units, like those of Sarvistan, have very elaborate architectural characteristics.15 To be sure, in all these examples the main domical sanctuary is provided with an ambulatory, but we have seen that the possibility of walking around the domed room is [7] clearly provided at Sarvistan. Furthermore, a recent study of legal Pahlavi texts pertaining to sanctuaries has clearly shown that there were several different kinds of temples and that different functions and purposes were attributed to each kind.16 Altogether, then, basing ourselves on the certain variety of religious buildings existing in Iran and on the peculiarities of the functions suggested by the architecture of Sarvistan, we can suggest that it was some sort of sanctuary, for which side-rooms and halls were used for precise ceremonies, whose characteristics are still quite uncertain.17 And the very peculiarity of the composition as well as its asymmetry could be explained by the fact that much in the religious development of Iran under the Sasanians involved new practices and new forms. Several years ago Monneret de Villard had already brought attention to a Syriac text which relates how, when no one 11 12 13 14 15 16 17

Basic bibliography in K. Erdmann, Das Iranische Feuerheiligtum (Leipzig, 1941); A. Godard, “Les monuments du feu,” Athar-é Iran, vol. III (1938); and below. The matter still needs investigation, but one may wonder whether the domed room behind the characteristic eyvan of so many palaces is not a small sanctuary. Survey, I, p. 553. L. Vanden Berghe, “Découvertes de Monuments Sassanides dans le Fars,” Iranica Antiqua, 1 (1961), pp. 175 ff. Preliminary report in Archäologische Anzeiger (1963), p. 637. J. P. de Menasce, Feux et fondations pieuses dans le Droit Sassanide (Paris, 1964). Most recent study on these problems in M. Molé, Culte, Mythe et Cosmologie dans l’Iran Ancien (Paris, 1963).

sarvistan: a note on sasanian palaces 297

had a clear model for the building of a church, palace features were used.18 Sarvistan could thus be considered as an instance in which a courtly architecture was transformed for religious purposes. Between the two possible interpretations of Sarvistan a specific choice is difficult to make. While we would tend to favor the second explanation over the first, the main point of this discussion lies elsewhere. On the one hand, it suggests that no final conclusion can be reached on the purpose of Sasanian buildings without full excavations. On the other hand, we hope to have made the methodological point that a building must be clearly understood in all of its component parts, and each part, as well as the whole, must be related to appropriate typological series before one can possibly try to explain its function and its purpose. A similar kind of problem is posed by the date to be given to Sarvistan. The attempt by Herzfeld to relate Sarvistan to a very specific group of constructions erected under Bahram Gur and described by Tabari19 is based only on the fortuitous coincidence of the meaning of the word “Sarvistan,” “gardens of cypresses,” and the mention of such gardens in the text. There is little doubt, of course, that the technique of construction of Sarvistan is more developed than that of the better-dated (c. AD 230) Firuzabad and of Bishapur (late third century), but beyond that we have almost no data [8] to judge the possible development of Sasanian architecture and the rhythm of its growth. Even the apparent lack of major stucco decoration, which would seem to separate Sarvistan from known monuments of the sixth and seventh centuries, is not entirely valid as an argument if we are right in thinking of it as a religious structure rather than as a palace. Altogether, then, this note concerning a moment of Near Eastern art which had been of so much concern to the late Kurt Erdmann would have fulfilled its purpose not so much if its conclusions are correct as if it challenges scholars to look anew at the still badly known and badly understood monuments of Sasanian Iran.

Postscript Since the writing of this article L. Bier has concluded a thorough survey of the site and come to some conclusions comparable to mine but to a different and much more likely date. L. Bier, Sarvistan, A study in Early Iranian Architecture (University Park, 1986).

18 19

U. Monneret de Villard, La Chiese di Mesopotamia (Rome, 1940), p. 28. Tabari, Geschichte der Perser und Araber, tr. Th. Nöldeke (Leiden, 1879), pp. 111–12.

Chapter XIX Notes on the Decorative Composition of a Bowl from Northeastern Iran*

Much has been written about the art of medieval Near Eastern ceramics, and a number of conclusions may be taken as reflecting a consensus of scholarly opinions and as corresponding to much of the objectively known evidence.1 At the same time, troublesome questions constantly arise as one attempts to understand any one apparent conclusion in depth or as one investigates what may be called the epistemological borders of an accepted statement, the fascinating gray area in which a generalization is true and yet not entirely satisfactory, either because it does not seem quite to account for all available information or because it raises too many additional problems. For instance, it seems fairly clear that Iraq, Egypt and northeastern Iran were the three major areas in which a new art of luxury ceramics developed in early Islamic times, and the time of the formation of this new art may have been as early as the last decades of the eighth century. While a certain number of common features and reciprocal influences can be shown to have existed in the decorative themes, the techniques and the functions of pottery in all three areas, the differences are on the whole far more striking, and, with the exception of one or two rather uncertain groups, the major types belonging to each region are comparatively well known to the student of Islamic art, even though there is no published systematic description of each series. What remains unclear, however, are the reasons why these particular provinces witnessed a major development in the art of ceramics while others did not. Is it simply a matter of insufficient exploration in other areas? Or are there some more fundamental regional differences in the artistic growth * First published in Islamic Art in the Metropolitan Museum of Art (1971), pp. 91–8. 1 This state of affairs is symbolized by the fact that for pottery we do possess several introductory studies, of which the most up-to-date and useful one is A. Lane’s Early Islamic Pottery (London, 1948). Like most similar studies it deals with a pottery that has primarily an aesthetic merit, i.e., with luxury or semi-luxury objects. The same apparent clarity is far from obvious when one turns to an archaeologically defined pottery, i.e., when the sum total of a moment’s ceramic production is considered. See, for instance, J. Sauvaget, “Introduction à l’étude de la céramique musulmane,” in Revue des Etudes Musulmanes, 33 (1965).

299

300 early islamic art, 650–1100

of various Islamic provinces? Furthermore, if the development of an art of ceramics in eastern Iran, Iraq and Egypt can be shown to be related, why are the ceramic series different from each other? If, on the other hand, there is no significant relationship between the pottery of the three areas, why is it that at approximately the same time the same technique exhibited such remarkable changes? Another kind of problem occurs when one attempts to date ceramic series with any degree of precision. Almost all museum labels and publications tend to put objects over a period of two centuries, e.g., ninth–tenth or twelfth–thirteenth centuries. While archaeological expeditions may provide occasional post quem or ante quem dates for certain groups, the very nature of medieval sites, even when they are untouched by contemporary robbers and excavated with appropriate care, usually precludes more precise dates. Even when a precise date can be given to some group of objects (as has been the case in a number of Fustat pits),2 archaeological evidence cannot limit the group to this date only. Thus for any one type we are compelled to accept, at least a priori, a comparatively long period of manufacture and utilization. But this very hypothesis of longevity makes both the formation of the type, and especially its decadence, abandonment, and replacement by another type, a particularly important problem for the solution of which appropriate mechanisms have not yet been developed. The difficulty is compounded when one considers the origin of most of our information. It consists on the one hand of masses of sherds, each of which provides only partial evidence about the finished object but can usually be fitted in coherent chronological or technical series, and, on the other hand, of more or less complete objects, often of uncertain origin. Although attempted in a few cases,3 the systematic correlation of these two sources of information is still in its infancy. Finally, a third kind of problem can be brought out. There is general agreement on the main techniques of Islamic pottery as well as on the curious fact that the luxury wares of the early centuries tended to be a production of bowls and of plates, i.e., of open shapes which emphasize a single circular surface to be decorated. It has also been generally recognized that interest in surface decoration [92] over generally indifferently shaped bodies is a major characteristic of Islamic ceramics. But comparatively little attention has been given to the themes of decoration themselves. Or rather, even if a number of individual topics have been studied, as in the instances of a celebrated plate in the Freer Gallery4 or of a specific type of

2 3

4

G. T. Scanlon, “Ancillary Dating Materials from Fustat,” in Ars Orientalis, 7 (1968). P. J. Riis and V. Poulsen, Hama, Les Verreries et Poteries Médiévales (Copenhagen, 1957); or, in a less systematic way, C. Wilkinson, “The Glazed Pottery of Nishapur and Samarkand,” in The Metropolitan Museum of Art Bulletin (November 1961). R. Ettinghausen, “The Iconography of a Kashan Luster Plate,” Ars Orientalis, 4 (1961).

decorative composition of a bowl from iran

301

Kufic inscriptions,5 there is nothing even approximating a corpus of ceramic designs; nor has there been any attempt at suggesting the formal processes by which artisans and painters organized the circular or other shapes they used. Yet, if we are to consider and understand some if not all the ceramic series of the Muslim Middle Ages as works of art to which the culture gave more than a utilitarian function, we should be able to present some hypotheses about the ways in which their decoration was organized, and about the aesthetic or iconographic meanings to be attributed to such themes as can be identified. It would be easy and in the context of this essay quite idle to multiply examples of apparently accepted conclusions whose implications or premises have not yet been worked out. And it is only proper that it be so, for any conclusion automatically becomes a hypothesis for further research and, tenuous though it may be at any one time, a conclusion must also consist of some sort of equilibrium between a concrete object and general hypotheses. It is with the aims of suggesting a few directions in the search for this balance and of proposing some hypotheses concerning the problem of defining ceramic decoration that this essay was undertaken. It will attempt, first of all, to describe and to explain the decorative composition on one recent acquisition of the Islamic department at the Metropolitan Museum. Then it will seek to widen the conclusions one may reach about this one object into broader considerations of the aesthetic principles involved in the art of Iranian ceramics in early Islamic times. The object involved (Fig. 1)6 is a deep bowl, 12 cm in height, 35 cm in diameter at the top, with a base diameter of 14.5 cm. The shape of the object is quite common; its size is somewhat larger than is usual, although parallels for such larger bowls can be found in eastern Iran, in Iraq, and in Egypt. The object is complete for the most part, and it is only in such places where the fragments were put together that there are traces of modern additions, but these do not affect the character of the decoration in any significant way. The technique of the object – design in brownish pigment painted over a yellow slip and under transparent glaze – is fairly typical of ceramic techniques of northeastern Iran. The technique and especially its color scheme suggests that the luster ceramics of Iraq were being imitated. Although it is not possible for the time being to prove that this particular object was actually made in Nishapur,7 such an origin is likely enough, as we shall see later on. But its point of origin is not important for our purposes, and it is simpler to consider it as one of the many ceramic types developed in northeastern Iran 5 6 7

L. Volov, “Plaited Kufic on Samanid Epigraphic Pottery,” Ars Orientalis, 6 (1966). It was first published by E. Grube, “The Art of Islamic Pottery,” The Metropolitan Museum of Art Bulletin (February 1965), p. 213, figs 5, 6. For the distinctions between Nishapur and Afrasiyab, see Wilkinson’s article and K. Erdmann’s earlier “Ceramiche di Afrasiab,” Faenza, 21 (1937), largely superseded now by Sh. Tashhodjaev, Hudozhistvennaia Polivnaia Keramika Samarkanda (Tashkent, 1967).

302

1 and 2 Bowl, painted and glazed earthenware. Persian, from Nishapur, ninth– tenth century

early islamic art, 650–1100

decorative composition of a bowl from iran 303

between the ninth and the eleventh centuries, more probably toward the end of that period. Like practically all these ceramics, the Metropolitan plate is decorated. On the outside (Fig. 2) groups of three vertical strokes divide the side of the bowl into five compartments, each compartment being then occupied by a roughly elliptical shape filled in desultory fashion with oblique strokes suggesting a sort of hatching. Although not the only type of design apparent on the sides of ceramic objects from early Islamic times in any one of the three major centers of production, it is a very common one. Its purpose is essentially that of emphasizing the shape of the object, but the more curious feature of this design – here and on many other objects – is its apparent carelessness, as though the decorator felt freer with his brush-strokes on the outer sides of his objects than on the inside. Whether this freedom of execution should be understood as an indication of the lesser importance medieval ceramicists gave to the exterior of their objects, or whether these marks in which one can observe rather curious variations in the number of strokes found in ellipses or circles should be interpreted as atelier marks with some internal significance, cannot be answered for the time being, and arguments exist for either interpretation. In any event the most important design on the object occurs inside the bowl, as is typical of almost all objects of that kind and time. If one [94] excepts a narrow border consisting of a much-simplified festoon typical of a large number of plates and bowls, this design consists of a unified composition set against a neutral background and framed in a single continuous circular edge. Seen as a whole, the bowl’s composition is not unique, but it is clearly distinguishable from several different compositions known in early Islamic pottery. It is quite different from such extraordinary designs as are found on a celebrated bowl in the Freer Gallery (Fig. 3), in which the background of the border becomes the design of the center.8 It is also different from designs which forcefully separate a wide ring-like border from a central medallion, as occur on many objects from northeastern Iran and from Egypt, for instance on a well-known bowl in the Metropolitan Museum (Fig. 4).9 In most of these objects the central medallion is composed around one or more axes, while the side border has a generally clockwise movement, especially when it involves writing. The composition of our object is also different from a third major type of composition found on early Islamic ceramics from all major areas, the single subject – an animal, a human figure, a scene, writing – set on a variety of backgrounds, from simply monochrome ones to a cluttering of themes from many origins as in a celebrated Nishapur group. 8

9

The object has been published by R. Ettinghausen, Medieval Near Eastern Ceramics in the Freer Gallery (Washington, 1960), figs 6, 7. A similar effect occurs also on Egyptian ceramics, Lane, Early Islamic Pottery, pl. 24 and, in a different way, pl. 25B. Lane, Early Islamic Pottery, pl. 15B; for Egyptian examples see pls 22B, 23A, etc.

304

early islamic art, 650–1100

3 Deep bowl, painted and glazed earthenware. Persian, from Nishapur, tenth century

Finally, it is different from a somewhat smaller number of designs which consist of an all-over pattern of a single motif repeated in more or less organized fashion. Yet, as I shall try to demonstrate, almost every one of these ways of composing on the circular shape of a bowl has affected the design of our object. In order to make this particular point clear we must first identify the individual elements which went into the making of the composition. First we have lines. These are of two kinds, a thick line that is always drawn directly on the neutral background and a thinner line generally used as a border or as an edge for some other unit. Second, we meet with a roughly circular element with a number of dots set on or just outside of the circle’s line and with a V-shaped, chevron-like unit in the middle. While one may imagine that the latter is a simplification of a natural subject like birds in flight, it is perhaps more accurate to interpret the whole motif as the breakdown into smaller units of a vegetal rinceau. This seems all the more likely since in three or four places on the bowl itself the circular units still seem to grow from each other, and also since there are a few fragments excavated in Nishapur that illustrate a number of intermediary steps from a real rinceau with buds and flowers or leaves to our schematic representation. The third element found on the Metropolitan plate has usually been called a

decorative composition of a bowl from iran 305

4 Bowl, painted and glazed earthenware. Persian, from Nishapur, tenth century

“peacock’s eye” because it does occur occasionally on wings of birds.10 Whatever the origins of the motif, it has become a vaguely circular form with a single darker dot somewhere inside the circle (but never in its center); on the Metropolitan object the peacock’s eye is always set in a field of stippling. Fourth, we encounter six silhouettes of birds. However simplified the silhouettes may be, they are clearly recognizable as birds and, while their outlines are very much alike, it should be noted that the two birds which are on either side of the main axis of the bowl have only four projecting prongs symbolizing both wings and tail, while the other four have five. Fifth, there are six units consisting of a stem crossed by two hatches and two or three small nodes or knobs; this stem divides itself [95] into three parts, and each part ends with a sharply pointed chevron made of unusually thick lines; there are some variants in the number of hatches and nodes. While here 10

For instance M. Pézard, La Céramique archaïque de l’Islam (Paris, 1920), p. cxix.

306

early islamic art, 650–1100

again we are dealing with a much-schematized and simplified motif, there is little doubt that it has a vegetal origin and that it derives from a rather common theme of a stem ending with three leaves, buds, or flowers. Finally, the sixth element of our bowl’s decoration consists of several letters in the center of the object; these letters are arranged in one combination repeated twice and in a small unit which could be understood as a ta’ marbutah; according to what has become a sort of conventional explanation of these letters, we have here either a repetition of a shorthand version of barakah or a simplification of barakah li-sahibihi.11 None of these six “phonetic” elements found on the Metropolitan Museum bowl is original, and numerous parallels can be found for any one of them on a large number of objects from northeastern Iran and from Iraq; some of them are even found in Egypt. It would be interesting to trace the growth and development of each of these features and to investigate whether they are of local origins or brought in from the West together with the luster technique which was imitated. But this is not our concern here and the only point of significance is that the specific combination of these themes which occurs here is not very common. It is known mostly on a small group of sherds and on one complete object found in or near Nishapur, of which, to my knowledge, only two have been published so far.12 Although it may be possible to reconstruct the exact compositional pattern found on the objects from which fragments only remain, the newly acquired Metropolitan bowl is only one of two in which this pattern is preserved in its entirety, and it is far more elaborate than the one published in 1937. This is especially so when one considers that element which gives to the object its greatest originality, the composition of its decoration. The center of the composition consists of a medallion in the middle of which a single band contains writing going from right to left. Above and below the inscription are found seven of our circular units with chevrons. The latter are so arranged that the ones above the writing point to the right whereas the lower ones point to the left. The circle is enclosed in an octagon and the area between the two geometric forms is filled with a single row of peacock’s eyes which acts as a sort of ring moving around the circle. The central medallion is thus an essentially geometrically conceived unit made of a circle and of a polygon; several movements or directions are given to this geometric unit, a circular one and two horizontal ones, from left to right and from right to left. Among the latter, however, the primary one is from right to left because it occurs twice and because it is the movement of the writing. From each of the hexagon’s six angles two thick lines are generated. One line leads all the way to the outer circle of the design; the other one stops 11 12

S. Flury in A. U. Pope, ed., A Survey of Persian Art (London, 1939), pp. 1748 ff. C. Wilkinson, “The Iranian Expedition, 1936,” The Metropolitan Museum of Art Bulletin (October 1937), fig. 18 and p. 18; see also idem (November 1961), fig. 16.

decorative composition of a bowl from iran 307

about two-thirds of the way, then turns back toward the first line so that, together with several thick lines coming from the outer circle and from the first line, it forms a rough and elongated diamond shape containing one of our tripartite stems. This diamond shape is itself surrounded by two areas filled with peacock’s eyes. What we have in this fashion is, repeated six times and apparently generated by the central hexagon, a vaguely geometric entity – a sort of panel – utilizing [96] one major new theme – the tripartite stem – and a fill already used in the medallion. This entity is given a counterclockwise direction by the thick points of the stem, but a clockwise direction is suggested by the position of the dots inside each peacock’s eye which tend to be at the same end of the eye. While the evidence for a clockwise movement by the peacock’s eyes is not systematic enough to be totally acceptable by itself, the intention for such a direction is confirmed by the chevrons which fill the spaces between our diamond-shaped entities and by the birds which are found in small triangular spaces at the edge of the design. Both birds and chevrons definitely indicate a clockwise movement for the whole design. At the same time, as was already pointed out, the two birds which are on the axis of the design (the axis being defined by the position and direction of the writing in the central medallion) differ from the rest of the birds. It is even possible, although less likely, that the smaller number of hatches and nodes found on the central tripartite stem is also to be related to an emphasis on the axis of the design as it is suggested by the central medallion. We can thus interpret the composition of our object as consisting of a central unit which generates a counter-clockwise whirl-like movement based on six vector-like – i.e., provided with a direction – units; these, however, contain in themselves or are set against an opposite, secondary clockwise movement. At the same time, to the predominant circular pattern there corresponds a horizontal axis emphasized by the writing and by minor modifications in the detail of both the clockwise and the counter-clockwise directed elements. Not one of the various aspects of this compositional pattern is unique in the bowl in the Metropolitan Museum, and one can easily find parallels to a relationship between central medallions with clear axes or with circular movement and clockwise (or even more rarely counterclockwise) borders (Figs 3 and 4), or to axially composed objects, or else to whirls generated from a center.13 But few if any examples show all these compositional devices at the same time and the other almost complete bowl done in the same technique has a far less sophisticated composition. Thus, 13

For the whirl and its possible symbolic meaning see R. Ettinghausen, “The Wade Cup,” Ars Orientalis, 2 (1957), pp. 341–56; examples of designs can be found in most of the numerous picture books dealing with Iranian ceramics. A convenient summary of drawings occurs in Tashhodjaev’s book; it is limited to material from Samarkand, but it does seem that Samarkand ceramicists were particularly fascinated with compositional problems of this order.

308

early islamic art, 650–1100

imperfect though it may be in many details – and I shall return to this point presently – the newly acquired Metropolitan bowl may serve as a particularly brilliant and complex example of the fascination exercised on early Islamic pottery painters, especially in northeastern Iran, by the problem of composing a decorative design inside a circle. Considering the nature of these objects, this concern is natural enough and it would be of considerable interest to pursue for other series in Iran and elsewhere the analyses of decorative compositions made so far only for Samarkand ceramics. Such a study would acquire particular significance if it could be shown that certain compositional devices tended to predominate in certain techniques or that decorative principles were only secondarily affected by techniques. We might then be able to determine far better than we can at the moment the sources of the taste which produced these ceramics. But two further conclusions are suggested by our brief analysis. The first one is of a rather abstract nature. Except for the silhouettes of the birds and, to a much smaller degree, the writing, none of the elements of decoration we have defined is by itself responsible for the composition it creates. Or, to put it another way, none of them contains in itself the form it gives to the design, in opposition, for instance, to concrete subjects like houses, personages, or whole scenes, which, because of the precise meanings presumably associated with them, tend to lead to simpler compositions.14 It is thus only when they are put in a certain relationship to each other that the various elements of our bowl acquire compositional sense. The patterns and purposes of the relationship between features, meaningless in themselves, are what determines the structure of the design or what we may call its grammar, that is the way in which an understandable sense can be given to an amorphous and finite number of individual signs. The great originality of our object lies in part in the fact that its grammar is of particular complexity and subtlety, while its “phonetic” elements are not only very few and very simple but also very roughly executed and at times downright careless. It is impossible, on the basis of one object alone, to say whether this carelessness was a wilful one serving to emphasize the “grammatical” structure and possibly the decorator’s poetic license in using the themes he had chosen or whether one must imagine behind our object a complex model comparable in both grammatical and phonetic subtlety to the Freer Gallery bowl [97] (Fig. 3). Until such time as more numerous specific studies of individual objects or of series of objects have been made, one may perhaps conclude more modestly that the dematerialization and the deconcretization of individual elements

14

This point is particularly apparent as one looks at Fatimid ceramics that developed a complex iconography or at the twelfth- and thirteenth-century pottery from Iran. Exceptions exist, of course, and it may be significant to note that an unclear presentation of concrete subjects occurs particularly frequently in Iraqi and northeastern Iranian pottery from the first period.

decorative composition of a bowl from iran 309

which characterizes the Metropolitan bowl should merely be used to illustrate a very characteristic aspect of early Islamic decorative art: the predominance of abstract axes and movements over specific motifs. The second conclusion concerns the nature of these motifs. What is striking, beyond their dematerialization, is what may be called their humility. These are not, even in their origin, heroic themes or noble floral compositions with demonstrable aesthetic or iconographic purposes; they are rather the small details of decoration, lines or fills and backgrounds found on other objects. Even letters and birds with which clear and precise meanings can be associated appear in a sort of casual and almost unreadable way. The transformation of these details into the main subject-matter of a design was perhaps one of the most original features of much of early Islamic art, especially in Iran. It may be compared to the way in which the architects of the Samanid mausoleum in Bukhara transformed the lowly brick into the main vehicle of the monument’s decorative design. There again, as we are only beginning to understand, very abstract geometric patterns of growth from a small number of basic principles were involved.15 We may then perhaps suggest that in much of early Islamic art – in Iran and elsewhere, with obvious differentiation due to a variety of factors – at least two formal languages were being created. One, concrete and precise like the writing of the most celebrated series of northeastern Iranian ceramics, tended to stress specific subjects, however complex may have been the compositions used for them. The other, looser and less particular in either details or in the choice of subjects, preferred to provide general impressions and general effects, usually of an abstract character, rather than to stress individual details. The Metropolitan bowl would be an illustration of this second tendency. It is obvious enough that much research must still be devoted to the vast number of objects available from the early medieval period in Iran before any of these conclusions, except perhaps the first one, can be anything but working hypotheses. Furthermore, the full understanding of the bowl itself requires a number of additional investigations that were beyond the purpose of this essay, whose objective was merely to draw attention to one of the several kinds of analyses with which we must approach the rich field of Islamic pottery. One last conclusion may, however, escape this hypothetical state. It is that the newly acquired bowl in the Metropolitan Museum is not only useful in defining the variety of ways in which early Islamic ceramicists ordered designs on the circular shape of their creations, but may occupy a significant place in our eventual understanding of how Islamic art was formed and what the modalities and rhythms of its growth have been.

15

The only study devoted to this problem is that of L. Rempel and G. A. Pugachenkova, Arhitekturnyi Ornament Uzbekistana (Tashkent, 1961).

Chapter XX A Tenth-Century Source for Architecture* O. Grabar and Renata Holod

Historians of Islamic art have usually been hampered in their research by the absence of easy availability of contemporary literary sources for monuments. Our understanding of the actual use, appreciation and evaluation of ancient artefacts or works of art still relies largely on today’s assumptions and constructs. Works dealing exclusively with lives of artists and the evaluation of their works are very rare. Dust Muhammad’s few pages on Persian painters have been used to establish the chronology of Persian paintings, while his critical comments form the beginning of a critical (and technical) vocabulary.1 Sadiqi’s memoirs illuminate his own drawings and miniatures and give us a glimpse of a complex artistic personality.2 Yet these two works, with their total preoccupation with criticism of painting and drawing, may be exclusively the products of their time, the sixteenth century, or of the tradition of royal libraries, particularly cultivated and encouraged in Iran, Central Asia and India in the fifteenth to the seventeenth century. They can also be understood as an extension of the traditional concern of poets and literati for the arts of the book. While mention of a treatise or manual on architecture did occur among the lost volumes of Rashid al-Din’s Universal History, the earliest extant treatise is a seventeenth-century Ottoman work, reflecting the unusually important position occupied by architects in the Ottoman court and, perhaps, their specific practices.3 Technical treatises, while not providing contemporary or critical information, are nevertheless crucial in providing the specialized contemporary knowledge and vocabulary. Abu’ al-Qasim’s treatise on ceramics [311] has been of unique * Firts published in Harvard Ukrainian Studies, Vol. III/VI (1979–80), pp. 310–19. 1 For Dust Muhammad’s treatise, see L. Binyon, J. V. S. Wilkinson and B. Gray, Persian Miniature Painting (1933; reprinted New York, 1979). Another example is Qadi Ahmad, whose Calligraphers and Painters was translated by V. Minorsky (Washington, DC, 1959). 2 For a discussion of Sadiqi’s memoirs, see A. Welch, Artists for the Shah (New Haven, 1976), pp. 41 ff. 3 The manual on architecture was written by Safar Efendi for the Ottoman chief architect Muhammad Agha. See R. Lewcock, “Material and Techniques,” in Architecture of the Islamic World, ed. by G. Michell (New York, 1978), p. 133.

311

312 early islamic art, 650–1100

importance in all discussions of Persian pottery of the thirteenth century.4 The eleventh-century treatise of Ibn Badis gives a detailed description of bookmaking with a precise terminology for the preparation of pens, inks, colors, papermaking and bookbinding.5 Other technical treatises have been noted in manuscript versions; only a minuscule number has been published.6 Thus the art historian must turn to other types of works, in which the description or evaluation of monuments or artefacts may be part of a historical narrative, a legal argument, a geographical description, a poetic metaphor, or a treatise on mathematics. Maqrizi’s Khitat Misr and Qadi al-Rashid’s Kitab al-Dhakha’ir wa al-Tuhaf provide a hitherto insufficiently explored but quite extraordinary dimension to the study of Cairo and the Fatimid court, describing in extensive technical detail the construction of the monuments of Cairo as well as the holdings of the Fatimid treasuries.7 Hisba manuals, regulating the production of various crafts, can contain valuable technical information. For example, the twelfth-century manual of Ibn ‘Abdun mentions sizes of construction materials, bricks, tiles, joists and beams prescribed for construction in Seville.8 Al-Kashi’s fifteenth-century treatise on arithmetic provides instructions for the measurement and computation of surfaces of arches, domes and stalactites, thereby revealing the sophisticated theoretical background on which Timurid architecture, with its precise modular system and its series of structural innovation, could have drawn.9 [312] Finally, the ethnographic record can be used for information, particularly on processes of manufacture. In his description of the building industry in Tunis, Revault assumes a basic continuity of construction techniques and nomenclature, and extrapolates from the present to describe the materials and construction of seventeenth-, eighteenth-, and nineteenth4 5

6

7 8

9

For the most recent translation and discussion of Abu’ al-Qasim’s treatise, see J. W. Allan, “Abu’ al-Qasim’s Treatise on Ceramics,” Iran, 11 (1973), pp. 111–20. Al-Mu‘izz Ibn Badis, “‘Umdat al-Kuttab wa ’uddat dhawi al-albab.” For translation, glossary and technical commentary, see M. Levey, Mediaeval Arabic Bookmaking and its Relation to Early Chemistry and Pharmacology, Transactions of the American Philosophical Society, n.s., vol. 52, pt 4 (Philadelphia, 1962). Two handbooks on carpentry are mentioned by Lewcock, “Material and Techniques,” p. 133. The tenth-century treatise of Abu’ al-Wafa’ al-Buzjani on what a craftsman should know of measuring has been translated by A. S. Krasnova, “Geometricheskie preobrazovaniia,” Akademia nauk SSSR/Institut istorii estestvoznania … v stranakh Vostoka, vol. 1 (1966). Al-Maqrizi, Khitat Misr (Bulaq, 1854); Qadi al-Rashid, Kitab al-Dhakha’ir, ed. M. Hamidullah (Kuwait, 1959). E. Lévi-Provençal, Séville Musulmane au début du XII siècle: Le traité d’Ibn Abdun sur la vie urbaine et les corps de métiers (Paris, 1947), p. 74. For a glossary of terms, see E. LéviProvençal, “Un document sur la vie urbaine et les corps de métiers à Séville au début du XII siècle: le traité d’Ibn Abdun,” Journal Asiatique, 224 (1934), pp. 177–299. Jamshid Ghiyath al-Din al-Kashi, Miftah al Hisab wa al-Risalat al-Muhitiyah, translated into Russian by B. A. Rosenfeld, with commentary by Rosenfeld and A. P. Iushkevych (Moscow, 1956), pp. 161–79.

a tenth-century source for architecture 313

century mansions.10 Similarly, contemporary Yemeni or Persian building practices and materials have been given a historical currency reaching far into the past, because they are considered traditional utilizations of regionally available materials and skills.11 While such argumentation may indeed be true, effort must continually be taken to check such information against the historical and archaeological record. The following pages deal with one text, found in a particularly unexpected place, which provides a description of building techniques in tenth-century Khorasan or Transoxiana, areas which have so often been discussed by Professor Pritsak. The Kitab al-Ba‘d wa al-Ta‘rikh (Book of Creation and of History) was composed in 355/965–6 by Mutahhar b. Tahir al-Maqdisi, who was from the city of Bust, now in southern Afghanistan.12 It is one of several “histories” – the most celebrated being Mas‘udi’s Muruj al-Dhahab13 – which sought, among other things, to provide a Muslim statement of the nature of the world and of history culminating in the Prophet’s Revelation.14 One of its chapters deals with the traditional argument for the existence of God; that is, that no creation is possible without a creator. In it al-Maqdisi compares the creation of the universe with the construction of a building, the making of a ship, and the weaving of a cloth. None of these existing “things” are possible without a maker, and thus the existence of the universe presupposes the existence of God. The argument itself is hardly an original one, as it is merely a modification and concretization of the broader theological notion of a Prime [313] Mover. To limit ourselves to the Muslim world, al-Ash‘ari, in his Kitab al-Luma‘, uses the examples of cotton being spun and woven and of a mansion (qasr) having been erected in a wasteland to demonstrate the same point.15 Further searches would no doubt uncover any number of other instances in which theologians, preachers, or just ordinary litterateurs utilized what certainly became a literary topos. What distinguishes al-Maqdisi’s version is the unusual precision of his description of the building and weaving processes. Since this precision is of little significance to the theological argument and otherwise 10 11 12

13 14 15

J. Revault, L’Habitation tunisoise: Pierre, marbre et fer construction et décor (Paris, 1978). Lewcock, “Material and Techniques,” p. 139; H. Wulff, The Traditional Crafts of Persia (Cambridge, Mass., 1966), p. 110 ff. Le livre de la création et de l’histoire, ed. and trans. by C. Huart, 6 vols (Paris, 1903); undated (?) Beirut facsimile of Huart’s edition. Little is known about our author; Huart was, in fact, at some pains to identify him (Journal Asiatique, 9th ser., 18 [1901], pp. 16– 21; Geschichte der arabischen Literatur, supp. 1, p. 222). See also F. Sezgin, Geschichte des arabischen Schrifttums, vol. 1 (Leiden, 1967), p. 337. Ed. and trans. by Barbier de Meynard and Pavet de Courteille (Paris, 1861–77). T. Khalidi, Islamic Historiography (Albany, NY, 1975), esp. chaps 2 and 3, with frequent references to our author. Al-Ash‘ari, Kitab al-Luma‘, trans. by R. J. McCarthy as The Theology of al-Ash‘ari (Beirut, 1953), p. 6 of text, p. 7 of translation.

314

early islamic art, 650–1100

unique,16 it is perhaps reasonable to suggest that it may reflect a somewhat more detailed practical knowledge of the two activities than would usually be expected. Too little is known about the author to use this information to imagine or reconstruct his life or social setting. But for architecture or weaving this text is a rare document on both the division of labor prevalent in Central Asia or northeastern Iran in Samanid times and the terminology for the various activities involved. This short note is limited to the section dealing with buildings. We will retranslate the text and then provide a few comparative observations drawn from the archaeological and ethnographic record. The text itself reads as follows (vol. 1, pp. 68–9; translation, pp. 61–3): If it were permissible to imagine the creation (huduth) of this world without a creator (muhdith), it would in fact be possible to imagine the existence of building (bina’) without a builder (bani), of a piece of writing (kitabah) without a writer (katib), of a design (naqsh) without a designer (naqqash),17 of an image (surah) without a painter (musawwir). It would, in fact, be permissible to the one who sees a solid residence (qasr)18 and a firm building (bind) to believe the following: (1) a pile (kumah) of earth (turab) was gathered together without a [314] gatherer (jami‘); it was then mixed (akhtalata) without a mixer until cohesive (alfaltaffa) and moist (nada); (2) then it was molded (ansabaka) into a brick (libn) of perfect proportion (taqdir) and admirable squareness (tarbi‘) without someone to plan it in advance (sabiq) and to fashion it (darib);19 (3) then the foundations (asas) of the residence (qasr) laid themselves out, its footings (qawa’id) strengthened themselves (tamakkana),20 its pillars (saqat) and transoms (a‘raq) rose up,21 so that its walls (haytam) could be extended (tawala) and its corners (arkan)22 completed; and mud

16

17

18 19 20 21

22

The point of uniqueness in so much detail is advanced with some reservation, since neither one of us is familiar with theological literature. Enquiries from colleagues more learned in these matters failed to elicit immediate parallels in Muslim or Christian literature, although the idea itself permeates such passages as 1 Corinthians 2:10 and ff., as has kindly been pointed out to us by Professor Leo Hines of Fitchburg State College, Massachusetts. There are several different ways of understanding the word naqsh, including “design, engraving, decoration.” In the absence of historical dictionaries, we chose the least specific meaning. The translation of qasr as “residence or mansion” seems more appropriate in this context than the usual “castle.” The more general bind was rendered simply as “building.” The terms sabiq and darib refer to pouring the mud brick mixture into the wood frame of the mold and beating it into its corners. See the commentary which follows. The exact activity involved here is unclear; most likely it still refers to the extension of the foundation into the first course of socle. See the commentary which follows. These two words are difficult to interpret. Saqat may be taken to mean tree trunk, whereas a‘raq (sing. ‘araqah) means “transom between two courses of brick stone.” In this context, the author is referring to a process of half-timber construction where the wood frame is raised before laying in brick. See the commentary which follows. Arkan normally means supports or pilasters. One can understand it simply in this fashion or, because a wood frame has already been mentioned by the author, a less common meaning, “corners,” may be intended to describe some fashion of reinforcing corners. Such a technique is documented in stone and rubble masonry, although it is

a tenth-century source for architecture

315

bricks (libn) flew into the air, landed on their (proper) sides (tarakamat‘ala hawdshiha), and arranged themselves in the most beautiful order; (4) then joists (judhu‘) and beams (jawa’iz) fell on their own according to the measurements of apartments (buyut)23 and of sectors (khitat) and were cut for building without anyone gathering them (trees) and cutting them; (5) then [wood] was hewn (antajara) without a hewer, sawed (antashara) without a sawyer, smoothed (asfana) without a plane; when [these wood pieces] are completed, the uneven parts straightened out, they rise on their own grooves (mughariz), transform themselves into ceilings over rooms (buyut), and their pillars (asatin) rise under them; (6) then a sheathing (safa’ih) covers them (ceilings), doors open and close on their own; (7) then the building is covered with lime (takallasa) and mud (tasayya),24 paved (taballata) and plastered (tajassasa); it is decorated (naqqasha) with different kinds of ornaments (tazawiq) and designs (nuqush). And so the work is finished, the building completed, its separate parts united in the best fashion and the most perfect arrangement. Not one of its partitions, bricks, or wooden beams appear without the viewer’s admiration for its wisdom and its purpose, all of this without the maker (fa‘il) who made it, the fashioner (sani‘) who fashioned it, the expert (sa‘i) who formed it, the planner (mudabbir) who planned it!

[315] Al-Maqdisi has identified seven steps in the process of construction clearly separated by the conjunction “then.” (1) The preparation of mud bricks: Before mud bricks can be formed, the earth not only needs to be thoroughly moistened and mixed, but an addition of straw and/or chaff is necessary to give it resiliency and to prevent it from crumbling. As observed in recent practice in Iran, the three components – mud, straw and water – are mixed with a hoe or by treading.25 Al-Maqdisi’s omission of straw in his description may indicate his unfamiliarity with the preparation process or, as is more likely, his assumption that this addition was common knowledge. His term jami‘ would then refer to the person treading the mud mixture. (2) Molding the mud brick: The shaping of mud brick involved using a wooden mold frame, packing the mud mixture into it and beating it into the corners of the mold.26 Al-Maqdisi’s terms refer to the specialized workmen who manufacture mud brick.

23

24

25 26

not commonly known in mud brick, half-timber construction. See the commentary which follows. The word bayt is another term with a long and complicated history. Since we are dealing with a single building, we prefer to interpret it in its early meaning of a discrete living unit within a larger complex; it could be as small as a single room (see below) or as large as a whole complex of rooms. See O. Grabar, R. Holod, J. Knudstad and W. Trousdale, City in the Desert (Cambridge, Mass., 1978), especially pp. 79–80. Although the process of covering the building with plaster is here described as having two parts – i.e., a covering with lime and then with clay – the author is actually referring to the mud plaster mixture found on most buildings studied. Wulff, Traditional Crafts, p. 109. The most complete discussion of the process is given in ibid.

316

early islamic art, 650–1100

(3) The laying of foundation and building of walls: The author does not specify the type of foundation that was laid. The simplest foundations could have been quite shallow stone rubble packed into a trench wider than the intended width of the walls, with a mud and lime plaster poured over the rubble course.27 However, since the author also mentions another term, qawa‘id, “footings”, or, even more likely, “socle”, the building in question could have been larger, requiring somewhat more extensive foundations. Since the author was presumably describing a manner of building in Khorasan and perhaps Transoxiana, it is interesting to compare his statements with the foundation techniques known through archaeological or architectural studies. A common way of preparing the area on which a building would stand was to level it and make a platform of hardened earth by tamping it (pisé) or by flooding it several times. Such a platform was then edged with rough cut stone. Footings for walls which would bear additional weight were laid in rough cut stone with rubble brick and mud lime mortar core. A variety of footings and foundations could be found within one monument.28 The footings for the walls [316] were built above the ground to about fifty centimeters as a socle. It is on this socle that the wood framing elements would be erected. Studies on Iranian and Central Asian architecture have been concerned largely with brick (actually mud or baked) building and vaulting techniques and have neglected the important role played by wood. Not only was it used for scaffolding and tie-beams in great monuments,29 but its use in residential architecture appears to have been extensive. What al-Maqdisi is describing is a half-timber construction where the walls, with their openings for doors (and windows), are framed out in wood. However, he appears not to be describing a half-timber construction where there are any diagonal crossbracing members or where the vertical members are too numerous to allow for horizontal brick coursing.30 At Lashkari Bazar, mud brick as well as pisé was laid in courses with tie-beams inserted at infrequent intervals. The author of our text appears to be referring to an intermediate technique which used wood framing rather extensively, but still laid brick in courses rather than wedging them between the cross-bracing. Further proof of 27 28

29 30

Cf. ibid., p. 111. Cf. N. B. Nemtseva, “Medrese Tamgach Bogra-Khana v Samarkande,” Afrasiyab, 3 (Tashkent, 1974), pp. 11–113; V. A. Nilsen, Arkhitektura Srednei Azii V–VII vv. (Tashkent, 1966), p. 217; N. M. Bachinskii, Antiseismika v arkhitekturnvkh pamiatnikakh Srednei Azii (Moscow and Leningrad, 1949), pp. 16–20; D. Schlumberger et al., Lashkari Bazar: L’Architecture, Mémoires de la délégation archéologique française en Afghanistan, vol. 18 (Paris, 1978), p. 15. Cf. Schlumberger, Lashkari Bazar, p. 16; Nemtseva,”Medrese Tamgach Bogra-Khana,” p. 113; Bachinskii, Antiseismika, p. 25. For examples of such techniques, see Bachinskii, Antiseismika, p. 30; T. V. Rapoport, “O progressivnykh traditsiiakh v arkhitekture narodnogo zhilishcha Uzbekistana,” Arkhitekturnoe nasledstvo, 13 (1961), p. 209.

a tenth-century source for architecture 317

horizontal coursing may be the reference to the corners (arkan) of the walls, which seem to have been interlaid in bond. (4) Preparation of parts of the superstructure: That wood superstructures were utilized throughout history in seemingly treeless areas is confirmed by archaeological and ethnographic evidence. Apparently, wood was available wherever there was irrigated land. At Dalvarzin Tepe, a Kushan site in southern Uzbekistan, numerous charred fragments of wood beams were discovered.31 The famous seventh- to eighth-century site of Penjikent used wood in its superstructures, supports and decoration.32 Although Lashkari Bazar is better known for its rich collection of vaults, some traces of wood superstructures, destroyed by the conflagration in the mid-twelfth century, have been recorded.33 There are a few [317] monuments, such as the masjid-i jami‘ of Khiva, which still retain some of their early (tenth- to twelfth-century) elements.34 According to the ethnographic record, wood joists and beams are probably the most ubiquitous superstructure in domestic architecture throughout Iran and Central Asia.35 (5) The assembling of the ceiling: Although al-Maqdisi seems to be describing domestic architecture, the edifice is not a simple peasant house. Because it has a ceiling, it must be a mansion or urban residence. The wooden pieces mentioned by the author are battens which were grooved to receive panels of wood mosaic.36 They have also been recorded in traditional domestic architecture of Tajikistan, the Ferghana valley, and the Hindu Kush. What is even more important for our purposes is the discovery of charred remnants of a paneled mosaic ceiling in the reception hall of a residence at Dalvarzin Tepe.37 Such a find indicates the continuity of a building tradition and allows us to utilize comparative examples from the ethnographic record with more certitude. The rooms described in our text must have had spans larger than the available beams. Their ceilings were supported by wooden pillars or columns. For certain areas of Central Asia, a typology and chronology of wood elements has been developed.38 In Iran, such material still needs to be

31 32 33 34 35

36 37 38

G. Pugachenkova, Dalvarzin Tepe (Tashkent, 1978), p. 196. V. L. Voronina, “Arkhitekturnyi ornament drevnogo Piandzhikenta,” in Skulptura i zhivopis drevnogo Piandzhikenta (Moscow, 1959), pp. 107–8. Schlumberger, Lashkari Bazar, pp. 16, 98. V. L. Voronina, “Kolonny sobornoi mecheti v Khive,” Arkhitekturnoe nasledstvo, 11 (1958). Bachinskii, Antiseismika, p. 32; Rapoport, “O progressivnykh traditsiiakh,” p. 215; V. L. Voronina, Narodnaia arkhitektura severnogo Tadzhikistana (Moscow, 1959); Wulff, Traditional Crafts, pp. 110–14. Wulff, Traditional Crafts, p. 87, and Rapoport, “O progressivnykh traditsiiakh,” p. 215, describe the construction of traditional ceilings in detail. Pugachenkova, Dalvarzin Tepe, p. 195 and fig. 129. Cf. Voronina’s works on Ferghana and Penjikent, and her “Order v narodnoi arkhitekture Tadzhikistana,” Arkhitekturnoe nasledstvo, 19 (1972), pp. 183–91.

318

early islamic art, 650–1100

studied, although the restoration studies of the Ali Kapu and the Chihil Sutun palaces in Isfahan now provide important information on the role of wood in Safavid architecture.39 Mention should also be made of the relatively early (eleventh- to twelfth-century) wood structures in Abyaneh, north of Isfahan, and in Turan Pusht, west of Yazd. Little information is available on wood structural members in the region of Bust, although remnants of carved and painted wood consoles found at Lashkari Bazar might indicate a larger role for wood than is usually supposed. (6) Roofing and setting of doors: The sheathing mentioned by the [318] author implies a flat roof, although it is not described in detail. Most likely, the covering consisted of reed matting or ceiling boards.40 It was subsequently covered by several layers of a mud, lime and straw mixture. The thickness of the roof varied from region to region. The doors are set into their frames, already built during the construction of the wood skeleton, through prepared socket holes.41 (7) Finishes: Although the author has indicated two separate processes for finishing the walls of the building (takallsa wa tasayya), they were probably rendered in one step with a mixture of mud, straw, and lime. Such a procedure is well attested in archaeological and ethnographic sources.42 The term taballata can refer to any kind of paving. The author may have had a specific material in mind, such as the variety of baked paving bricks that were recovered from Lashkari Bazar.43 The finish coat of gypsum plaster would have been applied only in sheltered areas, while the decoration mentioned by the author could have consisted of carved plaster and paint similar to that found at Lashkari Bazar. The type of building described belongs to domestic rather than monumental architecture, in spite of its size and decoration. Examples of domestic architecture or contemporary descriptions of houses are very rare for this time. This text, therefore, is an important addition to the historical record. Yet several difficulties are apparent. Although the author is supposed to be from Bust, he may not have had a local building in mind. The information on construction techniques recently made available for Lashkari Bazar indicates a more extensive use of pisé for walls, intermixed with courses of mud brick, and much less wood than indicated by the author. Most striking

39 40 41 42

43

C. Zander, Travaux de restauration des monuments historiques en Iran (Rome, 1968). These are mentioned by Wulff, Traditional Crafts, p. 111, and Rapoport, “O progressivnykh traditsiiakh,” p. 209. Wulff, Traditional Crafts, p. 86. Cf. Schlumberger, Lashkari Bazar, p. 15; P. I. Kostrov, “Tekhnika zhivopisi i konservatsiia rospisi drevnogo Piandzhikenta,” in Zhivopis’ drevnogo Piandzhikenta (Moscow, 1954), pp. 161–2; Zander, Travaux de restauration, p. 334; Rapoport, “O progressivnykh traditsiiakh,” p. 218; Wulff, Traditional Crafts, p. 112. Schlumberger, Lashkari Bazar, p. 17.

a tenth-century source for architecture 319

at Lashkari Bazar and at Bust, however, is the variety of vaults, domes and arches found not only at the royal palace, but also in the mansions (residences) which lay along the river.44 Since al-Maqdisi took pains to note every part of the construction process, his omission of vaults may be significant. Either he was describing construction techniques outside the immediate area of Bust (thereby [319] infirming somewhat the importance, tenuous as it is, of his life there) or vaulting was restricted to particularly expensive and monumental buildings. Without more texts and more monuments, we cannot decide which is the case. A provisional solution would be to consider that the text isolates patrician housing, to use Professor Bulliet’s terminology,45 which stood somewhere between princely and peasant buildings. While the description of the building process is quite precise and in large part corresponds to archaeological and ethnographic data, the author was not necessarily a builder himself. His observations are rather those of someone whose memory or experience of a house (his house?) being built was triggered by a routine theological argument. The fact that the text is in Arabic removes us from the specialized vocabulary probably used by the building trades in Khorasan and Transoxiana. By the same token, however, it suggests the existence of a terminology with wider currency than in a single region. But historical dialectology and the formation of an Arabic koiné are topics which have rarely been investigated.46 If our comments on this apparently unusual passage have a special point, it is to draw attention not only to the need for wider social and lexicographic studies, but also to the fact that more evidence for such studies, even highly specialized ones, exists than has been imagined.

44 45 46

Ibid., pp. 16, 90–94. R. Bulliet, The Patricians of Nishapur (Cambridge, Mass., 1972). Compare, for instance, our text with Ibn Khaldun’s Muqqadimah, ch. 5, nos 24 and 25; F. Rosenthal, trans., vol. 2 (New York, 1958), pp. 357 and ff.; A. Wafi, ed., vol. 3 (Cairo, 1960), pp. 932 and ff. Ibn Khaldun is much more academic in his discussion, but his general vocabulary is very close to Maqdisi’s. On the other hand, his technical vocabulary reflects another time and another place.

Index

This is an index of proper names only. English standard spelling was preferred to transliterations or to French spelling. The article “al” is not considered for alphabetization. ‘Abbasids 4, 22, 36, 46, 49, 63, 64, 66, 67, 101, 131, 132, 134, 135, 137, 139–41, 143–6, 148, 156, 157, 215, 219, 220, 235–9, 247, 250, 254 ‘Abd al-Malik 20, 23, 48, 64, 72, 96, 142, 144, 145, 147, 178, 185 ‘Abd al-Mu’min 102 ‘Abd al-Rahman II 261 ‘Abd al-Rahman III 271 Abou Gosh 124 Abraha 62 Abraham 61, 88, 89, 93 Abu al-Fida 128 Abu al-Qasim 311 Abu ‘Ubaydah 3, 6, 24 Abu Yazid ai-Bistami 78 Abyaneh 318 Achaemenids 176, 226 Affoula 124 Afrasiyab 249 Aghtamar 47 Agra Taj Mahal 69, 71, 72, 91, 98, 267 Aleppo 80, 95, 98, 99, 135 Firdaws madrasah 98 Alexander the Great 237 Alexandria 3 Alfonso VII 272 Algiers 76 Almohads 102, 257 ‘Amman 131, 136, 151, 188 al-Andalus 264, 267, 270, 272, 274 ‘Anjarr 134, 175, 188 Antioch 135, 166, 179

‘Aqabah 10, 133 ‘Arabah, wadi 133 Artzruni 47 Ascalon 126 al-Ash’ari 313 Assyrians 176 Aswan 244 ‘Atlit 124 Ayyubids 137, 160 Azar 93, 233 Azhar 244, 268 al-’Aziz 237 al-Azraq 133 al-Azraqi 57, 59, 60, 61, 62, 63, 64, 65, 67 Baghdad 36, 46, 66, 67, 132, 156, 220, 244, 245, 270 Bahram Gur 148, 291, 297 al-Baladhuri 142, 148 Balikh 138 Bamiyan 43 al-Basasiri 244 Basra 175 Bethlehem 45 Church of the Nativity 44 Beth-Shan 124 Bishapur 291, 294, 297 Bistam 78 Bryas 46 Bukhara 270, 309 Kalayan minaret 52, 53 Samanid mausoleum 309 Buran 236

321

322

index

Burgos 271, 272 Las Huelgas 272 Bust 318, 319 Buyids 141, 246, 254 Byzantium 4, 5, 11, 16, 18, 20, 23–7, 28, 36, 38, 41, 43, 45, 48, 55, 135, 142, 146, 179, 216, 236, 247, 250, 254 Cairo 19, 49, 50, 53, 73, 76, 95, 101, 128, 215, 217, 218, 226, 227, 241, 243, 244, 250, 254, 262, 268, 270, 272, 312 ‘Amr mosque 268 al-Aqmar mosque 76, 98, 244, 245 Ibn Tulun mosque 262, 268, 270 Hakim’s minaret 80 Muhammad Ali mosque 76 Qala’un’s mausoleum 75 Qarafa mosque 244 Salih Tala’ii mosque 245 Sultan Hasan madrasah 84 Carmathians 244 Carolingians 179 China 69, 173, 180 Constantin Porphyrogènete 142 Constantine 36 Constantinople 36, 46, 179, 237, 245, 263 Grand Palais 142, 166 Cordoba 5, 76, 79, 102, 250, 257, 259, 262, 263, 264, 265, 269, 270, 272 Great Mosque 257, 258, 259, 268, 269, 272 maqsurah 74 Ctesiphon 179, 294 Dalvarzin Tepe 317 Damascus 7, 19, 22, 25, 26, 27, 40, 45, 128, 131, 133, 135, 138, 142, 154, 175, 180, 194, 257, 263, 264 Great Mosque 7, 8, 9, 10, 26, 44, 48, 79, 91, 178 (Grande Mosquée), 186, 193 Hospital of Nur al-Din 36, 39 John the Baptist church 7 Mausoleum of Baybars 22 Darb Zubayda 67 David 91

Delhi Qutb-minar 80, 81 Dioscorides 5, 30, 34 Dura-Europos 13 Emmaus 124, 125 Ephesus 90 Ethiopia 95 Euphrates 10, 133, 138, 139, 175 Eutychius 3 al-Farabi 84 Fatehpur Sikri 267 Fatimids 5, 19, 36, 44, 49, 58, 79, 142, 144, 146, 160, 215–27, 229, 231–8, 240, 241, 243–54, 262, 312 Ferghana 317 Fez 73 Firuzabad 294, 297 Fustat 218, 222, 300, 303, 308 Gagik 47 Galen (pseudo-) 33 Galilee, sea of 131 Geniza 239, 243 Ghassanids 4 Ghawr 135 al-Ghazali 54 Ghaznavids 43, 98 Ghazni 244 Granada 272 Alhambra 38, 76, 83, 84, 85, 271, 272, 273 Hagar 61 al-Hajjaj 23, 143, 144, 147 al-Hakam 261, 263, 264 al-Hakim 229, 244, 246 Hama 8, 124, 125, 126, 175 al-Hariri 5 Harran 34 Harun al-Rashid 67, 236 Hawran 10, 133, 135, 136 Hedwig glass 252 Heraclius 3, 6 Herat 270 Hijaz 136, 264

index

Hindukush 317 Hira 125 Hisham 132, 134, 138, 143, 147 Hisn Maslamah 138 Humaymah 134, 136, 157 Ibn ‘Abbas 63 Ibn ‘Abdun 312 Ibn al-Arabi 58 Ibn Aziz 53, 231 Ibn Badis 312 Ibn Battutah 58, 76, 128 Ibn al-Bawwab 101 Ibn Idhari 263 Ibn Jubayr 58, 76 Ibn Khaldun 58 Ibn Khurdadhbih 138 Ibn Miskawayh 52, 54 Ibn Muqlah 101 Ibn al-Nadim 100 Ibn Tulun 262, 268 Ibn Zamraq 84 Ibn Zubayr 64, 66 Idrisi 263 Ifriqiyah 177, 215 Ilkhanids 101 India 43, 48, 49, 69, 76, 95, 96, 98, 173, 252, 267, 311 Isfahan 69, 76, 246, 272, 318 Ali Qapu 318 Chihil Sutun 74, 318 Great Mosque 69, 75, 98 Hasht Bihisht 74 Jurjir façade 246 Ishmael 89 Islamabad King Faisal Mosque 100 Istanbul 76, 272 Hagia Sophia 74 Suleymaniye 77 Topkapi Saray 34, 272 Jabal Says 12, 156, 175, 188 al-Jahiz 142 Jam 80, 98 Java 86 al-Jazari 5, 239

323

Jazirah 138, 139, 154 Jeddah 182 Jericho 131 Jerusalem 8, 22, 25, 27, 58, 67, 81, 89, 95, 125, 128, 157, 175, 188, 237, 246, 247, 262, 268 Aqsa Mosque 4, 8, 44, 45, 74, 218, 231, 246, 268 Dome of the Rock 4, 8, 20, 22, 25, 44, 45, 48, 67, 69, 70, 71, 72, 75, 97, 156, 175, 179, 180, 193, 262 Franciscan tower 82 Haram al-Sharif 8, 81, 246 Holy Sepulchre 75, 81, 237, 246 Jewish Temple 71, 88, 89 Lutheran tower 82 Russian tower 82 John of Damascus 45 John the Baptist 91 Justinian II 23, 44 Kairouan 257, 260, 262 al-Kashi 84, 312 Khabur 138, 139 Khalid al-Qasri 145 Khilwah 133 Khirbat al-Mafjar 12, 13, 15, 18, 19, 25, 28, 29, 31, 49, 50, 74, 128, 131, 134, 141, 142, 144, 149, 154, 156, 175, 179, 188, 191, 192, 210 Khirbat al-Minyah 12, 13, 14, 26, 107, 108–16, 118–21, 123, 125, 126, 127, 128, 129, 131, 132, 134, 154, 156, 175, 188 Khiva 317 Khorasan 133, 147, 177, 110, 244, 316, 319 Khorezm 132 Khosro 40, 179, 184 al-Khuza‘i 60 Kish 294 Konya 76 Küçükyalı 46 Kufah 139, 156, 157, 177, 188 Kunar Siah 296 Kushan 317

324

index

Lakhmids 4 Lashkari Bazar 98, 249, 316, 317, 318, 319 Ma’bad 146 Madinah 7, 27, 76, 89, 95, 97, 144, 247, 262 Mosque of the Prophet 268 Madinah al-Zahra 249, 270 al-Mahdi 63, 65, 66, 67, 157 Mahmud of Ghazna 43 Ma’in 44 Mamluks 6, 22, 101, 116, 118, 124, 128, 129, 254, 272 al-Ma’mun 61, 65, 67, 236 al-Mansur 63, 66, 67, 228 al-Maqdisi 59, 313, 315, 316, 317, 319 al-Maqqari 263 al-Maqrizi 53, 142, 218, 231, 237, 244, 245, 254, 257, 312 Maslamah b. ‘Abd al-Malik 138 al-Mas’udi 265 Mecca (Bakka) 57, 58, 59, 60, 61, 62, 63, 65, 66, 67, 89, 90, 95, 96, 97, 157, 247, 262, 264 Black Stone 64 Ka‘ba 57, 60, 61, 64, 65, 66, 67, 89, 90, 95, 180, 247, 265 Maqam Ibrahim 60, 61 Masjid al-Haram 57, 59, 60, 63, 89 Zemzem well 60, 63, 66 Monophysites 45 Morocco 6, 19, 76, 245, 272 Mshatta 12, 13, 15, 16, 55, 131, 134, 136, 141, 151, 152, 153, 154, 155, 156, 157, 175, 188, 189, 191, 210 Mu‘awiyah 40, 145 Mughals 69, 91, 98, 272 al-Mukhtar 144 al-Muktafi 247 al-Mustansir 233, 237 Mutahbar b. Tahir al-Maqdisi 313 al-Mutawakkil 157 al-Muwaqqar 157, 188 Nadir Shah 43 Nasir-i Khusraw 218, 233, 236, 239, 246, 251

Nazareth 128 Negeb 133 Nicaea 45 Nishapur 50, 53, 76, 249, 270, 300, 301, 303, 304, 305, 306 Ottomans 49, 69, 76, 102, 129, 272, 311 Palermo 50, 248, 2271 Capella Palatina 49, 52, 218, 234, 248 Palmyra 10, 131, 133, 135 Pamplona 271 Penjikent 317 Piazza Armerina 16, 142, 166 Pisa griffin 233, 238 Pliny 137 Plotinus 54 Potifar 89 Ptolemies 216 Qadi al-Rashid 312 Qal’a of the Beni Hammad 125 Qarmatians 101 Qasr el-Abyad 127 Qasr al-Hallabat 137 Qasr al-Hayr, East 12, 15, 17, 18, 28, 29, 30, 32, 131, 132, 138, 156, 175, 180, 188, 194, 195, 196, 197 Qasr al-Hayr, West 127, 131, 134, 141, 143, 144, 154, 156, 175, 179, 188, 194, 195, 196 Qasr al-Tuba 156, 188, 190 Qasr ibn Wardan 11 Qasr-i Shirin 294, 296 Qasr Kharana 188 Qastal 188 Qayrawan 76, 268 Great Mosque 78 Qays 138 Qaytbay 128 al-Qazwini 127 Qinnasrin 3, 24 Quraysh 60, 66 Qus 244 Qusayr 53 Qusayr ‘Amrah 4, 13, 16, 26, 27, 28, 29, 33, 49, 74, 131, 133, 142, 143, 149, 154,

index 325

159, 160–71, 175–8, 179, 181–4, 188, 189, 194, 196, 198–211 Rabat 76, 244 Ramallah 98 Ramlah 9, 157, 174 Raqqah 132, 133, 138 Rashid al-Din 311 Rayy 270 Ravenna 38 Riyadh 99 Rome 4, 25, 142, 151, 179 Pantheon 74 Rumi 27 Rusafah 10, 132, 134, 135, 138, 142, 143, 188 al-Sabi 142 Sadiqi 311 Safavids 272, 318 Safed 128 al-Saffah 63 Salonica 178 Samanids 309, 314 Samarkand 308 Samarra 48, 52, 53, 141, 142, 156, 159, 219, 225, 229, 233, 245, 249, 268 San’a 62, 100 Castle of Ghumdan 62 San Gimignano 73 Sangbast 80 Saragossa 272 Aljaferia 257 Sarvistan 291, 292, 293, 294, 295, 296, 297 Sasanians 29, 52, 147, 148, 186, 95, 197, 206, 226, 237, 291, 293, 294, 295, 296, 297 Seljuqs 5, 69, 133, 215, 216, 219, 220, 224, 232, 241 Severus ibn al-Muqaffa 152 Seville 272, 273, 312 Alcazar 272 House of Pilate 273 Shah Jahan 72 Sheba, Queen of 88, 92 Shi’ism 245, 247, 255

Sicily 216, 254, 272, 274 Sikandra Tomb of Akbar 91 Siraf 76 Soghd 180, 196 Solomon 88, 91, 92, 93, 94, 144 Spain 6, 36, 48, 69, 86, 167, 215, 243, 244, 257, 263, 267, 268, 271, 272, 273, 274 Spalato 16, 142 Tabari 142, 148 Tabgha 109 Taghlib 138 Tajikistan 317 Takht-i Sulayman 84, 296 Tarim basin 196 Tehran al-Ghadir Mosque 99 Theodore 45 Theodore Abu Qurra 48 Theophanes 45, 138 Theophilos 36, 45, 46 Tiberias 107, 109, 128, 131 Timurids 101, 312 Tinmal 244 Tirmidh 249 Tivoli 142 Toledo 272, 273 Bab Mardum 272 El Transito 273 Santa Maria la Blanca 273 Transoxiana 243, 313, 316, 319 Tulunids 220, 238, 247, 253, 254 Tunis 260, 312 Turan Pusht 318 Turfan 100 Turkmenistan 133 Turks 5, 6 Tyre 126 Ukhaydir 132, 155, 156, 191 ‘Umar 3, 6, 40, 64, 136, 137 Umayyads 4, 8, 13–16, 18–20, 24, 36, 38, 41, 44, 48, 49, 52, 64, 66, 67, 91, 97, 127, 128, 129, 131–5, 137–49, 151–7, 159, 160, 167, 173, 175, 176, 178, 180,

326

index

182–5, 187–91, 194, 195, 208, 210, 215, 220, 231, 236, 238, 246, 247, 257, 260, 263, 264, 269–72 ‘Uthman 64, 102, 263, 269 Venice 254 Vienna 33, 35, 37 Wahb ibn Munabbih 60 al-Walid I 20, 23, 24, 25, 26, 27, 28, 64, 65, 66, 107, 127, 144, 145, 146, 147, 148, 149, 152, 192

al-Walid II b. Yazid 142, 152 Wasit 139 Yazd 318 Yazid b. Abd al-Malik 44, 45, 148 Yazid b. Mu’awiyah 146 Yazuri 231 Yemenis 60, 61, 313 al-Zahir 246 Zakariyah 91 Zuleika 89