Eastern Approaches to Byzantium: Papers from the Thirty-third Spring Symposium of Byzantine Studies, University of Warwick, Coventry, March 1999 [Reprint ed.] 0754603229, 9780754603221, 1138263435, 9781138263437

Table of contents :
Abbreviations vii
List of Figures X
Preface xiv
Introduction / Antony Eastmond xvi
1. Speros Vryonis Jr. / 'The Decline of Medieval Hellenism in Asia Minor and the Process of Islamization from the Eleventh through the Fifteenth Century': The book in the light of subsequent scholarship, 1971-98 1
Section I. Byzantium's eastern frontier
2. Jonathan Shepard / Constantine VII, Caucasian openings and the road to Aleppo 19
3. Catherine Holmes / 'How the east was won' in the reign of Basil II 41
4. Jean-Claude Cheynet / La conception militaire de la frontière orientale (IXe-XIIIe siècle) 57
Section II. History writing in the east
5. Carole Hillenbrand / Some reflections on Seljuq historiography 73
6. Robert W. Thomson / The concept of 'history' in medieval Armenian historians 89
7. Stephen H. Rapp Jr. / From 'bumberazi' to 'basileus': writing cultural synthesis and dynastic change in medieval Georgia (K'art'li) 101
Section III. Byzantines
8. Liz James / Bearing gifts from the east: imperial relic hunters abroad 119
9. Catherine Jolivet-Lévy / Art chrétien en Anatolie turque: le témoignage de peintures inédites à Tatlarin 133
Section IV. Georgians
10. Zaza Skhirtladze / Newly discovered early paintings in the Gareja desert 149
11. Brigitta Schrade / Byzantium and its eastern barbarians: the cult of saints in Svanet'i 169
12. Giorgi Tcheishvili / Georgian perceptions of Byzantium in the eleventh and twelfth centuries 199
13. David Buckton / Stalin and Georgian enamels 211
Section V. Armenians
14. Lynn Jones / The visual expression of power and piety in medieval Armenia: the palace and palace church at Aghtamar 221
15. Helen C. Evans / Imperial aspirations: Armenian Cilicia and Byzantium in the thirteenth century 243
Section VI. Seljuqs and Turkomans
16. Rustam Shukurov / Turkoman and Byzantine self-identity. Some reflections on the logic of titlemaking in twelfth- and thirteenth-century Anatolia 259
17. Pamela Armstrong / Seljuqs before the Seljuqs: nomads and frontiers inside Byzantium 277
Index 287

Citation preview

EASTERN APPROACHES TO BYZANTIUM

Society for the Promotion of Byzantine Studies Publications 9

EASTERN APPROACHES TO BYZANTIUM Papers from the Thirty-third Spring Symposium of Byzantine Studies, University of Warwick, Coventry, March 1999

edited by

Antony Eastmond

ROUTLEDGE

Routledge Taylor & Francis Group

LONDON AND NRW YORK

First published 2001 by Ashgate Publishing Published 2016 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN 711 Third Avenue, New York, NY 10017, USA

Routledge is an imprint of the Taylor & Francis Group, an informa business Copyright © 2001 by the Society for the Promotion of Byzantine Studies Hon. Secretary, James Crow, Dept of Archaeology, The University, Newcastle-upon-Tyne NE1 7RU 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. Published by Variorum for the Society for the Promotion of Byzantine Studies British Library Cataloguing-in-Publication Data Eastern Approaches to Byzantium: Papers from the Thirty-Third Spring Symposium of Byzantine Studies, University of Warwick, Coventry, March 1999. (Publications for the Society for the Promotion of Byzantine studies ; 9) 1. Byzantine Empire - History -Congresses. 2. Byzantine Empire- HistoriographyCongresses I. Eastmond, Antony II. Society for the Promotion of Byzantine Studies III. Byzantine Studies Symposium (33rd: 1999: University of Warwick, Coventry) 949.5'02 U.S. Library of Congress Cataloging-in-Publication Data The Library of Congress Card Number was preassigned as: 00-108825 ISBN 13: 978-0-7546-0322-1 (hbk) Typeset by Wileman Design, Farnham, Surrey, UK SOCIETY FOR THE PROMOTION OF BYZANTINE STUDIES- PUBLICATION 9

Contents

Abbreviations List of Figures Preface

vii X

xiv

Introduction

Antony Eastmond

1.

The Decline of Medieval Hellenism in Asia Minor and the Process of Islamization from the Eleventh through the Fifteenth Century: The book in the light of

Speros Vryonis Jr

subsequent scholarship, 1971-98

Section I

xvi

1

Byzantium's eastern frontier

2.

Jonathan Shepard

3.

Catherine Holmes

4.

Jean-Claude Cheynet

Constantine VII, Caucasian openings and the road to Aleppo 'How the east was won' in the reign of Basil II La conception militaire de la frontière orientale (IXe-X IIIe siècle)

19 41 57

Section II History writing in the east 5.

Carole Hillenbrand

6.

Robert W. Thomson

7.

Stephen H. Rapp Jr

Some reflections on Seljuq historiography The concept of 'history' in medieval Armenian historians From bumberazi to basileus: writing cultural synthesis and dynastic change in medieval Georgia (K'art'li)

V

73 89

101

CONTENTS

VI

Section III

Byzantines

Bearing gifts from the east: imperial relic hunters abroad Catherine Jolivet-Lévy Art chrétien en Anatolie turque: le témoignage de peintures inédites à Tatlarin Liz James

8. 9.

Section IV

119

133

Georgians

10.

Zaza Skhirtladze

11.

Brigitta Schrade

12.

Giorgi Tcheishvili

13.

David Buckton

Newly discovered early paintings in 149 the Gareja desert Byzantium and its eastern barbarians: 169 the cult of saints in Svanet'i Georgian perceptions of Byzantium 199 in the eleventh and twelfth centuries Stalin and Georgian enamels 211

Section V Armenians 14.

Lynn Jones

15.

Helen C. Evans

The visual expression of power and piety in medieval Armenia: the palace and palace church at Aghtamar 221 Imperial aspirations: Armenian Cilicia and Byzantium in the thirteenth century 243

Section VI Seljuqs and Turkomans 16.

Rustam Shukurov

17.

Pamela Armstrong

Index

Turkoman and Byzantine self-identity. Some reflections on the logic of title­ making in twelfth- and thirteenthcentury Anatolia 259 Seljuqs before the Seljuqs: nomads and frontiers inside Byzantium 277 287

List of abbreviations

AB ABAW Abuldaze, Monuments

Analecta Bollandiana Abhandlungen der bayerischen Akademie der Wissenschaften I. Abuladze, ed., Dzveli k'art'uli agiograp'iuli literaturis dzeglebi (Monuments of old Georgian hagiographical literature) 3 vols (Tbilisi, 1963-71)

AJA

American Journal of Archaeology

AKKGWG

Abhandlungen der philologisch-historische Klasse der Königliche Gessellschaft der Wissenschaften zu Göttingen

AnatStud ANSMN AP

Anatolian Studies American Numismatic Society. Museum Notes

BAR BHL

British Archaeological Reports Bibliotheca hagiographica latina antiquae et medinae aetatis

BK BMGS

Bedi Kartlisa Byzantine and Modern Greek Studies

BS

Byzantina sorbonensia

ByzF ByzSlav BZ CahArch CCM

Byzantinische Forschungen Byzantinoslavica Byzantinische Zeitschrift Cahiers Archéologiques Cahiers de civilisation médiévale

CFHB CSCO CSHB

Corpus fontium historiae byzantinae Corpus scriptorum christianorum orientalium Corpus scriptorum historiae byzantinae

DChAE

A e X t îo v r fjç X p io n a v iK f iç ’A p x c u o À o y iK fjç 'E r a ip e ia ç

DOP El

Dumbarton Oaks Papers Encyclopedia of Islam, (1st edn: Leiden, 1913-38;

’A p x e io v TJo v t o v

vii

viii

LIST OF ABBREVIATIONS

2nd edn: Leiden, I960-)

GOxTR GRBS JOB JSAS JTS KC LCI

Greek Orthodox Theological Review Greek Roman and Byzantine Studies Jahrbuch der österreichischen Byzantinistik Journal of the Society of Armenian Studies Journal of Theological Studies S. Qauxch'ishvili, ed., K'art'lis c'xovreba (The Annals of Georgia), vols 1-2 (Tbilisi, 1955,1959) Lexikon der christlichen Ikonographie, eds E. Kirschbaum and W. Braunfels (Rome, Freiburg, Basel and Vienna, 1968-76), 8 vols

LThK

Lexikon für Theologie und Kirche

MGH NC OCA OCP

Monumenta Germaniae Histórica

ODB PG PL PO

Numismatic Circular Orientaba Christiana analecta Orientaba Christiana periodica A. Kazhdan, ed., The Oxford Dictionary of Byzantium (Oxford, 1991) J.P. Migne, ed., Patrologiae cursus completes, series graeca (Paris, 1857-66) J.P. Migne, ed., Patrologiae cursus completes, series latina (Paris, 1844-80) R. Graffin and F. Ñau, eds, Patrología Orientalis (Paris, 1904-)

PS REArm REB REGC RN

Palestinskii sbornik Revue des études arméniennes, new series Revue des études byzantines Revue des études géorgiennes et caucasiennes Revue numismatique (3rd series; 6th series)

SBAW

Sitzungsberichte der bayerischen Akademie der Wissenschaften. Philosophisch[-philologische] und historische Klasse Sources chrétiennes John Skylitzes, Synopsis historiarum, ed. I. Thurn, CFHB 5 (Berlin and New York, 1973)

SC Skylitzes

SMOMPK

Sbornik materialnov dlia opisania mestnostei i piemen kavkaza

SPBS

Society for the Promotion of Byzantine Studies (Publications)

SSCIS

Settimane di studio del centro italiano di studi sull'alto medioevo Theophanes Continuatus, ed. I. Bekker, CSHB

Theoph. Cont.

(Bonn, 1838)

LIST OF ABBREVIATIONS

IX

TIB

Tabula Imperii Byzantini

TM VV WZKM ZKOIRGO

Travaux et Mémoires Vizantiiskii Vremennïk Wiener Zeitschrift für die Kunde des Morgenlandes Zapiski Kavkazskago otdela imperatorskago russkago geograficheskago obshchestva Zbornik Radova Visantoloskog Instituta

ZRVI

List of figures

9.1 9.2 9.3 9.4 9.5

9.6

10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9

Tatlarin, église B. Abside nord: la Déisis de la conque (photo: author) Tatlarin, église B. Abside nord: la Théotokos dans la niche d'autel (photo: author) Tatlarin, église B. Abside sud: la Théotokos entre Michel et Gabriel, Joachim et Anne (photo: author) Tatlarin, église B. Schéma de la composition du mur ouest de la nef nord (drawing: author) Paris, Bibliothèque Sainte-Geneviève, Ms 56 (Commentaire sur les Psaumes de Pierre Lombard), fol. 185: Binité du Psautier (drawing: author) Gül§ehir, Kar§i kilise. Tableau infernal du mur ouest (photo: author) Gareja desert, T'et'ri Udabno. View of cave complex from south (photo: author) Gareja desert, T'et'ri Udabno. Principal church: plan and sections (drawing: N. Bakhtadre) Gareja desert, T'et'ri Udabno. Apse: Cross in mandorla (photo: author) Gareja desert, T'et'ri Udabno. Below apse: Virgin from Presentation in the Temple (photo: author) Gareja desert, Mravalcqaro. Small aisled church: plan and sections (drawing: M. Kiknadze) Gareja desert, Mravalcqaro. Schema of painting in north bay (drawing: author) Gareja desert, Mravalcqaro. North bay: Enthroned Virgin and Child (photo: author) Gareja desert, Mravalcqaro. North-west corner of naos: stylite and donor (drawing: author) Gareja desert, Mravalcqaro. Inscription of painter in sanctuary (drawing: author) x

LIST OF FIGURES

10.10 10.11 10.12 10.13

11.1

11.2 11.3

11.4

11.5

11.6

11.7 11.8 11.9 11.10

xi

Gareja desert, St Dodo. Small domed church. Apse: archangel and tetramorph (photo: author) Gareja desert, Sabereebi, Chapel 7. Apse: detail (photo: author) Gareja desert, Sabereebi, Chapel 8. Apse: detail (photo: author) Gareja desert, Camebuli. Small aisled church: detail of Deesis in sanctuary (photo: author) Cross of St George in Seti/Mestia (eleventh century). The 1.25 m high cross showing nine scenes of the martyrdom of Saint George was erected in front of the chancel. The archangel Gabriel (below) is a later addition probably replacing a lost inscription (photo: Rolf Schrade) St George in Nakip'ari. Wall painting of St. George tortured on the wheel (1130) (photo: Rolf Schrade) Cross in the treasury of St. George in Svip' P'ari (twelfth or thir­ teenth century). On the lower arm of the 2.40 m high cross are Sts George, Demetrios, Theodore, Mercurios, Panteleimon and Procopios (photo: Rolf Schrade) Enkolpion of Saint George from St. George in Ip'xi (second half of the twelfth century). The triptych is chased in gold and decorated with enamel, pearls and a carnelian. The inscription on the reverse names Mik'ael IV Mirianisdze, catholicos-patriarch at the beginning of Queen T'amar's reign as its donor. According to the eighteenth-century historian Vaxushti Bagrationi, it was brought from the patriarchal church of Svetic'xoveli to Svanet'i during the Mongol invasions of the thirteenth or fourteenth century (photo: Rolf Schrade) Icon of Saint George (twelfth century) from Labechina with an inscription of the Svan nobleman Ioane Vardanisdze. From: G. Chubinashvili, K'art'uli ok'romchedloba VIII-XVIII saukuneebisa (Georgian Goldsmith's art of the VIII-XVIII centuries) (Tbilisi, 1957), ill. 56 Chased icon of Saint George killing Diocletian from Seti/Mestia (eleventh century) with military saints on the side borders. The two inscriptions read: 'Saint George of Seti' (left) and 'Saint George' (right) (photo: Rolf Schrade) The former monastery of St. Quiricus (Lagurka) in Kala during the feast of the saint on 28 July (photo: Rolf Schrade) Lagurk'a monastery. Wall painting of St Quiricus dying on the staircase of Diocletian's throne (1111) (photo: Rolf Schrade) Chased icon of Saint Quiricus in the treasury of Lagurk'a monastery (first third of eleventh century) (photo: Rolf Schrade) Chased icon of Saint Barbara (eleventh century) from St. Quiricus

xii

11.11

11.12

11.13

11.14

14.1 14.2 14.3 14.4 14.5 14.6 14.7

14.8 15.1

LIST OF FIGURES

(Lagurka) in Ieli (now Mestia museum). The borders are adorned with representations of the deesis (above), the martyrs and mili­ tary saints George and Procopios (left and right) and the physi­ cian saints Cosmas, Panteleimon and Damian (below) (photo: Rolf Schrade) Painted icon of the Theotokos and Saint Barbara (tenth or eleventh century) from St George in Nakip'ari (photo: Rolf Schrade) Wall painting of the lower church of the Saviour in Laghami (tenth or eleventh century) with two male saints, probably Artemios and George, on the southern and Saint Barbara on the western wall (photo: Rolf Schrade) Silver head of the Svan flag Lem (thirteenth century), a donation of Grigol Kopasdze, prior of the monastery of St. George in Seti. The 'Archangel of Ugh(vi)ri' (Mt'avarangelozi ughrisa) as patron of the corresponding community is invoked together with 'Saint George of Seti' and 'Saint Jonas of Latali' as helper of whole Svanet'i (photo: Rolf Schrade) Chased icon of the Archangel Michael from the church of the Archangel of Muxeri (thirteenth or fourteenth century) (photo: Rolf Schrade) Aghtamar, church of the Holy Cross. East façade, portrait of Gagik Artsruni (photo: author) Medallion depicting al-Moqtadir (after B. Spuler and J. SourdelThomine, eds, Die Kunst des Islam) Aghtamar, church of the Holy Cross. West façade, portrait of Gagik Artsruni presenting his church to Christ (photo: author) Aghtamar, church of the Holy Cross. West façade, detail of frieze (photo: author) Aghtamar, church of the Holy Cross. Distribution of the lower fresco cycle (drawing: author) Aghtamar, church of the Holy Cross. Partial distribution of the upper fresco cycle (drawing: R. Freyman) Aghtamar, church of the Holy Cross. The Expulsion from Paradise and the Seraph Guarding Paradise (after N. Thierry, 'Le cycle de la création et de la faute d'Adam à Alt'amar', REArm 17 (1983)) Aghtamar, church of the Holy Cross. East façade (photo: author) Erevan, Matenadaran, Ms. 8321, fol. 25: portrait of Prince Levon (courtesy of the Matenadaran Institute of Ancient Armenian Manuscripts, Erevan, Armenia)

LIST OF FIGURES

15.2

15.3

15.4

15.5

15.6

17.1 17.2 17.3 17.4

xiii

Jerusalem, Armenian Patriarchate, Ms. Jerusalem 2660, fol. 288: portrait of Prince Levon and his bride princess Keran (courtesy of the Library of Congress, Washington, DC) Paris, Bibliothèque nationale de France, Ms. Coislin 79, fol. I (2bis) V: portrait of Michael VII Doukas (repainted as Nikephoros III Botaneiates) and Maria of Alania (Bibliothèque nationale de France: photographic plate) Jerusalem, Armenian Patriarchate, Ms. Jerusalem 2563, fol. 380: portrait of King Levon II and Queen Keran with their children (courtesy of the Library of Congress, Washington, DC) New York, Pierpont Morgan Library, leaf from Ms. 740, [Ms. M. 1111]: portrait of Archbishop Hohannes presenting baron Oshin and his family to the Virgin and Christ child (courtesy of the Pierpont Morgan Library, New York) Venice, Mekhitarist Library, ms. 107,: portrait of King Levon IV rendering justice (courtesy of the Library of Congress, Washington, DC) Map of southern Anatolia showing places mentioned in text Section through one-handled cooking pot/mug (drawing: author) Section through two-handled jar, showing rounded base (drawing: author) Almost complete two-handled jar, showing stick-burnished lines and rounded base (photo: author)

Preface

This volume arises from the Thirty-third Spring Symposium of Byzantine Studies - Eastern Approaches to Byzantium - held at the University of Warwick in March 1999. The Symposium was organised in the optimistic (and naive) belief that it would be possible to bring together scholars from across Europe, the Caucasus and America who could speak on all aspects of life and cultural exchange on the Byzantine frontier. I had not at that point considered what this would all cost. That the Symposium did succeed in attracting speakers from Austria, France, Georgia, Germany, Ireland and Russia, as well as from across the UK and USA to do just that stands entirely as a tribute to all those who were prepared to fund such an extravagant idea. It is therefore with great pleasure and a debt of grat­ itude that I thank the Humanities Research Centre of the University of Warwick, the Leventis Trust and the Hellenic Foundation for their extra­ ordinary generosity. I received further grants from the British School of Archaeology at Ankara, the Society for the Promotion of Hellenic Studies, and the Centre for East Roman Studies at the University of Warwick. These contributed to speakers' travel costs and to the subsidies needed for students to be able to attend a symposium at a market-driven conference university. Those speakers who were able to fund themselves are my particular heroes. The support of John Smedley and Variorum/Ashgate for the Symposium Feast lubricated many discussions. A number of speakers were not able to include their papers in this volume, but I would like to thank Anthony Bryer, Hugh Kennedy, Alexei Lidov, Michael Rogers and Rachel Ward for their contributions to the Symposium itself. As ever, the communications given by other scholars were integral to the Symposium, and abstracts can be found in the Bulletin of British Byzantine Studies 26 (2000), 75-91. In the organisation of the symposium I was blessed by the help of my mestumretukhutsesi, Susan Dibben, and of my pareshni, Duncan Givans and Ian Kelso. Stephen Hill provided an exhibition on Eski Giimu§, derived from material in the Michael and Mary Gough archive that he curates at Warwick University.

xiv

PREFACE

xv

The swift publication of this volume owes much to the efficiency of Kirsten Weissenberg at Ashgate, and also to all the contributors who heeded my incessant nagging to turn in copy and answer queries within tight deadlines with efficiency and (a façade at least of) good humour. Finally, a word needs to be said about transliteration. The papers in this volume employ Arabic, Armenian, Georgian, Persian, Syriac and Turkish in addition to the normal Greek and other Western and Slavonic languages. Given this, the pursuit of consistency has proved a thankless task and an elusive goal. Any major discrepancies and alternatives are listed in the index.

Introduction

Antony Eastmond The papers collected in this volume derive from Eastern Approaches to Byzantium, the Thirty-third Spring Symposium of Byzantine Studies, which was held at the University of Warwick in March 1999. They examine Byzantium from both sides of its eastern frontier. Concentrating on the period of the re-conquest and subsequent loss of the eastern provinces from the ninth to thirteenth centuries, they look at aspects of Byzantine policy and actions all along the frontier, and also at the actions and reactions of the peoples that they encountered: principally those of the Armenians, the Georgians and the Seljuqs, but also of the Syriacs and the other Turkoman tribes. Rather than concentrate on relations between Byzantium and one of its neighbours, the Symposium discussed ques­ tions of interaction among all the peoples on the eastern frontier of the Byzantine world. The importance of the eastern frontier to the development of Byzantium from the ninth to thirteenth centuries has long been estab­ lished in modern Byzantine histories, whether characterized in terms of over-expansion in the tenth century (with consequent effect on the economy and resources of the empire), of bad management in the eleventh century (the idea that the annexation of the Armenian kingdoms led to the loss of a buffer zone which opened the way to the Seljuq inva­ sions), or of consumption of scarce manpower and resources in the twelfth century in a futile and 'misguided' attempt to regain what had been lost. To a lesser extent, it has been recognised that the processes of interaction on the eastern frontier, and the introduction of new popula­ tions - mostly Armenians - into Byzantium as the empire swallowed up their territories also had an impact on Byzantine society. One of the prin­ cipal aims of the Symposium was to bring together new research into the peoples to the east of Byzantium, but the results of this are twofold. The first is to build up a greater understanding of each of these neighbouring societies whose history impinged on that of Byzantium so much in this period. It is this aspect that most of the papers concentrate on. The second is to use this knowledge to help understand Byzantium itself, by examinxvi

INTRODUCTION

XVII

ing the nature of the encounters and exchanges, whether military, politi­ cal, cultural or ideological, and the impact that they had on both sides of the frontier. These results can be gleaned from the volume as a whole, and it is these that point to the direction for future research. Any attempt to examine the whole length of the frontier in terms of history and imperial policy, of artistic and cultural production, of the perceptions, interactions and influences of all the peoples and cultures involved is an enormous undertaking. At the Symposium it was only possible to scratch the surface of the many ways in which all the peoples of the east interacted with Byzantium and with each other over this period. One of the key points that emerges from this volume lies not in the common threads and themes which link many of the papers, but in the diversity of materials and approaches and the many areas of differ­ ence between the papers. This book demonstrates the fragmentation of ideas, policies and identities along the eastern frontier of the Byzantine empire. From this mosaic of religions, histories, cultures and peoples we can build up a fuller picture of the role that the eastern frontier, and the peoples who lived along it, played in the formulation of the middle Byzantine empire, as well as the impact of Byzantium on its neighbours. Many of the papers in this volume deal with topics that lie outside the traditional (but artificial and self-imposed) 'borders' of Byzantine studies; some, indeed, argue that the influence of Byzantium in certain aspects of the history and culture of its neighbours was minimal. However, all the papers are important to furthering our understanding of the Byzantine empire. In those cases where an aspect of the relationship between Byzantium and one of its neighbours is argued to be slight, it is the absence or denial of the relationship that becomes the key to analysis. The study of why Byzantine models were not adopted or deemed useful in these particular contexts provides insights into the limits of Byzantium, particularly in its self-defined roles as bastion of Christianity and ultimate power on earth. These studies can help us to examine the ways in which Byzantium's institutions and ideologies were constructed for its own particular needs; this is a task that cannot be so easily undertaken from within Byzantium itself. In the cases where Byzantine ideas and models were not adopted or adapted abroad we can see the very frontiers of Byzantine identity itself: here the margins (from a Byzantine point of view) can define the centre. Elsewhere, it is in the varying nature of the relationships and exchanges revealed that the importance of these papers for the study of Byzantium lies. Some present familiar material in an alien context, such as the cult of St George now seen through Svan eyes, or the pseudo-Byzantine titles proclaimed by Turkoman rulers; others examine alien approaches to familiar problems, whether it be the promotion of royal power as

xviii

ANTONY EASTMOND

presented in tenth-century Vaspurakan, or the methods and aims of the writing of history in the east. All enable us to look at the ways in which these issues were handled in Byzantium itself in a new light. They also raise methodological issues as the possibilities opened up by access to different types of material in different contexts can provide models for the ways in which these issues are handled by Byzantinists. The contrasts and similarities throw Byzantium into sharper relief. The opening chapter by Speros Vryonis charts how far studies of Anatolia and the eastern frontier have developed since the publication, thirty years ago, of his seminal book, The Decline of Medieval Hellenism in

Asia Minor and the Process of Islamization from the Eleventh through the Fifteenth Century. In it he notes that some disputes have been resolved and some lacunae filled, but most of all he makes clear how much more there is still to be done. What is required, ultimately, is a composite, compara­ tive study of the processes and effects of cross-cultural exchange all along the frontier, and this volume provides some of the much-needed evidence that will lie at the heart of any such work. Individually, the case studies here stand alone as studies of particular problems, but together they provide fragmentary glimpses of the types of material that must be gath­ ered as well as the questions that must be asked of them. They build up a picture of the complex and disparate societies that developed to the east of Byzantium. And these, in turn, must inform our understanding of Byzantium itself. The first section on Byzantium's eastern frontier presents the view from within the empire. The chapters here explore the problems Byzantium faced as it tried to negotiate, advance and defend its frontier. Jonathan Shepard and Jean-Claude Cheynet write of the vagaries of imperial policy and their effects, of the many different policies required all along the iron-, tier, and of the ways in which these policies had to change over the centuries to adapt to new circumstances. From these, the frontier emerges as an ambiguous, changing space. Catherine Holmes's study of the local intermediaries, often of Muslim origin, who governed the border regions under Basil II demonstrates not only the porous nature of the border, but also the impossibility of providing an easy definition of the empire and its regional administrators. How were these officials regarded by Constantinople, and how did they view themselves? What impact did they have on contemporary perceptions of the empire? All demonstrate how Byzantium itself was changed by the re-conquest of the east, not only in the obvious ways one would expect from the inclusion of new lands and new peoples within the frontiers of the empire, but also in more subtle ways: the changing dynamic of Byzantine society itself, and the impact it had on Byzantine culture. Further evidence for these changes is provided by Catherine Jolivet-

INTRODUCTION

xix

Lévy and Pamela Armstrong from art and archaeology respectively. They write about the populations on the 'wrong' side of the frontier - of Greek Orthodox Christians living in Seljuq-controlled Cappadocia in the thir­ teenth century, and of nomadic Turkomans in Byzantine Lycia more than a century earlier. In each case they raise questions of the process of nego­ tiation (or lack of it) between these populations and the dominant culture around them, and of the survival of aspects of their indigenous cultures. Jolivet-Lévy demonstrates the degree to which the Orthodox Christians in thirteenth-century Cappadocia were able to maintain cultural ties with the empire of Nicaea, but also hints at the transformation wrought by Seljuq overlordship. The chapters on the Armenians, Georgians and Seljuqs present very different approaches to the study of this interaction between east and west in Anatolia and the Caucasus: the delicate balance between Byzantium, Persia, and local indigenous cultures. They are particularly valuable for the evidence they provide of changing perceptions of Byzantium, which can be related back to changes within Byzantium itself. These issues come across clearly in the three chapters on history writing in the east. The studies of Seljuq, Armenian, and Georgian history-writing traditions, by Carole Hillenbrand, Robert Thomson and Stephen Rapp respectively, outline the differing agendas and cultural ties which shaped much of the writing of history to the east of Byzantium. They demonstrate the moral and cultural conceptual structures that each of these peoples brought to their own history, and the ways in which they have shaped how modern historians must investigate them. These chapters force us re­ evaluate our use of these texts as sources for Byzantine history, and they present new methodological frameworks for the interpretation of the events and actions they describe. The question of changing perceptions of Byzantium is made most explicit in Giorgi Tcheishvili's chapter on Georgian attitudes to Byzantium, which reflects the changes relations underwent as the political, economic and military position of both states developed in the eleventh and twelfth centuries. The other chapters on these peoples to the east employ material evidence to explore aspects of their cultures. The two chapters on Armenia concentrate on élite culture, and on the promotion of rulers and the public projection of their political ideologies, first in tenth-century Vaspurakan, and then in thirteenth-century Cilicia. Lynn Jones's chapter on the visual promotion of king Gagik Artsruni in his palace church at Aghtamar presents a nuanced view of his concept of his own power, as it balanced Islamic caliphal iconography with local means of expressing Christian orthodoxy; she argues for a minimal impact of Byzantine models of power. Helen Evans's examination of the royal image presented by Levon Het'umid of Cilicia three hundred years later shows

XX

ANTONY EASTMOND

how far the geopolitical situation had changed. Now Byzantine imagery, rather than Islamic, lay at the core of the royal image, but this was being supplemented, if not supplanted, by the vigorous Crusader states to the south and by contact with the new power in the region, the Mongols. Both papers provide perceptive accounts of these different Armenian states that force us to re-evaluate their histories, and to revise the role played by the neighbouring cultures to which they turned in the formation of their royal identities. But both also allow us to gain new insights into the Byzantine state: they reveal much about the status of the empire and of the ways in which it was perceived by those who lived around it. These studies of Armenian royal imagery and the embodiments of Christian power that they present provide a control against which we can judge Byzantium itself, and its internal perceptions of its own power. To compare the image of Gagik on the west facade of Aghtamar with the contemporaneous portrait of emperor Alexander in Hagia Sophia in Constantinople is to witness the stark disparity between Byzantine pretensions and Armenian experience. It reveals how fragile the glorious vision of imperial Byzantine power was at the start of the tenth century among its Christian neighbours as the empire began its re-conquest of the east. Alexander's overbearing pomp carries no weight compared to the lure of caliphal power. Equally - if at first sight paradoxically - Levon Het'umid's more Byzantine appearance in the manuscripts of Cilicia demonstrates an analogous disparity in the thirteenth century. Byzantium does now provide the essential visual and material attributes for the display of power, but these can now be appropriated precisely because of the collapse of Byzantine power in the east over the previous 150 years; a collapse now made absolute by the sack of Constantinople in 1204 and the consequent fragmentation of the empire. Armenian Cilicia provides us with yet another vision of 'Byzantine' power to add to those promoted by the successor states of Nicaea, Epiros and Trebizond. The fall of Constantinople in 1204 opened up Byzantium to redefinition, and we must remember to include those 'non-Byzantine' attempts to appro­ priate it and to lay claim to the mantel of imperial authority in any account of the development of Byzantium in the thirteenth century. In contrast, the chapters on material culture in Georgia by Zaza Skhirtladze and Brigitta Schrade look to very different levels of society away from the heartlands of a royal court. They present material from the most remote regions of the country, at its two climatic extremes. They move from the art produced by the ascetic monks who lived in the rockcut monasteries of the semi-arid Gareja desert (on the modern frontier between Georgia and Azerbaijan) in the eighth to tenth centuries to that commissioned by the minor, local élites of the mountainous province of Svanet'i, high in the Caucasus, in the eleventh to thirteenth centuries. In

INTRODUCTION

xxi

both cases the evidence - published here for the first time - concerns the promotion and veneration of cults, and the liturgical and ritual use of art among these societies. The focuses of this veneration - Sts George and Barbara, Sts Quiricus and Julitta, archangels, the cross, the Theotokos are familiar throughout the Byzantine world, but here they emerge filtered through local customs and adapted to suit local needs. Again these studies reveal differing perceptions of Byzantium, but this time of its spiritual authority rather than its political power. They demonstrate the early residual authority of Syria and the Holy Land as an alternative source for spiritual leadership and the subsequent, but never complete, domination of Byzantium, as transmitted through the monasteries of Mount Athos. In the past, the discussion of this type of material has often tended to run to extremes: representing the cults either as provincial, corrupted forms of Byzantine cults, or as specifically local creations, requiring nationalistic interpretations. However, the more syncretistic studies here run along much more sensitive lines, and the value of the material they produce for our understanding of Byzantium is enormous, not least because it fills in gaps in our knowledge of what was happening in Byzantium itself (whether in the case of church decoration in the eighth century, where the artistic losses in Byzantium are so great, or in fleshing out details of the ways in which a Christian cult could function within a particular society). However, it is also in the very process of change, accretion and adaptation, in the differences between the veneration of St George or the archangel Michael in Byzantium and Svanet'i and the rest of Georgia, that their value lies. The mutability of these cults, and their ability to transform themselves to suit local needs parallels a similar ability within Byzantium itself - think only of the specifically local aspects of the cults of figures such as St Demetrios in Thessaloniki, or St Eugenios in Trebizond. These studies provide further comparative evidence for the ways in which cults operated throughout the east Christian world. Much work in recent scholarship has begun to focus on the intricacies of Byzantine society, on the diversity of its populations, on the networks of alliances and cultures that it contained, and on its different regional affiliations. From these a more detailed picture is emerging of the complex nature of Byzantium, of its varying perceptions of itself, and of the fragmented nature of Byzantine identity. By approaching Byzantium from the east, this volume seeks to advance that process and to expand the range of cultures that must be included in that picture of Byzantium. As the papers in this volume all show, there was no single eastern fron­ tier, but rather a series of borders - physical and mental - that continually overlapped with one another.

This page has been left blank intentionally

1. The Decline of Medieval Hellenism in Asia Minor and the Process of Islamization from the Eleventh through the Fifteenth Century.

,

The book in the light of subsequent scholarship 1971-98 Speros Vryonis Jr Perhaps it would be appropriate first to try to answer the question of why the book was ever written. One reason was the stimulus of my recently finished dissertation, which entailed an analysis of the Byzantine sources that dealt with Turkish nomadic incursions and settlements in Asia Minor and the perceptible economic and institutional decline in substantial parts of Anatolia during that time.1 By 1959, the broad question of what happened to the Byzantine, Armenian, Georgian, and Syriac Christian sectors of Asia Minor in the face of the Turkish tribal raids, conquests and settlements crystallized into the project that led to seven years of research and writing that resulted in the book. From this research I arrived at the following conclusions: 1.

2.

3.

The Greek-speaking cultural element of Asia Minor constituted a vital historical society, and thus the final conquests and Islamization of the peninsula represent something more than a negative historical event. The Turkish tribes and sedentary Islamic sultans did not completely subdue the peninsula until the latter half of the fifteenth century; and the processes of the military and cultural transformation of the region's Christian population took place over a long period. The nature of the conquests and settlements caused major dislocation and partial destruction to Byzantine society. The older Byzantine rule was shattered and replaced (as was that of the sultanate of Konya) by smaller unstable political entities that were, for very extensive

1 S. Vryonis, The Internal History of Byzantium in the Time of Troubles. (1057-1081)' (unpublished dissertation, Harvard, 1956). From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

1

2

4. 5.

6.

7.

SPEROS VRYONIS JR

periods, at war with one another. This condition, together with Muslim hegemony, exercised a corrosive action on structural bonds that had once held Christian communities together, thus preparing the Christian populace for religious conversion. Yet as of the mid-thir­ teenth century the Christians of Asia Minor seem to have constituted a substantial portion of Anatolian demography. The political and historical events of these conquests destroyed the Church as an effective social, economic and religious institution. Christian society, which had been subjected to such repeated disloca­ tions and isolated from its religious and political centre in Constantinople, and further often deprived even of its hierarchical leadership by the policies of the new small states, was ripe for absorp­ tion into this newly arrived and dynamic Islamic society. This was largely the work of traditional Islamic institutions, heavily supported by the political and economic patronage of the new Turkish emirs, sultans and their followers. These Muslim institutions (based on the former economic wealth and possessions which the conquests had transferred from the Christian to the new conquering society), the most important of which were the dervish orders, consti­ tuted, perhaps, the central missionising and syncretising force that brought Christians into the Islamic religion. Finally, Byzantine culture in Anatolia, though effaced at the level of formal culture by Islamic formal culture, nevertheless exercised an important role in aspects of Turkish popular culture and technology.

It was clear from the outset that such an inordinately ambitious under­ taking, encompassing the vast scholarly domains of Byzantinology, Turkology, Islam, Armenology and Balkanology, could not be anywhere near complete. The preface of the book noted this in partial but great amplitude. It observes, as of the time of the completion of the text some 32 years ago, the need for new, detailed histories of the Seljuqs, the 'empires' of Nicaea and Trebizond, and of the very numerous, smaller Turkish emirates. Much of the history of Byzantine Anatolia had not been systematically reconstructed and analysed. Substantial works on the process of Islamization of the Armenians, Georgians and Syriac Christians, as well as of the Greek-speaking population, were lacking. Finally, comprehensive studies of the folklore and popular cultures of the Anatolian Muslims and Christians were in their infancy: 'The number and extent of desiderata surely indicate how imperfect and insufficient the present work must be. I have concentrated on the fate and Islamization of the Greek speaking population in Anatolia to the exclu­ sion of the other Christian groups.' Ten years after its publication, I assayed an analysis of twenty reviews

THE DECLINE OF MEDIEVAL HELLENISM

3

which had been published in a broad range of journals.2 Having purposely waited for one decade before considering the positive and negative remarks of the twenty reviewers, I finally had to come to terms with the observations, and profited from indications of factual errors such as the nature of the ecclesiastical pentad, the location of a particular town, and various other isolated data. The basic theses and propositions, however, were not basically or effectively disrupted. Certainly these will have to be either restated or completely altered in the light of future research. In this paper I shall indicate some of the new and vigorous develop­ ments in this broad area dealing with the vast changes in Asia Minor primarily after the onset of the Turkish invasions, but also the general character of Byzantine Anatolia before that time. The recent Greek trans­ lation of my book carries a new preface in which the effort was made to give a representative sampling of this newer scholarly literature, and reference was made to 121 such contributions which the scholarly realms of Byzantinology, Armenology, Arabology and Turkology have produced.3 However, a systematic combing of the Byzantinische Zeitschrift bibliography fascicules and of other journals indicated that these 121 new references represent somewhere between 5 and 10 per cent only of a much larger scholarly production. In the years between the handing over of the manuscript to the press and 1981, when the review of the reviews was carried out, over 1,000 bibliographical cards were added to my ongoingAsia Minor bibliography. Since that time the number has doubled. Since then there have appeared the works of Angold and Langdon on the state of Nicaea,4 those of Karpov and Bryer on Trebizond.5 Still in the realm of Byzantine Asia Minor, there have appeared recently two works on Byzantine Asia Minor, and on the themes of Asia Minor.6 The most impressive of all the group projects is undoubtedly that of the Austrian Academy of Sciences, which has engaged in a long-term effort to establish 2 S. Vryonis, 'The Decline of Medieval Hellenism in Asia Minor and the Process of Islamization, the Book and its Reviews Ten Years Later', GOxTR 2 2 /2 (1982), 225-85. 3 S. Vryonis, E parakme tou mesaionikou Ellenismou ste Mikra Asia kai e diadikasia exellenismou (llos-15os aionas), trans. C. Galatariotou (Athens, 1996), 1-8. 4 M. Angold, A Byzantine Government in Exile. Government and Society under the Lascarids of Nicaea (1204-1261) (Oxford, 1975). J.S. Langdon, Byzantium's Last Imperial Offensive in Asia Minor. The Documentary Evidence for and Hagiographical Lore about John III Ducas Vatatzes' Crusade against the Turks, 1222 or 1225 to 1231 (New Rochelle, NY), 1992. 5 A. Bryer, The Empire of Trebizond and the Pontos (Aldershot, 1960). S. Karpov, L'imperio di Trebisonda, Venezia, Genova e Roma 1204-1261 (Rome, 1986). 6 N. Oikonimides, ed., Byzantine Asia Minor (6th-12 c.) (Athens, 1998); V. Vlysidou, E. Kundoura, S. Lampakes, T. Lounges and A. Savvides, E Mikra Asia ton thematon (Athens, 1990).

4

SPEROS VRYONIS JR

a historical geography of Byzantine Asia Minor.7 This represents the first major and massive input to our knowledge of the Byzantine Anatolian landscape, its road systems, and the location of towns and villages. All of these works touch the fundamental questions of demography and ecclesiastical administration in Byzantine Asia Minor. In particular they have raised the question of the significance, or lack thereof, of the ecclesiastical notitiae episcopatuum in Asia Minor. Do these documents refer to actual historical conditions in the bishoprics and metropolitanates of Asia Minor and to the presence in them of their hierarchs? Are they to be used in attempting to establish a list of 'towns' in Asia Minor? Or, are they simply pieces of mandarin antiquarianism without any reference to the historical realities in the province? The twelfth-century commentaries of Balsamon are comments on the historical realities of the structure of bishoprics and metropolitanates and their decline, and the inability to appoint hierarchs in many of them.8 The controversy as to whether or not there was continuity in late antique towns was set by the illustrious Byzantinist Alexander Kazhdan.9 At that time, when he was writing under the onus of a rigid state Marxist materialistic interpretation of historical evolution, Kazhdan chose the seventh to ninth centuries as the period in Byzantium which constituted the era of transition from a slave-based to a feudal society. According to the Marxist scheme this transitory stage to feudalism was accompanied by the disappearance of slavery - the basis of the ancient city - and so led to the disappearance of cities and towns. He argued further that this disappearance of urban life was manifested in (a) the decline in the number of coins found at excavations, in coin hoards and in museum collections; and (b) the complete disappearance of merchants, craftsmen and trade. The text of Theophanes, alone, would have been sufficient to disprove all of this. And indeed Ostrogorsky utilized that text to identify settle­ ments which Theophanes names as towns.10 But there are much more

7 This includes, of course, the TIB series and related volumes: F. Hill and M. Restle, Kappadokien (Kappadokia, Charsianon, Sebasteia und Lykandon) (Vienna, 1981). K. Belke and M. Restle, Galatien und Lykaonien (Vienna, 1984). F. Hill, Das byzantinischen Strassen-system in Kappadokien (Vienna, 1977). F. Hellenkemper and F. Hill, Neue Forschungen in Kilikien (Vienna 1986). A. Machatschek and M. Schwartz, Bauforschungen in Selge (Vienna, 1991). K. Belke and N. Mersich, Phrygien und Pisidien (Vienna, 1990). K. Belke, Paphlagonien und Honorias (Vienna, 1996). 8 G.A. Rhalles and M. Potles, Syntagma ton theion kai ieron kanonon II (Athens, 1852-59), 388. J. Darrouzes, Notitiae episcopatuum Ecclesiae Constantinopolitanae (Paris, 1981). 9 A. Kazhdan, 'Vizantiiskie goroda v VII-XI vv', Sovietskaia Arkheologiia 21 (1954), 164-88; A. Kazhdan, Derevniia i gorod v Vizantii IX-X vv (Moscow, 1960). 10 G. Ostrogorsky, 'Byzantine Cities in the Early Middle Ages', DOP 13 (1959), 47-66.

THE DECLINE OF MEDIEVAL HELLENISM

5

specific texts which refer to: (a) the money economy and circulation of coin; (b) the existence of merchants and craftsmen; (c) lively maritime and land trade; (d) an important trade in slaves which was taxed, in gold coin, at Abydos and in the Duodecanese. He testifies that the theme troops were paid in gold coin in both Strymon and Armeniakon. The legal codes of the time (the Sailor's Law, Farmer's Law, and the Ecloga) all speak of salaries, fines and dowries in gold coin.11 The studies by Clive Foss have shown damages to cities during the late Persian invasions. From these he surmises the general collapse of Byzantine urban society.12 John Haldon has given a tight summary and interpretation of a number of studies dealing with the problem of what happened to the late ancient town, and of the towns as demographic centres of some content and of different administrative nature.13 The point, and it had been made in great detail by the late Bulgarian historian and archaeologist, Velizar Velkov, was that the late ancient city had already taken on the basic insti­ tutional forms of town life in the early Byzantine middle ages.14 Haldon has, however, seized on the extreme importance of the internal decline of the late antique town's local patriotism, its funds, and its urban political institutions, and their assumption by the organs of the state and church. His careful analysis of the raids, first of the Persians and then of the Arabs, contributes to the picture of a further dislocation of a portion of town life. The degree to which the smaller and militarily harassed towns lost their commercial nature is a very difficult problem precisely again because of the lack of inscriptional and textual material, and because the archaeological material remains suggestive rather than conclusive. The attempted parallel between these small demographic centres and such super-cities as Constantinople, Baghdad and Damascus is perhaps not fruitful and is faulty methodologically. The reliance on the Huddud al-Alam, a twelfth-century pastiche of different and contradictory data even on the existence of these small towns (kastra) or of their disap­ pearance, is ill placed.15 Finally the comparison of the Arab razias in the

11 On these texts see S. Vryonis, 'An Attic Hoard of Byzantine Gold Coins (668-741) from the Thomas Whittemore Collection and the Numismatic Evidence for the Urban History of Byzantium', Z R V I8 (1963), 291-300. 12 C. Foss, 'Archaeology and the 'Twenty Cities' of Byzantine Asia Minor', A]A 81 (1977), 469-88. 13 J. Haldon, Byzantium in the Seventh Century. The Transformation of a Culture (Cambridge, 1990). 14 V. Velkov, Cites in Thrace and Dacia in Late Antiquity (Studies and Materials) (Amsterdam, 1977) is the more expanded version of the Bulgarian first edition. 15 V. Minorsky, Hudud al-Alam. The Regions of the World: A Persian Geography, 372AH 982AD (London, 1937), 40,157.

6

SPEROS VRYONIS JR

vast space of Asia Minor during the seventh and eighth centuries with the Turkish nomadic raids on the towns and villages of eleventh- to twelfthcentury western Asia Minor represents a comparison of two phenomena that are different militarily, economically and socially. The question of commerce (local or regional) and of craft specialization within these kastra agglomerates is one about which so far one can only make assumptions. But these kastra were characterized by specialized institutions: notably the presence of state and military officials and, prob­ ably, of ecclesiastical hierarchs. One can, I think, say that this differenti­ ated them from villages. The question of the continuity of towns as demographic agglomerates, with appropriate socio-economic differentiation, with a lively commerce and with specialized trades, still awaits further examination. Archaeology and perhaps sigillography will inform us. An additional and important problem is the question of the great reli­ gious change which led from Anatolia's various paganisms to Christianity of various sorts. Here the magnum opus of Frank Trombley,16 for the first time since the basic analytical synthesis of Johannes Geffcken,17 attempts a scholarly continuity of the latter's work more or less where it left off chronologically, but with new approaches and with a very broad geographical coverage that includes areas of Anatolia. Trombley, also a long-standing student of the late antique and early Byzantine evolution of and changes in town life, he has closely integrated the approach to religious change with the urban and rural structures of Anatolia. This was a most important area which was but briefly touched upon in the opening chapter of The Decline of Hellenism, and so the more recent work on Byzantine urban and religious history in Anatolia has taken on much form that was previously lacking. In the realm of economic history, the works of Michael Hendy and of Alan Harvey have presupposed an exaggeration of economic decline in the first half of the eleventh century in an effort to explain the catastrophic collapses after 1071 of the state in southern Italy and Anatolia.18 They suppose continued growth in such factors as demographic expansion and in the proposition that one should consider factors other than the interests

16 F. Trombley, Hellenic Religion and Christianity c.370-529, 2 vols (Leiden, 1993-94). H.-G. Beck, Vom Umgang mit Ketzern. Der Glaube der kleinen Leute und die Macht der Theologen (Munich, 1993). 17 J. Geffcken, Der Ausgang des griechsich-romischen Heidentums (Heidelberg, 1920); trans. S. MacCormack, The Last Days of Greco-Roman Paganism (New York, 1978). 18 A.L. Harvey, Economic Expansion in the Byzantine Empire 900-1200 (Cambridge, 1989); M. Hendy, Studies in the Byzantine Monetary Economy c. 300-1450 (Cambridge, 1985); idem. The Economy: A Brief Survey', in S. Vryonis, ed., Byzantine Studies. Essays on the Slavic World and the Eleventh Century (New Rochelle, NY, 1992), 141-52.

THE DECLINE OF MEDIEVAL HELLENISM

7

of the state. But the two seem to differ on the degree to which this was true of eleventh- and twelfth-century Anatolia. Hendy acknowledges: Between c.800 and c.1200, the empire relied in essence upon two large and peninsular land masses, although at different stages of the period, different territorial balances obtained: during the ninth and tenth centuries it relied essentially on Anatolia, with the Balkans as something of a makeweight. During the twelfth it relied on the Balkans with Anatolia as makeweight. During the eleventh century alone it possessed, and was therefore able to exploit, both peninsulas, in theory giving it an exploitative maximum.19

Then he tries to differentiate between the predominant raised plateau of inner Anatolia, which he says will have favoured ... the existence of a largely pastoral based economy, and relatively light population density. In contrast the sporadic and largely allu­ vial plains of the peripheries will have permitted a more varied cultivation and a denser - more urbanised - population. The basic socio-economic distinction will have been reflected in the fiscal resources available to the state: it is on the whole easier and more profitable to tax a sedentary/agri­ cultural population than an at least potentially moveable/pastoral one'.20

Then he turns to prosopographic or military/aristocratic indications of large landed estates and economic production from them. For him they were located in the central and non-central Anatolian plateau. Now this is contradictory, for he has just described the plateau as primarily produc­ tive in pastoral activity rather than agricultural and so less likely to produce revenue for the state. A second observation is the clear fact that this aristocracy, including the Boilas and the Taronites, spread to the eastern regions as well, where there was also a local semi-Byzantinized aristocracy before the invasions. Further, many of these aristocratic families fled their estates in central and western Anatolia with the disruptive nomadic raids of the eleventh century. Harvey, though a proponent and one of the originators of the theory of an economic expansion in Byzantium from 900 to 1200, differentiates much more sharply than does Hendy as to the extent of this proposed expansion: While the documentary evidence offers glimpses of short-term localised fluc­ tuations, the general trend of expansion applied in the European provinces. There was an important contrast, however, between Europe and Asia Minor, where there was probably a major fluctuation in the demographic trend in the late eleventh and early twelfth centuries owing to the Turkish invasions.

19 Hendy, The Economy', 143. 20 Ibid.; Hendy, Studies, 136.

8

SPEROS VRYONIS JR The upward trend in the western part of Anatolia was resumed in the mid twelfth century and continued in the thirteenth century.21

Leaving aside the problems of the economy in early eleventh-century Asia Minor, it is crystal clear from the texts that the nomadic raids, conquests and settlements in Asia Minor during the second half of the eleventh century and throughout most of the twelfth brought great upheaval both to the unprotected countryside and to many towns, and to talk of economic expansion even in the most expansive sense of the word is to force the body into a suit that does not fit.22 There are many more interesting problems dealing with Byzantine Asia Minor, and even more interesting scholarly propositions in the newer scholarship, but as this is a short illustrative survey of what the new scholarship offers, it is time to say something about the contribution of Armenology and those who, though not Armenologists, have studied the problems of Armenian Anatolia and the relations of Armenians to Byzantium. I shall point to but four of these new contributions which so enrich our understanding of the Armenian component in Byzantine and early Turkish Anatolia. These are: (1) the appearance of a significant portion of Armenians in the ranks of the Byzantine aristocracy, and the degree to which and the rate at which they were absorbed; (2) a nuanced history of the Chalcedonian Armenians as governors and military leaders in some of the absorbed Armenian provinces; (3) the increasing number of translations of Armenian historical sources into Russian or western European languages23 and finally (4) a finer calibrated approach to Armenian phenomena in general in geographical regions, chronological periods, or in cultural manifestations. The first of these, Armenians in the Byzantine aristocracy and the rate at which many of them were Byzantinized, has been examined, most importantly by Kazhdan,24 in a study which grew out of his magisterial earlier volume for the study of Byzantine aristocratic families in Asia Minor.25 Kazhdan has analysed and identified some 388 such individuals among 38 Armenian families with Armenian origins or intermarriage connections, and concluded that these particular families constituted some 10 to 19 per cent of the 300-odd Byzantine aristocratic families. Further, Kazhdan studied the indications of their absorption into various

21 Harvey, Economic Expansion, 245. 22 Vryonis, Decline, 166-7, for charts of the towns and villages that suffered destruction, pillaging, enslavement, massacre, or siege. 23 N. Adontz, Armenia in the Period of Justinian, trans. N. Garsoian (Lisbon, 1970). 24 A. Kazhdan, Armiane v sostave gospodvuiushchego klassa vizantiiskoi imperii v XI-XII vv (Moscow, 1975). 25 A. Kazhdan, Sostav gospodstvuiushchego Klassa v Vizantii XI-XII vv (Moscow, 1974).

THE DECLINE OF MEDIEVAL HELLENISM

9

segments of Byzantine society. The rate at which they were absorbed was established at two to three generations by Jean-Claude Cheynet.26 As for the second area of new research, which concentrates on the Chalcedonian Armenians, that is the Tzatoi or Haik Hrom, Viada Arutiunova-Fidanian in particular has elaborated their socio-historical role in Asia Minor as a bridge between Armenia and Byzantium.27 She takes the province of Taron and its earlier incorporation into the Byzantine Empire as an illustration of her concept of the 'contact zone', which displays the interaction of Byzantine and Armenian societies and the administrative, political and cultural effects on both societies. In an earlier work, she examined the nature, society and political activities of this group both in Byzantine society and during the beginning of the collapse of Byzantine authority in eastern Asia Minor in the later eleventh century.28 In a very crucial area of textual sources, the last twenty-five years have seen enormous progress in the study, and above all in the translation into Russian and western European languages, of a very significant number of Armenian primary sources, a process which opens up their historical problems and treasures for all those of us who have not, alas, mastered Armenian. These translations include the texts of Moses Khorenatsi, Yovhannes Drasxanakertc'i, Aristakes of Lastivert, the History of the House of the Artsrunik, Kirakos of Gandzak, Ghazar P'arpaci, P'austos Buzandic, T'ovma Metsobetsi, the history of Sebeos, as well as the histor­ ical colophons of medieval Armenian manuscripts.29 Finally there is a fourth scholarly process in motion by which scholars have begun to move away from more general works on Armenian histor­ ical phenomena. Here one might mention two things. First, Richard Hovannisian of UCLA has inaugurated a number of scholarly conferences that have as their goal the systematic diachronic and synchronic exami­ nation of the traditional Armenian lands, one by one. The proceedings of the first such conference, which centred upon the Armenian province of Taron, is presently in the press. Second, a valuable publication exploring relations between Armenia and Byzantium has been edited by Nina Garsoian.30 The volume presents new ideas on the nature of the relation 26 J.-C. Cheynet, 'Du pronom au patronyme: les étrangers à Byzance (Xle au Xlle siècles)', in N. Oikonomides, ed., Studies in Byzantine Sigillography (Washington, DC, 1987), 57-66. 27 V. Arutiunova-Fidanian, Armiano-Vizantiiskaia kontaknaia zona (X-XI w ) (Moscow, 1994). 28 V. Arutiunova-Fidanian, Armiano-khalkidonity na vostochnykh granitsakh Vizantiiskoi Impeni (XI v) (Moscow, 1980). 29 Details of Armenian histories, translations, and secondary literature published up to 1992 may be found in R.W. Thomson, A Bibliography of Classical Armenian Literature to 1500 AD (Turnhout, 1995). 30 N. Garsoian, ed., L'Arménie et Byzance: Histoire et Culture (Paris, 1996).

10

SPEROS VRYONIS JR

between Armenia and Byzantium in a much more analytical manner that proceeds by breaking up larger themes and subjects and examining them step by step. It demonstrates the present process of examining of Byzantino-Armenica in Asia Minor with greater specificity. The third area of new scholarly production relevant here is that which deals with the Turkish and Muslim Anatolians from the eleventh through the sixteenth century, an area to which modern Turkish scholarship has dedicated the greatest attention, but to which western and Russian schol­ arship has also paid significant, though not so extensive, regard. The successors of Mehmet Koprulii, in the first instance, and of Ismail Uzunqar^ili in the second, have executed the basic research and work in establishing the fields of Seljuq, Beylik, and Ottoman history, and the history of religion during these periods as a serious, scholarly endeav­ our.31 Among the most prominent of these are: Mehmet Koymen32 and Osman Turan on the history of the Great Seljuqs and the Seljuqs of Rum respectively;33 Oman Barkan on Ottoman defterology;34 Abdul Baki Golpinarli, who laid the foundation for the study of the mystic orders of the Mevlevis and Bektashis and therefore for the mystical tradition in Asia Minor, inclusive of Yunus Emre;35 Tayyib Gokbilgin on Ottoman history and on the nomadic tribes settled in the Balkans;36 and finally 31 M. Kòprùlù, Alcune osservazioni intorno all'influenza delle istituzioni bizantine sulle isti­ tuzioni ottomane (Rome 1953); idem, Islam in Anatolia after the Turkish Invasion, trans. G. Leiser (Salt Lake City, 1993); idem, The Origins of the Ottoman Empire, trans. G. Leiser (Albany, 1992); idem, Turk edebiyatinda ilk mutasavviflar (Ankara, 1966). I. Uzunqar§ili, Osmanli Tarihi 8 vols (Ankara, 1972-83); idem, Osmanli devieti teqkilatindan: Kapukulu Ocaklan, 2 vols (Ankara, 1943-44); idem, Osmanli devletinin ilmiye teqkilati (Ankara, 1965); id., Osmanli devieti teqkilatina medhal: Biiyuk Selqukliler, Anadolu Selquklileri, Anadolu Beylikleri, Ilhaniler, Karakoyunlu ve Akkoyunlularla memluklerdeki devlet teqkilaltina bir bakiq (Ankara, 1970). 32 M. Koymen, Alp Arslan zamam. Selquklu asker (Ankara, 1967); idem, Alp Arlsan zamam. Selquklu kultur muesseseleri (Ankara, 1975); idem, Alp Arlsan zamam. Selquklu saray teqkilati ve hayati (Ankara, 1966). 33 O. Turan, Selquklular zamanda Tiirkiye. Siyasi tarih, Alp Arslan'dan Osman Gazi'ye (1071-1318) (Istanbul, 1971); his edition and exigesis of the three rare Seljuq waqf documents are of first-rate importance: Selquk Devri Vakfiyeleri: I: '§emseddin Altun-Aba, vakfiyesi ve hayati', Belleten 11 (1947), 197-235; II: 'Mùbarizeddin Er-Toku§ ve vakfiyesi', Belleten 11 (1947), 415-29; III: 'Celaleddin Karatay vakiflan ve vakfiyeleri', Belleten 12 (1948), 17-171. 34 O. Barkan, XV ve XVI asilarda osmanli imparatorlugunda zirai ekonominin hukuki mali esaslan. I Kanunlar (Istanbul, 1945); idem, 'Essai sur les donneés statistiques des registres de recensement dans l'empire ottoman aux XVe et XVIe siècles', Journal of the Economic and Social History of the Orient 1 (1958), 7-36; idem, Suleymaniye Carni ve imaret inqaati (1550-1557) 2 vols (Ankara, 1972-79). 35 A.B. Golpinarli, Meviana Celaleddin. Hayat,felsefesi, eserleri, eserlerinden seqmeler (Istanbul, 1959); idem, Mevlanadan soma Mevlevilik, 2nd edn (Istanbul, 1983); idem, Yunus Emre Hayati (Istanbul, 1936); idem, Menakib-i Hunkiar Had Bektas-i Veli (Istanbul, 1956). 36 T. Gokbilgin, XV-XVI asirlarda Edirne ve Paqa Livasi vakiflar-miilkler-mukataalar (Istanbul, 1952); idem, Rumeli'de Yuriikler, Tatarlar ve Evlad-i Fatihan (Istanbul, 1957).

THE DECLINE OF MEDIEVAL HELLENISM

11

Halil Inalcik, active for the better part of six decades over the entire gamut of Ottoman history and institutions, from the transformation of the Beyliks into the kernels of sedentary governments and the rise of the Ottomans down to the end of the nineteenth century.37 Most of the new scholarship on Turkish and Muslim Anatolia, not available at the time of the publication of The Decline of Medieval Hellenism, was published by this second generation of Turkish scholars. The third Turkish scholarly generation has profited from all this and is in turn adding its own accomplishments. The first major lacuna before 1971 was in the realm of the Turkish mili­ tary conquests, in some cases re-conquests, of Anatolia over the 400-year period covered by the Konya sultanate, the numerous Beyliks and the Ottomans. An enormous step forward was made by Turan, based on a masterful coverage of the Persian, Arab and Turkish sources, which rendered Claude Cahen's Pre-Ottoman Turkey useless for any scholar who can read the 700 dense pages of the Turkish text. Turan's work had been preceded by a long line of smaller individual studies and the publications of crucial Arabic and Persian texts dealing with both the Seljuq waqfs and the contemporary court historiography. Of particular importance to the contents of The Decline of Medieval Hellenism are the works on Islamic mysticism, the mystic saints and their various dervish orders, all of which played such an important part in the lives of Muslim, Christian and Jewish Anatolians. The older works of Golpinarli have been re-published, often expanded, and they remain an inexhaustible treasury for students of Anatolian phenomena and in particular on the subjects of Djalal al-Din Rumi and the Mevlevi dervish order. The publication of the second edition of the hagiographical work of Eflaki presents scholars with the single richest work for the cultural and prosopographical history of Anatolia in the latter thirteenth and early fourteenth centuries.38 Of comparable historical significance is the publication of Elvan Qelebi's text, which is rich in the heterodox dervish traditions that go back to Baba Ilyas and the revolt of the tribal Babais.39 Somewhere between epic and dervish hagiography is AbuT-Hayr-Rumi's text.40 Indeed, the brunt of research on Anatolian dervish Islam has shifted from the great 37 H. Inalcik, Fatih devre uzerinde tetkiler ve vesikalar (Ankara, 1954); idem, 'Osmanli-lar'da raiyyet rusumu', Belleten, 23 (1959), 575-610; idem, 'Istanbul', in El2 4, 224-48. 38 T. Yazici, §ams al-Din Ahmad al-Aflaki al-Arifi Manakib al-'Arifin, 2 vols (Ankara, 1976-80). He had previously published a two-volume translation of the Persian text into modern Turkish, Ariflerin Menkibeleri (Istanbul, 1973). 39 I. Eriinsal, Y. Ocak, Elvan Celebi, Menakibu'l kudsiyye fi mensaibi'l iinsiyye (Baba llyas-i Horasani ve sulalesinin menkabevi tarihi) (Istanbul, 1984). 40 §. Akalin, ed., Ebul-Hayr-i Rumi, Saltuk-name, 2 vols (Istanbul 1988-89).

12

SPEROS VRYONIS JR

emphasis which Golpinarli had given to the more urban and Persianoriented circle of Rumi and the Mevlevis to these heterodox orders that were dominant in the rural areas occupied by recent Turkmen and Mongol arrivals in Anatolia. European scholarship continues to focus on Rumi and the Mevlevis in a limited productivity which, however, boasts the extraordinary works of Anna Marie Schimmel.41 Returning to the more rural mystical orders, the importance of the cryp­ toshaman, but Islamized Turkmen Baba was long ago identified. The systematic and convincing development of this theme in Seljuq and Beylik Anatolia has been the fruit of the labours of a small third genera­ tion of Turkish scholars. The role in the Islamization and conquests of Asia Minor had been pointed to in The Decline of Medieval Hellenism, but it was incomplete, given the lack of editions of many of the texts, as well as the lack of the proper analysis of the entire body of the menaqibnames as a genre with their agendas, myths and mythopoiesis. Even the taxonomy of the silsiles of the sheikhs was not secure. It is to these Turkmen shaman-Islamic Babas that Ahmet Ya§ar Ocak has turned, and to whom we are indebted for his major breakthrough as regards rural dervish heterodoxy in Asia Minor.42 The research of Ocak about the Babai revolt of 1240-41, as well as the study of a principal source for it, the Menaqibname of Elvan (Jelebi, began in the 1970s. His concentration on these two subjects has cleared the ground of fundamen­ tal confusion and error, and has led to a proper identification of who the leader of the revolt, formerly known as Baba Rasul, was. Ocak's analysis of the spiritual domain in both the Babai revolt and Elvan Qelebi has enabled him to confirm the presence of elements from the central Asiatic shaman tribal tradition of popular religiosity and lead­ ership. Thus the rural village and/or nomadic heterodox popular religion retains a vital tribal element under its diaphanous Islamic veil during the long evolution and reformation of the religious colouration of Anatolia.43 Ocak's works enable scholars to take a more even-handed approach to the rural in contrast to the urban orders; further it underlines and demon­ strates the liveliness of the central Asiatic religious substratum in these rural orders, and restores to their accomplishments the importance they 41 In her outpouring on Rumi one should note first and foremost the brilliant literary analysis of her The Triumphal Sun. A Study of the Works of Jalaloddin Rumi (London, 1978). Also, Mystical Dimensions of Islam (Chapel Hill, NC, 1975); idem, I Am Wind You are Fire. The Life and Work of Rumi (Boston and London, 1992). 42 A.Y. Ocak, Babailer Isyam, Aleviligin tarihsel altyapisi. Yakut Anadolu Islam-Tiirk heterodoxinin teqekulii (Istanbul, 1989; 2nd edn 1996). 43 On the typology of this heterodox saint and hagiography see his Turk halk inanqlannda ve edebiyatinda evliya menkabeleri (Ankara, 1984). Also his Kiiltur tarihi kaynagi olarak menkibnameler. Metolojik bir yaklamaqim (Ankara, 1997).

THE DECLINE OF MEDIEVAL HELLENISM

13

deserve in our considerations of the manner by which Islamization was effected in Asia Minor, from Christianity to Islam, and to what forms of Islam. Another event, certainly the most sensational in the raids, invasions and settlements of the Turks in Asia Minor, has brought new research to the fore. The publication of Claude Cahen's article on Manzikert estab­ lished itself in the realms of Byzantine, Turkish, Islamic and Armenian studies as the authoritative reconstruction of the events leading up to, during, and after the battle.44 In my dissertation and again later in the book, I demonstrated that Cahen had rejected the only surviving eyewit­ ness account of the battle (that of Michael Attaleiates). I clearly pointed to the weakness and indeed the untenability of Cahen's evaluation of the primary sources. This prompted a much more expanded and detailed evaluation of all 27 sources for the battle, a project that I carried out over many years. Crucial for the study of the battle was the decision of Turkish scholars to join in the national celebration of its 900th anniversary. There was an outpouring of articles and books on the battle itself, bibliographies and proceedings of conferences, there was a military exercise which in some way attempted to reconstruct the battle itself, and even a dramatic play was commissioned for the occasion.45 Of all this, perhaps the single most valuable production was the reprinting of the basic Islamic chronicle texts accompanied by a short but critical preface and the Turkish translations of the nine Arabic and four Persian accounts.46 Aside from the very convenience of bringing these texts together in one volume, the other important contribution of this volume was the preface, which for the first time described analytically the unreliability of most of this body of materials, a fact which Cahen had ignored when he vaunted the value of the Islamic sources. Years later, I once more turned to finishing the translation of the four Persian sources on the battle and then began the writing of a comprehen­ sive work on the battle of Manzikert. The first part proceeded to a system­ atic analysis of the following linguistic groups of medieval chronicles /histories as linguistic entities: Greek, Arabic, Persian, Ottoman, Armenian, Syriac, Latin, plus two translations into demotic 44 C. Cahen, 'La campagne de Mantzikert d'après les sources musulmanes', Byz 9 (1934), 613-42. 45 For the dramatization of the event in modern Greek and Turkish literature: K. Kyriazes, Romanos ofTetartos Diogenes (1982; repr. 1991); M. Faruk Gürtünca, Büyük hakan Alp Arslan. 4 perdelik manzum piyes. Malazgirt zafiri'nin 900 Yildonumu Münasebetiyle EnstitU tarafindan açilan yanimada kazanan eserler No.l (Istanbul, 1971). 46 F. Siimer and A. Sevim, Islam kaynaklarma gore Malazgirt savait (metinler ve çevirileri) (Ankara, 1971). See also the chapter by C. Hillenbrand in this volume (Chapter 5).

14

SPEROS VRYONIS JR

Greek and Bulgaro-Slavic. Simultaneously, I published three articles which attempted to present some of the results of this research. The first established a hierarchy of historical reliability among the six Greek, and then among the eight Arabic, sources, and then evaluated each of these two linguistic groups comparatively.47 The end result was to adjudge the Greek sources as far superior to those in Arabic, and among the Greek, to adjudge the long and coherent account of Attaleiates not only the most reliable, but also probably the only reliable source for the battle itself. The second concern of this study was to see if these two separate linguistic bodies of writings overlapped in any section so that one confirmed the other. The only real overlap, in great detail, had to do with the eight day captivity of Romanos in the camp of the sultan Alp Arlsan. Indeed, in relation to this episode the overlapping is almost complete as to structure, and the supposed philosophical conversation between the emperor and the sultan is almost identical. Finally, the Arab source is more detailed than the Greek. This raised the question of what might be the official sources behind the Greek and the Arab testimonies. In a second study, I examined the problem of establishing at what point modern scholarship began to approach the history of the battle on serious methodological grounds, with some indirect contact with the main Greek source, Attaleiates, and on the fragment of the Arab Seljuq historian Ghars an-Nime, again through indirect contact via much later plagiarizing sources.48 The key was the use of these Greek and Arab plagiarizing sources by both LeBeau and Gibbon in the late eighteenth century. The main thrust of this second article is a hierarchical analysis of the Persian, Armenian, Syriac and western sources. The end result of this investigation is that mythopoiesis already observed in the Arabic chronicles, becomes much more literary and fantastic in the Persian chronicles, and much more provincial and ethnic in the Armenian, Syriac (with the exception of Bar Hebraeus) and Norman productions. A third article, to appear soon, returns to the captivity of Romanos IV, in which the Greek and Arabic sources seem to coincide in the most remarkable manner.49 The article concludes:

47 S. Vryonis, The Greek and Arabic Sources on the Battle of Mantzikert', in Vryonis, ed., Byzantine Studies, 25-140. 48 S. Vryonis, 'A Personal History of the History of the Battle of Mantzikert', in Oikonomides, ed., Byzantine Asia Minor, 225^14. 49 The Greek and Arabic sources on the Eight Day Captivity of the Emperor Romanos IV' (forthcoming). A fourth study will also appear this year, entitled 'Michael Attaleiates, Michael Psellos and the Blinding of the Emperor Romanos IV'.

THE DECLINE OF MEDIEVAL HELLENISM

15

it is in this ... aspect... that the Arab sources are more detailed and richer ... Thus we must pay special attention to the account of Sibt Ibn al-Jawzi, which is plagiarised from Ghars an-Nime, for an understanding of what transpired in the camp of the sultan and in the exchanges and negotiations of the two monarchs. At the same time it is only Attaleiates (and the Greek sources dependent on him) who tell us how long this encounter lasted: 8 days. While the project of this history of the battle of Manzikert continues, I am also undertaking research into the social and mystic world of thir­ teenth- and fourteenth-century Anatolia as reflected in the writings of Eflaki (1286-81, d. 1360).50

50 Three short articles have, to date, appeared: The Muslim Family in 13th—14th Century Anatolia as Reflected in the Writings of the Mawlawi Dervish Eflaki', in E. Zachariadou, ed., The Ottoman Emirate (1300-1389) (Rethymnon, 1993), 213-223; 'Man's Immediate Ambiance in the Mystical World of Eflaki, the Mawlawi Dervish, Qonya (1286-1291, d. 1360', Symmeikta 9 (1994), 365-377; The Political World of the Mevlevi Dervish Order in Asia Minor (13th-14th Century) as Reflected in the Mystical Writings of Eflaki', in Philhellen: Studies in Honour of Robert Browning (Venice, 1996), 411^119. A fourth study is to appear this year, in Turkey: The Economic and Social Worlds of Anatolia in the Writings of the Mawlawi Dervish Eflaki'.

This page has been left blank intentionally

Section I Byzantium's eastern frontier

This page has been left blank intentionally

2. Constantine VII, Caucasian openings and the road to Aleppo Jonathan Shepard It is widely held that the destruction of the invading army of the amir of Melitene in 863 marked something of a turning-point in Byzantium's position vis-à-vis its eastern Muslim neighbours. In the words of Georg Ostrogorsky, 'the tide turned and there began the era of Byzantine attack in Asia, an offensive which opened slowly but which for the second half of the tenth century moved forwards with ever increasing celerity'.1 And yet, as Ostrogorsky himself implies, substantial territorial gains were only made almost a century after the Battle of Bishop's Meadow, and this raises the question why 'the systematic advance ... on [Byzantium's] eastern frontiers'2 took so long. It was only in the mid-920s that a series of initiatives which have been seen as beginning 'the epoch of conquest' was launched. According to one authority, the emperors, 'supported by the substantial resources of an extensive empire, were equipped to pursue, after 926, a project of reconquest of eastern Anatolia, of Cilicia and of northern Syria'; the prime target of the earliest offensives, Melitene, was taken over directly, in 934, and then, according to another scholar, 'Byzantium's eastward drive was resumed'.3 That a 'new spirit' of confidence4 infused the imperial élite in the mid920s is undeniable. The oration celebrating the peace treaty with the Bulgarians of 927 shows clear awareness in court circles of the repercus­ sions which security in the west will have on the empire's position in the east: 'all things are made new and sparkling ... Only the sons of Hagar

1 G. Ostrogorsky, A History of the Byzantine State (Oxford, 1968), 227. 2 Ostrogorsky, Byzantine State, 237. 3 EJ2 IX, s.v. Sayf, 107 (Bianquis); J.D. Howard-Johnston, 'Crown Lands and the Defence of Imperial Authority in the Tenth and Eleventh Centuries', ByzF 21 (1995), 86; R. Jenkins, Byzantium. The Imperial Centuries AD 610-1071 (London, 1966), 245. 4 S. Runciman, The Emperor Romanus Lecapenus and His Reign (repr. Cambridge, 1963), 135. From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

19

20

JONATHAN SHEPARD

mourn and shall mourn, who are bereft of heart at the mere echo of our concord'.5 Yet while accepting that Byzantine decision-makers had many other calls on their time and resources besides those from the eastern approaches in the late ninth and early tenth centuries, one may legiti­ mately question whether they had a clear-cut concept of 'systematic advance' in the first place, and also whether these conceived of the eastern borders as being susceptible to uniform treatment: they did, after all, sprawl for over 700 kilometres, taking in a variety of climates, creeds and forms of settlement. More fundamentally, one may wonder whether many persons in imperial circles had clearly formulated ambitions of - or positive interest in - widespread territorial expansion even in the later 920s. One should beware of reading back a consistent policy of territorial expansion on a broad front from the events of the later 950s and 960s, when massive armed forces were deployed, many strongholds and forti­ fied population centres together with substantial tracts of land were subjugated, and arrangements for maintaining such acquisitions had perforce to be made. This is not to rule out the existence of some sort of conscious 'policy' and sets of 'policy-options' at the imperial court. Nor is it to deny that a series of military initiatives was launched upon the decision of the central government from the mid-920s onwards, of a frequency, range and panache which had few precedents in the middle Byzantine era. But it is to question whether these campaigns amounted to much more than the safeguarding of Byzantium's eastern approaches and reducing the risk of further damaging incursions from the side of the Muslims. They were essentially reactive or pre-emptive, intended to take out troublesome thorns in the flesh such as Melitene and Theodosioupolis6, to break the power of particular amirs or communities and above all to bring about stability in the borderlands by means of, in effect, neutralization or demil­ itarization. There was not, I suggest, much imperial hunger for direct annexation of lands as such. Interest in acquisition of key strongholds or fortified towns was rather keener, but this was on a highly selective basis. It did not necessarily involve full-scale occupation by a wholly 'Roman' garrison or the maintenance of a linear frontier. By the same token, there was a greater predisposition towards loose hegemony and the forging of personal connections between the emperor and individual leaders and other notables in the border regions. From this perspective, the longrange strikes of Domestic of the Schools John Kourkouas (and earlier generals) far beyond the enemy's forward bases should be seen primarily 5 I. Dujcev, 'On the Treaty of 927 with the Bulgarians', DOP 32 (1978), 280-81. See also A.A. Vasiliev, Byzance et les Arabes II/l (Brussels, 1968), 261. 6 On the importance of Melitene, see J.-Cl. Cheynet's chapter in this volume (Chapter 4).

THE ROAD TO ALEPPO

21

as adjuncts to diplomacy, ways of 'showing the flag' and instilling in farflung élites some respect for the emperor and for those other potentates apparently enjoying his favour. This is not to deny that, besides serving as ancillary to 'diplomatic' measures, such strikes could also perform a straightforwardly military function - humiliating or breaking the power of a persistently belligerent commander, or serving to distract an invad­ ing force and induce it to return to defend its own country, on the lines recommended in Skirmishing.7 But if stability and de facto neutralization were at a premium, tribute and revenues from lucrative centres of commerce and population were of more concern than extensive tracts of land. Such a measured approach towards territorial annexation was not confined to eastern affairs. The quite limited size of middle Byzantine armed forces relative to the number of potential trouble spots is both a reflection and a cause of the basic cautiousness of the emperors' stance.8 Their caution reflected awareness of what the empire's resources and administration (for all the rhetoric of world leadership) could actually afford. But it also sprang from apprehensions about the proclivities of sizeable contingents of well-trained 'professional' soldiers, once raised. John Haldon has suggested that one reason why only modest numbers of full-time troops were stationed at Constantinople through the early and middle periods was fear of coups d'état. Episodes such as the mustering of soldiers not far from the capital to confront Symeon of Bulgaria in 913 and Domestic of the Schools Constantine Doukas's subsequent bid for the throne gave substance to such fears.9 I would suggest that similar appre­ hensions as to loyalties partly explain the dispersal of full-time troops in what were usually smallish units in the provinces, and that they had some bearing on general funding arrangements and tactics in the border­ lands.10 This was a systemic problem, neither peculiar to any one borderzone nor necessarily the product of agglomerations of family groupings 7 Skirmishing, 20, ed. and tr. G. T. Dennis, Three Byzantine Military Treatises, CFHB 25 (Washington DC, 1985), 218-23. 8 The scale of locally and centrally maintained armed forces and the means of maintain­ ing them are discussed by M. Whittow, The Making of Orthodox Byzantium 600-1025 (London, 1996), 183-93; J. Haldon, 'Military Service, Military Lands and the Status of Soldiers: Current Problems and Interpretations', DOP 45 (1993), 13-20,44-7 and n. I l l , 66; J. Haldon, Warfare, State and Society in the Byzantine World, 565-1204 (London, 1999), 78,101-6 and n. 68 on 314, 115-28. 9 J. Haldon, 'Strategies of Defence, Problems of Security: the Garrisons of Constantinople in the Middle Byzantine Period', in G. Dagron and C. Mango, eds, Constantinople and its Hinterland, SPBS 3 (Aldershot, 1995), 149-52; Theoph. Cont., VI.2-3, 381-3; J. Shepard, 'Symeon of Bulgaria - Peacemaker', Annuaire de l'Université de Sofia 'St. Kliment Ohridskï, Centre de Recherches Slavo-Byzantines 'Ivan Dujcev' 8 3 /3 (1989) [1991], 20. 10 The underlying assumption of Skirmishing is that a commander must fend for himself,

22

JONATHAN SHEPARD

in certain border regions. Military coups could be launched in and around Constantinople as well as from the borders. This fact of political life does much to explain why Bulgarian power, once orchestrated by a capable Christian ruler, so unnerved imperial decision-makers. In times of tension with Symeon, too many troops needed to be stationed too 'close to home', while senior commanders such as Constantine Doukas already had legit­ imate access to the Great Palace. These considerations throw some light on the consistency as well as the vacillations in the line taken by successive regimes towards the eastern approaches during the first half of the tenth century. As to the nature of that line, three propositions and a paradox will be offered here. First, it was primarily to Armenian-born potentates and adventurers and to Caucasian strong-points that Byzantine statesmen looked for an improve­ ment in their strategic position in the east, not to the lands beyond the Taurus Mountains or in the Middle Euphrates basin or Syria. And even there they were highly selective, focusing on the north-eastern border­ lands of 'Iberia' (Georgia), the strongholds along the northern and south­ ern branches of the Upper Euphrates and other river valleys running in a generally east-west direction, and on the castles and castle-holders strad­ dling the mountainous region of Taron, between the existing borders and Lake Van. Much depended on gaining the active co-operation of local, nominally 'subordinate', notables. Second, the aim of these initiatives was essentially defensive, to halt the raids launched from Muslim foreposts and to bar access to Muslim raiders through gaining control of a few choke-points. The creation of themes such as Mesopotamia and Chozanon between the two branches of the Euphrates11 and the offen­ sives against bases such as Melitene were primarily means to the same end, rather than conceived as a curtain-raiser to further expansion. Third, while there were economic, religious and cultural attractions to prompt intervention in the Caucasian lands, the underlying reason for prioritiz­ without counting on rapid reinforcement: Skirmishing, 4:156-9; 5:160-1; 7:162-3; 11:182-3; 12: 186-7; 16: 202-3; 17: 204-9; 19: 214-15; 20: 218-19; J. Haldon and H. Kennedy, The Arab-Byzantine Frontier in the Eighth and Ninth Centuries: Military Organisation and Society in the Borderlands', ZRVI 19 (1980), 101-5; Whittow, Orthodox Byzantium, 178; Haldon, Warfare, 105, 112-15, 117. See also Cheynet's chapter in this volume. 11 Constantine VII Porphyrogenitus, De administrando imperio, 50 [henceforth = DAÍl, ed. and trans. G. Moravcsik and R.J.H. Jenkins, CFHB 1 (Washington, 1967), 238-9; N. Oikonomidès, Les listes de préséance byzantines des IX et X siècles (Paris, 1972), 247, line 11; 267, line 16; 349, 359 (commentary); N. Oikonomidès, 'L'organisation de la frontière orientale de Byzance aux Xe-X Ie siècles et le Taktikon de l'Escorial', Actes du XlVe congrès international des études byzantines I (Bucharest, 1974), 235-7, repr. in Documents et études sur les institutions de Byzance (VIIe-XVe s.) (London, 1976), no. 24; J.D. Howard-Johnston, 'Byzantine Anzitene', in S. Mitchell, ed., Armies and Frontiers in Roman and Byzantine Anatolia, BAR International Series 156 (Oxford, 1983), 240, 256-7.

THE ROAD TO ALEPPO

23

ing them was a negative one: maximum gains of security, prestige and resources could be made by the emperor acting directly for himself, with minimal recourse to intermediaries or sizeable forces of Byzantine-raised and -salaried troops. Finally, the paradox: although ties were forged with potentates, and a number of key enemy attack-bases were acquired or dismantled, the most extensive actual gains of the mid-tenth century were made elsewhere, in Cilicia and the Middle Euphrates basin, and not in Caucasia. If the aforesaid three points were valid, this contrast would seem to be somewhat anomalous. I suggest that the explanation lies less in long-hatched plans for blanket expansion eastwards than in the persis­ tent spirit of militancy shown by a particular border amir, Sayf ad-Daula. This prompted Constantine VII to marshal and train unprecedentedly large forces and send them on the road to Aleppo, the amir's new-found power-base. Of the three points, the first has been mooted by other scholars,12 and to belabour it further may seem superfluous. But I do not think that the emperors' focus on Caucasia has been considered sufficiently in light of the other two propositions or that the selective nature of their interest has been elaborated upon. The fact that a great deal of the campaigning led by John Kourkouas in the later 920s and the 930s was carried out across the Armenian lands is well known, but it has not been linked with the inten­ sity of the coverage that certain western and southern regions of Armenia receive in the De administrando imperio (henceforth DAI). Constantine begins the section devoted to them thus: 'it is right that you [i.e. his son Romanos II] should not be ignorant of the parts towards the rising sun, for what reason they became once more subject to the Romans, after they had first fallen away from their control'.13 One should first note that, judging by the scope of what follows, Constantine regards the 'parts' towards the east as consisting essentially of the Caucasian lands. There is no inkling that the other regions still under Muslim control further south were likewise to be regained. It is equally noteworthy that Constantine acknowledges a certain continuity of policy between himself and his predecessors in relation to the eastern parts, whereas he repudiates Romanos Lekapenos's policy towards northerners such as the Bulgarians.14 His indications that Leo VI and Romanos Lekapenos alike had taken an interest in a number of Armenian individuals, leading fami­ lies and strategic points correspond fairly closely with the main thrust of 12 Notably by Whittow, Orthodox Byzantium, 315-18. In territories further south, once wrested from Muslim amirs later in the tenth century, considerable recourse to local power structures, or at least to officials of local origin, is discernible; see C. Holmes' Chapter 3 in this volume. 13 DAI, 43: 188-9. 14 DAI, 13: 72-5.

24

JONATHAN SHEPARD

Kourkouas's incursions of the later 920s and 930s: along the valleys of the northern and southern branches of the Upper Euphrates and the Upper Araxes towards power-bases such as Dvin and the strongholds round Lake Van.15 Above all, the sheer unusualness of the fact that Constantine purveys detailed, fairly accurate and up-to-date material on districts such as Taron and the environs of Lake Van needs to be taken into account. Admittedly, the treatment is not without its lacunae, and even on Taron there is a certain imbalance.16 But, as R.J.H. Jenkins noted, this material stands out as being more coherent and containing more details relevant to current affairs in the mid-tenth century than almost any other section.17 This should surely be connected with the fact that the four 'Caucasian chap­ ters' - unlike any other section of the DAI - amount to a kind of 'programme' for future initiatives to be carried out when opportunities arose.18 The 'Caucasian chapters' all deal to a greater or lesser extent with the control of kastra, as being key to predominance in a region, an assumption shared by contemporary Armenian chronicles. This is made explicit by Constantine in the case of Artanuj: together with its hinterland, it 'is a key to Iberia and Abasgia and the Mischians'.19 Kastra did, of course, vary among themselves. Artanuj, reportedly possessing extensive suburbs Tike a provincial city' and yielding 'immense customs revenues' from its wide-ranging trading connections, was a different proposition from some of the smaller fortified towns in the vicinity of Lake Van that are the subject of chapter 44. And a few of the kastra coming into the emperor's sights lay well to the east, for example Dvin, long a symbol of Muslim hegemony over the Armenians.20 But there seems to be a consis­ tent pattern of selection at work: Constantine is particularly interested in places that are sources of disturbance, whether located in the western

15 Vasiliev, Byzance et les Arabes I I /1, 261^1, 266-73. 16 J. Shepard, 'Imperial Information and Ignorance: a Discrepancy', ByzSlav 56 (1995), 110. See also P. Karlin-Hayter, 'Krikorios de Taron', Actes du XIV congrès international des études byzantines II (Bucharest, 1975), 345-7, 353. 17 DAI, 12 (general introduction); See also R.J.H. Jenkins, ed., De administrando imperio: II, Commentary (London, 1962), 3-5. 18 The distinctive characteristics of the 'Caucasian chapters' (43-6) are a principal ground for the view that the DAI's preceding chapters (14^12) were originally compiled for a differ­ ent kind of work, the hypothetical Peri ethnon: Jenkins, DAI: Comm., 2-5 (Jenkins). See also K.N. Iuzbashian, Armianskie gosudarstva epokhi Bagratidov i Vizantiia IX-XI w . (Moscow, 1988), 128,130. 19 DAI, 46: 216-17. On Artanuj's position astride various trade routes, see H.A. Manandian, The Trade and Cities of Armenia in Relation to Ancient World Trade, trans. N.G. Garsoïan (Lisbon, 1965), 145-6. The significance of kastra is also highlighted by Cheynet's chapter in this volume. 20 Ef2 II, 678-9 (Canard).

THE ROAD TO ALEPPO

25

borderlands, such as Theodosioupolis, or further afield, around Lake Van.21 By the same token he describes in detail key positions from where imperial officials or co-operative local notables could survey and manip­ ulate affairs to the empire's advantage. And he is interested in those strongholds which may block the passage of Muslim raiders, performing a role not unlike that of the kleisourai on the south-eastern approaches of the empire.22 Every one of the 'Caucasian chapters' presents a quasi-legal case for the emperor's rights to a particular stronghold or set of strong­ holds, invoking written records and recent events. One might be inclined to dismiss the DAPs accounts of the exchanges between emperors and local notables over their respective entitlements to such places as Theodosioupolis, the forts of Mastaton and Avnik (both near Theodosioupolis) and 'the country of Apoganem' in Taron as exam­ ples of Porphyrogennetan pedantry and lack of control of source mater­ ial; likewise with its account of the squabbles of members of the leading family of Taron over 'the house of Barbaros' at Constantinople and other possessions.23 But I suggest that a certain rationale underlies such detailed coverage. Even in western and southern Armenia, where some sort of expansion was being contemplated, annexation and direct rule were viewed as merely one possible option: hence Constantine's attention to particular families and individuals whose co-operation might be valu­ able. And if the active support of local dynasts was deemed indispensable to the expansion of imperial influence amongst the eastern approaches, it was all the more desirable that the emperor's actions should be construed as legal. Parading scruples about property rights and formal agreements would not, of course, necessarily convince those against whom the emperor was pressing a claim, but it burnished the empire's image as a haven of order and stability, 'a protective bastion against the enemies' in the (admittedly, self-serving) words of catholicos Yovhannes (John) Drasxanakertc'i.24 The apparent punctiliousness about legal entitlement and written deeds was a means of reassuring local notables not directly involved in the disputes that their property and social positions were safe under the emperor's wing. Equally, as the 'Caucasian chapters' indicate, individual notables in key areas sometimes handed over voluntarily strongholds or their 'country' to the empire.25 Seeing that 21 Vasiliev, Byzance et les Arabes I I /1, 249, 284; El2 II, s.v. Erzurum, 712 (Inalcik). 22 DAT, 50: 238-41; Oikonomides, 'L'organisation', 346-7; Haldon, Warfare, 79,114. 23 DAI, 45: 208-13; 43: 190-97. 24 Yovhannes Drasxanakertc'i, History of Armenia, LIV.38, trans. K.H. Maksoudian (Atlanta, GA, 1987), 193. The Byzantines' invocation of a written testament when eventually annexing the kingdom of Ani in the eleventh century was noted by K.N. Iuzbashian, 'Skilitza o zakhvate aniyskogo tsarstva v 1045 g.', VV 40 (1979), 79-80, 85, 89-90. 25 DAI, 43: 196-7; 46: 216-19.

26

JONATHAN SHEPARD

this might be done peaceably, without much outlay of resources, the emperor had every reason to pose as the champion of deeds of gift and testaments. If the DATs 'Caucasian chapters' imply the empire's reliance on the co­ operation and sometimes the armed forces of the local élites, they also reveal the extent of the concessions needed to secure them. This is evident in Constantine's exposition of his ongoing dispute with the Iberians over their entitlement to Theodosioupolis. Although Theodosioupolis was of outstanding strategic significance and the target of direct attacks by the Byzantines - including a seven-month siege by Kourkouas - both Romanos Lekapenos and Constantine VII were apparently ready to contemplate its capture and occupation by locals, in this case the Iberian kuropalates, Ashot II Bagratid and certain members of his family. It is the legal tangle that the successive emperors' promises engendered that takes up much of chapter 45. Constantine discloses without comment that he had been willing to make even more sweeping concessions to Ashot than Romanos had been: Constantine had issued a chrysobull conceding that Ashot could hold 'in mastery and lordship' 'all the places of the Agarenes which both he and his nephew ... may be able by their own power to reduce'.26 In the event the Byzantines managed to seize Theodosioupolis for themselves in 94927 and Constantine is, through his explication in the DAT, laying down the line that Ashot has contributed nothing to its capture and thus has no right to invoke the terms of the chrysobull. But in so doing he lets slip how frequently recourse to local co-operation was made both at central and field-commander level. John Kourkouas, as Domestic of the Schools, reportedly handed over the newly captured kastron of Mastaton to Ashot's brother, the magistros Bagrat, after Bagrat had sworn 'a written oath that he would retain it and never give it up to the Saracens'.28 The episode is cited as an instance of Iberian untrustwor­ thiness - Bagrat promptly turned the kastron over to the Saracens - but it also suggests the importance of the local assistance which Kourkouas received. Bagrat had taken part in the campaign against Theodosioupolis and asked for the kastron when Kourkouas was on the point of with­ drawing. It may well be that the offer was readily accepted because the Domestic of the Schools was ill provided to garrison the stronghold with troops of his own. Such an interpretation of Kourkouas's decision is supported by the cautiousness and awareness of the conditional nature of local notables' backing which pervade the 'Caucasian chapters' of the DAI. They are

26 DAI, 45: 210-11. 27 Vasiliev, Byzance et les Arabes I I /1, 318-19. 28 DAI, 45: 212-13.

THE ROAD TO ALEPPO

27

manifest in the tale of Romanos Lekapenos's acceptance of the offer of Ashot Kiskasis to hand over Artanuj and hasty despatch of a small élite corps to take possession of it. Not only did this draw the emperor into a dispute between Ashot and his son-in-law, Gurgen; it also united Gurgen with the cousins against whom he had been at odds in their indignation over Romanos's intervention.29 They wrote, threatening to 'put off our state of submissiveness to your ... majesty and make common cause with the Saracens', since they would be obliged to mobilize not only to regain Artanuj but also 'against Romania itself'.30 Romanos is depicted as being 'terrified' at the prospect of these notables joining forces with the Saracens, and at once backed down. Whether or not Constantine was trying to pin all the blame for the fiasco on Romanos, he does not belittle Romanos's fears as to the potential seriousness of a Georgian-Muslim alliance against the Byzantines. The likelihood of this materializing was clearly slighter around the time of writing than it had been in the 920s, when, probably, the bid for Artanuj was made. Even so, the episode is related as a kind of cautionary tale, warning how attempts at outright annexation, even at the invitation of some local notables, could be counter-productive. Similar considerations, a sense of the need to cater for the sensibilities and inter­ ests of the locals, inform Constantine's recommendations about Theodosioupolis. He has just demonstrated at length how the kuropalates Ashot has done nothing to deserve any of the strongholds recently regained there, and in 'strict justice' has no right to control on either side of the Upper Araxes. Constantine none-the-less recommends that the river be the dividing-line between what is directly 'beneath our imperial majesty' and the possessions of the Iberians. He invokes the opinion of John Kourkouas 'of blessed memory' but he also notes that the kuropalates himself had proposed the Araxes for the borderline.31 Here as elsewhere in the western Caucasian lands, he intimates, it was prudent to keep the locals happy, while taking care to retain the legal and moral 'high ground'. In the light of successive emperors' care to maintain a patina of legality and need for at least some local co-operation, one might turn to the ques­ tion of the emperors' recognition of the title of 'prince of princes' and of the title's significance after Ashot I Bagratuni of Armenia received the crown of a 'king' from the caliph during the 880s.32 But here it may suffice to note that different emperors responded to the opportunities presented by this concession of the caliphate in different ways, according to circum­ 29 DAI, 46: 219-20. On these individuals, see C. Toumanoff, Studies in Christian Caucasian History (Georgetown, 1963), 495-6. 30 DAI, 46: 220-21. 31 DAIf 45: 212-13. 32 DAI: Comm., 158-9 (Runciman); Iuzbashian, Armianskie gosudarstva, 67-71; A.E. Redgate, The Armenians (Oxford, 1998), 174-5.

28

JONATHAN SHEPARD

stances. Imperial recognition of a prince as 'prince of princes', which could apparently be accompanied by despatch of a crown,33 did not betoken plans for formal annexation. Bestowal or withdrawal of the title served rather as a bargaining chip,34 and also as a quasi-legal device for claims to lordship over those who had not submitted to the emperor directly. The argument, as cited by Constantine VII regarding the Qaisites of the Lake Van region, was that even if they were now refusing to pay the emperor tribute, their predecessors had been under the dominion of the 'prince of princes'; and 'since the prince of princes is the doulos of the emperor of the Romans ... , it is obvious that the fortified towns, town­ ships and territories of which he is lord also belong to the emperor of the Romans'.35 Mention of the Qaisite amirs and Lake Van leads to our second main point: that the exceptional cases where some sort of 'expansion' was consistently sought prove the rule that earlier tenth-century emperors were governed primarily by considerations of the security of 'Romania' as it was. If Constantine VII's explanations as to why the forts and fortified towns around Van should come under imperial dominion sound convo­ luted, their very prolixity attests the importance attached to this particu­ lar area. In respect of certain kastra there - as of nowhere else on any front - he expressly recommends repossession in terms which seem to mean direct occupation rather than a form of indirect rule: 'the emperor should get [them] back, as they are his property'.36 Constantine sketches in part of the background. The Qaisite amirs in control of the kastra had been forced by Kourkouas's raids to pay tribute to the emperor from around 931, but the current head of the clan, 'Apelbart' (Abu'l-Ward) is not paying tribute for them to the emperor.37 This development is connected with an episode not expressly mentioned by Constantine: in 940 Sayf adDaula made one - just conceivably two - incursions into the region and, beside the shore of Lake Van, exacted oaths of submission from leading Armenian princes, as well as from the Qaisite amirs.38 Sayf took over a 33 Kirakos Gandzakets'i, History of the Armenians, trans. R. Bedrosian (New York, 1986), 72; H.C. Evans, 'Kings and Power Bases', in J.-P. Mahé, R.W. Thomson, eds., From Byzantium to Iran. Armenian Studies in Honor of Nina Garsoian (Atlanta GA, 1997), 488. 34 As, for example, when Romanos Lekapenos offered it to an unidentified potentate as an alternative to migration to Byzantium and command of the themes of his choice: J. Darrouzès, L.G. Westerink, Théodore Daphnopatès. Correspondance (Paris, 1978), 54-5. 35 DAI, 44: 200-1. See A. Ter-Ghewondyan, The Arab Emirates in Bagratid Armenia, tr. N.G. Garsoian (Lisbon, 1976), 79-81. 36 DAI, 44: 204-5. 37 M. Canard, Histoire de la dynastie des Ha'mdanides de Jazîra et de Syrie (Paris, 1953), 740^41; DAI: Comm., 169 (Runciman); Ter-Ghewondyan, Arab Emirates, 82. 38 Vasiliev, Byzance et les Arabes I I /1, 285-8; DAI: Comm., 169 (Runciman); TerGhewondyan, Arab Emirates, 84-6; Whittow, Orthodox Byzantium, 319-20.

THE ROAD TO ALEPPO

29

number of strongholds, including Bitlis, from Amir Ahmad.39 Ahmad who, as lord of other fortresses such as Chliat and Arzes had been paying tribute to the emperor, was subsequently killed by AbuT Ward and his fortresses came under AbuT Ward's control. These kastra are singled out as suitable candidates for repossession by Constantine VII, writing a few years later. In other words, Constantine is reacting to recent setbacks, in which Byzantine arrangements for indirect rule and tribute-payments had unravelled and the long-term solution now appeared to be direct occupation of the forts. There is therefore no reason to doubt the rationale given for this course of action, an essentially defensive one: 'If the emperor holds these three kastra, Chliat, Arzes and Perkri, a Persian [i.e. Muslim] army cannot come out against Romania, since they are between Romania and Armenia and serve as a barrier'.40 If Sayf ad-Daula's intervention in southern Armenia and its repercus­ sions prompted Constantine's specific recommendations, a general fear of incursions from the central lands of the caliphate accounts for the stand­ ing interest in the Lake Van region that successive emperors showed. As Mark Whittow pointed out, there were few traversable passes north from the central lands and the most convenient of these, the Bitlis pass, lay not far from Lake Van.41 One may connect with this fact of geo-politics both the DAI's objective of establishing a direct presence around Lake Van and its interest in the leading family of the intervening region, Taron. As the case of the Qaisite amirs suggests, direct occupation was a last resort, when looser tributary arrangements with local potentates seemed to have broken down. The store set by such arrangements is suggested by the DAI's genealogies of the Qaisite clan and specifications as to which amir held which kastron, data presumably amassed over the years.42 A somewhat different instance of a nodal point on the eastern approaches is that of Melitene. This fortified town lay at the hub of communications routes, not least the valley of the southern branch of the Upper Euphrates leading to Lake Van.43 As a centre for Muslim raiders, it had long been a thorn in the flesh of Romania. So it is not surprising that Melitene was Romanos Lekapenos's first target once the pressure from

39 Vasiliev, Byzance et les Arabes I I /l, 287-8; Ter-Ghewondyan, Arab Emirates, 85; Whittow, Orthodox Byzantium, 319. 40 DAI, 44: 204-5. On these and other strongholds in the Van region, see DAI: Comm., 167-70 (Runciman); T.A. Sinclair, Eastern Turkey. An Architectural and Archaeological Survey I (London, 1987), 175-80, 200-202, 206-7, 264-8, 275-6, 279-81, map facing 326; T.A. Sinclair, The Site of Tigranocerta. IF, REArm 26 (1996-7), 71 (map); 73, 79. 41 Whittow, Orthodox Byzantium, 200. 42 DAI, 44: 200-203. 43 F. Hild and M. Restle, Kappadokien, TIB 2 (Vienna, 1981), 233-5; Sinclair, Eastern Turkey III (London, 1989), 1-5, 8-10, 3 5 ^ 2 , map facing 150.

30

JONATHAN SHEPARD

Symeon of Bulgaria eased. What has received less attention is the fact that here, too, direct rule and occupation do not seem to have been the emperor's first preference. Repeated devastation of Melitene's hinterland eventually induced the amir, Abu Hafs, to negotiate and he obtained a logos, presumably guaranteeing immunity from further attacks, from the emperor. He and fellow notables of the town were allowed to remain in power, so long as they rendered tribute and liaised with Byzantine forces. According to Theophanes Continuatus, they even took part in triumphs in Constantinople, 'leading Agarene prisoners-of-war, which was an astonishing and extraordinary sign of the godless Agarenes' misfor­ tune'.44 It was only after the death of Abu Hafs and the occupation of Melitene by the forces of one of the Hamdanid clan that Byzantium resumed its offensives. Eventually, in 934, the town surrendered and its citizens were offered the choice of converting to Christianity and keeping their property or of staying true to Islam and being forced to leave. The emperor was clearly interested in tapping the wealth of Melitene and its fertile surrounding plain and, according to Theophanes Continuatus, 'the emperor has caused many thousands [of pieces] of gold and silver to be paid in tribute (dasmophoreisthai) from there each year'.45 But while the town was taken under the emperor's wing as a kouratoreia, Muslim amirs apparently continued to perform administrative functions there until at least 961, and only from the 970s is there unimpeachable evidence of a Byzantine strategos assigned to Melitene.46 In fact, the town seems to have been partially depopulated and the double circuit of walls was razed to the ground, so that it became an open city. The demolition was later called 'an insensate counsel' by a Syriac writer47 and so it might appear, were one to suppose that Melitene's capture was intended to preface system­ atic Byzantine advances southwards or eastwards from a strategic communications hub. But these actions made sense if the Byzantines had few ambitions to annex territory and were primarily intent on neutraliz­ ing centres that had served as bases for the most damaging raids. Imperial propaganda now celebrated Saracen humiliations and the extension of imperial borders, but high on the actual agenda for the empire's approaches south and south-west from Melitene were the less glamorous goals of order and security. 44 Georgius Monachus Continuatus, in Theoph. Cont., 907; Theoph. Cont., VI.24, 416; Canard, Hamdanides, 734-5; Vasiliev, Byzance et les Arabes I I /l, 267-8. 45 Theoph. Cont., VI.24, 416-17; Howard-Johnston, 'Crown Lands', 86-93. 46 Canard, Hamdanides, 772, 804 and n. 190; Oikonomides, Listes, 265, line 21; F. Tinnefeld, 'Die Stadt Melitene in ihrer späteren byzantinischen Epoche (934-1101 Y, Actes du XIV congrès international des études byzantines II (Bucharest, 1975), 436. On the role of tribute and the prob­ able nature of the kouratoreiai at Melitene and elsewhere, see Holmes' chapter in this volume. 47 Michael the Syrian, Chronique, ed. and trans. J.-B. Chabot, III (Paris, 1906), 123.

THE ROAD TO ALEPPO

31

Our third main point concerns the attractions that might be held to have drawn Byzantium's rulers towards Caucasia as a suitable sphere for enhancing influence and establishing a presence, in preference to regions beyond the Taurus and Anti-Taurus ranges, during the first half of the tenth century. Several well-known phenomena might be adduced as positive grounds for such prioritization, including the migration of numerous Armenian individuals to Byzantium and the Armenian origin of significant numbers of army commanders and members of the ruling elite; the hopes of some Byzantine Church leaders for reunion with the Armenian Church; the economic vitality and size of population of Armenia's towns; and the signs of respect for the empire and its culture on the part of members of local elites in the Armenian lands.48 Few of these attractions were to be found in the 'no-man's-land' of the Taurus ranges or in the ribat accommodating zealous if inexpert ghazis that studded the Cilician plain beyond.49 And it is undeniable that the afore­ mentioned phenomena had a bearing on the emperors' interests. One might note, for example, the tracing of Basil I's ancestry back to the Arsacid dynasty in the Life of Basil overseen by Constantine VII;50 the letter of catholicos Yovhannes Drasxanakertc'i to the young Constantine in which he asks for military intervention to halt the oppression of the Muslim 'wicked beasts' and 'rescue the inheritance which is yours', mooting the association of Yovhannes' own 'flock' with 'the universal flock of your reasonable sheep ... [to] pursue their lives under the aegis of Roman supremacy, just like the people of Italy and all of Asia';51 or the rather less celebrated letter preserved in the name of Gagik Artsruni, ruler of Vaspurakan, which floats the idea of union between the Chalcedonian and the non-Chalcedonian Churches.52 It is from the milieu of the Artsruni that we have some of the most spectacular

48 I. Brousselle, 'L'intégration des Arméniens dans l'aristocratie byzantine au IX siècle', in H. Ahrweiler and N. Garsoïan, eds, L'Arménie et Byzance, BS 12 (Paris, 1996), 43-52; N.G. Garsoïan, 'The Problem of Armenian Integration into the Byzantine Empire', in H. Ahrweiler and A.E. Laiou, eds, Studies on the Internal Diaspora of the Byzantine Empire (Washington, DC, 1998), 53-124 esp. at 61-6, 73-8, 93-9; Redgate, Armenians, 223-5, 252-3. 49 Constantine VII, Three Treatises on Imperial Military Expeditions, ed. and trans. J.F. Haldon, CFHB 28 (Vienna, 1990), 90-91, 126-7, 128-9, 130-31, 174 (commentary); Haldon, Kennedy, 'Arab-Byzantine Frontier', 83, 106-15; Haldon, Warfare, 77 and n. 22, 177-9, 277; C.E. Bosworth, 'The City of Tarsus and the Arab-Byzantine Frontiers in Early and Middle 'Abbasid Times', Oriens 33 (1992), 276, 281-6, repr. in The Arabs, Byzantium and Iran. Studies in Early Islamic History and Culture (Aldershot, 1996), XIV. 50 Theoph. Cont., V .2-3,212-16; Brousselle, 'L'intégration des Arméniens', 43—4; Garsoïan, 'Armenian Integration', 66, n. 59. 51 Drasxanakertc'i, History, LIV.56-7, 66, trans. Maksoudian, 195-6, 197; Whittow, Orthodox Byzantium, 219-20. 52 Maksoudian, 'Introduction', to Drasxanakertc'i, History, 23; Redgate, Armenians, 223.

32

JONATHAN SHEPARD

evidence of familiarity with imperial ideology and iconography. The positioning of the scenes of the Baptism of Christ and the Transfiguration opposite the king's gallery in the church of the Holy Cross at Aghtamar seems to carry the same message as the almost contemporary mosaics opposite the emperor's gallery in St Sophia: that the ruler's authority came from God.53 In the same milieu of the Artsruni in southern Armenia, albeit two generations later, at the end of the tenth century, Gregory Narekac'i enthused about the emperors' control of the elements and of land and sea. He described the empire as 'stretching in the like­ ness of the firmament ... across the vast surface of the entire earth'.54 Kindred themes of the ruler as placed in charge of all God's creation appear in the decoration at Aghtamar and the celebrated bath-house of Leo the Wise.55 The basic contrast - of a predominantly Christian subject-population in most of Caucasia as against the culture, society and economy of Cilicia and parts of Mesopotamia geared to incessant raiding - is probably the single most important reason for imperial prioritization of Caucasian affairs. However, direct appeals such as that of Yovhannes were excep­ tional, and the DAI's cautionary tale about Artanuj suggests awareness of the unforeseeable and undesirable consequences that attempts at actual occupation might precipitate. A common stock of political culture did not necessarily make Armenian princes willing formally to subsume their political structures fully within the emperor's. Rather, ambitious leaders such as Gagik Artsruni looked to the basileus's court for convenient symbols and iconography with which to legitimize and perpetuate the self-determination of their own regimes. They were declaring their own special affinity with Christ's kingdom at least as much as flaunting links with the emperor. Such ambitions would have posed an obstacle to outright imperial annexation. But if the emperor's aspirations fell short of substantial territorial expansion and aimed primarily at securing order on his eastern approaches, stances such as Gagik's were not necessarily unacceptable. Nominal imperial hegemony could be upheld through the grant of titles, gifts and payment of stipends, offsetting the cost to a poten­ tate of any tribute exacted. The emperor did not need to take on the administration or direct defence of a whole district. In fact, the underly­ ing attraction of Caucasia may well have been essentially negative, arising 53 C. Jolivet-Lévy, "Présence et figures du souverain à Sainte-Sophie de Constantinople et à l'église de la Sainte-Croix d'Aghtamar', in H. Maguire, ed., Byzantine Court Culture from 829 to 1204 (Washington, DC, 1997), 236-7, 244-5; L. Jones, The Church of the Holy Cross and the Iconography of Kingship', Gesta 33 (1994), 108, 110, 113-15. For more detail see the chapter by L. Jones in this volume. (Chapter 14). 54 J.-P. Mahé, 'Basile II et Byzance vus par Grigor Narekac'i', TM 11 (1991), 562,563,571-2. 55 Jolivet-Lévy, 'Présence', 239^11, 243^1.

THE ROAD TO ALEPPO

33

from the geo-political problems of the empire: the emperor's limited reserves of full-time military manpower and abiding apprehensions as to the loyalty of Roman forces trained for attack. In other words, Caucasian affairs held out particular appeal for emper­ ors because the local elites were not, for the most part, Muslim; even when they were, there was at least a possibility of forging 'diplomatic' ties and tributary relationships with them, as at Melitene and with the Qaisite amirs around Lake Van. Above all, the Caucasian theatre was singularly well suited for the practice of diplomacy. The emperor could largely handle matters for himself, doling out titles and money, forging personal relationships with leading notables and securing ties through marriages between them or their sons and ladies in his own court circle.56 The emperor, with a minimum of intermediaries, could periodically engage them - and their military manpower - to carry out specific tasks on behalf of himself. Both Romanos and Constantine could contemplate the capture and occupation of so important a raiding base as Theodosioupolis by the Iberian kuropalates. If stability on the eastern approaches and the security of Romania proper rather than the annexa­ tion of further territories were the emperors' principal concerns, such reliance on various forms of local surrogates made good sense. It was compatible with aspirations towards indirect influence or even occupa­ tion at key strategic points, such as the kastra near Lake Van or along the valleys leading there. Moreover, outright occupation of a town could be achieved with only a modest-sized Roman force: that sent to take over Artanuj numbered only 300 men. In mountainous terrain a stronghold could easily close a pass and smallish garrisons were potentially highly cost-effective. Furthermore, it is unproven that the expeditionary forces which John Kourkouas repeatedly led into Caucasia and Mesopotamia to great acclaim57 were on an unprecedented scale, qualitatively or quanti­ tatively differing from those directed by Basil I and Leo VI to sack Muslim strongholds and ravage their hinterlands. After those expedi­ tions, too, spoils and prisoners had been paraded and court-orations delivered in conscious evocation of ancient Roman triumphs. But, then too, attempts at gains more durable than the devastation and debilitation of enemy attack-bases and retribution for Muslim raids were far and few between.58 56 DAI. 43: 188-97; 44: 198-9; 45: 206-7; 46: 214-23; DAI: Comm., 168; A. Carile, Il Caucaso e l'impero bizantino (secoli VI-XI)', SSCIS 43.1 (Spoleto, 1996), 64-6; K. Yuzbashian, 'Les titres byzantins en Arménie', in Ahrweiler, Garsoïan, eds., L'Arménie et Byzance, 218-20. 57 Theoph. Cont. VI.41, 426-7; Canard, Ha'mdanides, 731-53; Vasiliev, Byzance et les Arabes I I /l, 257-73, 284, 295-6. 58 Some of Basil I's expeditions seemingly aimed at permanent recovery of strongholds, but the focus was on troublesome bases, notably Melitene and Tarsus: Theoph. Cont. V.40,

34

JONATHAN SHEPARD

This leads to our postulated paradox. Substantial territorial gains were made soon after work on compiling the DAI ceased, and in a quarter receiving short shrift from the imperial handbook, the south-eastern approaches.59 To suppose that any paradox exists is to beg a question: might not the plugging of Muslim attack-routes across Armenia and taking out of Theodosioupolis have been preparatory to massive assaults across the Taurus? The major expeditions in 948-9 and the sacking of Hadat and Marash might be seen as the first steps in a new forward policy to conquer northern Syria. Lacking space for a full rebuttal, we will merely present an alternative reconstruction, turning on two sets of events. It was, I suggest, 'events' that precipitated a greater military build-up than had been attempted for a long time, a change in military tactics and the long­ term occupation of enemy territory in the Middle Euphrates basin. These two series of events were setbacks, in fact personal and political humiliations, for the emperor: the disastrous failure of the Cretan expedi­ tion of 949, rapidly followed by the incessant inroads into Asia Minor mounted by Sayf ad-Daula. In the opening years of his rule as senior emperor, Constantine staked his prestige on recovering Crete, thereby placing himself in the honourable if unsuccessful tradition of his father's efforts in the eastern Mediterranean, while at the same time showing up the inertia of his detested predecessor, Romanos. Crete was, in fact, an especially 'imperial' concern. The continuing raids and acts of piracy in the Aegean which a Muslim-controlled Crete sanctioned impinged fairly directly on life in the capital: the island appeared its 'all too near neigh­ bour and standing foe' to Liutprand, after his voyage to Constantinople in 949.60 A nexus of diplomatic ties needed to be woven by the emperor in order to isolate the Cretans from other Muslims, and to secure Byzantium's other fronts while its armed forces focused on Crete. Constantine's diplomatic démarches in the later 940s were to a large extent aimed at forming these ties.61 At the same time, the corralling together of disparate units of troops, many of them theme-soldiers from across the empire, and arranging for their pay, equipment and transportation were 269-70; V.46-51, 277-87; Vasiliev, Byzance et les Arabes II /l, 44-8, 82-93, 99-103; Whittow, Orthodox Byzantium, 314 and n. 6; J. Shepard, 'Emperors and expansionism: from Rome to Middle Byzantium', in D. Abulafia and N. Berend, eds, Frontier Societies in the Middle Ages (Aldershot, 2001), forthcoming. 59 Constantine, reviewing the newly instituted commands of the south-east, highlights the initiatives taken by more or less local, self-reliant figures such as Melias: DAI, 50: 238-41. Armenians led by Melias played a key role in the reduction of Melitene: Theoph. Cont. VI.24, 416; Vasiliev, Byzance et les Arabes I I /1, 258, 267-9; G. Dédéyan, 'Mleh le grand, stratège de Lykandos', REArm 15 (1981), 73-102. 60 Liutprand of Cremona, Antapodosis, 1.11, in Opera, ed. J. Becker, MGH in usum schol. (Hanover and Leipzig, 1915), 9. 61 Vasiliev, Byzance et les Arabes I I /l, 322-31.

THE ROAD TO ALEPPO

35

acts of co-ordination which the emperor probably supervised himself. A list of units and supplies refers in passing to arrangements for certain transports, 'as God will guide the holy emperor'.62 A saint's Life confirms the personal, high-profile interest which Constantine took in the expedi­ tion.63 The disastrous outcome - 'of which everyone knows'64 - was thus politically damaging as well as humiliating for Constantine, in so far as the expedition had been intended to display his competence and enjoy­ ment of divine favour. The second series of events, Sayf ad-Daula's incursions, may be traced to motives not utterly dissimilar to the emperor's.65 Sayf was a newcomer to Aleppo, lacking local connections and in need of spectacular gestures through which to legitimize his rule over the disparate inhabitants of the region. A particular problem was the influx into the settled zone on the edge of the north Syrian desert of large numbers of 'Arab, Bedouin tribes­ men who regarded trading caravans and markets as fair game for raiding.66 Partly as a means of providing alternative employment for the Bedouin, partly to place himself at the head of a movement generally deemed admirable and legitimizing in the Islamic world, and partly out of sheer conviction, Sayf unleashed a series of large-scale raids into Byzantine territory from 950 onwards. Most of his warriors were lightly armed and he relied on speed and surprise, tactics that did not make for the capture of fortified towns or seriously threaten Byzantium's strategic position.67 But they did serve to draw attention to Sayf as champion of a revitalized jihad and to humiliate the emperor, feats which the numerous poets and orators in Sayf's entourage proclaimed with éclat. Already during the expedition of 950 the most prolific of the poets, Mutanabbi, described Sayf as aiming for 'the emperor's life' and boasted that if the Byzantine commander-in-chief retreated, 'We will give him a rendezvous on the Bosphoros!'68 Such bluster probably exaggerates Sayf's expecta­ 62 Constantine VII, De cerimoniis aulae byzantinae, 11.45, ed. 1.1. Reiske, I (Bonn, 1829), 670. 63 Vita S. Pauli iunioris, 28, ed. H. Delehaye, in T. Wiegand, ed., Der Latmos (Milet, III/l) (Berlin, 1913), 122; Skylitzes, 245-6; Vasiliev, Byzance et les Arabes I I /1, 340-41. 64 Vita S. Pauli, 28,122. 65 Constantine's use of coins as legitimizing propaganda was noted by T.E. Gregory, The Political Program of Constantine Porphyrogenitus', Actes du XVe congrès international des études byzantines IV (Athens, 1980), 128-30. Sayf likewise emphasized his piety for political purposes on coin legends: R.J. Bikhazi, The Struggle for Syria and Mesopotamia (330-58/941-69) as reflected on Hamdanid and Ikhshidid Coins', ANSMN 28 (1983), 154-6, 159-63,171-2. 66 A.J. Cappel, The Byzantine Response to the 'Arab (10th to 11th Centuries)', ByzF 20 (1994), 116-18. 67 The limitations of Sayf's resources were highlighted by H. Kennedy, The Prophet and the Age of the Caliphates (London, 1986), 279-80. 68 Vasiliev, Byzance et les Arabes 11/2 (Brussels, 1950), 307-8 (Mutanabbi). See, on technical

36

JONATHAN SHEPARD

tions, but it does convey a renewed spirit of daredevilry and the desire to inflict spectacular losses. The damage which this and subsequent raids through the 950s inflicted became notorious in the capital, even if word of his insults only reached the ears of a select number of army commanders and courtiers at Constantinople.69 Constantine VII responded to the challenge with measures for which there were abundant precedents in Byzantine tactics, diplomacy and political theatre. Sayf's armies were ambushed as they withdrew laden with plunder through the passes in the manner prescribed in Skirmishing. In fact, the tactics then used against Sayf by the Domestic of the Schools, Bardas Phokas, form the subject-matter of this treatise, which declares him a pastmaster of the kind of guerrilla warfare formerly needed to defend the south-eastern themes.70 Byzantine armies were also sent on major retaliatory expeditions, to try to distract the enemy from his raids, for example an incursion as far as Nisibis in 952, when large numbers of prisoners were taken. And expeditions were sent to demolish the fortifi­ cations of the forward bases where warriors were marshalled before raids, for example Hadat and Marash (which had already been Byzantine targets in 948). The aim was not occupation, but rather to demilitarize them and make the marshalling of further expeditions against the empire more difficult. Byzantine engineers' dismantling of the fortifications of bases near the Anti-Taurus passes or on routes leading to the passes, and the fighting of battles against Muslim troops who tried to stop them or to carry out rebuilding work, were characteristic of the warfare with Sayf for several years.71 Constantine's other main response to Sayf ad-Daula's incursions took the form of diplomatic overtures. Embassy after embassy was despatched to propose truces and prisoner-exchanges of the sort that had been negoti­ ated between Constantinople and Baghdad for generations. Constantine probably hoped to cool Sayf's martial ardour by treating with him as supreme protector of the Muslims in the borderlands, a role that had, before the exchange supervised by Sayf himself in 946, belonged to the Abbasid caliph. The poetic accounts of Sayf's reaction to the embassies sending 'replies' in the form of his cavalry72 - are borne out by his actions. aspects of one of the poems of Mutanabbi composed during this campaign, A. Hamori, The Composition of Mutanabbi's Panegyrics to Sayfal-Dawla (Leiden, 1992), 8-9, 25-6, 28, 47-8, and, on other poets in Sayf's circle, EÏ2 IX, 103-4 (Bianquis). 69 Sayf's notoriety - unusual for a Muslim border amir - is suggested by the requests for prayers circulated among monastic houses by an imperial official, Sayf's expedition 'being now at the gates': J. Darrouzès, Epistoliers byzantins du X siècle (Paris, 1960), 146. 70 Skirmishing, 148-9. 71 Vasiliev, Byzance et les Arabes I I /1, 347, 352-6, 361. 72 Vasiliev, Byzance et les Arabes II /2, 336 (Mutanabbi).

THE ROAD TO ALEPPO

37

Thus the Byzantine proposal of a prisoner-exchange in 953 was rebuffed and a Muslim expedition reached almost as far as Melitene. The counter­ attack that Bardas Phokas led into Syria was intercepted and crushingly defeated: Bardas himself barely escaped capture and was dealt a wound on the forehead so severe 'that he was to carry around the thick scar for the rest of his life'.73 This detail is furnished not by Sayf's broadcasters but by a Byzantine chronicle: the indignities which the emperor's generals and, in effect, the emperor himself were undergoing were patent, and enduring. Constantine persisted with this combination of measures for several years. The counter-attacks that he mounted required sizeable numbers of troops who had to be maintained for lengthy periods on campaign, while losses, in the form of casualties and capture by the enemy, were heavy. It is from the mid-950s - and not, to my knowledge, earlier - that Arabic sources begin to remark not merely on the large size of the emperor's armies but also on the presence of heavily armed cavalry, kataphrakts, and the heterogeneity of the armies.74 The enlisting of unprecedented numbers of foreigners and recruitment of larger numbers of Byzantineborn soldiers full-time had, I suggest, much to do with the phenomenon of Sayf ad-Daula. The facts that his political base in Aleppo, was set well back from the 'front line' and that he could draw on ample reserves of enthusiastic if lightly armed Bedouin made it necessary for the emperor to raise additional troops, too. The workers employed to dismantle enemy fortifications required protection during their labours and this involved deployment of infantrymen. It is probably no coincidence that texts on combined cavalry and infantry operations proliferate from the mid-tenth century. What seems to be the earliest text, the Syntaxis armatorum quadrata, datable to around the beginning of the 950s, prescribes very briefly the workings of a hollow square of infantry that acted as a mobile base for cavalry.75 Subsequent texts are fuller and presuppose ample reserves of cavalry and infantry that will often be operating on enemy territory. The flurry of revisions and refinements of the hollow square, as also of wedge-shaped heavy cavalry formations, implies that they had only recently come into intensive use and that experimentation was still

73 Skylitzes, 241. 74 Some Arabic writings had long given inflated figures for Byzantine invading armies or for Byzantium's forces overall: Vasiliev, Byzance et les Arabes II/2, 8-9 (Tabari); Whittow, Orthodox Byzantium, 188-91; Haldon, Warfare, 102-03 and n. 68 on 314. But the detailed descriptions of large, heterogeneous Byzantine armies for the 950s are fuller, with the impli­ cation that such armies were an innovation: Vasiliev, Byzance et les Arabes II/2, 125 (Ibn Zafir); 161 (Ibn al-Atir); 243^1 (Dahabi); 331, 333-4, 338-9 (Mutanabbi); 364 (Abu Firas). 75 E. McGeer, 'The syntaxis armatorum quadrata: a Tenth-Century Tactical Blueprint', REB 50 (1992), 221-3, 226-8; E. McGeer, Sowing the Dragon's Teeth. Byzantine Warfare in the Tenth Century (Washington, DC, 1995), 184, 258-9.

38

JONATHAN SHEPARD

under way.76 That major changes in tactics occurred during the 950s is declared by Theophanes Continuatus. Morale was raised and training improved, so that soldiers ceased to 'hide themselves or ... turn tail as used to be their custom'. Instead they 'made their sojourn [in enemy terri­ tory] as if they were in their own country' and, with good courage, 'all advanced keenly against the enemy ... with all their might destroying utterly the Agarenes'.77 This offers, albeit in caricature, an outline of the guerrilla tactics long employed against Muslim raiders in contrast with the new offensive stance. The chronicle credits the change to Nikephoros Phokas, who succeeded his father Bardas as Domestic of the Schools in, probably, 955. This is a simplification, in that 'dogging and pouncing' remained an option in resisting Hamdanid raids, while Byzantine expe­ ditions had long been harrying Cilicia and Mesopotamia. But it registers awareness that significant change took place during the 950s. Constantine VII sought to present himself as au courant with the new tactics and to demonstrate the value of imperial wisdom, gained from the writings of ancient tacticians. Several codices containing collections of classical tactical works were written at his initiative or in the ambit of his court.78 Already, in a speech composed for distribution in 950, he declared his desire to go forth on campaign himself,79 and the compila­ tion of the Treatise on Imperial Expeditions is an earnest indication of his intention to lead from the front. It was probably re-edited and made ready for practical application in the mid-950s.80 But in the event Constantine stayed in his capital and Sayf showed himself still able to run rings round Byzantine defence forces. In the spring of 956, he swept through the area of Anzitene and led a separate expedition as far as Charsianon later the same year. Facing repeated reverses in what had developed into a kind of duel with 'the wicked and impious Hamda[nid]',81 Constantine resorted to various ceremonial devices to conjure up an aura of success. Hence the reception at Constantinople of a relic, the hand of John the Baptist, abducted from Muslim-held Antioch;82 the triumph celebrated after what seems to have been a quite minor naval battle;83 and, most spectacularly, the act of ritual trampling

76 McGeer, Dragon's Teeth, 184-8, 259, 261^4, 272-5, 283-8; Haldon, Warfare, 218-20, 222. 77 Theoph. Cont. VI.41, 459-60. 78 C.M. Mazzucchi, 'Dagli anni di Basilio Parakimomenos (Cod. Ambr. B 119 Sup.)', Aevum 52 (1978), 281. 79 H. Ahrweiler, 'Un discours inédit de Constantin VII Porphyrogénète', TM 2 (1967), 399; Mazzucchi, 'Basilio', 298 and n. 95. 80 Constantine VII, Three Treatises, 51-3 and n. 35 (introduction). 81 Darrouzès, Epistoliers, 146. 82 Skylitzes, 245. 83 Theoph. Cont. VI.29, 453; M. McCormick, Eternal Victory (Cambridge, 1986), 165-6.

THE ROAD TO ALEPPO

39

upon Sayf's cousin AbuT 'Asha'ir which Constantine performed before crowds in the Forum of Constantine in 956, a public humbling with clear overtones of ancient Roman glories and continuity from the first Constantine.84 There was, as Michael McCormick noted,85 a strong element of illusion and morale-boosting behind these ceremonies. Constantine had staked his prestige on putting an end to Sayf ad-Daula's depredations, and yet the raids went on. He therefore paraded through Constantinople such symbols of victory as he could muster and demonstrated that he could still attract powerful new relics for its protection. His overall strategy was still essentially one of 'containment' rather than acquisition of territory beyond the Taurus or Anti-Taurus. Thus in the summer of 957 the Byzantines succeeded in sacking the fortress of Hadat, destroying the fortifications and allowing the citizens safe conduct to Aleppo. They seem then to have withdrawn without attempting to occupy the position permanently.86 Seemingly, they were trying to create a kind of demili­ tarised zone to the south of the Anti-Taurus. Nonetheless, given the scale and great expense of the military build-up as well as the impact which the kataphrakts seem to have had on Muslim formations, Byzantine counter­ measures against Sayf might be expected to have become bolder. It could be that leading army families such as the Phokades, who had suffered casualties in the campaigning, were eager to mount longer-range offen­ sives against Sayf.87 But there is no reason to doubt the indication of an Arabic source that the decision to raise the stakes - to take the road to Aleppo - was made by Constantine himself, as a last resort. According to a commentary on a poem by Abu Firas, after enduring incessant raids and having had truce proposals countered with unheard-of demands, the emperor made peace agreements with the rulers of neighbouring peoples and raised a huge army to send against Sayf 88 This was the army directed against Samosata in 958 which, in contrast with previous expeditions, occupied the town and maintained it. I suggest that a decision had now been taken to deliver a knock-out blow against Sayf, striking at his seat of authority, Aleppo, where an elaborate palace had been built. Samosata was seen as an important forward base for this purpose, protruding into the Hamdanid domain, yet having communications along the Euphrates 84 Skylitzes, 241; McCormick, Eternal Victory, 159-63. 85 McCormick, Eternal Victory, 159-60,188. 86 Vasiliev, Byzance et les Arabes II/2, 97-8 (Yahya of Antioch); Vasiliev, Byzance et les Arabes I I /1, 361. 87 On the Phokas family's leading role in warfare against Sayf, see J.-Cl. Cheynet's appen­ dix, 'Les Phocas', in Le traité sur la guérilla (De velitatione) de l'empereur Nicéphore Phocas (963-969), eds. G. Dagron, H. Mihâescu (Paris, 1986), 298-306, 315. 88 Vasiliev, Byzance et les Arabes, II/2, 368 (Ibn Halawaih).

40

JONATHAN SHEPARD

valley with Byzantine forts in the region of Anzitene and with Melitene.89 Constantine may not have intended the permanent occupation of Samosata or the other towns that were seized during the drive towards Aleppo. The focus was on Aleppo, or rather on shattering Sayf's reputa­ tion as its ruler and protector. The Byzantines' eventual capture of Aleppo in 962 did strike a blow at Sayf's prestige from which it never fully recovered. Although their occu­ pation was brief, the palace was sacked and much of the town burnt; several sections of the walls had been destroyed during the siege.90 But by 962 the emperor who had launched the drive against Sayf was dead, and the military juggernaut directed at Aleppo had acquired a momentum of its own and prestige and political power for its field-commanders. Constantine's choice of the parakoimomenos, Basil, to command the assault on Samosata suggests that considerations of political security were ever­ present, for all his anxiety to deal Sayf's power base an irreparable blow.91 He was signalling his intention to lead operations himself shortly before he fell ill and died, in the autumn of 959.92 It is doubtful whether Constantine's ambitions in the late 950s were any more intent on acquir­ ing territory in Mesopotamia or Syria than they had previously been. He was using massive armed retaliation as a last resort, after the failure of repeated attempts to make truces or divert Sayf from raiding, and his prime concern seems to have been to restore a degree of order and secu­ rity on the eastern approaches. However, the inhibitions on full exploita­ tion of the empire's mounting material and demographic resources for military organisation had been lifted, and what began as a kind of puni­ tive operation turned into a drive for outright annexation of towns and territories in the Middle Euphrates basin and, subsequently, Cilicia. Under a reluctant, diplomacy-loving emperor, new forces of expansion­ ism were released in directions which he probably never intended.93

89 Theoph. Cont., VI.44,461; Canard, Hamdanides, 794-5 and n. 166; Vasiliev, Byzance et les Arabes I I /l, 363; Oikonomidès, 'L'organisation', 289; Hild, Restle, Kappadokien, 125-6. 90 Canard, Hamdanides, 657-8, 814. 91 Theoph. Cont, VI.44, 461; Mazzucchi, 'Basilio', 299; Whittow, Orthodox Byzantium, 346. 92 Skylitzes, 247. 93 Such a scenario is compatible with, and may even be corroborated by, the likelihood that Byzantine emperors, upon gaining mastery over Muslim regions from Melitene south­ wards, inclined to leave existing administrative structures intact and to employ locally born officials. This would be all the more understandable if Byzantine decision-makers had not, in the mid-tenth century, envisaged extensive conquests and lacked the ambition or person­ nel to take over the administrative burdens that permanent occupation, disparate popula­ tions and new funding commitments would entail: see Holmes' chapter in this volume.

3. 'How the east was won' in the reign of Basil II

Catherine Holmes Immortalized by his sobriquet 'the Bulgarslayer', the Byzantine emperor Basil II (976-1025) is most famous for his long military campaigns against Bulgaria. In contrast, the emperor's policy towards the empire's eastern neighbours was more usually characterized by peaceful diplomacy than by warfare.1 Yet, while the eastern frontier remained a low military prior­ ity for most of Basil's reign, it was not a region that could be safely neglected. In the decades immediately preceding Basil's reign, Byzantine armies had taken advantage of the waning powers of the Abbasid caliphate and extended Byzantine territorial boundaries into Cilicia, northern Syria and northern Mesopotamia. The result was a radical rede­ finition of a Byzantine east which for the previous three centuries had been limited to the Anatolian plateau. None-the-less, when Basil II came to the throne in 976, few of the territorial gains of this rapid expansion had been fully consolidated.2 In this chapter I want to ask how Byzantine authority in the newly conquered eastern territories was consolidated during Basil's reign. Given the geographical size of the region in question and the chronological length of the reign, I shall discuss only one dimen­ sion of the eastern frontier experience. Rather than analysing the empire's dealings with neighbouring states or the military administration of the frontier itself,3 I shall focus on relationships between Constantinople and 1 A detailed analysis of the empire's dealings with its eastern neighbours, both Muslim and Christian, during Basil's reign is offered by J.H. Forsyth, 'The Chronicle of Yahya ibn Sa'id alAntaki' (unpublished Ph.D. thesis, Ann Arbor, MI, 1977), chaps. 7-9. See J.-Cl. Cheynet's chapter in this volume (Chapter 4) for Basil's greater military attention to the west rather than the east. 2 For more background see the chapters by J. Shepard and J.-Cl. Cheynet in this volume, (Chapters 2 and 4). 3 For the military organization of the eastern frontier in this period see N. Oikonomidès, Les Listes de préséance byzantines des IXe et Xe siècles (Paris, 1972), 344-6, 354-63; N. From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower Flouse, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

41

42

CATHERINE HOLMES

the diverse populations who lived on the empire's eastern periphery, above all the inhabitants of the Muslim emirates annexed by Basil's predecessors in the second half of the tenth century. My principal ques­ tion will be whether Constantinople sought to control the periphery directly or whether it admitted a more flexible and devolved relation­ ship.4

Economic background It is important at the beginning of this discussion to establish the key background context against which political relationships between the Constantinopolitan centre and eastern periphery developed during the later tenth and early eleventh centuries. In the first section of this chapter I shall argue that this context was a cycle of severe economic contraction on the eastern frontier followed by a swift return to prosperity. Moreover, I shall stress the extent to which this recovery was generated by a hetero­ dox frontier population that included large non-Greek-speaking, nonChalcedonian, and even non-Christian communities. It is clear that the eastwards advance of Byzantine armies during the tenth century caused significant damage to the economy of the former Muslim emirates. This damage was frequently deliberate, with the inten­ sive raiding of rural hinterlands often forming the prelude to the conquest of important urban centres. During campaigns against the emirate of Melitene in the 920s and 930s, Byzantine forces repeatedly 'destroyed the surrounding hamlets and villages by fire'.5 The same strategy was adopted by the emperor Nikephoros Phokas in Cilicia and northern Syria in the 960s.6 Contemporary Arab geographers and historians report at length on the incidence of depopulation caused by the Byzantine Oikonomidès, 'L'organisation de la frontière orientale de Byzance aux Xe-X Ie siècles et le taktikon de l'Escorial', Actes du XlVe congrès international des études byzantines I (Bucharest, 1974), 285-302; in the same volume H. Ahrweiler, 'La frontière et les frontières de Byzance en Orient', 216-19; H.J. Kühn, Die byzantinische Armee im 10. und 11. Jahrhundert. Studien zur Organisation der Tagmata (Vienna, 1991), 158-69; W. Treadgold, Byzantium and its Army 284-1081 (Stanford, CA, 1995), 114-15. 4 This chapter owes a large debt both in terms of argument and evidence to G. Dagron, 'Minorités ethniques et religieuses dans l'Orient byzantin à la fin du Xe et au XIe siècles: l'im­ migration syrienne', TM 6 (1976), 177-216. Although I do not agree with all Dagron's conclu­ sions, I hope that my chapter will draw attention to the immense importance of his article for the history of the medieval Byzantine east. For more on the idea of loose hegemony rather than direct control of the frontier see Shepard's chapter in this volume; see Cheynet's chapter for military administration of the frontier under Basil. 5 Theoph. Cont., 415. For more on Byzantium's policy towards Melitene see Shepard's chapter in this volume. 6 Yahya ibn Sa'id al-Antaki, Histoire, ed. and trans. I.Kratchkovsky and A.Vasiliev, PO 18 (1923), 826.

'HOW THE EAST WAS WON'

43

advance, as Muslims who would not convert to Christianity were forced to leave conquered territories. When Tarsos in Cilicia fell in 965, many Muslims left the city for Antioch. When Antioch itself then surrendered in 969, these refugees moved on to the Syrian coastal town of Balanias.7 Yet, although widespread devastation and depopulation may have been the immediate consequence of Byzantine conquest, economic contraction appears to have been relatively short-lived. Outgoing popu­ lations were often replaced by in-comers, many of whom were nonChalcedonian Christians. Some years ago Gilbert Dagron used tenth- and eleventh-century chronicle evidence contained within the twelfth-century history of Michael the Syrian to demonstrate that by the later 950s large numbers of monophysite Syrians were beginning to migrate to the former emirate of Melitene.8 It is also clear that many Armenians came to live in former Muslim-controlled territories. By the final decade of the tenth century their numbers in Cilicia and Syria were such that Armenian monophysite episcopal sees were established at Tarsos and Antioch.9 In northern Syria monks professing various eastern Christian faiths were increasingly to be found in the countryside. Armenian monasteries were recorded in the Amanos mountains during the reign of Basil himself.10 Evidence from eleventh-century manuscript colophons and the archaeo­ logical record indicates that Georgian Chalcedonian monks were active in northern Syria.11 They were certainly present at, and may have controlled, the monastery of St Symeon Stylites the Younger, on the Wondrous Mountain.12 However, it was not purely eastern Christian migrants who contributed to the demographic and economic recovery of the Byzantine east. While Arab historians and geographers report that Byzantine conquests often 7 Yahya, PO 18, 797; al-Mukaddasi, The Best Divisions for Knowledge of the Regions, trans. B.A. Collins (Reading, 1994), 147; Dagron, 'Minorités ethniques', 180-81. 8 Dagron, 'Minorités ethniques', 189-90; Michael the Syrian: Chronique de Michel le Syrien, Patriarche Jacobite d'Antioche (1169-99), ed. and trans. J.B. Chabot (Paris, 1905-10), 125-7. 9 Stephen of Taron: Des Stephanos von Taron armenische Geschichte, trans. H. Gelzer and A. Burckhardt (Leipzig, 1909), 196; N.G. Garsoïan, 'Armenian Integration into the Byzantine Empire', in H. Ahrweiler and A.E. Laiou, eds, Studies on the Internal Diaspora of the Byzantine Empire (Washington, DC, 1998), 56-7. For the migration of Armenians to Byzantium see Shepard's chapter in this volume. 10 Armenia and the Crusades in the Tenth to Twelfth Centuries: the Chronicle of Matthew of Edessa, trans. A.E. Dostorian (Lanham, NY, 1993), 47-8. 11 W.Z. Djobadze, Materials for the Study of Georgian Monasteries in the Western Environs of Antioch-on-the-Orontes (Louvain, 1976); W.Z. Djobadze, Archaeological Investigations in the Region West of Antioch-on-the-Orontes (Stuttgart, 1986). 12 Djobadze, Archaeological Investigations, 204-11. By the end of the eleventh century Saint Nikon of the Black Mountain also noted the presence of Chalcedonian Armenian monks at St Symeon's monastery: Garsoïan, 'Armenian Integration', 106-8.

44

CATHERINE HOLMES

entailed the mass exodus of Muslims from the former emirates, either as fugitives or as enslaved prisoners of war, it is clear that many Muslims remained under Byzantine rule, or returned to their former homes after the conquests were over. In some cases the price of their residence in the empire was conversion to Christianity.13 However, while conversion was preferred, it may not have been mandatory. Writing in the final decades of the tenth century, the Arab geographer Ibn Hawqal lamented that many Muslims resided in Byzantine territory happy to pay a head tax.14 In 1048/9 Ibn Butlan, an Arab Christian doctor from Baghdad travelling in northern Syria, observed a mixed Christian and Muslim population cultivating a flourishing countryside near Antioch. In one village on the road between Aleppo and Antioch he noted the presence of a mosque as well as four churches. When he arrived at the port of Laodikeia he observed that while the town's main mosque had been converted into a church, the local Muslim population were able to worship in another mosque; they also retained their own judge (qadi).15 The prosperity of the Byzantine east in the later tenth and early eleventh centuries was also encouraged by commercial exchange with Muslims outside the empire. The best evidence of the importance of long-distance trade across the eastern frontier comes from the Treaty of Safar drawn up between Byzantium and its northern Syrian client state of Aleppo in 969. According to the commercial clauses of this agreement, goods conveyed by overland caravan included gold, silver, silk, precious stones, linen, Greek brocades and animals.16 So important was this caravan that when civil war broke out in the later 970s between Basil II and the general Bardas Skleros, imperial and rebel forces fought a battle in the passes of the Taurus mountains to gain control of it.17 Sea-borne trade between Byzantium and the empire's Muslim neighbours also seems to have been significant. The wreck at Serge Liman, which sank with a cargo of glass off the coast of Asia Minor, provides archaeological evidence for maritime trade between Byzantium and the Islamic east during Basil's reign itself. Copper coins of Basil II and gold quarter dinars of the contemporary Fatimid caliph al Hakim (996-1021) were found on board.18 13 For example at Melitene: A.A. Vasiliev, Byzance et les Arabes II/2 (Brussels, 1950), 154. 14 Ibn Hawqal, La Configuration de la terre, trans. J.H. Kramers and G. Wiet (Beirut and Paris, 1964), 186. 15 The Medico-Philosophical Controversy between Ibn Butlan of Baghdad and Ibn Ridwan of Cairo, ed. and trans. J. Schlacht and M. Meyerhof (Cairo, 1937), 54-7. 16 W. Farag, The Truce of Safar AH 359', offprint from the Eleventh Spring Symposium of Byzantine Studies (University of Birmingham, 1977). I am grateful to lohn Haldon for providing me with this. See also M. Canard, 'Les relations politiques et sociales entre Byzance et les Arabes', DOP 18 (1964), 52. 17 Skylitzes, 32. 18 G.F. Bass, 'A Medieval Islamic Merchant Venture', Archaeological News 7 (1979), 84-94.

'HOW THE EAST WAS WON'

45

Trade with the eastern world beyond the territorial borders of the empire may also provide the context for the economic recovery of the former emirate of Melitene. The city itself lay at the crossroads of impor­ tant trading routes. To the east lay Armenia, source of highly prized silks and woollen products;19 to the south the Djazira, 'the source of all supplies for Iraq'.20 Amida, located on the headwaters of the Tigris, was an established entrepot for Byzantine linens.21 Less than 200 kilometres down river was Djazirat ibn Umar, situated on the junction for routes between Armenia, Byzantium, Mayafaraqin and Arran.22 Several hundred kilometres further downstream lay Mosul and Takrit.23 The Takritan connection may have been of particular significance for the pros­ perity of Melitene. In his study of the migration of Syrian monophysites to the Byzantine eastern frontier, Dagron calculated that by the early eleventh century there were fifty-six Syrian churches in and around Melitene.24 Many of their patrons came from Takrit. Of these, the most famous were the Banu Abu Imran. Their wealth was such that they lent Basil II enough money to support an entire Byzantine field army when he stayed in Melitene during the winter of 1022.25 Their prosperity almost certainly derived from their position as merchants on the Tigris trading route. Not only did members of Banu Abu Imran live in Melitene and Takrit; others were to be found in Mosul.26

Tenth-century imperial pragmatism Thus far this chapter has argued that the agricultural and commercial activity of a variety of eastern Christian and Muslim settlers was funda­ mental to the recovery of the eastern periphery of the Byzantine empire in the later tenth and eleventh centuries. But the more important question for our purposes is how did the imperial authorities in Constantinople deal with this heterodox frontier population? In this section I shall argue that successive tenth-century emperors, above all Basil II, usually adopted a pragmatic approach to local governance, choosing to work with rather than against the ethnic and religious plurality of the Byzantine east. 19 Al-Mukaddasi, 329-31; Ibn Hawkal, 338; Al-Tanukhi, Table Talk of a Mesopotamian Judge, trans. D. Margoliouth (London, 1922), 137. 20 Al-Muqaddasi, 124. 21 Ibid., 133. 22 Ibn Hawkal, 219. 23 Ibid., 209, 223; al-Muqaddasi, 111. 24 Dagron, 'Minorités ethniques', 194. 25 Michael the Syrian, 144-5; Bar Hebreus: The Chronography of Gregory Abu l Faraj, the Son of Aaron, the Hebrew Physician, Commonly Known as Bar Hebreus, ed. and trans. E.A. Wallis Budge (London, 1932), 178. 26 Ibn Hawkal, 209.

46

CATHERINE HOLMES

One of the most visible manifestations of this imperial pragmatism is the extent to which emperors themselves often took practical steps to encourage the settlement and commercial enterprises of the heterodox frontier populations. In the mid-960s Nikephoros Phokas (963-69) encouraged the Syrian Jacobite patriarch to move to northern Mesopotamia from Antioch in order to escape persecution from Melkite (Chalcedonian Arab) Christians in northern Syria. The patriarch's migra­ tion then inspired a more general movement of Syrian Christians to Melitene and its hinterland.27 Syrian migrants, especially merchants from Takrit, were also drawn to Melitene by more prosaic imperial induce­ ments: the availability of tax breaks.28 Imperial authorities were even anxious to retain Muslim populations within the conquered territories. Nikephoros Ouranos, Basil II's supreme commander in the east during the first decade of the eleventh century, recommended that if enemy cities surrendered voluntarily, the local inhabitants should be allowed to keep their possessions, while the leaders of the urban élite should receive presents.29 Nor was this ethnic inclusiveness necessarily mere official propaganda. An eleventh-century Iraqi chronicler explicitly praised Basil II for his justice and affection for Muslims, his willingness to keep out of Muslim territory and his kindness to Muslims who entered his.30 Pragmatism also seems to have been the hallmark of the way in which imperial authorities chose to administer the former emirates. Although the evidence is patchy and comes mainly from lead seals, emperors such as Basil appear to have been willing to acknowledge the logic that in regions where the everyday language of economic and fiscal exchange was not Greek, maximum benefit was likely to accrue from minimal administrative change. Of course, most modern discussions of adminis­ tration in the Byzantine east usually focus on the military organization of the frontier, tracing the development of large regional duchies manned by professional troops from the centre under the hegemony of a doux or katepan, and the significance of small border themes staffed by Armenian infantry and light cavalry-men.31 However, underneath this military

27 Dagron, 'Minorités ethniques', 186-204; Michael the Syrian, 130-32. 28 Bar Hebreus, 178. For imperial interest in the commercial potential of the east see Shepard's chapter in this volume. 29 E. McGeer, Sowing the Dragon's Teeth: Byzantine Warfare in the 10th Century (Washington DC, 1995), 158. 30 Eclipse of the Abbasid Caliphate, ed. and trans. H. Amedroz and D. Margoliouth (Oxford, 1920-1), VI, 119. For the pragmatic and rational governance of the frontier in accordance with the manpower and financial resources available to Byzantium see in this volume Cheynet and Shepard. 31 See, for example, Oikonomides, Les Listes, 344-6, 354-63; Kühn, Die byzantinische Armee, 158-69; Treadgold, Byzantium and its Army, 114-15; see also Cheynet in this volume.

'HOW THE EAST WAS WON'

47

administrative tier, which may itself have been much more flexible and ad hoc in its arrangements than contemporary Byzantine bureaucratic taktika imply, I would suggest that there persisted a civil administration very similar to that which had existed under the previous Muslim regimes32. One strong sign that pre-existing administrative structures were left intact by tenth-century emperors comes from the fact that the seals of judicial and fiscal officials ubiquitous elsewhere in the Byzantine empire are rarely to be found on the eastern frontier. For example, there are hardly any extant seals from the east of kommerkiarioi (customs officials), despite the clear importance of trade in these regions and the explicit imperial desire to promote commerce visible in areas such as Melitene.33 Moreover, even in those instances where seals of bureaucrats familiar elsewhere in the empire are found in an eastern context, it seems possible that the primary function of their owners was to act as the overseers of indigenous tax collectors and judicial officers. This hypothesis certainly seems to be the best explanation for the survival from the eastern frontier of a large number of seals belonging to kouratores or episkeptitai, officials closely associated in the rest of the Byzantine empire with imperial estate management. Hitherto it has usually been argued that the appearance of such seals in an eastern context indicates that large areas of the Muslim emirates were turned into crown estates managed directly by imperial officials.34 Yet I would argue that in the east kouratores were not estate offi­ cials at all, but instead plenipotentiary figures placed at the head of an infrastructure of indigenous administrators. In this role they acted more as the guarantors of tribute than as the managers of imperial immovable property. The viability of this association between kourator and tribute is most strongly supported by the historical account of the creation of the kouratoreia at Melitene, established when that city was conquered in 934: 'they

32 See Shepard's chapter in this volume for the incidence of local dignitaries left in charge at Melitene from the 920s to 961. 33 Thus far I have managed to find only one example from an eastern context: G. Schlumberger, Sigillographie de l'empire byzantin (Paris, 1884), 312, no. 157. 34 N. Oikonomidès, 'L'évolution de l'organisation administrative de l'empire byzantin au XIe siècle', TM 6 (1976), 138; Kaplan, Les Hommes et la terre à Byzance du VIe au XIe siècle (Paris, 1992), 316-17; J.D. Howard-Johnston, 'Crown Lands and the Defence of Imperial Authority in the Tenth and Eleventh Centuries', ByzF 21 (1995), 88-94. Howard-Johnston provides a detailed list of such officials from the eastern frontier, to which should be added: John, spatharokandidatos and kourator of Antioch (J.C. Cheynet, 'Sceaux byzantins des musées d'Antioche et de Tarse', TM 12 (1994), no. 47); Euthymios Karabitziotes, exaktor, krites of Hippodrome, Seleukeia and kourator and anagrapheus of Tarsos (N. Oikonomidès, Studies in Byzantine Sigillography III (Washington, DC, 1993), 192); John Hexamilites, krites of Seleukeia and kourator of Tarsos (J. Nesbitt and M.Braunlin, 'Selections from a Private Collection of Byzantine Bullae', Byz 68 (1998), no. 13).

48

CATHERINE HOLMES

[the Byzantines] captured and razed it to the ground not only Melitene but also its neighbouring cities and districts which were highly produc­ tive and very fertile and yielded many other revenues. Having then turned Melitene into a kouratoreia, the emperor had many thousands of [pounds] of gold and silver raised annually in revenues from there'.35 The term used by Theophanes Continuator to convey the sense of the revenues raised on an annual basis at Melitene is Saofjoçopeîodai : The principal meaning of Saafjôç in Greek is 'tribute'.36 Although no other source comments explicitly on the imperial kouratoriai in the Byzantine east, there is literary evidence which indicates that the payment of tribute was how the imperial authorities most readily conceived of the reward they could expect from the conquest of Muslim emirates. This expectation is most clearly stated in the case of the campaign to annex Antioch. The city itself was conquered in the autumn of 969. However, in the summer of the previous year, Byzantine armies had softened up the city's hinter­ land with a large punitive raid. As the main army withdrew north for the winter, small Byzantine garrisons were left behind, occupying a ring of fortifications in the mountains and roads that surrounded the city.37 From these bases Byzantine commanders were under instructions to lead daily raids on the countryside around Antioch in order to force the inhabitants within the city to surrender. In a tribute-related context, it is striking that the later tenth-century historian, Leo the Deacon, argued that the objec­ tive of this strategy was to compel Antioch to become tributary (hypospondos) to the Byzantines.38 The principle that Constantinople may have chosen to control the eastern frontier through tribute-raising arrangements with local popula­ tions may also help to explain the important but rather ambiguous posi­ tion in the historical record of basilikoi. Whenever basilikoi are discussed by modern historians they are uneasily compared to kouratores, that is to say as officials with a role in estate or fiscal administration.39 However, the 35 Theoph. Cont., 417. For more on kouratoreiai see Cheynet's chapter in this volume. 36 H.G. Liddell and R. Scott, A Greek-English Lexicon, 2nd edn. (Oxford, 1968), 370. Theoph. Cont., 417; see Shephard's chapter in this volume. 37 These fortresses included Baghras, at the centre of a new theme in the Amanos moun­ tains called Mauron Oros, and Qalat Siman, the fifth-century monastery of St Symeon Stylites: Yahya, PO 18, 816; Skylitzes, 271-2; J.C. Cheynet, C. Morrisson and W. Seibt, Sceaux byzantins de la collection Henri Seyrig (Paris, 1991), no. 183; W.B.R. Saunders, 'Qalat Siman: a Frontier Fort of the Tenth and Eleventh centuries', in S. Freeman and H. Kennedy, eds, Defence of the Roman and Byzantine Frontiers, BAR International Series (Oxford, 1986), 291-305. 38 Leo the Deacon: Leonis Diaconi Caloënsis Historiae Libri Decern, ed. C.B. Hase, CSHB (Bonn, 1828), 73-4: .... coûte Ka0’ÉKâaTT]v ette^e Agcoeoi kcù KaTaBponaîç kcù èttitittiBeîcov BiapTrayaîç rr\v ’A v t îo x o u TaiTEivcoocoai, koù eiç ànrixaviav Beivt^v KaTaKÂEÎaavTEç kcù aKouaav àvayKaacoai 'Pconaioiç yEVÉoôai ÙTrôcrrTovBov. 39 Ahrweiler, 'Recherches', 73-4; J.C. Cheynet, 'L'apport arabe à l'aristocratie byzantine des Xe-X Ie siècles', ByzSlav 61 (1995), 141-2.

'HOW THE EAST WAS WON'

49

careers of the two basilikoi in the later tenth century indicate that such offi­ cials could be key intermediaries upon whom the Byzantine authorities in Constantinople depended in order to mobilize the resources of the great former emirates. One of these basilikoi was Kulayb, whose career is predominantly known from the chronicle of the Arab Christian historian Yahya ibn Sa'id al-Antaki. Kulayb was a Christian Arab, possibly a Syrian monophysite, who entered Byzantine service in 975 when he surrendered two fortresses in northern Syria to the emperor John Tzimiskes; in return he was given the high-ranking title of patrikios and appointed the basilikos of Antioch. During the revolt of Bardas Skleros, at the beginning of Basil's reign, Kulayb surrendered Antioch to the rebels, and was transferred to the position of basilikos in Melitene instead. When Skleros returned to Melitene from exile in Baghdad a decade later in 987, Kulayb was still basilikos of the city.40 If other evidence is aggregated with Yahya's testi­ mony, then Kulayb is transformed from a fairly anonymous frontier char­ acter into a linchpin of local politics and diplomacy during the first decade of Basil's reign. When a diplomatic envoy, Ibn Shahram, was sent in 981 by the Buyid emir of Baghdad to Byzantium to discuss Skleros's exile in Iraq, he met Kulayb. In his report of his meeting Ibn Shahram indicated that Kulayb alone of the rebel Skleros party had received a pardon from the emperor. It is clear from this report that it was Kulayb's ability to ensure the annual delivery of the tribute from the client state of Aleppo in northern Syria that had guaranteed his personal status on the frontier.41 Further signs that Kulayb was a high-profile figure on the fron­ tier, well rewarded by authorities at the centre, comes from the fact that he was able to sponsor the high-profile Syrian monastery of Bar Gagai near Melitene in 987/8.42 A second basilikos of critical political importance at the start of Basil's reign was Ubayd Allah, another Christian Arab. In 976 he used his posi­ tion as basilikos of Melitene to surrender the city to the rebel army of Bardas Skleros. This action enabled Skleros himself to sequester the fiscal revenues of the former emirate, and to declare revolt openly against the emperor. Taken into the service of Skleros, Ubayd Allah became Kulayb's successor as basilikos at Antioch.43 Basil II was only able to regain Antioch for the imperial 'side' in 977/8 when he promised to make Ubayd Allah governor (wilaya) of the city for life.44 Yahya ibn Sa'id's account of Ubayd 40 Yahya, PO 23 (1932), 369, 373, 420. 41 Eclipse of the Abbasid Caliphate VI, 23^4. 42 Michael the Syrian, 125-6; Dagron, 'Minorités ethniques', 192, 197. Sigillographical evidence indicates that Kulayb also had a son called Bardas: G. Zacos, Byzantine Lead Seals II, compiled J. Nesbitt (Berne, 1985), no. 371. See also Cheynet, 'L'apport arabe', 141-2. 43Yahya, PO 23, 373. 44 Ibid., 375-7; V.Laurent, 'La chronologie des gouveneurs d'Antioche sous la seconde domination byzantine', Mélanges de l'Université Saint-Joseph de Beyrouth 38 (1962), 231-2.

50

CATHERINE HOLMES

Allah's actions during the civil war at Antioch indicates the breadth of his powers. As basilikos he was clearly not simply a functionary with respon­ sibility for fiscal and judicial matters, but was also in charge of an armed garrison. Once he had defected to the emperor, he defended Antioch against attack by a Skleros army. Furthermore, he suppressed a revolt by local Armenians.45

Limited transformations during the reign of Basil Taken together Kulayb and Ubayd Allah's careers demonstrate the degree to which Basil II was dependent during the early years of his reign on local figures from former Muslim regimes to mobilize the resources of the frontier emirates, and to facilitate diplomatic relations with the empire's eastern neighbours. In some senses, however, once the civil wars which plagued the first thirteen years of Basil's reign came to an end in 989, the nature of the emperor's dependence on intermediaries began to change. From this point on key functionaries on the frontier seem to have been drawn from the ranks of Constantinopolitan administrators rather than from the representatives of previous Muslim regimes. The trajectory of this change is most easily traced in the ecclesiastical and secular history of northern Syria and Cilicia as it is reported by the historians Yahya ibn Sa'id and Michael the Syrian. In Yahya's account of the civil wars of the early years of Basil's reign, it becomes clear that another key intermediary figure on the eastern frontier was Agapios, patriarch of Antioch. Agapios's rise to power began during the first Skleros revolt. When Theodore, the incumbent patriarch of Antioch, died in May 976, Agapios, the bishop of the neighbouring Byzantine client city of Aleppo, travelled to Constantinople to persuade the emperor to appoint him as Theodore's replacement. In return he promised to compel Ubayd Allah, the rebel basilikos of Antioch, to declare for the emperor.46 Despite Agapios's relatively junior status, Basil and his advisers were so desperate to regain control of Antioch from the Skleros party that they agreed to this plan. In the first instance it was a deal that worked to Agapios's advantage. On his return to the east he persuaded Ubayd Allah to defect. In January 978 he himself was then installed as patriarch.47 During the next decade he used the authority he had been granted by Constantinople to secure his own position in the locality. At the heart of his policy was the promotion of the Antiochene Melkite church at the expense of local Syrian monophysites. According to later Syrian histori­ ans, Agapios burnt the books of Syrian churches, forced local notables to 45 Yahya, PO 23, 378. 46 Ibid., 375-6. 47 Ibid., 377.

'HOW THE EAST WAS WON'

51

have their children re-baptized as Chalcedonians, and then deployed these converts as local clergy in the countryside.48 However, Agapios' power as a mediator between locality and centre proved to be short-lived. Twelve years later Basil II decided to extricate himself from dependence on local figures such as Agapios. In 989 Agapios was accused of colluding with another rebel family, the Phokades, summoned to Constantinople and secluded in a suburban monastery. During the 990s, the emperor began to extend his authority even more energetically into the localities. In 996 Agapios was officially deposed.49 His replacement was a Constantinopolitan, John the chartophylax of Hagia Sophia.50 Soon John was joined in the east by another Constantinopolitan functionary Nikephoros Ouranos, who, as kraton of the east, exercised supreme mili­ tary command over the whole frontier.51 Yet, while the appointment of Constantinopolitan figures to positions of senior command on the frontier represented a change in the balance of power between centre and periphery in favour of the former, there is little sign that the basic structure of governance in the Byzantine east was revolutionized during the second half of the reign of Basil. Instead there is persuasive evidence that, underneath a thin tier of centrally appointed officials, the quotidian management of the eastern frontier remained in the hands of indigenous functionaries. As a result there was little change in the basic tribute relationship between locality and centre that had characterized Byzantine administration in the east since the middle of the tenth century. Evidence for only limited changes to fron­ tier administration during the second half of the reign of Basil II, and indeed during the rest of the eleventh century, comes both from the careers of those senior officials who were dispatched to the east from Constantinople, and from our knowledge of the minor officials on the periphery itself. Turning first to the careers of Constantinopolitan officials dispatched to the east, it is striking how many continued to fulfil the intermediary and plenipotentiary role previously undertaken by local notables such as Kulayb and Ubayd Allah. One such official was Nikephoros Ouranos, kraton of the East. Now at one level, Ouranos's duties were primarily mili­ tary in nature. Shortly after his arrival in Antioch he accompanied Basil II on a campaign to annex the princedom of Tao in western Georgia in 48 Michael the Syrian, 131-2. 49 Yahya, PO 23, 428. 50 Ibid., 445-6. 51 Skylitzes, 345; Yahya, PO 23, 400, 460, 466-7 J. Nesbitt and N. Oikonomidès, Catalogue of Byzantine Seals at Dumbarton Oaks and in the Fogg Museum of Art III (Washington, DC, 1991-96), no. 99.11; E. McGeer, Tradition and Reality in the Taktika of Nikephoros Ouranos', DOP 45 (1991), 139-40).

52

CATHERINE HOLMES

1000.52 The following year he returned to Tao to repel an incursion led by Gurgen, prince of Inner Iberia.53 Several years later he won a victory over an Arab dervish insurrectionist called al-Acfar.54 Yet Ouranos also had the expertise to take on much wider frontier duties than those of a mere military commander. He was able to call upon extensive experience in administrative and diplomatic affairs. Ouranos' early professional life had been spent in Constantinople within the imperial palace and the upper echelons of central administration. By 982 he was keeper of the imperial inkstand, a position requiring competence in the handling of sophisticated documents including imperial chrysobulls.55 His knowl­ edge of the administrative practices and court politics of Constantinople was so well regarded that during the mid- to later 980s he was appointed epitropos, or lay guardian, of the Athonite monastery of the Lavra, a posi­ tion which must have brought him experience in acting as an intermedi­ ary between the interests of a locality and central government.56 He was also a skilled diplomat. In 982 he was sent to Buyid Baghdad to negotiate the release into Byzantine hands of the rebel general Bardas Skleros.57 Furthermore, Ouranos was not the only official from the capital during the second half of the reign of Basil who was drafted into a frontier plenipotentiary role that demanded a full portfolio of competences. After Ouranos was posted to Antioch, he summoned his friend and correspon­ dent Philetos Synadenos to Tarsos.58 Although the later eleventh- or early twelfth-century manuscript in which the Synadenos-Ouranos correspon­ dence appears tells us that Philetos was krites of Tarsos, the responsibili­ ties which he undertook when he arrived in the east may have extended more widely than those of a judge.59 If, for example, Synadenos was vested with the same offices held by other senior officials at Tarsos in the tenth and eleventh centuries, then it is likely that his real responsibilities were as krites of the neighbouring theme of Seleukeia and kourator of the 52 Yahya, PO 23, 460. 53 Stephen of Taron, 212. 54 Yahya, PO 23, 466-7. 55 Several of Nikephoros's own letters seem to date from the period when he was still keeper of the imperial inkstand: V. Laurent, Le Corpus des sceaux de l'empire byzantin II, L'Administration centrale (Paris, 1981), 102. 56 Ouranos' appointment post-dates 984 and pre-dates 999: P. Lemerle, A. Guillou, N. Svoronos and D. Papchyrssanthou, eds, Actes de Laura I: Des origines à 1204, Archives de l'Athos V (Paris, 1970), 19-20, 45-6, 52; McGeer, 'Tradition and Reality', 130-31. 57 Eclipse of the Abbasid Caliphate VI, 25-34; Skylitzes, 327; Yahya, PO 23, 400^102. 58 'The very famous magistros, Ouranios [the heavenly onel, made me come': Darrouzès, Épistoliers byzantins, 257. 59 Synadenos' letters appear in MS 706 from the monastery of St John on Patmos: Darrouzès, Epistoliers byzantins, 9-12.

'HOW THE EAST WAS WON'

53

former emirate of Tarsos.60 Furthermore, while it is dangerous to read substantive meanings into the elusive literary artefacts which passed between senior officials such as Synadenos and Ouranos, it is possible that an elliptical allusion to the incompatibility of learning and the bearing of arms contained in one of Philetos's letters to Nikephoros may reflect the wide range of duties, including military service, that officials on the frontier were expected to undertake in imperial service. If this is so, Philetos implies that Ouranos was better equipped than himself: On the one hand I have lost the capacity to be wise and to be called wise, and on the other, I am completely inexperienced in the bearing of arms, the rattling of a spear, the moving and shooting of an arrow, and the shaking of a spear against the enemy, and as much as is required to make war against the foe - for I am not hardhearted or very daring, but someone undaring and feeble - I have failed at both: for I am now neither wise, nor daring, in the face of the enemy. And so tell me who I am, wise Strategos. As for me, what I had I have thrown away, what I had not, I am unable to take hold of, and that which I am, as you see, I have lost'.61

Moving beyond the careers of individuals such as Ouranos and Synadenos, there is further evidence that the appointment of Constantinopolitan officials to senior positions on the frontier in the second half of Basil's reign did little to change the basic tribute relation­ ship underpinning Byzantium's governance of its eastern provinces. Although this evidence comes from a region outside the former Muslim emirates, and from a slightly later period than Basil's reign, it demon­ strates that Byzantine administration continued to depend on indigenous officials deep into the eleventh century. The evidence in question is to be found in the Caucasian katepanate of Iberia, created from the princedom of Tao in western Georgia annexed during the second half of the reign of Basil II. Some decades later, during the reign of Constantine Doukas (1059-67), the senior Byzantine commander on this stretch of the frontier, the katepan Bagrat Vxkac'i, introduced a series of tax concessions for the northern Armenian city of Ani. Notice of these arrangements is inscribed on the west wall of the city's cathedral. The inscription itself was written in contemporary vernacular Armenian, and could thus be read by the local inhabitants. More important, it lists the officials who were expected

60 The evidence here comes from the tenth- and eleventh-century sigillographical record: Eustathios Romaios, krites of Seleukeia and megas kourator of Tarsos: K.M. Konstantopoulos, BuÇavnaKà ¡joXußSoßouXXa toû èv 'AQrjvais ’E ôvikoû NomoidariKou Mouoeiou (Athens, 1917), no. 147a; Euthymios Karabitziotes exaktor, krites of the Hippodrome and Seleukeia and kourator and anagrapheus of Tarsos and John Hexamilities, krites of Seleukeia and kourator of Tarsos (see above note 34). 61 Darrouzès, Épistoliers byzantins, 255.

54

CATHERINE HOLMES

to execute the kateparís decree. These functionaries are called tanuters. All are identified as local Armenians who held modest Byzantine titles: Mxit'ar hypatos, Grigor spatharokandidatos, and Sargis spatharokandidatos. One modern Armenian historian has suggested that these tanuters were the managers of local economic and fiscal affairs, whose duties resembled those performed by functionaries known as rais within towns with large Muslim populations. That is to say, they acted as the spokesmen for their own communities within different quarters of the city, and were respon­ sible for managing the fiscal relationship between those communities and the local representative of centralized authority.62 As such they were the key intermediaries around whom a tribute-based system of taxation could operate.

Conclusion This chapter has argued that imperial realism and pragmatism won the Byzantine east for Constantinople in the later tenth and eleventh centuries. Successive emperors encouraged the settlement and commer­ cial activities of the diverse populations of the frontier, and were willing to administer these regions through indirect means, which left indigenous bureaucratic structures largely intact. There are signs that these principles continued to underpin centre-periphery relations throughout the reign of Basil II and deep into the eleventh century. Taken as a whole, this is a picture of close co-operation between Constantinople and the populations of the eastern periphery. Yet this is a model that also contradicts the widely held belief that during the eleventh century an orthodox Chalcedonian centre became increasingly unable and unwilling to assim­ ilate a heterodox frontier. According to this view, it was the centre's failure to assimilate the periphery that persuaded many frontier popula­ tions to support Turkish and Crusading armies rather than Byzantine forces during the later eleventh and early twelfth centuries.63 This chapter is not the place to undertake an entirely fresh analysis of the role of ethnic and religious division in the political turmoil that surrounded the collapse of the eastern half of the Byzantine empire in the years after the battle of Manzikert. However, I would like to conclude with one brief thought about the relationship between administrative accommodation and ethnic conflict.

62 K.N. Yuzbashian, 'L'administration byzantine en Arménie aux Xe et XIe siècles', REArm 10 (1973-74), 179-81. 63 S. Vryonis, 'Byzantium: the Social Basis of Decline in the Eleventh Century', GRES 2 (1959), 169-72; Dagron, 'Minorités ethniques', passim. A corrective to a model of irreconcil­ able conflict between centre and periphery has recently been offered by P. Magdalino, The Byzantine Background to the First Crusade (Toronto, 1996), 18-33.

'HOW THE EAST WAS WON'

55

One of the cases most frequently cited to support the argument that the eleventh century witnessed the irretrievable breakdown between an Orthodox centre and heterodox frontier is the story of the arrest, impris­ onment and death of the Syrian patriarch John Bar Abdoun during the short reign of Romanos III (1028-34). The fundamentals of the case can be established from the chronicle of Michael the Syrian. In 1029, less than five years after the death of Basil II, the hierarchy of the Syrian Church was denounced in Constantinople by the Chalcedonian metropolitan of Melitene. Imperial messengers were then sent to Chrysoberg, the krites of Melitene, ordering him to detain the Syrian patriarch and his senior bishops. The Syrian clergy were arrested, conveyed to Constantinople and interrogated. As a result of their ordeal some of the bishops adopted a Chalcedonian position; those that did not remained in prison. The patri­ arch himself died in captivity.64 Yet we must ask to what extent this story necessarily demonstrates the increasing oppression of the periphery by an intolerant, Orthodox centre. Other details from the events surrounding the detention of the Syrian church hierarchy indicate that local secular officials of Constantinopolitan origin could still be enthusiastic promoters of the well-being of the popu­ lations under their tutelage. Chrysoberg, the krites, is almost certainly the owner of an unpublished seal in the Dumbarton Oaks collection belong­ ing to John Chrysoberges. The information contained in the legend on this seal indicates that Chrysoberges was not only krites of the city, but also held the offices of kourator and the kankellarios of the genikon.65 Thus, he emerges from the sigillographical evidence as a single official vested with a full portfolio of judicial and fiscal duties. Moreover, like other frontier functionaries such as Philetos Synadenos and Nikephoros Ouranos, he clearly exercised military duties: when he arrested the patriarch he dispatched nine soldiers to undertake the task.66 Like Ouranos and Synadenos, his pedigree was that of a Constantinopolitan administra­ tor.67 Yet, despite his Constantinopolitan roots, the krites could still act as the representative of the best interests of those he administered. According to Michael the Syrian, Chrysoberg arrested the Syrian patri­ arch with considerable reluctance, and only after trying to persuade local 64 Michael the Syrian, 14CM5. 65 John Chrysoberges, spatharokandidatos, kankellarios of the Genikon, krites, anagrapheus and kourator of Melitene (Howard-Johnston, 'Crown', 89, n. 41). 66 Michael the Syrian, 14(M1. 67 For seals of various tenth- and eleventh-century members of the Chrysoberges family active in civil and ecclesiastical administration see Schlumberger, Sigillographie, 285,313 and Nesbitt, Oikonomidès, Catalogue of Byzantine seals I, nos 1.30, 71.13; Laurent, Le Corpus des sceaux, no. 335; Zacos, Byzantine Lead Seals II, no. 57; J.C. Cheynet, 'Sceaux byzantins', nos 63-4.

56

CATHERINE HOLMES

Syrian dignitaries that the patriarch should leave the city for his own safety.68 Various synodal decrees issued by the church authorities in Constantinople in the decade following the arrest and interrogation of the Syrian churchmen corroborate the notion that many local secular officials were reluctant to disturb the heterodox populations of the locality. As Dagron noted, these decrees deplore the willingness of such officials to contemplate marriages between members of different Christian denomi­ nations and their acceptance of the testimony of Syrian Christians in legal cases.69 This evidence indicates that the pragmatism of secular officials rather than episodes of persecution may have been more typical of the exercise of Constantinopolitan authority in the Byzantine east in the decades which followed Basil's death. Although vacuums of imperial legitimacy in Constantinople, such as the short and unpopular reign of Romanos III, could occasionally be manipulated by local agitators such as the Chalcedonian metropolitan of Melitene, officials like Chrysoberg were usually able to use their access to power in the centre to arbitrate successfully between and on behalf of the frontier populations. Certainly the persecution of Romanos Ill's reign seems to have had little effect on the long-term prosperity of the Syrian community in Mesopotamia during the eleventh century. As Dagron has stressed, Syrian monasteries continued to be founded in the Melitene region until the arrival of the Turks.70 How then can we explain the loss in the later eleventh century to the Turks of the heterodox east that Basil himself had won, if not in terms of irretrievable breakdown between a Chalcedonian Constantinople and a heterodox frontier? Without too much flippancy, perhaps I can suggest that the answer may lie partly in the success of the devolved relationships that had been fostered so enthusiastically by emperors such as Basil II. In the predominantly peaceful relationships which typified Byzantine deal­ ings with its eastern Muslim neighbours during the reign of Basil and his immediate successors, the frontier was able to flourish with a minimal Constantinopolitan presence in the locality. However when the more belligerent Turks arrived, this slim-line presence simply proved to be inadequate.71

68 Michael the Syrian, 141 69 Dagron, 'Minorités ethniques', 204. 70 Ibid., 193. 71 See Cheynet's chapter in this volume for the extent to which the arrival of Turkish nomads in the mid- to late eleventh century shattered the pre-existing frontier equilibrium, and for the importance of civil war within Byzantium during the 1070s for the eventual collapse of the position of the empire in the east.

4. La conception militaire de la frontière orientale (IXe-XIIIe siècle) Jean-Claude Cheynet En Orient, durant les IXe-X IIIe siècles, la frontière a été organisée selon la conception que les empereurs et leurs conseillers se faisaient de l'empire,1 le jugeant en expansion ou sur la défensive, les forteresses y tenant un rôle décisif. Les empereurs eurent le choix entre deux types de défense, aux conséquences stratégiques et sociales bien différentes.

Les «frontières naturelles» et la paix durable Les empereurs, quoiqu'on ait dit de certains d'entre eux, comme Manuel Comnène, accusé de démesure, surent toujours estimer les capacités réelles de l'empire. Les éloges impériaux aussi bien que les traités mili­ taires suggèrent que l'empire, dans leur pensée, était borné par des limites naturelles, si l'on peut risquer cet anachronisme de forme, et cette pensée s'enracinait dans la tradition romaine. A coup sûr, les éléments du relief ont toujours facilité l'établissement d'une frontière, que ce fût le Taurus, le Rhodope ou les forêts de Bulgarie. Dans l'esprit des stratèges byzantins, le Danube, en Europe, constituait l'obstacle à ne pas franchir, sauf à mener quelques raids isolés pour prendre à revers les adversaires installés dans les Balkans, et l'Euphrate, pour l'Orient, assumait la même fonction.2 Bien

1 Ce point a déjà été abordé à plusieurs reprises: entre autres, H. Ahrweiler, 'La frontière et les frontières de Byzance en Orient', dans Actes du XÎVe congrès international des études byzantines I (Bucarest, 1974), 209-230. 2 Dans un discours prononcé en 1047, Jean Mauropous précise que les armées intervinrent tardivement contre Tornikios révolté, occupées qu'elles étaient à guerroyer, au-delà de l'Euphrate, contre les barbares (Abu'l swar de Dvin), (P. De Lagarde, Johannis Euchaitarum metropolitae quae in codice Vaticano graeco 676 supersunt, AKKGWG 28 (1882), n. 186, 179). Lorsque les tacticiens byzantins exposent les méthodes de franchissement d'un fleuve, ils prennent en référence le Danube et l'Euphrate (G.T. Dennis, éd. et tr., Three byzantine mili­ tary treatises (Washington, DC, 1985), 62. From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

57

58

JEAN-CLAUDE CHEYNET

après l'installation des Turcs en Asie Mineure, l'Euphrate, souvent associé au Tigre, reste la référence pour les empereurs victorieux. Jean II reçoit les éloges de Théodore Prodrome pour être le premier empereur à avoir ramené la cavalerie byzantine sur les rives du fleuve.3 En Orient, lors de l'expansion du Xe siècle, il n'y eut pas de tentative sérieuse de mettre la main sur la Palestine et l'Egypte, dont la population comptait encore beaucoup de chrétiens. Rappelons aussi combien les empereurs Comnène attachèrent de prix à la reconnaissance par les Croisés de la souveraineté byzantine sur Antioche et sa région, alors qu'ils ne revendiquèrent pas le royaume de Jérusalem, et n'insistèrent pas non plus pour obtenir Édesse. Dans le Caucase, les limites restèrent incer­ taines, car les empereurs hésitèrent un temps à absorber des Etats dont la population était largement chrétienne, mais en partie monophysite, et préférèrent prendre à leur service les aristocrates caucasiens, sur la base du volontariat. Au XIe siècle, cette conception s'exprime clairement car le programme de conquête a été pleinement accompli. Alors que l'armée byzantine est encore toute puissante, la volonté d'expansion fait défaut, sauf en Arménie où il s'agit d'atteindre le Caucase pour interdire aux émirs musulmans voisins les approches des hauts plateaux.4 Les partisans de l'expansion ne se recrutent pas dans les rangs des militaires qui savent combien il en coûte de tenir des pays qui ne correspondent pas à l'aire traditionnelle de l'empire.5 Les plus intéressés à l'acquisition de nouvelles terres seraient plutôt les fonctionnaires civils de la capitale, car ils gèrent à leur profit les cura tories qu'on y établit, particulièrement développées sur la frontière orientale. L'empereur Romain Argyre, issu de ce milieu, fut du reste le seul à envisager une reprise des conquêtes en Orient.6 L'«l'illusion de la paix durable»7 implique une adaptation de l'armée à une politique fondamentalement défensive. Il fallait donc évaluer, sans se tromper, le niveau du danger extérieur. Le surestimer, c'était entretenir, donc payer des troupes en surnombre; le sous-estimer, c'était la porte

3 Theodoros Pródromos, Historische Gedichte, éd. W. Hôrandner (Vienne, 1974), 256, 262, 304. 4 Ces acquisitions n'étaient pas jugées vitales pour l'empire. Face aux Turcs, certains généraux pensaient qu'il fallait se replier sur les vieux thèmes romains et abandonner à leur sort les thèmes arméniens, peuplés d'hérétiques, qui servaient de glacis. En revanche il ne fut jamais question de se replier en deçà du Danube, même aux pires moments des invasions ouzes ou petchénègues. 5 Michel Psellos, Chronographie, éd. É. Renauld (Paris, 19672) II, 114. Isaac savait que «pour de telles annexions, il est besoin de beaucoup d'argent, de bras vaillants et d'une réserve suffisante». 6 Psellos I, 35-6. 7 P. Lemerle, Cinq études sur le XIe siècle byzantin (Paris, 1977), 267-71.

LA FRONTIÈRE ORIENTALE

59

ouverte aux envahisseurs. Les successeurs de Basile II se sont efforcés de mener une politique rationnelle, qui s'inspire, en fait, des choix du souverain que les historiens considèrent souvent comme le plus grand empereur de l'époque médio-byzantine. La possession des forteresses, citadelles ou villes protégées par leurs murailles, résidence des autorités, constitue bien l'enjeu principal, car leur chute entraîne celle des campagnes. De ce point de vue, la similitude entre l'Occident et l'Orient, dans ce domaine comme dans bien d'autres, est frappante. Mais il ne faut pas oublier que la forteresse impose un investissement coûteux pour deux raisons: l'entretien des murs, parti­ culièrement nécessaire dans un pays sismique et le paiement d'une garni­ son permanente. En temps de paix prolongée, les murs ne sont plus réparés et les forteresses restent vides d'hommes, pour des raisons d'économie.8

L'évolution historique Jusqu'en 976

Les opérations en Orient sont assez bien décrites par les chroniqueurs byzantins et musulmans et par les auteurs de traités mili­ taires, notamment le Traité sur la guérilla, écrit sous le patronage de Nicéphore Phocas. Dans la première moitié du siècle règne la guerre acritique. Ce type de combat est trop bien connu pour être ici largement développé. Les populations de la région frontière sont habituées aux raids et prennent leurs dispositions en conséquence. Les paysans cachent leurs récoltes dans des silos souterrains et se réfugient avec leur bétail en des lieux reculés ou fortifiés, ou même dans une véritable forteresse. En cas de surprise, les dégâts sont relativement mineurs. Du côté musulman, on se contente, même au temps de Sayf ad-Daula, d'accomplir le devoir du djihad, sans esprit de conquête. Du côté des Byzantins, l'optique est différente et les terres de l'Est sont à prendre. Comme nous l'avons dit, acquérir un territoire supposait d'en tenir les kastra. Ainsi s'explique l'acharnement du domestique des Scholes, Jean Kourkouas, à s'emparer de Mélitène. Dès la première moitié du siècle, les Byzantins commencèrent donc à grignoter les territoires musulmans et arméniens. L'organisation de la frontière est remarquable­ ment décrite dans les taktika du Xe siècle. De nombreux thèmes furent créés, caractérisés par leur faible étendue, et tous constitués autour d'une ville fortifiée qui donne, en règle générale, son nom à la circonscription.9 8 Sous Isaac II, les révoltés bulgares s'installèrent dans les forteresses de Mysie, dépourvues de garnison, ce qui leur donna une solide position pour reconstituer un État bulgare (Nicetae Choniatae Historia, ed. LA. Van Dieten (Berlin et New York, 1975), 394). 9 N. Oikonomidès, éd. et trans., Les listes de préséance byzantines des IXe et Xe siècles (Paris, 1972), 358-9.

60

JEAN-CLAUDE CHEYNET

Les combattants engagés dans la guerre acritique appartiennent aux thèmes. Les tagmata interviennent rarement, parce que apparemment ils ne sont pas stationnés dans des zones proches des combats. Comme l'a bien vu Gilbert Dagron,10 parmi les soldats des thèmes, il faut mettre à part les troupes d'élite qui entourent les stratèges et ne se distinguent guère, dans le fonds, des professionnels composant les tagmata, et le reste des soldats, sans doute les plus nombreux, qui semblent de médiocre qualité et que le stratège emploie à des tâches moins difficiles, comme la défense des défilés. Les tagmata ne combattent qu'en deux cas précis: d'abord, lorsque les adversaires, Sayf ad-Daula notamment, ont réuni des troupes en nombre tel que les stratèges des thèmes menacés ne peuvent pas ouvertement s'opposer à l'émir. Ensuite, lorsqu'une grande expédition est décidée en territoire adverse pour occuper un point stratégique, tel Adata par exemple, qui contrôle la passe homonyme. La défense de la zone frontière est donc assurée par des Byzantins, Grecs ou Arméniens. Les habitants du cru, fussent-ils des immigrés, tirent profit de leurs victoires ou supportent les conséquences de leurs échecs. Ce positionnement à la frontière convient particulièrement aux officiers qui encadrent ces troupes. Tous sont issus de l'aristocratie militaire rési­ dant dans ces thèmes proches de la frontière, les Phocas, Maléïnoi, Sklèroi, Argyroi, Kourkouas, pour ne citer que les plus influents d'entre eux. Ils mènent des guerres locales, largement à leur guise, loin de Constantinople, alors que les empereurs, depuis Basile Ier, ne conduisent plus les armées en Orient. La défense s'organise donc en profondeur: en avant, un no man's land où passent les armées ennemies, puis une large bande où stationnent les troupes aguerries et rapides des acrites, en arrière, les grands thèmes aux effectifs encore importants; enfin, ultime objectif, la capitale, qui, sauf une démonstration de force de Syméon, le tsar bulgare, n'est plus vraiment menacée par voie de terre au Xe siècle et abrite moins souvent des soldats.

De 976 à 1071 La frontière orientale fut bouleversée par les conquêtes de Nicéphore Phocas et de Jean Tzimiskès. Ils s'emparèrent des grandes cités fortifiées qui avaient constitué les bases de leurs adversaires: Tarse, Adana, Mopsueste, et surtout Antioche avec sa citadelle et sa formidable enceinte. Après d'infructueuses tentatives, dont celle que Basile II entre­ prit en personne contre Tripoli, les limites sud du duché d'Antioche restèrent fixes. Sa frontière orientale fut de même stable pendant un siècle. 10 G. Dagron et H. Mihâèscu, Le traité sur la guérilla de l'empereur Nicéphore Phocas (Paris, 1986), 275-83.

LA FRONTIÈRE ORIENTALE

61

Au XIe siècle, un stratège des villes pareuphratiques, Georges Maniakès, réussit à s'emparer d'Édesse, puissamment fortifiée autour de sa citadelle, d'où il résista à la contre-attaque des chefs arabes locaux. Cette annexion ne prenait pas place dans un plan d'extension systéma­ tique des frontières de l'empire au-delà de l'Euphrate. Elle venait simple­ ment à point pour venger la malheureuse expédition de Romain III contre les Alépins. Les plus grands changements se produisirent en Arménie avec l'annexion du Vaspurakan et des principautés jusque-là indépen­ dantes, notamment le royaume d'Ani. Nous avons vu que dès le Xe siècle avaient été constitués de nombreux petits thèmes frontaliers. Nicéphore Phocas les multiplia en s'emparant de la Cilicie: thèmes de Tarse, d'Anazarbe, de Mopsueste, du Mauron Oros et même, pour quelques mois, d'Antioche, places logeant toutes de nombreuses garnisons. La plus importante fut établie à Tarse car cette cité assuma, pour les territoires nouvellement libérés, la charge que supporta Antioche après 969. On n'a pas assez souligné le poids décisif de Tarse pour les hommes de la frontière, le même que celui de Mélitène au siècle précédent, celui d'un objectif idéal dont la destruction mettrait fin aux attaques contre leurs thèmes. La disparition des Etats-tampons arméniens et syriens plaça les Byzantins directement en contact avec les Fatimides au sud et les Turcs à Test. En conséquence, le gros des armées orientales fut établi sur les fron­ tières. Conserver de petits thèmes comportait le risque de ne pouvoir faire face à une attaque d'envergure, tant la coordination d'unités modestes et dispersées risquait de prendre du temps. Or, si la menace extérieure avait diminué d'intensité, il restait quelque danger face aux émirs d'Azerbaïdjan et surtout face aux califes fatimides, qui firent preuve de dynamisme pendant près d'un siècle. C'est ainsi que s'explique la grande réforme des duchés ou catépanats, sans doute entreprise par Jean Tzimiskès. En dehors de son avantage militaire, une telle réforme permit à l'empereur, parvenu au pouvoir par un coup d'État, de surveiller les officiers mis en place par son prédécesseur: trop nombreux pour être totalement remplacés, ils se trouvaient du moins commandés par des hommes sûrs. Les anciens stratèges subsistèrent, subordonnés aux ducs, mais leurs thèmes furent parfois regroupés, on connaît au moins un stratège d'Adana, Mopsueste et Anazarbe, qui commandait de ce fait toute la Cilicie.11 Les capitales de ces nouvelles et vastes circonscriptions servaient de forteresses de repli et de réserve, en arrière de la frontière. L'une d'elle, Antioche, reprit son ancienne fonction du temps des musulmans, celle de

11 Lecture corrigée par W. Seibt, Die Bleisiegel in Österreich I, Kaiserhof (Vienne, 1978), 261.

62

JEAN-CLAUDE CHEYNET

place de protection (elle comptait au nombre des 'Awâsin), même à l'égard d'autres capitales de thème comme Édesse. Ce dispositif rappelle donc le système musulman établi au lendemain de la conquête face aux Byzantins. Cette mission est soulignée par la présence fréquente du domestique des Scholes. Ainsi, en 1035, le catépan d'Édesse, attaqué par un groupe d'émirs, dont celui d'Amida, fut sauvé par les tagmata station­ nés à Antioche, conduits par le domestique des Scholes, Constantin.12 Sur la frontière du nord-est, une autre ville jouait à coup sûr le même rôle qu'Antioche, mais on ne saurait déterminer laquelle, Sébastè, Théodosioupolis, ou Mantzikert. On peut penser qu'après la prise d'Ani, c'est cette dernière ville qui servait de grande base arrière. Lorsque le Turc Ibrahim înâl mena un grand raid en 1048, ce furent le duc d'Ibérie et celui de Grande Arménie (Ani) qui l'arrêtèrent à Kaputru. Sur la frontière même coexistaient de petits thèmes byzantins face à de petits émirats musulmans, voire quelques États chrétiens du côté de la Géorgie. Telle est la situation décrite par Kékauménos. D'un côté le stratège byzantin, qui dispose de peu de forces, doit se défier de son visà-vis, privé lui aussi des moyens de le réduire, mais susceptible de prendre sa ville par ruse.13 La frontière s'étant déplacée de cent ou cent cinquante kilomètres et étant devenue linéaire avec des bornes précises, les charmes de la guerre acritique s'estompaient; les raids ennemis n'étaient plus à craindre et la mobilisation permanente n'était plus de mise. Les populations de Cappadoce, du Charsianon, voire du Lykandos, étaient trop éloignées de la nouvelle frontière pour espérer lancer des razzias en territoire ennemi, car elles n'avaient pas accompagné massivement l'expansion vers l'est. De même, l'encadrement aristocratique n'avait pas suivi, puisque les empereurs veillèrent à ne pas redistribuer les terres conquises. De plus, les territoires voisins du duché d'Antioche étaient dans l'ensemble solide­ ment tenus par les Fatimides et les émirs locaux avaient bien garni les villes qui formaient un réseau dense, ne laissant pas espérer l'effet de surprise sans lequel il n'est pas de bon butin. Les populations de la nouvelle frontière n'auraient donc pas eu grand-chose à gagner si elles avaient guerroyé avec leurs voisins, sinon quelque bétail pris aux bédouins qui maintenaient dans cette région la tradition des razzias. Il n'y a donc plus de zone vide, même si les bédouins continuent à se faufiler entre les points fortifiés. Nicéphore Ouranos n'eut aucun mal à repousser leur coalition la plus forte, qui avait regroupé 6,000 hommes autour des Noumérites. 12 Skylitzes, 400. 13 G.G. Litavrin, Sovety i rasskazy Kekavmena (Cecaumeni consilia et narrationes) (Moscou, 1972), 168-70.

LA FRONTIÈRE ORIENTALE

63

De surcroît, les populations locales n'avaient aucune tradition militaire. Il s'agit en majeure partie de chrétiens dont l'occupant arabe avait, selon la loi islamique, assuré la protection en échange d'un désarmement. Il ne semble pas que le système du recrutement thématique ait été appliqué à ces nouveaux territoires, comme en témoigne la réforme de Monomaque en Ibérie, lorsqu'il supprima l'armée locale. Dans le duché d'Antioche, de nombreuses garnisons arméniennes furent établies dès l'époque de Tzimiskès et sont en tout cas bien attestées sous Basile II. Cette conception de la frontière répondait bien au modèle mis en place par Basile II. Ce type de défense n'était plus fondé sur la solidarité des grandes familles d'Orient avec leurs soldats et leurs paysans, mais confié à des généraux issus de lignées souvent moins prestigieuses. Ils suivaient les instructions générales données depuis Constantinople, qui laissaient aux ducs le soin de mener des opérations locales. L'initiative était repassée de la frontière au centre. La frontière, défendue par un réseau dense de forteresses, formait une sorte de ligne «Maginot». Les vieux thèmes romains étaient dépourvus de toutes troupes, hormis celles cantonnées durant l'hiver. Cette limite était conçue pour un monde stable, or les Turcs Seldjoukides, qui unifièrent à leur profit une grande partie du Proche-Orient, remplacèrent des émirs aux moyens d'action relativement limités. Mais ils ne constituaient pas une menace grave, car ils visaient au plus des rectifications de frontières avec un empire qu'ils considéraient comme éternel.14 Face aux Turcs nomades, le bilan est mitigé: les garnisons étaient incapables d'empêcher les raids en profondeur, puisqu'une fois franchi le rideau défensif fronta­ lier, plus aucun obstacle n'interdisait aux ennemis de piller au loin, jusqu'à Chônes, par exemple. Cependant les troupes byzantines réus­ sirent à plusieurs reprises à intercepter et détruire des bandes turques, mais d'autres, il est vrai, lui échappèrent.15

De Mantzïkert à la Première Croisade Michel Psellos, dans une lettre adressée à un stratège resté pour nous anonyme, résume admirablement la situation qui s'est développée au cours de la seconde moitié du XIe siècle:

14 Alp Arslan, victorieux, demanda simplement à Romain Diogénès un réaménagement de la frontière: Aristakès de Lastivert, Récit des malheurs de la nation arménienne, tr. et comm. M. Canard et H. Berbérian (Bruxelles, 1973), 128. 15 Avant de condamner ce système, il faut se rappeler qu'au siècle précédent, à l'apogée de la guerre acritique, Sayf ad-Daula avait mené des raids profonds et regagné sans encom­ bre ses bases de départ. Face à l'armée du sultan seldjoukide Toghrul Beg, Mantzikert fut remarquablement défendue par Basile Apokapès. Alp Arslan s'empara d'Ani davantage en raison d'une dissension parmi les assiégés, que par sa supériorité en poliorcétique.

JEAN-CLAUDE CHEYNET

64

il semble ... que les plaines et les collines du territoire romain étaient cachées comme par des voiles. Certains étaient naturels, les précipices, les montagnes et les fleuves, les autres construits de main d'homme, les villes et les forteresses. Le barbare qui poussait son cheval jusqu'à ces obstacles, à leur vue, retenait sa monture et n'osait pénétrer. C'était comme une digue qui leur fermait l'accès à la forteresse. Lorsqu'une telle digue s'est rompue, toute la partie adverse s'est engouffrée chez nous, comme les flots d'un fleuve jusque-là canalisé. Maintenant rien ne sépare le pays romain du pays barbare, mais tout semble mêlé et les gens vivent dans la confusion des races. C'est pour cela qu'ils luttent avec nous, tantôt par une guerre sur l'Euphrate, tantôt par une autre sur l'Istre.16

En Orient, le schéma s'applique avec la plus grande pertinence. Après Mantzikert, il faut distinguer, ce qui, en règle générale, n'a pas été suff­ isamment perçu, entre la frontière «arménienne» et la frontière «syri­ enne». Cette dernière, qui s'étendait depuis Harput jusqu'à la côte méditerranéenne, resta intacte pendant une dizaine d'années et aucune infiltration d'importance ne se produisit dans cet espace bien protégé par le réseau de forteresses mis en place autour d'Antioche à l'époque précé­ dente. Membidj ne retomba pas sous autorité musulmane avant le règne d'Alexis Comnène. La majeure partie des garnisons étant d'origine arménienne, il est naturel que des chefs issus de cette ethnie, en principe tous chalcédoniens, tels Philarète Brachamios, Gabriel de Mélitène, ou Basile Apokapès, un Géorgien, aient pris la direction des opérations. En revanche, au sud de la Mer Noire, les Turcs se déversèrent le long des grands axes routiers jusqu'à la Marmara. On peut se demander pourquoi ils ne furent pas arrêtés. Tout d'abord il y eut bien des tentatives: ainsi Trébizonde, sans doute un temps perdue, fut reprise par Théodore Gabras qui poursuivit son avance vers le sud, en direction de Païpert. Les forteresses du thème des Arméniaques furent occupées par des troupes franques, celles de Crispin, puis celles de Roussel de Bailleul qui s'op­ posèrent avec efficacité aux mouvements des Turcs, à la satisfaction de la population locale. Mais Roussel, engagé dans une guerre civile, fut ramené, prisonnier, par Alexis Comnène. La ville la plus disputée, Sébastè, n'a pas cependant constitué, comme Antioche, un véritable bastion. Bien que nous n'ayons pas d'information précise, notamment des chroniqueurs grecs, nous avons une idée de l'acharnement des combats qui se déroulèrent autour de cette ville par le Dânishmendnâmeh. Ce récit épique, qui ne saurait fournir ni une chronologie précise des événements, ni même une description exacte des adversaires en présence, encore que le rôle de Théodore Gabras y soit mentionné, met en valeur l'importance

16 E. Kurtz et F. Drexl, Michaelis Pselli scripta minora magnam partem adhuc inedita I (Milan, 1 9 4 1 ), 23 9 .

LA FRONTIÈRE ORIENTALE

65

qu'ont accordée les Turcs de Dânishmend à la possession de Sébastè et des villes du thème des Arméniaques. D'autres facteurs encore ont joué. Nous sommes dans une région où la population arménienne est nombreuse et il est possible, quoique non prouvé, qu'elle se soit sentie moins attachée aux autorités constantinopolitaines. Enfin, la raison déterminante ne relève pas de la défense des fron­ tières, mais de la politique intérieure byzantine; les adversaires de tous bords, engagés dans des guerres civiles, ont fait appel à des auxiliaires turcs pour l'emporter, ce qui les a très vite introduits en Asie Mineure occidentale, et même en Europe. La plus lourde responsabilité pèse sur Nicéphore Mélissènos qui avait placé des garnisons turques dans de nombreuses villes de Galatie, de Bithynie, dont Nicée même sans doute, et qui offrit ainsi aux Turcs des citadelles qu'ils auraient été incapables de prendre par leurs propres moyens. À partir de Mantzikert et plus encore de 1081, date de la révolte de Mélissènos, la frontière ne présente plus de limite bien déterminée par une ligne de forteresses. Elle se distingue de l'époque des guerres acritiques en ce qu'on n'y perçoit plus une zone de guerre en avant d'un terri­ toire pacifié et en principe hors d'atteinte de l'ennemi. Aucun fonctionnaire des administrations centrales, pas même celle du fisc, ne pouvait, dans les années 1080, dresser un tableau des territoires appar­ tenant à l'empire. C'est l'époque des fameux toparques,17 Dabatènos en Paphlagonie, Bourtzès en Phrygie. Les Turcs seldjoukides recevaient en permanence des renforts turcomans venant d'Iran, mais n'étaient proba­ blement pas plus aptes qu'auparavant aux guerres de siège. Des forte­ resses, bien situées, restèrent contrôlées par leurs défenseurs pendant des dizaines d'années, le plus fameux exemple étant celui de Kyzistra gardée par les frères Mandalai jusqu'à ce que l'Arménien Thoros le Roupénide s'en emparât en 1111, mais Makrès aurait aussi tenu longtemps Kavala et un certain Constantin, Iconium.18 L'administration militaire renonça aux grandes circonscriptions territoriales: les ducs furent cantonnés à la défense d'une cité et de son territoire immédiat, retrouvant les attribu­ tions des stratèges de petits thèmes du Xe siècle. La frontière orientale a donc quasiment disparu au cours du règne d'Alexis Comnène. Jean d'Antioche dressait en 1091 un bilan négatif des dix premières années du règne d'Alexis Comnène, affirmant, non sans exagération, que l'empire était réduit à l'Ouest comme à l'Est aux murailles de Constantinople. Il fallut mettre au point la défense des terri­

17 Le toparque n'appartient pas à la typologie officielle des fonctionnaires. Le terme désigne un détenteur d'autorité plus au moins autonome. 18 C. Cahen, 'Seldjoukides de Rum, Byzantins et Francs d'après le Seldjuknameh', dans Mélanges Henri Grégoire III (1951), 97.

66

JEAN-CLAUDE CHEYNET

toires restés byzantins. Près de Constantinople, Alexis se contenta de dresser une barrière face aux Turcs de Nicée sur la côte sud de la Marmara, appuyée sur Nicomédie et sur la nouvelle forteresse de Kibôtion. La mer constituait la zone de manœuvre indispensable à ce type de défense. Or les Turcs n'avaient pas de traditions navales. On comprend mieux les efforts désespérés d'Alexis Comnène visant à détrui­ re la flotte construite par Tzachas, à l'aide de spécialistes grecs, non seule­ ment parce que celui-ci menaçait les îles et éventuellement la capitale, mais surtout parce qu'il coupait les dernières lignes de défense en Asie Mineure. En effet, Alexis Comnène avait conservé des bases en Orient. Sans doute les territoires byzantins du Taurus furent-ils perdus lorsque le sultan seldjoukide Soliman les prit à revers, c'est-à-dire sur leur front le moins défendu, puisque l'ennemi venait traditionnellement de l'est et non de l'ouest. Antioche tomba, non pas à la suite d'un siège, mais grâce à un raid surprise. Alexis Comnène contrôlait encore des têtes de pont sur la côte sud de la mer de Marmara, des villes égéennes, très certainement Antalya et Chypre, qu'on peut considérer, et c'est ainsi que les Byzantins la voyaient, comme une grande base arrière. Il ne faut pas juger que toute l'Asie Mineure était perdue dès 1081.19

De la Première Croisade aux Lascarides Le passage des Croisés redonna une nouvelle mobilité à la frontière vers l'intérieur des terres. Une recon­ quête devenait possible et fut entreprise selon les principes habituels: le contrôle des places fortes, à commencer par les ports, puisque la domina­ tion navale des Byzantins n'était plus contestée depuis l'élimination de Tzachas. Ainsi furent reprises Sinope, Smyrne, Laodicée de Syrie, et à partir de là, l'arrière-pays. Après une génération d'absence, Alexis Comnène fit campagne en Orient et son action définit la frontière qui devait perdurer jusqu'aux Lascarides. Il reprit le plat pays jusqu'à ce qu'il rencontrât une trop forte résistance. À ce moment-là, il renonçait à tenir un pays incertain et en évacuait la principale richesse, la population, créant une zone vide d'hommes entre les territoires solidement byzantins et ceux tenus par l'administration régulière de l'état seldjoukide.20 Les successeurs d'Alexis, ses fils et petit-fils Jean et Manuel Comnène, ne changèrent pas fondamentalement de politique. Sous Manuel, la fron­ tière prit son aspect durable. La forteresse représentait toujours l'enjeu

19 Voir en dernier lieu J.-Cl. Cheynet, La résistance aux Turcs en Asie Mineure entre Mantzikert et la Première Croisade, dans EY'f'YXIA Mélanges offerts à Hélène Ahrweiler (Paris, 1998), 131-47. 20 Alexis Comnène transféra en territoire byzantin la population de Philomélion qu'il ne pouvait pas défendre: Anne Comnène, Alexiade, éd. et trans. B. Leib (Paris, 1945), III, 203.

LA FRONTIÈRE ORIENTALE

67

capital, où se réfugiaient les populations menacées par les raids, certes modestes, mais fréquents, des Turcomans. Manuel, poursuivant cet objec­ tif, releva Atramyttion, restée en ruines depuis les razzias turques de la fin du XIe siècle, Pergame et Chliara, créant le thème au nom nettement évocateur de Néokastra. La stratégie de Manuel était claire: défendre, en Asie Mineure, la partie la plus utile et la plus rentable du point de vue fiscal que sont les vallées du Méandre et du Sangarios, les plaines côtières de la Mer Noire, avec Trébizonde comme point d'appui, et celles de la Méditerranée, Pamphylie et Cilicie, avec Chypre comme base arrière. Le plateau central lui échappait donc. Toutefois Manuel voulait d'une part s'assurer la communication directe avec la Cilicie et d'autre part refouler progressivement les nomades turcs de la zone non contrôlée de la fron­ tière. Comme au temps de son père Jean II, cette politique exigeait de s'appuyer sur des places fortes d'où l'armée byzantine pouvait rayonner. C'est lorsqu'il tenta de prendre Amasée au nord, Dorylée en ruines, qu'il fit relever, et Kotyaeion que Manuel subit ses plus graves revers. Myrioképhalon marquait l'abandon des prétentions sur le plateau central, mais Manuel veilla avec succès à ce que son œuvre ne fût pas anéantie. En 1177, les villes du Méandre résistèrent à une offensive de grande enver­ gure lancée avec vingt-quatre mille hommes. A la mort de Manuel, tout l'Orient était doté d'un réseau hiérarchisé de places fortes réparties en profondeur et appuyées sur le rebord du plateau anatolien. Il est assez difficile de dater les nombreux restes de forteresses qui subsistent, notamment en Bithynie.21 Les travaux datent-ils de l'époque de Manuel, dont on connaît l'activité de construction autour de Malagina,22 ou des efforts des Lascarides pour tenir un pays dont ils espéraient faire la base du renouveau byzantin après 1204? Si on avait demandé à Manuel, au sommet de sa puissance, quelles étaient les fron­ tières de l'empire en Orient, il aurait répondu qu'il avait rétabli les fron­ tières du siècle précédent, puisque successivement l'émir danishmendide, le sultan seldjoukide, les princes d'Arménie, le prince d'Antioche s'étaient déclarés ses douloi, reconnaissant ainsi Manuel pour maître. On voit bien la fragilité de cette prétention qui ne reposait plus sur une souveraineté directe sur la population. La défense continua d'être assurée sous les Comnènes et les Anges par des soldats des tagmata. Seuls s'étonneront de ce maintien du système antérieur ceux qui jugent l'armée de métier moins efficace que les milices locales. Les empereurs avaient une opinion contraire et préféraient

21 C. Foss, 'The Defenses of Asia Minor against the Turks', GOxTR 27 (1982), 145-205, repris dans Cities, Fortresses and Villages of Byzantine Asia Minor (Aldershot, 1996), no. V. 22 C. Foss, 'Byzantine Malagina and the lower Sangarius', AnatStud 40 (1990), repris dans Cities, no. VII.

68

JEAN-CLAUDE CHEYNET

toujours les soldats professionnels, dont de nombreux étrangers. La proportion de ces derniers nous échappe, mais il est possible que le nombre de Grecs enrôlés dans les tagmata n'ait pas été négligeable au XIIe siècle.23 La présence d'une armée de métier ne permit pas la naissance d'une nouvelle génération d'acrites, à la différence de l'époque où les Arabes menaçaient l'Asie Mineure. Il fallut attendre le repli des Byzantins dans les provinces de Bithynie et des Thracésiens après 1204, pour que la frontière appartienne de nouveau aux soldats acrites qui répondaient avec succès aux raids limités des nomades turcomans et s'enrichissaient même au cours des opérations. La reprise de Constantinople et les contraintes budgétaires qui en découlèrent mirent fin à cette expérience, pour le plus grand malheur des possessions byzantines d'Asie Mineure. On construisit encore de nouvelles forteresses. Angélokastron fut érigée par Isaac Ange pour bloquer les raids turcomans dans la haute vallée du Méandre; mais en fait rien ne changea dans la stratégie byzan­ tine, qui se maintint constamment sur la défensive. Compte tenu des circonstances, les guerres civiles entre Andronic et les villes de Bithynie, le passage difficile de la Troisième Croisade, le dynamisme accru des Turcomans et l'apogée des Seldjoukides, les Byzantins sauvèrent l'essen­ tiel, même si la Cilicie leur échappa, ce qui était inévitable puisque la base-arrière, Chypre, était elle-même tombée aux mains d'un usurpateur. L'administration byzantine subdivisa les thèmes de Manuel en petites unités pour des raisons peu claires, puisque à la différence du siècle précédent, le territoire n'était pas bouleversé par les invasions. Peut-être s'agissait-il d'une mesure de protection du pouvoir impérial qui avait été précédemment emporté par une armée venue d'Orient, celle d'Andronic Comnène, et parfois menacé, ainsi par le grand domestique, Jean DoukasVatatzès, révolté précisément contre ce même Andronic. Rien de neuf donc, sinon le rétrécissement du territoire byzantin avec la chute de quelques places fortes, Dadibra, Gangres, Sôzopolis; Antalya est de plus en plus isolée. De petites entités indépendantes, tirant profit de l'affai­ blissement de l'autorité centrale et de l'abondance de troupes turques prêtes à s'engager, apparaissent sur la frontière, Philadelphie autour de Mangaphas, le thème de Mylasa-Mélanoudion, sous Michel Ange Doukas. Le territoire impérial se réduit, mais l'essentiel des terres riches est encore sauvegardé ainsi que la continuité territoriale le long de la Mer Noire, bien que les Seldjoukides approchent dangereusement de la côte.

23 LoLorsqu'Isaac II Ange parvint au pouvoir, il reçut l'appui de nombreux soldats orien­ taux, en large partie des autochtones (Nicétas Chôniatès, 357).

LA FRONTIÈRE ORIENTALE

69

Conclusion L'organisation de la frontière orientale correspond à l'image que Byzance se fait de ses adversaires: ou bien la frontière forme une zone mobile, destinée à avancer vers l'Orient, ou bien elle demeure figée, puisqu'aucune expansion n'est envisageable. A partir du Xe siècle, en obtenant la supériorité stratégique, les empereurs renoncèrent à la défense en profondeur, qui avait caractérisé l'armée des thèmes, et choisirent de développer une sorte de glacis protecteur, constitué d'un réseau frontalier de forteresses, appuyé sur quelques grandes places d'armes situées en arrière du front. À la différence de l'époque romaine, ce furent les meilleures troupes, pas très nombreuses, qui y furent stationnées. Cette solution, logique dans un monde devenu moins dangereux, se révéla inopérante face à un adver­ saire, inconnu et sous-estimé, qui réussit à percer cette cuirasse. Il ne faut pas condamner a posteriori l'aveuglement des responsables, car dans les thèmes d'Occident, où les mêmes principes avaient été appliqués avec une moindre rigueur, faute d'avoir connu la paix assez longtemps, la résistance fut finalement plus efficace: il y avait en effet davantage de forteresses en état de s'opposer aux ennemis qui avaient franchi les premières lignes de défense. Les empereurs donnèrent souvent la priorité à la défense d'une frontière, européenne ou asiatique, par rapport à l'autre: Nicéphore Phocas et Jean Tzimiskès se préoccupèrent de l'Orient, tandis que Basile II fit avant tout porter ses efforts sur l'Occident. Après un siècle purement défensif, au XIIe siècle, Jean II privilégia l'Orient, tandis que Manuel, durant la plus grande partie de son règne, se tourna vers l'Occident. Enfin, l'organisation de la défense reflète fidèlement l'évolution des rapports entre le pouvoir central et les populations de la frontière ainsi que leurs cadres.

This page has been left blank intentionally

Section II History writing in the east

This page has been left blank intentionally

5. Some reflections on Seljuq historiography

Carole Hillenbrand Introduction In the eleventh century the Seljuqs, a Turkic dynasty of nomadic origin, conquered large areas of the eastern Islamic world, parts of central Asia, Iran, Iraq and Syria, as well as new lands in Anatolia. The Seljuq rulers quickly presented themselves as upholders of Sunni Islam. Their empire remained broadly unified until 1118; thereafter, centrifugal forces inher­ ent in the nomadic heritage of the Seljuqs fragmented their polity. Seljuq history is interwoven with the history of several other empires and dynas­ ties - Byzantium, Seljuq successor states in Syria such as the Zengids and Ayyubids, the Anatolian Turcoman dynasties, the Fatimids of Egypt who were the main opponents of the Seljuqs in the eleventh century, the Crusader states, and other groups in central Asia. In Anatolia the Seljuqs formed part of the mosaic of Turcoman dynasties vying for power on the eastern Byzantine borders and they interacted with Byzantium from the eleventh to the early fourteenth century. Given the time span of Seljuq power and their vast empire, it is not surprising that their history should have been written from a number of perspectives in widely varying terri­ tories.

Definitions and parameters 'Seljuq historiography' will be defined in this chapter as the historiogra­ phy of the Great Seljuqs (c. 1030-1194), who provided the principal inspi­ ration for subsequent historiographical traditions in Anatolia and Syria. Of course, the historiography of the Seljuqs of Anatolia (c. 1077-1307) is inextricably linked to the eastern tradition, as indeed is the historiography of the Seljuq successor states in Syria and that of the Mamluks of Egypt. Indeed, Seljuq traditions - cultural, military and governmental - were From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

73

74

CAROLE HILLENBRAND

translated wholesale into Anatolia and Syria. So too was the art of writing history. 'Historiography' will be taken to denote the various Islamic genres of writing which touched more or less directly on the history of the Seljuqs - 'Universal Histories', dynastic histories and town chronicles.1 The term 'Seljuq historiography' does not, however, mean historical writing about the Seljuqs by the Seljuqs themselves in some form of Turkic language. The written history of the Seljuq Turks exists only in Arabic and Persian sources, composed by Arabic and Persian court officials or religious scholars with their own concerns and preoccu­ pations and within their own historiographical traditions.2 Unfortunately, there are no extant Islamic sources from the period 1050 to 1150 which record the history of the Seljuqs. As was often the case in the medieval period, there is a sizeable time lag between events and their first appearance in the historical record.

The Persian historiographical tradition The first historical works in New Persian appeared suddenly in the middle of the tenth century in Khurasan and Central Asia, areas under the rule of the Samanid dynasty.3 They were written by court officials, religious scholars and others. It was usually the case that such historians engaged in some other activity as well as history.4

1 Auxiliary historical works such as biographical dictionaries, 'Mirrors for Princes', geographical works and encyclopedias, and even poetry, are often valuable resources for Seljuq history. 2 Comparatively little work has been done on Seljuq historiography. As Humphreys rightly points out; This literature has not been adequately studied', see R.S. Humphreys, Islamic history: a framework for inquiry (Princeton, 1991), 165. The pioneer work of Cahen on this subject is still useful for the names of writers and their works; see C. Cahen, The histo­ riography of the Seljuqid period', in B. Lewis and P. M. Holt, eds, Historians of the Middle East (London, 1962), 59-78. See also idem, 'History and historians', in M.J.L. Young, J.D. Latham and R.B. Serjeant, eds, The Cambridge history of Arabic literature. Religion, learning and science in the ‘Abbasid period (Cambridge, 1990), 165-8. The research of Kafesoglu was also a land­ mark; see I. Kafesoglu, 'Tiirkiye'de Selquklu tarihqiligi' in Cumhuriyetin 50. yilina armagan (Istanbul, 1973), 83-92: trans. G. Leiser as 'Seljuk historiography in Turkey', International Journal of Turkish Studies 3 (1985-86), 131-6. 3 In the field of Persian historical writing important new research has been carried out very recently by Julie Meisami. See, for example, Julie Scott Meisami, 'Why write history in Persian? Historical writing in the Samanid period', in C. Hillenbrand, ed., Studies in honour of Clifford Edmund Bosworth II: The Sultan's Turret (Leiden, 1999), 348-74; and above all, eadem, Persian historiography to the End of the Twelfth Century (Edinburgh, 1999). See also M.A. Waldman, Towards a Theory of Historical Narrative. A Case Study in Perso-Islamicate Historiography (Columbus, 1980), and K.A. Luther, 'Islamic rhetoric and the Persian histori­ ans, 1000-1300AD', in Studies in Near Eastern Culture in memory of Ernest T. Abdel-Massih (Ann Arbor, 1990), 90-98. 4 See Meisami, 'Why write history in Persian?', 348.

SELJUQ HISTORIOGRAPHY

75

An important aspect of the Persian historiographical tradition is its obvious didacticism. Scholars have been dismissive about the value of Persian historiography on the Seljuqs, criticizing it for presenting diffi­ culties for those in search of 'factual data' with which to reconstruct the past.5 But this is to misinterpret the purpose of such a tradition, which aims to present exemplary history and in particular to show the nature of true kingship. The influence of the Fiirstenspiegel genre is very strong.6 This dates from the Sasanian period in the first instance and in fact even further back, to Baylonian times. But here it is given an Islamic dress. Kings are chosen by God for the good of mankind. They are endowed with divine charisma (farr-i ilahi) and they dispense justice in accordance with God's decrees. This 'rhetorical' history seeks out striking examples to reinforce these themes and uses direct discourse and interpolated material, such as Qur'anic and hadith quotations. Those who write this kind of history are concerned to show the latent of patterns of the events themselves and to elucidate their underlying 'meaning'. It is difficult to identify the audience for such works. The device of the sultan being given counsels in Persian on how to rule stretches credulity in the case of the early Turkish-speaking Seljuq sultans, although as time went on they may well have understood something of the works dedi­ cated to them. It is more likely that these works were intended for a Persian-speaking intellectual elite, in Iran and later in Anatolia. The Persian historical sources on the Seljuqs were clearly intended for decla­ mation, not just for reading, and they are literary in style and structure, forming part of the belles-lettres tradition. Meisami suggests convincingly that such works were targeted at those who did not know Arabic and that the promotion of Persian historical writing may well have been a means for sovereigns to legitimize their claims to rule in the eastern Islamic world.7 It is important to stress the approach and value of the Persian historio­ graphical tradition both for its own intrinsic importance within the eastern territories of the Seljuq empire and also because it was this Persian tradition - not the Arabic one - which was transplanted lock, stock and barrel into Seljuq Anatolia and which persisted there until Ottoman times. In his history of the early Ottoman sultans entitled Hasht Behesht (Eight 5 The comments of Gibb are typical of this negative judgement: 'History became a work of artifice ... Their [the works of history] bombast and lack of judgement make the most unfavourable impression' El1, Supplement, art. 'ta'rikh', 239). For similar views, see also G. von Grunebaum, Medieval Islam (Chicago, 1962), 282-3, and I. Afshar, in the introduction to his edition of Nishapuri, Saljuqnama (Tehran, 1953), 6. 6 For a recent overview of the Fiirstenspiegel genre, see C.E. Bosworth, 'Administrative literature', in The Cambridge History of Arabic Literature. Religion, learning and science, 165-8. 7 Persian historiography, 364.

76

CAROLE HILLENBRAND

Paradises)8 composed in Istanbul in the early sixteenth century, Idris Bitlisi writes in the same spirit of didacticism and in the same high-flown, sophisticated Persian style that Ravandi had used several centuries earlier, first in Iran and then in Konya. The genres used by Seljuq historians writing in Persian fall into two broad categories - dynastic histories and local chronicles. The key sources which deal with the Seljuq rulers of Iran and Iraq - Nishapuri and Ravandi - are dynastic histories which cover the whole sweep of events from the legendary origins of the dynasty until its collapse. Local chroni­ cles of individual cities, such as the work of Ibn Funduq,9 are of some additional interest but inevitably they are rather narrow in their focus. A rarity in Seljuq historiography, but largely outside the time frame of this chapter, is Ibn Bibi's history of the Seljuqs of Anatolia, written in 1281, which takes the form of memoirs and deals with the years 1192-1280.101

Two key Persian authors on the Seljuqs Nishapuri (d. 1187) The work of Nishapuri, the Saljuqnamaru became the Urtext of all later Persian histories of the Seljuqs, such as those of Ravandi, Hamdallah Mustawfi, Mirkhwand and others. Nishapuri worked at the Seljuq court where he was the tutor to two Seljuq princes. He wrote his strongly pro-Seljuq chronicle around 1175, probably shortly after the accession of the last Seljuq ruler of Iran, Tughril. His dilemma was that of the highly educated Persian bureaucrat attempting to come to terms with the Turkish presence. He charts the history of the Seljuq dynasty from their rise to power in Khurasan until his own day, when Seljuq decline is all too apparent. Despite its straightforward and unembellished style, the work is heavily impregnated with didactic elements. This is particularly true of the short preface in rhyming prose in which Nishapuri reflects on kingship in general and the government of the Seljuqs in particular. Although the Seljuqs were Turkish nomads, he places them within the ethos of Perso-Islamic kingship. According to Nishapuri: 'Kings are shep­ herds of the flock and the protectors of creatures from all kinds of calamity and dread/12 Present rulers should learn from studying the past:

8 This vast work, available in many manuscripts, has still not been edited. 9 Ibn Funduq, Tar'ikh -i Bayhaq, ed. A. Bahmanyar (Tehran, 1965). 10 Ibn Bibi, Al-awamir al-'ala'iyya fi'l -umur al-'ala'iyya, published as Histoire des Seldjoucides d'Asie Mineure d'après Ibn Bibi (Recueil de textes relatifs a l'histoire des Seldjoucides), iii, ed. M.Th. Houtsma (Leiden, 1902); trans. H. Duda, Die Seltchukengeschichte des Ibn Bibi (Copenhagen, 1959). 11 Nishapuri, Saljuqnama, ed. I. Afshar (Tehran, 1953). This is not always an accurate edition. A new edition is being prepared by A. Morton. 12 Ibid., 9.

SELJUQ HISTORIOGRAPHY

77

There were no sultans greater or kinder to the flock or more worthy in their regard for the people than the kings of the House of Seljuq.' Writing in the period of Seljuq decline, Nishapuri laments the lapse in piety and good government which characterized the rule of earlier Seljuq rulers: 'Were the kings of this age to emulate their exemplary conduct, this would result in the strengthening of (both) religion and rule and the establishment of the realm on firm foundations.'13 The work ends in panegyric mode; Nishapuri expresses the hope that: 'all the regions of the world will be under the command of the Seljuq sultan'.14

Ravandi (flourished until the early thirteenth century)

Ravandi lived at the end of the Seljuq period. His only surviving work is the Rabat al-Sudur, a dynastic history of the Great Seljuqs.15 Ravandi would have wished to dedicate his work to a Persian Seljuq ruler but after the demise of the dynasty in Iran proper in 1194, he looked for patronage from Konya, wanting his book to be in the 'name of a Seljuq sultan'.16 Indeed, he went there personally to present his work. The reorienting of Ravandi's history towards Anatolia is a clear sign that early thirteenth-century Persian writers considered the Anatolian Seljuq dynasty to be the new champions of Sunni Islam, and Konya the centre for the continuation of Persian scholarly traditions. Ravandi's work is inflated and derivative; much of it is based on that of Nishapuri. In the past scholars have not rated it highly. They do not like its padding and excess material.17 Indeed, entire chapters are devoted to non-historical information. The work is a 'compilation' and the crucial fact is that it does not have 'history' as its prime aim. It is divided into three parts - a long introduction, a historical section and finally the non-histor­ ical chapters on topics as varied as chess, horsemanship and calligraphy. The style is full of rhetorical devices. There is no denying that these get in the way of 'straight history'. Interpolations include Qur'anic quotations, hadith, proverbs, poetry (especially the Shahnama) and edifying anecdotes. Ravandi's angle is apparent from the very beginning. The introduction praises the Seljuq Turks, devout Muslims who defend the True Faith with the sword. Sultan Kaykhusraw is praised as 'the fruit of the tree of Seljuq: a tree whose root is the strengthening and propagation of the faith'.18 13 Ibid., 10; the second quotation is translated by J. Meisami in Persian historiography, 230-1. 14 Ibid., 83. 15 Ravandi, The Rabat -us-Sudur zva 'Ayat -us-Surur, ed. M. Iqbal (London, 1921). 16 Ravandi, Rabat al-sudur, 62. Originally he had dedicated the work to Sultan Sulayman II (d.1204) but with the latter's death he shifted his panegyrics to his successor, Kaykhusraw (Ibid., 19-38). 17 See the comments of Iqbal in the introduction to his edition. 18 Ravandi, Rabat, 29.

78

CAROLE HILLENBRAND

The principles are those of the 'Mirrors for Princes' genre. The histori­ cal part serves to legitimize Seljuq rule.19 Ravandi's work is better struc­ tured than that of Nishapuri, who serves up a continuous and plain account. Ravandi uses events as paradigms. For example, when the early Seljuq vizier, al-Kunduri, comes to a grisly end, this may be interpreted as a presage of the ultimate decline of the Seljuq dynasty.20 The exemplary nature of Ravandi's account of the battle of Manzikert is clear from the issues which he selects for emphasis: a slave captures the Byzantine emperor, a small force successfully takes on the enormous Byzantine army, the eventual blinding of Romanos IV Diogenes can be seen as a metaphor for the darkness of unbelief. Above all, God's inexorable purpose for the world involves the ultimate victory of Islam. All in all then, the work is intended to show the sultan how to rule,21 and it deploys both precept and example for this purpose. Thus the book has a sermonizing tone, and Ravandi knows just how to milk an event so as to point a moral and adorn a tale. Present rulers would do well to derive lessons from a study of the past: 'When they become aware of that and they read and come to know about the life and behaviour of each one [of the past rulers], they will choose that which is the epitome of the virtuous conduct of those who have gone before (them).'22 The conclu­ sion of the book looks forward to a revival of Seljuq fortunes in Anatolia. Ravandi's history is, then, above all, a well-structured work - aimed at edifiying and at showing recurring patterns and eternal truths. Despite its intrinsic linguistic difficulty, it probably enjoyed popularity amongst the Persian intelligentsia in Seljuq Anatolia and was translated into Turkish in the reign of the Ottoman sultan Murad II (1421-51).

The Arabic historiographical tradition The thirteenth century witnessed a surge of historical material about the Seljuqs,23 much of which is still extant. The Arabic historiographical tradi­ tion comprised a variety of genres - 'Universal Histories' from the Creation until the time of the author, dynastic histories and town chroni­ cles.24 This Arabic tradition is much more diverse and territorially wide­

19 Especially vis-à-vis their predecessors, the Ghaznavids, and the Abbasid caliphate. 20 Ravandi, Kahat, 117-18. 21 J. Meisami, 'Ravandi's Rabat al-sudur: History or Hybrid', Edebiyat, n.s. 5, 201. 22 Ravandi, Rabat, 65. 23 See the table in the appendix to this chapter. 24 Mention should also be made of the various biographical dictionaries of the thirteenth and fourteenth centuries. Noteworthy in the Seljuq context is the vast biographical dictio­ nary of Ibn al-'Adim, Bughyat al-talab, which contains 8,000 entries. This author quotes frequently from lost sources of the Seljuq period, such as al-Hamadhani.

SELJUQ HISTORIOGRAPHY

79

spread than the Persian one, owing to the predominance of Arabic within the medieval Islamic world as the language of prestige. It allowed both a plain and an ornate style. The 'Universal Histories' and local chronicles are written in an unpretentious style and annalistic format, and this pattern endured for centuries. The tradition of high-flown rhymed prose in Arabic historiography, as epitomized in the writings of Saladin's biographer, 'Imad al-Din al-Isfahani, was short-lived.25 The great Mamluk historians of the fourteenth and fifteenth centuries followed the simpler way of writing.

Three key Arabic authors on the Seljuqs Al-Bundari (flourished in the 1220s)

'Imad al-Din al-Isfahani, famous as a historian of Saladin and the conquest of Jerusalem, made in 1183 a revised Arabic version of the memoirs of the years 1072-1134 written in Persian by Anushirwan (d. 1138/9), the vizier of the 'Abbasid caliph, alMustarshid, and later of the Seljuq sultan Mas'ud. This Arabic reworking, written in high rhetorical rhymed prose, has come down to us in an abridged version made by al-Bundari, a fellow Persian who, like 'Imad alDin, had moved to Syria. The work of al-Bundari contains important information but it is difficult to read because of its retention of much of his predecessor's style.26

Akhbar al-dawla al-saljuqiyya The authorship and provenance of the only Arabic dynastic history of the Seljuqs, entitled Akhbar al-dawla al-saljuqiyya, remain controversial.27 It is an interesting and often racy chronicle cover­ ing all periods of Seljuq rule. The author's attitude to the Seljuqs is gener­ ally favourable. His sources must have been very varied. This is mirrored in the wide range of Arabic styles found in his text. At times the writing is simple and unvarnished; on other occasions the language is full of hyperbole. It is clear that the author is taking passages from earlier sources and piecing them together. His aim was to emphasize the Islamic credentials of the Seljuqs.

Ibn al-Athir (d. 1233)

Mention should be made, above all, of Ibn al-Athir and his Universal History, an annalistic work which provides, among much else - for the entire Islamic world was the author's purview - the most coherent account of the Great Seljuq sultans. It also represents a high

25 For a recent study of Im ad al-Din al-Isfahani, see L. Richter-Bernburg, Der syrische Blitz. Saladins Sekretär zwischen Selbstdarstellung und Geschichtsschreibung (Beirut, 1998). 26 Al-Bundari, Zubdat al-nusra wa nukhbat al-'usra, ed. M.Th. Houtsma (Leiden, 1889). 27 Akhbar al-dawla al-saljuqiyya, ed. M. Iqbal (Lahore, 1933). For a long time the work under this title was attributed to al-Husayni but it seems that this valuable source in its present form is in fact an abridgement of a longer history called Zubdat al-tawarikh which was written by al-Husayni. But more research needs to be done on this historiographical problem.

80

CAROLE HILLENBRAND

point of Muslim historiographical achievement. His presentation of infor­ mation is clear and balanced, and his narrative is enlivened by a notable breadth of vision.28

The differences between the Arabic and Persian historiographical traditions The Arabic tradition on the Seljuqs, as exemplified in the Universal History of Ibn al-Athir or the local chronicles of Syria, with its sober and unpretentious style, makes it much easier to use than the Persian works in the form and style of Ravandi. The generally annalistic framework of the Arabic tradition makes the extraction of 'facts' much easier for histo­ rians and there is less poetry to obstruct or demoralize researchers. This is not, of course, to suggest for one moment that the poetry is not serving an important function in the view of the medieval author who chooses to place it at key points in his text. The Persian tradition, an 'ethical-rhetori­ cal historiography', is without doubt literary in nature, a branch of belleslettres (adab) which aims to entertain and instruct and which was designed to be read aloud and relished as a literary performance.29

Two case studies In keeping with the focus of this volume, two specific episodes of Byzantine history, as seen through the eyes of the Islamic sources, will now be considered. They will, it is hoped, serve as models on the basis of which the generalizations about Seljuq historiography made in this chapter can be tested.

The battle of Manzikert, 1071 The battle of Manzikert is recounted in numerous Islamic sources which come from a wide geographical area and chronological span.30 Just two of these accounts will now be considered. 28 Ibn al-Athir, Al-Kamil fi'l -tar'ikh, ed. C.J. Tornberg (Leiden, 1853), XII, 124-6. Regrettably Ibn al-Athir does not specify what his sources are, although some of them can be deduced. For example, he used a lost Persian source, the Maliknama. It is not clear whether Ibn al-Athir knew Persian or not. Certainly, as Richards points out, Cahen gives no reasons for his assertion that 'he [Ibn al-Athir] was apparently ignorant of Persian' (see Cahen, ' Historiography', 71; D.S. Richards, 'Ibn al-Athir and the later parts of the Kamil: a study of aims and methods', in D.O. Morgan, ed., Medieval historical writing in the Christian and Islamic worlds [London, 1982], 88). 29 Some of the Persian historiographical works written in Seljuq Anatolia, such as those of al-Anawi and Aqsarayi, remain relatively unexploited presumably because of their inherent linguistic difficulty. 30 See the contribution by S. Vryonis Jr in this volume (Chapter 1). Both he and I are preparing books on the battle of Manzikert. See also El2, art. 'Malazgird' (C. Hillenbrand); S. Vryonis Jr, 'A personal History of the History of the Battle of Manzikert' in N. Oikonomides, ed., Asia Minor (Athens, 1998), 225^14.

SELJUQ HISTORIOGRAPHY

81

(i) The narrative of Ibn al-Qalanisi (d. 1160) The first extant Muslim account of the battle, that of Ibn al-Qalanisi writing from Damascus, dates from the middle of the twelfth century.31 It is short and unadorned. It may be summarized as follows. The Seljuq sultan Alp Arslan defeated the Byzantine emperor at Manzikert. The Byzantine army was vast in size, whilst the Seljuq forces were smaller.32 The emperor was taken prisoner. He was subsequently released after a peace treaty had been signed. He was sent back to Byzantium, where he was blinded. It is difficult to assess whether this account is fragmentary because the author lacked more detailed information or because the battle, although worthy of some mention, was outside the focus of a local town chronicle. The propaganda potential of this narrative is left unexploited. Certain key elements which will recur in later accounts are, however, already estab­ lished here: the superior numbers of the Byzantine army, the victory of Alp Arslan, the capture and release of Romanos Diogenes and his subse­ quent fate in Byzantium. It is noteworthy that the date is not mentioned. The style of the passage is unpretentious. Ibn al-Qalanisi shows no special interest in the course of the battle itself. His account, short as it is, focuses rather on the events before and after it. Above all, the notion of Manzikert as a hinge of Islamic history is simply not there.33 (ii) The narrative of Ibn al-Jawzi (d. 1200) The account of Manzikert given by the Baghdadi scholar Ibn al-Jawzi is full of new 'details' and elaborates some of the core themes found in the account of Ibn al-Qalanisi.34 Fully fledged Islamic themes are now incor­ porated into the account. Alp Arslan addressses his troops like Elizabeth before the coming of the Armada.35 His speech to the Seljuq troops is a carefully constructed set piece, positioned for dramatic effect before the battle itself: When it was the time for the prayer on the Friday, the sultan prayed with the troops. He called on Almighty God and made humble supplication. He wept and abased himself and he said to them [the troops]: 'We are with reduced numbers of men. I want to throw myself at them [the Byzantines] 31 The sources on which Ibn al-Qalanisi drew for this account are unknown but it is likely that he had access to the work of the Baghdadi historian, Ghars al-Ni'ma b. Hilal al-Sabi' (d. after 1077). 32 The figures given are absurdly inflated for both sides: 600,000 in the Byzantine army and 400,000 Seljuq forces. 33 The nationalist interpretation of Manzikert as the decisive moment which shaped the history of Turkey is a largely twentieth-century one. 34 Ibn al-Jawzi, Al-Muntazam fi ta'rikh al-muluk wa'l-umam, ed. unidentified (Hyderabad, 1940), VIII, 261-3. 35 Muntazam, VIII, 262.

82

CAROLE HILLENBRAND when prayers are being said for us and for the Muslims on the pulpits. I will either achieve my aim or I will go as a martyr to Paradise. So he amongst you who wants to follow me let him follow me and he who wants to leave let him leave/

The speech ends with the pronouncement: Today I am only one of you and I am fighting alongside you. He who follows me and gives himself to Almighty God, Paradise and booty will be his. He who leaves [the battle­ field] his due will be the Fire and ignominy'.36 The role of the speech is to point to the significance of the battle in Islamic terms and to link it to God's wider purpose for the world - the ultimate victory of Islam over Christianity. In the manner of Thucydides, the Muslim chronicler 'puts into the mouths of the speakers the senti­ ments he thinks proper to their situation'.37 Alp Arslan is presented as the paradigm of the Muslim warrior of the faith fighting in the path of God, in the hope of the reward of Paradise. The awesome impedimenta of the Byzantine army are stressed - 400 waggons carry weapons, saddles, ballistas and mangonels - in order to enhance the victory of the depleted Seljuq forces. Ibn al-Jawzi gives a precise place and date for the battle during which Romanos is captured by a slave and brought to Alp Arslan. He is released after terms for peace have been arranged. The natural antithesis between the mighty Byzantine emperor and the Muslim slave who captures him is exploited for all it is worth. Romanos and Alp Arslan converse in a stylized discourse conducted in high Arabic. Alp Arslan, the unlettered Turkish nomad, treats the Byzantine emperor honourably and releases him. Thus the eventual victory of the Seljuq sultan is not just a military triumph over the powerful Byzantine empire. The symbol of the captured Byzantine emperor is very potent: Christianity is subjugated to Islam. It is interesting to note that this speech, put into the mouth of the Turkish sultan Alp Arslan for the year 1071 by a late twelfth-century Baghdadi Arabic author, is a prototype for the Ottoman gazanama genre.38 This was a popular genre in Ottoman times, depicting a campaign of the ruler in which he is shown as the warrior sovereign, the sultan of the jihad fighters, extending the frontiers of the Ottoman empire and the House of Islam. The sultan's public prayers before the battle inspire his troops to

36 Ibid. 37 K. Egan, Thucydides, tragedian', in R.H. Canary and H. Kozicki, eds, The writing of history (Madison, 1978), 79-80. As Egan points out, 'the relationship with what was actually said on particular occasions varies considerably', ibid., 77. 38 See A.S. Levend, Gazavatnameler ve Mihaloglu Ali Beyin Gazavatnamesi (Ankara, 1956), 15-177.

SELJUQ HISTORIOGRAPHY

83

victory.39 Here, then, are the germs of what in Ottoman times became a fully fledged historiographical genre intended to entertain and edify. It is difficult to say how far back this historiographical tradition may go. Certainly speeches with Qur'anic resonances, positioned at moments of high psychological tension, have a long pedigree in Islamic historical texts. It is significant, moreover, that Ibn al-Jawzi was a famous preacher and prolific writer of religious treatises, and it is always possible that the speech may have been inserted by him for the first time into the account of the battle of Manzikert. In this way, the Islamic credentials of the Seljuq Turks are reinforced; these relative newcomers to the faith are placed firmly within an Islamic context. The Turkish nomad is transmuted into a Muslim leader fighting jihad. The chronology of events is vague - the two sides seem to have met on a Wednesday but the battle itself is placed firmly on a Friday on which Alp Arslan is mentioned as praying with his troops before the outset of the battle. Unusually, however, amongst the Muslim sources as a whole, Ibn al-Jawzi manages to provide a date which corresponds to an actual Friday (27 Dhu'l Qa'da 463/26 August 1071). Friday, the Muslim day of prayer, becomes a topos in subsequent accounts of the battle. It should be noted that the account of Ibn al-Jawzi is reproduced in its broad outlines by most later Muslim authors who mention the battle.40 It might be said, therefore, that his is the canonical narrative. (iii) General comments The preceding Islamic accounts of Manzikert, and others which have not been mentioned here, reveal the following points. The Islamic accounts of the battle are very vague about the military details and the course of the battle itself. The date of it remains highly debatable. If we therefore attempt to reconstruct the battle from Muslim sources alone, we have very little concrete information. So far as raw data are concerned, the Muslim sources have little to offer about Manzikert as a battle. This rare interest41 on the part of Islamic historians in the relationship between the Islamic world and Byzantium stems rather from the propa­ ganda potential of the battle. Twelfth- and thirteenth-century Muslim historians eagerly seize on Manzikert as a focus for their accounts of the history of the Near East under the domination of the Turks, whom they

39 For a discussion of the Ottoman gazaname tradition, see C. Woodhead, 'Perspectives on Suleyman', in M. Kunt and C. Woodhead, eds, Suleyman the Magnificent and his age (London, 1995), 172-3. 40 For example, his grandson Sibt b. al-Jawzi, Ibn al-Athir and Ibn al-'Adim. 41 An exception to this generalization is Ibn Bibi, for whom Byzantine proximity and involvement in the affairs of the Anatolian Seljuqs occasions no surprise or explanation. It is a fact of life.

84

CAROLE HILLENBRAND

wish to endow with good Islamic credentials. A subsidiary but significant factor for Muslim writers who have lived through the heyday of Muslim jihad against the Crusaders is the association of this victory at Manzikert over a Christian enemy, the Byzantine emperor, with subsequent Muslim victories over the Crusaders in Outremer. In other words, Manzikert was perceived to be the first step in the process by which Turkish-led dynas­ ties defeated the Christians and proclaimed the triumph of Islam.

The fall of Constantinople to the Latins in 1204 The unusually wide hori­ zons of perhaps the greatest of the medieval Muslim chroniclers, Ibn alAthir, writing from Syria, prompted him to include under the year 600AH (1203/4) an interesting and very rare account of the Fourth Crusade.42 In it the destinies of three groups which concern Ibn al-Athir - the Byzantines, the Crusaders and the Seljuqs of Anatolia - converge. Ibn al-Athir relates that these Crusaders came originally to help Jerusalem but that they made their way to Constantinople because of the internal power struggle being enacted there. The Byzantines, in great distress at their treatment by the Crusaders, contacted the Seljuq sultan Sulayman b. Qilij Arslan, lord of Konya, but 'he found no way of acceding' to their request for help. Ibn al-Athir possesses some concrete information about the Latin conquest of Constantinople, although his sources are not mentioned, and he presents it in his usual unvarnished way. He mentions specifically the name of three Crusader leaders - the blind Doge of Venice whose 'horse had to be led' for him,43 the Count of Flanders, and the Marquis of Monferrat. He describes the three leaders casting lots for possession of Constantinople, and the conquest by 'the patriarch of Byzantium, Lashkari', of the lands east of Constantinople.44 On the surface this account, perhaps surprising for its inclusion at all, since Seljuq or indeed any Muslim involvement is minimal, is straightfor­ ward. Yet Ibn al-Athir, who rarely rises to great heights of invective or emotional outrage - a striking and well-known exception is his famous account of the Mongol invasion - is making here a few telling propaganda points as well as revealing a nuanced view of Byzantium and Christendom. Emphasizing the plunder and killing of the population of Constantinople by the Franks, he writes the following: 'A group of the notables of Byzantium entered the great church which is called Sophia. The Franks came to it and a group of priests, bishops and monks came out to them with the Gospel and the Cross in their hands to entreat thereby to 42 Ibn al-Athir, Kamil, XII, 124-6. 43 Ibid., 125. 44 Ibid.

SELJUQ HISTORIOGRAPHY

85

spare them. They did not heed them and they killed [them] and plun­ dered the church/45 There are strong echoes here of the accounts in the Islamic sources of the fall of Jerusalem to the Crusaders in 1099. But here in 1204 the victims are the Byzantines who, although they are fellow-Christians, are them­ selves victims of Crusader treachery and brutality. For Ibn al-Athir the familiar Christian enemy, Byzantium, has been ousted from the firing line of Islamic anti-Christian invective (directed before the coming of the Crusades at Byzantium) by a new and even more devilish foe, the Crusaders. Byzantium, a long-familiar entity in the Near East and a frequent player in Seljuq affairs, is treated almost sympathetically here. It should also be noted that mention of Byzantium in the Seljuq sources is not usually accompanied by the traditional minatory tags, such as 'May God curse them' or 'May God send them to perdition' which are employed regularly for the Crusaders from the second half of the twelfth century onwards.

General observations For the historian of Byzantium the Islamic sources which deal with Seljuq history are perhaps disappointing. Their concentration on Manzikert, with only occasional references to other Byzantine involvement in Syria and Anatolia, is a clear indication of the introverted priorities of Islamic historians. Important and protracted historical developments are not the concern of such writings. Complex processes - such as the impact of the Turcoman nomads on agriculture, the level of Seljuq Islamization, the situation of Christians under Seljuq rule in Anatolia - are left undis­ cussed. Chance warlike encounters, such as Manzikert probably was, are given retrospective validity by being portrayed as part of an overall strat­ egy, a deliberately expansionist policy on the part of the Seljuqs. Border skirmishes, raiding and nomadization are cast as jihad. Thirteenth-century Muslim writers living under Turkish rule, such as Ibn al-Athir, Ravandi and al-Husayni, hark back to the Seljuqs as the illustrious predecessors of their present regimes, just rulers with impeccable credentials, masters of their own destiny, zealous, newly converted Muslims who wage jihad against the infidel and heretic alike. Given this casting of the Turkish Seljuq sultans in 'heroic' Islamic mould at the hand of Arabic and Persian court officials and religious scholars, is it then possible to glean any insights at all from the sources into the effects of the Seljuq conquest on the Near East and, more espe­ cially in the context of this volume, on Anatolia? In other words, do the

45 Ibid.

86

CAROLE HILLENBRAND

sources tilt towards a corroboration of the views of Cahen or Vryonis, Wittek or Lindner? It is more in the evaluation of short anecdotes and asides, in hints rather than in overall conceptions, that some tentative comments may be made. The lifestyle of the Turcomans, be it in Iran, Syria or Anatolia, is not ignored by Muslim writers even within the framework of their overarching historiographical agenda of legitimizing and Islamizing the rule of the Seljuq sultans. Beneath the rosy surface there are suggestions that the coming of the nomadic Turks was difficult to bear;46 their military might was a fact of life, a 'necessary evil' to be channelled and utilized by those who knew better, the bureaucrats and religious leaders, in order to support the Islamic world and defend it by force of arms against all comers. The innate superiority and hostility felt by Arabs and Persians, especially initially, towards their Turkish overlords emerges in recurring clichés about Turkish drunkenness and brutal­ ity; the topoi of Turks building minarets out of the skulls of their enemies and then making the call to prayer from these structures, or drinking wine from goblets made of these skulls, spring to mind.47 There are also stories of Turks being kept outside city walls because of their disruptive impact on those inside. The evolution of the Turkish leadership from nomadic chiefs to seden­ tary rulers must have been slow, difficult and complex. In general, however, the sources elide these problems and negative comments about the Seljuqs have to be searched for laboriously in an otherwise favourable - if stereotyped - picture of them in both the Arabic and Persian historio­ graphical traditions.

46 This negative picture is, of course, corroborated by the much more overtly hostile testi­ mony of Georgian and Syriac sources. 47 Such a story is told of the Turkish commander, Tughtegin, who fought the Franks in the early twelfth century.

Mirkhwand Ahmad Ghiffari

Turkish translations of Ravandiand Ibn Bibi

1350-

Ibn al-Furat Al-Maqrizi Al-'Ayni

Rashid aI-Din AI-Yazdi Hamdallah Mustawfi

Aqsarayi Ahmad of Nigde Author of Tarikh-i al-i Saljuq

1300-1350

Ibn al-'Adim Ibn Shaddad Ibn Wasil

Ravandi

Ibn Bibi

Sibt b. al-Jawzi

Ibn Funduq Nishapuri

Al-Hamadhani Anushirwan

1250-1300

Al-Shaybani Bundari Ibn Abi Tayyi' Ibn al-Athir

Iran

Hilal al-Sabi Author of the Ghars al-Ni'ma Maliknama

Iraq

Al-Anawi

Ibn al-Azraq

Jazira

1200-1250

Ibn al-Qalanisi Al-Halabi 'Imad ai-Din al-Isfahani

Syria

1150-1200

Ibn Zafir

Egypt

Ibn Munqidh

Anatolia

Select table of medieval historical writing on the Seljuqs in Arabic and Persian

1100-1150

1050-1100

Appendix

al-Husayni

Gardizi Bayhaqi

Central Asia

88

CAROLE HILLENBRAND

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

6. The concept of 'history' in medieval Armenian historians Robert W. Thomson The Symposiarch's instructions were explicit. I was to make a contribu­ tion on the Armenian tradition of history writing, its aims and methods, so that when Byzantinists read an Armenian chronicle, they might have some idea about the context within which it was written. After some discussion I narrowed my theme to the tenth to thirteenth centuries, concentrating on three main points. What did Armenian historians of this period reckon the purpose of their efforts to be? What models did they use in composing their histories? And how did they view Armenia in the larger scheme of things? This period is clearly interesting for Byzantinists, since it covers Byzantine expansion into Armenia and the brief occupation of most of the country; the arrival of the Seljuqs and the disaster of Manzikert; and the campaigns of the Crusaders. It ends with the Mongol invasions. But from the point of view of Armenian historiography, the tenth century does not mark any new beginning or dramatic development. So first I shall offer a brief outline of traditional Armenian historiography, against which back­ ground the aims and methods of the writers of the tenth and later centuries will, I hope, become more comprehensible.1 Historical writing goes back to the first generation of literacy in Armenian - that is, to the beginning of the fifth century when a native script was invented by Mashtots.2 Before that time Armenians had not 1 For a fuller discussion of the early period in Armenian historical writing see J.-P. Mahé, 'Entre Moïse et Mahomet: réflexions sur l'historiographie arménienne', REArm 23 (1992), 121-53; and R.W. Thomson, The Writing of History: The Development of the Armenian and Georgian Traditions', II Caucaso: Cerniera fra culture dal Mediterraneo alia Persia (secoli IV-XI), SSCIS 18 (Spoleto, 1996), 493-514. Details of Armenian histories, translations and secondary literature published up to 1992 may be found in R.W. Thomson, A Bibliography of Classical Armenian Literature to 1500 AD (Turnhout, 1995). 2 For the life and work of Mashtots as known through the biography composed soon after From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2000 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

89

90

ROBERT W. THOMPSON

been illiterate; they simply wrote in languages other than their own - in Greek, Syriac, or Parthian. The invention of the script was a Christian enterprise. Mashtots and his disciples sought to provide as quickly as possible in the vernacular Christian texts translated from Greek and Syriac; but the secular traditions of late antiquity were not ignored completely. So quite soon after the death of Mashtots, thanks to his colleagues' labours, a potential Armenian author had many models at his disposal. Quite apart from a great mass of biblical, liturgical and patristic works, there were biographies and martyrdoms, as well as full-scale historical works like those of Eusebios, Socrates Scholastikos and Josephus, whose Jewish Wars was used by Moses Khorenatsi.*3 This is not the place to review all Armenian histories from the fifth century onwards. In brief, we can say that they were Christian works. Despite obscure references, nothing is known of pagan writings from Armenia.4 And when Armenian authors describe their pagan past, since they were writing generations, if not centuries, later, their recollections or imaginations were much distorted. Armenian historians thought of their compositions as having a moral as well as informative purpose. History was a record of God's providence in the world; the reader was to learn to avoid wickedness and follow the examples of virtuous conduct. A minor­ ity, notably Moses Khorenatsi and his imitators, emphasized secular fame; but Moses was a classicizing author, influenced by late antique rhetorical texts.5 It is important to remember that writings in the Armenian language were produced in the eastern part of the country - the much larger section under Iranian control after the division of 387.6 Until the end of the Sasanian era Armenian historians had to come to grips with the position of their country as part of the greater Iranian world, where their Christian allegiance sat uneasily with their political allegiance to the Persian shah.

his death by his pupil Koriwn see G. Winkler, Koriwns Biographie des Mesrop Mastoc', OCP 245 (Rome, 1994). 3 For Armenian translations of Greek and Syriac texts see C. Zuckerman, A Repertory of Published Armenian Translations of Classical Texts, with an Appendix by A. Terian; revised by M.E. Stone, Institute of African and Asian Studies; Hebrew University of Jerusalem (Jerusalem, 1995); L. Ter-Petrosian, Ancient Armenian Translations (New York, 1992: English and Armenian versions; Erevan, 1984: French version); J.-P. Mahé, Traduction et exégèse: réflex­ ions sur l'exemple arménien', Mélanges Antoine Guillaumont, Cahiers d'Orientalisme 20 (Geneva 1988), 243-55; Thomson, Bibliography, 29-88: 'Translations into Armenian'. 4 The historian Moses Khorenatsi is the first to refer to local pre-Christian records; see J.P. and A. Mahé, Histoire de l'Arménie par Moïse de Khorène (Paris, 1993), 35-9. 5 See the Introduction to R.W. Thomson, Moses Khorenatsi: History of the Armenians, Harvard Armenian Texts and Studies 4 (Cambridge, MA, 1978). 6 For an overview of Armenian history in the period covered in this chapter see vol. 1 of R.G. Hovannisian, ed., The Armenian People (New York, 1997).

'HISTORY' IN MEDIEVAL ARMENIAN HISTORIANS

91

This balancing act is the theme of two important works: Lazar Parpetsi writing around the year 500, and Elishe of disputed date. The latter's influential history of the Armenian revolt in 450 against Iran describes in detail the heroic deaths of the leaders in battle and the martyrdom of important prisoners taken to Iran. But for Elishe virtuous conduct is not seen in the terms of an early Christian martyrdom, where the salvation of an individual soul is at stake. Armenian moral virtue is linked to the survival of the nation. The Armenians are not fighting for Christendom, but for the survival of specifically Armenian traditions. They had a intense awareness of the danger of apostasy - which was indeed frequent and often politically motivated, as in Georgia. Elishe's model is the Maccabees, just as Moses Khorenatsi echoes attitudes of Josephus.7 The collapse of the old world order in the seventh century and the rise of Islam found their historian in Sebeos. This unknown author concen­ trates, like his predecessors, on the relationship of Armenians with Iran. But he was also familiar, from sources available at the Armenian patriar­ chate, with details of Byzantine history. He describes both the religious tensions between the churches, and the course of political and military events in late sixth- and early seventh-century Byzantium. His canvas is notably broader than that of his predecessors.8 Sebeos is also significant in that he is the first Armenian historian to attempt an explanation of dramatic historical change. All Armenian authors refer to God's providence; but Sebeos specifically draws on bibli­ cal prophecy to explain the rise of Islam and the rapid Muslim conquests. He is the first to equate Islam with the last of the four kingdoms of Daniel's vision, adducing numerous other prophetic utterances in support of the pre-ordained nature of this new phenomenon. He does not, however, offer any hope for the future such as might be found in the Apocalypse attributed to Pseudo-Methodios. The latter was composed a little later than Sebeos; and although its ideas penetrated Armenia, and apocalyptic notions gained wide diffusion, such expectations of deliver­ ance are not echoed by the main-stream historians.9 7 These themes are discussed in the Introductions to the English translations by R.W. Thomson: Elishe: History of Vardan and the Armenian War (Harvard Armenian Texts and Studies, 5) (Cambridge, MA, 1982); The History of Lazar P'arpec'i, Columbia University Program in Armenian Studies. Suren D. Feshjian Academic Publications 4 (Atlanta, 1991); Moses, as n. 5. Details of Armenian editions in Thomson, Bibliography. 8 See now R.W. Thomson and J. Howard-Johnston, The Armenian History attributed to Sebeos, Translated Texts for Historians (Liverpool, 1999). 9 For apocalyptic notions see further below, and in general: A. Sanjian, T w o Contemporary Armenian Elegies on the Fall of Constantinople, 1453', Viator 1 (1970), 223-61; N. Garsoi'an, 'Reality and Myth in Armenian History', The East and the Meaning of

92

ROBERT W. THOMPSON

Let us now turn to the period of our investigation and see how these earlier ideas and themes are echoed in later historians. The first to be discussed is Thomas Artsruni, writing early in the tenth century. His history deals explicitly with only one region of Armenia, just as earlier writers had concentrated on one particular noble family - Elishe and Lazar on the Mamikoneans, Moses on the Bagratids. Thomas's work is entitled 'History of the House of the Artsrunik'/ who were a leading noble family in the Lake Van area. It is intended as an encomium of their first prince to acquire royal status, Gagik, famous for his patronage of the church on the island of Aghtamar.10 Thomas is the first to mention that enigmatic historian Moses Khorenatsi by name or to quote from his book. Just as Moses had eulo­ gized the Bagratids, tracing their ancestry back to a Jewish antiquity, so Thomas, with greater plausibility, elaborates on the traditional Assyrian origin of his patrons.11 Here, then, is one characteristic of the majority of Armenian historians: they are often, though not always, concerned with a specific region, or one noble family; and they are often partisan in their efforts to boost the status of their patrons. A second characteristic of Thomas's work, echoed in later writers, is his adaptation of imagery from his predecessors to a later situation. His treat­ ment of the Muslim rulers in ninth-century Caucasia is modelled on the way Elishe portrayed the shahs of the fifth century, that is, in terms taken from the Books of the Maccabees. The impious Seleucids were first replaced by the Sasanians and then by the Muslims; while for the Jewish law is substituted Armenian tradition.12 As for the purpose of writing history, Thomas stresses secular fame; and for its methods he emphasizes veracity, reliability, strict chronology and elegance. These expressions come from Moses, who in turn had taken them from the rhetoricians of antiquity. Almost contemporary with Thomas is the catholicos John V Drasxanakertc'i, who served as patriarch from 898 to 924. Unlike those which preceded it, his history was written by a participant in events; it is

History, Università di Roma 'La Sapienza', Studi Orientali XIII (Rome, 1994), 117-45; R.W. Thomson, 'Constantine and Trdat in Armenian Tradition', Acta Orientalia Academicae Scientiarum Hungaricae 50 (1997), 277-89. 10 See the Introduction to R.W. Thomson, Thomas Artsruni. History of the House of the Artsrunik' (Detroit, 1985). The church is examined in detail in L. Jones's chapter in this volume (Chapter 14). 11 Moses, I, 23, is the first to associate the Artsruni family with the sons of Senekerim who fled to Armenia, as described in Isaiah, 37.38. 12 For the earliest examples in Armenian historiography of the Maccabees as a model see R.W. Thomson, 'The Maccabees in early Armenian Historiography', JTS 26 (1975), 329—41; reprinted in Studies in Armenian Literature and Christianity (Aldershot, 1994).

'HISTORY' IN MEDIEVAL ARMENIAN HISTORIANS

93

a record of his role in the politics of Armenia. Given his position, it is not surprising that many regions of the country feature in the history on occa­ sions when John's duties took him to those parts; for he travelled widely as peacemaker between the fractious nobles. His work is neither provin­ cial in outlook nor partisan in its presentation, though John is firm in supporting Armenian independence from Byzantium.13 Like Sebeos, John quotes several documents to which he had access. In his own case two letters stand out: one he received from Nicholas Mystikos, the patriarch of Constantinople, and one that he wrote himself to the emperor Constantine Porphyrogennetos. It is surprising, however, that John does not give any clear indication of the purpose of historical writing. In his preface he notes that reliable historical narrative is beneficial to ratio­ nal men. The epilogue is an exhortation to virtue, not a defence of the role of history. Reflections on the philosophy of history seem not to have been congenial to such a busy and preoccupied participant in events. As he impatiently notes in his preface: 'And now, let me not waste more time on other matters in this introduction, since wretched old age holds death ready at the door, and my anxiety impels me to narrate the disastrous calamities and the terrible turmoils that came upon and overwhelmed us.' The works I have mentioned so far are narrative histories with a broader or narrower sweep, in which various documents are quoted or rhetorical speeches of the author's invention inserted. Armenians were also familiar with a different style of historical text, which we might call the 'chronicle'. The most famous of these, Eusebios's Chronicle of the early fourth century, was translated into Armenian; not only did it become a prime source for information about the ancient world, its themes of chronological progression and strict dating greatly influenced Moses Khorenatsi, and through him all later Armenian historians.14 The earlier Chronicle of Hippolytos was also translated,15 though the surviving Armenian version has been expanded with later material.16 The distinc­ 13 See the Introduction to K. Maksoudian, Yovhannes Drasxanakertc'i. History of Armenia (Atlanta, 1987). 14 The date of the translation is debated, perhaps sixth century; see the Introduction to the translation by J. Karst, Die Chronik des Eusebios aus dem armenischen übersetzt, Griechischen Christlichen Schriftsteller 20 (Leipzig, 1911). Eusebios's Ecclesiastical History was translated earlier, from the Syriac version. For the genre of chronicle in Armenian see C. Hannick, 'La Chronographie grecque chrétienne de l'antiquité tardive et sa réception dans l'historiogra­ phie arménienne', in R.B. Finazzi and A. Valvo, eds, La diffusione dell'eredità classica nell'età tardoantico e medievale (Alessandria, 1998), 143-55. 15 See J.-P. Mahé, 'Quadrivium et cursus d'études au Vile siècle en Arménie et dans le monde byzantin', TM 10 (1987), 159-206, esp. 187. 16 The expanded text of Hippolytos was published as Ananun Zhamanakagrut'iwn ('Anonymous Chronicle') by B. Sargisean (Venice, 1904); trans. J. Markwart and A. Bauer in A. Bauer and R. Helm, Hippolytus Werke, IV: Die Chronik, Griechischen Christlichen Schriftsteller 36 (Leipzig, 1929).

94

ROBERT W. THOMPSON

tion between a chronicle and a history is not clear-cut. But one can distin­ guish lists of rulers and the ordering of events under successive dates from a work like that of Elishe or John Catholicos. The chronographical approach is first clearly marked in the work of Stepannos of Taron, known as Asolik, 'singer' or 'reciter', writing just after the year 1000.17 Like Moses, Asolik divides his work into three books. The first is a period of ignorance, that is, the time from Adam to the Armenian king Trdat before his conversion. Unlike Eusebios, Asolik does not offer comparative tables to enable us to compare what is happen­ ing in different kingdoms at the same time. He simply gives lists of the rulers of antiquity. Book two covers the period of illumination, that is, the history of Armenia from its conversion to the reign of the first Bagratid king, Ashot, at the end of the ninth century.18 Although his account becomes progressively more elaborate, this section is still based on lists. Only in the third book, devoted to the Armenian Bagratids, does Asolik abandon this schematic approach. As for the purpose of history, like his predecessors Asolik emphasizes the importance of transmitting a memorial of past events to future gener­ ations. The first to do this, he says, was Moses, author of the Pentateuch, whose example was followed by Greek historians, and then by the Armenians. Asolik was not the first Armenian writer to give a list of prior historians, whose works he says he is now supplementing with his own. That honour belongs to Lazar, who says that he is continuing the works of Agathangelos and Pawstos (the Buzandaran). But after Asolik it became common for Armenian historians to copy such lists, bringing them up to date. This does not mean that they necessarily consulted all their prede­ cessors, but merely that they saw their own role as part of a larger move­ ment. For Asolik, his book of historical records will enable his readers to follow the true path, avoiding pagan and heretical errors and looking forward to the renewal of Creation at the end of historical time. The first to devote his entire history to chronography was Samuel of Ani, whose work covers the period from Noah to 1180 in the form of tables and entries under successive years.19 Like Asolik, he refers to the threefold division of historical time into past, present and future. In his

17 E. Dulaurier, Etienne Acoghig de Daron. Histoire Universelle (Paris, 1883), gives a transla­ tion of Books I and II. Book III was translated by F. Macler, Etienne Asolik de Taron. Histoire Universelle (Paris 1917); also in Publications de l'Ecole des langues orientales vivantes, le série, tome 18 bis (Paris 1920). 18 Ashot I Bagratuni: sparapet in 855; Prince of Princes in 862; crowned 884; died c. 890. 19 Samuel's own work goes to 1180; it was later expanded by a series of continuators. The translation by M.F. Brosset, Collection des historiens arméniens II (St Petersburg, 1876), 348-483, gives only the tables from the birth of Christ to 1348; the Latin version in PG 19, 601-740, is complete.

'HISTORY' IN MEDIEVAL ARMENIAN HISTORIANS

95

preface he notes that T h e interest of mankind seems to be concerned more with the past and the future than with the present'. Regarding the future, he warns his readers about inquisitive speculation into what God has reserved for himself. As for the past, he is (to the best of my knowl­ edge) the first explicitly to compare different sources for a date and to change one in order to make his tables consistent.20 This raises a very interesting question of historical method, which I shall address when we come to the chronicler Vardan. An author who has attracted particular attention in Byzantine circles is Aristakes from Lastivert, north of Erzurum. Beginning in the year 1000, he describes the collapse of the medieval Armenian kingdoms, the brief period of Byzantine control, the attacks of the Seljuqs, and the dénouement at Manzikert.21 Conscious of continuing the work of Asolik, he indi­ cates that his history is to be a record, on the pattern of the old chroniclers. But he goes beyond that. His second purpose is to help the reader under­ stand the cause of those sins which brought calamities upon the Armenians, so that he may come to fear God, be penitent, and thus save his soul.22 Although Aristakes does not mention by name any previous Armenian historian except Asolik, a parallel here would be Elishe, who emphasizes the wicked conduct of the apostate Vasak and declares that by perusing the record of his villainy the reader may be deterred from following in his footsteps.23 More decisively than his predecessors, Aristakes sees a religious pattern to events and a way out. He takes a specifically Old Testament approach. The sins of Israel angered God, and he punished them through foreign races. Likewise, the backsliding of the Armenians is punished by foreign invasions.24 He does not hold out hope of a miraculous saviour, in the fashion of apocalyptic writers. But since disasters occur because of sin, repentance and confession will bring God back to the Armenians' side. So Aristakes urges a return to Christian morality, seeing nothing inevitable in what happened in the past or may come about in the future. As for the particular causes of the sins which brought down God's anger, Aristakes has strong views. He was a writer of conservative bent, viewing through rose-tinted spectacles 700 years of Christian Armenia. But in the eleventh century the recent growth of cities like Ani, Kars or 20 p.136 of the Armenian text, concerning the death of Constantine Porphyrogennetos; ed. A. Ter-Mikaelean, Hawak'munk' i groc' Patmagrac' (Ejmiatsin, 1893). 21 Translation and commentary in M. Canard and H. Berbérian, Récit des malheurs de la nation arménienne, Bibliothèque de Byzantion 5 (Brussels, 1973), with page references to the Armenian text, ed. K.N. Yuzbashyan, Patmut'iwn Aristakisi Lastivertc'woy (Erevan, 1963). 22 Aristakes, 145. 23 Elishe, 140. 24 Aristakes, 65-8; see also 113.

96

ROBERT W. THOMPSON

Artsn had been quickly followed by attack and destruction. Earlier Armenian historians had been ambiguous about city life, for cities had not played a major role in the social, religious, intellectual or political life of Armenia, whose social structure was quite different from that of the eastern Roman empire or the world of Islam.25 Aristakes sees a specific cause of the rapid moral decay which had brought prompt retribution: corruption through avarice, sparked by the profits of trade in cities. Salvation can only be obtained through collective action - not by force of arms, but by a return to the morality and piety of past days.26 The history of Aristakes reminds us that the influence of the Bible was all-pervasive in Armenian literature. First it was a historical source: the Old Testament, plus apocryphal books, presented the origins of mankind and God's subsequent dealings with the nation of Israel; the New Testament described the incarnation of Jesus Christ, the origins of the church and the promise of future judgement. In addition, the language of the Bible had a great influence in a literary sense, as a natural source of imagery. Old Testament themes in particular were drawn upon, both literally and metaphorically. Armenian historians saw parallels between the fortunes of Israel and the fate of their own country - hence the perva­ sive comparisons with the Maccabees. Likewise the utterances of the prophets were also of relevance to Armenia. We have already seen in Sebeos the adaptation of Daniel's image of the four beasts to the rise of Islam. Into this tradition falls the twelfth-century historian Matthew from Edessa, important for his description of the Turkish invasions and the Crusades.27 Matthew offers a dramatic picture of the first encounter of the Armenians with Turks in 1018. An Armenian force was sent against them, commanded by Dawit Artsruni, the son of Senekerim the king of Vaspurakan.28 The Armenians were forced to turn back since they were unfamiliar with the tactics of these mounted archers. Matthew's descrip­ tion of Senekerim's reaction is revealing: 'Sitting down, he examined the chronicles and utterances of the divinely inspired prophets, the holy teachers, and found written in these books the time specified for the coming of the Turkish troops. He also learned of the impending destruc­ tion and end of the whole world.' Matthew goes on to quote Isaiah.29 25 See N.G. Garsoian, The Early-Mediaeval Armenian City: An Alien Element?' The Journal of the Ancient Near Eastern Society 16-17 (1984-85) (Ancient Studies in Memory of Elias Bickerman), 67-83. 26 Aristakes, 74-6. 27 A. Dostourian, Armenia and the Crusades, 10th to 12th Centuries: The Chronicle of Matthew of Edessa (Lanham, MD, 1993). Texts and earlier translations in Thomson, Bibliography, 151-2. 28 Note the adoption of the Assyrian name; see above at n. 11. 29 Matthew, Book I, ch. 48.

'HISTORY' IN MEDIEVAL ARMENIAN HISTORIANS

97

Elsewhere he adapts to the Turks the imagery of the fourth beast in Daniel's vision, as used for the Muslims by Sebeos earlier.30 However, unlike Aristakes, Matthew does not draw any meaningful conclusion. Events may have been predicted; but they are not regarded as a trial or punishment which will induce the Armenians to mend their ways. So in a sense, history is the story of the inevitable. At least once, however, Matthew does suggest that relief may come in the future. In the year 1036/7 there occurred an eclipse of the sun, and the noted scholar John of Kozern was asked for an interpretation. In a long passage he describes future troubles. But then the valiant nation of the Franks will come and capture Jerusalem. (Matthew, writing after 1136, would have been aware of the capture of Jerusalem in 1099, but he puts this prophecy back to the days of John of Kozern.) Following this, according to Matthew's report concerning John, there would be further sufferings and the Persians, that is, the Turks, would grow strong. Eventually, however, the Roman emperor will be awakened and come against them. Then, under the control of the Roman emperor, the land will see prosperity once more.31 The idea that the Turks inflict sufferings on the Christians, but are then expelled by the Roman emperor who comes in person to the east and inaugurates a time of peace and prosperity, is a feature of the Apocalypse of Pseudo-Methodios. This was composed at the end of the seventh century in Syriac to explain the conquests of the Arabs. It was known in Armenia, though a complete Armenian version does not seem to be extant.32 There was also an earlier original Armenian prophecy which enjoyed later elaboration. The prophecy of the patriarch Nerses I is first attested at the end of the fifth century in the anonymous Buzandaran. It originally referred to the division of the Armenian kingdom into two spheres in 387. In the tenth century a Life of Nerses introduced later events supposedly foreseen by the fourth-century patriarch. Successive recensions of this text (which unfortunately have not yet been published or properly distin­ guished) updated the prediction: Nerses now foretold the abolition of the monarchy in 428, the Persian capture of Jerusalem and the captivity of the Cross in 614, the rise of the Ismaelites, the arrival of the Franks who recap­ tured Jerusalem, and the coming of the 'Archers'. This term usually refers to the Mongols, but is occasionally applied also to the earlier Turks. Finally, salvation will come from the Romans.33 30 Matthew, II, 108; cf. Sebeos, 141,177. 31 Matthew, II, 64. 32 Syriac text edited and translated by G.J. Reinink, Die syrische Apokalypse des PseudoMethodius, CSCO 540, 541 (Leuven, 1993). There is a long extract in Stepannos Orbelean, ch.32; tr. M.F. Brosset, Histoire de la Siounie par Stephannos Orbelian (St. Petersburg, 1864), 89-94. 33 Patmut'iwn srboyn Nersisi Part'ewi, Sop'erk' 1 (Venice, 1853); translation in J.-R. Emine,

98

ROBERT W. THOMPSON

These apocalyptic hopes for salvation, elaborated in popular texts and occasionally echoed in the historians, became linked to an old Armenian theme which takes us back to the conversion of King Trdat at the begin­ ning of the fourth century and the activity of Saint Gregory the Illuminator. All versions of these events, the history of Agathangelos as it survives in Armenian, and its earlier versions reflected in translations into Greek and Arabic, note that King Trdat and Saint Gregory visited Constantine and formed an alliance with that first Christian emperor. The progressive elaborations of that encounter do not concern us here,34 save that the notion of salvation from Rome as found in the Pseudo-Methodian Apocalypse was easily linked with this old tradition concerning Constantine. According to Agathangelos, Trdat had travelled to Rome with an escort of 70,000 Armenian troops. We now discover that some of these had remained in the Roman capital, and their descendants would join the forces of the new Constantine when he came to the east.35 But we are straying from our theme. Finally, I would like to mention briefly the attitude to history of Vardan, a noted scholar of the thirteenth century. His writings often give the impression of lecture notes written up in an anecdotal way. The Geography, for example, is little more than an organized listing of places with notes of antiquarian interest attached. His Commentary on Grammar is an explanation of technical terms, while his commentaries on scripture consist of short disconnected paragraphs. Vardan's Chronicle reflects his eclectic interests.36 It is divided into sections arranged chronologically, which recount important events taken from earlier histories. He touches on military, political and ecclesiastical topics, but offers no clear purpose to his compilation. He does emphasize that man was created with a rational mind; so by studying the events of this world one may prepare for an understanding of the eternal verities. But the point I wish to make here is that by the thirteenth century there already existed in Armenian a large number of histories and chronicles on which Vardan could draw. How did he cope with the plethora of infor­ mation and the discrepancies in his sources? His choice of highlights is not explained, though it clearly depended to

'Généalogie de la famille de saint Grégoire et vie de saint Nersès', in V. Langlois, ed., Collection des historiens anciens et modernes de l'Arménie II (Paris 1869), 21-41. 34 See Thomson, 'Constantine'. 35 See Sanjian, 'Two Contemporary Armenian Elegies', 259, translating lines 249-56 of the Elegy by Arakel of Bitlis. 36 See the Introduction to the translation in R.W. Thomson, 'The Historical Compilation of Vardan Arewelc'i', DOP 43 (1989), 125-226. In addition to the literature on Vardan in Thomson, Bibliography, 210-12, see S.P. Co we, 'The Background and Development of Vardan Arewelc'i's Cosmology', REArm 24 (1993), 75-87.

'HISTORY' IN MEDIEVAL ARMENIAN HISTORIANS

99

some extent on which particular historian he was following at any given point. More importantly, since there were often several sources on which he could draw, Vardan had to cope with the differences in their accounts. He prefers the inclusive method. That is, instead of following one source to the exclusion of others for any given episode, he combines the infor­ mation to be gleaned from several previous writers. In the process he creates a new version of events, details from different origins being put together for the first time. This may not seem a very critical approach to the reconciliation of conflicting accounts. But I think that Vardan was simply adapting to the writing of history his method of biblical commen­ tary. Just as different commentators, each with something of his own to say about a particular passage of the Bible, may be brought together in the hope that a better understanding of the holy text may emerge, so in the composition of a historical narrative, all accounts of earlier writers may be harmonized to create a fuller picture of events. Gospel harmonies are an obvious parallel, and Tatian's Diatessaron had had a strong impact in Armenia.37 In this brief presentation I could not even mention by name all the Armenian writers whose histories or chronicles may be of interest to a modern scholar. Nor did I turn to other kinds of texts which might supplement the historians' views about the meaning and purpose of history. But in sum, by medieval times the main features of Armenian historical writing had already been formed. Later writers could draw on long-standing traditions regarding Armenian origins, their individual Christian heritage with its emphasis on Jewish parallels, and the scope and purposes of historiography. Emphasis on chronology was not new; but the formal chronicle did gain in popularity. Apocalyptic hopes of deliverance became widespread, and the interpretation of prophecy began to figure more prominently. Events of the period forced Armenian historians to broaden their scope, since by now they had to come to grips with Crusaders, Turks and Mongols, as well as their old acquaintances, the Byzantines and Muslims. By the tenth century the truly formative histories that influenced Armenian perceptions of their own destiny had already been written. The works I have briefly touched upon, however, have a greater appeal to modern historians precisely because of their broader scope.

37 For the influence of the Diatessaron on Armenian authors see S. Lyonnet, Les origines de la version arménienne et le Diatessaron, Biblica et Orientalia 13 (Rome, 1950), 57-98. The Commentary on the Diatessaron by Ephrem (or one of his circle) was translated from Syriac into Armenian at an early date; see L. Leloir, Commentaire de l'Evangile concordant ou Diatessaron, traduit du syriaque et de l'arménien, SC 121 (Paris, 1966), and further references in Thomson, Bibliography, 46-8.

This page has been left blank intentionally

7. From bumberazi to basileus: Writing cultural synthesis and dynastic change in medieval Georgia (K'art'li) Stephen H. Rapp Jr The genesis of medieval Georgian historical literature is inextricably tied to three crucial historical 'moments'. First, written Georgian literature was enabled by the Christianization of the region and particularly by the conversion of the king of the core eastern Georgian region of K'art'li (Iberia) around the year 337. Because contemporary Christianity was a faith based upon the written word, a crucial component of the process of conversion was the invention of a local script so that ecclesiastical texts could be transmitted to the various Georgian peoples. Second, as Zoroastrianism and Sasanid Persia increasingly posed a danger in the late sixth and early seventh century, the K'art'velian Crown actively sought an alliance with Christian Byzantium. At the same time, the Armenians, who generally distrusted the Byzantines, openly questioned the ortho­ doxy of the K'art'velian prelate and in 608/9 the Armenians set in motion an ecclesiastical schism with their northern neighbours. The K'art'velians were consequently compelled to re-evaluate their Christian past. Thus, later in the seventh century appeared the first (extant) local written narra­ tive describing the conversion of K'art'li, a text having a distinctly K'art'velian perspective that denied Armenian involvement. The third impulse leading the K'art'velians to compose history was the hegemonic rivalry for Caucasia between great Eurasian powers of Persia and Byzantium. In the sixth century, probably around the year 580, Sasanid Persia dissolved K'art'velian kingship, eliminating K'art'velian indepen­ dence. It was during this lengthy interregnum stretching to 888 that Georgian historiography first appeared. The earliest K'art'velian histori­ ans do not identify themselves, nor do they reveal their motives for writing history, though their works were clearly a tool to celebrate the independence of bygone years and even a call for indigenous kingship to From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

101

102

STEPHEN H. RAPP JR

be restored at once. Many neighbouring peoples - including the Armenians, Caucasian Albanians and the Persians - were at the same time exploring their historical origins in both written and oral traditions. The contest for Caucasia waged by Persia and Byzantium had lasting consequences. One of these repercussions was the reorientation of the selfidentity of the K'art'velian élite from the Near East dominated by Persia to the Christian Byzantine empire. This chapter will investigate this poorly studied shift in so far as it affected the development of and is expressed in medieval Georgian historiography. As will become abundantly evident, this hegemonic struggle and the resultant transformation of aristocratic identity is the proper context for exploring the genesis and evolution of Georgian historical writing between the early ninth and mid-thirteenth centuries. These two issues are intertwined: one cannot be divorced from the other. So as to understand the situation around the year 800, when the K'art'velians initially began to write down their history, we must return to the end of the sixth and the start of the seventh centuries, at which time the rulers of the eastern Georgian district of K'artTi embarked upon a grand experiment. For centuries their lands, and most of Caucasia for that matter, had been intimately associated with the greater Persian common­ wealth. Though the Achaemenid and Sasanid kings had sometimes cast their political hegemony over the isthmus, it was the propagation of Persian culture throughout western Asia that proved more enduring. The various Georgian, Armenian and Albanian peoples were integrated into the Near Eastern civilization dominated by Persia, but at the same time they were nevertheless distinct.1 All three, for example, possessed their own languages, though Georgian and Armenian - and presumably Albanian, too - shared a lexical core with the Iranic languages. This distinctiveness was further reinforced by the Christianization of the Caucasian monarchies beginning in the fourth century. Because local alphabets and literature were invented by Christians so as to strengthen the faith, fifth- and sixth-century Caucasian sources provide a Christian perspective that is often anti-Persian (and anti-Zoroastrian) in tone. Through careful scholarly inquiry, however, it is possible to expose the Persian substratum of ancient and medieval Caucasian society.2

1 These are the three dominant peoples of early medieval Caucasia. Transcaucasia' reflects a Russian perspective and is anachronistic for the pre-modem period. 2 C. Toumanoff, Studies in Christian Caucasian History (Washington, DC, 1963), convinc­ ingly argued that pre-modern Caucasia constituted a distinct socio-cultural unit. An extra­ ordinary proof of this is the recent identification of two MSS (N/Sin-13 and N/Sin-55) at St Catherine's monastery on Mount Sinai written in Georgian, Armenian and Albanian: Z. Alek'sidze, Albanuri mcerlobis dzegli sinas mt'aze da misi mnishvneloba kavkaziologiisaVvis (Tbilisi, 1998), with English text, 'Georgian-Albanian Palimpsest on Mt. Sinai and Its Relevance to Caucasian Studies', 31-42.

FROM B U M B E R A Z I TO B A SILE U S

103

It is not altogether obvious from contemporary evidence how the Georgians, Armenians and Albanians were able to reconcile their Persian heritage with Christianity, a religion that after the fourth century became synonymous in Persian eyes with the hostile Byzantine empire. The ensuing two or three centuries together constitute the formative period of Christian Caucasia. Antagonisms with Persia increased, and by the end of the sixth century Sasanid military campaigns reached a fevered pitch. The monarchy of K'artTi was fatally weakened already in the 520s. In 580 it finally succumbed, thus duplicating the extermination of the Arsacids of Armenia in 428. Within a decade, 'presiding princes' arose to fill the K'art'velian power vacuum. The earliest of these presiding princes were recognized by Constantinople; they surely drew upon the prestige of Constantinople and must have hoped that the Christian emperor might intervene on their behalf. However, K'artTi's traditional orientation towards the Near East was largely maintained. One tangible relic of this situation is the circulation of Sasanid silver coins in Caucasia. For whatever reason, some of the Sasanid coinage circu­ lating (minted?) in K'artTi was modified. We know neither the identity of those who executed these alterations nor the quantity of coins affected. But the modifications almost certainly could not have been carried out with the knowledge of the presiding princes.3 Significantly, the overall appearance of Sasanid coinage was not altered: it still featured the head of the Great King on the obverse and a Zoroastrian fire-altar flanked by two attendants on the reverse. At first, only the (Georgian) abbreviation of the presiding prince's name was added to the obverse, adjacent to the likeness of the Persian king. The Great King himself might have allowed the adjustment, for it might be deemed as showing the presiding princes' subjection to him. However, it is possible that the local presiding prince wished to show that he shared in Persian royal authority. The presiding prince Step'anoz I (r. c. 590-627) carried this numismatic innovation a step further. Not only was his name inscribed next to the figure of the Great King, but the detail of the fire-altar depicted on the reverse was diminished and the resultant vague image was adorned with a Christian cross. Coins circulating during the reign of prince Step'anoz II (r. 637/642-c. 650) perpetuated the trend. The importance of these socalled K'art'velo-Sasanid coins is twofold. On the one hand, they reflect Christian eastern Georgia's persisting connection to the Persian world in the sixth and seventh centuries. On the other hand, they simultaneously demonstrate that the K'art'velians were becoming increasingly aware of the growing religious divide with the Persians. With Christianization had come new opportunities for alliance with Byzantium. As a consequence, 3 E.A. Pakhomov, Monety Gruzii (Tbilisi, 1970), 17-36.

104

STEPHEN H. RAPP JR

K'art'velian loyalties and even K'art'li's place with the Eurasian world had to be reconsidered. How did K'art'velians understand their place within Eurasia? One of the richest sources for this question is the medieval Georgian historical corpus known as K'art'lis c'xovreba, informally known as the 'Georgian Royal Annals' or 'Georgian Chronicles'.4 Nearly all surviving pre-modern Georgian-language histories are preserved in it. Owing to the late manu­ script tradition, it is not possible to fix with certainty the origin of the corpus. However, its earliest incarnation may have already been produced at the start of the ninth century when its oldest component texts were first written down. K'art'lis c'xovreba's thirteen medieval histories may be classified into two broad groups.5 The corpus's 'pre-Bagratid' texts were composed before the ascension of the Bagratid dynasty in 813 a d and generally portray the local monarchy in Persian terms and in a Near Eastern context, while its 'Bagratid' histories were produced during the Bagratid regime and paint local kingship in explicitly Christian and especially Byzantine colours. This distinction is crucial for understanding not only the development of Georgian historical writing but also wider Byzantino-Georgian relations. Because Georgian-language histories were written on behalf of kings and because they are narrowly concentrated on kingship,6 it is impossible to speak about Georgian historiography without investigating its depic­ tion of royal authority, hence the focus here. Additionally, it must be emphasized that this chapter is not intended to be a comprehensive analysis of the diverse Eurasian elements present in K'art'velian kingship. Rather, I endeavour to sketch as succinctly as possible the dominant Persian-Near Eastern and then Christian-Byzantine paradigms and frameworks that characterized the two periods.

4 For the critical edition of the corpus, see S. Qauxch'ishvili, ed., K'art'lis c'xovreba, vols 1-2 (Tbilisi, 1955,1959) [hereafter KC]; vol. 1 republished as K'art'lis c'xovreba: The Georgian Royal Annals and Their Medieval Armenian Adaptation, S. Rapp intro., vol. 1 (Delmar, NY, 1998). The majority of volume 1 was translated by R.W. Thomson: Rewriting Caucasian History: The Medieval Armenian Adaptation of the Georgian Chronicles, the original Georgian texts and the Armenian adaptation (Oxford, 1996). On the corpus, see also: Toumanoff, Studies, introduction; and idem, 'Medieval Georgian Historical Literature (Vllth-XVth Centuries)', Traditio 1 (1943), 139-82; I. Javaxishvili, Dzveli k'art'uli saistorio mcerloba (V-XVIII ss.) (Old Georgian Historical Writing, 5th-18th Centuries), reprinted as vol. 8 of his T'xzulebani (Collected Works) (Tbilisi, 1977); and M. Lort'k'ip'anidze, Ra aris k'art'lis c'xovreba? (What is k'art'lis c'xovreba?) (Tbilisi, 1989). 5 The dating, identification, and even the very number of these texts remain controversial. For an elaboration of the view expressed here, see my 'Imagining History at the Crossroads: Persia, Byzantium, and the Architects of the Written Georgian Past', unpublished Ph.D. dissertation, University of Michigan (Ann Arbor, 1997), UMI reprint no. 9722070. 6 Compare contemporary Armenian histories that express aristocratic concerns and perspectives.

FR O M

BUMBER AZI T O BASILEUS

105

Although their Armenian neighbours had been composing history since the fifth century, the K'art'velians of eastern Georgia first consigned their oral historical traditions to writing in the early ninth century, a decade or two before the establishment of Bagratid authority in K'art Ti. The ecclesiastical schism with the Armenians and the ensuing effort to put into writing a distinctly K'art'velian tradition partially explain the appearance of a native Georgian historiography, as do increasing contacts with Byzantine civilization (and its premium upon the written word), not to mention the desire by some K'artVelians to restore local royal author­ ity. Three pre-Bagratid histories are extant: (1) The Life of the Kings; (2) The Life of Vaxtang Gorgasali; and (3) a brief untitled continuation of Vaxtang's biography.7 These were composed in the period c. 790-813 and are uniquely preserved within the corpus K'art'lis c'xovreba.8 The Life of the Kings addresses the earliest period of K'art'velian history and commences with the legendary ethnogenesis of the K'art'velian people and their Caucasian neighbours.9 The foreign traditions exploited by its anonymous author are indicative of the Eurasian influences that circulated in contemporary Caucasia. For example, its author interpolated the origin of central Caucasia into the genealogy of peoples found in the Judeo-Christian tradition, specifically in the biblical account of Genesis. He inserted K'art'los, the legendary eponym of the K'art'velians, into the Armenian tradition about their own eponym Hayk (the Georgian Haos). And most prominently, the historian intentionally set early K'art'velian history within the context of a Near East dominated by Persia, and not within the framework of the Mediterranean world of Greek civilization and the Roman and then Byzantine empires. This is made evident from the start: the eponyms of Caucasia are reputed to have rebelled against the first king upon the earth, Nimrod the Persian.10 According to The Life 7 The last two are collectively known as C'xorebay vaxtang gorgaslisa. Not all scholars recognize them as separate texts, and enormous debate surrounds their dating and identifi­ cation. The vast majority of experts in Georgia envisage the three works as products of the eleventh century. The Life of the Kings is usually attributed to the eleventh-century arch­ bishop Leonti Mroveli. 8 K'art'lis c'xovreba has come down to us in a relatively late manuscript tradition: its oldest Georgian-language variant was copied in the late fifteenth century. The corpus's oldest surviving redaction is a copy of its medieval Armenian-language adaptation (Patmut'iwn Vrac'), produced between 1274 and 1311. For a description of the manuscripts, their interre­ lation, and relevant scholarly literature, see Rapp, 'Imagining History, introduction' and excursus B, 712-28 (where the major variants are described individually). 9 An alternative tradition is preserved in the independent corpus Mok'c'evay k'art'lisay: Rapp, 'Pre-Christian History in the Georgian Shatberdi Codex: A Translation of the Initial Texts of Mok'c'evay k'art'lisay ('The Conversion of K'art'li')', Le Museon 112 /1 -2 (1999), 79-128. 10 Cf. Gen. 10.8-9 where it is said that Nimrod was the first upon the earth to be a mighty man. See also I Chron. 1.10.

106

STEPHEN H. RAPP JR

of the Kings, the epic struggle pitting the Caucasians against the Persians was relived countless times, especially at those moments when the Persians were compelled to engage their Turanian adversaries in the far east. Moreover, Faridün and several other Great Kings celebrated in the Persian epic cycle are prominently featured in The Life of the Kings. Simply put, eastern Georgia's history has been made incomprehensible without that of Persia and the Near East, though our anonymous ninth-century author emphasizes that the Caucasians often resisted outright political domination by the Persians and that K'art'velian culture was distinct despite its semblance to its Persian analogue.11 The depictions of two early K'art'velian monarchs vividly demonstrate the linkage of pre-Bagratid kingship to Persia. Significantly, this Persianlike image was not immediately altered with the conversion of the K'art'velian monarchy and their subjects to Christianity beginning in the third decade of the fourth century. Thus the alleged first indigenous (and semi-mythical) K'art'velian king P'arnavaz (r. ?299-?234 b c ) and the Christian hero-king Vaxtang I Gorgasali (r. 447-522 a d ) are both portrayed in a manner which would have been familiar to peoples throughout the greater Persian world. The Life of the Kings relates that P'arnavaz became the first king of the K'art'velians during the time of Alexander the Great.12 Although his father was a native K'art'velian, P'arnavaz's mother was a Persian from Isfahan. Following the murder of his father and uncle at the hands of Alexander's ruthless agents, the remaining members of P'arnavaz's family concealed themselves in the rugged peaks of the Caucasus. Despite the predicament, P'arnavaz was destined for glorious things. The ninthcentury historian describes him as wise (gonieri), an enterprising horse­ man (mq'edari shemmart'ebeli), and a skilful hunter (monadire q'elovani). He also benefited from good fortune, a concept denoted by the Old Georgian term sue. His very name indicates that he possessed farnah, the 'royal glory' that in the Persian imagination earmarked legitimate rulers. It 11 While the historical accuracy of specific events related in The Life of the Kings may be cast into doubt, contemporary Classical and Persian evidence, including Herodotus and Sasanian inscriptions, firmly situates Caucasia within the Persian sphere of influence. In addition, Toumanoff (Studies) and others have conclusively demonstrated that the social structure of pre-modern Caucasia was thoroughly Persian. For example, the countryside was dominated by Persian-like noble houses, and remained so down to the Russian conquest of the nineteenth century. There can be little doubt, then, that ancient and early medieval K'art'li was attached to the Persian world. Though its details might be later contrivances, The Life of the Kings reflects early K'art'li's connection to the near east. Thus the extraordinary research of N.G. Garsoian that has exposed Armenia's links to Persia must now be adjusted to comprehend neighbouring K'art'li: see esp. her collected works in Armenia between Byzantium and the Sassanians (London, 1985). 12 For the story of P'arnavaz, see The Life of the Kings, KC 1:20-26.

FR O M

BUMBERAZI T O BASILEUS

107

should be stressed that the Persian basis of P'arnavaz's name is not unique in Georgian history: before the rise of the Bagratids in the ninth century, K'art'velian royal nomenclature was overwhelmingly Persian.13 P'arnavaz became cognizant of his destiny during his stay in exile. First, he had a dream in which a ray of sunlight wrapped itself around his waist. Standing before the sun, P'arnavaz anointed himself with its essence.14 It should be remembered that solar imagery was central to Persian understandings of kingship and religion. Later, after P'arnavaz had awoken, the future king went hunting, a favourite activity of the Persian elite. A wounded deer led him to a hidden cave, inside which P'arnavaz subsequently found an inconceivable treasure. In a quest to restore his family name and to liberate his people from Azon, Alexander's tyrannical local governor, the emboldened P'arnavaz returned to K'art'li. Having gained a comprehensive victory over Azon, P'arnavaz declared himself the mep'e, or monarch, of the K'art'velians.15 Reminiscent of the primordial kings of the Persian epic, P'arnavaz endeavoured to create the institutional bases of his society. He allegedly promoted the use of the K'art'velian idiom of the Georgian language (k'art'uli) and even invented the Georgian alphabet; a patently false claim, for in reality it was devised by Christians in the late fourth/early fifth century a d . The author of The Life of the Kings makes the association with Persia even more explicit. He describes the king's alleged establishment of a network of regional governors or erist'avis (literally, 'the heads of the army/people') as having been predicated upon a Persian model: 'In this fashion did P'arnavaz order everything [in his realm by] imitating the kingdom of the Persians.'16 In some respects, K'art'li had become a Little Persia. While many similarities between the culture and society of K'art'li and Persia may be explained by a common genesis and evolution, there are instances, like P'arnavaz's creation of an administrative apparatus, of direct appropriation. The anonymous author of The Life of the Kings also applied the Persianlike description of kingship to P'arnavaz's pre-Christian successors. These 13 See the fundamental study by M. Andronikashvili, Narkvevebi iranul-k'art'ul enobrivi urt'iert'obidan, vol. 1 (Tbilisi, 1966), with extensive English summary, 'Studies in Iranian-Georgian Linguistic Contacts', 547-71. 14 D. Rayfield, The Literature of Georgia: A History (Oxford, 1994), 54-5. 15 Although P'arnavaz anointed himself with the essence of the sun (in his dream), The Life of the Kings also states that P'arnavaz declared his loyalty to the king of Asurastan, i.e., the Seleucids, from whom he reportedly received royal insignia. The deification of P'arnavaz (Mihran venerates his tomb) suggests a Hellenistic influence. Classical sources also infer a historical link between the Hellenistic Seleucids and the early kings of K'art'li, yet it is a Persian, and not a Hellenistic or some alternative, model of kingship that conditioned preBagratid historical writing. 16 The Life of the Kings, KC 1:254.

108

STEPHEN H. RAPP JR

images pervade the text, which concludes precisely with the conversion of the monarchy to Christianity in the first half of the fourth century. But what happened to this imagery following Christianization? It is commonly assumed that K'art'li was suddenly and forever thrust into the orbit of Christian Byzantium upon the conversion of King Mihran (Mirian) III c. 337 at the hands of the holy woman Nino. But local sources do not show the early Christian K'artVelian kings imitating Constantine 'the Great', nor do their authors describe their royal subjects in Eusebian terms. Rather, the Persian-like formulation of local kingship persisted, although Christianity was added to the mix, thus demonstrating the fluidity of self-identity existing along the Perso-Byzantine frontier. Overlooking the Mtkvari River in the Georgian capital Tbilisi now stands a statue of the early medieval king Vaxtang Gorgasali. He sits on a gigantic horse; behind him is Metexi church and before him, on the other bank of the river, is Sioni, the patriarchal cathedral. This sculpture, like the Vaxtang of popular memory, is surrounded by the Church. Yet the historical Vaxtang's association with Christianity was considerably more complicated. Vaxtang Gorgasali is the quintessential example of the melding of Christianity and the Persian-like culture of medieval eastern Georgia. Though Vaxtang ruled in the late fifth and early sixth centuries, his anonymous biography, The Life of Vaxtang, was written down in the initial years of the ninth century. Like P'arnavaz before him, his parentage was mixed and reflects K'art'li's connections to the Near East: he was the son of the K'art'velian king Mihrdat V (r. 435^7) while his mother, Sagduxt, was a Christianized Persian. While contemporary Armenian and Byzantine sources describe Vaxtang as wielding limited powers,17 the later Life of Vaxtang portrays him as an epic monarch marshalling unri­ valled authority and majesty.18 As his statue in Tbilisi suggests, he is regarded today as one of the leading heroes of the Georgian past. Though many of the details of the Georgian-language Life of Vaxtang are legendary, literary devices and outright fabrications, the source is an important historical document because of the light it sheds on the constel­ lation of Eurasian influences that were present in ninth-century Caucasia, 17 The historians are Ghazar (Lazar) P'arpec'i and Procopius respectively: Rapp, "Imagining History", 428-37. P"arpec"i's testimony is particularly important because he had a first-hand knowledge of K'artVelian affairs through his residency at a noble estate on the Armeno-K'art'velian frontier. 18 See B. Martin-Hisard, "Le roi géorgien Vaxt'ang Gorgasal dans l'histoire et dans la légende", in Temps, Mémoire, Tradition au Moyen Age: Actes du XHIe Congrès de la Société des Historiens médiévistes de l'enseignment Supérieur Public (Aix-en-Provence, 1983), 207-42; S. Kakabadze, Vaxtang gorgasali da misi xana (Vaxtang Gorgasali and his age) (Tbilisi, 1994); and V. Goiladze, Vaxtang gorgasali da misi istorikosi (Vaxtang Gorgasali and his historian) (Tbilisi, 1991), with Russian summary, 205-7.

FR O M

BUMBERAZI T O BASILEUS

109

and in particular, Georgia's equipoise between the civilizations of Byzantium and Persia. The most prominent Persian dimension of the imagined Vaxtang is his conspicuous portrayal as a Sasanian hero-king.19 Even in The Life of the Kings, pre-Christian K'art'velian rulers were believed to have had a corps of champion duellists at their command. These fantastic warriors were termed bumberazni. Before opposing armies would plunge into warfare, each camp would dispatch a champion bumberazi to engage in fierce hand-to-hand combat, often duelling on horseback. Some pre-Christian K'art'velian kings are said to have personally participated in such contests. It is striking that, in the Georgian tradition, bumberazni are confined to the Persian world. With one possible exception,20 the bumber­ azni found in K'art'lis c'xovreba are Persians, Armenians, Georgians and Iranic north Caucasians, all of whom belonged to the Persian common­ wealth. Indeed, the imagined Vaxtang publicly identified Persia as 'the land of heroes and giants'.21 This exalted status was never bestowed upon Rome or Byzantium in pre-Bagratid historiography.22 More than any other K'art'velian monarch, the Christian Vaxtang is clothed in bumberazi imagery. Vaxtang himself is said to have worsted several extraordinary champions, some of whom the author explicitly named for dramatic flair. In one instance, the imagined Vaxtang not only vanquished the king of Sindet'i in a physical bumberazi engagement, but in an intellectual one as well, in which the contestants hurled eloquent parables at one another.23 The imagined Vaxtang employs monotheistic

19 Although Vaxtang himself was a K'art'velian, his royal authority is described in terms reminiscent of the Great Kings of Sasanid Persia. The Great King is said to have sent a letter to Vaxtang which began: 'From Hurmazd, the King of Kings, to Vaxtang, Varan-XuasroTang, the valiant King of the Ten Kings' (The Life of Vaxtang, KC 1:158). At least within the purview of greater Caucasia, Vaxtang was a veritable shahanshah, a 'king of kings'. Scholars have noted parallels in the descriptions of the imagined Vaxtang and Bahram Gur (in the Shahnama): W.E.D. Allen, A History of the Georgian People From the Beginning Down to the Russian Conquest in the Nineteenth Century (London, 1932; repr. 1971), 77; and I.A. Orbeli, 'Bakhram Gur i Azade', in his Izbrannye trudy (Erevan, 1963), 549-554. 20 Polykarpos, a Greek, defeated by Vaxtang. Significantly, Polykarpos is never adorned with the heroic badge 'bumberazi'. 21 The Life of Vaxtang, KC 1:1709. 22 What is more, Vaxtang publicly declared the K'art'velian kings had been 'descended from the hero Nimrod, who before [other] kings became renowned on earth' (The Life of Vaxtang, KC 1:16112_13). It should be recalled that the medieval K'art'velians regarded Nimrod to have been the first king of the world and ethnically a Persian. References in The Life of Vaxtang might be related to the story about Nimrod in the apocryphon by Ps.Methodios. See also P.W. van der Horst, 'Nimrod after the Bible', in his Jewish World of Early Christianity, Novum Testamentum et Orbis Antiquus 14 (Fribourg-Gottingen, 1990), 220-32. I wish to thank S. Harvey for kindly providing this reference. 23 The Life of Vaxtang, KC 1:188-95.

110

STEPHEN H. RAPP JR

and Christian imagery in his parable, referring prominently to 'God the creator of all' and Christ. Vaxtang fabulously claims to have 'delivered Jerusalem, the holy city, where trod the feet of our Lord Jesus Christ ... And then I rescued all Christendom from ruin ...'. Yet when he uttered these words, he was campaigning in Sindet'i with his kinsman, the Great King of Sasanid Persia. At the same time, Vaxtang also dismissed the tension between Zoroastrianism and Christianity by reasoning that 'although the Persians are not in the true religion yet they know God the creator and believe in the spiritual life'.24 These are not the words of a king who had proclaimed his full allegiance to Byzantium (which Vaxtang had not done) but rather of one negotiating a tolerable and bene­ ficial medium between the antagonistic civilisations of Persia and Byzantium. Indeed, Vaxtang is depicted in his Life as being the 'mediator' (shuamdgomelobay) between the rival Persian and Byzantine emperors.25 But how was Vaxtang, a Christian king, to reconcile the Persian-like features of his image, ideology and behaviour? We lack Georgian histori­ cal sources written during Vaxtang's lifetime, but his biography alerts us to the persistence of Persian imagery in the ninth century, some five centuries after the Christianization of the K'art'velian Crown. I would argue that this later representation was fundamentally identical to that of Vaxtang's own day. Numerous contemporary sources confirm K'art'li's ancient Persian heritage and, if nothing else, why would K'art'velian authors suddenly invent a Persian-flavoured past? While the text's description of bumberazi skirmishes brightly illuminates K'art'li's connec­ tion to Persia, its champions have been modified slightly so that they were adorned with generic Christian attributes. Thus, immediately before and after grappling with his opponents, Vaxtang prayed to his Christian God, lauded the consubstantial Trinity, and attributed all success to the Christian faith. Vaxtang's feet might have been planted in the Persian world, but his soul belonged to the Christians. It is likely that the histori­ cal Vaxtang's subjects would have expected this, but we can assert with certainty that his biographer's ninth-century audience would have iden­ tified with the imagined Vaxtang's straddling of the Near Eastern and Byzantine worlds. Vaxtang's anonymous ninth-century biographer did not regard the K'art'velians' dual affiliation to the Persian and Christian realms as being necessarily incompatible. But what were the K'art'velians to make of the Byzantine emperor's claim to stand at the head of all Christendom? We are told that two Byzantine monks confronted the imagined Vaxtang about his invasion of Pontos and the abuses that his troops had inflicted 24 Ibid., 192-3, trans. Thomson, Rewriting Caucasian History, 209. 25 See Rapp, Imagining History', 421-3.

FR O M

BUMBERAZI T O BASILEUS

111

upon local Christians. At the crucial moment, and like P'arnavaz before him, Vaxtang had a dream: he saw a crown resting on the arms of a cross; in return for the crown he was compelled to recognize the supremacy of the Byzantine emperor. When the dream was interpreted for him by the (partisan) holy men, he opened negotiations with Constantinople and a peace was concluded. This vision ultimately persuaded Vaxtang to acknowledge the special status of the Byzantine emperor within Christendom. On his deathbed, the imagined Vaxtang summoned his nobles and implored: 'Do not despise our house, nor abandon the friend­ ship of the Greeks [that is, Byzantines]'.26 Even at this dramatic moment, however, the imagined Vaxtang did not reject his realm's intimate connection to the Persian cultural world. Vaxtang lived and died a bumberazi, but a Christian bumberazi. Just a few years after the composition of the three pre-Bagratid histo­ ries,27 a new princely dynasty assumed power. In 813, the K'art'velian Bagratid family seized the office of presiding prince. Ashot I (r. 813-30) inaugurated a dynasty that would occupy the throne for a thousand years down to the Russian conquest of the nineteenth century. At first, the Bagratids did not make history writing a priority. Political consolidation and ideological development were more pressing matters. In 888 they resuscitated K'art'velian kingship, which had been in abeyance since its abolition by the Persians in the sixth century. Then Bagrat III (d. 1014), thanks to his fortunate genealogy and the foresight of the kuropalates Davit' of Tao/Tayk' and his adviser Ivane Marushis-dze, became the first ruler of a politically unified and integrated Georgian kingdom in 1008. During the ensuing two centuries, Bagratid monarchs expanded their authority beyond the confines of Georgia itself, transforming Georgia into an imperial power. To make sense of this unprecedented state of affairs and to commemo­ rate their triumphs the Bagratids began to sponsor the work of historians. The 'Bagratid' period of Georgian historical writing thus commences in the first half of the eleventh century. The first order of business was to articulate the Bagratids' unique suitability to rule. Back in the early eighth century, the Armenian historian Moses Xorenac'i (Moses Khorenatsi) had traced Bagratid ancestry to a Jewish family.28 The Georgian branch of the Bagratids expanded this claim, fashioning itself as the direct biological offspring of the king-prophet David. As a consequence, no one, and no

26 The Life of Vaxtang, KC 1:203. 27 The third pre-Bagratid source, the untitled work by Ps.-Juansher, is an intermediate text that links the two historiographical periods: see Rapp, 'Imagining History', 486-92. 28 Moses Khorenatsi, History of the Armenians, 1.22, trans. R.W. Thomson (Cambridge, MA, 1978), 110-11.

112

STEPHEN H. RAPP JR

dynasty, could eclipse its legitimacy. The earliest Bagratid-era historical text, The Life and Tale of the Bagratids, written by Sumbat Davit'is-dze perhaps already in the 1030s, opens with a manipulated biblical geneal­ ogy tracing the descent of the Bagratids from Adam, through Noah, and including King David.29 The anonymous eleventh-century Chronicle of K'art'li also enunciates this Davidic claim.30 A comparison of the content and imagery of Bagratid-era histories and their pre-Bagratid counterparts exposes fundamental differences. First of all, in Bagratid texts Georgian territories other than K'art'li now received considerable attention. Increased information and interaction with outly­ ing Georgian regions like the kingdoms of Ap'xazet'i in the west and Kaxet'i in the east is duly reflected in contemporary historiography. But K'art'li remained the social, cultural, economic and political centre around which the unified Georgian kingdom gravitated and thus it retained a favoured status within the Georgian historical canon. In addi­ tion, and of immense importance for us, Georgia was no longer described as an integral component of the Near East. Bagratid historians abandoned both the imagery applied to P'arnavaz and Vaxtang as well as the general Persian context; instead they highlighted Georgia's connection with Christendom and especially with Byzantium. Reliable information about events in Byzantium became commonplace.31 This development may be partly explained by the fact that the Georgian Bagratids had risen to eminence with Byzantine support and aid. Moreover, their initial base in the south-western Georgian domains was situated along the border with Byzantine Anatolia. Byzantine influences 29 Georgian title: Cxorebay da ucqebay bagratoniant'a. Cf. the manipulated genealogy at the start of the pre-Bagratid Life of the Kings tracing the pedigree of the K'art'velian people. The Bagratids thus set themselves apart, and far above, their K'art'velian subjects. 30 Georgian title: Matiane k'art'lisay. Thomson, Rewriting Caucasian History, translates the title 'Book of K'art'li'. 31 Georgia's political relationship with Byzantium vacillated turbulently, though cultur­ ally the Georgians were becoming increasingly attached to the Byzantine commonwealth. The Georgian Church, for example, abandoned its traditional use of the Syro-Palestinian liturgy in favour of that of Constantinople. The Georgians were deeply affected by Byzantine theology, ecclesiastical art, hymnography and literature. The Byzantinization of Georgia occurred not only by means of a shared border, but also as the result of the pres­ ence of thousands of Georgian monks throughout Byzantine lands, including Syria, Palestine, Egypt, Cyprus, Anatolia, Bulgaria, on Mount Athos, and in the imperial capital itself. 'Diplomatic marriages' played a prominent role in the Byzantino-Georgian exchange: Bagrat IV (r. 1027-72) married Helena, niece of Romanos III Argyros (r. 1028-34); Bagrat's celebrated daughter, Mart'a, known in Byzantium as Mary 'of Alania', married Michael VII Doukas (r. 1072-78) and Nikephoros III Botaneiates (r. 1078-81). For Mart'a, see L. Garland, Byzantine Empresses: Women and Power in Byzantium, ad 527-1204 (London-New York, 1999), 180-86; and Garland and Rapp, 'Mary "of Alania": Woman and Empress between Two Worlds', forthcoming.

FR O M

BUMBERAZI T O BASILEUS

113

poured into this neo-K'art'li323and subsequently spread throughout the diverse Georgian regions by means of Bagratid expansion and parallel ecclesiastical consolidation (which frequently preceded the former). Three historical works written during the heyday of the medieval Georgian realm exemplify this trend: (1) The Life of King of Kings Davit' II, (2) The Histories and Eulogies of the Crowned (literally, of those Possessing Sharavandedi, for which see below), and (3) The Life of the Monarch of Monarchs Tamar 33 Like their predecessors, these histories largely focus on kingship. But in Bagratid texts, the very nature of royal authority and its context have been diverted from a Near Eastern paradigm towards a Byzantine one. Local rulers were no longer depicted as Persian-like herokings. Rather, the Bagratids and their historians and artists embraced the model of the basileus, or Byzantine emperor.34 Many of the Persian/Near Eastern elements of Georgian royal authority fell from usage while Byzantine imagery and contexts became increasingly common. To an unprecedented degree Bagratid historians embedded biblical quotations and allusions in their works, thus emphasizing the Crown's special attachment to the sacred. Historians now compared their monarchs to Christian-Byzantine heroes like Constantine; this is the case with King Davit' II (r. 1089-1125), who, in imitation of Constantine, summoned formal ecclesiastical councils, albeit only on an all-Georgian scale. Davit"s great-granddaughter Queen T'amar is said to have been 'a second Constantine' (r. 1184-1213);35 she is also compared to the king-prophet David, Abraham, and even Jesus.36 It is especially striking that no Bagratid monarch is likened to the K'art'velian heroes of old, including P'arnavaz and the Christian Vaxtang Gorgasali. From the time they seized power in K'art'li, the Bagratids proudly flaunted their Byzantine dignities and these were inflated sequentially

321 apply this term to the south-western "Georgian' domains between the eighth and tenth centuries. It was at this time that K'art'velian elites flooded into the region in order to escape the Arab occupation of K'art'li proper. The region was subsequently 'K'art'velized'. 33 Georgian titles: C'xorebay mep'et'-mep'isa davit'isi; Istoriani da azmani sharavandedt'anv, and C'xorebay mep'et'-mep'isa t'amarisi. The differences of title for the biographies of Davit' and T'amar are a product of translation: both titles use the Georgian mep'e, meaning 'monarch', 'king', 'ruling queen'. 34 Unfortunately, we lack the requisite sources from the early Bagratid period that would have exposed the transition from a Persian-like kingship to one more attuned to Byzantine imagery and ideology. 35 The Life of T'amar, KC 2:117. 36 A. Bryer, 'Preface: A View from the Byzantine Side of the River Akampsis', in K. Vivian, trans., The Georgian Chronicle: The Period of Giorgi Lasha (Amsterdam, 1991), x; and A. Eastmond, 'Royal Renewal in Georgia: The Case of Queen Tamar', in P. Magdalino, ed., New Constantines: The Rhythm of Imperial Renewal in Byzantium, 4th-13th Centuries, SPBS 2 (Aldershot, 1994), 289, who notes that T'amar's only female comparison is to Aphrodite.

114

STEPHEN H. RAPP JR

from kuropalates to sebastos. In the eleventh century, it is uncertain whether the emperor himself had sanctioned the Byzantine titles held by the Georgian kings; in any case, Georgian monarchs now believed it was their birthright to hold such titles, and some of these titles were made heredi­ tary on a unilateral basis. Titular inflation reached its zenith with Davit' II. Davit' believed himself even more esteemed than some mere caesar or sebastos. An inscription on a Georgian icon found at Mount. Sinai identi­ fies Davit' by the Greek basileus, the very title of the Byzantine emperor.37 Later, in the thirteenth century, the Georgian author of The Histories and Eulogies, a text about T'amar and her father Giorgi III (r. 1156-84), generically describes his work as a vasiloghrap'ia: a history dedicated to basileis, 'emperors'.38 Eventually, Davit' II discarded formal Byzantine honours once and for all, thus removing any implication that the Georgian monarch was inherently subordinate to the Byzantine emperor. After this action, the Bagratid kings increasingly adorned themselves with Georgian terms denoting 'autocrat' (for example, t'wt'mpqrobeli). The message was obvious: the Georgian Crown was not dependent upon the Byzantine emperor and, furthermore, the Georgian monarch was the equal of the emperor, at least within the purview of greater Caucasia.39 As we have seen, the Persian-like imagery of pre-Bagratid kingship was abandoned wholesale under Bagratid rule. Contemporary historians reformulated the highest grade of Bagratid kingship to evince expressly Christian concerns.40 The most extraordinary monarchs, like Davit' and T'amar, were recognized as Christian saints.41 In his Life, Davit' is explic­ itly said to have attained 'the grace of apostleship like Paul and the great Constantine'.42 Although no contemporary history specifically names 37 D. Kldiashvili, 'L'lcone de Saint Georges du Mont Sinai avec le portrait de Davit ATmasenebeli', REGC 5 (1989), 107-28. 38 Histories and Eulogies, KC 2:2lg. 39 This is not unlike the claims advanced on behalf of the Solomonids of Ethiopia. 40 Rapp, 'Imagining History', The Saint-King', 672-79. Consider several parallel develop­ ments: the head of the Georgian Church was not simply the kat'alikos, as he had been in the pre-Bagratid age, but now he was kat'alikos-patriarch. Byzantine innovations to the legend about St Andrew were embraced by the Georgians so as to endow their church with an apos­ tolic foundation. It is interesting that the Georgian Andrew legend appeared in contempo­ rary hagiography but was not incorporated into the historiographical tradition until the eighteenth century: for a translation of the relevant 'interpolation', see Thomson, Rewriting Caucasian History, 354-9. In terms of royal imagery, we have already seen how the Bagratids could be compared to the heroes of Christian Byzantium. 41 Three pre-Bagratid rulers also seem to have been considered saints. But precisely when Mirian, Vaxtang and presiding prince Arch'il II (r. 736-86) entered the ranks of the saints is not altogether clear, for it has not been definitely established when the Georgian Church formally instituted canonization. 42 M. Shanidze, ed., C'xorebay mep'et'-mep'et'a davit'isi (The Life of Davit' II) (Tbilisi, 1992), 20918_i 9; which supersedes the older KC 1:3544_5.

FR O M

BUMBERAZI T O BASILEUS

115

T'amar as a saint, in them she is afforded a special connection to God: her anonymous biographers could refer to her as 'the Mountain of God'43 and even as a model for monks.44 However, in the Vani Gospels, copied during her lifetime, T'amar is counted among the ranks of the saints 45 Though she is not endowed with such an exalted status in contemporary historical texts, The Eiistories and Eulogies could nevertheless identify her as the fourth member of the Holy Trinity.46 In his recent Royal Imagery in Medieval Georgia, Antony Eastmond has explored the theme of the Bagratid saint-king in Georgian art. He has demonstrated that of all Georgian secular figures to have been painted, only the Bagratids were regularly afforded haloes.47 These haloes were a graphical link between the Bagratids and the Christian God, a connection that could have been comprehended even by the illiterate masses. I concur with Eastmond, though I would also suggest that this halo motif artistically paralleled the Bagratid concept of sharavandedi: both were part of the sanctification programme of the Bagratids. Ironically, this compound word is based on the Middle Persian term shahravand, which itself is derived from the Persian shah, 'king', and the Avestan aurand, 'mighty, hero'.48 In the pre-Bagratid period, the word sharavandedi was used sporadically to denote the rite of consecration or coronation.49 In Bagratid times, sharavandedi meant the rays or even the corona of the sun. It was reserved as an honorific for the Bagratids and, on rare occasions, for important Christian saints. Many Bagratids were explicitly associated with the sun: Davit' II was 'the Sun of all Sovereignty';50 Queen T'amar was 'the Sun of Suns'51 and Tike the Sun [she] enveloped everyone in the 43 Histories and Eulogies, KC 2:21. 44 The Life of T'amar, KC 2:14717_lg. 45 Kekelidze Institute of MSS, Tbilisi, no. A -1335,1. 272: K'. Sharashidze, ed., Sak'art'velos saxelmcip'o muzeumis k'art'ul xelnacert'a aghceriloba: qop'ili saeklesio xelnacerebi (A kolek'c'ia) (Description of the Georgian Manuscripts in the State Museum of Georgia: formerly Ecclesiastical Manuscripts [A Collection]) vol. 4 (Tbilisi, 1954), 407-10. 46 Histories and Eulogies, KC 2:2522. 47 A. Eastmond, Royal Imagery in Medieval Georgia (University Park, PA, 1998), 162 and 198. 48 Traditionally sharavandedi is simply translated as 'kingship' (mep'obay) or as having to do with the royal crown (gwrgwni): Abuladze, Dzveli k'art'uli enis lek'sikoni (Lexicon of old Georgian language) (Tbilisi, 1973), 471 (with biblical references); and Z. Sarjveladze, Dzveli k'art'uli enis lek'sikoni (Lexicon of old Georgian language) (Tbilisi, 1995), 228-9 (with several references linking sharavandedi to the sun). Cf. the eighteenth-century lexicon of the scholar Sulxan-Saba Orbeliani, Lek'sikoni k'art'uli (Georgian lexicon), ed. Abuladze, vol. 2 (Tbilisi, 1993), 279 (whose first definition connects it with the rays of the sun). 49 The Martyrdom of Evst'ati, cap. 5, Abuladze, Monuments 1, 3733; trans. D.M. Lang, Lives and Legends of the Georgian Saints, rev. edn. (Crestwood, NY, 1976), 104: 'and God allotted [all Israel] a king and laid down the rite for his consecration' (emphasis mine). The oldest extant MS of the text has the corrupted sharadelobisay (Abuladze, apparatus criticus, 37, no. 39). 50 The Life of Davit', ed. Shanidze, cap. 1 5 ,16512_13. (KC 1:323 20). 51 Histories and Eulogies, KC 2:51.

116

STEPHEN H. RAPP JR

light radiated by her sharavandedi'.52 One of the histories about T'amar is titled Istoriani da azmani sharavandedi'ani, literally T h e Histories and Eulogies of Those Possessing Sharavandedi'. The Bagratids were not alone in applying solar imagery: not only did their K'art'velian predecessors take advantage of it, but so too did the Byzantines and various Near Eastern rulers. It is hardly surprising that the Georgians, wedged between the two, should do the same. In any event, I believe that the haloes illuminating the heads of Bagratids conflate the idea of the family's special link to the sacred and that of the affiliated notion of sharavandedi. And more than that, Georgian solar imagery, though partly inspired by the Byzantines, was textually repre­ sented by a Persian loan-word. The use of Arabic in legends on contem­ porary Georgian coins5253 and the fascination with Persian-style epics at the Bagratid court, most famously Shot'a Rust'aveli's Knight in the Panther's Skin ( Vep'xistqaosani), are further indications of Georgia's ongoing bond with the Near East. The Bagratid concept of sharavandedi thus bridged the Near Eastern and Byzantine worlds. Even during Georgia's 'Golden Age' the Bagratid monarchs had not entirely abandoned their ancient connec­ tions to the Near East. Although the early Bagratid kings were textually clothed in Byzantine imagery, after Davit' II we see the Bagratids adopting an image that is more generically Christian. Yet careful examination shows that select Near Eastern motifs were still present in the Bagratid conception of royal authority. Thus the balancing act between 'east' and 'west' was sustained even during the Bagratid age, though Georgia's primary orientation within the Eurasian world had been reformulated, at least as far as kingship is concerned, so that after the ninth/tenth century it was more clearly associated with the Christian Byzantine commonwealth.54

52 The Life of Tamar, KC 2:14731_32. 53 See Rapp, The Coinage of T'amar, Sovereign of Georgia in Caucasia', Le Muséon 106/3-4 (1993), 309-30, for the fusion of Georgian, Islamic and Byzantine motifs. Among prominent Byzantine ones is the labarum, which may have reached the Georgian court via Christian Rus'. 54 This association took an extraordinary turn following the capture of Constantinople by the Crusaders in 1204. At that time, some members of the Komnenos family made their way to the Georgian court. Queen T'amar eventually provided troops to assist these Komnenoi in their capture of Trebizond where a Byzantine kingdom was established. T'amar's eager­ ness to lend assistance to the Komnenoi was due in part to her own blood-link to the family: A. A. Vasiliev, The Foundation of the Empire of Trebizond (1204-1222)', Speculum 11 (1936), 3-37; Toumanoff, 'On the Relationship between the Founder of the Empire of Trebizond and the Georgian Queen Thamar', Speculum 15 (1940), 299-312; and M. Kurshankis, 'Relations matrimoniales entre Grandes Comnènes de Trébizonde et princes géorgiens', BK 34 (1976), 112-27. The Georgian family Andronikashvili (cf. Andronikos) daims descent from the Komnenoi.

Section III Byzantines

This page has been left blank intentionally

8. Bearing gifts from the east: imperial relic hunters abroad Liz James* Between the fourth and the sixth centuries, the Holy Land emerged from a relative obscurity to become a spiritual focus of the Byzantine empire. Through pilgrimage to the holy places or through contact with the holy in the form of holy people, living or dead, and through objects associated with the holy, Christians sought intimacy with the divine for personal reasons: faith, for healing and penance, for advice and revelation. One of the ways in which the east approached Byzantium, or, perhaps more accurately, Byzantium appropriated the east, was through the appre­ hending, movement and transmission of objects associated with the holy - Christian relics. The eastern provinces were despoiled methodically and enthusiastically, for the benefit of the New Rome, for they held the most ancient and venerable relics, the leavings of Christ, the Virgin, and a multitude of Old and New Testament figures. This chapter is framed around a simple question: what relics in Constantinople were brought there from the East by emperors and empresses? And why? Constantinople, when refounded by Constantine, was, unlike Rome, a relic-poor city, which possessed the remains of a handful of native martyrs from the pre-Constantinian persecutions, but nothing more spectacular. By 1204, through hard work and enterprise, it boasted one of the great relic collections of Christendom, particularly rich in the instruments of the Passion - the crown of thorns, the sponge and lance, the pillar at which Christ was scourged, stone from his sepulchre and a considerable quantity of wood from the True Cross.*1 It was rich also in the Virgin's clothes, a couple of heads of the Baptist and most of his * This chapter is for Jill Storer. 1 O. O. Meinardus, 'A Study of the relics of saints in the Greek church', OrChr 54 (1970), 267-9, lists the Passion relics in Constantinople. A. Frolow, La relique de la Vraie Croix, recherches sur le developpment d'un cult (Paris, 1961), discusses the relics of the True Cross. From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

119

120

LIZ JAMES

hair, and remains of an assortment of other saints. Indeed, it has been calculated that there were about 3,600 relics of 476 Greek saints in Constantinople, a figure which is generally perceived as substantial.2 So, which of these relics came from the east via the emperors to Constantinople, and why? It has been said that 'many of the relics in the city owe their presence to imperial interest and initiative'.3 However, of the 3,600, I calculate that, between Constantius and John V Palaiologos, only 38 relics, together with the instruments of the Passion, are directly, so to speak, associated with specific emperors bringing them into the city. This may simply be a result of how the sources record it: the 38 are all cases where the sources say explicitly 'and emperor so-and-so discovered/brought relic such-andsuch into the City'. Usually the sources use phrases like 'under emperor X' or 'in the time of so-and-so', which is not so clear; thus I have, for present purposes, discounted these. The relics-for-Constantinople story begins with Helena, mother of Constantine the Great, and the tale of the discovery of the True Cross. Whether true or false, that story set a pattern for what Leslie Brubaker has called imperial matronage: empresses imitating the pious actions of Helena in church building and in relic collection.4 To Helena is ascribed the discovery of the True Cross and the nails from the crucifixion, the building of a church in Jerusalem to house the cross and the sending back of part of the cross to Constantine in Constantinople.5 After Helena, there is a gap in recorded relic-hunting until Eudoxia, the wife of Arkadios, is recorded as taking the relics of various martyrs to the church of St Thomas in Drypia.6 Eudoxia's daughter, Pulcheria, and Pulcheria's sisterin-law, Eudokia, wife of Theodosios II, are also both involved in relic dealing. Either or both played a part in the translation of relics of St Stephen from Jerusalem to Constantinople.7 Pulcheria also received the relics of Sts Laurence and Agnes, was involved with articles of the 2 Meinardus, 'A study', 130-33. 3 J. Wortley, The Trier ivory reconsidered', GRBS 21 (1980), 383, gives approximately twenty-five cases of alleged relic acquisition by Constantinople (note that this is to the capital, not by the emperors) between 350 and 450. 4 L. Brubaker, 'Memories of Helena: patterns in imperial female matronage in the fourth and fifth centuries', in L. James, ed., Women, men and eunuchs. Gender in Byzantium (London and New York, 1997), 52-75. 5 For an account of Helena's actions and discussion of their historical accuracy or other­ wise, see J.W. Drijvers, Helena Augusta: the mother of Constantine the Great and the legend of her finding of the True Cross (Leiden, 1992). 6 John Chrysostom, Homily II, Dicta postquam reliquae martryrium, PG 63, 467-72. See also L.J. Wilson, 'The Trier ivory: a new interpretation', Byz 54 (1984), 602-14. 7 For Pulcheria and St Stephen's relics, see Theophanes, Chronographia, ed. C. de Boor (Leipzig, 1883-85), 86 (AM 5920); trans. C. Mango and R. Scott, The Chronicle of Theophanes

IM P E R IA L R E L IC H U N T E R S A B R O A D

121

Virgin's clothing, and was enabled through a vision to find the relics of the 40 martyrs of Sebaste which had been buried in the church of St Thyrsus within the city but sadly forgotten.8 Eudokia for her part sent from Jerusalem the icon of the Virgin painted by St Luke, the ur-icon of the Virgin, and brought back with her after her first trip there assorted other relics.9 Pulcheria's successor, Verina, played a role in the arrival of the Virgin's robe in Constantinople.10 And other empresses, Ariadne, Sophia, Constantina, wife of Maurice, Martina, Theodora the Khazar and Eirene-who-restored-the-icons, are also associated with relic dealing one way or another. The role of empresses in the translation of relics to Constantinople is graphically illustrated by the Trier ivory. This ivory plaque measuring 13.1 by 26.1 by 2.3 cm is now in the Trier Domschatz. It is clearly one side of a small box or casket, the rest of which has completely disappeared. It depicts a procession. Two bishops, seated on a cart drawn by two horses and holding a coffer proceed through a gateway, above which is an image of Christ. The procession is led by a group of male figures on foot and in imperial costume moving towards a standing female figure in imperial costume. She stands in front of a building which is clearly a church and equally clearly is still under construction. Behind, crowds of people line a colonnaded approach. There are no inscriptions of any sort on the ivory. It has been suggested that in the Trier ivory we have 'an accurate repre­ sentation of the historical circumstances [of the event] as they have come down in the sources'.11 Unfortunately, we have no idea what this accu­ rately represented event was. It is not known who the figures are, what the procession represents or even what date it is. All that can be said is (Oxford, 1997), 135-6. For Eudokia and St Stephen's relics, see Marcellinus Comes, Chronicle, in T. Mommsen, ed., Auctores Antiquissimi II, Chronica minore 2 (Berlin, 1884), 37-108 (yr 439); trans. B. Croke, Marcellinus Comes, Chronicle (Canberra, 1995), yr 439; E.D. Hunt, Holy Land Pilgrimage in the Later Roman empire a d 312-460 (Oxford, 1982), esp. ch. 10. 8 Sozomen, Ecclesiastical History: J. Bidez and G.C. Hansen, eds, Sozomenus Kirchengeschichte (Berlin, 1960), IX, 2; Chronicon Pascale, ed. L. Dindorf (Bonn, 1832), yr 541; trans. M. and M. Whitby, Chronicon Pascale (Liverpool, 1989). For Pulcheria's housing of relics of the Virgin, see K. Holum, Theodosian empresses (Berkeley, 1982), 142 and n. 120. 9 This story is recorded by Nikephoros Kallistos Xanthopoulos in the fourteenth century as excerpted from Theodore Lector's Ecclesiastical History: see PG 86,165A, trans. C. Mango, The art of the Byzantine Empire (Toronto, 1974), 40. See also the remarks of C. Walter, 'Iconographie considerations', in J.A. Munitiz, J. Chrysostomides and E. Harvalia-Crook, eds, The letter of the three patriarchs (Camberley, 1997), iv, noting that this story is not recorded anywhere else. 10 A. Wenger, 'Notes inédites sur les empereurs Théodose I, Arcadius, Théodose II, Léon I', REB 10 (1952), 47-59. 111. Kalavrezou, 'Helping hands for the empire: imperial ceremonies and the cult of relics at the Byzantine court' in H. Maguire, ed., Byzantine Court Culture from 829 to 1204 (Washington, DC, 1997), 59.

122

L IZ JA M E S

that it represents the arrival of some sort of relic in Constantinople, where it is greeted by an empress. Stylistically, the ivory has been dated to any point between the fifth and seventh centuries.12 As a result of this, almost any empress who had anything to do with relics has been advanced as a suitable candidate, from Helena, the mother of Constantine, to Theodora the Khazar, wife of Justinian II. The scene has been identified as repre­ senting Eudoxia and the procession to the church of St Thomas, or as representing Pulcheria with the relics of St Stephen, perhaps a piece produced some centuries after the event. The empress has also been iden­ tified as Verina, wife of the fifth-century emperor Leo I, in which case the church represents the Chalkoprateia, and the scene relates to the transla­ tion of the Virgin's robe; and as Martina, with Heraklios and the True Cross.13 It would be quite possible to make out a case for the image to represent the empress-regent Eirene, Constantine VI and the relics of St Euphemia. Although this question of identity is unlikely ever to be answered satis­ factorily, what is interesting is the number of empresses associated with it. As a scene of relic translation, it has proved possible to associate the Trier ivory with the majority of empresses from between the fourth and seventh centuries, who are all recorded as having relics brought to Constantinople from the east. In other words, relic-collecting was a stan­ dard action for many empresses. But why? What did empresses expect from relics? This question raises the further question, what were relics for? Pilgrimage carried with it the possibility of recreating the biblical past; the cult of relics also served to make the holy real and present, embodying the saint and serving as a visible testimony of the saint's continued presence, often in a place far removed from the relic's homeland.14 Jack Goody has isolated three factors of devotion: reverence, healing and protection, 12 For the most recent catalogue entry on the Trier ivory, see C. Stiegemann and M. Wernhoff, eds, Kunst und Kultur der Karolingenzeit (Mainz, 1999), cat. VIII/9. 13 For the empress as Eudoxia: Wilson, The Trier ivory'; as Pulcheria: K. Holum and G. Vikan, The Trier ivory, adventus ceremonial and the relics of St Stephen', DOP 33 (1979), 115-33, and also Wortley, Trier ivory reconsidered'; and most recently, L. Brubaker, The Chalke Gate, the construction of the past and the Trier ivory', BMGS 23 (1999), 258-85; as Verina: V. Grumel, 'A propos de la plaque d'ivoire du trésor de Trêves', REB 12 (1954), 187-90; as Martina, see S. Spain, 'The translation of relies ivory, Trier', DOP 31 (1977), 281-304. For the Virgin's robe: J. Ebersolt, Sanctuaires de Byzance (Paris, 1921), ch. 4, esp. 44-7; and N.H. Baynes, The finding of the Virgin's robe', in Byzantine studies and other essays (London, 1955), 240^17. Ebersolt also suggests that the reign of Arkadios and the relics of the prophet Samuel or the reign of Justinian might be the event commemorated in the Trier ivory, Sanctuaires, 14 and n. 3. 14 See E.D. Hunt, The traffic in relics: some Late Roman evidence', in S. Hackel, ed., The Byzantine Saint (Birmingham, 1981), 171-80.

IM P E R IA L R E L IC H U N T E R S A B R O A D

123

which are all essentially personal.15 Certainly in Byzantium, relics were reverenced, and employed for protection and healing. If honoured, the holy dead would use their powers to defend their worshippers.16 They also had further consequences. The discoverer of a relic could gain indi­ vidual credit; such a discovery and translation were clearly a sign of God's favour to the individual, and the individual might anticipate the powers of the holy dead being used to their benefit. Such personal sanc­ tity in Byzantium had considerable knock-on effects. It has long been recognized that emperors derived considerable prestige from the transla­ tion of relics. Similarly, where it has been suggested that 'there can have been relatively few occasions in the early Byzantine period when an empress translated relics to support her own basileia',17 her own imperial power, the reverse is true. In every case of relic translation I have mentioned, the empress's basileia was involved. This was not a cynical manipulation of relics for the sake of power; rather, as God was conceived of as intrinsic to every part of life in Byzantium, any action involving God and the holy was bound to have repercussions beyond 'simple' piety. Reverence, piety, healing and the rest could also add up to power and prestige, particularly for empresses who were so intimately linked with God's chosen emperors. In every case of relic recovery that I have cited above, the empress gained something from this action. Eudoxia, thanks to her troubled rela­ tionship with John Chrysostom, was always in need of a few signs of divine favour to indicate she was in the right.18 In Pulcheria's case, the translation of St Stephen's right hand is dated to 421, which coincidentally is the date of Theodosios II's war with Persia, a war which Holum has described as 'Pulcheria's crusade'.19 Whether or not that was the case, it was certainly a good moment for potent relics to come in to Constantinople. As for her involvement in the Virgin's robe and girdle, Pulcheria's cult of the Virgin, its potential role in her basileia and her 15 J. Goody, Representations and contradictions (Oxford, 1997). On relics as supernatural defenders, see N.H. Baynes, The supernatural defenders of Constantinople', AB 67 (1947), 165-77 and A.M. Cameron, The Virgin's robe: an episode in the history of early seventh century Constantinople', Byz 49 (1979), 42-56. 16 Baynes, 'Supernatural defenders'. E. Jeffreys also raised the question of the place of relics in the dedication of churches. Certainly churches were built for relics, but whether all churches had to contain relics is less clear despite the demands of Canon VII of the Seventh Ecumenical Council. 17 Holum and Vikan, Trier ivory', 133. 18 See Holum, Theodosian empresses, ch. 2 and J.N.D. Kelly, Golden mouth (London, 1995), passim. 19 K. Holum, 'Pulcheria's crusade ad 421-422 and the ideology of imperial victory', GRBS 18 (1977), 153-72; though see the arguments of R.W. Burgess, 'The accession of Marcian in the light of Chalcedonian apologetic and Monophysite polemic', BZ 8 6 /7 (1993-94), 47-68, against Holum's picture of Pulcheria.

124

L IZ JA M E S

quarrel with Nestorios are well-enough known not to need rehearsing again.20 The Forty Martyrs were discovered either in 434-46,21 the time when Pulcheria's rival, the empress Eudokia, might be thought to have been reaping the rewards of her pilgrimage to Jerusalem, or they were found in 451, which, coincidentally, was the year in which Pulcheria and Marcian were crowned, so another good moment for God's favour to be shown.22 As for Eudokia, whatever her motives for her trips to the Holy Land, the power and prestige offered by both the pilgrimage and the translation of relics was an opportunity not to be missed. And similar links between relics and prestige are apparent in the case of other empresses. To deal in relics was a sign of piety and piety was one of the qualities of a good empress. Verina's links with the Virgin's robe are clearly a significant factor in her appearance in sources as 'pious and orthodox' and may well have provided her with a sort of symbolic capital of prestige contributing to her status as dowager empress mother and her role in the disturbances of the early part of Zeno's reign. Similarly, Martina's links with the True Cross may have been deliberately engi­ neered to play down her position as incestuous wife of Heraklios. For all empresses, there was also the further benefit of association with the sainted Helena, mother of Constantine the Great. Of course emperors also collected relics and accrued prestige. Relic acquisition began with Constantine and appears to cease temporarily with Justinian II, resuming again in the eighth century with Eirene. Interestingly, it tends to be the spouses of those empresses who collected relics who also participate in this activity. Of the emperors directly linked with relic acquisition before Iconoclasm, only the wives of Constantius and the wives of Theodosios I are not recorded as having collected relics, and of these, Aelia Flacilla had a well-defined position of piety that perhaps needed less of the support offered by relics than her husband's. It was Constantius rather than Constantine who played the major role in the collection of relics in Constantinople. It was in his reign that the bodies of Andrew, Timothy and Luke arrived in the church of the Holy Apostles.23 Theodosios I had the bodies of the martyrs Terentius and

20 See, for example, Holum, Theodosian empresses. V. Limberis, Divine heiress. The Virgin Mary and the creation of Christian Constantinople (London and New York, 1994) relies exten­ sively on Holum's analysis. 21 Sozomen, EH, IX, 2. 22 Chronicon Pascale, yr 451. 23 See, for example, Chronicon Pascale, yr 357. For the debate about whether it was Constantine or Constantius who was responsible and the possibility that the translation took place in 336 see C. Mango, 'Addendum' to 'Constantine's mausoleum and the translation of relics', BZ 83 (1990), 434, repr. in Studies on Constantinople (Aldershot, 1993), V. Constantius also brought relics to Constantinople from Antioch. See Ebersolt, Sanctuaires, 40.

IM P E R IA L R E L IC H U N T E R S A B R O A D

125

Africanus brought to Constantinople and placed in the martyrion of St Euphemia. He also found the head of John the Baptist in a small village near Chalcedon, where it had arrived in the reign of Valens from Jerusalem via Cilicia. In 391, he wrapped it in his purple robe and brought it to the Hebdomon, where he built a church for it.24 However, it seems to have had little attention as, in the reign of Marcian, another head of the Baptist surfaced in Emesa. This was either brought to Constantinople by Justinian or by Michael III, depending on which source you read.25 The remains of the prophet Samuel were transferred to Constantinople in the reign of Arkadios.26 Theodosios II, at the rededication of Hagia Sophia in 415, seems to have had the relics of Joseph the son of Jacob, and Zacharias the father of John the Baptist installed.27 He also translated the relics of John Chrysostom, exiled by his father, to the Holy Apostles.28 Leo I brought in the bodies of the three Hebrews from the fiery furnace. Daniel the Stylite was later buried under them.29 Zeno had the body of St Barnabas, found on Cyprus clutching a copy of Matthew's Gospel written in Barnabas's own hand, appropriated.30 Justinian looked west and asked for relics of Peter and Paul from Rome, though he did not get them. But he did put the trumpets from Jericho and the well-head that Christ sat on to talk with the Samaritan woman into Hagia Sophia.31 Justin II had some of the hair of John the Baptist and the bodies of the Myrophores put into the church of Christ in Chalke, and built the oratory of St James, putting into it relics of the Holy Innocents, Symeon, Zacharias, father of the Baptist, and James the brother of Christ in it.32 Again, however, it is unclear which of these were already in the city and were being re-homed; Zacharias, for example, had previously been part of the Hagia Sophia collection. Heraklios rescued the True Cross, and in 614 Patriarch Niketas 24 Terentius and Africanus: Theodore Lector, Ecclesiastical History II, 62; head of the Baptist: Sozomen, EH, VII, 21, Chronicon Pascale, yr 391. 25 For Marcian, see Marcellinus, Chron., yr 453 and Chronicon Pascale, yr 453; for Justinian, see Acta Sanctorum, juin, vol. 4, p. 741; for Michael III see the Menologion of Basil; II Menologio di Basile II (Turin, 1907), II, pi. 420. Also Ebersolt, Sanctuaires, 80-81. 26 Chronicon Pascale, yr 406. 27 Chronicon Pascale, yr 415. 28 Theodoret, Historia Ecclesia, ed. L. Parmentier, rev. F. Scheidweiler, Kirchengeschichte (Berlin, 1954), V, 36 ,1-2. 29 Analecta Bollandiana XIII, 406-7. 30 George Monachos, Chronicon, ed. C. de Boor (Leipzig, 1904), II, 619; Georgius Cedrenus, Compendium Historiarum, ed. I. Bekker (Bonn, 1838-39), I, 618-19. 31 Narratio de S. Sophia, ch. 22 in T. Preger, ed., Scriptores originum Constantinopolitanarum I (Leipzig, 1901), 98-9, on the well-head and the trumpets from Jericho. Procopios, Buildings, I, VII, 2-10 records Justinian's (re-)discovery of the Forty Martyrs and his healing through their relics as a sign of God's special favour to the emperor. 32 Patria of Constantinople, III, ch. 148 in Preger, ed., Scriptores originum II, 263; A.M. Cameron, The cult of the Theotokos in sixth century Constantinople', JTS 29 (1978), 98.

126

L IZ JA M E S

brought the instruments of the Passion to Constantinople.33 Justinian II had the body and omophorion of St Anne brought to the city 34 It is no accident that practically all of these relics are relics of Old and New Testament characters, not of later saints and martyrs, power from the source of Christianity: the Holy Land and the Bible. It has been suggested that the eastern church venerated these Old Testament figures primarily because Constantinople had so many relics of them, in contrast to the west.35 This seems a rather chicken-and-egg view and begs the question of why emperors and empresses collected these relics in the first place, particularly as both the Theodosian Code and the Justinianic Code forbade any interference with the bodies of the dead.36 The existence of such laws may have played a part in the collection of objects such as the axe of Noah and the staff of Moses. The translation of relics implies compelling reasons, both specific to the relic and the collector and also to more general reasons of imperial authority.37 Relic collecting and re­ housing related to the imperial virtues of piety and philanthropia. However, any action which related to imperial virtues also related to imperial power. On the broadest possible scale, the translation of relics was to the greater glory of the God-chosen city of Constantinople. Further, many of these relics, collected by both emperors and empresses, and most notably the Passion relics with their direct link with Christ, disappeared into the palace churches. Here, of course, access was obvi­ ously limited. In other words, these relics were not collected for general use and veneration, but became part of the display of imperial power through ceremonial and ritual. They offered power, prestige, fame and spiritual advantage to those who could appropriate them. What is also suggestive is the nature of the sources which mention the translation of relics. Many are contemporary with the events they describe; others are not. Wortley has already outlined some of the issues around Theophanes's account of Pulcheria's involvement with the relics of St Stephen:38 reasons for this may relate to the Iconoclast dispute.

33 Chronicon Pascale, yr 614. 34 Patria III, ch. 79, Preger, ed., Scriptores originum, II, 244. 35 J. Wortley, Tconoclasm and Leipsanoclasm: Leo III, Constantine V and the relics', ByzF 8 (1982), 263. Wortley quotes John of Damascus on the hierarchy of the saintly: the Virgin, the Baptist, Apostles, martyrs, above all Stephen the first martyr, the ascetics and finally the Old Testament holy, De Fide Orthodoxa IV, PG 9 4 ,1168A. The relics acquired by emperors fit this hierarchy. 36 Codex Theodosianus IX 17.7 and Codex Justinianus III 44.14, which qualifies this by adding 'except with the permission of the emperor'. 37 As C. Mango, 'Constantine's mausoleum and the translation of relics', BZ 83 (1990), 51-62, argued in the case of the relics of Timothy, Andrew and Luke. 38 Wortley, 'Trier ivory'.

IM P E R IA L R E L IC H U N T E R S A B R O A D

127

George Monachus and Cedrenus are also late sources for which no corroboration exists. In a sense, these texts reflect the concerns of their times and may be writing these back into their perceptions of the past, perhaps to justify and validate the present. The authors of patriographic sources, for example, tried to make sense of their urban surroundings and the history of their city, and are concerned with images in particular, almost certainly again as a result of Iconoclasm.39 The fate of relics during Iconoclasm is not always clear. Iconophile sources make an attempt to blacken Constantine V as a destroyer of relics, yet, as John Wortley has shown, these stories may be more of a reflection of Iconophile propaganda, and Iconoclasm may well not have led to leipsanoclasm.40 John of Damascus felt compelled to defend the place of relics in the Orthodox faith, but whether his defence of them means that they were in genuine danger during Iconoclasm is a debatable point. After Iconoclasm, the ninth and tenth centuries saw a period of increased relic hunting. Leo VI acquired the body of Lazarus from Cyprus; Romanos Lekapenos and Constantine VII brought back the Mandylion and Constantine had the body of Gregory Nazianzenos installed in the Holy Apostles; in 956, he obtained the right arm of John the Baptist. Nikephoros Phokas recovered a lock of the hair of John the Baptist, matted with blood, and the Holy Brick, an acheiropoietos image of Christ; he also obtained fragments of the True Cross in 965 from Tarsos. John Tzimiskes obtained the sandals of Christ and yet more hair of the Baptist.41 As with the acquisition of relics before Iconoclasm, motive forces beyond 'simple piety' are apparent. In the case of Leo and Lazarus, Shaun Tougher suggests that this relic was obtained for reasons of personal 39 See A. Kazhdan's review of A. Cameron and J. Herrin, eds, Constantinople in the early eighth century: the Parastaseis syntomoi chronikai (Leiden, 1984) in BZ 80 (1987), 400-403, and A. Cameron, 'Byzantium and the past in the seventh century: the search for redefini­ tion', in J. Fontaine and J.N. Hilgarth, eds, Le septième siècle. Changements et continuités (London, 1992), esp. 257 and n. 18. On the Patria in particular, see G. Dagron, Constantinople imaginaire (Paris, 1984). 40 Wortley, 'Iconoclasm and Leipsanoclasm'. 41 For Leo and Lazarus, see Georgius Monachus Continuatus, ed. I. Bekker (Bonn, 1838), 860. Arethas and Patriarch Nikolaos both mention the translation: see R.H. Dolley, 'The histori­ cal significance of the translation of St Lazaros from Kypros to Byzantion', Byz 19 (1949), 59-71 and S.F. Tougher, The reign of Leo VI (Leiden, 1997), 201 and n. 53, correcting Dolley's dating. For the Mandylion, see Theoph. Cont., VI, 48, 432; Cedrenus, II, 319, and also A.M. Cameron, 'The history of the image of Edessa: the telling of a story' in Okeanos. Studies presented to I. Sevcenko, Harvard Ukrainian Studies 7 (1983), 80-94, and A.M. Cameron, 'The mandylion and Byzantine iconoclasm', in H.L. Kessler and G. Wolf, eds, The Holy Face and the paradox of representation (Bologna, 1998), 34-5. For Nikephoros Phokas, Cedrenus II, 364; Leo Diaconus, Historiae Libri X, ed. C.B. Hase (Bonn, 1828), IV, 10 (70-71). For Tzimiskes: Leo the Deacon, X, 4.

128

L IZ JA M E S

piety, for just as Lazarus had died and been resurrected three days later by Christ, so Leo had been imprisoned for three years and released to life as emperor through the will of God.42 This may be personal piety but it also carried a political message. The Mandylion came to Constantinople in 944. Depending on which source one reads, this was either the doing of Romanos Lekapenos or of Constantine VII Porphyrogennetos, and the sources stress the benefits each gained. Constantine, for example, gained great approbation for his action, a deed portrayed as being as great as Heraklios's recovery of the True Cross and one which boosted Constantine's claim to the throne. The image demonstrated its favour for Constantine in causing a man possessed by demons to call out that the kingdom was Constantine's. Theophanes Continuatus records too that Constantine was able to see the image which was indistinguishable to his brothers-in-law; this too must have served as a sign of divine favour for the legitimate emperor.43 Constantine used the arm of the Baptist to score his own personal points, the Baptist being of particular significance for the Macedonian dynasty and its claims to power.44 For the fifth-century empresses, relic hunting was easy, for the Holy Land, site of all the best relics, was part of the empire. Later emperors such as Heraklios, Nikephoros Phokas and John Tzimiskes had to liberate what they could though conquest. Nikephoros Phokas gained his relics during his campaigns of 968 in Syria. John Tzimiskes picked up the sandals and hair of the Baptist on campaign in the Euphrates in 975 and installed the sandals in the church of the Virgin in the Palace and the hair in the church of Christ in Chalke. In the same year, he took an icon of the crucifixion from Beirut to go into the same church. Such recoveries of sacred objects must have served as an indication of God's favour and intervention, rewarding his chosen emperors (both of whom were ruling in place of the legitimate heirs, Basil and Constantine, both minors) for their sterling efforts on behalf of the Christian empire. Relics were visible signs of the holy and proof of the Bible and of God's love for humanity. But they also seem to have acted as totems, as palla­ dia, as protectors. At the foot of his column, Constantine had put what Nicholas Kallistes called 'the inviolable treasure' of the capital: twelve baskets from the multiplication of the loaves and fishes, the axe that Noah used when building the ark, the two crosses of the thieves and the jar of perfume with which Christ was anointed in the Oratory, as well as the Palladium from Troy.45 Similarly, the Virgin's robe served as a palladium,

42 My thanks to Shaun Tougher for discussion on this point. 43 Theoph. Cont., VI, 48, 432. 44 See Kalavrezou, 'Helping hands', 67-79. 45 Kallistes, Hist. Eccl., VII, 49, PG 145, 1325. For the relics under the column of

IM P E R IA L R E L IC H U N T E R S A B R O A D

129

protecting Constantinople. As supernatural defenders, it is unsurprising that their proper place was with the human defender of the empire, God's chosen emperor. Significantly, as well, these are all objects associated with the holy rather than actual body parts. These relics retrieved by imperial relic hunters, which are overwhelm­ ingly of this particular biblical character, relate far more to protection and to imperial status under Christ than do the bodies of post-biblical saints. In this context, what was the relationship between relics and icons in Byzantium? John Wortley proposed that at no point in the history of Constantinople were the icons of the city 'ever so numerous or famous as her relics' and cited sources such as Fulcher of Chartres which mention relics but not icons; he also suggested that, for foreign pilgrims, it was relics not icons that mattered.46 This last is not totally surprising, and says little about Byzantine attitudes to icons and much about western attitudes to relics. Westerners and the non-Orthodox were not particularly inter­ ested in icons; Orthodox pilgrims, notably the Russian travellers to the city, were very interested in icons.47 Indeed, after the seventh century, furta sacra seems more of a western phenomenon - the most notorious thefts, of St Mark's body from Alexandria and of St Nicholas from Myra, are indeed relics moving east-west again, but from Byzantium.48 But how did the Byzantines relate relics and icons? Meinardus's figure of 3,600 relics cited above sounds considerable. However, Janin's Géographie Ecclésiastique de l'empire byzantin lists, at a rough count, about one thousand churches in Constantinople.49 To have an equivalent number of icons to relics leads to an average of fractionally over three icons per church, a figure which is not difficult to imagine. In other words, there were at least as many icons as relics in the city. While many of the major relics in Constantinople were in imperial possession and kept in the Great Palace, even the most wonder-working of icons, such as the Hodegetria and the Blachernitissa, were housed in more public churches, processed around the city and even accompanied the army on campaigns. Are icons a different form of devotion? In the case of the Virgin's robe, although it is her robe which is seen as the palladium of Constantinople, Constantine, see, among others, Cedrenus, 1,518,565, Patria II, ch. 20 in Preger, ed., Scriptores originum II, 161. 46 Wortley, 'Iconoclasm and leipsanoclasm', 253 and n. 2. 47 As just a skim through the accounts translated in G. Majeska, Russian travellers to Constantinople in the fourteenth and fifteenth centuries (Washington, DC,1984) reveals. Also see K.N. Ciggaar, Western travellers to Constantinople (Leiden, 1996). 48 For which see P.J. Geary, Furta sacra (Princeton, 1990). 49 R. Janin, Géographie Ecclésiastique de l'empire byzantin (Paris, 1969). My thanks to Tony Eastmond for calculating with me.

130

L IZ JA M E S

it is the Virgin herself who, appearing on the ramparts of the city in 626, saves it.50 It is her icon, not her robe, her relic, paraded round the walls which saves her people; it is she, not her robe, which appears on the walls.51 And how was the Virgin recognized? Not by her robe but through her likeness to her images, her icons. Her robe is certainly seen as a palladium, but do palladia and icons do the same thing? One key distinction, perhaps, is that an icon is deliberate whereas a relic does not have foreknowledge of its future status.52 Although it has been said that the rise of icons is concomitant with an increase in the veneration of relics,53 there appears to be a change before and after Iconoclasm in the perception of the relics of saints generally and in the role of the saint's body in Byzantium. The importance of the body of the saint is a feature of early Christian worship. A whole series of fifthcentury stories are told about the bodies of holy men. Rival tombs were constructed for the hermit Marcianus, but the saint asked to be buried in secret so that there could be no squabbling over his remains and no attempts to pass off forged relics.54 In the case of James of Cyrrestica, rivals from a local town and village hovered over the saint's deathbed, waiting to seize his body.55 When the holy man recovered, they were forced to leave empty-handed. Clearly, at this point, the body and posses­ sion of the body mattered and had power. But things changed. Antony Eastmond has raised the question of what happened to the body of Symeon Stylites.56 Sources place the body in both Constantinople and Antioch. After the fall of Antioch to the Persians and then the Arabs in the seventh century, all references to it from there disappear. Nor are there any references to it from Constantinople after the record of the transfer of relics by Leo I recorded in the Life of Daniel the Stylite. The body vanishes from the historical record. Tzimiskes does not retrieve it from Antioch in his re-conquest of 969; Antony of Novgorod, visiting the monastery and tomb of Daniel the Stylite, does not mention the body of the first stylite supposedly buried there. Why does it disappear? Is it a wane in the power of the saint's body compared to a 'rise' of icons? Scholars have perhaps been seduced by medieval western attitudes to relics, and the stress laid

50 Cameron, The cult of the Theotokos', and Cameron, The Virgin's robe'. 51 Though neither icon nor relic is given a role in the sieges of 674-8 and 717-8, as Wortley, 'Iconoclasm and leipsanoclasm', 254, n. 5 also noted. 52 A suggestion made by Hannah Hunt. 53 The suggestion is Wortley's in 'Iconoclasm and Leipsanoclasm'. 54 Theodoret de Cyr: Histoire des moines de Syrie, ed. and trans. P. Canivet and A. LeroyMolingen (Paris, 1977-79), III, 18. 55 Theodoret HR, XXI, 9. 56 A. Eastmond, 'Body vs. column: the cults of St Symeon Stylites' in L. James, ed., Desire and Denial in Byzantium, SPBS 6 (Aldershot, 1999), 87-100.

IM P E R IA L R E L IC H U N T E R S A B R O A D

131

on the incorruptibility of the body and the need for it to be whole at the Second Coming, into seeing this as, almost automatically, still the case in Byzantium.57 The Byzantine view on bodily resurrection seems less sharply defined.58 In this chapter, I have outlined those relics which are directly stated in the sources as acquired by imperial figures. The imperial relics I have looked at here are, in the majority, special cases. They are relics of Christ, the Virgin and Old and New Testament figures, not bodies of later saints. They were kept in the Palace and formed a part of imperial ceremony. They were palladia and were put into places where they could act accord­ ingly. Where relics and their translation can be seen as supporting impe­ rial power, no icon ever functions as a symbol for the state.59 Icons perhaps were intercessors in a way that relics could not be. Reliquaries themselves could become icons, as with the gold repoussé covering on the stone from Christ's tomb kept in the imperial palace, and as is the case with an icon in the Vatican which originally formed a casket lid. And as for the Mandylion, is it icon, relic, or both? Boundaries are blurred, but differences remain. Icons perhaps helped the 'eyes of the spirit' to see more clearly; certainly no saint was ever recognized from their relics, but from their icon. Where once relics were felt to embody the saint, that role seems to have been increasingly taken by the icon. Iconoclasm acted as a period when the nature of religious images became increasingly defined. As the icon came to stand more and more for the holy person depicted, how did the relic fit in? In the end, then, perhaps Eudokia won. She also sent back from Jerusalem to Pulcheria, along with various relics, an icon of the Virgin painted from life by St Luke, from which all icons of the Virgin were drawn. In 1204, the Hodegetria failed to defend her city. Where was her robe when it was needed? Its absence perhaps says much about the changing positions and significance of icons and relics. Which is a long way from the east, but perhaps the empire felt itself increasingly, distanced from its eastern Holy Land. 57 On western concerns, see C. Walker Bynum, 'Material continuity, personal survival and the resurrection of the body: a scholastic discussion in its medieval and modern contexts' in Fragmentation and redemption (New York, 1991), 239-98. 58 John Chrysostom, Homily 7 on I Thessalonians, PG 62, 435^10, discusses mechanical objections to the resurrection of the dead and suggests that the faithful need not worry about this issue and that it is a different body which will be resurrected. See D. Krausmuller, 'God or angels as impersonators of saints. A belief and its contexts in the "Refutation" of Eustratius of Constantinople and in the writings of Anastasius of Sinai', Gouden Hoorn 6 /2 (1998-99), 5-16, which touches on this theme. My thanks to Mary Cunningham for both of these references. Also T. Ware, The orthodox church (London, 1964), 238, 264. 59 See A. Weyl Carr, 'Court culture and cult icons in Middle Byzantine Constantinople' in Maguire, Byzantine court culture, esp. 83-4.

This page has been left blank intentionally

9. Art ch Art chrétien en Anatolie turque: le témoignage de peintures inédites à Tatlarin Catherine Jolivet-Lévy A l'orient de l'empire byzantin, au cœur du sultanat de Rûm, les pein­ tures du XIIIe siècle conservées en Cappadoce témoignent à la fois de la permanence de traditions locales et d'influences ponctuelles du monde turc - ce que l'on savait - mais aussi, comme j'essaierai de le montrer, de relations avec l'empire de Nicée. La village de Tatlarin se trouve à une vingtaine de kilomètres au sud de Gülçehir; plusieurs églises y sont connues, dont certaines ont été récem­ ment restaurées. Lorsque nous avons publié, en 1996,1 un complexe ecclésial situé près de l'entrée de la 'ville souterraine', seules les peintures de l'une des églises avaient été nettoyées. Aujourd'hui, celles de la seconde (ou église B) ont bénéficié d'une restauration, qui a révélé quelques sujets intéressants et, dans l'abside sud, une inscription dédicatoire, donnant le nom de la donatrice, la protopapadias Rodathys, et la date: 6723,3e indiction. Si celle-ci, qui correspond à l'année 1215, ne se rapporte précisément qu'au décor de la niche où elle se trouve, le reste des pein­ tures conservées est à l'évidence contemporain, comme le prouvent les liens - iconographiques et stylistiques - avec un autre décor de la région, daté de 1212, celui de Kar§i kilise, église située à l'entrée de Gulçehir.2 L'église se compose de deux nefs inégales voûtées en berceau, commu­ niquant entre elles, terminée chacune par une abside, au fond de laquelle

1 C. Jolivet-Lévy et N. Lemaigre Demesnil, 'Nouvelles églises à Tatlarin, Cappadoce', Monuments et Mémoires. Fondation E. Piot 75 (1996), 21-63. 2 Décor restauré en 1996, dont la description dans G. de Jerphanion, Une nouvelle province de l'art byzantin. Les églises rupestres de Cappadoce (Paris, 1925-42), t. II, 1-16, est incomplète. Nous en avons repris l'étude: C. Jolivet-Lévy, 'Images et espace cultuel à Byzance: l'exemple d'une église de Cappadoce (Kar§i kilise, 1212)', dans M. Kaplan, éd., Le sacré et son inscription dans l'espace à Byzance du IVe au XIIIe siècle, BS 18 (sous presse). La main d'un même peintre se retrouve dans certaines images de Tatlarin et de Kar§i kilise. From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

133

134

C A T H E R IN E JO L IV E T -L É V Y

une niche abrite l'autel.3 L'abside nord offre des aménagements liturgiques - chancels et niche de prothèse - qui n'apparaissent pas dans l'abside sud, suggérant pour les deux nefs une fonction différente: synaxe eucharistique au nord et, peut-être, offices commémoratifs au sud, où se trouve une tombe, creusée dans le sol en bonne place, à l'entrée du court bras voûté, perpendiculaire à la nef, qui s'ouvre au sud. Le programme iconographique de cette partie de l'église, partiellement conservé - cycle de la Passion,4 Jeunes Hébreux dans la fournaise5 - semble en accord avec cette fonction funéraire supposée. Nous n'analyserons pas ici toutes les peintures conservées, souvent de façon très fragmentaire, dans l'église, mais nous limiterons au décor des absides et à une composition très effacée du mur ouest de la nef nord.6 Arrêtons nous d'abord sur les programmes absidaux. La conque de l'abside principale, au nord, accueille une composition traditionnelle en Cappadoce depuis le XIe siècle: la Déisis (Figure 9.1). De type composite, elle associe au trimorphon des éléments des visions théophaniques. Le Christ en trône tient, posé sur le genou, un livre avec le verset de Jn 8.12: 'Je suis la lumière du monde. Qui me suit n'errera pas dans les ténèbres'. Il est entouré des quatre symboles des évangélistes, que désignaient, comme dans les peintures cappadociennes du Xe siècle, les participes empruntés à l'ekphonèse introduisant le Triple Sanctus de l'Anaphore.7 De part et d'autre sont représentées les puissances célestes (roues de feu, chérubin tétramorphe, séraphin hexaptéryge) et, à petite échelle sur les bords de la conque, les deux intercesseurs, Marie, au nord, et JeanBaptiste, en mélote, au sud. Deux figures agenouillées les précèdent: les prophètes Isaïe (purifié par le charbon ardent) et Ézéchiel, représentés ici, comme dans une série de compositions absidales cappadociennes du Xe siècle, moins comme témoins de la vision divine que pour montrer le pouvoir de purification et de sanctification de l'Eucharistie.8 3 Pour une description précise de l'architecture: Jolivet-Lévy et Lemaigre Demesnil, 'Nouvelles églises', 24-9. 4 La Cène occupe le tympan sud et la Trahison de Judas la voûte (côté est) de la petite chambre annexe, tandis que la Crucifixion sur le mur ouest de la nef, est encadrée par les Myrophores au sépulcre (voûte, côté sud) et l'Anastasis (côté nord). 5 Dans la voûte, côté ouest, de la petite annexe sud; la composition figure également dans le programme de Kar§i kilise, ainsi que la Cène, la Trahison de Judas, les Myrophores et l'Anastasis, ces dernières semblant dans les deux églises de la même main. 6 Signalons une image de donateur en prière auprès d'un saint militaire sur le mur sud, près de l'entrée primitive; d'autres 'portraits', pratiquement effacés, existaient dans le bras sud, dont au fond (mur sud), une grande figure en prière (la protopapadias?). Nous nous réservons de publier ailleurs l'étude de l'ensemble des peintures. 7 Sur ce thème: C. Jolivet-Lévy, Les églises byzantines de Cappadoce. Le programme icono­ graphique de l'abside et de ses abords (Paris, 1991), 338. 8 Cf. Jolivet-Lévy, Les églises byzantines, 339.

A R T C H R É T IE N E N A N A T O L IE T U R Q U E

135

La niche large et profonde, qui abrite l'autel, contient la représentation de la Théotokos avec l'Enfant (Figure 9.2), image de l'Incarnation directe­ ment associée à l'Eucharistie qui en renouvelle le mystère. D'une exécu­ tion pauvre et schématique, elle semble l'œuvre d'un autre artiste, au talent plus modeste, qui a aussi travaillé, en 1212, au décor de Kar§i kilise. La Théotokos trône de face, l'Enfant sur les genoux, encadrée par Anne et Joachim, figurés en tant que témoins du rôle essentiel de Marie comme instrument de l'Incarnation et donc du salut de l'humanité.9 Aux sigles traditionnels désignant la 'Mère de Dieu' s'ajoutent, entre les theopatores et la Théotokos, deux inscriptions tracées verticalement. A gauche se trouve le cryptogramme EÉÉE pour lequel le contexte iconographique suggère une lecture autre que la formule habituelle faisant référence à Hélène et à la Vraie Croix: peut-être 'Ecoocpopog ettegev EÙpr|Ka[iEV 'Ebé[x ('Satan est tombé, nous avons trouvé le Paradis')?10 A droite, l'inscription qualifie la Mère de Dieu de ôeî 7ToXiK..a, soit, peut-être, ô eî TToXuKÀEa ou qeî TioXuKuSà ('toujours très glorieuse').11 Au registre médian de l'abside, encadrant la niche d'autel, sont représentés les quatre évangélistes,12 en train d'écrire, groupés par deux comme dans une série de représentations dans la peinture murale et la miniature.13 Ils sont assis sur de riches trônes à dossier et un lutrin avec un rouleau déployé apparait à l'arrière-plan. Le rôle de témoins de l'Incarnation et de la divinité du Christ des évangélistes justifie leur inser­ tion dans le programme de l'abside, proches à la fois du Christ souverain de la Déisis, qui les surmonte (dans la conque), et de la Théotokos (dans la niche), qu'ils encadrent. En l'absence de coupole pouvant accueillir 9 Leur association à la Théotokos, comme leur type iconographique, sont conformes à une tradition bien attestée à Byzance au moins depuis le XIe siècle; des exemples dans D. Mouriki, 'H TTavayia xai oi O eottcîtopeç, àtpqyrmaTiKri OKf|vfi f| EiKovoaTiKri irapàaTaaq', DChAE 5 (1969), 31-52; D. Mouriki, The Mosaics of Nea Moni on Chios (Athènes, 1985), I, 148-9; E. Kitzinger, The Mosaics of the Church of St. Mary's of the Admirai in Palermo, DOS 27 (Washington, DC, 1990), 136-8; M. Panayotidi, 'Oi ToixoypacpÎEç tou 'Ayiou TECûpyiou AaSpqvou ottî NâÇo', DChAE 16 (1991-92), 148. Pour des exemples en Cappadoce d'Anne et Joachim associés au programme du bêma: Jolivet-Lévy, Les églises byzantines, 61, 81, 124, 132,197,199, 212, 274, 291. 10 G. Babic, 'Les croix à cryptogrammes peintes dans les églises serbes des XIIIe et XIVe siècles', dans S. Dufrenne, éd., Mélanges Ivan Djucev. Études de civilisation (Paris, 1979), 6. 11 Pour TToÂuKÂEris et ttoâukuBtîç, plutôt que cxeî ttoXù KaXâ. 12 Les deux de gauche, figurés âgés, sont sans doute Jean et Matthieu, les deux autres, à droite, Luc et Marc. 13 Citons, en nous limitant à quelques exemples dans la peinture monumentale, les représentations de Karanlik kilise (Gôreme), Myrioképhala (Crète), la Mavriotissa de Castoria - où ils figurent comme à Tatlarin dans l'abside - et Lagoudéra: A. Nicolaides, 'L'église de la Panagia Arakiotissa à Lagoudéra, Chypre: étude iconographique des fresques de 1192', DOP 50 (1996), 53-7, où l'on trouvera la bibliographie sur les portraits des évangélistes.

136

C A T H E R IN E JO L IV E T -L É V Y

l'image du Pantocrator entouré, dans les pendentifs, par les quatre évangélistes, le programme traditionnel a donc été concentré dans l'ab­ side, la voûte en cul-de-four recevant l'image du Christ et la niche d'autel celle de la Vierge. On peut même se demander si la raison d'être de cette niche profonde abritant l'autel, qui ailleurs est souvent simplement accolé à la paroi, n'était pas de ménager un espace propice à la mise en valeur d'un programme absidal en réduction: on aurait ici un nouvel exemple de concertation entre architectes et peintres, montrant de surcroît combien l'image de la Théotokos devait paraître indispensable dans le programme iconographique du sanctuaire. Sous les évangélistes se tenaient quatre évêques frontaux, aujourd'hui très effacés: Grégoire de Nysse et Basile à gauche, Jean Chrysostome (?) et une figure non identifiable à droite. Dans la niche nord14 est représenté le Christ Emmanuel en buste, selon une iconographie bien attestée en Cappadoce au XIIIe siècle, qui est liée à la fonction de prothèse de cette niche: les commentateurs de la liturgie assimilent la prothèse à Bethléem ou à la crèche et interprètent la préparation des oblats comme le symbole de la Première Venue du Christ.15 Le programme iconographique de cette abside nord est complété à l'in­ trados de l'arc absidal par deux figures d'archanges tournés vers l'abside, et, sur la face interne des piédroits, au-dessus des chancels, par deux images frontales de saints médecins, très effacés, probablement Cosme et Damien.16 Dans l'abside sud ne sont bien conservées que les peintures associées à la dédicace de la protopapadias, dans la niche d'autel, tandis que quelques fragments peu distincts suggèrent de restituer une Déisis dans la conque. Comme celle de l'abside nord, la niche d'autel de l'abside sud a reçu un décor centré sur l'image de la Théotokos entre ses parents, mais au sein d'une composition plus savante et exécutée dans un style de bien meilleure qualité (Figure 9.3). Au-dessus d'une bande ornementale et de l'inscription dédicatoire de la protopapadias, trône la Vierge, l'enfant sur le bras gauche; celui-ci déploie un rouleau portant le texte de Jn 8.12: 'Je suis la lumière du monde. Qui me suit n'errera pas dans les ténèbres ...', que

14 Sous la niche prothèse: deux figures à petite échelle, qui semblent non nimbées et pour­ raient être des donateurs. 15 Cf. Jolivet-Lévy, Les églises byzantines, 158; N.B. Teteriatnikov, The Liturgical Planning of Byzantine Churches in Cappadocia, OCA 252 (Rome, 1996), 85-6. En Cappadoce, l'image du Christ Emmanuel est liée à la prothèse à l'Archangélos de Cemil, Tatlarin, Güzelôz no. 4 (Mavrucan), Saint-Eustathe d'Erdemli, Bezirana kilisesi (Jolivet-Lévy, Les églises byzantines, 158, 233, 249, 274, 317), ainsi que dans l'église no. 1 (ou A) de Tatlarin, figure récemment nettoyée et encore inédite. 16 Pour d'autres exemples de cet emplacement oriental des anargyres: Jolivet-Lévy, Les églises byzantines, 79 et passim; voir aussi Jolivet-Lévy, 'Images et espace cultuel'.

ART CHRÉTIEN EN ANATOLIE TURQUE

137

pointe l'index de Marie, soulignant la valeur de ce verset. Au-dessus de la tête de Marie, la 'Main du Seigneur', symbole de Dieu le Père, sortant d'un segment de ciel, bénit. L'inscription tracée sous la main divine cite le Psaume 109.3: ek yacrrpôç Tipô éoacpôpou èyÉv[q]aa oe ('du sein avant l'aurore je t'ai engendré'). De part et d'autre de la Théotokos, Michel et Gabriel en loros impérial tiennent le dossier du trône.17 Joachim (bénis­ sant, le rouleau à la main) et Anne (en maphorion rouge, tenant une croix, l'autre main ouverte sur la poitrine), dont la place est inversée par rapport à l'abside nord mais conforme à l'usage habituel, sont représentés debout de face, de part et d'autre. Cette composition est intéressante à la fois par la densité de son contenu doctrinal et par les liens qu'elle semble révéler avec Nicée. Le verset 3 du Psaume 109, en rappelant la conception du Logos avant le temps, souligne la divinité de l'enfant, co-éternel au Père,18 Logos incarné figuré dans les bras de la Vierge, tandis que la Main de Dieu bénissante transmet au Fils la gloire du Père. Véritable proclamation visuelle du mystère de l'Incarnation, le décor de la protopapadias affirme l'identité du Logos préexistant et du Dieu fait chair, Jésus Christ, conformément au dogme de l'Eglise orthodoxe et à la liturgie, qui cite à plusieurs reprises ce verset le jour de la fête de la Nativité du Christ.19 Malgré des différences évidentes, la composition évoque irrésistiblement la mosaïque absidale, bien antérieure, de l'église de la Dormition de Nicée, seul exemple, à ma connaissance, qui associe la Main de Dieu, l'inscription du Psaume 109.3 et l'image de la Théotokos avec l'enfant.20 L'importance du Psaume 109, Tune des pierres angulaires de la chris­ 17 Comme on le voit dans plusieurs représentations à partir des XIIe—XIIIe siècles, par exemple dans la chapelle dite de la Vierge du monastère de Saint-Jean le Théologien à Patmos ou sur des icônes du Sinai: K. Weitzmann, 'Thirteenth-Century Crusader Icons on Mount Sinai', Studies in the Arts at Sinai (Princeton, 1982), 297-8. Faut-il voir dans cette fami­ liarité des archanges la volonté de mettre l'accent sur la divinité de l'enfant? 18 Sur l'exégèse du Ps. 109: M.-J. Rondeau, 'Le Commentaire des Psaumes de Diodore de Tarse et l'exégèse antique du Psaume 109/110', Revue de l'histoire des religions 176 (1969), 533,153-88,177 (1970) 7-33; C. Mango, The Chalkoprateia Annunciation and the pre-eternal Logos', DChAE 17 (1994), 170. 19 Cf. J. Mateos, Le Typicon de la Grande Église, OCA 165 (Rome, 1962), 152,154,156,160; J. Grosdidier de Matons, éd., Romanos le Mélode, Hymnes, SC 110 (Paris, 1965), 88-9: 'Celui qui, sans mère, fut engendré par le Père avant l'aurore, aujourd'hui sans père, a pris chair en toi sur la terre ... '. 20 Sur la mosaïque absidale de Nicée, voir en dernier lieu Mango, 'The Chalkoprateia Annunciation', 168-70; l'inscription du Psaume 109.3, qui se rattache à la première phase du décor de l'abside (fin VIIe siècle?), fut conservée au-dessus de l'image de la Théotokos (probablement postérieure à 843) jusqu'à la destruction de l'église au début de ce siècle. Rappelons que dans le décor du narthex, postérieur à 1065, Joachim et Anne, les évangélistes, le Christ et Jean Baptiste étaient représentés associés à l'image de la Théotokos: Th. Schmit, Die Koimesis Kirche von Nikaia (Berlin et Leipzig, 1927), 48-56.

138

CATHERINE JOLIVET-LÉVY

tologie du Nouveau Testament, dans les débats théologiques, suggère de mettre ce programme en rapport avec les discussions contemporaines, hypothèse que pourrait confirmer la présence dans l'église de Tatlarin d'une autre image, peut-être aussi inspirée par le Psaume 109 (Figure 9.4). Peinte à grande échelle sur toute la largeur de la paroi ouest de la nef nord, cette seconde composition, très mal conservée, se présente comme un unicum dans l'iconographie orientale. Deux figures du Christ, toutes deux de type Pantocrator, avec nimbe crucifère, partagent le même trône, Tune plus grande, un livre fermé sur le genou, semblant tenir par l'épaule celle plus petite, qui siège à sa droite; deux anges en adoration les enca­ drent. L'image est ambiguë et toute tentative d'interprétation aléatoire. Le dédoublement du Christ peut difficilement être compris - à moins de proposer une lecture hérétique (nestorienne) - comme la visualisation de ses deux natures, humaine et divine, déclarées inséparables par le concile de Chalcédoine. Il traduit plutôt le Logos préexistant, Dieu le Père - sous l'aspect du 'visible du Père', le Christ21 - et le Logos incarné, le Fils, qui siège à sa droite, la différence de taille marquant la subordination du second, le trône commun leur consubstantialité; l'état de conservation n'autorise pas la restitution de la colombe de l'Esprit Saint, qui trans­ formerait cette Majesté binitaire en Trinité, mais elle ne permet pas non plus de l'exclure totalement. La représentation de l'intronisation du Fils à la droite du Père - ou du Logos incarné à la droite du Logos préexistant - se fonde sur plusieurs passages scripturaires, en particulier sur le psaume 109.1: Te Seigneur dit à mon Seigneur: Siège à ma droite, tes ennemis j'en ferai l'escabeau de tes pieds', et l'Épître aux Hébreux 1.3-4: 'ce Fils, qui soutient l'univers par sa parole puissante ... s'est assis à la droite de la majesté dans les hauteurs, devenu d'autant supérieur aux anges que le nom qu'il a reçu en héritage est incomparable au leur'. Ecrits patristiques et traités théologiques soulignent l'importance de cette notion de trône partagé par le Fils et le Père, en particulier dans le contexte de la défense du dogme orthodoxe.22 En Occident, l'image du Père et du Fils trônant côte à côte illustre le Psaume 109.1 dès l'époque carolingienne, pour devenir surtout fréquente à partir du dernier tiers du XIIe siècle;23 dans le Commentaire sur les 21 Plus couramment figuré à Byzance en Ancien des Jours (Daniel 7.7, 22), type icono­ graphique qui souligne l'éternité du Verbe, et non en Pantocrator comme ici. 22 Cf. C. Schönborn, 'Dieu veut rester homme à jamais', Communio IX, I (janv.-fév. 1984), 29-44; C. Markschies, 'Sessio ad Dexteram. Bemerkungen zu einem altchristlichen Bekenntnismotiv in der christologischen Diskussion der altchristlichen Theologen', dans M. Philonenko, éd., Le Trône de Dieu, Wissenschaftliches Untersuchungen zum Neuen Testament 69 (Tübingen, 1993), 252-317; K. Corrigan, Visual Polemics in the Ninth Century Psalters (Cambridge, 1992), 44-5, 80. 23 Sur cette iconographie, voir E.H. Kantorowicz, 'The Quinity of Winchester', Art Bulletin

ART CHRÉTIEN EN ANATOLIE TURQUE

139

Psaumes de Pierre Lombard de la Bibliothèque Sainte-Geneviève de Paris (fin XIIe-début XIIIe siècle),24 les deux Personnes, toutes deux à nimbe crucifère et assises sur le même trône, sont unies par le même geste qu'à Tatlarin (Figure 9.5).25 A partir de la fin du XIIe siècle, la colombe de l'Esprit Saint rejoint la Majesté binitaire, le Psaume 109 devenant alors un emplacement privilégié pour la représentation de la Trinité. En Orient, vers la même époque, l'illustration du psautier du Musée Bénaki (Athènes, Vitr. 34.3.30, fol. 130v) réunit l'Ancien des Jours (dans une mandorle) et le Fils trônant à sa droite, hors de la mandorle,26 mais il s'agit d'une image exceptionnelle et il faut attendre le XIVe siècle pour que se répande l'iconographie de la Trinité synthronoi:27 le Christ Ancien des Jours et le Christ Pantocrator, de même taille, côte à côte sur le même siège et accompagnés de la colombe de l'Esprit Saint.28 Les autres images byzantines, binitaires et trinitaires, qui peuvent être versées au dossier, font encore ressortir la spécificité de la peinture cappadocienne, qu'il s'agisse de la miniature du Paris gr. 923, illustration littérale d'Actes 7.55-56,29 de celle du Prologue de l'évangile de Jean du Paris, gr. 64,30 de la juxtaposition des trois aspects du Logos (Ancien des Jours, Pantocrator,

29 (1947), 73-85; F. Boespflug et Y. Zaluska, 'Le dogme trinitaire et l'essor de son iconogra­ phie en Occident de l'époque carolingienne au IVe Concile du Latran (1215)', CCM 3 7 /3 (1994), 210-20. 24 Ms. 56, fol. 185: Boespflug et Zaluska, 'Le dogme trinitaire', PI. Vb. 25 Sur l'ambiguïté de ce geste: Boespflug, Zaluska, 'Le dogme trinitaire', 220. 26 A. Cutler et A. Weyl Carr, 'The Psalter Benaki 34.3. An unpublished illuminated manu­ script from the family 2400', REB 34 (1976), 298-9, fig. 16. 27 Création byzantine indépendante ou dérivée des images binitaires occidentales? Cf. H. Gerstinger, 'Über Herkunft und Entwicklung der anthropomorphen byzantinisch-slawis­ chen Trinitäts-Darstellungen des sogenannten Synthronoi- und Paternitas- (Otéchestwo) Typus', in Festschrift W. Sas-Zaloziecky (Graz, 1956), 79-85; M.S. Rozycka, 'Observations sur l'image de la Sainte Trinité dans l'art byzantin: l'illustration du Psaume 109.1.', Mélanges d'histoire byzantine offerts à Oktawiusz Jureivicz à l'occasion de son soixante-dixième anniversaire, Byzantina Lodziensia 3 (Lodz, 1998), 201-8. 28 Cf. le psautier serbe de Münich (Bayer. Staatsbibl., slav. 4): H. Belting, éd., Der Serbische Psalter. Faksimile-Ausgabe des Cod. Slav. 4 der Bayerischen Staatsbibliothek München (Wiesbaden, 1978), 236. Attestée également à Mateic au XIVe siècle, cette formule trinitaire synthronoi se rencontre au XVe siècle (Kalenic, Dragalevci: S. Dufrenne, 'Images du décor de la prothèse, REB 26 [1968], 304) et devient surtout fréquente dans la peinture post-byzantine de l'école crétoise (cf. par ex. N. Chatzidakis, Icons of the Cretan School 15th-16th century [Athènes, Musée Benaki, 1983], no. 23; G. Galavaris, The Icon in the Life of the Church. Doctrine, Liturgy, Devotion [Leiden, 1981], 11-12, PI. IV). 29 Fol. 40r, l'Ancien des Jours trône, le Christ est debout à sa droite, tous deux dans la même mandorle: K. Weitzmann, The Miniatures of the Sacra Parallela. Parisinus Graecus 923 (Princeton, 1979), 190, fig. 490. 30 Fol. 158v, le Christ Ancien des fours et le Christ Pantocrator trônent chacun dans une mandorle: F. Boespflug et Y. Zaluska, 'Note sur l'iconographie du Prologue de Jean', Recherches de science religieuse 8 3 /2 (1995), 293-303.

140

CATHERINE JOLIVET-LÉVY

Emmanuel), attestée dans les manuscrits et, à partir du dernier tiers du XIIe siècle, dans la peinture monumentale,31 ou encore des images dites de Paternité, dans lesquelles le Christ (Emmanuel ou Pantocrator) est sur les genoux de l'Ancien des Jours.32 L'image de Tatlarin, caractérisée par le trône commun, la différence d'échelle entre les deux figures christiques et le geste qui les unit, est à l'évidence plus proche des représentations occidentales que de l'iconographie byzantine; la connaissance au XIIIe siècle d'un modèle occidental est envisageable, mais la prise en compte des discussions théologiques contemporaines en Orient33 suggère une interprétation spécifique. Au XIIIe siècle, l'Église orthodoxe est toujours confrontée aux héré­ tiques majeurs, monophysites surtout (Jacobites ou Arméniens), et à l'Islam: la divinité du Christ, le mystère de la Trinité, l'Incarnation et la Rédemption sont au cœur des discussions; en témoignent les traités polémiques - le Trésor de la foi orthodoxe de Nicétas Choniates, rédigé vraisemblablement à Nicée au début du XIIIe siècle, la Réfutation d'un Agarène du moine Barthélémy d'Édesse ou encore la Controverse sur la foi d'Euthyme, écrite dans la région de Mélitène.34 Mais des dissensions exis­ tent aussi au sein même de l'Orthodoxie, portant sur l'union des deux natures du Christ, sur l'égalité des trois personnes de la Trinité et, surtout, sur le sens à donner à la parole du Christ 'Mon Père est plus grand que

31 Par exemple dans le Paris gr. 74, fol. 167 (XIe siècle), aux Saints-Anargyres de Castoria, à Nerezi, etc.; cf. S. Der Nersessian, 'Recherches sur les miniatures du Parisinus Graecus 74', JÔB 21(1972), 109-17; S. Tsuji, The Headpiece Miniatures and Genealogy Pictures in Paris gr. 74', DOP 29 (1975), 165-203; Rozycka, 'Observations sur l'image de la Sainte Trinité', 206-8. 32 Sur ce thème: A. Heimann, 'L'iconographie de la Trinité. I. Une formule byzantine et son développement en Occident', L'art chrétien (oct. 1934), 39-41; Gerstinger, 'Über Herkunft und Entwicklung'; S.A. Papadopoulos, 'Essai d'interprétation du thème iconographique de la Paternité dans l'art byzantin', CahArch 18 (1968), 121-36. Citons aussi ici la représentation des trois figures identiques, sur le même trône, dans les Homélies de Jacques de Kokkinobaphos (Paris gr. 1208; Vatican gr. 1162): A. Heimann, 'L'iconographie de la Trinité. II. Les trois personnes divines sous une même forme humaine', L'art chrétien (nov. 1934), 19-30. 33 Pour les répercussions sur l'iconographie byzantine des débats théologiques de l'époque comnène: G. Babic, 'Les discussions christologiques et le décor des églises byzan­ tines au XIIe siècle', Frühmittelalterliche Studien 2 (1 9 6 8 ) , 3 6 8 - 8 6 ; D. Mouriki, 'Oi t o ix o ypaq>ÎEç t o u TrapEKKÂriaiou Tfjs M ovfjç t o u ’ Ico à v v o u t o u © E O À ôy ou aTqv F i â T i io ', DChAE 1 4 (1 9 8 9 ) , 2 0 5 - 6 3 ; D. Mouriki, 'Al (3i(3AiKai ttpooeikovîoeiç Trjç T T a v a y ia ç e ’iç t ô v T p o û Â Â o v Trjç TTe p i |3âétttou t o u M u o T p â ', ’ApxaioÀoyiKÔv A e à t iô v 2 5 ( 1 9 7 0 ) , M e X étcu 2 1 7 - 5 1 ; A.L. Townsley, 'Eucharistic Doctrine and the Liturgy in late Byzantine Painting', Oriens Christianus 5 8 (1 9 7 4 ), 1 3 8 - 5 3 ; S.E.J. Gerstel, Beholding the Sacred Mysteries. Programs of the Byzantine Sanctuary (Washington, DC, 1 9 9 9 ), 4 4 - 7 . 34 Cf. A. Th. Khoury, Les théologiens byzantins et l'Islam. Textes et auteurs (VlIIe-XIIIe s.) (Louvain et Paris, 1969), 249-58, 259-93, 294-309.

ART CHRÉTIEN EN ANATOLIE TURQUE

141

moi' (Jn 14.28).35 L'interprétation de cette phrase de l'évangile avait déjà opposé orthodoxes et hérétiques des premiers siècles, lors des controver­ ses sur la consubstantialité du Père et du Fils. Le débat reprend au XIIe siècle, importé d'Occident, les théologiens latins professant une infériorité du Fils dans son humanité, position jugée légitime par l'empereur Manuel Comnène, mais récusée par nombre de théologiens grecs. La majorité d'entre eux s'accordent sur le fait que le Christ n'a pas pu parler de sa nature humaine concrète comme telle, car elle partage la puissance du Verbe et reçoit la même adoration que lui: elle est l'égale en gloire du Père. Au synode de 1166, Manuel fait voter une formule de conciliation: en déclarant 'Mon Père est plus grand que moi', le Christ a voulu parler de sa nature créée et concrète, suivant laquelle il a, entre autres, souffert. L'infériorité relative du Fils ne doit s'entendre que du Fils de l'Homme, du Verbe fait chair; elle réside dans sa capacité à assumer forme humaine dans l'Incarnation et à souffrir avec l'humanité. En même temps est proclamée l'égalité en gloire des deux natures du Christ: 'Éternel souvenir à quiconque dit que l'humanité du Christ a été exaltée par son union avec la divinité, qu'elle mérite d'être l'objet de la prière et d'être assise à la droite du Père ... sans mélange des propriétés de chaque nature'.36 Malgré ce concile, les discussions théologiques sur l'infériorité du Christ repren­ nent de plus belle après la mort du patriarche Luc (1169) et un nouveau synode est réuni en 1170; les métropolites de Cappadoce sont présents à ces débats.37 Après la mort de Manuel Comnène (1180), qui avait fait graver l'édit conciliaire sur des dalles de marbre à Sainte-Sophie,38 la discussion se poursuit et elle durera jusqu'à l'époque du patriarche Michel Autoreianos (1208-14), qui fera une ultime et vaine tentative de révision.39 Peut-on interpréter la composition de Tatlarin comme un équivalent visuel de la définition du synode de 1166, qui fut insérée peu après dans le Synodikon? Le Père (par nature non représentable, figuré par le Christ) est plus grand que le Fils - Verbe incarné, maits ce dernier, consubstantiel au Père, mérite de trôner à sa droite et de recevoir la même adoration que

35 Sur cette controverse: Niketas Choniates, 'Dogmatikè Panoplia', PG 140, 201-81; trad. H.J. Magoulias, O City of Byzantium. Annals of Niketas Choniates (Detroit, 1984), 120-21,183; J. Rosenblum, trad.,Jean Kinnamos. Chronique (Paris, 1972), 162-6; J. Gouillard, 'Le Synodikon de l'Orthodoxie. Édition et commentaire', TM 2 (1967), 216-26; S.N. Sakkos, ’O riarrjp p o u p E i'Ç œ v p o u ê o T i'v . A'. Kpirua) K E i p é v o u m i É p p r j v E Î a . B'. ’EpiSEç K a i o û v o S o i Kara r ô v i ( 3 ’ aicbva, 2TTOu8aaxr)piov m i ’EKKXqcrnKrjç VpapuaToÂoyiaç 7-8 (Thessalonique, 1968). 36 Canon 4: Mansi, t. 22, col. 4. 37 Cf. Sakkos, ’O ü a rq p pou peîÇojv pou èotîv. B'., 98-104 (tableaux). 38 Cf. C. Mango, 'The Conciliar Edict of 1166, DOP 17 (1963), 315-30. 39 L. Petit, 'Documents inédits sur le concile de 1166 et ses derniers adversaires', VV 11 (1904), 465-93; en particulier 477 (n. 1) pour la poursuite de la polémique 'vers la fin du patriarcat de Michel'.

142

CATHERINE JOLIVET-LÉVY

lui; la communauté de trône, la présence des deux anges en prière enca­ drant la Majesté binitaire expriment leur égalité d'honneur: l'infériorité relative du Fils est comme 'compensée' par sa consubstantialité. La participation de la Cappadoce, au début du XIIIe siècle, aux querelles théologiques byzantines pourrait expliquer aussi une image infernale assez insolite, conservée à Kar§i kilise (Figure 9.6).40 L'iconographie en est caractérisée par le sort fait aux représentants du clergé -les damnés poussés dans les ténèbres infernales sont des évêques et des prêtres, à l'exclusion de toute autre catégorie sociale - et par le rôle joué par Judas. Derrière un groupe serré de prélats, trois vieillards, dont un grand démon tire la barbe, semblent dans le giron d'une très grande figure rouge, identifiée par une inscription à Judas, qui, la corde au cou, est tiré par le maître de l'Enfer. Bien que la critique des ecclésiastiques soit un topos de l'iconographie du Jugement dernier, leur présence ici exclu­ sive est unique, comme l'est aussi l'image de Judas recevant les évêques dans son sein: pécheur et damné exemplaire, Judas est le modèle par excellence du traître, du cupide et de l'hérétique.41 Dans une région qui fut de tous temps propice aux dissidences religieuses et dans une scène qui est le lieu privilégié de l'expression des réalités contemporaines, les prélats condamnés pourraient bien représenter ceux qui, accusés d'hérésie, furent anathématisés lors des controverses contemporaines. Quelle que soit leur signification précise, les peintures de l'église B de Tatlarin montrent donc que la Cappadoce, au début du XIIIe siècle, loin d'être coupée du monde byzantin, auquel elle n'appartient plus poli­ tiquement, participe aux débats qui déchirent alors l'Orthodoxie et est en relation avec l'empire de Nicée. L'empereur Théodore Lascaris est d'ailleurs mentionné dans la dédicace de Kar§i kilise, dont la donatrice, Irène, s'est fait représenter, avec ses deux filles, à côté de sainte Théodotè, martyre de Nicée, tandis que le saint patron de cette ville, Tryphon,42 figure dans la voûte de la nef. Au maintien de relations avec le monde byzantin, s'ajoutent l'influence hypothétique, mais plausible, de modèles occidentaux, et celle - ponctuelle - de l'art seldjoukide contemporain,43 dont nous n'avons pas eu l'occasion de parler ici. Malgré les problèmes d'interprétation qu'ils suscitent, les monuments de Cappadoce constituent ainsi un témoignage intéressant sur la complexité de la réalité historique et artistique au début du XIIIe siècle en Anatolie. 40 Jolivet-Lévy, 'Images et espace cultuel'; reproduction en couleurs dans: C. Jolivet-Lévy, La Cappadoce, mémoire de Byzance (Paris, 1997), 105. 41 D'où l'intérêt pour l'histoire de Judas dans les psautiers à illustrations marginales: A. Grabar, L'Iconoclasme byzantin (Paris, 19842), 228. 42 C. Foss et J. Tulchin, Nicaea: A Byzantine Capital and its Fraises (Brookline, MA, 1996), 6, 104-8 (Tryphon), 7,115 (Théodotè). 43 Sensible en particulier dans la représentation des dragons que terrassent Georges et Théodore, à Tatlarin comme à Kar§i kilise.

ART CHRETIEN EN ANATOLIE TURQUE

143

Fig. 9.1. Tatlarin, eglise B. Abside nord: la Deisis de la conque

(photo: author)

Fig. 9.2. Tatlarin, eglise B. Abside nord: la Theotokos dans la niche d'autel

(photo: author)

144

CATHERINE JOUVET-LEVY

Fig. 9.3. Tatlarin, eglise B. Abside sud: la Theotokos entre Michel et

Gabriel, Joachim et Anne (photo: author)

Fig . 9.4. Tatlarin, eglise B. Schema de la composition du mur ouest de la

nef nord (drawing: author)

ART CHRETIEN EN ANA TOllE TURQUE

145

Fig. 9.5. Paris, Bibliotheque Sainte-Genevieve, Ms 56 (Commentaire sur les Psaumes de Pierre Lombard), fo!' 185: Binite du Psautier (drawing: author)

Fig. 9.6. Giil§ehir, Kar§l kilise. Tableau infernal du mur ouest (photo:

author)

This page has been left blank intentionally

Section IV Georgians

This page has been left blank intentionally

10. Newly discovered early paintings in the Gareja desert Zaza Skhirtladze The monastic complex known in historical sources as Garejis k'ueqana The land of Gareja' - is spread over almost 120 km12 of the south-eastern corner of Georgia, and comprises more than twenty rock-cut monasteries and hermitages.1 The description of the monastic complex as a 'land' reflects not only its scale, but also its significance for the country. The first monastery, the Lavra, was established in the mid sixth-century by one of the so-called Thirteen Syrian Fathers', St Davit', and his disciples, and was adapted to the environment and severe climate of the semi-desert. Over the centuries other monasteries and hermitages were established in other regions of the desert, and it became one of the largest spiritual centres in Georgia. Gareja was also the site of intense literary activity as well as of an original, local school of wall painting. Close links between the Garejan monasteries and spiritual centres elsewhere in the Byzantine and east Christian worlds can be traced in historical sources and hagiog­ raphy, as well as in epigraphy and art.2 1 KC 2:272ff. 2 G. Chubinashvili, Peshchernye monastyri Davit' Garedji (Tbilisi, 1948); S. Amiranashvili, Istoriia gruzinskoi monumental'noi zhivopisi (Tbilisi, 1957); G. Abramisvili, Davit' garejelis c'ikli k'art'ul kedlis mxatvrobashi (The cycle of Davit' Garejeli in Georgian monumental painting) (Tbilisi, 1972); A. Vol'skaia, Rospisi srednevekovykh trapeznykh Gruzii (Tbilisi, 1974); A. Vol'skaia, 'Garedzhiiskaia zhivopisnaia shkola. Rospisi Bertubani', in Srednevekovoe iskusstvo. Rus. Gruziia (Moscow, 1978), 92-105; A. Vol'skaia, 'Peintures anciennes de Garedja', IV International Symposium on Georgian Art (Tbilisi, 1983), offprint; A. Vol'skaia, 'Rospisi peshchernykh monastyrei Davit'-Garedji', in Gareji, Proceedings of Kakheti Archaeological Expedition 8 (Tbilisi, 1988), 130-55; T. Virsaladze, 'Osnovnye etapy razvitiia gruzinskoi srednevekovoi monumental'noi zhivopisi', II International Symposium on Georgian Art (Tbilisi, 1977), offprint; T. Shevjakova, Monumental'naia zhivopis' rannego srednevekovia Gruzii (Tbilisi, 1983); Z. Skhirtladze, Sabereebis p'resk'uli carcerebi (Fresco inscriptions of Sabereebi) (Tbilisi, 1985); G. Gaprindashvili, Gareji (Tbilisi, 1987); A. Eastmond, Royal imagery in medieval Georgia (University Park, PA, 1998). From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

149

150

ZAZA SKHIRTLADZE

The majority of the monasteries and hermitages in the Gareja desert have been known for a long time, and have been studied by several gener­ ations of scholars and restorers beginning from the early twentieth century. It is thanks to their efforts and work that, despite the establish­ ment in the desert of a Soviet military training ground (polygon), many of the rock-cut chapels and their frescoes have been saved from destruction, while all those which were impossible to safeguard have been thoroughly recorded with measured drawings, photographs and copies. Despite all this work, the sheer scale and diversity of the desert means that much remains to be studied here. This is especially true of monu­ ments which have been discovered only in the last few years. The great majority of the newly recorded complexes is connected with the early stage of the development of monastic life in the rocky desert and provides unique evidence, thanks to their architecture, mural painting and numer­ ous historical inscriptions. In recent years, members of the Gareja Research Centre have been working in the monasteries in the western and extreme eastern peripheral parts of the desert and have prepared complete surveys of the rock-cut monasteries of T'et'ri Udabno, Mravalcqaro, Camebuli, Pirukugmari, Patara Kvabebi, Satorge, and Berebis Seri. Thanks to this, it has become possible to renew the compilation of a systematic corpus of the entire monastic complex a task which was first started fifteen years ago. The primary concern of the surveys is to record the newly discovered monas­ teries in as full a manner as possible. Among recently studied monasteries and their wall paintings, which are connected with the early period of history of the desert, two should be singled out.

T'et'ri Udabno T'et'ri Udabno is the least well known of the rock-cut monasteries in the western part of Gareja.3 The main body of the monastery was cut into the upper portion of the massive rocky projection of the T'et'ri Udabno ridge, approximately fifteen kilometres west of the Lavra of St Davit'. The few caves that are still preserved here are the only remnants of a once larger monastery, the main parts of which were cut into the upper part of the massive projections of the rocky mountain (Figure 10.1). Among the most important of the remaining caves are two aisled churches, distinguished 3 Pr Preliminary notes about this rock-cut complex were published in the newspapers Sakartvelos Respublika (21 December 1996), 4; and Svobodnaia Gruziia (6 January 1997), 4; cf. also Z. Skhirtladze, T'et'ri udabnos eklesiis moxatuloba' (The painting of T'et'ri Udabno church), XXXII Scientific Conference of the Institute of the History of Georgian Art, Abstracts (Tbilisi, 1997), 22-3.

EARLY PAINTINGS IN THE GAREJA DESERT

151

by their archaic architectural forms, which are located about thirty metres apart. A church with a northern annex cut into the right side of the rock was the principal of the rock-cut structures of the monastery. Almost nothing is now left of this church, apart from the chancel, which is partially filled by a landslide. However, it is possible to reconstruct much of the architectural appearance of the church from this (Figure 10.2). A lancet-like chancel apse with a narrow and disproportionately high triumphal arch and a distinct horseshoe plan indicates that the inner space of the church had a distinct vertical emphasis. It is hard to state for sure whether the church was domed, since nothing tangible is left of the naos. The only indications of the size and shape of the naos are provided by the dimensions of the northern annex, the length of which indicates that the church must have had a relatively large naos, elongated along its east-west axis. However, this does not preclude the possibility of the exis­ tence of a dome. The chancel of the church was completely plastered and painted. The rock, a granular sandstone, is very soft in this part of the T'et'ri Udabno hills, and it disintegrates at a touch. Not long before its discov­ ery, a large mass of rock had fallen down from the upper part of the ridge, and buried the south-west part of the northern annex. In due course, the church itself shared the same fate: all that is now left is the chancel, pierced by a large vertical crack and on the edge of a precipice. The small part of the church that was preserved was open to the elements, and because of this the painting there has gradually been washed away. It deteriorated until it reached a state where the plaster could fall down at any moment. Without emergency measures it was inevitably destined to destruction. For this reason it was decided to remove the fresco from the chancel and transfer it to Tbilisi. In 1997, members of the Gareja Studies Centre, together with English and French colleagues, established a project which, in addition to the complete recording of the lesser known monas­ teries in the desert, also envisaged the conservation of the most damaged wall paintings. With financial support from INTAS, it became possible to undertake emergency restoration work to save the T'et'ri Udabno church apse decoration.4 In December 1998, a group of restorers under Merab Buch'ukuri removed the apse fresco. In the course of restoration, it became possible to reconstruct the general scheme of the chancel decora­ tion, and a number of important iconographic features were noted. The conch of the chancel in T'et'ri Udabno bears a large radiant cross inscribed into an oval seven-coloured mandorla (Figure 10.3). The central part of the mandorla is white, followed by two wide bands, the first brownish-red, the second grey. The next bands are narrower: dark grey, 4 INTAS Project 96-0022 (details at http://www.intas.be)

152

ZAZA SKHIRTLADZE

cherry red and then white. The outer band is a turquoise-green strip. The upper part of mandorla is almond shaped; the cross is brown with a large circular turquoise jewel in the middle; and the distinct outlines of a large green jewel are also discernible at the ends of its upper arm. After clean­ ing here the image of a tablet was revealed with a three-line inscription in asomtavruli (old Georgian uncial script): IILQ RC"kCJ'rITirlrb1purplede TQvhlCChCJ) - 'Jesus of Nazareth king of the Jews'. Four purple rays of light radiate from the central crossing of the cross. The upper part of the mandorla is represented against the background of the starry sky, painted with lapis lazuli. The stars are circular, but their backgrounds have been completely lost over the course of time. Within the circles, rays of light are depicted in the form of white quadrifolates. In total, seven stars are depicted to the right of the mandorla; it can be assumed that the same number would have been depicted in the left part of the sky. On either side of the mandorla, in the lower half of the conch, palm trees with outstretched branches are depicted against the background of a wide 'earth' strip, which has now acquired a light ochre colour. Below the conch, the walls were covered by a thin layer of white rock sediment, brought down by the flow of rain over its surface. Before the removal of the fresco fragments, the restorers cleaned the whole area of the chancel, in the course of which a large, icon-like composition of the Presentation in the Temple was revealed immediately below the cross and mandorla. It comprises three figures: to the left is depicted Simeon holding the child Jesus, while the Virgin is represented to the right (Figure 10.4). Images similar to the T'et'ri Udabno conch composition are known in early monuments of the Christian east. In connection with Gareja, it is noteworthy that, alongside other provinces of the Christian east, this tradition is most frequently seen on monuments in Syria, where numer­ ous examples of embellishing the conch with relief and painted or mosaic images of the cross are known.5 5 In the churches of Arnas, Qefr Zeh, Korkaya, Mar Sovo, El Adha stands an equal-armed relief cross without mandorla or any additional details: G. Bell, 'Churches and Monasteries of the Tur Abdin', in M. van Berchem and J. Strzygowski, eds, Amida (Heidelberg, 1910), 250-56, pi. V, fig. 1; G. Bell, Churches and Monasteries of the Tur Ahdin, intro. M. Mango (London, 1982), pi. 99-101,121,158; J. Lassus, Sanctuaires chrétiens de Syrie (Paris, 1947), 301; Ch. Ihm, Die Programme der christlichen Apsismalerei vom vierten Jahrhundert bis zur mitte des achten Jahrhunderts (Wiesbaden, 1960), 76-7, 210, kat. LVI-LVIII; M. Mundell, 'Monophysite Church Decoration', in A. Bryer and J. Herrin, eds, Iconoclasm (Birmingham, 1977), 65,66, fig. 9-10); in the fifth-century conch mosaic of the south-east chapel of Resafa cathedral the lumi­ nous cross was inscribed into a wide, circular, crown-like mandorla, while the rest of the conch area was decorated with leafy scrolls (Lassus, Sanctuaires, 299-300, fig. 109; J. Kollwitz, 'Die Grabungen in Rusafa', in Neue deutsche Ausgrabungen in Mittelmeergebiet und im vorderen Orient (Berlin, 1959), 65; Ihm, Programme, 78, 210, kat. LIX, Abb. 14; Mundell,

EARLY PAINTINGS IN THE GAREJA DESERT

153

The representation of the cosmic cross surrounded by the luminous glory was not only established in the east Christian church over a long period, but seems to have been marked with special significance. The geography of this ancient tradition can be mapped throughout Cappadocia,6 Thrace7 and Coptic Egypt.8 Local traditions in Georgia provide many manifestations of a special attitude towards the cross. It is recorded from the very first days of the conversion of the country in the fourth century.9 As a result, it has always seemed strange that this has never been reflected in the history of apse decoration in early medieval Georgia: up to now no apse compositions crowned by the cross have been found in Georgia. The discovery at T'et'ri Udabno of an apse programme centred on the cross is therefore of great importance. Indeed, the general picture of the veneration of the cross in Georgia seems more complete taking it into account. At the same time, this newly discovered fresco may be examined within a far vaster context. Christian artistic tradition links the image of the cross on the starry sky 'Monophysite church decoration', 67-8, where the church is dated to the sixth-seventh centuries; cf. also Th. Ulbert, 'Die Basilika des heiligen Kreuzes in Resafa-Sergiopolis', in Resafa II (Mainz am Rhein, 1986), 85-7,132-3, Abb. 52, Taf. 33! - this reconstruction differs to a certain extent from that presented by Lassus). The same solution is found on represen­ tation of the cross on the eastern part of the vault in the church of the monastery of Mar Gabriel near Kartmin (Ihm, Programme, 78, 209, kat. LV; E.J.W. Hawkins and M. Mundell, 'The mosaics of the monastery of Mar Samuel, Mar Simeon and Mar Gabriel near Kartmin', DOP 27 (1973), 285-7, fig. 7,18-23; Mundell, 'Monophysite church decoration', 66). 6 Cf. for example, Chapel 5 in Güllü dere and the church of Niketas the Stylite in Çavu§in, Hagios Stephanos in Cemil, Mezarlar alti kilise and Karshibechak in Avcilar, Kapili vadisi kilisesi in Karacaoren, Ipral dere near Ürgüp (M. Restle, Die Byzantinische Wandmalerei in Kleinasien (Recklinghausen, 1967), I, 142; III, figs. 342-3; G. P. Schiemenz, 'Die Kapelle des Stylites Niketas in den Weinbergen von Ortahisar', /ÔB 18 (1969), 239-58; N. Thierry, 'L'église peinte de Nicétas, stylite, et Eustrate, clisurarque ou fils de clisurarque', XIV Congrès international des études byzantines 3 (Bucharest, 1976), 451-5; idem, 'Matériaux nouveaux en Cappadoce', Byz 54 (1984), 318-20, 351-3; idem, Haut Moyen Age en Cappadoce: les églises de la région de Çavu$in I (Paris, 1983), 1-33,182-9; C. Jolivet-Lévy, 'Peintures byzan­ tines inédites de Cappadoce', Archéologia 229 (1987), 40^13; eadem, Les églises byzantines de Cappadoce. Le programme iconographique de l'abside et des ses abords (Paris, 1991), 44-6, 53-6, 70-72, 7 5 -6,161-3,171-3, 209, pl. 4, 36v 53v 65v 100, 101,104!, 129). 7 S. Eyice and N. Thierry, 'Monastère et source Sainte de Midye en Trace Turque', CahArch 20 (1970), 56, fig. 12; J. Lafontaine-Dosogne, 'Pour une problématique de la peinture d'église byzantine a l'époque iconoclaste', DOP 41 (1987), 324. 8 K. Wessel, Koptische Kunst. Die Spatantike in Agypten (Recklinghausen, 1963), 175. 9 G. Chubinashvili, Pamiatniki tipa Dzhvari (Tbilisi, 1948), 3-5, 18-25; N. Aladashvili, Monumental'naia skulptura Gruzii (Moscow, 1976); N. Chubinasvili, Khandisi (Tbilisi, 1976), 9-10; J. Lafontaine-Dosogne, 'Recherches sur les programmes décoratifs des églises médié­ vales en Géorgie en relation avec la peinture monumentale Byzantine', II International Symposium on Georgian Art (Tbilisi, 1977), offprint, 6; N. and M. Thierry, 'Peintures du Xe siècle en Géorgie méridionale et leurs rapports avec la peinture byzantine d'Asie Mineure', CahArch 25 (1975), 94; E. Privalova, Rospis Timotesubani (Tbilisi, 1980) 15-16,159,162.

154

ZAZA SKHIRTLADZE

in the first instance with the dome, since it was perceived as an Arch of Heaven, an image of the cosmos, eternally marked by the cross. However, two opposing viewpoints exist concerning the interpretation of apse compositions crowned by the cross. The first of these emphasizes the eschatological essence of such programmes, based on the idea of the Second Coming, and indicative of the Last Judgement. The second links these programmes with the idea of eternal triumph. According to this, an apse composition crowned by the cross is a manifestation of the heavenly, eternal glory of the Saviour, which was obtained through his passion on the cross of Calvary. Both of these interpretations were greatly influenced by the rise of the veneration of the True Cross, which goes back to the epoch of Constantine the Great.10 The picture seems to be more complicated in case of the conch compo­ sitions, since in these cases the cross is included within a larger iconographic programme. There are a number of noteworthy parallels to the wall paintings of T'et'ri Udabno, in which the radiant cross, inscribed in a mandorla, is represented within Paradise and surrounded by various symbolic images, as, for example, at Nola, Fundi and San Apollinare in Classe.11 The iconographic programme of the apse painting at T'et'ri Udabno seems to support the second interpretation of this type of imagery. Naturally, based on just this one image, it would be hard to answer all the questions raised by the early apse programmes containing the cross image. Nevertheless, the fact that the cross at T'et'ri Udabno is specifically identified as the cross of Calvary is of great significance. As a result, the argument based on the priority of the triumphal aspect of similar programmes acquires additional support. The juxtaposition of the image of the Eternal Glory of the Cross and that of the Presentation in the Temple, the latter associated with the ideas of sacrifice and salvation, is not unusual; far more so is the depiction of these two compositions together in the apse. A further striking feature is the iconography of the Presentation, showing Simeon with the Christ Child in

10 This question is discussed at greater length, with full bibliography, in Z. Skhirtladze, "Under the Sign of the Triumph of Holy Cross: The original decoration of Telovani Church and its iconographic programme', CahArch 47 (1999), 101-18. 11 Nola: PL 61,3369; C. Davis-Weyer, Early Medieval Art (300-1150). Sources and Documents (New Jersey, 1971), 22; cf. F. Wickhoff, 'Das Apsismosaik in der Basilica des H. Felix zu Nola. Versuch einer Restauration', Römische Quartalschrift 3 (1889), 158-76; Ihm, Die Programme, 80, 179-81, cat. XXXVI; J. Engemann, 'Zu den Apsis-tituli des Paulinus von Nola', Jahrbuch ß r Antike und Christentum 17 (1974), 30-31. Fundi: PL 61, 330; Davis-Weyer, Early Medieval Art, 23; Engemann, 'Zu den Apsis-tituli', 22, 26-9, fig. 5; Ihm, Die Programme, 80-81, cat. XXXVII, fig. 17. San Apollinare in Classe: F.W. Deichmann, Ravenna. Hauptstadt des spätantiken Abendlandes 1 (Wiesbaden, 1969), 181ff.

EARLY PAINTINGS IN THE GAREJA DESERT

155

his hands. This iconography is generally considered to be connected with the post-iconoclastic period.12 However, the chronology of the image at T'et'ri Udabno can be narrowed down by the existence of graffiti scratched on the fresco, which can have been executed no later than the ninth century. This raises new issues, which must remain the subject of separate discussion. The immediate sources for these newly discovered paintings in Gareja is unclear, and the exact date of their execution cannot easily be determined. The image of the victorious cross inscribed in the luminous glory and represented within paradise finds its place in early chronology, and reflects far older artistic tendencies. An interest in archaic traditions has long been recognized as a distinctive feature in Gareja, and so it is possible that this painting could belong to a later period, although it is more reasonable to connect the creation of the apse composition with initial stage of existence of the monastic complex itself. Thus the date of execution of the murals can presumably be ascribed to the first centuries of monastic life in Gareja - probably to the seventh-eighth centuries. Finally, it should be pointed out that although it is difficult to deter­ mine whether both parts of the T'et'ri Udabno apse decoration were created simultaneously, it may be assumed that they were. While it cannot be excluded that the two parts were painted at separate times, it is quite clear that in terms of the general iconographic scheme and theolog­ ical concept, as well as the artistic rendering, the T'et'ri Udabno apse paintings were conceived and perceived as one indivisible entity.

Mravalcqaro Mravalcqaro is the extreme western monastery of Gareja.13 The rock-cut complex, situated twenty-five kilometres to the west of the Lavra of St Davit' is mainly limited to the ninth-tenth centuries. It was at this period that local monks-masons carved out a number of aisled and domed churches at the complex, as well as several cells. Among the two aisled churches at the monastery (both cut no later than the second half of the ninth century), one is large with an additional northern compartment and southern porch. Its prototype is to be sought in the original architectural form of the oldest church in Gareja: the church of the Transfiguration in the Lavra of St Davit'. Far more original is a minute church, located at the eastern end of the cave sequence. The apse here is characterized by a 12 H. Maguire, The Iconography of the Symeon with the Child in Byzantine Art', DOP 34-5 (1981), 261-6. 13 Cf. L. Mirianashvili and T. Djodjua's brief reports in Dzeglis Megobari, no. 1 (1996), 42-3 and no. 3 (1996), 41-3.

156

ZAZA SKHIRTLADZE

distinct horseshoe-shaped form, and the naos is articulated by unusually deep bays (Figure 10.5). Domed churches, discovered nearby and still half-filled with earth, are also distinguished by their original design. A small church in the western part of the monastery, with a deep apse and a square naos articulated by arched bays and a conical dome, belongs to a group of cruciform-domed churches, which can be found in various of the early medieval monaster­ ies of Gareja, notably Sabereebi,14 and Dodo.15 The second domed church, located a little to the east of the first, was carved with unique ingenuity to exploit the non-load-bearing nature of rock architecture: here, the dome, which stands over a large naos with high proportions, is formed of three superimposed cylinders, the diameter and height of each of which decrease upwards. Among the rock-cut structures of the complex, only the small aisled church bears any paintings. The fresco decoration, which is now consid­ erably damaged, contains a series of images which were already estab­ lished early on in the history of Garejan painting: Christ in Majesty, a votive, icon-like image of the Virgin and the Child, a stylite saint, holy women, and donors (Figures 10.6-10.8). The murals were executed simul­ taneously with the excavation of the church, in the mid-ninth century. The fact that the church is dedicated to the Virgin is testified to not only by the large image of her placed in a prominent location, opposite the entrance (Figure 10.7), but also by the painted and scratched inscriptions containing appeals to her. The image of the Virgin was beginning to be established in the local school of mural painting from the ninth-tenth centuries; however, at that time she was depicted in the apses of lateral chapels. This image, represented as a kind of votive 'icon', acquired a significance similar to that of an apse image: it was specially decorated with a plant foliage border along the upper, semi-circular part of the bay, which forms a frame around the image (Figure 10.6). The segment of sky with the outstretched blessing right hand of God, which is depicted on the bay vault, acts as an accent crowning the whole composition. Of the two donor figures represented in the church, one, larger figure is depicted frontally at the north end of the west wall of the naos, facing the apse with his hands raised in prayer (Figure 10.8). Only fragments are left of a long explanatory inscription. Vague outlines of the second donor figure are discernible on the north wall of the western bay of the church. This figure's prayers are directed to the west wall, where a large image of a saint was represented.

14 G. Gaprindashvili, 'Skal'nye peshchernye khramy i ikh znachenie v razvitii monumental'noi arkhitektury Gruzii', Peshchery Gruzii 12 (1988), 10-12, fig. 21-6. 15 G. Chubinashvili, Peshchernye monastyri, pi. 126.

EARLY PAINTINGS IN THE GAREJA DESERT

157

Among nearly sixty explanatory, dedicatory and memorial inscriptions, painted and scratched, one is of special importance: this is a fragmentary nine-line inscription by the painter, written in asomtavruli. The inscription is located in the lower part of the composition of the Theophany, on the north wall of the chancel (Figure 10.9). Its location and content place it in a group of such dedicatory inscriptions in Gareja, all of which date from this period, which other painters also placed in the chancel: comparable cases are found in two murals in the Sabereebi cave complex. The minute size of the church determined the form of its decoration. The painterly modelling used in the treatment of the faces, and the complicated system of the vestment design, both reveal the hand of a trained master, who modified classical prototypes of early date in accor­ dance with the artistic tendencies established in Georgian art at that period. The affinity of the frescoes with the murals of Sabereebi church 6 is evident; it is possible that both were executed by the same workshop or artists.16

The study of the monasteries in the peripheral parts of the Gareja desert has considerably broadened our knowledge on the geographical extent and character of monastic life there. It is obvious that already in the early stage of its history, rock-cut monasteries were spread on both sides of the Lavra of St Davit', over a vast territory both to the north-east (Sabereebi, Pirukugmari, Kvabebi) and to the south-west (T'et'ri Udabno, Mravalcqaro, Camebuli). The establishment of cave complexes in the western part of Gareja, similar to the case of the other major monasteries, goes back to the early Middle Ages. Remnants of the first centuries of monastic life in the main monasteries of Gareja are preserved in far smaller quantities compared to other epochs. One of the main reasons for this is the history of continuous inhabitation and major building activity and alterations taking place in the principal monasteries established by St Davit' and his disciples (the Lavra, St Dodo, and Nat'lismc'emeli [St John the Baptist]). Equally, the ascetic nature of the lifestyle of St Davit' and his disciples suggests that the scale of monastic building cannot have been large in the early centuries. However, unlike the principal monasteries, those in peripheral areas of the desert have preserved almost unaltered seventh- to tenthcentury structures, paintings and inscriptions, as they were never enlarged or modified in later centuries, and were often abandoned. The character and scale of monastic life changed considerably in later centuries. The ninth-eleventh centuries represents a new period of efflo­ 16 Sheviakova, Monumental'naia zhivopis', 11, pi. 32, 33; A. Volskaja, Rospisi, 134-5.

158

ZAZA SKHIRTLADZE

rescence in Gareja. Indeed, it was at this time that new complexes were founded, old churches and cells renewed and enlarged, and many of them now decorated with murals. These changes taking place in the desert coincided with the most significant stage in the history of Georgia, the unification of the country with its consequent political and cultural developments. Many of these tendencies may be connected with the activity of the ninth-century monk, St Hilarion K'art'veli. According to the notes preserved in various versions of the Life of the saint, Hilarion began his activity in the Gareja desert. Later, after having travelled to Palestine, where he stayed in the Lavra of St Saba for several years, in the mid-ninth century (c. 850-55) he returned to Gareja and undertook significant work for the enlargement and transformation of monastic life there.17 Beginning from this period, earlier ascetic traditions were replaced by new principles of monastic organization; relations with various monastic centres of the Christian East (Antioch, Sinai and Cappadocia) became both broader and more intense; and the number of pilgrims visiting Gareja greatly increased. These changes principally affected the Lavra of St Davit' and the monasteries of the central region of the desert (St Dodo, Udabno, Nat'lismc'emeli), and only to a lesser extent the monasteries in the periph­ eral western and especially eastern parts. This is discernible in several monasteries in Gareja, most clearly in Mravalcqaro, Sabereebi, Pirukugmari and the complexes at Berebis Seri. The changes introduced were diverse and considerable: they affected the organization of the monasteries and the churches themselves, and the architectural rendering of subsidiary and dwelling structures. From this time onwards, it is possi­ ble to trace a well-established architectural physiognomy of the churches, cells, refectories, burials and subsidiary structures, which perfectly fit the surrounding environment. All these elements correlate to form a homo­ geneous planning system for the monasteries. Changes are especially discernible in church architecture: alongside aisled chapels, traditional for Gareja from the very beginning, cruciform-domed-type churches become more common and appear in numerous variations. Finally, the wide dissemination of the practice of painting churches and caves indicates the changes taking place in the life of the desert. Similar to other monuments of early date in Byzantium, interiors are still only partially decorated, but now the paintings are located in different areas of the church: the chancel, dome and northern bay. This, in its turn, is indica­

17 Abuladze, Monuments, 2:9-37; 3:208^18; 4:134-54,356-8; M. Dolakidze, Ilarion k'art'velis c'xovrebis redakciebi (Versions of the Life of Hilarion K'art'veli) (Tbilisi, 1974), 37-57, 68-77 (with earlier bibliography); B. Martin-Hisard, 'La pérégrination du moine Géorgien Hillarion au IXe siècle', BK 39 (1981), 101-38.

EARLY PAINTINGS IN THE GAREJA DESERT

159

tive of the gradual process of development of the complete system of decoration. Alongside the diversity of the artistic rendering of murals, the selection of the iconographic themes, which were firmly established in the local school of painting in the course of the centuries, is distinguished by its stability. Thus, in the ninth- to tenth-century murals of four of the churches in Sabereebi, the selection of the iconographic themes is actually homogeneous: the chancel bears diverse versions of the Maiestas Domini composition, which is based on the two-tiered scheme of the Ascension. The northern wall usually bears a heraldic composition of warrior saints or the Crucifixion. The iconography of the latter reveals abundant links with prototypes originating from the artistic traditions of the Holy Land. Thus, in church 7, the composition is considerably enlarged, and comprises - in addition to the figures of the crucified Christ, thieves and Roman centurions - a symbolic image of Mater Ecclesia, the representation of the Temple of Jerusalem, personifications of the Sun and Moon, figures of the women of Jerusalem, the Virgin and St John, and the episode of the dividing of Christ's robes arranged around the whole perimeter of the upper part of northern cross-arm.18 Stylistically a number of different trends can be seen in the early paint­ ings of Gareja. One group of frescoes (for example in Dodo monastery)19 reveals the use of prototypes sharing Hellenistic traditions. At the same time, in accordance with the general trend of Georgian art of the eighth-ninth centuries, these prototypes have been transformed by local craftsmen through linear expressiveness, which, beginning from this time, became characteristic throughout the country (Figure 10.10). The difference can be better seen in the frescoes of two churches at Sabereebi, nos. 7 and 8, which were executed within a few years of each other during the tenth century. The first of the two lavishly reflects links with the Hellenistic heritage (Figure 10.11), while the second shows a distinct striving towards a more ornamental, decorative treatment of form: forms created by means of homogeneous colour spots outlined by a contour line indicating the vestment folds, all reminiscent of cloisonné enamel art (Figure 10.12). The colouristic arrangements play a significant part in the artistic rendering of early wall paintings in Gareja: a palette formed of ochre with reddish-brown accents is characteristic of the frescoes in Sabereebi church 5 and a small, aisled church in Camebuli monastery (Figure 10.13); in the murals of the domed church in St Dodo monastery and in Sabereebi 18 Z. Skhirtladze, "Early Medieval Georgian monumental painting: Establishment of the system of church decoration', OrChr 81 (1997), 190, fig. 14; cf. T. Velmans, 'Une variante orig­ in a l de la Crucifixion de type palestinien', in C. Moss and K. Kiefer, eds, Byzantine East, Latin West, Art-historical studies in honor of Kurt Weitzmann (Princeton, 1995), 309-15. 19 Sheviakova, Monumental'naia zhivopis', 14-16, pi. 48-50.

160

ZAZA SKHIRTLADZE

church 7 preference is given to dark blue, lapis lazuli, cinnabar and orange colours; and in the decoration of Sabereebi church 8, the eye is drawn by distinct turquoise and purple colour spots, as well as intense dark yellow tints. Beginning from the ninth century, as Georgia began the gradual process of liberation from invaders and the formation of a united kingdom, a tendency towards displaying the local aspects of Christian culture can be discerned. Based on early literary sources and ecclesiastical legends, Lives of local saints were created, followed by the formation of their iconography and life cycles. All these are reflected in the Garejan murals. Life cycles and separate images of St Davit' Garejeli naturally occupy the leading place here. However, the local school also shows a special interest in the iconography of St Nino, apostle of Georgia: for example, in the northern annex of the main church of Udabno monastery, scenes of the life of St Davit' Garejeli were replaced by four episodes of St Nino's Life, when the second layer of painting was executed in the twelfth-thirteenth century.20 Evidence shows that the scale of monastic building activity decreased considerably in the peripheral rock-cut complexes of Gareja beginning from the eleventh century; and after the second half of the thirteenth century life gradually ceased here and caves were abandoned. This could have been caused by natural disasters (earthquakes), but was certainly exacerbated by the invasions of the Seljuqs from the second half of the eleventh century, during which the land near the river Iori and RustaviGardabani were turned into the winter shelter of the invaders' army.21 After this, the development of monastic life in this part of the desert seems to have been impossible. For the most part only the large monas­ teries of Gareja - the Lavra of St Davit', St John the Baptist, St Dodo, Bertubani, Udabno - still continued their existence. Beginning from the eleventh century the transition of the monastic complex into the hands of the royal, Bagrationi court changed its monas­ tic organization as well as the character of the architecture and art of the monasteries. From this period onwards, a new stage in the history of Gareja desert had begun.

20 Amiranashvili, Istoriia, 52-3. 21 C'xorebay mep'et'-mep'isa davit'isi (Life of Davit', king of kings) KC 1:320-22; trans. R.W. Thomson, ed., Rewriting Caucasian History. The Medieval Armenian Adaptation of the Georgian Chronicles. The Original Georgian Texts and the Armenian Adaptation (Oxford, 1996), 310-14.

EARLY PAINTINGS IN THE GAREJA DESERT

161

Fig. 10.1. Gareja desert, T'et'ri Udabno. View of cave complex from south

(photo: author)

162

ZAZA SKHIRTLADZE

Fig. 10.2. Gareja desert, T'et'ri Udabno. Principal church: plan and sections (drawing: N . Bathtadze)

Fig. 10.3. Gareja desert, T'et'ri Udabno. Apse: Cross in mandorla (photo: author)

EARLY PAINTINGS IN THE GAREJA DESERT

Fig. 10.4. Gareja desert, Tet'ri Udabno. Below apse: Virgin from

Presentation in the Temple (photo: author)

Fig. 10.5. Gareja desert, Mravalcqaro. Small aisled church: plan and

sections (drawing: M. Kiknadze)

163

164

ZAZA SKHIRTLADZE

Fig. 10.6. Gareja desert, MravaIcqaro. Schema of painting in north bay

(drawing: author)

Fig. 10.7. Gareja desert, MravaIcqaro. North bay: Enthroned Virgin and Child

(photo: author)

EARLY PAINTINGS IN THE GAREJA DESERT

Fig. lD.8. Gareja desert, Mrava1cqaro. North-west corner of naos: sty lite and donor (drawing: author)

Fig . lD.9. Gareja desert, Mrava1cqaro. Inscription of painter in sanctuary (drawing: author)

Fig . lD.lD. Gareja desert, St Dodo. Small domed church. Apse: archangel and tetramorph (photo: author)

165

166

ZAZA SKHIRTLADZE

Fig. 10.11. Gareja desert, Sabereebi, Chapel 7. Apse: detail (photo: author)

Fig. 10.12. Gareja desert, Sabereebi, Chapel 8. Apse: detail (photo: author)

Fig. 10.13. Gareja desert, Camebuli. Small aisled church: detail of Deesis in sanctuary (photo: author)

This page has been left blank intentionally

11. Byzantium and its eastern barbarians: the cult of saints in Svanet'i Brigitta Schrade Svanet'i (in Georgian) or Shvan (in Svan),1 which is situated in the northwest of Georgia, between the Central Caucasus range and the Colchian Plain, is an outstanding cultural landscape. The mountainous region in the valleys of the Enguri (Upper Svanet'i) and C'xeniscqali (Lower Svanet'i) appears rather isolated today. But this was not so in antiquity, when Svans controlled important pass-roads between the northern steppe and the Black Sea as 'rebellious barbarians' evoking the interest of Byzantine diplomacy during the Persian-Byzantine wars of the sixth century.2 Throughout the Middle Ages, Svanet'i maintained wideranging relations with the Georgian lowlands and the neighbouring mountainous regions, at a time when it was a more or less independent political unit among the Georgian kingdoms and principalities.3 In the course of its long Christian history,4 Svanet'i has become a kind of regional treasury as it could participate in cultural development but was geographically much safer than the Georgian lowlands with their endless raids and occupation. An incredible wealth of Christian art has survived in the more inaccessible Upper Svanet'i, mostly from the tenth

1 C. Gudjedjiani, L. Palmaitis and B.G. Hewitt, eds, Svan-English Dictionary (Delmar, NY, 1985), 272. Svan belongs with Georgian, Mingrelian and Laz in the South Caucasian (K'art'velian) language group: G. Deeters, G.R. Solta and V. Inglisian, 'Armenisch und Kaukasische Sprachen', Handbuch der Orientalistik 1, 7 (Leiden and Cologne, 1963), 4-8. Svan has no written form. 2 D. Braund, Georgia in Antiquity (Oxford, 1994), esp. 46-7, 290-91, 311-14. 3 Svan history in outlined in W.E.D. Allen, A History of the Georgian People (London, 1932); C. Toumanoff, Studies in Christian Caucasian History (Georgetown, 1963); D.M. Lang, The Georgians (Bristol, 1966), K. Salia, History of the Georgian Nation (Paris, 1983). A history of Svanet'i is to appear in B. Schrade, ed., The Art of Svanet'i (London, forthcoming). 4 In the sixth century, Svans are numbered by Procopios, Wars IV, eds J. Haury and G. Wirth (Leipzig, 1962/3), 2, 21-33, among the established Christians of the Caucasus. From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

169

170

BRIGITTA SCHRADE

to fourteenth centuries, including churches, wall paintings, liturgical objects, manuscripts, and about 600 painted and chased icons and crosses. The majority of the art is by local masters but also includes central Georgian and Byzantine works which have been dedicated to Svan churches or brought here for protection.5 They reflect not only the various relations and the social organization of Svanet'i during medieval period but reveal also a clearly defined local cult of saints.6 The most striking cult in Svanet'i is the cult of warrior saints, especially of Saint George. At the same time, we find a major cult of the martyr Quiricus and his mother Julitta, as well as of female saints, principally Barbara, and of archangels. The cults are evident in the dedications of churches, wall paintings and icons. Traces of them can also be found in folk tradition and local festivals. The existence of half-Christian, half-pagan feasts has often led to the assumption that Svans became Christians only in name: replacing the names of pagan gods with those of Christian saints, without changing their attributes.7 Thus, I want to consider the following questions: were the Svans "barbarians' in a double sense: eastern neighbours of the Byzantine empire and only 'half'-Christian? In what way were Svan cults influenced by the Byzantine cult of saints and how did Svans work up this ideology?

5 For a comprehensive survey of Christian art in Svanet'i, see: Schrade, Art, as well as earlier works, such as: N. Aladashvili, G. Alibegashvili and A. Volskaia, Zhivopisnaia shkola Svaneti (Tbilisi, 1983); R. Qenia and V. Silogava, Ushguli (Tbilisi, 1986). For chased icons, see also G. Chubinashvili, Gruzinskoe chekannoe iskusstvo, 2 vols (Tbilisi, 1959). For wall paint­ ings, see N. Thierry, 'Notes d'un voyage archéologique en Flaute-Svanétie', BK 37 (1979), 133-79, 38 (1980), 51-95; T. Velmans, Byzanz. Fresken und Mosaike (Zürich and Düsseldorf, 1999), 317. For Byzantine works of art in Svanet'i, see L. Khuskivadze, 'La staurothèque byzantine de la Svanéti', in C. Moss and K. Kiefer, eds, Byzantine East, Latin West. ArtHistorical Studies in honor of Kurt Weitzmann (Princeton, 1995), 627-32; L. Khuskivadze, 'Vizantiiskii krest iz Macxvarishi', Zograf 15 (1984), 24-40; also J.A. Cotsonis, Byzantine Figurai Processional Crosses, Dumbarton Oaks Byzantine Collection Publications 10 (Washington, DC, 1994), ill. 5a-d. For a survey of Svan church treasuries (without ill.), E. T'aqaishvili, Ark'eologiuri ek'spedic 'ia lechxum-svanet'shi 1910 cels (Archaeological expedition of 1910 to Lechkhum-Svanet'i) (Paris, 1937). 6 I am following here the concept of art as fundamental to the understanding of society proposed by R. Cormack, Writing in Gold. Byzantine Society and its Icons (London, 1985), 9-11. 7 J.B. Telfer, The Crimea and Transcaucasia II (London, 1876), 71. This opinion, widespread among travellers, reflects the situation in Svanet'i after an absence of at least two centuries of regular church life, which had led to a certain 're-paganization'. It is supported by ethnol­ ogists who interpret the pagan layer as of a relatively late time: V. Bardavelidze, Drevneishie religioznye verovania i obriadovoe graficheskoe iskusstvo gruzinskikh piemen (Tbilisi, 1957). According to other analyses, Christianity was deeply enrooted in Svanet'i, see esp. Obzor deiatelnosti Obshchestva vozstanovlenia provoslavnago khristianstva na Kavkaze za 1860-1910 gg (Tiflis, 1910), 42-60.

BYZANTIUM AND ITS EASTERN BARBARIANS

171

St George The first to be looked at is St George and the other warrior saints. Georgios, venerated as chief-martyr and high-ranked warrior saint all over the Christian world, held the top position of Christian saints in Georgia where he also became the patron of the country.8 His extraordi­ nary veneration in Svanet'i is still traceable: in the nineteenth century Svans applied to him and not to God when in dangerous situations, and what they feared most was punishment by the saint to whom they also offered sacrifices.9 In the church of St George in Kurashi, he was able to predict the death of the inhabitants of the village.10 Some feasts are still celebrated, like the Lichanishoba in Adishi (Hadish) which takes place on the first Sunday in August and where, near the little church of St George, people pray for a son while those families in which a boy had been born offer a flag to the saint. It might indeed seem that this mountain St George was a pagan god rather than a Christian saint. According to the Georgian historian Ivane Javaxishvili, as T'et'ri Giorgi ('white George') he replaced an ancient moon-god, the main deity of the old-Georgian pantheon.11 There is certainly a pagan layer also in the cult of St George, but this applies to a much higher degree to the East Georgian mountainous regions where Christianity was never entirely adopted.12 In Svanet'i, we find another image which is lucidated by the three names for St George: Jgraag, his cult denomination and his traditional name as church patron; Jgraag P'usnai in folklore; and Cminda Giorgi in inscriptions on icons and wall paintings.13

8 LThK 4 (1995), 476-8. For the association of St George with Georgia, see A.-D. v. den Brincken, Die 'Nationes Christianorum Orientalium' im Verständnis der lateinischen Historiographie (Cologne, 1973), 103-25. 9 E. Gabliani, Dzveli da axali Svanet'i (Svanet'i, old and new) (Tiflis, 1925), 125. 10 A. Stoianov, Puteshestvie po Svanetii, ZKOÎRQO 10, 2 (1876), 320-21. 11 I. Javaxishvili, K'art'veli eris istoria (History of the Georgian People) I (Tbilisi, 1960), 48-59; an English text of I. Javaxishvili, St. George the Moon-God is to be edited by S. Rapp. This theory is problematic: G. Peradze, 'Der Heilige Georg im Leben und in der Frömmigkeit des georgischen Volkes', Der Orient 12, 2 (Potsdam, 1930), 49-52, has already cast doubt on the connection of St George with the old-Georgian moon-god. Archaeological findings, Svan toponyms and the Svan ritual hymn, Lite, hint instead of a strong preChristian cult of the sun (Helios, Sol Invictus) which could have formed the basis for the cult of the Christian Saint George. 12 The Georgian chronicle describes the royal pressure, including heavy taxes and even 'the sword', that were applied: KC 1:124-6; trans. R.W. Thomson, Rewriting Caucasian History (Oxford, 1996), 138^10. On the syncretism of the cult of St George: I. Sourgouladze, Saint George dans les croyances religieuses géorgiennes (Tbilisi, 1983). 13 B. Schrade, 'Götter oder Heilige: heidnische und christliche Glaubensvorstellungen der Swanen', Georgica 4 (1999), 32-8.

172

BRIGITTA SCHRADE

Jgraag14 may be compared to the east Georgian T'et'ri Giorgi. But Georges Charachidze has noticed an important difference: while the ritual for the Khevsur-Pshav god, who is much more cruel than his Svan counterpart, focuses on the slaughter of the sacrificial animal, the kernel of the corresponding cult in Svanet'i is the common eating of the cooked meat.15 Thus the Svan cult is reminiscent more of the Christian agape than the blood-sacrifice of old pagan cults. The more Christian orientation of the Svan cult is demonstrated by our second St George: Jgraag P'usnai as well. P'usd/p'usnai means 'master, God', and p'usd are also churches dedicated to God.16 In Svan legends, Jgraag P'usnai helps the hunter Givergil against the Dalis, the pagan goddesses of hunting, thus appearing far stronger than the old gods of the Svan pantheon.17 St George and God merged into one person who stood above all pagan deities. In oral tradition, Jgraag is also the patron of churches dedicated to St George. We find at least one in each village. Inside the churches, in wall paintings, on icons and in the sult'a matiane, the community registers which are preserved in churches,18 the saint is called cminda (cmiday) Giorgi, his Georgian and Svan liturgical name. Here, we find our saint in his well-known Christian context, and this provides the key to under­ stand his veneration in Svanet'i. Apart from a few representations which show St George as a martyr with his passion, as on the eleventh-century cross of Seti and in the wall painting of Nakip'ari (1130), the saint is mostly represented as a holy soldier, in all iconographical variants: mounted and as standing figure, alone or' together with other military saints, especially Theodore and Demetrios. In combination with other saints, he also appears on crosses and in the border medallions of icons. A whole group of warrior saints is represented on one of the big crosses which were erected in front of the chancel barriers of Svan churches. It belongs to the treasury of St George in Svipi Pari and is dated to the twelfth or thirteenth century.19 The long row of military saints, George, 14 Also Jg(u)rag. On the origin of the name: A. Arabuli, 'Svanuri sakulto termini jgëràg) Etimologiuri dziebani (Tbilisi, 1990), 108-15. 15 G. Charachidzé, Le système religieux de la Géorgie païenne (Paris, 1968), 478. 16 T. Mibchuani, Dasavlet' sak'art'velos k'art'vel mt'ielt'a et'nogenezis, gansaxlebisa da kulturis istoriidan (On the history of the ethnogeny, distribution and culture of the west Georgian mountaineers) (Tbilisi, 1989), 318-20. 17 H. Fàhnrich, Der Sieg von Bachtrioni (Leipzig and Weimar, 1984), 47-53. 18 V. Silogava, Svanet'is cerilobit'i dzeglebi (Written monuments of Svanet'i) I (Tbilisi, 1986), 248-306. The earliest registers date to the thirteenth century. 19 Dating according to G. Chubinashvili, Iskusstvo I, 525, II, 473^1; and M. Axalashvili, X -X V ss. carcerebi svanet'is cheduri xelovnebis dzeglebze (Inscriptions of the X-XV centuries on monuments of Svan chased art) (Tbilisi, 1987), 37.

BYZANTIUM AND ITS EASTERN BARBARIANS

173

Demetrios, Theodore, Mercurios, Procopios and another holy soldier who bears the name of the physician saint Panteleimon,20 is accompanied by an inscription: 'Greatest among the martyrs, great chief-martyr, St George of Svip, protect and save us in both lives, be adorned and praised by your inhabitants of the valley Mai, Amen'.21 The military saints represented here are reminiscent of the Byzantine ones of the tenth to twelfth centuries.22 Indeed, what we have in Svanet'i is a reflection of Byzantine veneration of saint warriors of this period. At the same time, the Svan cult reveals changes in Georgia itself, connected with the unification of the country and the creation of a mighty Christian kingdom. The cult of these saints, originally venerated as martyrs and not as mili­ tary generals, must have been practised very early in Georgia, as Iberians and Laz possessed monasteries in the Syrian-Palestine region where these cults have their roots.23 Two fifth- and sixth-century floor inscrip­ tions in the Judean desert monastery of Bir al-Qut mentions Saint Theodore.24 And by the tenth century, a cult of saints is attested in the Georgian chronicle K'art'lis c'xovreba. It refers to an icon of St George which King Kostanti (Constantine) III of Ap'xazet'i (899-915/16) vener­ ated in the eastern Georgian cathedral of Alaverdi after his victory in Heret'i.25 The cult was advanced by his successor, Giorgi II (915-59), who ordered hymns and sermons from Ioane Minchi on Mount Sinai.26 This growing veneration is also reflected in the dedication of churches and their inscriptions.27 It seems that the flourishing of this cult was inspired by Byzantium, where warrior saints as well as the archangel Michael had become state 20 The combination of military and physician saints in Byzantine art is not unusual, being associated with healing powers: G. Vikan, A rt, Medicine, and Magic in Early Byzantium', DOP 38 (1984), 65-86. 21 Axalashvili, Carcerebi, 37-9, ill. 33-6. 22 G. Schlumberger, L'Épopée byzantine à la fin du dixième siècle (Aalen, 1969), 1 ,173, 585; II, 57, 273, 493; D. Buckton, ed., Byzantium (London, 1994), 147-8, ill. 160. 23 Procopios, De aed., eds, J. Haury and G. Wirth (Leipzig, 1964), V, 9. 24 G. Ceret'eli, Udvelesi k'art'uli carcerebi palestinidan (The most ancient Georgian inscrip­ tions from Palestine) (Tbilisi, 1960), 75-93. See also R. Mepisashvili and V. Tsintsadze, The Arts of Ancient Georgia (Leipzig and London, 1979), 42. 25 Matiane k'art'lisay (Chronicle of K'art'li), KC 1:264; trans. Thomson, Rewriting Caucasian History, 267. 26 M. Tarchnisvili, Geschichte der kirchlichen georgischen Literatur, Studi e Testi 185 (Vatican, 1955), 115-16. C. Toumanoff, 'The Kings of Abasgia', Le Muséon 69 (1956), 82, argues for Giorgi II against Tarchnisvili's Giorgi I. 27 N. Shoshiashvili, Aghmosavlet' da samxret' sak'art'velo (V-X ss.) K'Kart'uli carcerebis korpusi: Lapidaruli carcerebi I (Corpus of Georgian inscriptions in east and south Georgia, V-X centuries) (Tbilisi, 1980), 170, 243^1, 253^1, 271. V. Silogava, Dasavlet' sak'art'velos carcerebi. nakv. I (IX-XIII ss.), (Georgian inscriptions of west Georgia, part I, IX-XIII centuries) (Tbilisi, 1980), 197-200.

174

BRIGITTA SCHRADE

symbols by the tenth century.28 Their cult was supported by emperors like John Tzimiskes who renamed Euchaita to Theodoroupolis in honour of St Theodore who had supported his conquest of Silistria.29 Both Theodores, Tyron and Stratelates, who had been unified in one person until the ninth century,30 were now closely connected with the military expeditions of the empire.31 Demetrios, the local saint of Thessaloniki, was even devoted an imperial feast.32 The cult of St George was firmly established in the eleventh and twelfth centuries, with an elaborate iconography,33 based on the menologion of Symeon Metaphrastes.34 St George now passes for a protagonist of the empire, shown together with the emperor on the banner of the army.35 He is praised in hymns and sermons, among others by perhaps the best preacher of his time, John Mauropous, in the famous St George of Mangana in Constantinople.36 This cult was transmitted more or less directly to Svanet'i where, in the ninth to tenth centuries, Christianity was promoted anew by the pious rulers of the kingdom of Ap'xazet'i who also supported the conversion of the Alans on the northern side of the Caucasus.37 In this transitional period, when Ap'xazet'i and the east and south Georgian kingdoms and principalities were striving for hegemony with or against Byzantium, before they were united under the Bagratids at the beginning of eleventh century, Svanet'i maintained old traditions and received new ideas. Among the new ideas was the cult of warrior saints which must have

28 P. Schreiner, Aspekte der politischen Heiligenverehrung in Byzanz', in J. Petersohn, ed., Politik und Heiligenverehrung im Hochmittelalter (Siegmaringen, 1994), 372-7. 29 Schlumberger, Épopée 1 ,144-6. 30 H. Delehaye, Les légendes grecques des saints militaires (Paris, 1909), 11-43. 31 Constantine Porphyrogennetos, Three treatises on imperial military expeditions, ed. J.F. Haldon, (London, 1990), 120. Schreiner, Aspekte', 374-6. 32 Constantini Porphyrogeniti de cerimoniis aulae byzantinae, ed. J.J. Reiske (Bonn, 1829), 532-5. For Demetrios as patron of Thessaloniki, see Cormack, Writing in Gold, 50-94. 33 V. Likhacheva, 'The Iconography of the Byzantine Saint in the Illuminations of the Eleventh and Twelfth Centuries', in S. Hackel, ed., The Byzantine Saint, Suppl. Sobornost 5 (Birmingham, 1981), 195. L. Mariés, 'L'irruption des Saints dans l'Illustration du Psautier Byzantin', AB 68 (Mélanges Paul Peeters, II) (Brussels, 1950), 153-62. 34 H.-G. Beck, Kirche und theologische Literatur im Byzantinischen Reich, Handbuch der Altertumswissenschaft 12, 2,1 (Munich, 1959), 570-75. 35 According to an epigram of Michael Psellos, PG 122, 531. Schreiner, 'Aspekte', 374. 36 Beck, Kirche, 555-6. It is assumed that the Academy of Mangana was attended by Georgian scholars including Arsen Iqaltoeli and Ioane Petrici who were closely connected with King Davit' II: Tarchnisvili, Geschichte, 69, 204, 211. 37 Svanet'i was a principality of Ap'xazet'i from the end of the eighth century: C. Toumanoff, 'Caucasia and Byzantium', Traditio 27 (1971), 119. For promotion of Christianity in the mountains: Z. Didebulidze, Kulturnie vzaimosviazi narodov Gruzii i centralnogo Predkavkazia (Tbilisi, 1983), 38^19.

BYZANTIUM AND ITS EASTERN BARBARIANS

175

fitted in perfectly with the local ideology, as Svans had been famous as soldiers since antiquity. It was spread from the Byzantine-influenced south and west, Tao-KlarjetT and Ap'xazet'i,38 the ideological basis being delivered by the works of Euthymios and George the Hagiorites (Ep't'wime and Giorgi Mt'acmindeli), the outstanding Georgian monkfathers and translators of Mount Athos, who were active between the mid-tenth and eleventh centuries.39 While Euthymios translated the metaphrastic version of the passion of St George40 and warned his compa­ triots at home that the old, apocryphal text which evidently still existed at this time was forbidden by the church,41 George provided a Georgian translation of the Greek synaxarion which existed only 'in the church of Saint Sophia and the monastery of Studios'.42 It is this 'official' chief martyr and celestial soldier George who can be found in Svanet'i, where he is placed 'correctly' at the top of the group of military saints as in Svipi Pari,43 his passion being based on the metaphras­ tic version, as on the cross from Seti. In this form, as all-vanquishing heav­ enly warrior who could save people from every kind of danger, he became the hero of Svan men, their protector and ally in military and daily activi­ ties, represented in nearly all the churches even when they were not dedi­ cated to him. In wall paintings, we find St George often together with Theodore, mostly on horseback, in prominent locations: beside the entrance or leading from the entrance to the altar, or as standing 'protecting' figures.44 As protectors, they also appear on the façades, as at St George in Adishi in the twelfth to thirteenth century, where they are directed towards the entrance of the village, as if to ward off possible enemies. But above all, we find St George alone. This, and the overwhelming number of icons of him from the eleventh to fourteenth centuries, demon­ strates that Svans - like Georgians in general - followed, but did not simply copy, the Byzantine cult.45 They adapted it to their own needs,

38 G. Peradze, 'Die Problème der àltesten Kirchengeschichte Géorgiens', OrChr 29, 3 /7 (1932), 162. 39 Tarchnisvili, Geschichte, 126-74. 40 E. Gabidzashvili, Cminda giorgi dzvel k'art'ul mcerlobashi (St George in old Georgian liter­ ature) (Tbilisi, 1991), 178-99. 41 Tarchnisvili, Geschichte, 330-34; revised by Gabidzashvili, Cminda giorgi, 178-99. 42 P. Peeters, Le tréfonds oriental de l'hagiographie byzantine (Brussels, 1950), 205. 43 Delehaye, Légendes grecques, 2-3. 44 These motifs can also be found in Cappadocia: A.W. Epstein, 'Rock-cut chapels in Góreme Valley, Cappadocia: the Yilanli group and the Column Churches', CahArch 24 (1974), 116-18, figs 1, 2. The facing mounted saints are more accentuated in Svanet'i: T. Velmans, 'La koine grecque et les régions périphériques orientales du monde byzantin', JÔB 31/2(1981), 717. 45 About 60 chased and painted icons survive, not counting representations on crosses and in the borders of chased icons.

176

BRIGITTA SCHRADE

notably by avoiding the great variety of saints characteristic of Byzantine society.46 Thus, there are no churches in Svanet'i dedicated to the Byzantine 'state symbols' Theodore and Demetrios. The predominant position of St George after the eleventh century was obviously due to the important role of this saint within the united Georgian kingdom (Sak'art'velo) to which Svanet'i belonged as a princi­ pality (saerist'avo). It reached its Golden Age under Davit' II Aghmashenebeli (the Builder) (1089-1125) and his successors, especially Queen T'amar (1184-1212). This period was also one of military expan­ sion, and St George became the patron of the army and the official saint of the state. His cult was promoted through Byzantine iconographic and historiographic concepts: in the Georgian chronicle he is described as the heavenly leader of King Davit' at the battle of Didgori (1121) which led to a decisive victory over the Seljuqs,47 and an icon in the monastery of St Catherine on Mount Sinai shows him as mediator between the king and Christ.48 On a somewhat smaller scale, this is repeated in Svanet'i, whose rulers were proud of their position at the court and adopted official titles. At the same time, they persisted fiercely in maintaining their independence.49 Both are reflected in the cult of St George: he appears as protector of the Svan rulers as well as of Svanet'i itself. Two icons of the saint from the twelfth century are donations by the important nobleman, Ioane Vardanisdze.50 On one from Labechina in Lechxumi (formerly Lower Svanet'i), the donor bears the Greek title protostrator,51 and on the other from Mestia in Upper Svanet'i, he has his Georgian court titles: erist'av 46 For the variety of church patrons in Constantinople: R. lanin, Constantinople Byzantine (Paris, 1950), 286-472. 47 C'xorebay mep'et'-mep'isa davit'isi, KC 1:341; trans. Thomson, Rewriting Caucasian History, 332-3. 48 D. K'ldiasvili, 'L'icône de Saint Georges du Mont Sinai avec le portrait de Davit' Aymasenebeli', REGC 5 (1989) 107-28. A. Eastmond, Royal Imagery in Medieval Georgia (University Park, PA, 1998), 67-71, defines it as 'a purely Byzantine image of royal power'. The union of church and state is evident also in the triptych of St George from the treasury of Svetic'xoveli, now in Svanet'i, which belonged to the catholikos Mirianisdze who crowned Queen T'amar. 49 For the eleventh and twelfth centuries, the chronicles record two revolts against the Crown in which Svans were involved: one against Giorgi II, the father of Davit' II, (Chronicle of K'art'li, KC 1:315; trans. Thomson, Rewriting Caucasian History, 306); the other against queen T'amar (M. Brosset, Histoire de la Géorgie depuis l'antiquité jusqu'au XIXe siècle (St Petersburg, 1849), 411-29). 50 E. Takaïchvili, 'Antiquités géorgiennes', Byz 12 (1937), 197-204. Ioane Vardanisdze lived during the reigns of Giorgi III and T'amar. See also Axalashvili. Carcerebi, 31—4. 51 The title is abbreviated, as usual in Georgian inscriptions, and appears only on one icon. Thus it remains an interpretation by the author, supported by P. Peeters. On titles, see R. Guilland, Recherches sur les institutions byzantines I (Berlin and Amsterdam, 1967), 478-83.

BYZANTIUM AND ITS EASTERN BARBARIANS

177

(ruler, prince) of Svanet'i and mechurchlet'uxuc'esi (chancellor of the exchequer).52 While pleading on the icon of Labechina for himself and his family, Ioane implores the Mestia icon for heavenly support for 'Great Svanet'i' (didisa suanet'isat'wis). The protection of all Svanet'i is also evident on the thirteenth-century silver-head of the Svan banner Lem,53 where the saint is addressed as supporter of the 'united and pacified valley of Svanet'i' (ert'obilsa hevsa bedniersa suanet'isasa), together with the prophet Jonah and the 'localized' archangel, to whom I will return later. Finally, it should be noted that St George is often represented as a mounted saint killing not a dragon but a figure named 'impious king Diocletian'.54 This Georgian iconographical type is unknown in Byzantine art and has been explained as a symbol of the conflicts between Georgia and Byzantium which followed the struggles for territory and political power especially under Basil II (976-1025).55 It is not possible to resolve this question here, and in our context it is not that important whether the image did embody that conflict or was rather an appropriation of older eastern models for representing the victory of Christianity. What is signif­ icant is that Svans seemed to like the idea that the emperor could not kill the martyr, but that St George could triumph physically as well as ideo­ logically over him. This motif is widespread in Svan wall painting and on icons, and survived longer there than in the Georgian lowlands, where it was replaced from the eleventh century by the dragon motif, one of the official miracles of the saint.56

St Kvirike In addition to St George, we find in Svanet'i another martyr and heavenly protector whose cult was also of great importance. Called Cminda Kvirike in Georgian and Svan, he is better known in the west as St Quiricus.57 There are only two churches in Upper Svanet'i dedicated to this saint, 52 For Georgian court titles, see I. Surguladze, 'Die Hausordnung des Konigshofes', BK 31 (1973), 225-45. 53 Axalashvili, Carcerebi, 62-4. The Svan banner of which a copy is preserved in Mestia museum, has the form of a lion or wolf and is supposed to have been used as palladium. A similar banner in the form of a dragon is represented in the ninth-century Psalterium Aureum: L. Reau, La miniature (Melun, 1946), pi. III. 54 The earliest icons of this type in Georgia are dated to the tenth century: Chubinashvili, Iskusstvo I, 336-7, II, 34-5. 55 Chubinashvili, Iskusstvo, 368-73. On Georgian-Byzantine conflicts, see Salia, History, 141-8. 56 The last Svan icons of this type are dated to the thirteenth century: Chubinashvili, Iskusstvo II, ill. 195-200. Georgian texts of the dragon miracle exist from the eleventh century: Gabidzashvili, Cminda giorgi, 73-140. See also: W. Haubrichs, Georgslied und Georgslegende im friihen Mittelalter, Scriptor 13 (Konigstein, 1979), 214-15, n. 24. 57 Enciclopedia dei Santi. Le Chiese Orientali I, BSO (1999), 491, LThK 5, (1996), 1082.

178

BRIGITTA SCHRADE

but both were evidently important ones.58 The most famous belongs to a former monastery which, like the church itself, is called the Lagurka: 'the church of Kvirike'.59 Lagurka monastery, at the top of a steep hill in the village commune of Kala, is the most venerated place of Svanet'i.60 Its festival, Kvirikoba, cele­ brated on 28 July,61 is the main feast of Svans. The modest church has preserved an extraordinary rich treasury to which belong nine icons of St Kvirike, the only group of his icons still existing in Georgia.62 One shows him together with his mother Ivlita (Julitta). The legend of the holy martyrs Julitta and Quiricus is well known: Julitta, a noble widow from Iconium, fled together with her three-year-old son Quiricus to escape the Christians' persecution under Diocletian. They were captured in Tarsos, tortured and finally put to death. Quiricus suffered martyrdom before the eyes of his mother, who was decapitated shortly afterwards.63 It is worth recalling this Vita, as the two scenes which are depicted in the wall painting of Lagurka (in addition to two single representations of the saints)64 are not in line with the passion text. According to the legend, Quiricus is a little boy, and when represented together with his mother (as in the wall paintings of Kastoria)65 he is also depicted as a child. But the wall painting of Lagurka church, dated 1111,66 and the icons show him as a young adult martyr. What is the reason for this? It seems that we have here two different cults, or rather an ancient cult that was changed to suit new requirements. This change accompanied the replacement of the Jerusalem service rites by the Constantinopolitan rites, which took place at the time of the unification of the east and west 58 The other church belongs to the village Ieli. For the church treasury (now in Mestia museum), see Takaïchvili, Expédition, 231—4. 59 Lagurka is an analogous form of lamaria, 'the church of the mother of God', and is phonologically transformed from la-k(v)irike to la-gurk in Svan. I am very grateful to Prof. Z. Cumburidze for this information. Gudjedjiani, Dictionary, 130, gives another explanation (a Svan transformation of the Greek kuria, kuriake, 'church in the Lord's name'). The first expla­ nation is more plausible. 60 B. Schrade, 'Realität oder Fiktion: Das Kloster des Quiricus in Obersvaneti (Curzon Press, forthcoming). 61 This corresponds to 15 July of the old (Julian) calendar. A second feast was celebrated on Easter Saturday, Obzor, 53^4. 62 G. Chubinashvili, Tzobrazhenie sviatogo Kvirike v Gruzinskoi chekanke', Literaturnye Razyskania 2 (1944), 108, mentions nine icons. Another icon was discovered in neighbouring Ushguli: Qenia, Ushguli, 71. 63 Acta Sanctorum Jun. IV (Paris, 1867), 13-31. 64 Aladashvili, Zhivopisnaia shkola, 30-32, 56-77,131-3, Thierry, 'Voyage [I]', 153-8. 65 S. Pelekanidis, Kastoria (Thessaloniki, 1953), pi. 40b. In single representations he appears also as youth, L C I7, 242. 66 Aladashvili, Zhivopisnaia shkola, 132, reads the inscription as 1112.

BYZANTIUM AND ITS EASTERN BARBARIANS

179

Georgian churches under the patriarchate of Mc'xet'a, a process antici­ pating the unification of Georgia itself.67 This becomes evident in the case of Kvirike. His church is obviously older than the monastery. It can be dated to the tenth or eleventh century while the monastery buildings were erected later, probably at the begin­ ning of twelfth century, when the church was painted.68 Kvirike alone is the church patron, as is shown by all the inscriptions and also the icons.69 The oldest icon, dated to the first third of the eleventh century and evidently the prototype for the following ones,70 depicts Kvirike as a young martyr who resembles the martyr St George on an icon of the sixth century in the monastery of St Catherine on Mount Sinai.71 This is not mere coincidence or the mistake of a provincial master but obviously connected with an early version of the passion of Quiricus and Julitta. This Vita, which was condemned by the Gelasianum at the end of the fifth century along with an early version of St George's legend and later replaced by the 'purified' version of Metaphrastes's collection, treats Quiricus as the main person of the couple, a ripe martyr who fights hero­ ically for his faith. 72 This colourful, dramatic version survived in eastern churches for a long time, similar to the 'popular book' of St George.73 Like George, Quiricus appears in this as an invincible martyr who shocks the infidels with his rhetorical and physical strength, encouraging his mother who finally declares: 'From now on, you are my father, and I am your daughter'.74 An apocryphal version of this legend also existed in Georgia. It was criticized by Ep't'wime Mt'acmindeli at the turn of tenth century, a time when the Georgian Church was eager to prove its autocephaly to the

67 For the subordination of the West Georgian bishoprics to the East Georgian patriarchate see, B. Martin-Hisard, 'Kirche und Christentum in Georgien', in E. Boshof, ed., Geschichte des Christentums 4 (Freiburg, Basel, Vienna, 1994), 549, 557-61. 68 The dating is derived from inscriptions, architecture and archeological findings, see Schrade, 'Kloster'. 69 For fresco inscriptions, Aladashvili, Zhivopisnaia shkola, 31; for graffiti inscriptions, Silogava, Svanet'is cerilobit'i II, 281-92. 70 Chubinashvili, 'Izobrazhenie', 108-13. 71 K. Weitzmann, Die Ikone (Munich, 1978), 42, ill. 2. 72 A. Dillmann, Über die apokryphen Märtyrergeschichten des Cyriacus mit Julitta und des Georgius, Sitzungsberichte der Kgl. Preuss. Akad. d. Wiss. zu Berlin XXIII (Berlin, 1887), 339-56. 73 K. Krumbacher, Der heilige Georg in der griechischen Überlieferung, ABAW XXV 3 (Munich, 1911). 74 R. Basset, 'Le synaxaire arabe-jacobite', PO 17, 3 /5 (1923), 658-9. See also I. Guidi, 'Le synaxaire éthiopien', PO 7, 3 /2 (1950), 370-72, for 28 July, where the saint is invoked as mighty heavenly protector: 'death will not come to cattle and the pest will not kill men where a church has been erected in his honour'.

180

BRIGITTA SCHRADE

Greeks.75 But old traditions are not so easy to erase. This was true of Kvirike whose cult has early traces in the lowlands of Georgia.76 Until the eleventh century many churches were dedicated to him, and his name was popular with simple monks as well as high clergy and worldly rulers.77 He, like St George, even had his month, kvirikobis t've: July, the month of his martyrdom.78 He enjoyed the same high appreciation in Svanet'i, where he had also a memorial month of his own, ligurke t'ev.79 His cult survived here, while in the more central parts of Georgia the saint obviously shared the fate of martyrs whose legends and cults changed or were forgotten. Kvirike, the invincible martyr and heavenly protector who was connected ideologi­ cally with the chief martyr St George,80 was reduced in the now officially adopted version to the little child of the martyr Julitta and so lost his special meaning.81 This development can also be traced in the changes of Georgian memorial days. In lectionaries of the tenth century which reflect the Jerusalem rites of the seventh century and which were in use also in SvanetT, we find 15 July as his memorial day alone;82 in the calendarium of Sinaiticus 34 from the end of the tenth century, we have the couple Julitta-Quiricus, although for two days, 15 and 16 July (the second date was originally the memorial day of Julitta alone).83 And in the first third 75 Tarchnisvili, Geschichte, 330-33. On the autocephaly of the Georgian Church, MartinHisard, 'Kirche', 578-81. 76 St Kvirike (Kwirike) appears in the Palestinian meneon of the seventh-eighth century: A. Cagareli, 'Pamiatniki gruzinskoi stariny na Sviatoi Zemlie i na Sinai', PS 10 (1888), 161. He is also addressed on the stela of Usaneti, dated to the eighth century: Shoshiashvili, Aghmosavlet', 128-29. The foundation of a monastery of St Kvirike is mentioned in Grigol Merchule's Vita of Grigol Xandzt'eli: N. Marr, G. Merchul: Zhitie sv. Grigoria Khandztiiskogo (St Petersburg, 1911), 112. 77 Chubinashvili, 'Izobrazhenie', 97-8. For names of bishops and k'orepiskoposi (who were also political rulers) see Thomson, Rewriting Caucasian History, 399-400. For examples from Svanet'i: Silogava, Sak'art'velos carcerebi, no. 30-31. 78 The month of St George - giorgobis tve - is celebrated on 10 November as the main feast in Georgia, according to the old Jerusalem tradition, and not on 23 April: K. Kekelidze, Ierusalimskii kanonar VII veka (Tiflis, 1912), 282-6. Svans celebrate as first feast the April memorial day, thus following Greek tradition. 79 I. Nisharadze, 'Russko-svanskij slovar', SMOMPK 41 (1910), 143. 80 The names of both saints are connected in Georgian cult places, as at the east Georgian Kvirike cminda giorgi: H. Fähnrich, Lexikon der Georgischen Mythologie (Wiesbaden, 1999), 167. 81 K. Kekelidze, Etiudebi dzveli k'art'uli literaturis istoridaan (Studies in the history of old Georgian literature) 5 (Tbilisi, 1957), 186-7. See also the text (in Georgian) in Chubinashvili, 'Izobrazhenie', 100-104. 82 Kekelidze discovered two lectionaries in Svanet'i, one in Latali (Latal) and the other in the Lagurka monastery. In the Lagurka ms, the pages around the feast of Kvirike are torn out. But the analogous lectionary from Latali (Kekelidze, Kanonar, 120) gives 15 July for Kvirike alone, the accompanying troparion and reading refer to the heavenly warrior (1 Cor. 9, 24-7) and the child brought to Jesus in Kapernaum (Math. 17:24-18:10). 83 G. Garrite, Le calendrier palestino-géorgien du Sinaiticus 34 (Xe siècle) (Brussels, 1958), 78-9.

BYZANTIUM AND ITS EASTERN BARBARIANS

181

of the eleventh century, in the Short Synaxarion of Ep't'wime Mt'acmindeli, 15 July is dedicated definitely to both Julitta and Quiricus.84 That the cult of Kvirike alone survived for longer in Svanet'i, where we find inscriptions and representations at least until the fourteenth century,85 may have several reasons. One could be that Svans, as we have seen, used older lectionaries. But above all, it resulted from a certain conservatism of Svan society which had long and stable traditions in both social and political structures. The independent communes, called valleys (xevi) in old texts, with elected elders at their top, had main churches and thus main saints which were addressed on behalf of the community for help and assistance. St Kvirike was embedded in this system and could not just be transformed. Even T'evdore, the painter of Lagurka, who called himself 'the king's painter' (mep'isa mxatvrisayt'a) and so was obvi­ ously connected with the royal court of Davit' II, paid tribute to this; or had to, as his painting was ordered by the local community: 'In the name of God, this Holy Church was painted and adorned for glory and prayer by the nobles of this valley (amis xevisa aznaurt'a) and all who built it ... Holy Kvirike, be praised and have mercy ...'.86 T'evdore had already represented the saint as an adult martyr in his first wall painting in Ip'rari (Ip'ral, 1096), in the same commune of Kala, and repeated it in his next commission at the church of St George in neighbouring Nakip'ari (1130). In this church Kvirike appears together with St Demetrios on the chancel barrier, as he had at Ip'rari.87 Thus T'evdore's masterly paintings, which reflect contemporary ideas, at the same time respect the older texts, and underline the visible role of Kvirike as patron of a highly venerated church, helper of the valley of Kala and one of the celestial protectors of the Svans. It may seem as if the high esteem accorded to Kvirike suppressed the role of his mother. But this was not the case. On the contrary, she, in her turn, fitted in perfectly with the row of female martyrs which we find not only in the paintings of T'evdore but in many other Svan wall paintings, too.

Female saints The female element in Svan churches comes as no surprise, as Georgians

84 Kekelidze, Kanonar, 297-310. 85 In the Lagurka church he is invoked in graffiti inscriptions until the thirteenth century: Silogava, Svanet'is cerilobiti II, 281-92. See also his representation in the upper church of Laghami (Schrade, Art) and his icon from the Lagurka treasury (T'agaishvili, Ek'spedic'ia, 199, no. 52) from the fourteenth century. 86 Aladashvili, Zhivopisnaia shkola, 30-55, esp. 31. 87 Aladashvili, Zhivopisnaia shkola, 77-101, esp. 79.

182

BRIGITTA SCHRADE

have a long tradition in venerating female saints. Among the rich hagiographic literature (the first passions of saints being translated from Greek, Syriac, Armenian and Arabic) exist special collections, materiki (materikon), of female saints' passions.88 Whole groups of female martyrs were also included in mixed hagiographic collections, like Sts Febronia, Christine and Catherine in a tenth-century manuscript from Sinai.89 From the end of the tenth century, we know of images of women saints in Tao. Inscriptions in Oshki mention Sts Marina, Irene, Catherine, Barbara, Thekla, Helena and the Georgian Nino. Sts Barbara and Marina are found in the cathedral of Ishxani.90 From Tao the veneration of women saints may have spread to Svanet'i. The monastic movement in Tao, revived by Grigol Xandzt'eli at the end of the eighth century, had close contacts with Georgian monasteries and convents in Constantinople, Palestine and other parts of the East Christian world.91 It delivered also the basis for the cultural development of remote regions where Christian culture was promoted in a combina­ tion of Jerusalem and Byzantine traditions.92 Both traditions are reflected in Svanet'i, where female saints are often represented in whole groups, together with warrior saints, as in Tao and Cappadocia.93 But compared to the great variety of saints in Cappadocia from the tenth century onwards,94 the range in Svanet'i is much more restricted. We find fewer saints, principally Sts Barbara, Catherine, Julitta, Marina, Christine, Irene and Thekla. And what is remarkable in compar­ ison to Tao is that there are no traces of a cult of Nino, the apostolic saint who converted East Georgia in the early fourth century. And this, despite the fact that the cult of St Nino was closely connected with Queen T'amar who, in her turn, was highly venerated in Svanet'i.95 The absence of Nino 88 Tarchnisvili, Geschichte, 392-4. 89 G. Garitte, Catalogue des manuscrits géorgiens littéraires du Mont Sinai, CSCO Subs. 9 (1956), 15-26. 90 E. Taqaishvili, Arkheologicheskaia ekspedicia 1917-go goda v iuzhnye provincii Gruzii (Tbilisi, 1952), 38, 64, 55, fig. 69. On the dating of the frescoes, see N. Thierry, Teintures du Xe siècle en Géorgie méridionale et leur rapports avec la peinture byzantine d'Asie mineure'. CahArch 24 (1974), 86-105. 91 For the web of Georgian contacts throughout the empire: R. Morris, Monks and laymen in Byzantium, 843-1118 (Cambridge, 1995), 81-2,137, 282. On monasteries: Peeters, Tréfonds, 203-7. 92 Grigol Xandzt'eli had contacts to Jerusalem and Constantinople, Martin-Hisard, 'Kirche', 566-8. 93 Cf. the church of St Barbara in Soganli: G. de Jerphanion, Une nouvelle province de l'art byzantine III (Paris, 1934), figs 189, 2-3, and others. 94 M. Restle, Die Byzantinische Wandmalerei in Kleinasien II and III (Recklinghausen, 1967), esp. nos 201-50. 95 For the ideological link between St Nino and Queen T'amar in Georgian art, Eastmond, Imagery, 119-21,199.

BYZANTIUM AND ITS EASTERN BARBARIANS

183

may be explained by the fact that she was not responsible for the conver­ sion of the Svans, this being officially ascribed to the apostle Andrew (and, unofficially, in Svan understanding, to Christ himself).96 Her feast may never even have played a role in their religious life, as it is absent from the old Svan lectionaries.97 The most venerated female saint in Svanet'i was Barbara, the fourthcentury virgin martyr from Heliopolis.989Her Svan name is Barbal, " And apart from the Theotokos (Lamaria), she is the only female church patron in Svanet'i. It has been supposed that St Barbara was venerated as the Christian replacement of an old sun goddess.100 But the elements of her cult are rather those of a popular Christian saint, which implies a syncretistic form of veneration. Her feast (Barblash), which is celebrated on her memorial day, is of Christian origin.101 And her great popularity in Svanet'i where many churches were dedicated to her,102 is not unusual but corresponds to her veneration in Byzantium, where she was invoked as helper in cases of disease and the hour of death, being even connected with astral prophecy.103 The same functions may be deduced from her Svan popular cult,104 and icons where St Barbara appears together with military and physician saints, as on an eleventh-century chased icon from Ieli, or with the Theotokos on a painted icon of the tenth or eleventh century from Nakip'ari.105 This last icon and her representations in wall paintings

96 The mission of Andrew in the Caucasus is legendary and only appears in the eleventhcentury Georgian-Antiochian controversy on the autocephaly of the Georgian Church: K. Kekelidse, "Die Bekehrung Géorgiens zum Christentum', Morgenland 18 (1928), 9-15. For the Svan tradition, Obzor, 44. 97 Her feast on 14 January is fixed in Georgian menaeia, and appears in the tenth-century calendarium of Ioane Zosime: Garitte, Calendrier, 36, 44. 98 LThK 1 (1993), 1401-2. 99 Also Barbol: V. Bardavelidze, Svanur xalxur dgheobat'a kalendari (Svan folk daily calen­ dar) (Tbilisi, 1939), 190. 100 V. Bardavélidzé, 'Chant sacré svane "Barbal Dolaschi"', BK 11-12 (1961), 188-91. M. van Esbroeck, 'Legenden in der Geschichtsschreibung', in B. Schrade and Th. Ahbe, eds, Géorgien im Spiegel seiner Kultur und Geschichte (Berlin, 1998), 84-5, opposes Bardavelidze's opinion and demonstrates the pure Christian meaning of Barbara, who is ideologically connected with Sts Christine and Irene. 101 On 4 December, according to the old calendar (Bardavelidze, Kalendari, 1-5). 102 Some of the churches, described as 'old ones', no longer exist, like those in the once rich and important communes of Mulaxi and Laxiri: T'aqaishvili, Ek'spedic'ia, 251-2, 262. 103 As church patron in Constantinople: Janin, Constantinople, 290, 296, 299; as deathhelper: D. H. Kerler, Die Patronate der Heiligen (Ulm, 1905), 432; and for prophecy, H. Delehaye, Les légendes hagiographiques, Subsidia Hagiographica 18 (1927), 150-51. 104 She is venerated as helper in case of disease: Bardavelidze, Kalendari, 190. 105 Dated by Thierry, 'Voyage [I]', 161: eleventh/twelfth century; and by T. Sakvarelidze, A. Alibegashvili, Georgian Icons (Tbilisi, 1994), 155: tenth century.

184

BRIGITTA SCHRADE

resemble her impressive image in the lower church of the Saviour in Laghami, probably a sepulchral chapel of the ninth or tenth century.106 In older literature, this church ensemble is called a monastery, and I think that we should look here for the basis of her cult in Svanet'i, which might have been distributed and maintained by monasteries. It is certainly not haphazard that one of her churches, in Xe,107 is situated not far from Lagurka monastery which was, at least partly, inhabited by nuns.108 Another church exists in the neighbouring commune of Ushguli (Ushkul), which was also an important political and church centre, and is known from later sources as a refuge place for noble women from other parts of Georgia.109 St Barbara is often represented together with St Catherine of Alexandria. Both are included in the iconographic structure of the only representation of a Georgian king in Upper Svanet'i, the coronation image of Demetre I (1125-54), at Mac'xvarishi (painted in 1140). St Catherine, with all the attributes of her royal birth, plays an official role as holy martyr, creating a connection between heaven, king and local nobles.110 It was certainly this role as female mediator between God and the world and the expectation of heavenly help from women saints for earthly affairs, which was important to Svans, and a reason to maintain the cult of female saints - more so as women had an equal position in Svanet'i, where cults were divided between men and women as were the duties of daily life.111 The Theotokos, the patron saint of all Georgia under the Bagratids,112 was included in this cult and often represented, but could never displace the women martyrs. This could have been a tribute to the Jerusalem traditions. Svans were closely connected with monasteries in the Holy Land where they also copied manuscripts for their churches at 106 Thierry, 'Voyage [I]', 176-7, fig. 26. Barbara also appears in an extensive row of women saints in a fourteenth-century wall painting in the upper church. Archaeological excavations of this church are in preparation. 107 For wall paintings: T. Velmans, 'L'église de Khé, en Géorgie', Zograf 10 (1979), 71-82. 108 As no typika or other written sources from the monastery exist, its history can be deduced only from inscriptions. They recall the daughter of the Imeretian King Bagrat (1660-81) who lived here under her nun's name, Tekla. She is mentioned also in the dedica­ tory inscription of an icon of the Theotokos from neighbouring Cvirmi where are repre­ sented Sts Catherine and Marina: T'aqaishvili, Ek'spedic'ia, 218-20. 109 K. Machabeli, 'Ein Werk der europäischen Kleinkunst in Swanetien', Georgica 20 (1997), 132-3. Svans believed that Queen T'amar was buried in the monastery of Ushguli, the inhab­ itants of Xe assumed it was in the church of St Barbara: Bartolomej, 'Poezdka v Volnuiu Svanetin', ZKORGO 3 (1855), 168-9. 110 For a convincing analysis: Eastmond, Imagery, 71-91. 111 Men were responsible for hunting and meat products, they invoked the heaven for help while women had to take care of bread and milk products and invoked the earth: M. Tchartolani, in Schrade, Art. 112 Eastmond, Imagery, 54.

BYZANTIUM AND ITS EASTERN BARBARIANS

185

home.113 Thus, women saints appear in Svanet'i as female counterparts of the military saints together with whom they were mostly also repre­ sented. Barbara-Barbal seems to have been the female equivalent of George-Jgraag: a popular saint who guaranteed support and help espe­ cially for Svan women.

Archangels Saints in Svanet'i not only had a protective function, they also contributed in structuring society, being themselves 'structured' within a system that was accepted and understood by everyone. At the beginning of the twen­ tieth century, religious ceremonies were still opened by an elder with the formula: 'To the glory of God, the Almighty ... to the glory of the Creator ... to the glory of Michael and G abriel... to the glory of St George ...!' Then were enumerated all Svan churches with their patrons.114 Within this fixed hierarchy, God the Almighty stood at the top and was addressed by mediators.115 Those were, in the first place, the archangels and St George. This final section examines the role played by the holy archangels. Analogous to St George, archangels appear as patrons of many churches, as protectors of the altar, the holiest place of the church, and in a great iconographic variety as holy warriors and imperial guards in wall paintings and on icons. The archangels, in Svan t'aringzel, t'arglezer,116 were treated in Svanet'i, as in other parts of the Christian world, as saints. Their cult as mediators between the heavenly and the earthly was established early in Georgia,117

113 The colophon of the Gospels of Lahili (fourteenth century) mentions the Svan Iovanc, son of K'sc'xiani, who copied this manuscript in the church of the Resurrection in Jerusalem for the church of Muxeri and sent it by the deacon of P'xotreri to Svanet'i: Silogava, Svanet'is cerilobiti I, 57. Svans as pilgrims in the Holy Land are mentioned still in the seventeenth century by Patriarch Dositheos of lerusalem: M. Brosset, 'De l'État réligieux et politique de la Géorgie jusqu'au XVIIe s.', Bull Scientifique publ. par l'Académie lmp. de Saint-Pétersbourg V, 111-12(1839), 239. 114 Obzor, 53. Also V. Bardavelidze, K'art'uli (svanuri) saceso grap'ikuli xelovnebis nimushebi (Examples of Georgian [Svan] graphie art) (Tbilisi, 1953), 76-7. 115 Behind this can be detected the early Christian idea that the veneration of saints means veneration of Christ who is the only mediator (1 Tim. 2, 5). 116 From Georgian mt'avarangelozi. For variants, Gudjedjiani, Dictionary, 106. 117 Inscriptions at Pantiani (east Georgia; second half of the seventh century) and on the stela of Usanet'i (eighth century) both invoke the archangels Michael and Gabriel. Shoshiashvili, Aghmosavlet', 99-100, 127-8, fig. 35. In west Georgia, where Greek was the official liturgical language until the tenth century, we find Georgian inscriptions which mention Michael and Gabriel (Sxieri) or the archangels in general (Zemo Krixi) from the ninth century onwards: Silogava, Sak'art'velos carcerebi, 32-4, 40-^3. In the reign of T'amar, the cult of the archangels was spread from Georgia to the north Caucasus: J. Knobloch, Homerische Helden und christliche Heilige in der kaukasischen Narthenepik (Heidelberg, 1991), 53.

186

BRIGITTA SCHRADE

and played an important role in the imperial ideology of Byzantium. Archangel Michael, in particular, was closely connected with the imperial cult. As protector of the capital and patron of Byzantine emperors, he reached the status of a 'national' saint, being more appreciated than Christ or the Theotokos.118 This imperial concept, with its splendid pictorial realizations and perhaps also the close connection of the celestial archistrategos Michael with military saints, seems to have influenced the Georgian and Svan representations of the archangels.119 But in their cult, Svans made their own choice according to older models and their special needs.120 The archangels, above all Michael and Gabriel, appear often together, or they are generalized. In this form, they were made the protectors of a valley, village or church and addressed as heavenly helper for the community. In some cases, as on the Svan flag Lem, or on a thirteenth- or fourteenthcentury icon of the Archangel Michael, they are even invoked for all Svanet'i and Georgia: 'Saint Archangel of Muxeri,121 not being made (built) by the hand of man, bless the Bagrationi kings (mep'eni bagratunianni) and the Dadiani,122 and the nobles (didebulni), and united Georgia (ert'obili sak'art'velo), and all the Svans {ert'obilni suanni), and the valley of Latali, and increase your village Laila and all who venerate you. Amen.'123124 Archangels were localized institutions, like the other saints and also Christ as church patron (Mac'x(o)var)}24 because they merged with their church into one single unit. Being legal institutions, they could be addressed as oath-saints and as guarantors in treaties and official docu­ ments. As protectors and heavenly guides of the communes or all Svanet'i, they were assigned even a political function. In this form, they also became an unrenounceable part of social and political life. They were venerated in their representations, which acted as the core of their cult.

118 According to Schreiner, 'Aspekte', 372-3. 119 We find the archangels in Svan wall paintings and on icons in all the iconographie vari­ ants known from Byzantine art: mostly in court dress, with loros, labarum, and sphera: V. Lixacheva, 'Dve ikony arxangelov iz Verxnei Svanetii', PS 17 (80) (1967), 159-66. 120 For the early veneration of archangels and their magic function cf. M. Tatic-Djuric, Das Bild der Engel (Recklinghausen, 1962), 8-34. The same veneration is reflected in Svan cults, where archangels appear as protectors of mountain peaks and are invoked for the birth of a son, Bardavelidze, K'alendari, 191. 121 Muxer(i) is a once important church belonging to the village of Laila in the commune of Latali. 122 Princes of Mingrelia who expanded their sovereignty over Lower Svanetl when the united Georgian kingdom fell apart after the thirteenth century. 123 Axalashvili, Carcerebi, 64-7. 124 For images: T. Velmans, 'L'image de la Déisis dans les églises de Géorgie', CahArch 29 (1980-81), 47-102.

BYZANTIUM AND ITS EASTERN BARBARIANS

187

The icons as images of the 'imagined' saints125 were equal to relics and holy to such a degree that they were not allowed to be taken away from churches, sometimes even being fixed with chains.126 But this popular veneration does not mean that Svans were barbarians in the sense that they had not become Christians. They had found their own access to ideas which they used according to their needs. Their saints were expected to work miracles, to maintain law and order, to be moral support, help and protection as in the Byzantine empire and elsewhere in the Christian world.

125 Cf. the cult of the 'Saint Imagined' St Demetrios in Thessaloniki: Cormack, Writing in Gold, 50-94. 126 This is reported by D. Bakradze, Svanetia, ZKOIRGO 6 (1861), 55.

188

BRIG ITT A SCHRADE

Fig. 11.1. Cross of St George in Seti/Mestia (eleventh century). The 1.25 m

high cross showing nine scenes of the martyrdom of St George was erected in front of the chancel. The archangel Gabriel (below) is a later addition, probably replacing a lost inscription (photo: Rolf Schrade)

BYZANTIUM AND ITS EASTERN BARBARIANS

Fig. 11.2. St. George in Nakip'ari. Wall painting of St George tortured on

the wheel (1130) (photo: Rolf Schrade)

189

190

BRIGITTA SCHRADE

Fig . 11.3. Cross in the treasury of St George in Svip'i P a' ri (twelfth or thirteenth century). On the lower arm of the 2.40 m high cross are Sts George, Demetrios, Theodore, Mercurios, Panteleimon and Procopios (photo: Rolf Schrade)

BYZANTIUM AND ITS EASTERN BARBARIANS

191

Fig. 11.4. Enkolpion of Saint George from St. George in Ip'xi (second half of the twelfth century). The triptych is chased in gold and decorated with enamel, pearls and a carnelian. The inscription on the reverse names Mik'aeI IV Mirianisdze, catholicos-patriarch at the beginning of Queen T'amar's reign as its donor. According to the eighteenth-century historian Vaxushti Bagrationi, it was brought from the patriarchal church of Svetic'xoveli to Svanet'i during the Mongol invasions of the thirteenth or fourteenth century (photo: Rolf Schrade)

192

BRIG ITT A SCHRADE

Fig. 11.5. Icon of Saint George (twelfth century) from Labechina with an inscription of the Svan nobleman Ioane Vardanisdze. From: G.

Chubinashvili, K'art'uli ok'romchedloba VIII-XVIII saukuneebisa (Georgian Goldsmith's art of the VIII-XVIII centuries) (Tbilisi, 1957), ill. 56.

Fig. 11.6. Chased icon of Saint George killing Diocletian from Seti/Mestia

(eleventh century) with military saints on the side borders. The two inscriptions read: 'Saint George of Seti' (left) and 'Saint George' (right).

BYZANTIUM AND ITS EASTERN BARBARIANS

Fig. 11.7. The former monastery of St. Quiricus (Lagurka) in Kala during the feast of the saint on 28 July (photo: Rolf Schrade)

Fig. 11 .S. Lagurka monastery. Wall painting of St Quiricus dying on the staircase of Diocletian's throne (1111) (photo: Rolf Schrade)

193

194

BRIGITTA SCHRADE

Fig. 11.9. Chased icon of Saint Quiricus in the treasury of Lagurka

monastery (first third of eleventh century) (photo: Rolf Schrade)

195

BYZANTIUM AND ITS EASTERN BARBARIANS

Fig. 11.10. Chased icon of Saint Barbara (eleventh century) from

st.

Quiricus (Lagurka) in Leli (now Mestia museum). The borders are adorned with representations of the deesis (above), the martyrs and military saints George and Procopios (left and right) and the physician saints Cosmas, Panteleimon and Damian (below).

Fig. 11.11. Painted icon of the Theotokos and Saint Barbara (tenth or eleventh century) from St George in Nakip'ari (now Mestia museum).

196

BRIG ITT A SCHRADE

Fig. 11.12. Wall painting of the lower church of the Saviour in Laghami, Mestia (tenth or eleventh century) with two male saints, probably Artemios and George, on the southern and Saint Barbara on the western wall (photo: Rolf Schrade)

Fig. 11.13. Silver head of the Svan flag Lern (thirteenth century), a donation of Grigol Kopasdze, prior of the monastery of St. George in Seti Mestia. The 'Archangel of Ugh(vi)ri' (Mt'avarangelozi ughrisa) as patron of the corresponding community is invoked together with 'Saint George of Seti' and 'Saint Jonas of Latali' as helper of whole Svanet'i (photo: Rolf Schrade)

BYZANTIUM AND ITS EASTERN BARBARIANS

Fig. 11.14. Chased icon of the Archangel Michael from the church of the

Archangel of Muxeri (thirteenth or fourteenth century) (photo: Rolf Schrade)

197

This page has been left blank intentionally

12. Georgian perceptions of Byzantium in the eleventh and twelfth centuries Giorgi Tcheishvili The attitude of Georgians towards Byzantium in the Middle Ages is a subject which has received scant scholarly attention,1 although Georgian sources contain a wealth of information regarding the study of the contemporary perceptions of Byzantium. The information for the eleventh and twelfth centuries is better than for any other period in the history of Georgian-Byzantine relations. The bulk of the information is derived from Georgian literary sources: hagiographies, chronicles, typika, charters, ecclesiastical acts, and panegyrics. Inscriptions, coins and wall paintings also fill gaps in our information and can add many valuable illustrations to it. The Georgian sources of the eleventh and twelfth centuries provide a vivid image of contemporary Byzantium and Byzantines. This image is constructed on comparative characteristics of the Byzantines and Georgians, as well as parallels from the Bible and Greek mythology. Though not always unbiased, the image illuminates the basic attitudes of Georgians towards Byzantines, who were the object of admiration as well as of criticism. However, it is important to distinguish between the ideal image of Byzantium as model of power and Orthodoxy, and that produced through direct dealings with the empire. In the Georgian sources, Byzantium is variously named as saberdznet'i (Greece), elada (Hellas), dasavlet'i (the west), and sameup'o (the empire); and Byzantines are called berdzenni (Greeks) and elinni (Hellenes). It can be seen from this that the Georgian terms designating Byzantium and Byzantines differ from Byzantine and eastern (Persian, Armenian, Syrian, Arab, Seljuq) terminology, and bear similarity to western European 1 G. Dragon, 'Contemporary Image and Influence of Byzantium', in K. Fledelius and P. Schreiner, eds, BYZANTIUM: Identity, Image, Influence (Copenhagen, 1996), 63-5, stressed the importance of this problem at the XIX International Congress of Byzantine Studies. From Eastern Approaches to Byzantium, ed. Antony Eastmond. Copyright © 2001 by the Society for the Promotion of Byzantine Studies. Published by Ashgate Publishing Ltd, Gower House, Croft Road, Aldershot, Hampshire, GU11 3HR, Great Britain.

199

200

GIORGI TCHEISHVILI

terms. However, in Georgia, unlike Europe, Greece/Byzantium and Greeks/Byzantines never gained a negative or pejorative meaning; indeed, the opposite was the case. From the ethnic name berdzeni (a Greek person) is derived the Georgian word brdzeni (a wise man, philosopher). Greek wisdom was always renowned among the Georgians, and Greek education remained an object of admiration even after Byzantium lost its political prestige. This attitude is well illustrated in The sermon on the Living Pillar' by the twelfth-century catholicos Nikoloz Gulaberisdze: T h e Greeks are proficient orators, philosophers and historians, and, like celestial creatures, they surpass other nations with the scope of their intellect'.2 The Byzantine empire was regarded with great prestige in Georgia, principally because of its position as the centre of Christian civilization. Byzantium was 'the most pious nation'3 and its subjects were 'the true Orthodox', who held the honour of being the Georgians' masters on issue of faith.4 In the Georgian sources Byzantium is defined as 'a country of pilgrimage'. Byzantium had been venerated as a country of pilgrimage for the Georgians from as early as the fifth century, when it had been encouraged by the activities of Peter the Iberian.5 This image acquired greater lustre from the end of the tenth century. In both hagiographic literature and in the royal chronicles of the Georgian court, Byzantium and its capital, Constantinople, were the subject of extravagant praise. They were the terrestrial Paradise;6 the Zion and Jerusalem, from where the law and the word of the Lord went forth.7 Those who sought ascen2 V. Karbelashvili, e