Dariali: The 'Caspian Gates' in the Caucasus from Antiquity to the Age of the Huns and the Middle Ages: The Joint Georgian-British Dariali Gorge Excavations and Surveys of 2013–2016

By Eberhard Sauer

The Huns, invading through Dariali Gorge on the modern-day border between Russia and Georgia in AD 395 and 515, spread terror across the late antique world. Was this the prelude to the apocalypse? Prophecies foresaw a future Hunnic onslaught, via the same mountain pass, bringing about the end of the world. Humanity’s fate depended on a gated barrier deep in Europe’s highest and most forbidding mountain chain.



Centuries before the emergence of such apocalyptic beliefs, the gorge had reached world fame. It was the target of a planned military expedition by the Emperor Nero. Chained to the dramatic sheer cliffs, framing the narrow passage, the mythical fire-thief Prometheus suffered severe punishment, his liver devoured by an eagle. It was known under multiple names, most commonly the Caspian or Alan Gates.



Featuring in the works of literary giants, no other mountain pass in the ancient and medieval world matches Dariali’s fame. Yet little was known about the materiality of this mythical place. A team of archaeologists has now shed much new light on the major gorge-blocking fort and a barrier wall on a steep rocky ridge further north. The walls still standing today were built around the time of the first major Hunnic invasion in the late fourth century – when the Caucasus defences feature increasingly prominently in negotiations between the Great Powers of Persia and Rome. In its endeavour to strongly fortify the strategic mountain pass through the Central Caucasus, the workforce erased most traces of earlier occupation. The Persian-built bastion saw heavy occupation for 600 years. Its multi-faith medieval garrison controlled Trans-Caucasian traffic. Everyday objects and human remains reveal harsh living conditions and close connections to the Muslim South, as well as the steppe world of the north. The Caspian Gates explains how a highly strategic rock has played a pivotal role in world history from Classical Antiquity into the twentieth century.

• Language: English
• Publisher: Oxbow Books
• Release date: Apr 30, 2020
• ISBN: 9781789251937

Eberhard Sauer

Eberhard W. Sauer is Professor of Roman Archaeology at the University of Edinburgh. Since 2012 he has been the Project Leader in the ERC Persia Project, which has involved fieldwork on Sasanian military monuments in Iran, Georgia and Oman. His many publications include Sasanian Persia between Rome and the Steppes of Eurasia (ed, 2017) and Persia’s Imperial Power in Late Antiquity: the Great Wall of Gorgān and Frontier Landscapes of Sasanian Iran (2013).

Book preview

DARIALI: THE ‘CASPIAN GATES’ IN THE CAUCASUS FROM ANTIQUITY TO THE AGE OF THE HUNS AND THE MIDDLE AGES

DARIALI: THE ‘CASPIAN GATES’ IN THE CAUCASUS FROM ANTIQUITY TO THE AGE OF THE HUNS AND THE MIDDLE AGES

THE JOINT GEORGIAN-BRITISH DARIALI GORGE EXCAVATIONS & SURVEYS OF 2013–2016

BRITISH INSTITUTE OF PERSIAN STUDIES ARCHAEOLOGICAL MONOGRAPHS SERIES VI, VOLUME 1

BRITISH INSTITUTE OF PERSIAN STUDIES ARCHAEOLOGICAL MONOGRAPHS SERIES VI, VOLUME 2

Generously supported by the ERC Persia and its Neighbours Project

EBERHARD W. SAUER, LANA CHOLOGAURI, ANA GABUNIA, KRISTEN HOPPER, DAN LAWRENCE, EVE MACDONALD, MARJAN MASHKOUR, FIONA A. MOWAT, DAVIT NASKIDASHVILI, KONSTANTIN PITSKHELAURI, SETH M.N. PRIESTMAN, LYUDMILA SHUMILOVSKIKH, ST JOHN SIMPSON, ANTHI TILIAKOU, YOSHINARI ABE, CATHY M. BATT, GEORGE GAGOSHIDZE, BRIAN GILMOUR, DAVID P. GREENWOOD, CATRIONA PICKARD, IMOGEN POOLE, VALENTIN RADU, LISA SNAPE, J. RILEY SNYDER, EMMANUELLE STOETZEL, SARIEH AMIRIBEIRAMI, MARTINA ASTOLFI, IAN BAILIFF, SANAZ BEIZAEE DOOST, ENRICA BONATO, BENOÎT CLAVEL, KARYNE DEBUE, DELPHINE DECRUYENAERE, ANNAMARIA DIANA, HOMA FATHI, RUSUDAN JAJANIDZE, SAFOORA KAMJAN, TEHREEM KAINAAT, ROYA KHAZAELI, ELENA KRANIOTI, MICHEL LEMOINE, GRAHAM PHILIP, CATHERINE SHUPE, RYUJI SHIKAKU, BEN STERN & SCOTT STETKIEWICZ

Published in the United Kingdom in 2020 by

OXBOW BOOKS

The Old Music Hall, 106–108 Cowley Road, Oxford, OX4 1JE

and in the United States by

OXBOW BOOKS

1950 Lawrence Road, Havertown, PA 19083

© Oxbow Books and the individual authors 2020

Hardcover Edition: ISBN 978-1-78925-192-0

Digital Edition: ISBN 978-1-78925-193-7 (epub)

Kindle Edition: ISBN 978-1-78925-194-4 (mobi)

A CIP record for this book is available from the British Library

Library of Congress Control Number: 2019954229

All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopying, recording or by any information storage and retrieval system, without permission from the publisher in writing.

For a complete list of Oxbow titles, please contact:

UNITED KINGDOM

Oxbow Books

Telephone [phone]

Email: [e-mail]

www.oxbowbooks.com

UNITED STATES OF AMERICA

Oxbow Books

Telephone [phone], Fax [phone]

Email: [e-mail]

www.casemateacademic.com/oxbow

Oxbow Books is part of the Casemate Group

Front cover: Dariali Fort dominating the gorge and controlling traffic across Europe’s highest mountain range.

Laser scanning by Malkhaz Datukishvili & Giorgi Kvaratskhelia

Finds conservation by Nino Kamkamidze & artefact drawings by Emzar Lomtadze

Major contributions to fieldwork and recording and analysis of contexts, structures and finds by Graeme Erskine, David Gagoshidze, Ketevan Gotsiridze, Emanuele Intagliata, Koba Koberidze, Steve Usher-Wilson, Mark Andrews, Maria Buczak, Marc Heise, Gabriela Ingle, Julian Jansen Van Rensburg & Przemysław Polakiewicz

Digitisation of excavation plans and sections by Silvia Perini, Graeme Erskine, Annamaria Diana, Emanule Intagliata & Julian Jansen Van Rensburg, in collaboration with Eberhard W. Sauer. Plates by Silvia Perini, Rosalind MacDonald, Seth M.N. Priestman & Fiona A. Mowat

Contents

Volume 1

Acknowledgements

Section A: Preliminaries

Chapter 1. Introduction

1.1. Summary

1.2. Geography

1.3. Site chronology

1.4. Ethnic/political terms

1.5. Technical notes

Section B: Excavations and survey

Chapter 2. Late antique buildings occupied to the Late Middle Ages: life over one millennium on Dariali Fort (Trench F)

2.1. Introduction

2.2. The late antique fort walls and intramural life (late fourth to mid-seventh centuries AD/Phase 3)

2.3. Mid-seventh/early eighth-century power vacuum (Phase 4)

2.4. The early medieval zenith of activity (eighth to tenth/eleventh centuries AD/Phase 5)

2.5. The late medieval Georgian castle and renewed activity between the thirteenth to fifteenth centuries AD (Phase 6b)

2.6. Gunmen’s backyard: modern activity in the west of the fort (Phase 7b)

Chapter 3. Towering over the northern approaches: late antique buildings, medieval food storage and modern military (Trench Q)

3.1. Introduction

3.2. The late antique fort walls and intramural life (late fourth to mid-seventh centuries AD/Phase 3)

3.3. The seventh to early eighth century: plus ça change, plus c’est la même chose: political upheaval and archaeological continuity (Phase 4)

3.4. Resurgence: dry-stone building boom and life from the eighth to the tenth/eleventh centuries (Phase 5)

3.5. A derelict backyard in the medieval castle (Phases 5b–6b)

3.6. Attempting to halt Soviet aggression (Phase 7b)

Chapter 4. Barrier, bastion and aqueduct: sondages and surveys on and around Dariali Fort (Trenches L, X and O)

4.1. A tower at the southern approaches to the Dariali Fort: Trench L (Phases 3–5)

4.2. Buried beyond reach: in search of the earliest occupation debris at the base of Dariali Fort’s steep western cliffs: Trench X (Phases 5–7)

4.3. A modern military shelter: Trench O (Phase 7b)

4.4. A gated road-blocking wall west of the fort (Phases 3a–5b/c?)

4.5. Vital water supply: piped water from the mountains above and staircase to the river below (Phase 3a–d?)

Chapter 5. Extramural areas south of the fort: two-and-a-half millennia of traffic and two millennia of food production in the shadow of the rock (Trenches P and M)

5.1. Trench P: 2,500-year-old campfires to medieval gardening and modern shelters at the base of the tower-crowned rock (Phases 2a–7b)

5.2. Extramural food production throughout the first millennium: Trench M (Phases 2b–5)

Chapter 6. Dariali early medieval cemetery (Trenches E, G and AB)

Eberhard W. Sauer, Anthi Tiliakou, Catherine Shupe, Annamaria Diana, Elena Kranioti and Konstantin Pitskhelauri

6.1. Introduction

6.2. Garden plots or arable fields? Pre-cemetery activity in the area of Trenches E, G and AB (Phases 2/3–5)

6.3. Trench E: collective graves of the first phase (5a–b) of the cemetery

6.4. Trench G: individual adult, adolescent and child graves of the cemetery’s later phase (5b–c)

6.5. Trench AB: individual child graves of the cemetery’s later phase (5c)

6.6. Burials beyond our trenches (Phase 5b–c)

6.7. The chronology of the cemetery (Phase 5)

6.8. The early medieval cemetery and the religion of the garrison (Phase 5)

6.9. The early medieval cemetery, the ethnicity of the garrison and its connections to the outside world (Phase 5)

Chapter 7. The Caspian Gates? Bakht’ari fortified ridge: first line of defence and northernmost barrier (Trench Y/Phase 3)

Chapter 8. Medieval Gveleti Fort: valley-blocking cliff-top bastion and royal refuge from the Mongols (Trenches C, D, N, U, V and W)

8.1. Introduction

8.2. Living on top of the cliff-edge in early modern times: Trench C (Phase 7a)

8.3. Trench D: early medieval to early modern occupation of Gveleti Fort (Phases 5–7a)

8.4. Trench N: early modern housing next to the lower fort’s stone tower (Phase 7a)

8.5. Trench U: activity at the approaches to the upper fort in the era of royal residency (Phase 6b)

8.6. Trench V: medieval occupation south-west of the church (Phase 6a)

8.7. Trench W: a late medieval stone house (Phase 6b)?

8.8. The history of Gveleti Fort (Phases 5–7)

Chapter 9. Elusive migration-era burials and enigmatic stone cairns: fieldwork near Gveleti Cemetery and in the Amali Valley (Trenches A, B, H, I, J, K, R, S, T, Z and AA)

9.1. In search of Gveleti Cemetery (Phases 3–7)

9.2. The enigmatic Amali Cairns and the power of nature: Trenches Z and AA

Chapter 10. Landscape investigations in the Dariali Pass

Kristen Hopper, Dan Lawrence, Lisa Snape, Lana Chologauri, Seth M.N. Priestman, Lyudmila Shumilovskikh, Konstantin Pitskhelauri and Graham Philip

10.1. Introduction

10.2. Landscape investigations in Khevi

10.3. The landscape survey in Khevi

10.4. Archaeological and historical landscapes of Khevi

10.5. Terrace field systems

10.6. Discussion

10.7. Dariali Pass Survey Site Gazetteer

10.8. Sedimentary descriptions from terrace field investigations

Volume 2

Acknowledgements

Section C: Specialist contributions: finds, building materials, biological and environmental evidence and scientific dating

Chapter 11. Provisioning and supply across an ancient frontier: the late antique and medieval ceramic sequence from the Dariali Gorge in the High Caucasus

Seth M.N. Priestman

11.1. Introduction

11.2. Phase distribution and changing assemblage composition

11.3. Chronological development of the assemblage

11.4. Other changes in assemblage composition

11.5. Other assemblages

11.6. Discussion

11.7. Class catalogue

11.8. Vessel types

11.9. Petrographic analysis and raw material provenance

Enrica Bonato and Seth M.N. Priestman

11.10. Residue analysis of cooking pots by GCMS

Ben Stern and Seth M.N. Priestman

11.11. Context dating from ceramic finds

Chapter 12. Fragment of a ceramic vessel with an ancient Georgian inscription discovered at Dariali Fort

George Gagoshidze

Chapter 13. Vessel glass from the Dariali Fort

Fiona Anne Mowat

13.1. The assemblage from Dariali

13.2. Methodology

13.3. Fabric classes

13.4. Chemical analysis

13.5. Site phasing and vessel circulation

13.6. Glass recycling, cullet and trade at Dariali

13.7. Descriptive catalogue by fabric type

13.8. Conclusion

Chapter 14. Report of chemical compositional characterisation of glass fragments excavated from Dariali Fort (Georgia) by non-destructive X-ray fluorescence analysis

Yoshinari Abe and Ryuji Shikaku

14.1. Materials and methods

14.2. Results and discussion

14.3. Conclusion

Chapter 15. The small objects and other finds

Lana Chologauri, Ana Gabunia, Fiona Anne Mowat, Seth M.N. Priestman, Eberhard W. Sauer and St John Simpson, with an appendix by Scott Stetkiewicz

15.1. Introduction

15.2. Finds from the Dariali and Gveleti Forts and extramural areas

15.3. Finds from the cemetery

15.4. Conclusion

15.5. Appendix: Slag from the Dariali Fort

Scott Stetkiewicz

Chapter 16. The sword from Grave G9 in the cemetery south of Dariali Fort: analytical and technological study and assessment

Brian Gilmour

16.1. Introduction and preliminary description

16.2. Analysis and technology of the sword blade

16.3. Discussion and conclusion

Chapter 17. Ceramic building materials from Dariali Fort

Seth M.N. Priestman

Chapter 18. Mortars from Dariali Fort and nearby fortifications

J. Riley Snyder and Martina Astolfi

18.1. Introduction

18.2. Materials and methods

18.3. Results

18.4. Discussion

18.5. Conclusion

Chapter 19. Human skeletal remains

Anthi Tiliakou, Catherine Shupe, Elena Kranioti and Annamaria Diana

19.1. Introduction

19.2. Methodology

19.3. Taphonomy and state of preservation

19.4. Biological sex, age, ancestry and stature

19.5. Non-metric traits

19.6. Dentition

19.7. Dental attrition

19.8. Palaeopathology

19.9. Dental Enamel Hypoplasia (DEH)

19.10. Caries, Dental calculus, Periodontitis, Periapical cavities and Ante-mortem Tooth loss

19.11. Osteoarthritis (OA)

19.12. Periostitis/Chronic Periostitis

19.13. Entheseal changes

19.14. Trauma

19.15. Conclusion

Chapter 20. Dariali Cemetery stable isotope analysis

Catriona Pickard

20.1. Isotope analysis

20.2. Materials and methods

20.3. Results and discussion

20.4. Conclusion

Chapter 21. Herding and hunting in the highlands from the Sasanian to late medieval periods

21.1. The archaeozoology of the Dariali Gorge

Marjan Mashkour, Sarieh Amiri, Homa Fathi, Roya Khazaeli, Karyne Debue, Delphine Decruyenaere, Sanaz Beizaee Doost, Benoît Clavel, Safoora Kamjan, Rusudan Jajanidze and Eberhard W. Sauer

21.2. Fish remains

Valentin Radu and Eberhard W. Sauer

21.3. Microvertebrates

Emmanuelle Stoetzel, Lyudmila Shumilovskikh, Karyne Debue, Michel Lemoine and Marjan Mashkour

Chapter 22. Plant remains

Lyudmila Shumilovskikh and Imogen Poole

22.1. Charcoal

Imogen Poole and Lyudmila Shumilovskikh

22.2. Archaeobotany

Lyudmila Shumilovskikh

22.3. Vegetation history

Lyudmila Shumilovskikh

Chapter 23. Archaeomagnetic studies of features excavated in Dariali Gorge

Cathy M. Batt, David P. Greenwood and Tehreem Kainaat

23.1. Abstract

23.2. Introduction

23.3. Background

23.4. Archaeomagnetic sampling in the field

23.5. Archaeomagnetic measurements

23.6. Results

23.7. Interpretation of the archaeomagnetic results and comparison with the global geomagnetic field model

23.8. Previous archaeomagnetic studies in the region

23.9. Summary and conclusion

23.10. Further work

Chapter 24. Luminescence dating and micromorphological assessment

Lisa Snape and Ian Bailiff

24.1. Introduction

24.2. Field observations

24.3. Discussion

Section D: History

Chapter 25. The history of the Dariali Gorge

25.1. Prehistoric colonisation, Cimmerian invasion and the earliest visits to Dariali Rock (third millennium to third century BC/up to Phase 2a)

25.2. From obscurity to world-fame: Iberians, Iranians and Romans in the Dariali Gorge (Phase 2b: second century BC to fourth century AD)

25.3. The late antique fort (Phase 3)

25.4. Power vacuum: mid-seventh century collapse of Sasanian rule to eighth-century Islamic conquest (Phase 4)

25.5. The gates and their garrison in the Early Middle Ages (Phase 5)

25.6. Abandonment and resurgence: Dariali and Gveleti Forts in the High and Late Middle Ages (Phase 6a–b: eleventh to late fourteenth or fifteenth centuries)

25.7. End of the Middle Ages to Soviet Invasion: Dariali Gorge in the second half of the First Millennium (Phase 7)

Section E: Appendices and Conclusion

Appendices. Landslides, the location of the gates and imperial landscapes: notes on historical geography

IA hostile environment: landslides and their effect on settlement patterns in the gorge

II Where were the gates? A French eyewitness to the narrowness of the gorge

III Investigations of ancient canal systems in Central and Eastern Georgia

Kristen Hopper, Dan Lawrence, Konstantin Pitskhelauri and Graham Philip

Conclusion

Acknowledgements & the history of the Dariali Gorge (in Georgian)/წინასიტყვაობა და დარიალის ხეობის ისტორია ანტიკურ ხანაში Abridged and translated by Davit Naskidashvili/დავით ნასყიდაშვილი

Bibliography

We dedicate this book to Professor Vakhtang Licheli, Dr Manana Odisheli and Professor Michael Vickers, without whose kind support this project would not have been possible.

fig. 1:1: Professor Michael Vickers and Dr Manana Odisheli on Dariali Fort in 2013.

fig. 1:2: Professor Vakhtang Licheli in Trench F in 2014.

Acknowledgements

Our fieldwork in the Dariali Gorge, jointly carried out by Ivane Javakhishvili Tbilisi State University, the Universities of Edinburgh and Durham and our partners at the University of Bradford and the CNRS at Paris, has been generously funded by the European Research Council (ERC), as part of our ‘Persia and its Neighbours’ project. We are most grateful to the ERC for having enabled us to venture to a site which epitomises, perhaps more than any other, the contact and conflict of Persia with its most powerful neighbours in the north: the Alan, Caspian or Caucasian Gates¹ – arguably the most famous fortification on Persia’s northern frontiers.

It would not have been possible to accomplish, or even commence, this project without the kind help of many scholars and colleagues. We are indebted to Professor Michael Vickers and Dr Manana Odisheli for their crucial support from the start. When the first author conceived the idea of excavations at the ‘Caspian Gates’ just a few weeks before a journey intended to explore fieldwork opportunities in Azerbaijan in summer 2012, he knew nobody in Georgia except for Professor Vickers – and without his advice would not have succeeded in putting this idea into practice, let alone within less than a year. On 28 August 2012 Professor Vickers, Dr Odisheli, Davit Naskidashvili, Lika Sekhniashvili and Sophie Zhghenti took Eberhard Sauer on a journey to the Dariali Gorge – a one-day reconnaissance mission, into the lay of the land and access arrangements in borderland territory, that paved the way for four seasons of fieldwork. In addition to their advice on a wide range of academic and practical issues, we are also very grateful to Professor Vickers and Dr Odisheli for their wonderful hospitality whilst in Georgia. It was also in late August 2012 that they put Eberhard Sauer in touch with Professor Vakhtang Licheli, who subsequently kindly secured an agreement of understanding and a permit and offered much help with practical and academic matters. Professor Licheli also liaised with colleagues and drew recent archaeological work to our attention.

The late Iveta Naskidashvili very kindly accommodated members of the team on many occasions between 2013 and 2017 and always made us feel very welcome. Her warm-hearted support for our project and our team contributed much to its success, she will be sorely missed by all of us, and we will always fondly remember her.

The Persia project may never have started and certainly would have been less successful without the late Professor Tony Wilkinson, the co-PI on the Persia project, who offered much support for, and advice on, our work in Dariali, if sadly not living to see its completion.

Dr Davit Mindorashvili directed and published a pioneering fieldwork project at Dariali Fort and Dariali and Gveleti Cemeteries,² kindly shared his unique expertise with us when taking us on a guided tour of these sites on 17 April 2013 and subsequently offered support throughout the project. He also helped us find parallels for artefacts, notably those of bone, and offered advice on the ceramic assemblage. We are indebted to Iago Kazalikashvili, the director of Stepantsminda Museum, for sharing his knowledge on the location of the Gveleti and Gigias Satibi Cemeteries with us. Professor David Braund has also kindly offered important advice when the project was still at planning stage.

We are indebted to our reviewers, Dr James Howard-Johnston and Professor Richard Payne, for their invaluable advice on many aspects of this long report that has substantially improved it, whilst we are responsible for the imperfections that remain. We are very grateful to Dr Darejan Katcharava for her kind comments on the Georgian version of the historical conclusions. Dr David Reich accepted samples from Dariali Cemetery for DNA analysis, to be published separately once processed. We are grateful to Dr Irina Shingiray for initiating this and to Professor Elio Brancaforte for safely bringing the samples to America. Dr Shingiray also alerted us to several important Russian and Azeri publications. Professor Dmitry S. Korobov (at a workshop in Oxford organised by Dr Shingiray) and Nick Evans kindly drew our attention to the important Kasar Wall. Professor Korobov also alerted us to the Khilaki Wall, that may be highly significant for the broader context of the Caucasian defensive system, and sent us key publications as well as relevant photos, cartographic material and information. Dr Denis Beletskiĭ provided information on a small church at Gveleti. Warwick Ball came to visit in 2014 and drew our attention to important paintings of the gorge by Pierre Blanchard, which we were subsequently able to reproduce with the kind permission of Dadiani Palaces History and Architectural Museum. Ana Gabunia and Lana Chologauri liaised with the museum. The late Professor Edmund Bosworth kindly responded to an enquiry about the Islamic sources on the Dariali Gorge and Dr Tim Greenwood alerted us to an important passage in the work of the Armenian historian Łazar. We are also grateful to Dr Ian Colvin for his academic advice. Stephen Copp kindly advised us on techniques to create a 3D terrain model of the site, later accomplished by Dr Malkhaz Datukishvili and Giorgi Kvaratskhelia. Professor R.R.R. Smith and Dr Julia Lenaghan provided a photo of the Prometheus relief from Aphrodisias, Dr Friederike Naumann-Steckner (Römisch-Germanisches Museum, Cologne) kindly granted permission to reproduce a photo of the Mansuetus inscription from Cologne.

To Nino Kamkamidze, we are very grateful for her professional conservation of the sword and numerous other finds and her reconstruction of ceramic vessels. Emzar Lomtadze kindly accomplished drawings of our finds and arranged shoring within our trenches. The numerous plans and sections of our trenches have been digitised by Dr Silvia Perini, Graeme Erskine, Dr Annamaria Diana, Dr Emanuele Intagliata and Dr Julian Jansen Van Rensburg. Dr Perini in particular deserves credit for the final versions of all the plans and sections included in the volume. Dr Perini, Rosalind MacDonald, Dr Seth Priestman and Dr Fiona Mowat accomplished plates of the finds.

Numerous libraries, notably at Edinburgh, Oxford and London, were consulted and we are grateful to their staff. EWS would also like to thank Aldona Judina for Russian lessons which proved immensely useful. We are indebted to Dr Cameron Petrie of the British Institute of Persian Studies and Cambridge University and to Mette Bundgaard, Dr Julie Gardiner, Declan Ingram, Clare Litt and Jessica Scott of Oxbow for their kind efforts in seeing the book through the press and all their editorial efforts, support and advice. Dr Sacha Jones kindly copy-edited chapters 21.1 and 21.3.

Many colleagues in Edinburgh University have offered much support, notably Bill Bruce, Professor Ewen Cameron, Fiona Carmichael, Anna Gibbons, Lindsay Hampton, Karen Howie, Professor Alvin Jackson, Professor Gavin Kelly, Stephanie Knox, Susan McIntosh, Lesley McLean, Jennifer Mills, Angela Noble, Brian Pacey, Dr Joanne Rowland, Dr Ulrike Roth and Louise Todd.

Being part of our greater Persia and its Neighbours Project, we are also immensely grateful to our colleagues in Iran, Dr Jebrael Nokandeh, Hamid Omrani Rekavandi, Dr Mohammad-Reza Nemati, Bardia Shabani and Professor Meysam Labbaf-Khaniki, as well as to our partners in Oman, Professor Nasser Said al-Jahwari and Dr Derek Kennet. Examining Sasanian frontiers in detail in three separate territories has helped us to understand how the system worked as a whole.

Otar Tsamalaidze, the director of Kazbegi National Park, very kindly granted permission to work within the National Park and supported our project. We are very grateful to the Patrol Police and Border Police departments of the Ministry of Internal Affairs as well as to border patrols on duty for facilitating our fieldwork at Dariali Fort and the visits of academic colleagues on many occasions. The monks of the Monastery of the Archangels St Michael and St Gabriel and the nuns in the newly-built nunnery kindly offered help, advice and hospitality throughout our project. We are indebted to our excellent cooks, Irina Kushashvili and Bibino Tsiklauri, our host, Stephane Pitskhelauri, our driver Viktor Avsajanishvili, our housekeeper Nelly Khaikashvili, our workmen and, most importantly, to our talented and extremely hard-working team, accomplishing amazing work in difficult circumstances, undeterred by challenging weather conditions and exposed to torrential downpours on many occasions. We provide a list of the team below.

Dariali team members, taking part in fieldwork and/ or finds processing in Georgia (Figs 1:3–1:6): Sarieh Amiribeirami (2015–2016), Mark Andrews (2016), Giulia Bellato (2016), Maria Buczak (2015–2016), Tika Bulashvili (2016), Carlos Cáceres Puerto (2016), Lana Chologauri (2013–2016), Malkhaz Datukishvili (2014– 2015), Karyne Debue (2016), Delphine Decruyenaere (2015–2016), Annamaria Diana (2013), Graeme Erskine (2013–2015), Homa Fathi (2014–2016), Ana Gabunia (2013–2016), David Gagoshidze (2013 and 2016), Ketevan Gotsiridze (2013–2015), David Greenwood (2013–2016), Marc Heise (2014), Kristen Hopper (2013–2014), Gabriela Ingle (2013 and 2016), Emanuele Intagliata (2013 and 2015), Erin Irwin (2016), Rusudan Jajanidze (2016), Julian Jansen Van Rensburg (2013), Stuart Johnston (2015), Nino Kamkamidze (2014– 2016), Roya Khazaeli (2014–2016), Koba Koberidze (2013 and 2015–2016), Elena Kranioti (2013), Giorgi Kvaratskhelia (2014–2015), Dan Lawrence (2013–2014), Zurab Lemondzhava (2014–2015), Emzar Lomtadze (2014–2016), Eve MacDonald (2013–2016), Rosalind MacDonald (2015–2016), Marjan Mashkour (2014 and 2016), Alexandre Mesarkishvili (2015–2016), Anahita Mittertrainer (2015), Fiona Mowat (2013–2016), Amiran Nadirashvili (2015), Davit Naskidashvili (2015–2016), Claudia Nunes Caldeira (2015), Konstantin Pitskhelauri (2013–2016), Przemysław Polakiewicz (2016), Seth Priestman (2013–2016), Graham Ritchie (2016), Kim Ruf (2014–2015), Rusudan Rusishvili (2015–2016), Eberhard Sauer (2013–2016), Lyudmila Shumilovskikh (2014–2016), Catherine Shupe (2014 and 2016), Lisa Snape (2014–2015), Scott Stetkiewicz (2014–2015), Leigh Stork (2014), Anthi Tiliakou (2013–2014) and Steve Usher-Wilson (2013–2014 and 2016).

Find assistants and workmen: Niko Alibegashvili (2016), Vasiko Alibegashvili (2016), Vasil Alibegashvili (2014–2015), Vazha Alibegashvili I (2014–2016), Vazha Alibegashvili II (2015–2016), Shavlego Arabuli (2015), Viktor Avsajanishvili (2013–2016), Andro Burduli (2015), Lamara Chabukashvili (2016), Irakli Chopikashvili (2014), Pavle Chopikashvili (2014), Fido Efkhoshvili (2013), Vakho Goshiashvili (2015), Nika Kavtarashvili (2015), Shavlego Kavtarashvili (2016), Oliko Khaikashvili (2016), Bachuki Kukishvili (2014), Archili Kushashvili (2014), Bachuki Kushashvili (2013), Davit Kushashvili (2015), Edisheri Kushashvili (2014–2015), Gega Kushashvili (2014), Giorgi Kushashvili (2013), Gogi Kushashvili (2013–2015), Lukhumi Kushashvili (2013–2014), Rezo Kushashvili (2013), Gvtiso Marsagishvili (2014–2016), Tornike Naneishvili (2015), Gela Pitskhelauri (2014), Giorgi Pitskhelauri (2013), Gogi Pitskhelauri (2013– 2015), Juba Pitskhelauri (2015), Manuchar Pitskhelauri (2014–2015), Vazha Pitskhelauri (2015–2016), Gocha Tekturmanidze (2013–2016) and Jura Valishvili (2013).

fig. 1:3: Dariali team members, 2013.

fig. 1:4: Dariali team members, 2014 at Trench F.

fig. 1:5: Dariali team members, 2015 (photo by Lana Chologauri).

fig. 1:6: Dariali team members, 2016. All photographs feature only part of the team in each season, as some could come for parts of the season only, and some were busy elsewhere when the group photos were taken.

Notes

1In classical Greek, Latin and modern English, we often find the plural ( Πύλαι , Portae and Gates) used for a single gated entrance or defensible narrow mountain pass resembling a gate in function or even shape, such as the Dariali Gorge, in part explained by the singular of the Greek word ( πύλη ) being just a single wing of a double-winged gate (Liddell and Scott 1940, 1553–54; Buchwald 2009, 6–7 no. II.A). In Arabic, the word Bāb (باب) tends to be used for a gate or passage, unlike the Greek and Latin terms normally translated with the singular, ‘Gate’. To avoid confusion, we use consistently the plural in this report, i.e. the Alan, Caspian or Caucasian Gates.

2Mindorashvili 2005; information on these important investigations was kindly provided by Professor Vakhtang Licheli and members of our Georgian team, who also summarised key sections of the report.

SECTION A

P

RELIMINARIES

1

Introduction

1.1. Summary

Dariali Gorge was the most famous mountain valley in the ancient and early medieval world. The principal route across the Central Caucasus, Europe’s highest mountain range, led through the gorge, connecting the Steppes of Eurasia with Transcaucasia and the wider Near East. The narrow gorge, known in Antiquity as the ‘Caspian Gates’ or ‘Alan Gates’, resembled a natural gate and was controlled via gated barriers. It shared the name ‘Caspian Gates’ with the coastal Derbent route on the Caspian Sea shore and a pass in the Alborz Mountains.¹ Whilst one sympathises with Pliny the Elder who thought it was a great error to apply the name to Dariali, more appropriately labelled the ‘Caucasian Gates’ or the ‘Gates of Hiberia’ (i.e. Iberia),² the geographer’s advice fell on deaf ears. In classical literature, Dariali Gorge is mostly called the ‘Caspian Gates’. Indeed, where it is clear which northern mountain pass authors of the first millennium AD referring to the ‘Caspian Gates’ have in mind, it was more frequently Dariali than the two competitor sites further east. Locally, and in the Near East more broadly, the ‘Alan Gates’ gained currency, still the name of the gorge today: Dar-i-al, i.e. ‘Door of the Alans’. There were other toponyms too,³ but we are using in our title the two under which it reached world fame: the ‘Caspian Gates’ and ‘Dariali’.⁴

The riches of the south attracted numerous northern invaders to our Transcaucasian traffic route, the Cimmerians, Alans, Huns and Khazars – as well as the armies of Tsarist and Communist Russia and Hitler’s Germany in modern times. The major powers of the south, Rome, Sasanian Persia and the Caliphate, similarly strived for direct or indirect control and launched military expeditions deep into hostile alpine terrain. Little wonder that it came to the attention of more writers than any other mountain valley in the first millennium. Described by the tenth-century Arab author Mas’udi as one of the most famous forts of the world, Dariali Fort and the associated gates had already been known to Tacitus, Ptolemy, Procopius and numerous other classical and medieval authors. It also features on inscriptions in the royal heartlands of Sasanian Iran.

So dramatic was the landscape of the gorge that it became a place of legend. The mythical fire thief Prometheus was believed to have suffered excruciating punishment, chained to its sheer cliffs, his liver devoured by an eagle. Invasions by the Huns via the gorge spread terror across the ancient Near East, and the belief arose that it was Alexander the Great who had built the famous gates, acting as the agent of God to contain the Hunnic threat to the very existence of humanity. Apocalyptic literature envisaged a future breach of these floodgates by the Huns, their onslaught on civilisation bringing the world to an end. The gorge featured famously and prominently also in late antique negotiations between the great powers of Persia and Rome.

The gorge’s unrivalled literary fame contrasts sharply with its archaeological obscurity. A small number of articles and Davit Mindorashvili’s book on his pioneering fieldwork in the gorge,⁵ all in Georgian or Russian, are devoted to the physical remains of the world-famous gates – and there is the odd passing reference and short description here and there. It seemed vital to us to explore what archaeology could contribute to its history. When was the gorge blocked via gated barriers? Where were they? How were they guarded? What role did they play in wider imperial Persian strategy to use mountains and barriers to safeguard internal prosperity and to use northerners as allies against Rome?

Our fieldwork (from 2013 to 2016 within the framework of the ERC Persia Project)⁶ gave us more than we bargained for. Dariali’s strategic significance lasted well beyond the end of Antiquity, and our research has arguably shed as much light on medieval Dariali as on late antique defensive efforts. But it is only in the longue durée that we can understand a gorge that has been a borderland over most of the last two millennia and more. Twentieth-century gun emplacements and fourth-century fort walls served similar purposes – to exercise a stranglehold over unwanted transgression and to control human traffic and trade. The border of Georgia is today where the border of Iberia had been 2,000 years ago – and where the border of Sasanian Persia, the Caliphate and the Emirate of Tbilisi had been. Modern geography, explored by our landscape team, helps us to understand the challenges involved in supplying the garrison with food one-and-a-half millennia ago.

fig. 1:7: Now in ruins and partially buried under stone collapse, travellers or hostile forces approaching Dariali Fort from the north, faced a once imposing barrier-wall running from the fort (left) to a rock cliff.

Pollen records show that the gorge was populated already before the time of the earliest recorded invasions in the eighth century BC – shedding light on how expansion of settlement into alpine valleys may have paved the way for the earliest Transcaucasian traffic. The hinterland of the gorge was, it seems, as densely settled as natural constraints allowed for much of the later first millennium BC and the first millennium AD – centuries before it had reached its greatest fame, and throughout its time in the international limelight. Around the late fourth century, construction started on the late antique fort – that was perhaps largely completed in the early fifth century. Around that time also a road-blocking wall was built to the west of the fort and a further barrier, of similar architecture, on a precipitous ridge one kilometre north of the fort – then perhaps the narrowest part of the gorge and the famed Caspian Gates. This fits well with the sources attesting Persian fort construction in the later fourth or fifth century, and we have been able to refine its chronology. Construction works may have started shortly before, and in any case not long after, the most momentous of all Hunnic raids into Near Eastern lands. It was via Dariali Gorge that the Huns invaded in AD 395, penetrating deep into Asia Minor, Mesopotamia and Levantine lands. Completion of the Dariali barrier (fig. 1:7) ensured there was no repetition of destructive raiding and may have opened the door to Persia employing its new northern allies in proxy wars against Rome.

fig. 1:8: The outer fort walls were built in Late Antiquity.

fig. 1:9: Approximate extent of the Sasanian Empire in the AD 470s, if some of the sites and monuments plotted are later. Our ERC project also explored, jointly with colleagues in Iran and Oman, the Gorgan and Tammisheh Walls, fortifications in their hinterland and Qal’eh Iraj in Iran and the Fulayj Fort in Oman.

The late antique architects built a bulwark in durable, high-quality mortar masonry (fig. 1:8) that stood for more than a millennium with only dry-stone walls added during the next six centuries of dense occupation. The thoroughness of the late antique workforce erased all potential traces of earlier occupation on the hilltop, with no more than plough soils south of the fort hinting at earlier occupation. The garrison was tidy, with little intramural rubbish disposal and no precious objects left behind. Pottery and food were brought to the fort from the south – but there were also strong northern/local connections. Politically, the garrison reportedly served southern masters in Late Antiquity, and for a period in the late fifth/early sixth century also northern warlords, but economically it appears to have been well integrated into a Transcaucasian supply network. No early burials have been found, perhaps pointing to Zoroastrians manning the gates. A small cemetery at Gveleti, some 2.5 km south, with a Sasanian coin and Pahlavi-inscribed gem set in a fingering, may have been for an armed allied force.

Efforts were made to seal off or control other Transcaucasian traffic routes too, and Persia’s demands for, literally, tons of Roman gold to secure the Caspian Gates and keep the Huns at bay were excessive for Dariali alone, if a drop in the ocean for the system of Sasanian northern defences as a whole. Probably dating back to the late fourth century, Dariali Fort was one of the earliest elements of a massive fortification programme in the Sasanian Empire’s northern frontier zone, completed by the sixth century (fig. 1:9).⁷ Much grander barriers included the Ghilghilchay, Derbent, Tammisheh and Gorgan Walls and perhaps also the K‘lisura Wall, not to mention numerous forts and fortresses all over the Persian realm. By around the mid-fifth century, Sasanian military infrastructure dwarfed that of the Eastern Roman Empire. The combined size of Sasanian military compounds exceeded that of Dariali Fort several thousand times. Yet, skilfully positioned in a mountain pass, it was vital in Persia’s endeavours to use the Caucasus as a natural barrier – the first empire to do so. None of the western gold reached Persian frontier troops, who were probably paid in kind, nor did Persia require western support for maintaining its northern defences. The main challenges lay in logistics and keeping the long supply lines secure – only possible with Iberian cooperation. Dariali Fort was a disproportionately small and cheap, if extraordinarily effective, part of Persia’s northern defences.

We do not know who held the fort after the mid-seventh century demise of the Sasanian state. Traceable building activity had come to a virtual standstill from the sixth to the eighth or ninth century – evidence for lesser investment and the durability of existing infrastructure. The same tidiness as in Late Antiquity prevailed amongst fort occupants also in the later seventh century – suggesting perhaps that descendants of the Sasanian garrison remained in residence and followed similar traditions. A sudden influx of northern-style ceramics may hint at political and economic realignment: no longer an outpost of a southern empire, the garrison sought new alliances.

A marked change in cultural behaviour follows. From around the early eighth century, for some 300 years, copious amounts of rubbish are dumped within the walled compound – pointing to new occupants following practises not seen before in the fort. They chose to live in the solid late antique buildings, gradually adding basic dry-stone annexes, perhaps to compensate for some basements becoming inhabitable – rubbish disposal necessitating repeated raising of floor surfaces to prevent infiltration of muddy water. It may also have been in the early eighth century (or in any case not long before or after) that a cemetery was established – also occupied for some 300 years.

Who were Dariali Fort’s new occupants in the later first millennium? The sources refer to a Muslim army conquering the fort in c. AD 727 and holding it for more than two centuries, into the AD 940s or beyond. Yet, archaeology paints a more complex picture than textual accounts and casts doubt that the occupants were exclusively Muslim or observed strict Muslim custom. We found probable wine jars, evidence for occasional pork consumption, a Georgian graffito featuring the Christian name Gogi (George) and collective burials strikingly resembling Christian family tombs from elsewhere in Georgia. Khazar-style bone carvings featuring animals are likely evidence for the presence of northerners living in the shadow of the fort walls. Pottery supplies come from the north or local sources as well as the south. This does not mean that the sources are lying. Glass, a rarity in pre-medieval times, is now flooding in from Islamic lands and traded beyond the gates. Other rarities include fragments of a chlorite cooking pot probably from the Arabian Peninsula and a polished shark vertebra perhaps from the Persian Gulf. There are individual burials, the heads turned in the direction of Mecca. Those found are not amongst the earliest tombs, but not all the cemetery has been excavated – and digging of probably Muslim burials continues at least until the later tenth century. Towards the turn of the millennium, graves with rich northern grave goods, one with a sword and three arrowheads, another with numerous beads and jewellery, mark the arrival of northerners practising distinctly non-Islamic traditions. Also around AD 1000, a man dies of a fatal sabre(?) attack.

Those buried in the early medieval cemetery (many not living beyond childhood and some surviving into their twenties or thirties or a little beyond) often suffered periods of stress or malnutrition – no doubt a result of supply problems at a site often cut off from the outside world through snow, ice, mud avalanches and rockfall. Occasionally, they enjoyed food from the south. The author Mas’udi, writing in the 940s, attests supply convoys from Islamic Tbilisi, venturing through infidel lands to the fort. His testimony is corroborated by plant remains and isotope analysis, the latter suggesting that southern supplies did not cease even after northerners had taken up residence at the fort.

The evidence is complex, but it seems that early medieval Dariali had a multi-faith and multi-ethnic garrison, that accepted newcomers – perhaps a sign that different faith groups joined forces. Their common cause was to offer protection from northern raiding in exchange for essential supplies and some precious commodities from the south. The situation evolved as power constellations changed – and there will have been periods when there was close cooperation with northern forces. Dariali Fort was hard to storm militarily, but utterly dependent on food provisions. It may be for these reasons that (during the last three centuries of the first millennium) many generations of the same core garrison occupied the fort without, it seems, ever being ousted, but living in symbiosis with those in control of the supply lines and admitting some into their circle. It is for this reason that similar cultural behaviour lasted from the eighth to tenth or early eleventh centuries and that we see both northern and southern imports and influences, without any one religious or ethnic group being solely in residence.

From the late tenth or early eleventh to the thirteenth or fourteenth century, there is no archaeological evidence for the occupation of Dariali Fort – though, not necessary proving total abandonment, as a much-reduced garrison could have been accommodated in buildings outside the area excavated. Indeed, the fort features in the sources and was, from the twelfth century onwards, under Georgian control. In contrast to Dariali Fort, we found traces of high medieval occupation on Gveleti Fort, some 3.6 km further south. This prominent natural bastion had seen low-level activity since the late first millennium as well as in the eleventh and/or twelfth century when it may have served as a local refuge. We witness a new surge of activity at both forts in the late thirteenth or fourteenth century. A tall building on the summit of Dariali Fort, then nearly a thousand years old, was re-occupied, new dry-stone annexes added and rooms paved with flagstones. At a similar time, we also see stone construction on Gveleti Fort. In the early fourteenth century, Georgia’s King Davit VIII reportedly fled to Gveleti Fort pursued by the forces of the mighty Ilkhanid Empire. Gveleti’s impregnable natural defences saved the king’s life. This may have stimulated a period of investment into this mountain castle that continued to be occupied into the sixteenth or seventeenth century, whilst there is no archaeological evidence for activity on Dariali Fort beyond the fourteenth or early fifteenth century.

In the early nineteenth century, ruined Dariali Fort was largely destroyed following its occupation by Ossetian rebels against Tsarist Russia, who were reportedly starved into submission. Russian explosives were meant to ensure that Dariali would never again stand in the way of Russian imperialism. Yet a gun cartridge case of 1916, from a material extraction pit next to a modern bunker, suggests that it did so once more, if unsuccessfully. This modern ammunition suggests that the bunker, and perhaps some of the other modern military trenches, may date to the period shortly after its First World-War production – when the Georgian ‘David’ may have used Dariali Fort to take a stance against the Russian ‘Goliath’. Georgia’s brief period of independence was crushed by the Red Army in 1921, breaking through the gorge. Today, the narrowest stretch of Dariali Gorge forms again the northern border of Georgia, as it had been the border of the Kingdom of Iberia, Persia, the Caliphate, the Emirate of Tbilisi and medieval Georgia – 2,000 years of continuity determined by geography.

Long before being a thorn in the flesh of the Russian Bear – the gorge-blocking bastion’s fame had spread thousands of kilometres to imperial capitals and residences. No larger than a football field, Dariali Fort’s unique topography, controlling a bottleneck on one of the few viable routes across the massive mountain chain separating Europe from Asia, has allowed it to punch well above its weight.

1.2. Geography

Dariali Fort, the focal point of our studies, is located c. 120 km north, as the crow flies, of Tbilisi, the capital of modern Georgia and c. 10 km ENE of the 5,047 m high Mount Kazbek. To reach it, the traveller coming from the south first has to ascend the 2,379 m-high Jvari or Cross Pass before descending c. 1,000 m in altitude. Snow, ice and avalanches impede travel in winter and spring and there is a permanent danger of rock fall and mudslides, which sometimes block the route. Not an easy crossing, alternative routes pose even greater difficulties and dangers. The Transcaucasian traffic highway via the Cross Pass and Dariali Gorge is generally considered the most viable across the Central Caucasus and has been the main traffic artery between Central Transcaucasia and the Steppes of Eurasia for over two millennia. Our project focused on the narrowest part of this Transcaucasian highway and the barriers and bastions built here to control movement and prevent hostile forces from crossing the mountains (fig. 1:10).

Dariali Fort, arguably the nerve centre of defensive efforts, crowns a conical hill, rising to roughly 1,390 m above sea-level a little over 100 m above the Tergi/Terek River (fig. 1:11).⁸ Melting snow turns this river, skirting the east side of the fort, for much of the year into a dangerous, fast-flowing torrent, with a current strong enough to move rocks and to drown any would-be trespasser. Traffic inevitably had to follow one bank of the river except where there were bridges which, in the landslide-prone vicinity of the fort, were invariably wiped away within a few years at most. Prior to the nineteenth century, traffic in this part of the gorge followed the left/western side of the river and all travellers had to pass a narrow passage between Dariali Fort and a steep rock-cliff, controlled via a gated barrier. On the opposite side of the Tergi a few hundred metres north of the fort, the rapid waters of the unfordable river washed along unsurmountable rock cliffs – blocking all traffic along the river’s right bank. It was at this point on the river’s left bank, just under one kilometre NNW of the fort, that we find one of the narrowest points of the gorge. Prominent Bakht’ari Ridge, towering over the narrow passage to the west of the river, was fortified here to control traffic.

fig. 1:10: Map of key sites in the Dariali Gorge by Kristen Hopper. Sites classified by type and period of first occupation or construction, as established by archaeological evidence. (Only surviving/excavated monuments have been plotted, and the map should not imply that these were necessarily the earliest. Late antique fortifications may well have replaced more ancient forts, as suggested by early cultural layers at the base of Dariali Rock and written sources.) Elevation data: SRTM 30 DEM (courtesy of the US Geological Survey).

South of the Dariali Fort (3.6 km), Gveleti Fort towers over another narrow passage, as well as forming an impregnable retreat. Whilst Dariali Fort was probably built in the late fourth century, at the postulated site of a more ancient fortification, occupation at Gveleti Fort appears to have commenced 300–500 years later, and the surviving walls probably belong to the second millennium. Not relevant for the ancient history of the gorge, Gveleti Fort was occupied simultaneously with Dariali Fort in the fourteenth century (and probably already at the end of the first millennium). Further south, there is a small undated dry-stone fortification at the village of Tsdo, probably a local refuge and lookout post and certainly not a valley-blocking fort.

Three cemeteries have been excavated in the northern parts of the gorge. Dariali Cemetery (late seventh/eighth to late tenth/early eleventh century) is south of Dariali Fort and clearly associated. Gigias Satibi Cemetery was the northernmost, but its exact position is hard to pinpoint. It was certainly over a kilometre NNE of Bakht’ari Ridge, maybe even 2 km. It has yielded eighth/ninth-century grave goods and was perhaps associated with a further control point at a narrow section of the gorge. The only known pre-medieval burial ground is Gveleti Cemetery (sixth to seventh or eighth centuries) c. 2.5 km south of Dariali Fort. The distinctly northern-style burials may belong to a northern auxiliary force. Only the former of these three burial grounds was explored within the framework of our project, whilst Gigias Satibi is now in Russia and inaccessible to us and attempts to locate further graves at Gveleti failed to yield any pre-modern burials.

fig. 1:11: Dariali Fort, with its prominent road-blocking wall north-west of the fort and the two trenches excavated on top of the fort: Trench F immediately south-west and Trench Q north-west of the centre. Note also the separate small rock south of the main plateau (defended by a separate bastion which is not visible on the plot). Laser-scanning plan by Malkhaz Datukishvili and Giorgi Kvaratskhelia plotted onto satellite imagery (Source: Esri, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo, and the GIS User Community). (Note that the map was compiled in 2015, prior to further extension of Trench Q.)

Whilst all our excavations (excluding only small sondages dug into terraces by the landscape team) took place within a five-kilometre stretch of the gorge, the landscape survey covered more ground and reached further south, in order to reveal the wider landscape setting of defensive installations across the ages and the food supply networks necessary to sustain them. A study of canal systems in eastern Georgia has been included in an appendix, being of significance for our wider Persia and its Neighbours Project.

1.3. Site chronology

Radiocarbon, archaeomagnetic and OSL samples

We used three different laboratories for radiocarbon dating. We are grateful to the following colleagues for dating samples and answering our enquiries: Diane Baker, Professor Tom Higham and Professor Christopher Ramsey and their colleagues (Research Laboratory for Archaeology and the History of Art, University of Oxford), Professor Gordon Cook, Dr Elaine Dunbar and Dr Brian Tripney (Scottish Universities Environmental Research Centre) and Professor Tomasz Goslar (Poznan Radiocarbon Dating Laboratory). We owe the identification of dated animal bones to Marjan Mashkour, Sarieh Amiribeirami, Karyne Debue, Delphine Decruyenaere, Homa Fathi and Roya Khazaeli. Charcoal samples were identified by Imogen Poole and Lyudmila Shumilovskikh. Human bone was identified by Annamaria Diana, Elena Kranioti, Cate Shupe and Anthi Tiliakou. Lisa Snape and Ian Bailiff provided OSL dates for Trench F and landscape features.⁹

Archaeomagnetic dates have been provided by Cathy M. Batt, David P. Greenwood and Tehreem Kainaat of the University of Bradford. The summary tables only list date ranges that are consistent with the radiocarbon dates, whilst their separate report lists all possibilities since 1000 BC.¹⁰

Bayesian modelling was accomplished by Eberhard W. Sauer and Cathy Batt, using the OxCal programme. Eberhard Sauer produced the final versions and selectively tested results in the BCal programme, whilst Cathy Batt provided essential advice on the programme, tested various sequences independently, reassuringly coming to similar results. Eberhard Sauer takes the responsibility for all potential imperfections, whilst the task could not have been accomplished without the advice of Cathy Batt. Eberhard Sauer is grateful to Professor Gordon Cook, Dr Derek Hamilton and Dr Tony Krus for having drawn his attention to the imperfections of his initial interpretation of radiocarbon sequences, their advice and for taking the time for a meeting. They have not seen or scrutinised the final results and are not to be blamed for any shortcomings. In the light of budgetary constraints towards the end of the project, the complexity of the stratigraphy and pressures of time, it was decided not to commission a professional Bayesian statistician. Having modelled our data in various ways, variants of models mostly produced similar results and constantly made us rethink our interpretation, increasing our confidence that the models are reasonably reliable.

Where there are few samples in a sequence, and/or if they are far apart in time, the modelled dates are close to the unmodelled dates, as one might expect. In cases of numerous well-stratified samples for a shorter period (e.g. the early medieval casemate) modelling has often allowed us to narrow down dates considerably, without producing results that are stratigraphically or chronologically unconvincing. In the text, we mainly cite the modelled dates (at 95.4% confidence), but the tables always list unmodelled as well as modelled dates (whenever there is a sequence that permitted us to produce Bayesian models). They also indicate if there is a high probability of a sample dating to a shorter period within the modelled or unmodelled range. These raw data will allow readers to independently assess our chronology, whether they wish to be cautious, and base their argument on broad date ranges with small margins of error, or prefer narrow date ranges, even if there is a greater chance that the real date may fall outside. Samples that are in no certain stratigraphic relation to other samples are shown as single plots. Whenever samples are in a clear stratigraphic sequence, they are shown as multi-plots, often spanning several phases. Tables of results, by contrast, have been split into phases, to make it easier for the reader to find relevant dates. The stratigraphic information presented will allow Bayesian statisticians to test, review and refine the models in future.

Phases

Our chronology is largely based on radiocarbon dating and stratigraphy. Occasionally, artefact typology, archaeo-magnetic and optically stimulated luminescence dating has helped to refine dating. Once typical artefact assemblages for different times were identified or typical masonry styles, such information has been used to date otherwise undated or poorly dated contexts.

We have resisted the temptation to date phases on the basis of textual evidence, and few historically recorded events can be directly linked with structures and deposits. It is nonetheless permissible to note that the start and end of some archaeologically identified phases may well coincide with turning points recorded in literary sources. The earliest traceable horticultural and agricultural activity (phase 2b) at the base of Dariali Rock roughly coincides with the earliest contemporary references to the pass defences. The earliest surviving structures and deposits on top of the plateau (phase 3a) are probably related to the historically attested construction of a fort by the Persians (but none of the sources provide a precise date for this event, and it is archaeology that helps to refine it). It is tempting to think that phase 4 marks the mid-seventh to early eighth century Post-Sasanian to Pre-Islamic interlude, but there is (as always) some leeway as to the precise start and end of the phase. Phases 5a-c are different to all previous phases, marked by much heavier intramural rubbish deposition, construction of dry-stone structures (as opposed to the earlier, late antique mortar walls and the virtual abandonment of building in the transitional phase) and the establishment of a cemetery (whilst there are no known graves before). One is inclined to think that the start of phase 5a may mark the arrival of new occupants, perhaps the historically attested Islamic garrison in c. AD 727 – even if there were almost certainly various ethnic groups and religions represented amongst fort occupants over the next three centuries (phase 5a-c).

There is a limit to the precision of our dating, but we have been able to differentiate the phases listed in Table 1:1. Not all deposits and structures can, of course, be assigned to a single phase. These phases are widely used in the finds reports and plans and facilitate systematic examination of developments. In the overall synthesis, we have used them more sparingly to avoid excessively technical language. It should be clear from the dates cited what phase is discussed and vice versa.

Chronological terms

The chronological terms used in this volume require a brief definition, as there is no agreement in scholarly literature on the precise duration of Late Antiquity and the Middle Ages. Any attempt to subdivide history into neat periods is, of course, subjective and dependent on what developments in religious, cultural, military, technological, administrative and political history are considered most significant in defining change, not to mention that such changes can affect different territories centuries apart. Furthermore, there will almost always be more change over a long period of time, however defined, than over a much shorter period bridging the postulated end and start of two successive phases of history. Such broad chronological terms nevertheless help us to describe complex developments without unnecessarily convoluted language. They may also stimulate thought, if properly defined, as to what constitutes significant change in world history.

Table 1:1: Phases used in the report.

In this monograph, key chronological terms are defined as follows:

Late Antiquity (phase 3): the period from the reign of Diocletian (r. AD 284–305), with its wide-ranging reforms, to the late AD 630s/mid-seventh century when the Caliphate replaced the Roman and Sasanian Empires as the dominant military and economic power in the Near East and the Mediterranean. Only the last two-and-a-half centuries of Late Antiquity are represented in the archaeological strata at Dariali Fort.

The term ‘(Eastern) Roman Empire’ (or just ‘Rome’) is preferred here to ‘Byzantine Empire’, notably in a pre-medieval context, as it was the only direct successor to Rome, and contemporary authors still define the empire as ‘Roman’. The Empire’s Roman heritage, not the location of its capital city, determined imperial identity. After the Empire’s division into a western and eastern half, the terms ‘Roman Empire’ or ‘Rome’ refer principally to the Eastern Roman Empire. The Eastern successor state to Rome was the beneficiary of the Caucasian defences, the victim when they failed, or were used by the masters of the pass to employ northerners in proxy wars against Rome. The Eastern Roman Empire was also sometimes a paymaster or contributor to the upkeep of the Caucasus fortifications (in theory at least, if in practise the money may have been used for other purposes) – and it was the only Roman Empire to survive beyond the late fifth century. In conventional Georgian chronology, phase 3 forms part of the Early Middle Ages (whereas in this report the term ‘Early Middle Ages’ is only used for phases 4–5). It should be noted that the narrower definition of the term Middle Ages (phases 4–6, i.e. mid-seventh to fifteenth centuries) as used in this report, as opposed to the broader definition in conventional Georgian chronology (c. fourth-eighteenth centuries, corresponding roughly to our phases 3–7a), is not meant to imply that we disagree with conventional Georgian chronology. As stressed above, any definition of the chronological parameters of the Middle Ages is to some extent arbitrary. Whilst the broader Georgian definition of the term ‘Middle Ages’ is as logical as the one used here, we felt that it would reduce the risk of misunderstanding to adopt a narrower definition, for two reasons:

1. This is internationally now more widely used.

2. The history of Dariali Fort from the fourth to the seventh centuries differs markedly from its history in the seventh/eighth to tenth/eleventh centuries. Describing the former as ‘late antique’ and the latter as ‘early medieval’ will make it clear what we refer to, as opposed to using the same term for both.

Middle Ages (phases 4 to 6): the period from the end of (Late) Antiquity, as defined above, to the fifteenth century and the Renaissance. This is broadly subdivided into the following eras: early medieval (mid-seventh to mid-eleventh centuries/phases 4–5), high medieval (phase 6a, mid-eleventh to thirteenth centuries and not well represented in our excavations) and late medieval (phase 6b, late thirteenth or fourteenth to mid-fifteenth centuries).

The first of these three medieval sub-phases is by far the longest, lasting for c. 400 years. As far as Dariali is concerned, this could be further subdivided into the interlude (phase 4), between the end of the Sasanian era in the mid seventh century (even if in all likelihood effective Sasanian control would have ceased a little before) and the Islamic conquest of c. AD 727, and the occupation of the fort by a new, reportedly Muslim, but in reality, probably a multi-faith garrison for the next three centuries or so (phase 5). Activity on the fort, attested archaeologically, seems to come to a temporary end toward the end of the tenth or in the first third of the eleventh century. Northern influences increase as time progresses, but we do not know if the fort stayed under formal Muslim control until this point or changed ownership a few years or decades before. After little activity in the High Middle Ages (phase 6a), exclusively traceable at Gveleti Fort, there is renewed occupation in both forts in the Late Middle Ages (phase 6b).

Post-medieval/Modern Era (phase 7): c. mid/late fifteenth century to the present. In conventional Georgian chronology, phase 7a still forms part of the Late Middle Ages (whereas in this report the term ‘Late Middle Ages’ is only used for phase 6b).

1.4. Ethnic/political terms

As we use the term ‘Roman’ in a political and legal, but not in an ethnic, sense, even for times when Rome or Italy more broadly was not a part of the Roman Empire, it is important to stress that the same applies to many other ethnonyms. It is clear that many northerners, notably larger groups, such as the Sarmatians and later the Huns, Turks/Khazars, were not ‘monolithic’ and included or absorbed other ethnic groups. Similarly, ‘Eranshahr’ (the Empire of the Iranians) under the Sasanian dynasty, included Iranians and non-Iranians. Centred (as already under the Achaemenids) on the Persis province in southwest Iran, ancient authors use the term ‘Persian’ for ethnic Iranians/speakers of Persian as well as for all inhabitants of the empire or parts thereof, just as the term ‘Roman’ has more than a single meaning. ‘Sasanian’ is a term that refers to a dynasty, if frequently applied to the state and its inhabitants in modern literature. By contrast, modern scholarship would not refer to the inhabitants of the Roman/Eastern Roman Empire by dynastic terms, such as ‘Constantinian’ or ‘Theodosian’ – perhaps because dynasties did not last long enough.

Ancient and medieval authors frequently conflate or confuse ethnic groups