Landscape Archaeology in Southern Caucasia. Finding Common Ground in Diverse Environments: Proceedings of the Workshop Held at 10th Icaane in Vienna, April 2016 (Orea) 9783700182047, 370018204X

Table of contents :
Contents
Preface by the Series Editor
Landscape Archaeology in Southern Caucasia:An Introduction
Moving in the Mountains: GIS and Mapping thePhenomenology of Travel through the South Caucasus
Patterns of Settlement and Metallurgy in Late Bronze–EarlyIron Age Kvemo Kartli, Southern Georgia
The Archaeological Landscape of the Hrazdan River Basinduring the Late Bronze–Early Iron Age
A Survey Project on the Borderlands of Turkey – Armenia –Nakhchivan – North-western Iran: Preliminary Report of the2014–2016 Surveys on the Middle Araxes Basin
Prehistory from the Ploughsoil: Interpreting ArtefactDistributions from Intensive Survey in the Highlands ofSamtskhe-Javakheti, Southern Georgia
Phasis and its Landscape: Preliminary Report of theArchaeological Survey of the Lower Stream of the Rioni RiverDelta
Samshvilde: Multidisciplinary Approaches to aHistorical City of Central Transcaucasia
Medieval Underground Shelters (Darnebi)of South-west Georgia
Approaching Landscapes of Infrastructure: Methods andResults of the Vayots Dzor Silk Road Survey
Approaches to Understanding Provincial Structure in theEarly Islamic Caucasus: Historic Landscape Characterisationin the Kura Plain
Index
Authors

Citation preview

William Anderson – Kristen Hopper – Abby Robinson (Eds.) Landscape Archaeology in Southern Caucasia Finding Common Ground in Diverse Environments

Österreichische Akademie der Wissenschaften Philosophisch-historische Klasse

Oriental and European Archaeology Volume 8 Series Editor: Barbara Horejs

Publications Coordinator: Ulrike Schuh

William Anderson – Kristen Hopper – Abby Robinson (Eds.)

Landscape Archaeology in Southern Caucasia Finding Common Ground in Diverse Environments

Proceedings of the Workshop held at 10th ICAANE in Vienna, April 2016

Accepted by the Publication Committee of the Division of Humanities and the Social Sciences of the Austrian Academy of Sciences: Michael Alram, Bert Fragner, Hermann Hunger, Sigrid Jalkotzy-Deger, Franz Rainer, Oliver Jens Schmitt, Peter Wiesinger and Waldemar Zacharasiewicz

Picture on the opposite page: Field survey in Samtskhe-Javakheti, April 2017 (photo: G. Khaburzania)

This publication has undergone the process of anonymous, international peer review. The paper used for this publication was made from chlorite-free bleached cellulose and is aging-resistant and free of acidifying substances.

English language editing: Hazel Harrison Graphic design: Angela Schwab Cover design: Mario Börner, Angela Schwab

All rights reserved. ISBN: 978-3-7001-8204-7 Copyright © 2018 by Austrian Academy of Sciences, Vienna Printing: Prime Rate Kft., Budapest Printed and bound in the EU https://epub.oeaw.ac.at/8204-7 https://verlag.oeaw.ac.at

ORIENTAL AND EUROPEAN ARCHAEOLOGY Vol. 1

B. Horejs – M. Mehofer (eds.), Western Anatolia before Troy. Proto-Urbanisation in the 4th Millenium BC? Proceedings of the International Symposium held at the Kunsthistorisches Museum Wien, Vienna, Austria, 21–24 November, 2012 (Vienna 2014).

Vol. 2

B. Eder – R. Pruzsinszky (eds.), Policies of Exchange. Political Systems and Modes of Interaction in the Aegean and the Near East in the 2nd Millennium B.C.E. Proceedings of the International Symposium at the University of Freiburg, Institute for Archaeological Studies, 30th May–2nd June, 2012 (Vienna 2015).

Vol. 3

M. Bartelheim – B. Horejs – R. Krauß (eds.), Von Baden bis Troia. Ressourcennutzung, Metallurgie und Wissenstransfer. Eine Jubiläumsschrift für Ernst Pernicka (Rahden/Westf. 2016).

Vol. 4

M. Luciani (ed.), The Archaeology of North Arabia. Oases and Landscapes. Proceedings of the International Congress held at the University of Vienna, 5–8 December, 2013 (Vienna 2016).

Vol. 5

B. Horejs, Çukuriçi Höyük 1. Anatolia and the Aegean from the 7th to the 3rd Millennium BC. With contributions by Christopher Britsch, Stefan Grasböck, Bogdana Milić, Lisa Peloschek, Maria Röcklinger and Christoph Schwall (Vienna 2017).

Vol. 6

M. Mödlinger, Protecting the Body in War and Combat. Metal Body Armour in Bronze Age Europe (Vienna 2017).

Vol. 7

Ch. Schwall, Çukuriçi Höyük 2. Das 5. und 4. Jahrtausend v. Chr. in Westanatolien und der Ostägäis. Mit einem Beitrag von Barbara Horejs (Vienna 2018).

Contents

7

Contents Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Preface by the Series Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Kristen Hopper – William Anderson – Abby Robinson Landscape Archaeology in Southern Caucasia: An Introduction . . . . . . . . . . . . . . . . . . . . . . . 11 Lara Fabian Moving in the Mountains: GIS and Mapping the Phenomenology of Travel through the South Caucasus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Nathaniel L. Erb-Satullo Patterns of Settlement and Metallurgy in Late Bronze–Early Iron Age Kvemo Kartli, Southern Georgia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Manuel Castelluccia The Archaeological Landscape of the Hrazdan River Basin during the Late Bronze–Early Iron Age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Ayhan Yardimciel – Mehmet Ali Özdemir – Mehmet Işıklı A Survey Project on the Borderlands of Turkey – Armenia – Nakhchivan – North-western Iran: Preliminary Report of the 2014–2016 Surveys on the Middle Araxes Basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 William Anderson – Michelle Negus Cleary Prehistory from the Ploughsoil: Interpreting Artefact Distributions from Intensive Survey in the Highlands of Samtskhe-Javakheti, Southern Georgia . . . . . . . . . . . . 83 Davit Naskidashvili Phasis and its Landscape: Preliminary Report of the Archaeological Survey of the Lower Stream of the Rioni River Delta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 David Berikashvili Samshvilde: Multidisciplinary Approaches to a Historical City of Central Transcaucasia . . . . . 105 Abby Robinson – Giorgi Khaburzania Medieval Underground Shelters (Darnebi) of South-west Georgia . . . . . . . . . . . . . . . . . . . . . 119 Kathryn J. Franklin – Astghik Babajanyan Approaching Landscapes of Infrastructure: Methods and Results of the Vayots Dzor Silk Road Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Paul Wordsworth Approaches to Understanding Provincial Structure in the Early Islamic Caucasus: Historic Landscape Characterisation in the Kura Plain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

8

Contents

9

Preface by the Series Editor The 8th volume of the OREA series about Landscape Archaeology in Southern Caucasia represents the proceedings of a workshop held at the International Congress on the Archaeology of the Ancient Near East (ICAANE) in 2016. This 10th anniversary of the ICAANE took place from 25th to 29th of April in Vienna, hosted and organized by the Institute for Oriental and European Archaeology (OREA) of the Austrian Academy of Sciences. Altogether 800 participants from 38 different countries found their way to Vienna to celebrate the 10th anniversary of ICAANE with 8 scientific sections, 28 workshops, round tables, a huge poster exhibition and a special section about “Cultural Heritage under Threat”. The topics of the 10th ICAANE covered traditional, as well as new fields, in relation to stateof-the-art approaches and methodologies. The general themes of transformation and migration, cultural landscapes, religion and rituals, environmental shifts, contextualized images, economies and societies, Islamic archaeology, as well as current excavations and field reports, have been discussed in large sections published as the 10th ICAANE proceedings with the Harrassowitz Publishing House. A special element was new scientific input discussed in the additional 28 workshops, focused on more detailed questions in relation to our broad scientific fields. The engaged discussions of internationally high-ranked experts with young scholars was essential for the success and open atmosphere of the 10th ICAANE in Vienna. I would like to thank W. Anderson, K. Hopper and A. Robinson, who not only organised the workshop about Landscape Archaeology in Southern Caucasia. Finding Common Ground in Diverse Environments, but also edited these proceedings as a volume for the internationally peerreviewed OREA series. The editors brought together 16 authors for ten contributions focussing on different aspects of Caucasian Landscape Archaeology, accompanied by a detailed introduction providing a good overview for an audience not familiar with this particular region. The editors succeeded in their aims to present and reflect on some of the current approaches in Landscape Archaeology based on site and regional studies in Armenia, Azerbaijan, Georgia, parts of eastern Turkey and northwest Iran. By drawing attention to the Southern Caucasus as a zone of cultural contacts with its particular environmental and cultural conditions, the editors and authors offer a new perspective of this fascinating region to a broader readership. Moreover, their consideration of future research directions demonstrate the potential of southern Caucasian archaeology and the impact of “Finding Common Grounds in Diverse Environments”. My sincere thanks for financial support of the conference go to several Austrian and international institutions, which are The Austrian Federal Ministry of Europe, Integration and Foreign Affairs, the University of Vienna, the City of Vienna, the Vienna Science and Technology Fund (WWTF), the Institute for Aegean Prehistory (INSTAP), the Austrian Orient Society HammerPurgstall and the Austrian Academy of Sciences. For the publication of this volume, I would like to thank Ulrike Schuh for the coordination, Angela Schwab for the layout, Hazel Harrison for English language editing and the Publishing House of the Austrian Academy of Sciences. Barbara Horejs Vienna, 2 March 2018

10

Contents

Landscape Archaeology in Southern Caucasia: An Introduction Kristen Hopper * – William Anderson ** – Abby Robinson ***

The State of Research Landscape archaeology is a complex and mutable term that evades a straightforward definition. As a practice, it encompasses a number of different theoretical and methodological approaches. Most of these involve an attempt to understand the development of a place through time, investigating how it is shaped by natural events and cultural actions and, in turn, how these factors influence human activities.1 It tends to imply a regional – as opposed to a site-based – approach, is holistic in outlook and accounts for relational aspects of time and space. Yet, the study of a specific site within the context of its local surroundings may also take the form of landscape archaeology. Therefore, it does not entail a determined or agreed set of procedures but is typified by a range of methods that are tailored to different environmental and archaeological conditions and research questions. These methods generate different forms and resolutions of information – from the fine grained, dealing with individual features, artefacts and ecofacts, to the general and large scale, incorporating whole regional systems. There is a vast, and still growing, body of literature concerned with the archaeological landscapes of the Near East, particularly in Mesopotamia, Anatolia and the Levant.2 This research has focused on understanding not just regional settlement patterns but the interaction between humans and their environments at multiple temporal and geographic scales. However, it is only in the last 20 or so years that we have seen increased interest in applying similar methodologies and approaches (including intensive and extensive surveys, satellite remote sensing and GIS-based analyses) and particularly in adopting diachronic approaches to research projects in Southern Caucasia.3 This volume offers a chance to present and reflect on some of these approaches as they are currently being practised in Southern Caucasia.

Durham University, [email protected]. Landscape Archaeology in Georgia Project, [email protected]. *** University of Melbourne, [email protected]. 1 Ashmore – Knapp 1999, 2; Anschuetz et al. 2001; Wilkinson 2003, 3–4. 2 See Wilkinson 2003; McPhillips – Wordsworth 2016. 3 See Marro 2004; Alizadeh – Ur 2007; Badalyan et al. 2008; Smith et al. 2009; Badalyan et al. 2010; Ristvet et. al. 2011; Rova et al. 2011; Birkett-Rees 2012; Castelluccia et al. 2012; Lyonnet et al. 2012; Ricci 2012; Greene – Lindsay 2013; Lindsay – Greene 2013; Ristvet et al. 2013; Anderson et al. 2014; Egeland et al. 2014; Hammer 2014a; Stöllner – Gambashidze 2014; Castelluccia 2015; Earley-Spadoni 2015; Hughes 2015; Petrosyan et al. 2015; Sauer et al. 2015; Batiuk et al. 2017, Erb-Satullo et al. 2017; Fabian 2017; Franklin et al. 2017; Intagliata – Naskidashvili 2017; Lawrence – Wilkinson 2017; Carminati 2018; Negus Cleary et al. 2018; Khaburzania – Robinson 2018. Though this is not an exhaustive list, the publications cited here represent the output of a significant number of international collaborative projects that have involved regional survey or that have embraced landscape methodologies at multiple scales. It includes the output of many current projects in the region including the following: The Archaeological Exploration of Barda‘a Project (AEB): Archaeological survey of the Late Antique and early Islamic city of Barda‘a, Azerbaijan – Oxford University in association with the Nizami Ganjavi Programme *

**

12

Kristen Hopper – William Anderson – Abby Robinson

Southern Caucasia is defined for the purposes of this volume as the land between the Black and Caspian Seas including and adjacent to the Greater and Lesser Caucasus ranges. This encompasses Armenia, Azerbaijan, Georgia, parts of eastern Turkey and north-west Iran (Fig. 1).4 The practice of archaeology in the Caucasus region (specifically in Armenia, Azerbaijan and Georgia) is reflective of political conditions that have prevailed there over the past two centuries. Having been an antiquarian pursuit in the 19th century, influenced by imperial nations, especially Tsarist Russia, under the Soviet Union, archaeology became a science of culture history at a pan-regional scale, with explicitly Communist and Marxist-inspired aims.5 Since the break-up of the Soviet Union, archaeology in the independent Caucasus states (Armenia, Azerbaijan and Georgia) has developed along different trajectories.6 Whilst the region’s Soviet-era archaeology followed a more or less consistent set of methodological, theoretical and analytical tenets, archaeology in the independent Caucasus states from 1991 onwards, not to mention autonomous areas that have since emerged, is characterised by diversity. This divergence is not only manifested in differences between nation states but is also apparent within those states, reflected in the variable scales and research aims of projects being undertaken and the methods and approaches that they use. The recent and growing interest in the archaeology of Southern Caucasia has been influenced not only by the opening up of the region following the break-up of the Soviet Union but more recently by the increasing difficulties encountered in working in other parts of the Near East, especially as the result of political unrest. Furthermore, where before 1991 nearly all archaeology was funded centrally and publicly, we now see an array of different funding sources that include not only governments but also overseas public institutions, universities from within and outside the region, multinational corporations and private sponsors. Some projects benefit from large-scale and multiple sources of funding, while others operate on very limited budgets. The internationalisation of research and the involvement of scholars trained in other fields – both geographical and disciplinary – have contributed to the diversity of archaeological approaches and methods currently practised in the region. This is a contributing factor in the direction towards a transnational archaeology, where international cooperation is becoming more the rule than the exception.

4

5 6

for the study of languages and cultures of Azerbaijan and the Caucasus; Project ArAGATS – Cornell University, The Institute of Archaeology and Ethnography, NAS Republic of Armenia; Georgian-Australian Investigations in Archaeology (GAIA) – University of Melbourne and the Centre for Archaeological Research (Tbilisi) and associated projects including the Landscape Archaeology in Georgia (LAG) Project and the Archaeological Survey in Samtskhe-Javakheti (ASSJ) Project; the Georgian-Italian Shida Kartli Archaeological Project – Ca’ Foscari University of Venice and the Georgian National Museum; Gadachrili Gora Regional Archaeological Project Expedition (GRAPE) Project – Georgian National Museum, University of Toronto; Lerik in Antiquity Archaeological Project – the University of Pennsylvania, Azerbaijan National Academy of Sciences; Mil Plain Survey – Deutsches Archäologisches Institut, Azerbaijan National Academy of Sciences, Institute of Archaeology and Ethnography; The Darial Pass Survey, part of the project Persia and its Neighbours: The Archaeology of Late Antique Imperial Power in Iran – University of Edinburgh, Durham University and Tbilisi State University; and the current survey projects by Kafkas and Ardahan Universities in the Ağrı and Iğdır and Ardahan regions of eastern Turkey. The region is referred to by multiple names in the literature (e.g., the South Caucasus, Transcaucasia, Caucasia; see Kohl – Tsetskhladze 1995, n. 2; Rubinson – Smith 2003, 8, n. 1;). We have allowed authors to follow their preference. Likewise, terms denoting cultural, ethnic, historical groupings are at the discretion of the author(s). Attaining consistency is not possible or desirable. Which term is appropriate for the region depends on the subject and time period being studied (Rapp 2012). Chernykh 1995; Shnirelman 1995; Sagona 2010; Lozny 2017; Sagona 2017. Gamkrelidze 2004 for Georgia; Smith 2005; Lindsay – Smith 2006 for Armenia; Khatchadourian 2008; Sagona 2010; Sagona 2017.

Landscape Archaeology in Southern Caucasia: An Introduction

13

Finding Common Ground: Aims and Objectives This volume is the result of a workshop that was held at the 10th ICAANE in Vienna on 28th April, 2016, aimed at bringing together scholars engaged in archaeological survey and landscape analysis in Southern Caucasia. It was intended to stimulate conversations on research aims, data analysis and management, methodological issues (survey methodologies, multi-scalar and multidisciplinary data integration, GIS analyses), and long standing thematic and historical debates. Furthermore, it was designed to encourage discussions about how we could, as a community of scholars, promote communication and exchange of information between ourselves and with others working in different survey regions and across modern borders. The physical geography – above all, the Greater and Lesser Caucasus mountain ranges and the Kura/Mtkvari and Aras/Araxes river systems – has been fundamental in shaping the forms of human occupation in the region. It has also greatly influenced both our methodologies and our interpretations of past landscapes. Its mountainous geography, and the strategic and defensive advantages it often provided, have also contributed to the characterisation of Southern Caucasia as a periphery, particularly in the story of successive powerful empires originating to its east and west. However, this pejorative designation is debatable as it is only valid from the perspective of other, supposedly superior cultural areas.7 The region is perhaps better characterised as an important zone of cultural contact between the Near East, Anatolia and Central Asia that, in part, due to geography (especially the Caucasus mountain ranges) still maintained an important and local trajectory of development. The presentations at the workshop detailed a wide range of approaches that are also reflected in the contributions to this volume. In terms of scale, they include regional surveys such as those undertaken by Yardimciel, Özdemir and Işıklı in the Middle Araxes basin, and Robinson and Khaburzania in Samtskhe-Javakheti province in south-west Georgia, to site-specific investigations, such as that led by Berikashvili at Samshvilde. Moreover, Wordsworth’s chapter also demonstrates the usefulness of a multi-scalar approach. The contributions engage with multiple forms of evidence – architectural, artefactual, environmental and textual. Furthermore, the range of methodologies that is represented includes extensive and intensive ground survey, feature mapping and spatial analysis, remote sensing, and the study of artefacts and materials. While the chronological focus of individual projects is sometimes specific, as a whole these contributions represent a broad time scale – from the Neolithic through to the medieval period. Periodisation is the source of much debate in the region. As an example, the term ‘medieval’ can be used to cover a period of more than a millennium, from the 4th to the 18th century AD. This includes centuries long before and after it would generally be used in other regions, including neighbouring Persia and Anatolia. As in our approach to regional names and cultural groupings, we have encouraged a spirit of plurality and not sought consistency for terms that are themselves debated and contested. However, as implied by the title of this volume, we are very much concerned with how we find common themes to research and debate. Therefore, it is important that we are explicit about our approaches and methods, our research questions and our terminology and consider what impact our choices have had upon our results and interpretations. Themes of this Volume The varied approaches represented here highlight the relationship between past and present landscapes. In many cases, the contributions provide a long-term perspective. The recognition of the palimpsest nature of the landscape is clear in the contribution by Anderson and Negus Cleary who explore the relationship between artefact distribution and historical and modern agricultural

7

Rubinson – Smith 2003; Smith 2005; Kohl 1988; Khatchadourian 2013; Khatchadourian 2014.

14

Kristen Hopper – William Anderson – Abby Robinson

activities (e.g., manuring, terracing) in Samtskhe-Javakheti, and how that can contribute to our understanding of changing land-use patterns over the long term. Wordsworth also explores the relationship between recent, particularly Soviet-period, landscape transformations and archaeological landscapes around Bərdə in Azerbaijan. He investigates how we can use concepts such as Historical Landscape Characterisation, commonly applied in the UK and increasingly elsewhere,8 to identify pre-Soviet land-use patterns. Franklin and Babajanyan also pick up on the theme of Soviet-period landscape reorganisation along the Silk Road in the Vayots Dzor region of Armenia and consider how it can affect our understanding of the archaeological record. However, they also rightly point out how the significant transformations brought about in the Soviet period are only some of many anthropogenic changes to have occurred over millennia that shape our current perceptions of this landscape. The rapidly changing landscape and the impact of Soviet and post-Soviet industrialisation have increased the demand for heritage management in Southern Caucasia. The impact of agricultural intensification (particularly deep ploughing, collectivised agriculture and earth moving in advance of irrigation schemes), heavy industry (particularly of extractive and resources industries), the construction of infrastructure such as roads, and the increasing importance of international tourism have brought new challenges and also new opportunities. Landscape archaeology has an important contribution to make in the way that these challenges are approached, through informing public policy and balancing community and heritage sector interests that may include environmental conservation, the reconstruction and maintenance of sites, equitable and sustainable tourism, or local economic growth. Construction and development works at a local, national and international scale, ranging from urban building to multi-national resources extraction and transportation, led to the emergence of cultural heritage management and ‘rescue archaeology’ in the region. Whilst archaeology and heritage are primarily managed by the state across Southern Caucasia through museums and heritage agencies, there has more recently been a growth in private sector heritage management that operates on the basis of developers funding research and salvage of archaeologically important sites that are threatened with destruction. In Southern Caucasia, by far the largest of these projects was triggered by the BP-funded construction of the Baku–Tbilisi–Ceyhan pipeline that crosses Azerbaijan, Georgia and eastern Turkey. The resulting archaeology generated significant new information on numerous sites, several of which were excavated and presented in publicly accessible reports.9 Yet, while the damage caused by this development may have been offset by the often high-quality archaeological work completed, these salvage works raise the tension between sitebased and landscape-based approaches. With the site as the focus, contextualising these results within a wider ‘landscape’ framework can be difficult. Whether this can ever be reconciled in the context of developer-funded rescue archaeology is a matter of debate, and one beyond the scope of this volume, but it does underline the vital importance of foregrounding landscape research at times of rapid social and economic change. Another technological innovation that has influenced the practice and methods of archaeology in the region is the expanded use of remote sensing. Although aerial photography is by no means new to archaeology, the availability of free or low-cost satellite imagery on platforms such as Google Earth has provided archaeologists with an accessible and immensely useful tool for both site discovery and site monitoring, but one that brings a new set of methodological and ethical challenges.10 Satellite remote sensing is used by a number of the authors in this volume to inform survey methodologies, locate archaeological sites and investigate the relationship between settlements, activity areas and ancient features. Ground-truthing the results of remote sensing exercises through survey can help to create predictive models for site location (e.g., Erb-Satullo), while the

8 9 10

E.g., Turner – Crow 2010. Taylor et al. 2011. Myers 2010.

Landscape Archaeology in Southern Caucasia: An Introduction

15

use of multiple types of historical and modern satellite imagery, combined with historical maps and documents, can help us to track landscape changes over the course of the 20th century and show how this has influenced our reading of the archaeological record (see chapters by Wordsworth and Naskidashvili). Many of the studies in this volume (Castelluccia, Erb-Satullo, Franklin and Babajanyan, Naskidashvili, Robinson and Khaburzania, Wordsworth, and Yardimciel et al.) demonstrate the value of integrating historical texts and legacy data – from earlier surveys, excavations and maps – into new contexts. In addition to the fresh insights that are provided through the reanalysis of this material, research of this kind is also useful in introducing non-English language publications to a wider academic audience. Furthermore, these sources, in addition to local knowledge, can be crucial for identifying archaeological sites and features that have been affected by modern activities or are not visible on satellite imagery, such as the underground structures discussed by Robinson and Khaburzania. In terms of methodologies, a wide variety of approaches are represented here. However, methodologies that involve extensive and intensive pedestrian survey and consider sites and features within the wider context of the landscape are favoured over survey techniques that focus exclusively on ‘sites’ as units of investigation. There is also clear attention being paid to how topography and environment influence not only our methodologies but also the outcomes, and ultimately the comparability of our data. Erb-Satullo, for example, explores the relationship between metallurgical activities and settlement during the Late Bronze Age and Early Iron Age in Kvemo Kartli (Mashavera and Debeda river valleys), arguing that, in part, hilltop sites of the LBA/EIA in this region were positioned in relation to ore deposits and the desire to exploit these resources or control access to them. Anderson and Negus Cleary discuss the usefulness of intensive survey in upland environments and clearly demonstrate how such techniques can be adapted to specific environmental zones to provide new information on historical settlement and land use. In his paper, Naskidashvili highlights challenges in employing intensive pedestrian survey in western Georgia (also noted by Erb-Satullo); due to marshy conditions, it was only possible in cultivated fields, but the distribution of artefacts was heavily influenced by flooding. The need to adapt survey methodologies to specific environmental zones while retaining transparency about methodologies to ensure comprehension and comparability of data is clearly demonstrated. Finally, several of the papers consider what could be termed ‘landscapes of movement’. It is hard to overstate the influence of the region’s major mountain chains and rivers on social, political and economic developments. Fabian examines how the mountains acted as both a barrier to and a conduit of movement for ancient communities and how we can explore this concept through the use of GIS-based analyses. By taking into consideration both the results of leastcost path analysis and historically documented route systems, more nuanced models of movement can be developed which are capable of recognising change through time. Fabian also touches on the tension between imperial and local perceptions, in this case of space, a major theme in empire studies more generally. Continuing to address the idea of movement, Franklin and Babajanyan draw our attention to how we can use network or infrastructural analyses to widen our perspectives on the Silk Road in the Vayots Dzor region and shift our thinking away from linear models of connected nodes.

Future Directions Considering the discussions that ensued from the workshop and the papers that comprise this volume, there are several key themes that we propose as avenues for further research. One of these is the relationship between agricultural and pastoral land use and, by extension, the relationship between upland and lowland environments and the communities who inhabit them. As Wordsworth touches on in the Bərdə example, we have an underdeveloped understanding of pasture lands

16

Kristen Hopper – William Anderson – Abby Robinson

Fig. 1   Map of Southern Caucasia. The numbers correspond to the geographical areas discussed in the relevant chapters in this volume: 3. Mashavera-Debeda region; 4. Hrazdan River basin; 5. Middle Araxes basin; 6. Samtskhe-Javakheti Highlands; 7. Colchean Lowlands; 8. Samshvilde; 9. Aspindza-Akhalkalaki region; 10. Vayots Dzor region; 11. Bərdə (Base map SRTM 30m DEM, available from the US Geological Survey)

and the role of pastoralism in many ancient economies.11 Furthermore, particularly for highland environments, we need to understand better the changing relationship between pastoral and agricultural land use through time. Anderson and Negus Cleary demonstrate the importance of this in their observation of the changes in land use zoning between the prehistoric and medieval periods. Focussing on methodology and techniques of recording and presenting data, we also need to consider both the comparability of our data and its preservation and dissemination. Indeed, the increasing use of GIS and remote sensing is producing easily sharable datasets. However, it is vital to continue to be explicit about our methodologies and their impact on the data generated. Also, with the increasing amount of data that is generated through survey and remote sensing, we must consider how ‘big’ (particularly virtual) datasets will be maintained and managed. Further on the topic of data sharing, the use of open source repositories where spatial and other data related to a project can be accessed and augmented by other scholars (as discussed by Franklin and Babajanyan) is increasingly important. On this theme, it should be noted that a national cultural heritage database for Georgia, discussed during the workshop, has recently been launched in Tbilisi.12 A public user interface such as this will enable local and international scholars to access the in-

11 12

Hammer 2014b. http://memkvidreoba.gov.ge/ (last accessed 29 Jan. 2018). This database is the product of a collaboration begun in 2013 between the National Agency for Cultural Heritage Preservation of Georgia and the Norwegian Cultural Heritage Directorate: it is intended as a resource for the heritage sector and other government agencies but also for research projects nationwide.

Landscape Archaeology in Southern Caucasia: An Introduction

17

formation already in the database and to share their own data. This and similar enterprises represent first steps along the way to ensuring that the interests of cultural heritage management and research projects dovetail in the design and implementation of our projects and our databases. Continued engagement with legacy data is also crucial for both research and heritage management. Historical aerial photograph archives, such as those housed at the Centre for Archaeological Research in Tbilisi, provide a diminishing physical resource that requires attention in order to preserve the vast amount of information it represents. Increasingly, legacy data (especially historical aerial photographs) represent the only record of no longer extant archaeological features destroyed or irrevocably altered by modern land use practices. Finally, we should emphasise that a spirit of engagement and collaboration between scholars within and outside the Caucasus region and the free exchange of ideas is central to furthering this agenda. This is the common ground that we seek when approaching Southern Caucasia’s diverse, complex and fascinating landscapes. In this spirit, we have chosen to dedicate this volume to the memory of Tony Sagona and Tony Wilkinson. These two archaeologists were important figures in our lives as intellectual guides and as friends, and they have enriched the lives of many others whom they met and worked alongside. One of the remarkable qualities that both demonstrated was their ability to forge connections across national, linguistic and social boundaries. Their contributions to the archaeology of the Near East are monumental and will be felt for many years to come. Acknowledgements: We would like to thank everyone who participated in the session at the 10th ICAANE in Vienna on April 28, 2016, most especially those who presented papers: Eleonora Carminati (University of Melbourne), Lara Fabian (University of Pennsylvania), Alan F. Greene (Stanford University), Giorgi Khaburzania (National Agency for Cultural Heritage Preservation of Georgia), Damjan Krsmanovic (University of Leicester), Walter Kuntner (University of Innsbruck), Dan Lawrence (Durham University), Ian Lindsay (Purdue University), Davit Naskidashvili (Iv. Javakhishvili Tbilisi State University), Kathryn O’Neil Weber (Cornell University), Christopher Ratté (University of Michigan, Ann Arbor), Andrea Ricci (German Archaeological Institute), Lisa Snape-Kennedy (Durham University) and Paul Wordsworth (University of Oxford). Furthermore, we are grateful to all those who contributed to this volume, to Giorgi Khaburzania for helping to organise the session, and to Dan Lawrence for helpful comments on the introduction.

References Alizadeh – Ur 2007 K. Alizadeh – J. Ur, Formation and destruction of pastoral and irrigation landscapes on the Mughan Steppe, northwestern Iran, Antiquity 81, 311, 2007, 148–160. Anderson et al. 2014 W. Anderson – J. Birkett-Rees – M. Negus Cleary – D. Krsmanovic – N. Tskvitinidze, Archaeological survey in the south Caucasus (Samtskhe-Javakheti, Georgia). Approaches, methods and first results, Anatolia Antiqua 22, 2014, 11–33. Anschuetz 2001 K. F. Anschuetz – R. H. Wilshusen – C. L. Scheick, An archaeology of landscapes. Perspectives and directions, Journal of Archaeological Research 9, 2, 157–211. Ashmore – Knapp 1999 W. Ashmore – A. B. Knapp, Archaeologies of Landscape. Contemporary Perspectives (Oxford 1999). Badalyan et al. 2008 R. S. Badalyan – A. T. Smith – L. Khatchadourian – P. S. Avetisyan, Village, fortress and town in Bronze and Iron Age Southern Caucasia. A preliminary report on the 2003–2006 investigations of Project ArAGATS on the Tsaghkahovit Plain, Republic of Armenia, Archäologische Mitteilungen aus Iran und Turan 40, 2008, 45–105. Badalyan et al. 2010 R. S. Badalyan – A. T. Smith – L. Khatchadorian, Project ArAgats. 10 years of investigations into Bronze and Iron Age sites in the Tsaghkahovit Plain, Republic of Armenia / ArAgats projesi. Ermenistan Tsaghkahovit Ovası’nda 10 yıldır sürdürülen Tunç ve Demir çağ araştırmaları, Türkiye Bilimler Akademisi Arkeoloji Dergisi 13, 2010, 263–276.

18

Kristen Hopper – William Anderson – Abby Robinson

Batiuk et al. 2017 S. Batiuk – M. Jalabadze – A. Graham – I. Koridze – K. Abu Jayyab – C. Savulov, The Gadachrili Gora Regional Archaeological Project. 2016 Preliminary Report, Anatolica 43, 2017, 173–202. Batmaz et al. 2018 A. Batmaz – G. Bedianashvili – A. Michalewicz – A. Robinson (eds.), Context and Connection. Essays on the Archaeology of the Ancient Near East in Honour of Antonio Sagona, Orientalia Lovaniensia Analecta 268 (Leuven 2018). Birkett-Rees 2012 J. Birkett-Rees, Power and Presence. Landscape and Tenure in Middle Bronze Age Central Transcaucasia, Ancient Near Eastern Studies 49, 2012, 61–94. Carminati 2018 E. Carminati, Landscape management during the Early Kurgan period. The Bedeni Plateau case study, in: Batmaz et al. 2018, 271–276. Castelluccia 2015 M. Castelluccia, The evolution of the archaeological landscape of the Armenian highland during the Iron Age, Ancient Civilizations from Scythia to Siberia 21, 2015, 302‒368. Castelluccia et al. 2012 M. Castelluccia ‒ R. Dan ‒ R. La Farina ‒ A. Petrosyan ‒ M. Raccidi, The first season of the Kotayk Survey Project. Preliminary report, Aramazd. Armenian Journal of Near Eastern Studies 7, 2, 2012, 28‒35. Chernykh 1995 E. N. Chernykh, Postscript: Russian archaeology after the collapse of the USSR. Infrastructural crisis and the resurgence of old and new nationalisms, in: Kohl – Fawcett 1995, 139–148. Earley-Spadoni 2015 T. Earley-Spadoni, Landscapes of warfare. Intervisibility analysis of early Iron and Urartian fire beacon stations (Armenia), Journal of Archaeological Science, Reports 3, 2015, 22–30. Egeland et al. 2014 C. P. Egeland – B. Gasparian – D. Arakelyan – C. M. Nicholson – A. Petrosyan – R. Ghukasyan – R. Byerly, Reconnaissance survey for Palaeolithic sites in the Debed River Valley, northern Armenia, Journal of Field Archaeology 39, 4, 2014, 370–386. Erb-Satullo et al. 2017 N. Erb-Satullo – B. Gilmour – N. Khakhutaishvili, Copper production landscapes of the Late Bronze–Early Iron Age South Caucasus, Journal of Anthropological Archaeology, 47, 2017, 109–126. Fabian 2017 L. Fabian, Numismatic communities in the northern South Caucasus 300 BCE–300 CE. A geospatial analysis of coin finds from Caucasian Iberia and Caucasian Albania, in: H. Teigen – E. Seland (eds.), Sinews of Empire. Networks of the Roman Near East and Beyond (Oxford 2017) 29–63. Franklin et al. 2017 K. Franklin – T. Vorderstrasse – F. Babayan, Examining the late medieval village from the case at Ambroyi, Armenia, Journal of Near Eastern Studies 76, 2017, 113–138. Gamkrelidze 2004 G. Gamkrelidze, On the history of archaeology in Georgia, Journal of Georgian Archaeology 1, 2004, 208–217. Greene – Lindsay 2013 A. Greene – I. Lindsay, Mobility, territorial commitments, and political organization among Late Bronze Age polities in southern Caucasia, Archeological Papers of the American Anthropological Association 22, 1, 2013, 54–71. Hammer 2014a E. Hammer, Highland fortress-polities and their settlement systems in the southern Caucasus, Antiquity 88, 2014, 757–774. Hammer 2014b E. Hammer, Local landscape organization of mobile pastoralists in southeastern Turkey, Journal of Anthropological Archaeology 35, 2014, 269–288.

Landscape Archaeology in Southern Caucasia: An Introduction

19

Hughes 2015 R. Hughes, The Archaeology of a Colchian Landscape. Results of the Eastern Vani Survey (PhD Diss., University of Michigan, Ann Arbor 2015). Intagliata – Naskidashvili 2017 E. Intagliata – D. Naskidashvili, Forgotten borderlands. Guria and Adjara survey project, Heritage Turkey 7, 2017, 15–16. Khaburzania – Robinson 2018 G. Khaburzania – A. Robinson, Nature and local knowledge combine in powerful strongholds. Preliminary findings of the archaeological survey in Samtskhe-Javakheti project, in: Batmaz et al. 2018, 303–324. Khatchadourian 2008 L. Khatchadourian, Making nations from the ground up. Traditions of classical archaeology in the South Caucasus, American Journal of Archaeology 112, 2, 2008, 247–278. Khatchadourian 2013 L. Khatchadourian, An archaeology of hegemony. The Achaemenid Empire and the remaking of the fortress in the Armenian Highlands, in: G. E. Areshian (ed.) Empires and Diversity. On the Crossroads of Archaeology, Anthropology, and History (Los Angeles 2013) 108–145. Khatchadourian 2014 L. Khatchadourian, Empire in the everyday: a preliminary report on the 2008–2011 excavations at Tsaghkahovit, Armenia, American Journal of Archaeology 118, 1, 137–169. Kohl 1988 P. Kohl, The northern “frontier” of the Ancient Near East. Transcaucasia and Central Asia compared, American Journal of Archaeology 92, 4, 1988, 591–596. Kohl – Fawcett 1995 P. Kohl – C. Fawcett (eds.), Nationalism, Politics and the Practice of Archaeology (Cambridge 1995) 139–148. doi: 10.1017/CBO9780511558214 Kohl – Tsetskhladze 1995 P. Kohl – G. Tsetskhladze, Nationalism, politics, and the practice of archaeology in the Caucasus, in: Kohl – Fawcett 1995, 149–174. Lawrence – Wilkinson 2017 D. Lawrence – T. J. Wilkinson, The northern and western borderlands of the Sasanian Empire. Contextualizing the Roman/Byzantine and Sasanian frontier, in: E. W. Sauer (ed.), Sasanian Persia between Rome and the Steppes of Eurasia (Edinburgh 2017) 99–125. Lindsay – Greene 2013 I. Lindsay – A. Greene, Sovereignty, mobility, and political cartographies in Late Bronze Age Southern Caucasia, Journal of Anthropological Archaeology 32, 2013, 691–712. Lindsay – Smith 2006 I. Lindsay – A. T. Smith, A history of archaeology in the Republic of Armenia, Journal of Field Archaeology 31, 2, 2006, 165–184. Lozny 2017 L. R. Lozny (ed.), Archaeology of the Communist Era. A Political History of Archaeology of the 20th Century (Cham 2017). Lyonnet et al. 2012 B. Lyonnet – F. Guliyev – B. Helwing – T. Aliyev – S. Hansen – G. Mirtskhulava, Ancient Kura 2010–2011. The first two seasons of joint field work in the Southern Caucasus, Archäologische Mitteilungen aus Iran und Turan 44, 2012, 1–196. Marro 2004 C. Marro, Upper Mesopotamia and the Caucasus. An essay on the evolution of routes and road networks from the Old Assyrian Kingdom to the Ottoman Empire, in: A. Sagona (ed.), A View from the Highlands (Louvain 2004) 91–120.

20

Kristen Hopper – William Anderson – Abby Robinson

McPhillips – Wordsworth 2016 S. McPhillips – P. Wordsworth (eds.), Landscapes of the Islamic World. Archaeology, History, and Ethnography (Philadelphia 2016). Myers 2010 A. Myers, Camp Delta, Google Earth and the ethics of remote sensing in archaeology, World Archaeology 42, 2010, 455–467. Negus Cleary et al. 2018 M. Negus Cleary – N. Tskvitinidze – W. Anderson – D. Krsmanovic – J. Birkett-Rees, Mapping the Vilayet of Gurjistan. The southern Caucasus highlands in the late medieval–early Ottoman periods, in: Batmaz et al. 2018, 381–405. Petrosyan et al. 2015 A. Petrosyan ‒ R. Dan ‒ R. La Farina ‒ M. Raccidi ‒ M. Castelluccia ‒ B. Gasparyan ‒ A. Babajanyan, The Kotayk Survey Project (KSP). Preliminary report on 2014 fieldwork activity, Aramazd. Armenian Journal of Near Eastern Studies 9, 1, 2015, 58‒68. Rapp 2012 S. Rapp, The Sasanian world through Georgian Eyes. Caucasia and the Iranian Commonwealth in Late Antique Georgian Literature (Farnham 2012). Ricci 2012 A. Ricci, Archaeological landscape studies. The Mil-Qarabağ plain and the Kvemo Kartli survey projects. A preliminary account of the first two field seasons (2010–2011), contribution in: Lyonnet et al. 2012, 127–145. Ristvet et. al. 2011 L. Ristvet – V. Baxşeliyev – S. Aşurov, Settlement and society in Naxçivan. 2006 excavations and survey of the Naxçivan Archaeological Project, Iranica Antiqua 46, 2011, 1–53. Ristvet et al. 2013 L. Ristvet – V. Bakhshaliyev – H. Gopnik – S. G. Ashurov, The origins of political complexity in Naxçıvan. Excavations and survey at Oğlanqala 2008–2010, in: A. Mehnert – G. Mehnert – S. Reinhold (eds.), Austausch und Kulturkontakt im Südkaukasus und seınen angrenzenden Regıonen in der Spätbronze-/Früheisenzeit, Schriften des Zentrums für Archäologie und Kulturgeschichte des Schwarzmeerraumes 22 (Langenweißbach 2012) 281–290. Rova et al. 2011 E. Rova – M. Puturidze – Z. Makharadze, The Georgian-Italian Shida Kartli archaeological Project. A report on the first two field seasons 2009 and 2010, Rivista di Archeologia 34, 2011, 5–30. Rubinson – Smith 2003 K. S. Rubinson – A. T. Smith, Archaeology in the Borderlands, in: A. T. Smith – K. Rubinson (eds.), Archaeology in the Borderlands. Investigations in Caucasia and Beyond (Los Angeles, CA 2003) 1–8. Sagona 2010 A. Sagona, Past and present directions in the archaeology of the Transcaucasus / Transkafkasya Arkeolojisinde Geçmiş ve Güncel Yönelimler, Türkiye Bilimler Akademisi Arkeoloji Dergisi 13, 2010, 143–157. Sagona 2017 A. Sagona, The Archaeology of the Caucasus. From Earliest Settlements to the Iron Age (Cambridge 2017). Sauer et al. 2015 E. W. Sauer – K. Pitskhelauri – K. Hopper – A. Tiliakou – C. Pickard – D. Lawrence – A. Diana – E. Kranioti – C. Shupe, Northern outpost of the Caliphate. Maintaining military forces in a hostile environment (the Dariali Gorge in the Central Caucasus in Georgia), Antiquity 89, 2015, 885–904. Shnirelman 1995 V. A. Shnirelman, From internationalism to nationalism. Forgotten pages of Soviet archaeology in the 1930s and 1940s, in: Kohl – Fawcett 1995, 120–138. Smith 2005 A. T. Smith, Prometheus unbound. Southern Caucasia in prehistory, Journal of World Prehistory19, 4, 2005, 229–279.

Landscape Archaeology in Southern Caucasia: An Introduction

21

Smith et al. 2009 A. T. Smith ‒ R. S. Badalyan ‒ P. Avetisyan, The Foundations of Research and Regional Survey in the Tsaghkahovit Plain, Armenia, The Archaeology and Geography of Ancient Transcaucasian Societies 1 (Chicago 2009). Stöllner – Gambashidze 2014 T. Stöllner – I. Gambashidze, The gold mine of Sakdrisi and the earliest mining and metallurgy in the Transcaucasus and the Kura-valley system, in: G. Narimanishvili (ed.), Problems of Early Metal Age Archaeology of Caucasus and Anatolia (Tbilisi 2014) 102–124. Taylor et al. 2011 P. M. Taylor – C. R. Polgase – N. Museyibli – J. M. Koller – T. A. Johnson, Past and Future Heritage in the Pipelines Corridor. Azerbaijan, Georgia, Turkey (Washington D.C. 2011). Turner – Crow 2010 S. Turner – J. Crow, Unlocking historic landscapes in the Eastern Mediterranean. Two pilot studies using Historic Landscape Characterisation, Antiquity 84, 323, 216–229. Wilkinson 2003 T. J. Wilkinson, Archaeological Landscapes of the Near East (Tucson 2003).

Moving in the Mountains

23

Moving in the Mountains: GIS and Mapping the Phenomenology of Travel through the South Caucasus Lara Fabian * Abstract: Running across the isthmus that separates the Near East from the Eurasian Steppe, the Caucasus Mountains dominate both the physical and imagined landscapes of the surrounding regions, shaping the lives of those in their shadow. Perceptions about mountains and mountain peoples have long undergirded the study of the South Caucasus, but recent developments in the study of landscape offer new avenues for considering interactions between humans and mountains. Of particular interest are the affordances that the rugged topography offers those moving around the zone, since mountains both constrain motion but also offer opportunities to those who know them well. This paper examines broad-scale movement possibilities through the South Caucasus, using Least Cost Path (LCP) networks to model movement corridors and to consider the experiential interaction between people and mountainous zones. These analyses provide insights into the relative ease or difficulty of pursuing particular routes. When used in conjunction with other sources of archaeological and historical data, the LCP networks help to tease out the various choices available to local residents and external actors as they navigated the challenging topography. Keywords: Environment; GIS; Landscape archaeology; Least Cost Path; Mountains; South Caucasus

Introduction: Mountains and Mobility in the South Caucasus From the medieval Silk Road to today’s oil pipelines, the South Caucasus has long been a conduit linking the Near East, Anatolia and the Eurasian Steppe. It is a crossroads par excellence. At the same time, the region is often defined in terms of its liminality: on the edge of empires ancient and modern, an unremitting frontier at the dividing line of the continents. Paradoxically, the centrality of the South Caucasus as a transit space stems from the fact that it seems to lie perpetually along the seams of an empire. This paper argues that the defining topographical feature of the region – its mountains – is the key to understanding both its liminality and its connective centrality. Although the concepts of ‘crossroads’ and ‘conduits’ imply mobility and connectivity, archaeological and historical scholarship of the South Caucasus has tended to downplay both intraand inter-regional interaction. To some extent, this is a result of geographic divisions within the discipline that have relegated the zone to a perpetual periphery,1 which is the terminus of others’ space rather than an actor in its own right. Meanwhile, as a consequence of modern geopolitics, macro-scale patterns entwining the territory’s physical geography and its social and political life in antiquity have been obscured in favour of the historicising study of a sequence of polities (Urartu, Caucasian Albania, Caucasian Iberia, Armenia etc.), abutting each other either spatially or temporally, controlling a fixed territory.2 Thinking about movement through space is a valuable step in pushing against these tendencies and developing more nuanced cartographies.3 Given the topography, this requires a consideration of how mountains shape mobility. Although often seen as inhibiting movement, the constraints introduced by the rugged topography actually funnel travellers along specific routes and may

* 1 2 3

Seminar für Alte Geschichte, Albert-Ludwigs-Universität Freiburg, [email protected]. A. Smith 2005. For the history of this as it relates to the Classical period, see Khatchadourian 2008. On the problems with traditional maps of ancient polities, see M. Smith 2005.

24

Lara Fabian

“encourage greater communication than a broad open plain”.4 To model this, I draw on the concept of ‘friction of terrain’ from the work of the anthropologist James Scott.5 The idea recognises that control of different types of terrain requires different degrees of energy investment. Highlands are particularly resource-intensive because they are difficult to traverse and therefore difficult to supervise. Friction of terrain is not based solely on topography but is also socially produced, shaped by the interaction between intrinsic landscape characteristics and efforts to either minimise friction through distance-reducing technologies (such as bridges or paved roads), or maximise it through the obstruction of such technologies.6 It does not apply equally to all agents: pedestrians with good knowledge of the terrain will face less friction than a state organism using wheeled transportation to move foodstuffs.7 Friction of terrain is therefore a type of ‘geographic resource’ available to highland populations seeking to limit access to and control of their territory.8 Although the concept has been applied most often to highland territories’ attempts to elude imperial power, I suggest that the residents and political authorities in the South Caucasus leveraged their topography both to resist external forces and to engage with their neighbours. This paper begins by exploring understandings of mountainous environments in science and the social imagination, framing the ways that these have moulded perceptions about the South Caucasus as well as lifeways within it. Then, I discuss the potentials of a GIS-based analysis which uses a slope-based measure of friction of terrain to model route networks through the mountains, identifying spaces that attract movement. This technique uses quantitative approaches to develop essentially phenomenological cartographies that reflect interactions between humans and their environment.9 The resulting insight into the human experience of highland space can be triangulated against other bodies of archaeological and historical data, developing rich context for the choices that the people and societies of the South Caucasus have made in navigating their rugged landscape. The ‘Highland Imaginary’ and Mountain Realities Mountainous territories have recently been the subject of increased interest, as climate change has called attention to the importance of their ecosystems.10 Archaeologists working in the Mediterranean and Near East have been slow to participate in this conversation, with highland research falling outside of the traditional sphere of interest and presenting technical challenges for the discipline.11 Therefore, although discussions of mountains have been a mainstay of Urartian studies for some time,12 their full explanatory power in the broader study of the South Caucasus is still on the rise.13 Despite their low visibility in modern scholarship, mountains have been a topic of conversation since the time of Herodotus, with the South Caucasus featuring from the earliest accounts. These popular approaches to highlands have had lasting ramifications. In Greek and Latin literature, mountains were understood most basically as territorial borders. This is clearest in the work of Strabo, for whom mountains, rivers and seas divided the world into its constituent parts.14 They were also seen

4 5 6 7 8 9 10 11 12

13 14

Rubinson – Smith 2003, 5. Scott 2009. Scott 2009, 166. Scott 2009, 43–44. Sorge 2014, 37. On using GIS to examine the affordances of physical geography, see Llobera 2001; Gillings 2012. Messerli – Ives 1997b; UNEP WCMC 2002. Glatz – Casana 2016. For example, Zimansky 1985; Zimansky 1995; Biscione 2002. Contrast this to the situation in the Andes, where archaeologists have been developing and debating theories about the interplay between altitude and cultural development for many decades. For the roots of this Andean research, see Van Buren 1996, and for a recent overview, see Contreras 2010. For example, Lindsay – Greene 2013; Hammer 2014. Strabo, II.5.17; Pothecary 2005, 177.

Moving in the Mountains

25

as technical challenges to be mastered by a fit society.15 At the same time, in antiquity’s ‘geographic imaginary’, highlands served a tropic function as ‘an alter-ego of urban civilization’.16 This manifested itself in the opposition of lowland and highland peoples, with the former seen as the standard bearers of civilisation, while the latter were warlike bands threatening stability.17 Furthermore, particularly at the edges of the known world, highlands were spaces of alterity where men came face to face with myth.18 In Greek and Roman thought broadly, then, mountains had a dual identity. They were manifestly physical, able to dominate and define space through their sheer mass; but they also had a spiritual reality that derived from their status as ‘othered’ places. The depiction of the South Caucasus in Greek and Latin sources is archetypal. The Caucasus range, according to Herodotus, has ‘more and higher mountains than any other range’ (ἐὸν ὀρέων καὶ πλήθεϊ μέγιστον καὶ μεγάθεϊ ὑψηλότατον).19 In Strabo’s tour through the South Caucasus, the ‘warlike portion’ (οἱ … μάχιμοι) of Iberian society has its home in the highlands,20 in contrast to their more civilised neighbours dwelling in the lowlands.21 The highlands were also the territory of the Amazons, who even Strabo suggests may be more myth than reality.22 Echoes of these perceptions are clear even in relatively recent literary presentations of the territory, with the Caucasus acting as a ‘land of fairytales’ in Russian thought of the nineteenth century.23 Most directly, it was through the works of Russia’s national poet, Alexander Pushkin, that the physical territory of the Caucasus came to serve as a modern metonym for romantic notions about mountain people: their freedom, brutality and fierce otherness.24 This literary landscape has deemphasised the actual strategies of people grappling with their physical space, favouring instead mountain tropes. More recently, modern natural sciences research has brought new approaches to the study of mountains, articulating specific ways in which mountains differ from other landforms and the ramifications of these differences on cultural processes. One of the most striking characteristics of mountains is the disproportionate ecological diversity of highlands, which can contain disparate ecological zones in extremely close physical proximity because of their range of altitudes.25 Highlands are hotbeds not only of biodiversity but also of cultural and linguistic diversity.26 And they are, at least today, uniquely vulnerable to conflict: over half of all wars and armed conflicts in the late twentieth century involved mountainous areas.27 The Caucasus exhibits many of these characteristics. Conflict, a defining literary trope since antiquity, remains an ongoing political concern in the region.28 Linguistic diversity has also been a persistent hallmark. The tenth-century Arab geographer al-Masʿudi deemed the area jabal alalsun, ‘the mountain of tongues’.29 A millennium earlier, Strabo mentions that twenty-six languages were spoken in the territory of Albania.30 Recent work on ‘vertical bilingualism’ in the Caucasus, meanwhile, is explicit in connecting its unique linguistic landscape to the patterns of mobility in the physical landscape.31

15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Gschnitzer 1996. Meissner 1996, 369. For example, Shaw 1990. Evans 1999. Herodotus, I.203. Strabo, XI.1–5. Strabo, XI.3.3. Strabo, XI.5.1, 5. Layton 1986, 483. Layton 1986; Hokanson 1994; Grant 2005. Jeník 1997; Körner 2002. Stepp et al. 2005. Libiszewski – Bächler 1997. Hunter 2006; Sagramoso 2007. Catford 1977, 283. Strabo, XI.4.6. Nichols 2014, 41. For other recent scholarship on the linguistics of the Caucasus, see Comrie 2008.

26

Lara Fabian

Amid this flurry of general mountain research, however, the precise definition of a ‘mountainous region’ has proved elusive. Mountains are, after all, relative: a 100m elevation along a coastal plain may be deemed a ‘mountain’ by local residents but would not merit notice were it found in the Alps. It was only in the late twentieth century that researchers began to attempt a quantitative answer for the question, ‘what makes a mountain a mountain’?32 Developed using newly available global elevation data, one widely cited global mountain model combines absolute elevation (all land over 2500m) with slope and terrain roughness data to automatically classify both high peaks and lower-elevation mountains for the entire planet.33 Despite this there is a sense among those who study mountains that such quantitative definitions do not capture the essence of the landform. The most recent textbook on mountains said bluntly, ‘a landform is considered a mountain when local people rate it as such’.34 Or, in the words of the premier early-twentieth-century mountain scholar: ‘A mountain, strictly speaking, is a conspicuous elevation of small summit area. A plateau is a similar elevation of larger summit area with at least one sheer side. An essential and yet indefinite element in the definition of a mountain is the conspicuity. Conspicuity, like height, is a relative matter, and depends upon the personal evaluation or the standard by which it is measured… Mountains should be impressive; they should enter into the imagination of the people within their shadows.’35 The issue, at first glance, is whether a ‘mountain’ is the product of absolute or relative topography. But, underpinning this is the idea that phenomenological encounters between mountains and humans are central to the status of the landform. A mountain is not a mountain purely because of its topography, but because of the emotional response of humans to the topography. This is, in a sense, analogous to the understanding of mountainous spaces within Greek and Latin literature, which rated both the physical mass and the spiritual power as intrinsic characteristics. The persistence of this duality suggests that there is an essential experiential component to mountains. Modelling Movement Potentials Friction of terrain is one attempt to grapple with human encounters with these spaces: to capture the phenomenological dimensions of movement. Although the concept was largely deployed qualitatively by Scott,36 a slightly different usage of ‘friction’ has featured in quantitative analyses of human movement for the last two decades, through the study of Least Cost Paths (LCPs). LCP analysis uses GIS to model movement choices through a landscape by analysing travel decisions in terms of time and difficulty of transit, assuming that an iterative decision-making process identifies the most advantageous route.37 In this system, friction is calculated as a function of slope and, together with other travel-impeding or travel-attracting features of landscape including ecological and even cultural elements,38 it becomes a ‘cost surface’ across which the least cost route for a specific means of transit can be algorithmically derived. Recently, LCP analysis has turned its attention from the consideration of single routes to the study of LCP networks.39 These network models can be in the form of ‘from everywhere to everywhere’ Cumulative Cost Path (CCP) maps, which reflect travel potential across a space.40

32 33 34 35 36 37 38 39 40

Funnell 2001, ch. 1; Debarbieux 2009. Kapos et al. 2000. Price et al. 2013, 6. Peattie 1936, 3–4. Although, see Scott 2009, ch. 2. Bell – Lock 2000; Conolly – Lake 2006; Howey 2007; Llobera et al. 2011; Surface-Evans – White 2012. Herzog 2013a, 184. Whitley – Burns 2008; Llobera et al. 2011; Murrieta-Flores 2012; White – Barber 2012; Herzog 2013b; Verhagen 2013. White – Barber 2012.

Moving in the Mountains

27

Fig. 1   Comparison of South Caucasus maps showing elevation (left) and slope-based cost surface (right). Darker values represent higher elevation (left) or higher cost (right) (graphics: L. Fabian)

Fig. 2   Grid of study points across landscape (graphics: L. Fabian)

Fig. 3   Least Cost Paths from each study point to each other study point (graphics: L. Fabian)

A CCP map aggregates a large number of LCPs, recording the number of LCPs to traverse each point in the landscape – the frequency with which that route appears in the model. High frequency routes are travel-attracting corridors, loci of movement favourable for pedestrian traffic. This technique is particularly promising in the case of mountainous topographies, where routes are restricted in marked ways. The construction of such a network can for the South Caucasus be broken down into the following steps:41 1. Creating a cost surface that reflects terrain friction and any other relevant factors through the zone such as hydrology (Fig. 1).42 2. Seeding non-site sample points across the zone (Fig. 2). 3. Generating LCPs for movement from each study point to every other one across this cost surface (Fig. 3). 4. Aggregating all the LCPs and ranking them according to how often a specific path was chosen to generate the CCP map.

41 42

See also description in Verhagen 2013. The level of the Caspian Sea can fluctuate dramatically within a short timespan. The analysis presented here is based on the current sea level of ca. –27 mASL. A highstand of ca. –25–23 mASL seems to have occurred ca. 500 BC (Kislov et al. 2014, 51; Kroonenberg et al. 2008, 30). This would have changed the coastline of the Kura-Aras lowlands, but without serious consequences for the analysis presented here. For a summary of recent paleoclimate research on Caspian Sea, see Kislov et al. 2014. For dramatic hypotheses about coastline changes based on ancient literary source, see Murav’ev 1983. If accurate, this reconstruction would present problems for the model described here, however recent research on historic Caspian Sea levels argues against Murav’ev’s reconstruction.

28

Lara Fabian

Fig. 4   One possible Cumulative Cost Path map for the South Caucasus, showing aggregated LCPs along with presumed capital cities of the Parthian-Roman period polities of Armenia, Caucasian Iberia and Caucasian Albania. Here ‘frequency’ reflects the number of LCPs that run through a given point in space. Thicker, darker routes correspond to higher numbers of LCPs (graphics: L. Fabian)

One possible CCP model for pedestrian travel in the South Caucasus is shown in Figure 4.43 But, before discussing observations gleaned from the model, a few words about its limitations are in order. LCP analysis assumes that humans moving in a landscape are following optimal routes, which are thought to have been ‘learned’ over repeated encounters with the landscape.44 However, poor judgment of topographical characteristics such as slope and distance together with incomplete geographical knowledge of human agents could have prevented route optimisation, particularly in environments where travel was more sporadic.45 Furthermore, and especially relevant to the current investigation, the dominant algorithms for quantifying pedestrian choices may not accurately handle the exigencies of especially rugged topographies.46

43

44 45 46

The model was constructed using elevation data drawn from the 3 arc-second SRTM DEM available from the USGS, which was resampled to a resolution of 200m because of processing limitations involved in calculations over such a large zone (1100 × 850km). The cost surface took into account slope using Tobler’s hiking equation (Tobler 1993) to assign slope-based movement costs using the Path Distance function. It also included hydrological data as well as a preferential inducement for movement through mountain passes, discussed below. 518 study points were seeded in a regular grid across the zone (see fig. 2), and the final CCP map was an aggregation of over 268,000 LCPs generated using ArcGIS’s Flow Accumulation tool including both short- and long-distance paths (see fig. 3). Surface-Evans – White 2012, 2; Herzog 2013a, 180. Freundschuh 1998. Pingel 2010.

Moving in the Mountains

29

In the construction of this particular South Caucasus model, the combination of cost structure used to analyse slope costs, the steepness of the Caucasus and the coarseness of the elevation data led initially to implausible LCPs.47 To mitigate these problems, the cost surface was then modified to include a preferential treatment of mountain passes.48 The resultant model reflects travel-attracting corridors across the territory for periods of the year when the mountain passes were traversable, with the assumption that people were using those passes. The model does not distinguish between populations (local and non-local, for example) and does not reflect distancereducing interventions, such as the building of bridges. Advances in the quantification of human movement in the future will undoubtedly reveal more accurate information about travel potentials in the territory. But even the presently available techniques give a glimpse into some of the logics of the space. Movement Corridors and Historical Route Systems Unlike accounts of travel through the South Caucasus based on written sources, which privilege very specific types of movement (usually either military campaigns or long distance trade activity), the cumulative cost path map is a general index of the phenomenon of movement.49 It is, in a sense, democratic. It does not describe routes per se, but rather represents how accessible a path was for those who chose to use it, relative to other available options. The South Caucasus CCP model picks out east–west riverine corridors along the Kura and Aras rivers as the most conducive spaces for movement in the zone. Intersecting with these, a coastal path along the Caspian as well as a mountain pass through the Greater Caucasus (through the Dariali Gorge) are the most accessible routes leading north from the South Caucasus.50 Both of these north–south routes are well attested in ancient and more modern accounts of travel through the territory, although each has seasonal constraints. The model also suggests that movement can flow through a number of smaller north–south passes through the Greater Caucasus.51 While the Caspian coast attracted movement, the coast of the Black Sea was less favourable.52 The Lesser Caucasus in general are more accessible than the Greater Caucasus, leading to the exploitation of a wider variety of paths through those mountains. Nevertheless, the Aragats Plain serves as a focal point of movement, at the nexus of both north–south and east–west routes. Incorporating historical and archaeological data is critical to the deployment of the South Caucasus CCP model. Written accounts of travel through the Caucasus from antiquity are rare, maps still rarer, and the identification of toponyms presents challenges.53 Nevertheless, these sources have been leveraged in the construction of speculative road and route maps.54 A diachronic view of the reconstructed routes suggests that east–west axes across the South Caucasus were relatively

47

48

49

50

51 52 53 54

The initially calculated paths almost never traversed the Greater Caucasus, not even taking advantage of the largest and best-known mountain pass at Dariali Gorge. Given the ample archaeological and ethnographic evidence for the exploitation of this pass, the model was failing to capture actual movement patterns. The passes were identified using Jenness’ Topographic Position Index (2006), a landform classification tool. Landform classification uses comparative topographical characteristics to identify passes rather than ethnographic accounts. The identified spaces were then given an artificially low friction. But, see fn. 8 on the difficulty of integrating GIS with landscape phenomenology approaches in the spirit of Tilley (1994). Coastline changes would have altered the course of the coastal route, but iterations of the model run on different coastline configurations suggest that the presence of such a route is persistent. See fn. 42. The use of some of these passes in antiquity has been suggested: see Braund 1994, 44. Supporting Marro’s (2004) suggestion about the importance of maritime networks in this area. Murav’ev 2014. Manandian 1954; Hewsen 2001; Marro 2004.

30

Lara Fabian

stable over time,55 with the most persistent route following the Aras River, while another route often cut across the Kura Plain.56 There is more temporal variation in the choice of north–south routes, particularly through the Lesser Caucasus. This can be explained, perhaps, by the difficulty of north–south movement expressed in the CCP analysis. In the absence of any truly good choices, there was more room for experimentation based on the priorities and needs of different times and travellers. Interestingly, two of the routes that can be identified on the Late Antique Peutinger Table through the South Caucasus were north–south corridors running from Artashat into Caucasian Iberia.57 The exploitation of these routes reflects a distinct energy expenditure that correlated to their importance for imperial actors motivating the official route system, although these routes are rarely mentioned in later accounts of travel and are relatively inaccessible in the CCP analysis. The absence of most of the riverine routes from the Peutinger Table, meanwhile, is curious. In any case, itineraries represent an imperial vision of space. It is harder to access the local use patterns of the space cartographically. But, if one simply considers the position of the capital cities of the dominant Parthian-Roman period political authorities in the territory (Caucasian Iberia, Caucasian Albania and Armenia), one sees that the local polities were relating to these corridors in starkly different ways (Fig. 4). Mtskheta and Artaxata, the capitals of Caucasian Iberia and Armenia respectively, were both located at or near nodal points of route corridors. In contrast, Albania’s presumed capital (Gabala) was remote from any corridor, along less prominent routes.58 Different logics, clearly, guided the occupation of these spaces. Conclusions The experiential relationship between people and mountains is central to life in the South Caucasus, and therefore to our study of it. In this paper, I have focused on movement patterns, but I would argue that the affordances of the topography reach farther than this – interacting with all facets of cultural production. However, movement patterns are a good starting point. Thanks to advances in computing power and the development of quantitative studies of human movement, we have techniques that allow us to build heuristic models that explore how the topography may have been experienced. We can then use these heuristic models together with archaeological and historical datasets, creating new avenues for understanding how ancient populations were interacting with their space. Future consideration of wider bodies of archaeological material within the framework of movement potentials will help to clarify the varying goals of these spatial logics. In the case of the South Caucasus, this brief exploration of potentials together with historical route choices points out that we should pay particular attention to north–south routes, as the variability of choice makes these more expressive of the priorities of users. The north–south routes, many of which traverse smaller mountain passes, could have been the site of local territorial power, with residents able to help or hinder those attempting to pass through. This examination also highlights the different positions occupied by local polities, with some located at accessible juncture points and others positioned along less popular routes. Territorial friction, after all, is not predictive, but is rather a geographic resource that can be strategically deployed in pursuit of different goals.

55 56 57 58

Marro 2004, maps 1–6. Marro 2004, 106. Hewsen 2001, 65; see also the work of Manandian 1954. For the possible importance of these less prominent routes, see Gregoratti 2013; Fabian 2017.

Moving in the Mountains

31

References Bell – Lock 2000 T. Bell – G. Lock, Topographic and cultural influences on walking the Ridgeway in later prehistoric times, in: G. Lock (ed.), Beyond the Map. Archaeology and Spatial Technologies (Amsterdam 2000) 85–100. Biscione 2002 R. Biscione, The Iron Age settlement pattern. Pre- Urartian and Urartian periods, in: R. Biscione – S. Hmayakyan – N. Parmegiani (eds.), The North-Eastern Frontier. Urartians and Non-Urartians in the Sevan Lake Basin (Rome 2002) 351–370. Braund 1994 D. Braund, Georgia in Antiquity. A History of Colchis and Transcaucasian Iberia, 550 BC–AD 562 (Oxford 1994). Catford 1977 J. C. Catford, Mountain of Tongues. The languages of the Caucasus, Annual Review of Anthropology 6, 1977, 283–314. Comrie 2008 B. Comrie, Linguistic diversity in the Caucasus, Annual Review of Anthropology 37, 1, 2008, 131–143. Conolly – Lake 2006 J. Conolly – M. Lake, Geographical Information Systems in Archaeology (New York 2006). Contreras 2010 D. A. Contreras, Landscape and environment. Insights from the prehispanic Central Andes, Journal of Archaeological Research 18, 3, 2010, 241–288. Debarbieux 2009 B. Debarbieux, Mountains. Between pure reason and embodied experience. Philippe Buache and Alexander von Humboldt, in: D. Cosgrove – V. della Dora (eds.), High Places. Cultural Geographies of Mountains, Ice and Science (London, New York 2009) 87–104. Evans 1999 R. Evans, Ethnography’s freak show. The grotesques at the edges of the Roman earth, Ramus 28, 1, 1999, 54–73. Fabian 2017 L. Fabian, Numismatic communities in the northern South Caucasus 300 BCE – CE 300. A geospatial analysis of coin finds from Caucasian Iberia and Caucasian Albania, in: H. Teigon – E. Seland (eds.), Sinews of Empire (Oxford 2017) 37–69. Freundschuh 1998 S. M. Freundschuh, The relationship between geographic scale, distance and time as expressed in natural discourse, in: M. J. Egenhofer – R. G. Golledge (eds.), Spatial and Temporal Reasoning in Geographic Information Systems (New York 1998) 131–142. Funnell 2001 D. C. Funnell, Mountain Environments and Communities (New York 2001). Gillings 2012 M. Gillings, Landscape phenomenology, GIS and the role of affordance, Journal of Archaeological Method and Theory 19, 4, 2012, 601–611. Glatz – Casana 2016 C. Glatz – J. Casana, Of highland-lowland borderlands. Local societies and foreign power in the Zagros-Mesopotamian interface, Journal of Anthropological Archaeology 44, Part A, 2016, 127–147. Grant 2005 B. Grant, The good Russian prisoner. Naturalizing violence in the Caucasus Mountains, Cultural Anthropology 20, 1, 2005, 39–67. Gregoratti 2013 L. Gregoratti, The Caucasus. A communication space between nomads and sedentaries (1st BC–2nd AD), in: S. Magnani (ed.), Le aree montane come frontiere. Spazi d’interazione e connettività (Rome 2013) 525–540.

32

Lara Fabian

Gschnitzer 1996 F. Gschnitzer, Das Gebirge im Geschichtswerk des Polybios, in: E. Olshausen – H. Sonnabend (eds.), Gebirgsland als Lebensraum (Amsterdam 1996) 173–187. Hammer 2014 E. Hammer, Highland fortress-polities and their settlement systems in the Southern Caucasus, Antiquity 88, 341, 2014, 757–774. Herodotus Herodotus, Herodoti Historiae, N. G. Wilson (ed.) (Oxford 2015). Herzog 2013a I. Herzog, The potential and limits of optimal path analysis, in: A. Bevan – M. Lake (eds.), Computational Approaches to Archaeological Spaces (Walnut Creek, CA 2013) 179–211. Herzog 2013b I. Herzog, Least-cost networks, in: G. Earl – T. Sly – A. Chrysanthi – P. Murrieta-Flores – C. Papadopoulos – I. Romanowska – D. Wheatley (eds.), Archaeology in the Digital Era. Papers from the 40th Annual Conference of Computer Applications and Quantitative Methods in Archaeology (CAA) (Amsterdam 2013) 237–248. Hewsen 2001 R. H. Hewsen, Armenia. A Historical Atlas (Chicago 2001). Hokanson 1994 K. Hokanson, Literary imperialism, narodnost’ and Pushkin’s invention of the Caucasus, Russian Review, 1994, 336–352. Howey 2007 M. C. L. Howey, Using multi-criteria cost surface analysis to explore past regional landscapes. A case study of ritual activity and social interaction in Michigan, AD 1200–1600, Journal of Archaeological Science 34, 11, 2007, 1830–1846. Hunter 2006 S. Hunter, Borders, conflict, and security in the Caucasus. The legacy of the past, SAIS Review 26, 1, 2006, 111–125. Jeník 1997 J. Jeník, The diversity of mountain life, in: Messerli – Ives 1997a, 199–235. Jenness 2006 J. Jenness, Topographic Position Index (TPI) v. 1.3a. Extension for ArcView 3.x, 15 Sept. 2006. Online (last accessed 10 Jan 2018>. Kapos et al. 2000 V. Kapos – J. Rhind – M. Edwards – M. F. Price – C. Ravilious, Developing a map of the world’s mountain forests, in: M. F. Price – N. Butt (eds.), Forests in Sustainable Mountain Development. A State of Knowledge Report for 2000, IUFRO Research Series 5 (Wallingford 2000) 4–19. Khatchadourian 2008 L. Khatchadourian, Making nations from the ground up. Traditions of Classical Archaeology in the South Caucasus, American Journal of Archaeology 112, 2, 2008, 247. Kislov et al. 2014 A. V. Kislov – A. Panin – P. Toropov, Current status and palaeostages of the Caspian Sea as a potential evaluation tool for climate model simulations, Quaternary International 345, 2014, 48–55. Körner 2002 C. Körner, Mountain biodiversity, its causes and function. An overview, in: C. Körner – E. M. Spehn (eds.), Mountain Biodiversity. A Global Assessment (New York 2002) 3–20. Kroonenberg et al. 2008 S. B. Kroonenberg – N. S. Kasimov – M. Y. Lychagin, The Caspian Sea, a natural laboratory for sea-level change, Geography, Environment, Sustainability 1, 2008, 22–37.

Moving in the Mountains

33

Layton 1986 S. Layton, The creation of an imaginative Caucasian geography, Slavic Review 45, 3, 1986, 470–485. Libiszewski – Bächler 1997 S. Libiszewski – G. Bächler, Conflicts in mountain areas. A predicament for sustainable development, in: Messerli – Ives 1997a, 103–130. Lindsay – Greene 2013 I. Lindsay – A. Greene, Sovereignty, mobility, and political cartographies in Late Bronze Age southern Caucasia, Journal of Anthropological Archaeology 32, 4, 2013, 691–712. Llobera 2001 M. Llobera, Building past landscape perception with GIS. Understanding topographic prominence, Journal of Archaeological Science 28, 9, 2001, 1005–1014. Llobera et al. 2011 M. Llobera – P. Fábrega-Álvarez – C. Parcero-Oubiña, Order in movement. A GIS approach to accessibility, Journal of Archaeological Science 38, 4, 2011, 843–851. Manandian 1954 Я. А. Манандян, О торговле и городах Армении в связи с мировой торговлей древних времен (V в. до н. эры – ХV в. н. эры) ( Yerevan 1954). Marro 2004 C. Marro, Upper Mesopotamia and the Caucasus. An essay on the evolution of routes and road networks from the Old Assyrian Kingdom to the Ottoman Empire, in: A. Sagona (ed.), A View from the Highlands (Louvain 2004) 91–120. Meissner 1996 B. Meissner, Vorstellung der Griechen von den Bergen, in: E. Olshausen – H. Sonnabend (eds.), Gebirgsland als Lebensraum (Amsterdam 1996) 351–369. Messerli – Ives 1997a B. Messerli – J. D. Ives (eds.), Mountains of the World. A Global Priority (New York 1997) Messerli – Ives 1997b B. Messerli – J. D. Ives, Preface, in: Messerli – Ives 1997a, vi–xi. Murav’ev 1983 С. Н. Муравьев, Птолемеева карта Кавказской Албании и уровень Каспия, Вестник древней истории 163, 1, 1983, 117–147. Murav’ev 2014 С. Н. Муравьев, От Сурами до Цнори. Дороги античной Восточной Грузии по Tabula Peutingeriana и Космографии Равеннского анонима, in: Т. Н. Джаксон – А. В. Акопян (eds.), Πολύτροπος. Сборник научных статей памяти Аркадия Анатольевича Молчанова (1947–2010) (Moscow 2014) 55–82. Murrieta-Flores 2012 P. Murrieta-Flores, Understanding human movement through spatial technologies. The role of natural areas of transit in the Late Prehistory of South-western Iberia, Trabajos de Prehistoria 69, 1, 2012, 103–122. Nichols 2014 J. Nichols, The vertical archipelago. Adding the third dimension to linguistic geography, in: P. Auer – M. Hilpert – A. Stukenbrock – B. Szmrecsanyi (eds.), Space in Language and Linguistics (Berlin 2014) 38–60. Peattie 1936 R. Peattie, Mountain Geography. A Critique and Field Study (London 1936). Pingel 2010 T. J. Pingel, Modeling slope as a contributor to route selection in mountainous areas, Cartography and Geographic Information Science 37, 2, 2010, 137–148. Pothecary 2005 S. Pothecary, The European provinces. Strabo as evidence, in: D. Dueck – H. Lindsay – S. Pothecary (eds.), Strabo’s Cultural Geography. The Making of a Kolossourgia (Cambridge 2005) 161–179.

34

Lara Fabian

Price et al. 2013 M. F. Price – A. C. Byers – D. A. Friend – T. Kohler – L. W. Price, Mountain Geography. Physical and Human Dimensions (Berkeley 2013). Rubinson – Smith 2003 K. S. Rubinson – A. T. Smith, Archaeology in the Borderlands, in: A. T. Smith – K. Rubinson (eds.), Archaeology in the Borderlands. Investigations in Caucasia and Beyond (Los Angeles, CA 2003) 1–8. Sagramoso 2007 D. Sagramoso, Violence and conflict in the Russian North Caucasus, International Affairs 83, 4, 2007, 681–705. Scott 2009 J. C. Scott, The Art of Not Being Governed. An Anarchist History of Upland Southeast Asia (New Haven, Conn. 2009). Shaw 1990 B. D. Shaw, Bandit highlands and lowland peace. The mountains of Isauria-Cilicia, Journal of the Economic and Social History of the Orient 33, 2, 1990, 199. A. Smith 2005 A. T. Smith, Prometheus unbound. Southern Caucasia in prehistory, Journal of World Prehistory 19, 4, 2005, 229–279. M. Smith 2005 M. L. Smith, Networks, territories, and the cartography of ancient states, Annals of the Association of American Geographers 95, 4, 2005, 832–849. Sorge 2014 A. Sorge, Highland sanctuary and the state. Mountains as a political category in Mediterranean history, in: A. C. Dawson – L. Zanotti – I. Vaccaro (eds.), Negotiating Territoriality. Spatial Dialogues Between State and Tradition (New York 2014) 21–35. Stepp et al. 2005 J. R. Stepp – H. Castaneda – S. Cervone, Mountains and biocultural diversity, Mountain Research and Development 25, 3, 2005, 223–227. Strabo Strabo, Strabons Geographika, S. Radt (ed.) (Göttingen 2002–2011). Surface-Evans – White 2012 S. L. Surface-Evans – D. A. White, An introduction to the least cost analysis of social landscapes, in: D. A. White – S. L. Surface-Evans (eds.), Least Cost Analysis and Social Landscapes. Archaeological Case Studies (Salt Lake City 2012) 1–10. Tilley 1994 C. Y. Tilley, A Phenomenology of Landscape. Places, Paths and Monuments (Oxford 1994). Tobler 1993 W. Tobler, Three Presentations on Geographical Analysis and Modeling (Santa Barbara 1993). UNEP WCMC 2002 UNEP World Conservation Monitoring Centre (ed.), Mountain Watch. Environmental Change and Sustainable Development in Mountains (Cambridge 2002). Van Buren 1996 M. Van Buren, Rethinking the vertical archipelago. Ethnicity, exchange, and history in the South Central Andes, American Anthropologist 98, 2, 1996, 338–351. Verhagen 2013 P. Verhagen, On the road to nowhere? Least cost paths, accessibility and the predictive modelling perspective, in: F. Contreras Cortes – M. Farjas – F. J. Melero (eds.), Computer Applications and Quantitative Methods in Archaeology. Proceedings of the 38th Annual Conference on Computer Applications and Quantitative Methods in Archaeology (CAA) (Oxford 2013) 383–390.

Moving in the Mountains

35

White – Barber 2012 D. A. White – S. B. Barber, Geospatial modeling of pedestrian transportation networks. A case study from precolumbian Oaxaca, Mexico, Journal of Archaeological Science 39, 8, 2012, 2684–2696. Whitley – Burns 2008 T. G. Whitley – G. Burns, Conditional GIS surfaces and their potential for archaeological predictive modelling, in: A. Posluschny – K. Lambers – I. Herzog (eds.), Layers of Perception. Proceedings of the 35th International Conference on Computer Applications and Quantitative Methods in Archaeology (CAA) (Bonn 2008) 292–298. Zimansky 1985 P. E. Zimansky, Ecology and Empire. The Structure of the Urartian State, Studies in Ancient Oriental Civilization 41 (Chicago 1985). Zimansky 1995 P. E. Zimansky, Urartian material culture as state assemblage. An anomaly in the archaeology of empire, Bulletin of the American Schools of Oriental Research 299/300, 1995, 103–115.

Patterns of Settlement and Metallurgy in Late Bronze–Early Iron Age Kvemo Kartli, Southern Georgia Nathaniel L. Erb-Satullo * Abstract: Despite the large numbers of ancient mining and metal production sites in the South Caucasus, spatial analyses of metal production landscapes are far from common. Archaeological survey in the ore-rich foothills of Kvemo Kartli, southern Georgia, demonstrates the potential for integrating previous geological and archaeological research into present-day landscape archaeology projects. Combining reports of earlier excavations, satellite imagery and ground-level observations, 15 hilltop sites and 9 sites with possible mining activity were mapped. Systematic surface collection revealed that most of these hilltop sites contain at least some traces of Late Bronze–Early Iron Age (LBA– EIA, c. 1500–600 BC) habitation, with a number of sites having a predominantly LBA–EIA ceramic assemblage. The relative positioning of hilltop settlement and mining sites shows that settlement sites were frequently positioned in ways that controlled access to ore deposits. Metallurgical activity, evidenced by the presence of slag, was attested at several of these hilltop sites. In one or two cases, settlements were built directly adjacent to ore deposits. However, it remains an open question whether proximity to ore deposits and evidence of metal production at these sites meant that metal production was administered or managed by elites within those communities. Detailed investigation of the spatial contexts of metal production at all spatial scales is necessary for understanding the factors driving social and technological change during this period. Keywords: Craft production; Landscape archaeology; Survey; Metallurgy; Mining; Bronze Age; Iron Age; Caucasus

Introduction The South Caucasus region was a major centre of metal production since at least the 6th millennium BC.1 The region is exceptionally rich in evidence of metal production, from mining and smelting sites to casting and forging debris found at settlement sites and, in rare instances, in graves.2 To date, archaeometallurgical research in the region has generally focused on reconstructing technological chronologies, dating the origin and spread of techniques for making arsenical copper, tin bronze, iron, gold, etc.3 Less research has focused on examining how the metal economy interfaced with other aspects of social life. Although earlier projects in Georgia identified numerous mining and metal production sites4, landscape perspectives and spatial interpretations were a secondary focus, if they were considered at all. Recent research has begun to correct this imbalance, mapping landscapes of Late Bronze and Early Iron Age smelting sites in the densely vegetated subtropical environment of western Georgia (ancient Colchis).5 This work, which relied on local informants and field notes from previous Soviet-period investigations6, revealed an extensive network of copper smelting sites dating to the Late Bronze and Early Iron Age. However, the modern-day landscapes

School of Archaeology, University of Oxford; Department of Anthropology, Harvard University, [email protected]. 1 Courcier 2014. 2 E.g. Avalishvili 1974, 132–135; Chavleishvili 1991; Gambashidze et al. 2001; Mudzhiri 2011; Erb-Satullo et al. 2014; Papuashvili – Jibladze 2014; Erb-Satullo et al. 2015. 3 Kavtaradze 1999; Courcier et al. 2008; Stöllner – Gambashidze 2014; Courcier et al. 2016. 4 E.g. Gzelishvili 1964; Khakhutaishvili 1987; Chartolani 1989, 158–166. 5 Erb-Satullo et al. 2014; Erb-Satullo et al. 2015; Erb-Satullo et al. 2017; Gilmour et al. submitted. 6 Khakhutaishvili 1987. *

38

Nathaniel L. Erb-Satullo

Fig. 1   Map of the Mashavera and Debeda corridors, showing the approximate limits of the survey area and known ore deposits. Known ore deposits on the Georgian side of the border are based on maps produced by Nazarov (1966)

of western Georgia are ill-suited to typical methods of systematic archaeological survey, such as field-walking and remote sensing. As a result, while the surveys uncovered dozens of smelting sites, it was difficult to investigate metallurgical and settlement landscapes simultaneously. In order to explore spatial dimensions of metal production and its position within the broader society, I initiated a programme of survey in a metal-rich region of Kvemo Kartli, southern Georgia. The focus of the project was investigating the social context of copper and iron production in the Late Bronze and Early Iron Age. Numerous models of the rise of iron production cite differences in the relative organisation of the bronze and iron industries as key factors in explaining its spread.7 Yet explorations of spatial patterns of Late Bronze and Early Iron Age metal production are limited outside of the Levant and the Eastern Mediterranean. Kvemo Kartli is an attractive place to explore these issues for a number of reasons. Previous research had shown that the area has remains of both ancient settlements and metallurgical activity. In the late 1950s and early 1960s, Gzelishvili identified a number of metal production sites from varying periods, most of which he identified as the remains of iron production. One of these, Kvemo Bolnisi, he dated to the Late Bronze and Early Iron Age.8 More recent work by a GermanGeorgian team at the site of Sakdrisi has identified the earliest known gold mine in the world, dating to the late-4th to early-3rd millennium BC.9 Tantalising evidence of even earlier metallurgy comes

7 8 9

Mirau 1997; Veldhuijzen 2012; Yahalom-Mack – Eliyahu-Behar 2015. Gzelishvili 1964, 36–37. Stöllner et al. 2008; Stöllner – Gambashidze 2011; Stöllner – Gambashidze 2014.

Patterns of Settlement and Metallurgy in Late Bronze–Early Iron Age Kvemo Kartli, Southern Georgia

39

from nearby Arukhlo, where copper ores and a copper bead, likely dating to the 6th millennium BC, were discovered.10 Moreover, unlike the rain-drenched shores of the Black Sea, the lowland parts of Kvemo Kartli are more sparsely vegetated, making it possible to use a wider array of survey techniques, including remote sensing and systematic surface collection. The area targeted for survey was the lowland foothill zone of the Mashavera and Debeda rivers, close to the major Mashavera and Alaverdi ore deposits (Fig. 1).11 Topographically, the survey area sits astride two major routes of travel between the Kura River basin and the Lesser Caucasus highlands. The goals of the 2013 and 2014 survey seasons were to explore the relationship between the settlement landscape and metallurgical activities and test field methodologies for surveying hilltop sites with varying surface conditions. Of particular interest was the question of whether Late Bronze Age and Early Iron Age (c. 1500–600 BC)12 sites were positioned in ways to control or monitor access to ore deposits. The topographic disposition of ore deposits and trans-regional routes offers the possibility of probing these questions. A practice of settlement construction oriented solely towards controlling movement between the highlands and the lowlands would manifest itself in the building of hilltop sites only along major avenues of movement. On the other hand, if access to (and perhaps control of) ore deposits were an important concern, we should expect to see hilltop sites positioned in ways that prioritise access to ore deposits, even when they lie away from major routes of travel. Methods of Survey The record of previous archaeological work meant that a range of different methodologies were available for locating and mapping Bronze and Iron Age archaeological sites. Apart from Kvemo Bolnisi, few Late Bronze and Early Iron Age settlements have been investigated in the vicinity of the survey area, but several LBA–EIA burials were excavated in the Debeda River valley south of the modern town of Marneuli.13 Published maps from the Soviet period are generally handdrawn sketch maps giving only the approximate positions of sites relative to major watercourses and modern towns. By comparing sketch maps with modern maps and Google Earth imagery, however, it was possible to find the probable locations of a number of previously identified sites, which could then be investigated on the ground. Although not all sites discussed by Gzelishvili were able to be rediscovered, a significant number were. Identification of new sites proceeded recursively. Survey began by visiting a list of possible sites from previous research and from close examination of Google Earth imagery. While some of the possible sites identified early in the survey lacked evidence of ancient occupation, those that did were then re-examined in the satellite imagery, and diagnostic features were noted. Re-examination of satellite imagery, coupled with observations of hills from the valley bottom, improved the accuracy of site identification.

10 11 12

13

Hansen 2012. Nazarov 1966; Mkrtchyan et al. 1968. The chronological definition of the term ‘Late Bronze Age to Early Iron Age’ requires some explanation as it applies in Georgia. Pottery characterising these periods, including black and grey burnished wares with a range of incised and impressed designs, is roughly consistent over the period. While scholars speak of a fully developed Iron Age beginning sometime in the early 1st millennium BC, these distinctions are typically made on the quantity and character of metal artifacts, not so much on the pottery (M. Puturidze, personal communication). In neighbouring Armenia, the presence of Urartian artifacts is a prime distinguishing characteristic between the Iron I and Iron II periods (Smith et al. 2009, 91), but evidence of Urartian contact, much less conquest, is far weaker in Georgia. For the purposes of this chapter, I use the term Late Bronze – Early Iron Age to refer generally to the period from about 1500–600 BC, while recognising that some sites with these types of ceramics may have been contemporary with fairly widespread use of iron (i.e. technically not so ‘early’ in the Iron Age). Avalishvili 1975.

40

Nathaniel L. Erb-Satullo

Fig. 2   Schematic diagram showing field-walking patterns for the two different types of collection units

Exploration of uninhabited hilltops (false positives) in the early stages of survey produced an improved understanding of how both inhabited and uninhabited hilltops appeared from the ground and the air. Even later in the survey, some hills suspected to be lacking in settlement were nonetheless explored to limit the impact of confirmation bias. The result of these efforts was that by the end of the survey, it was possible to identify where hilltop sites were located and, just as important, where they were not, with reasonable accuracy. In the satellite images and when viewed from the valley floor, the most diagnostic features of hilltop sites were the encircling walls and terraces. In most cases, individual linear stone features were not resolvable, but the change in slope resulting from a buried terrace or fortification wall was often clearly visible. Combined with valley-bottom observations of a hill’s profile, these features contributed to the high degree of accuracy in identifying settled hills even before visiting the site. This survey methodology admittedly privileges sites placed on prominent hills rather than in the valley bottom. Large-scale walking of agricultural fields is certainly feasible, particularly in springtime, but this approach was incompatible with the survey goals of the 2013 and 2014 seasons. Once a settled hilltop had been located, systematic surface collection was used to obtain a representative sample of ceramics, search for metal production debris and to determine the settlement’s approximate spatial extent. Two different methods of systematic collection were employed. Where it was deemed advantageous to correlate collection units with surface features, the method of fieldwalking involved mapping out an irregularly-shaped collection area.14 These irregular ‘elective’ survey areas were walked in 5m intervals along a particular orientation (Fig. 2a). This method had the advantage of being able to adapt the collection units to the topographic, erosional and vegetation profile of the site. This worked well for sites where the exposure and visibility of surface materials was highly variable, as in cases of modern digging or when sherd densities were relatively low. This method was less effective in surveying sites with uniformly high densities of ceramics, as it resulted in large numbers of sherds that did not provide additional value for the research goals. This method was employed for surface collections at Sites 2, 9, 13, 20 and 22. Another method involved laying out 10 ×10m grids at a regularly defined interval (in most cases, 40m) and then collecting ceramics within that area. Though slower than the elective collection units, this approach worked very well at sites with high densities of sherds and relatively consistent surface visibility, yielding a representative yet manageable sample of materials. The collection method involved collecting all ceramics and lithics from a 5 × 5m sub-area and collecting all diagnostic sherds (rims, bases and handles, as well as all sherds with decoration) from the full 10 × 10m square. A 1m transect spacing was used for the 5 × 5m grid, while a 2.5m spacing

14

See similar methods in Ur et al. 2013.

Patterns of Settlement and Metallurgy in Late Bronze–Early Iron Age Kvemo Kartli, Southern Georgia

41

Fig. 3   Map of LBA–EIA hilltop sites and mining activity in the survey area. Hilltop sites with finds of metallurgical slag are marked in red. Not all identifications of mining activity are equally certain, nor are all mining sites assumed to date to the LBA–EIA. See text for an evaluation of the evidence for mining at each of these sites, and a discussion of which sites provide the most convincing spatial associations between LBA–EIA pottery and mining/ metallurgical debris

was used for the 10m square (Fig. 2b). Collections of this type were employed at Sites 4, 16, 23, 25, 30, 33 (Mtsvane Gora) and 34. The use of a flexible strategy of surface collection has some limitations. Most notably, the different spacing of transects in the different collection methods means that comparisons of sherd densities between sites using these two methods are not possible. The closer transect spacing in the 10m collection squares means that coverage is better and measured densities would be correspondingly higher relative to the other method. However, other factors besides collection methods limit the reliability of using relative sherd densities between sites as a measure of relative intensity or duration of occupation. Most of the inter-site differences in sherd densities are likely the result of variable taphonomic processes at different sites, and so intersite comparisons are of limited value. The two surface collection methods were never used on the same site, so calculations of sherd densities are comparable with a single site. Results and Discussion Systematic survey of hilltop sites revealed a wide distribution of sites with traces of Late Bronze and Early Iron Age settlement (Fig. 3). At a number of sites, (e.g. Sites 2, 4, 25, 33, 34 and 36) Late Bronze Age and Early Iron Age pottery predominated. At others (Sites 13, 16, 20 and 22), LBA–EIA pottery was identified, but in smaller amounts relative to later Classical and/or medieval pottery.

42

Nathaniel L. Erb-Satullo

Fig. 4   Late Bronze and Early Iron Age sherds from hilltop sites. 1–3: Site 2; 4: Site 4; 5: Site 13; 6–9: Site 20; 10–12: Site 25; 13–15: Site 33; 16–17: Site 34

Late Bronze–Early Iron Age pottery was characterised by black and grey burnished sherds, often with incised lines or pattern-burnished decoration (Fig. 4). Stamped triangle decoration (Fig. 4.13), ‘zoomorphic’ or horned handles, and hole-mouth vessels (Fig. 4.4, 4.15) were more rare, yet they are particularly diagnostic of the period. While LBA–EIA ceramics are predominantly black and grey, coloration of ceramics was variable. Some lighter brown sherds were identified whose decoration strongly points to an LBA–EIA date (see Fig. 4.13). At sites with many diagnostic LBA–EIA sherds and few or no ceramics diagnostic of other periods, the large quantities of non-diagnostic grey, black and even light brown body sherds almost certainly date to the LBA–EIA. A number of LBA–EIA sites had walls surrounding tops of the hill (Fig. 5). Such walls followed the natural topography of the hill, resulting in long, relatively narrow enclosures at Site 34 and 13. In some cases, such as Sites 35 and 36, multiple encircling terrace walls were identified. The distribution of residential structures and craft production debris in relation to these fortifications is a major focus of ongoing investigations, as the placement of workshops and the question of ‘lower towns’ has implications for the organisation of craft production and the social structure

Patterns of Settlement and Metallurgy in Late Bronze–Early Iron Age Kvemo Kartli, Southern Georgia

Fig. 5   Walls surrounding hilltop forts. Some walls are visible only by a distinct change in slope, while others have well-preserved stone architecture. Scale bar in lower image is 30cm

43

44

Nathaniel L. Erb-Satullo

Fig. 6   Surface collection at Site 33 (Mtsvane Gora), showing distribution of ceramics and slag. The encircling wall is clearly visible, particularly on its northern side

of Late Bronze Age–Early Iron Age fortress-settlements.15 Surface collections in 2013 and 2014 discovered large quantities of ceramics on the slopes below walled hilltops. While in some cases, this might derive from refuse tossed down the side of the hill, in others, there is clearer evidence for structures on lower slopes. At Site 25, for instance, linear stone features were observed among stone collapse on the lower slopes of the hill, likely indicating the presence of stone-walled terrace houses. Ceramics from a collection unit, which covered some of those features, included a number of LBA–EIA sherds, though a couple of sherds that might post-date this period were also found. A number of sites had traces of metal production. Careful examination of Soviet period sketch maps led to the rediscovery of Gzelishvili’s Kvemo Bolnisi site (Site 2 in the new survey), which was characterised by abundant Late Bronze and Early Iron Age ceramics. Gzelishvili’s 1958– 1959 investigations describe a complex of ‘antique’ (i.e. Classical) and ‘pre-antique’ settlement and mortuary complexes as well as a significant number of metallurgical remains.16 He describes a metallurgical workshop on the southern edge of the settlement as relating to iron production, despite noting the presence of an ancient copper mine at the site. The specific dating of the settlement and workshop is based on ceramic comparisons with other sites in eastern Georgia, with the settlement belonging roughly to the 13th–7th centuries BC and the workshop dating to the earlier part of that range. According to Gzelishvili, Classical remains were limited to graves and some structures on the very top of the hill.17

15 16 17

Lindsay et al. 2010; Hammer 2014. Gzelishvili 1964, 31–38. Gzelishvili 1964, 32.

Patterns of Settlement and Metallurgy in Late Bronze–Early Iron Age Kvemo Kartli, Southern Georgia

45

When the site was revisited in 2013, it was clear that the site had experienced significant damage in recent years from a series of mining prospection trenches, which exposed ceramics, charcoal and architecture in section. Although not all features (walls, excavation trenches, grave locations, etc.) on Gzelishvili’s map of the site were found, it was possible to identify the general area of the metal workshop. Although no trace of the workshop structure remained, several pieces of slag were found near a modern mining prospection trench in the general area of the workshop. Contrary to Gzelishvili’s identification of workshop as an iron production workshop, greenish corrosion was visible in some places on the slags. Qualitative pXRF analysis of several slag samples revealed the presence of copper, producing spectra very similar to known copper slags from western Georgia.18 In addition, while ancient mine workings were not apparent, the site was covered in bright green copper ores, and numerous stone hammers, likely used for mining, were also identified. The greenish corrosion on the slags, coupled with the abundance of copper ore throughout the site, suggests that it was copper, rather than iron, that was produced at the workshop. Elsewhere in the Caucasus, other reinvestigations of alleged iron smelting sites have produced similar results.19 Another site containing overwhelmingly Late Bronze–Early Iron Age pottery and traces of metal production was Site 33 (Mtsvane Gora). Situated on an isolated hill set off from the main ridge, it is ideally positioned to control movement along the Debeda River, which offers a direct route into the highlands through the Alaverdi ore-bearing zone. The site was identified by its encircling wall, enclosing an area of about 0.6ha on the top of the hill. The wall is no longer preserved above ground level, but it is clearly visible as a change in slope, and stones are visible at the surface in some areas (Fig. 6). Surface collections at the site reveal a distribution of ceramics concentrated on the southern slopes of the hill. Ceramics from the site included a number of sherds particularly diagnostic of the LBA–EIA, including a hole-mouth vessel with incised rim, and body sherd impressed-triangle decoration (Fig. 4.13, 4.15). Surface collections in 2014 also recovered 14 fragments of metallurgical debris, mostly pieces of slag with one slagged crucible fragment and one partially melted/slagged stone (possibly part of a furnace or hearth construction). With the exception of the crucible rim, which had a small amount of greenish copper corrosion adhering to it, none of the recovered fragments had any macroscopic indications of copper production. Initial scientific analyses strongly suggest that iron metallurgy was practised at the site, and more detailed chemical and microscopic investigation of this debris is currently underway. The distribution of the slags on the surface is intriguing. Most slags were found on the southern slope of the hill, a result that may derive from the more eroded quality of the less vegetated southern slope. However, there is a clear association between a cluster of slags and the wall encircling the top of the hill. It is possible that a workshop was located just uphill from that cluster, and slags were tossed or eroded slowly downhill. Does the association of the wall and the metallurgical debris indicate that metal production was controlled or protected by those who built and maintained the fortifications? Somewhat analogous evidence of metal production at fortress complexes was found at the site of Gegharot in Armenia, where casting moulds, crucibles and possible pyrotechnological installations were discovered.20 Hilltops are not ideal locations for carrying out fuel-hungry pyrotechnological activities, since all wood, ore and other materials must be carried uphill.21 Still, it would be hasty to conclude unequivocally that the positioning of a workshop within or adjacent to a hilltop enclosure necessitates tight administrative control over production, not the least because the social organisation of these hilltop settlements remains an open question. Further investigation is necessary to tease out the social meaning of these spatial relationships.

18 19 20 21

See Erb-Satullo et al. 2014; Erb-Satullo et al. 2015; Erb-Satullo et al. 2017. Erb-Satullo et al. 2014. Badalyan et al. 2008, 64–65. A possible exception would be wind-powered furnaces (Killick 1991; Rehren 2011, 77–79), but there is no evidence for these kinds of furnaces in the Caucasus. Moreover, while it is slightly breezier at the top of the hill than at the bottom, the difference is not nearly enough to make wind-powered furnaces worthwhile.

46

Nathaniel L. Erb-Satullo

Fig. 7   Traces of ancient mining activity. Scale bars in the top row of photos are 1m, other scale bars are 10cm

Tenuous evidence of metal ore exploitation in association with LBA–EIA remains was identified at Site 9. A recent modern cut exposed a vein of ore containing orange, green and sparkly dark minerals, which are most likely a mixture of iron and copper ores, though samples of these minerals were not analysed (see Fig. 7 middle-right). On the same slope, surface collection recovered a number of black-burnished sherds and an additional chunk of the dark sparkly mineral similar to

Patterns of Settlement and Metallurgy in Late Bronze–Early Iron Age Kvemo Kartli, Southern Georgia

47

that found in the modern cut. Such evidence suggests that LBA–EIA metallurgical activity may have taken place at this site, but it is not as conclusive as cases where slag, mining tools and/or ancient mine workings are present. Other hilltop sites in the survey area contained some traces of metal production, but the association of that debris with LBA–EIA pottery was less clear. Systematic surface collection of materials from Site 23, a hilltop site with stone walls, yielded two small pieces of slag. Given the hilltop location, these fragments likely indicate that metal production occurred somewhere on the hilltop, as natural processes would not move slag uphill. However, the ceramic assemblage at the site indicates that in addition to LBA–EIA occupation, there was also occupation in later periods. A number of higher-fired red-orange sherds, some with reddish paint or slip, are perhaps the best indicator of this later occupation. Likewise, several pieces of slag were recovered from systematic survey at Site 22 (Kveshi Castle), a medieval fortress prominently positioned on a hill in the middle of the Mashavera Valley. While the majority of pottery recovered was medieval in date, a number of sherds dated to the Late Bronze–Early Iron Age were also identified. In cases such as Sites 22 and 23, it is difficult to argue for specific dates for production debris based on surface collections alone. Besides the discoveries of metal production in association with hilltop settlements, the survey did locate a number of mines away from settlements (Fig. 7). Gzelishvili reports the discovery of mines (allegedly of iron) in the area south of the modern town of Bolnisi.22 Plans of three mines were published, and dates between the 6th to 13th centuries AD were suggested, at least in part based on evidence of iron tools used in digging the mines. Using geological maps of ore deposits23, a number of mining adits (horizontal mine entrances) were discovered. Due to the lack of proper equipment, most of these adits were not explored to great depth, so it was difficult to correlate mine plans published by Gzelishvili and the mines found by the present survey. Unfortunately, none of the mining sites were included on Gzelishvili’s maps of the area. Several of the mines identified, however, were located near place names mentioned by Gzelishvili, so it is likely that at least some of the mines mapped by the current project correlate with earlier finds. On the surface, most mines were identified by their spoil heaps, which were often visible on the hill slope at a distance. Sites 7 and 8 all had spoil heaps consisting of sparkly iron ores (probably hematite), making them particularly distinctive (see images of Site 8 in Fig. 7).24 The four adjacent adits found at Site 14 may be more recent, given their state of erosion, but that is mostly conjecture.25 At least one mine lacking a directly associated settlement did appear to predate the use of iron mining tools. Site 31 is located between Sites 4, 20 and 25 and consists of a series of rocky outcrops with red-orange iron and green copper staining. A tell-tale stone hammer, made of rock inconsistent with the surrounding outcrop, was identified, strongly suggesting early working of these deposits (Fig. 7 bottom). However, several nearby mining adits (Sites 32 and 37) are probably more recent, based on their similarities to Sites 7–8 and 14–15. Of all ancient mining sites, the most tentatively identified is Site 24. Several depressions in a rocky outcrop were identified as possibly anthropogenic in origin, perhaps the sedimentfilled remains of ancient mine workings. On the surrounding hill slope, several rounded stone hammers and grindstones – easily differentiated from the surrounding bedrock – were found. Chipped obsidian was identified in a number of instances, but ceramic sherds were very rare

22 23 24

25

Gzelishvili 1964, 14–20. Nazarov 1966. Note that the presence of iron ores does not necessarily mean that the iron was the desired ore mineral. Polymetallic ore deposits are extremely common, and the desired minerals are usually removed from the site, not abandoned in the spoil heap. Gzelishvili describes the discovery of five mines (шахты) in the same general area, assigning them to the late medieval period (Gzelishvili 1964, 20). It is possible he was referring to the total combined number of adits at Sites 14 and 15, or it is possible that a fifth adit at Site 14 was not found by the current survey.

48

Nathaniel L. Erb-Satullo

and generally undiagnostic. Stone hammers may have been used for a range of purposes, and the presence of stone hammers at a settlement site is not in itself conclusive evidence of mining activity.26 However, stone hammers are frequently found at mining sites, and the lack of ceramics suggests either that there was little settlement at the site or that the main concentration of settlement was not located. Further examination of the site, which was visited only once in 2013, is necessary to elucidate its character further. The topographic disposition of ore deposits and settlement sites offers the possibility to test the hypothesis that settlements were positioned at least in part to control, monitor and/or facilitate access to ore deposits. As discussed above, a number of settlement sites have clear traces of metal production, while two (Sites 2 and 9) were directly adjacent to an ore deposit. While mines are difficult to date as a rule, the presence of LBA–EIA settlement directly adjacent to these ore outcrops is likely significant. Other hilltop sites were positioned in side valleys containing possible ore deposits. Site 23, for instance, is tucked up a small side valley of the Mashavera Gorge, virtually out of sight of the main route between the lowlands and the highlands. While positioned off the main route of travel along the Mashavera, Site 23 is nonetheless quite close to one of the major ore deposits of the Mashavera Valley, the Madneuli deposit, where a massive modern pit mine is located. Site 9 is positioned at the entrance to a valley with known iron deposits, and the finds of ore outcrop on the site itself suggest there may be additional unmapped small deposits of copper or iron. While no direct evidence of metal production was found at Sites 4, 13 and 16, this trio of sites commands access to the Poladauri Valley, which abuts the backside of the major Madneuli deposits mentioned above. Site 4 has a direct line of sight to the probable traces of ancient mining activity at Site 31, at a distance of just over 2km. Yet, while the Mashavera Gorge leads directly up into the highland plateaus, the access through the Poladauri Valley is blocked by the Somkheti range, making it a poor route of travel between the lowlands and the highlands (Fig. 1). The placement of hilltop sites, therefore, likely derives at least in part from a desire to exploit and perhaps control access to ore deposits. While a number of hilltop sites control major routes of interregional travel through the Mashavera and Debeda gorges, a substantial proportion do not. Were control of movement between the lowlands and the highlands the paramount concern in the placement of sites, one would expect a network of hilltop sites positioned only along main routes of travel. As this is clearly not the case, alternative explanation is required. The placement of settlement sites at or near ore deposits, coupled with the traces of pyrotechnological activity, suggests metallurgical factors influenced their placement. In this sense, Site 2 (Kvemo Bolnisi) is probably the best example of a site that is ill-positioned to control highland–lowland movement, yet well-positioned to protect access to an ore deposit. Indeed, the position of the site, virtually invisible from Mashavera Valley to the north, might have shielded it from unwanted attention. Still, metallurgical considerations were probably not the only factor governing the placement of most sites. The profusion of Late Bronze Age weaponry, the growth of massive fortresses and bioarchaeological evidence suggest that interpersonal violence and raiding was a defining feature of societies.27 Placement of fortified settlements at the mouths of dead-end valleys would have protected agricultural fields and unfortified settlements located within them. Settlements tucked far up in these valleys might have escaped the notice of potential aggressors passing through the region, and hilltop sites at their entrance would have served as lookouts.

26

27

For example, several stone hammers were identified at Site 25, but systematic survey did not uncover any evidence of smelting or smithing. While the stone hammers at Site 25 may have been used to mine copper at nearby Site 31, it is also possible that they may have served another purpose. For a catalogue of LBA weaponry, see Picchelauri 1997; for bioarchaeological evidence of trauma, see Marshall 2014, 268–271.

Patterns of Settlement and Metallurgy in Late Bronze–Early Iron Age Kvemo Kartli, Southern Georgia

49

Conclusions and Future Directions Survey of Late Bronze and Early Iron Age landscapes in Kvemo Kartli illustrates the value of incorporating previous research into new projects and using satellite imagery to locate new sites of interest. By carefully following up on previous research, we can better integrate new discoveries and legacy data to produce a more robust understanding of the past. For instance, correlation of earlier excavations with new observations at the site led to the identification of the metal produced at the Kvemo Bolnisi workshop as copper (or copper-alloys) rather than iron, as previously thought. Maps produced by previous geological surveys proved effective in locating areas of interest for discovering ancient mines, though even comparatively detailed maps contain omissions that may prove important to the study of early mining and metallurgy. Because modern mining companies focus on large deposits that are economically feasible to exploit, they may ignore small, rich, near surface deposits favoured by ancient miners, as may be the case at Site 9. The evidence that fortified hilltop sites were positioned, at least in part, to exploit metal resources suggests that metal production was a significant economic driver in the region by the Late Bronze Age. It is particularly interesting to note that in a couple of instances (Sites 2 and possibly 9) LBA–EIA settlement was situated at the ore deposit, suggesting the importance of spatial proximity in making a claim to the ore deposit. Such direct proximity means that such communities may have exercised exclusive rights to work such deposits, as it would have been impossible for external groups to work the mine without force or, at the very least, the acquiescence of the local community. However, the available evidence is silent on the question of whether metal production was controlled by certain groups within these communities, let alone the question of whether ‘control’ of deposits operated at the level of the individual settlement, the valley, or the region. Indeed, there is considerable latitude for interpretation concerning the social organisation of production within communities. One could imagine a model where elites exercised direct control over production, aided by the direct proximity of virtually the whole chaîne opératoire of production. On the other hand, one could equally imagine a model of community specialisation28, where multiple metalworking clans formed a more balanced power structure oriented towards protection of a shared resource. Resolving these issues requires teasing out the true nature of these fortified sites as social institutions. Were they residences of elites and their retainers who exercised direct control over nearby production from their hilltop fastnesses? Or are these fortifications, especially the more modestly sized ones, less extensions of elite authority and more communal constructions for mutual defence? The political significance of fortress complexes has been a major focus of research, particularly in the highland regions to the south of the survey area.29 Further research on fortified settlements in the hilly margins of the Kura Plain will provide a complementary lowland case study for these rich highland datasets, providing a more nuanced picture of the social contexts of Late Bronze and Early Iron Age metal production. Acknowledgements: Giorgi Bedianashvili and Kakha Kakhiani provided valuable assistance in organising fieldwork logistics. Bastien Varoutsikos, Kathryn O’Niel Weber, and Jimsher Chkhvimiani participated in the survey. Special thanks go to the Dmanisi excavations administration (director: David Lordkipanidze), which provided accommodation for the initial 2013 season. This research was supported by an ARISC Graduate Research Fellowship, a National Science Foundation Graduate Research Fellowship (DGE0644491 and DGE1144152), and the Harvard Department of Anthropology. Digital elevation models in figures 1 and 3 were provided by ASTER (a product of METI and NASA) and GTOPO30.

28 29

See Stark 1991. Lindsay – Greene 2013; Smith – Leon 2014; Smith 2015.

50

Nathaniel L. Erb-Satullo

References Avalishvili 1974 გ. ავალიშვილი / G. Avalishvili, ქვემო ქართლი ძვ. წ. I ათასწლეულშის პირველ ნახევარში / Kvemo Kartli dzv. ts. I atastsleulis pirvel nakhevarshi ( Kvemo Kartli in the First Half of the 1st Millennium BC) (Tbilisi 1974). Avalishvili 1975 გ. ავალიშვილი / G. Avalishvili, გვიანბრინჯაო-ადრერკინის ხანის სამარხები / Gvianbrinjao–adrerkinis khanis samarkhebi (Late Bronze– Early Iron Age burials), in: Javakhishvili – Japaridze 1975, 166–172. Badalyan et al. 2008 R. S. Badalyan – A. T. Smith – L. Khatchadourian – P. S. Avetisyan, Village, fortress and town in Bronze and Iron Age southern Caucasia. A preliminary report on the 2003–2006 investigations of Project ArAGATS on the Tsaghkahovit Plain, Republic of Armenia, Archäologische Mitteilungen aus Iran und Turan 40, 2008, 45–105. Chartolani 1989 Ш. Чартолани, К Истории Нагорья Западной Грузии в Доклассовой Эпохи (Tbilisi 1989). Chavleishvili 1991 ი. ჩავლეიშვილი / I. Chavleishvili, კოლხური ბრინჯაოს საწარმოო კერა ჩოლოქ-. ოჩხამურის ხერთვისში / Kolkhuri brinjaos satsarmoo kera cholok. Ochkhamuris khertvisshi (A Colchian bronzecasting production hearth from the Choloki-Ochkhamuri interfluve), სამხრეთ-დასავლეთ საქართველოს ძეგლები / Samkhret-Dasavlet Sakartvelos Dzeblebi (Monuments of Southwestern Georgia) XIX, 1991, 3–26. Courcier 2014 A. Courcier, Ancient metallurgy in the Caucasus from the sixth to the third millennium BCE, in: B. W. Roberts – C. P. Thornton (eds.), Archaeometallurgy in Global Perspective (New York 2014) 579–664. Courcier et al. 2008 A. Courcier – D. Kuparadze – D. Pataridze, Archaeometallurgical researches on the early beginnings of metallurgy (VIth–IIIrd millennia BC) in the Caucasus. An example of interdisciplinary studies, Metalla 15, 2008, 35–50. Courcier et al. 2016 A. Courcier – B. Jalilov – I. Aliyev – F. Guliyev – M. Jansen – B. Lyonnet – N. Mukhtarov – N. Museibli, The ancient metallurgy in Azerbaijan from the end of the Neolithic to the Early Bronze Age (6th to 3rd millennium BCE). An overview in the light of new discoveries and recent archaeometallurgical research, in: G. Körlin – M. Prange – T. Stöllner – Ü. Yalçin (eds.), From Bright Ores to Shiny Metals. Festschrift for Andreas Hauptman on the Occasion of 40 years Research in Archaeometallurgy and Archaeometry, Der Anschnitt Beiheft 29 (Bochum 2016) 26–36. Erb-Satullo et al. 2014 N. L. Erb-Satullo – B. J. J. Gilmour – N. Khakhutaishvili, Late Bronze Age and Early Iron Age copper smelting technologies in the South Caucasus. The view from ancient Colchis c. 1500–600 BC, Journal of Archaeological Science 49, 2014, 147–159. Erb-Satullo et al. 2015 N. L. Erb-Satullo – B. J. J. Gilmour – N. Khakhutaishvili, Crucible technologies in the Late Bronze– Early Iron Age South Caucasus. Copper processing, tin bronze production, and the possibility of local tin ores, Journal of Archaeological Science 61, 2015, 260–276. Erb-Satullo et al. 2017 N. L. Erb-Satullo – B. J. J. Gilmour – N. Khakhutaishvili, Copper production landscapes of the South Caucasus, Journal of Anthropological Archaeology 47, 2017, 109–126. Gambashidze et al. 2001 I. Gambashidze – A. Hauptmann – R. Slotta – Ü. Yalçin (eds.), Georgien. Schätze aus dem Land des goldenen Vlies. Exhibition catalogue Deutsches Bergbau-Museum Bochum (Bochum 2001). Gilmour et al. submitted B. J. J. Gilmour – M. Cox – N. Erb-Satullo – N. Khakhutaishvili – A. M. Pollard, Ancient Colchis and the origins of iron. Interim results from recent field survey work in Guria, western Georgia, in: N. Sekunda (ed.), Wonders Lost and Found: A Book to Celebrate the Archaeological Work of Professor Michael Vickers (submitted).

Patterns of Settlement and Metallurgy in Late Bronze–Early Iron Age Kvemo Kartli, Southern Georgia

51

Gzelishvili 1964 И. А. Гзелишвили, Железоплавильное Производство в Древней Грузии (Tbilisi 1964). Hammer 2014 E. Hammer, Highland fortress-polities and their settlement systems in the southern Caucasus, Antiquity 88, 2014, 757–774. Hansen 2012 S. Hansen, The copper bead, in: B. Lyonnet – F. Guliyev – B. Helwing – T. Aliyev – S. Hansen – G. Mirtskhulava, Ancient Kura 2010–2011. The first two seasons of joint fieldwork in the Southern Caucasus, Archäologische Mitteilungen aus Iran und Turan 44, 2012, 84–85. Javakhishvili – Japaridze 1975 ა. ჯავახიშვილი – ო. ჯაფარიძე / A. Javakhishvili – O. Japaridze (eds.), ქვემო ქართლის არქეოლოგიური ექსპედიციის შედეგები / Kvemo Kartlis Arkeologiuri Ekspeditsiis Shedegebi (Reports of the Kvemo Kartli Archaeological Expedition) (Tbilisi 1975). Kavtaradze 1999 G. L. Kavtaradze, The importance of metallurgical data for the formation of a Central Transcaucasian chronology, in: A. Hauptmann – E. Pernicka – T. Rehren – Ü. Yalçin (eds.), The Beginnings of Metallurgy. Proceedings of the International Conference ‘The Beginnings of Metallurgy’, Bochum 1995, Der Anschnitt Beiheft 9 (Bochum 1999) 67–101. Khakhutaishvili 1987 Д.А. Хахутаишвили, Производство Железа в Древней Колхиде (Tbilisi 1987). Killick 1991 D. Killick, A little-known extractive process. Iron smelting in natural draft furnaces, The Journal of the Minerals, Metals & Materials Society 43, 1991, 63–64. Lindsay – Greene 2013 I. Lindsay – A. Greene, Sovereignty, mobility, and political cartographies in Late Bronze Age Southern Caucasia, Journal of Anthropological Archaeology 32, 2013, 691–712. Lindsay et al. 2010 I. Lindsay – A. T. Smith – R. S. Badalyan, Magnetic survey in the investigation of sociopolitical change at a Late Bronze Age fortress settlement in northwestern Armenia, Archaeological Prospection 17, 2010, 15–27. Marshall 2014 M. E. Marshall, Subjected(ed) Bodies. A Bioarchaeological Investigation of Late Bronze Age – Iron I (1500–800 B.C.) Armenia (PhD Diss., University of Chicago, Chicago 2014). Mirau 1997 N. A. Mirau, Social context of early ironworking in the Levant, in: W. A. Aufrecht – N. A. Mirau – S. W. Gauley (eds.), Urbanism in Antiquity. From Mespotamia to Crete (Sheffield 1997) 99–115. Mkrtchyan et al. 1968 С. С. Мкртчян – К. Н. Паффенгольц – Э. А. Хачатурян, Алавердский Рудный Район ( Yerevan 1968). Mudzhiri 2011 Т. Муджири, Горнорудное производство в древней Грузии (Tbilisi 2011). Nazarov 1966 Ю. И. Назаров, Особенности формирования и прогноз глубинных (скрытых) месторождений медноколчеданной формации южной Грузии (Moscow 1966). Papuashvili – Jibladze 2014 რ. პაპუაშვილი – ლ. ჯიბლაძე / R. Papuashvili – L. Jibladze, ბრინჯაოს მეორადი მეტალურგიული წარმოების ცენტრი ყულევის ნამოსახლარიდან / Brinjaos meoradi metalurgiuli tsarmoebis tsentri qulevis namosakhlaridan (A centre of secondary metallurgical production at the Qulevi settlement), ძიებანი / Dziebani 22, 2014, 178–191. Piccelauri 1997 K. Picchelauri, Waffen der Bronzezeit aus Ost-Georgien (Espelkamp 1997).

52

Nathaniel L. Erb-Satullo

Rehren 2011 T. Rehren, The production of silver in South America, Archaeology International 13, 2011, 76–83. Smith 2015 A. T. Smith, The Political Machine. Assembling Sovereignty in the Bronze Age Caucasus (Princeton 2015). Smith – Leon 2014 A. T. Smith – J. F. Leon, Divination and sovereignty. The Late Bronze Age shrines at Gegharot, Armenia, American Journal of Archaeology 118, 2014, 549–563. Smith et al. 2009 A. T. Smith – R. S. Badalyan – P. S. Avetisyan, The Archaeology and Geography of Ancient Transcaucasian Societies, Volume 1. The Foundations of Research and Regional Survey in the Tsaghkahovit Plain, Armenia (Chicago 2009). Stark 1991 M. T Stark., Ceramic production and community specialization. A Kalinga ethnoarchaeological study, World Archaeology 23, 1991, 64–78. Stöllner – Gambashidze 2011 T. Stöllner – I. Gambashidze, Gold in Georgia II. The oldest gold mine in the world, in: Ü. Yalçin (ed.), Anatolian Metal V, Der Anschnitt Beiheft 24 (Bochum 2011) 187–199. Stöllner – Gambashidze 2014 T. Stöllner – I. Gambashidze, The gold mine of Sakdrisi and the earliest mining and metallurgy in the Transcaucasus and the Kura- valley system, in: G. Narimanishvili (ed.), Problems of Early Metal Age Archaeology of Caucasus and Anatolia (Tbilisi 2014) 102–124. Stöllner et al. 2008 T. Stöllner – I. Gambashidze – A. Hauptmann, The earliest gold mining of the ancient world? Research on an Early Bronze Age gold mine in Georgia, in: Ü. Yalçin – H. Özbal – A. G. Paşamehmetoğlu (eds.), Ancient Mining in Turkey and the Eastern Mediterrannean (Ankara 2008) 271–288. Ur et al. 2013 J. Ur – L. de Jong – J. Giraud – J. F. Osbourne – J. MacGinnis, Ancient cities and landscapes in the Kurdistan region of Iraq. The Erbil Plain Archaeological Survey 2012 Season, Iraq 75, 2013, 89–117. Veldhuijzen 2012 H. A. Veldhuijzen, Just a few rusty bits. The innovation of iron in the Eastern Mediterranean in the 2nd and 1st millennia BC, in: V. Kassianidou – G. Papasavvas (eds.), Eastern Mediterranean Metallurgy and Metalwork in the Second Millennium BC (Oxford 2012) 237–250. Yahalom-Mack – Eliyahu-Behar 2015 N. Yahalom-Mack – A. Eliyahu-Behar, The transition from bronze to iron in Canaan. Chronology, technology and context, Radiocarbon 57, 2015, 285–205.

The Archaeological Landscape of the Hrazdan River Basin during the Late Bronze–Early Iron Age Manuel Castelluccia * Abstract: This paper aims to evaluate the archaeological landscape patterns of the territory of the Hrazdan River basin, a crucially strategic position that links two of the most important areas of the Armenian highlands, the Lake Sevan basin and the Aras Valley. The chronological focus spans the Late Bronze Age and Early Iron Age (1500 BC to 800 BC), a period in which the lands south of the main Caucasus range were the scene of several noteworthy innovations evident in the archaeological record that include an increase in the quantity and variety of metalwork, the introduction of iron, the emergence of new pottery types and the growth of funerary evidence. Among the most important developments was the transformation of the socio-political structure of the local population, as seen in the settlement pattern, and the emergence of fortress sites, which shows a tendency toward militarisation of the society. In the Hrazdan Basin, these fortresses are formed by circuit walls of large, irregularly shaped boulders; they are situated at high points above the river gorge, on the foothills of the valley and, in the case of Tghit, on a mountain-top. The fortresses are representative of a social organisation based around military aristocratic elites which developed in Armenia during the Late Bronze to Early Iron Age. Keywords: Archaeological landscape; Fortresses; Late Bronze Age; Early Iron Age; Transcaucasia

Introduction During the Late Bronze Age (1500–1100 BC), remarkable material innovations occurred in the societies of southern Caucasia. Firstly, there was a massive increase – in both quantity and variety – in the production of metal objects. Several new types appeared, such as crescent-shaped bronze axes, one-cast daggers with bell-shaped pommel, bronze belts, horse-bits and various kinds of personal adornment. The production of iron objects began gradually in the Late Bronze Age and became more common from the tenth century onwards. Burials also greatly increased and burial types diversified. Several large burial grounds date to this period, where kurgans, dolmens, cistgraves and interments in simple grave-cuts exist side by side. However, cist-graves with low mounds and surrounding cromlechs are the most prevalent. Particularly prominent is a marked increase in settlements, notably fortified sites, termed ‘hillforts,’1 which are characterised by walls built from very large, irregularly shaped boulders. These hill-forts are spread across the mountainous areas of northern Iran, eastern Anatolia and southern Transcaucasia and several important studies have been devoted to the phenomenon.2 They are a clear sign of the initial development of socio-political complexity among the native population and signal an increasingly militarised landscape. During the Late Bronze–Early Iron Age most of present-day Armenia and the neighbouring regions of eastern Georgia, western Azerbaijan and Karabakh shared somewhat similar archaeological evidence, which has been interpreted as the result of a uniting, common culture. Several scholars have addressed the matter, and different interpretations, definitions and approaches have

Associazione Internazionale di Studi sul Mediterraneo e l’Oriente, Rome, [email protected]. Biscione 2009, fn. 1. 2 Mikaelyan 1968; Smith 1996; Smith ‒ Kafadarian 1996; Biscione 2002; Sanamyan 2002; Belli ‒ Konyar 2003; Biscione 2009. *

1

54

Manuel Castelluccia

Fig. 1   The Hrazdan River basin and the sites discussed in the text: 1. Bdžni; 2. Berdi dar; 3. Elar; 4. Kaghsi 1; 5. Kaghsi 2; 6. Kamaris; 7. Karashamb 1; 8. Karashamb 2; 9. Karmir Berd; 10. Lčašen; 11. Muchannat-tapa; 12. Tghit; 13. Tsitsernakaberd (GoogleEarth)

been proposed concerning this culture’s identification and labelling. Names that have been proposed are ‘Ganja-Karabakh’ or ‘Chodžali-Kedabek’ culture3, ‘Lčašen culture’4 or the ‘LčašenMetsamor/Tsitelgori horizon’5. Although there are several shared features, strong regional variations may also be oabserved; however, it is generally accepted that Lake Sevan and the neighbouring areas are the core zone. The Study Area The Hrazdan River basin occupied a strategic position, linking two of the most important economic areas of Armenia: the Aras Valley and the Lake Sevan basin. It has yielded remarkable remains of the Late Bronze–Early Iron Age cultures of the Armenian highlands. Hrazdan is the second largest river of the Republic of Armenia. It originates from Lake Sevan at an elevation of 1900m and flows southwards through the Kotayk province and Armenia’s capital, Yerevan, finally joining the river Aras on the present-day border with Turkey. For most of its course, it passes through the Kotayk plateau in a deep and picturesque gorge. The local geological formations consist of basalts and tuff emitted by the volcanoes in the Gegham range. Today, the river is extensively exploited to irrigate crops and for hydro-electric schemes, which strongly reduces its flow. Before entering into Yerevan’s outskirts, the river divides two diametrically opposed geographic environments (Fig. 1). To the right of the river lie the slopes of the Pambak and Tsaghkunyats mountain ranges and, further south, the isolated extinct volcano of Arayi Ler and the Yegvard plateau; to the left, the landscape is formed by an extensive basalt plateau with several extinct

3 4 5

Hančar 1934; Minkevič-Mustafaeva 1963; Džafarov 1984; Schachner 2001. Pogrebova 2011. Badalyan et al. 2009, 68‒93; Sagona 2012, 257.

The Archaeological Landscape of the Hrazdan River Basin during the Late Bronze–Early Iron Age

55

volcanoes that are part of the Gegham mountain range, one of the main sources of obsidian in Armenia. After flowing through Yerevan, the river passes through the Ararat Plain, one of the largest agricultural areas in the Armenian highlands, stretching from the foothills of Mount Aragats to the north to the Gegham ridge in the east and the base of Mount Ararat to the south. It is a lengthy fertile strip, about 100km long and ranging from 15 to 20 to 45km wide, and is crossed by the river Aras (which marks the modern border between Turkey and Armenia). The Hrazdan Basin served as an important communication route, especially by linking the Aras and Kura rivers. Just north of Lake Sevan flow the rivers Debed and Aghstev, which belong to the Kura River drainage basin. Given this geographical meeting point, it is not surprising that the valleys of the Hrazdan Basin have yielded prolific archaeological remains. In the past, at least three main roadways connected the Lake Sevan basin with the Ararat Valley: one ran beside the present course of the river, which is the same route as the modern Yerevan–Sevan highway. A second road probably crossed the Gegham range, which runs north–south parallel to the western shore of the lake, at a 2700m-high pass; the third followed a route further south through the Selim Pass.6 The area around the lower part of the river Hrazdan has been the site of several important discoveries in Caucasian archaeology. The Urartian centres of Karmir Blur and Erebuni, as well as the Early Bronze Age settlement of Shengavit, are located within the urban limits of Yerevan. Further south, the Hellenistic settlements of Dvin and Artaxata lie just a few kilometres away from the river. Archaeological Investigations of the Hrazdan River Basin and the Kotayk Plateau In recent decades, new archaeological projects have focused on the study of the archaeological landscape of the Hrazdan River basin and Kotayk plateau. The most notable is a joint Armenian-Austrian expedition investigating the Urartian settlements of Aramus, c. 20km north-east of Yerevan. In the north-western part of Kotayk province, along the river Marmarik, a survey was carried out by a joint expedition between the University of Idaho and the Institute of Archaeology and Ethnography of Armenia, but a report has not yet been published. Since 2013, a joint ItalianArmenian expedition has taken on the task of surveying the upper Hrazdan area, in order to clarify the urban settlement pattern between the Yerevan and Lake Sevan basins.7 This project involved collaboration between the Italian International Association of Mediterranean and Oriental Studies (ISMEO) and the Armenian Institute of Archaeology and Ethnography of the National Academy of Sciences of the Republic of Armenia. The fieldwork was organised firstly with a study of the area using remote sensing techniques, together with a review of the literature concerning the area. An especially valuable source of information is the National List of Monuments of the Republic of Armenia, an inventory of all known sites of cultural interest organised according to the municipalities in which they are located. Subsequently, sites have been visited in the field wherever possible and information gathered on site attributes and artefacts.8 The Armenian-Italian expedition achieved some important results, which will be further discussed below. However, in order to tentatively reconstruct the archaeological landscape of the whole Hrazdan River basin, first it is necessary to gather together all related literature, especially that of the Soviet period. The present review of the available archaeological evidence will start in relation to its southernmost part, that is, the Armenian capital, Yerevan. Reconstructing the archaeological landscape of modern-day Yerevan and its outskirts is not an easy task, since the expansion of the modern

6 7 8

Biscione – Dan 2011. Castelluccia et al. 2012; Petrosyan et al. 2015. Castelluccia et al. 2012, 28.

56

Manuel Castelluccia

Fig. 2   Tsitsernakaberd fortress (after Smith ‒ Kafadarian 1996, 26, fig. 3)

city has probably covered many archaeological sites. Fortunately, the Armenian scholar S. Esajan gathered together the archaeological evidence concerning the city’s urban limits in a book published during the Soviet period.9 With regard to the Late Bronze–Early Iron Age, within the modern city of Yerevan, only scarce traces of the period have been detected, although at least two fortresses are known. The first of these is located on the hill known as Tsitsernakaberd, on top of a steep promontory overlooking the Hrazdan. Unfortunately, the site was heavily damaged by construction activities which obliterated most of the structures, but fortunately it was mapped before the remains of the walls disappeared altogether and its main features have been published.10 No archaeological investigations were ever carried out at the site and most of the available evidence was unearthed during the building work: a pottery assemblage dating to the late second and early first millennium BC includes grey wares with incised decorations analogous to finds from the same period from numerous sites in Armenia.11 The fortress consists of two lines of defensive walls of ‘cyclopean’ masonry, with large unworked stones set in rough courses and smaller ones filling the spaces between them (Fig. 2). The first line has a perimeter of 281m enclosing 0.2ha; the second circuit encloses a small raised citadel of 0.03ha. The second fortress, Muchannat-tapa, has totally disappeared. It was located near the railway station of Yerevan but has now been covered, if not destroyed, by modern housing (Fig. 3). The archaeological site covered about 0.5ha and included remains from the third millennium BC. Most of the finds, however, date to the Early Iron Age and consist of pottery. It was partially investigated before the Second World War and several reports are available.12 The first excavations began in 1935 and continued the following year under the guidance of E. A. Bajburtjan. Three distinct cultural levels were unearthed, the middle one of which probably contained Urartian material. It is reported that the oldest layer contained monochrome

9 10 11 12

Esajan 1969. Smith ‒ Kafadarian 1996, 26‒28. Smith et al. 2009, 68-90. Bajburtjan 1937; Field ‒ Prostov 1937; Piotrovskij 1966, 23.

The Archaeological Landscape of the Hrazdan River Basin during the Late Bronze–Early Iron Age

57

Fig. 3   Probable site of the Muchannat-tapa fortress (GoogleEarth)

painted pottery,13 while the upper layer can be dated to the Hellenistic period. B. Kuftin also reported the discovery of a short sword and axes made of iron, red pottery with geometric designs in black, and a cylinder stamp seal bearing the image of a bird, probably Urartian.14 Moreover, he describes pottery dating to the third millennium.15 Subsequently Esajan published a complete summary of the archaeological evidence from the site.16 He identified a bronze belt with dotted decoration as coming from this site, but unfortunately further information is lacking.17 The presence of Urartian-style items suggests that this site was still in use during the Middle Iron Age. Traces of a settlement and a necropolis of pre-Urartian date are attested at Karmir Blur; scarce remains from the same period were also found in Erebuni.18 Just a few kilometres north of Yerevan lies the important site of Karmir Berd (literally ‘Red Hill’), also known in the archaeological literature under the names Tazakend, Kizil-Kala and Gaja-Charaba. It stands on a rocky promontory on the right bank of the gorge of the river Hrazdan and consists of a fortress and a large burial ground. The fortress is surrounded on three sides by the cliffs of the 80m-deep river gorge and is open only on the north-east side (Fig. 4), where there are massive fortification walls built with large blocks of basalt (Fig. 5). The fortress covers an area of about 3.5ha. It was reused in the mediaeval era and hence the oldest structures are partially hidden or destroyed. The fortress has yielded material from the mid-2nd millennium to the 7th century BC, plus the later occupation during the Middle Ages. In the burial ground, remarkable Middle Bronze Age discoveries have been made. Several past expeditions have concentrated on excavating the cemetery. It was excavated for the first time in 1896 by P. V. Čarkovskij, who unearthed 22 tombs. The site was investigated again in 1903 by M. Zachar’janc, who dug another 17 tombs,19 and the following year E. Rösler investigated 28 burials.20 Further investigations were carried out some decades later: B. Piotrovskij worked there in 1934,21 A. Martirosjan in 1962, L. Karapetjan

13 14 15 16 17 18 19 20 21

Bajburtjan 1937, 212‒213. Kuftin 1943, 59. Kuftin 1943, 120‒123, figs. 76‒78. Esajan 1969, 25‒29. Esayan 1984, pl. 8, no. 28. Ter-Martirosov 2012, 170. IAK 1909. OAK 1904, 98‒99. Piotrovskij 1949, 43‒44.

58

Manuel Castelluccia

a

b Fig. 4   a. Sketch of Čarkovskij of Karmir Berd (after Kušnareva 1960, 138, fig. 1); b. Plain of Karmir Berd (after Karapetjan 1972, 165, fig. 1)

Fig. 5   Remains of ‘cyclopean’ walls in Karmir Berd fortress (photo: M. Castelluccia)

carried out some soundings within the fortress in 1966,22 and finally, from 1965 to 1967, Esajan carefully investigated the necropolis, digging 73 graves.23 The Late Bronze–Early Iron Age cemetery consists largely of cist-graves lined with stone slabs and covered with small mounds. In some cases there is also a cromlech. The average length of the cists is 1.8m, while the average width is 0.8–0.9m. Generally only three sides of the cist are lined with stone slabs, while the northern side is formed instead of smaller, irregular stones. Other stone slabs are sometimes present on the bottom of the pit. Esajan recognised three different grave shapes: rectangular, square and round. In many graves the bones have not been preserved, although some did contain skeletal remains: the deceased usually lay on the left side, head to the north or east, in a crouched position. The burials are all single inhumations except for Tomb 3, where two skeletons were found, one lying on the left side and the other on the right, both with the head pointing eastwards. The grave goods were not particularly abundant, comprising small groups of pottery and some metalwork (Fig. 6).

22 23

Karapetjan 1972. Esajan 1969, 29‒52.

The Archaeological Landscape of the Hrazdan River Basin during the Late Bronze–Early Iron Age

59

Fig. 6   Finds from Karmir Berd (after Esajan 1976, pls. 21, 22, 26)

East of Karmir Berd there are two other fortresses, Kamaris24 and Elar,25 which show scarce traces of the Early Iron Age. North of Karmir Blur lies another remarkable site, Karashamb, known for its Middle Bronze Age necropolis, which also contains material from the Late Bronze Age. Near the large burial ground there is a fortress, but it is yet to be investigated. Near Karashamb, on the opposite side of the river, traces of a fortress have been attested on a rocky spur in the Hrazdan River gorge; it is referred to here as Karashamb 2.26 In the western part of the site stand the remains of huge ‘cyclopean’ retaining walls, while in the eastern section, several stone-built structures are still clearly visible. Much of the pottery collected from both site and hillside dates to the Early and Middle Bronze Age, but some sherds appear to be of Late Bronze Age date. According to a local villager, the easternmost part of the rocky spur was used in Soviet times as a basalt quarry. Despite the very scarce presence of Late Bronze Age pottery, it is plausible that the walls of this site date mostly to that period, since fortresses with masonry of that kind are typical of this epoch. Further to the north, a very important fortified settlement known as Tghit overlooks the village of Teghenik on a rocky outcrop of the Tsaghkunyats range; it was investigated by the ItalianArmenian team in 2014.27 The fortress provides a view stretching from the Gegham range to the Aras Valley, and overlooks the entire Hrazdan Valley. The few recovered sherds suggested a Late Bronze–Early Iron Age date, with traces of the medieval period as well. The fortress is rectangular in shape (Fig. 7). Single fortification walls run along the northern, southern and eastern sides; the eastern wall is built just above a vertical cliff overlooking the village of Teghenik. The western side is less steep, sloping down towards a large depression, and features a triple fortification system of walls. The upper line of fortification is the main one and measures about 70 × 100m; it is built with very large stones, with walls over 4m thick and pre-

24 25 26 27

Smith ‒ Kafadarian 1996, 25. Khanzadyan 1979. Castelluccia et al. 2012, 29. Petrosyan et al. 2015, 62.

60

Manuel Castelluccia

Fig. 7   Aerial view of the fortress of Tghit (GoogleEarth)

Fig. 8   Defensive wall of Tghit (after A. Petrosyan et al. 2015, pl. 22, no. 2)

The Archaeological Landscape of the Hrazdan River Basin during the Late Bronze–Early Iron Age

61

served up to 5m in height (Fig. 8). A large opening, probably part of a 6m-wide gate, is present in the middle of this upper fortification. To have such a wide gate would be unusual, since entrances are perhaps the weakest point of a fortress. Traces of smaller walls are preserved within the entrance itself; these may have belonged to an additional system of defence or perhaps pertain to a later period. On the first wall there are also three 12m-long buttresses, protruding about 3m from its face. The second defensive line consists of thinner walls with a width of about 2.5m and several circular towers. The gate of the second line is not in line with the main entrance but several metres to the south. This disposition increased defensive capabilities, since an attacking force, after breaking through the first entrance, would have moved toward the second gate, exposing its right side for several metres to defenders located on the top of the second wall. The third defensive wall line was the smallest, with walls 1.5–2.5m thick and, in several places, it has collapsed or is completely covered by vegetation. No entrance was identified. When visited, the inner part of the site was densely covered with high vegetation, making it impossible to see the layout of the inner buildings clearly. The buildings’ presence was, however, visible to some degree. There was a large oval-shaped body of water enclosed by an earth embankment reinforced by large stone blocks 190m north of the fortress; it measured 110 × 200m and was clearly a reservoir intended to supply water to the fort. A canal entering the lake from the north probably conveyed runoff from the slope generated by rainwater and melting snow. Another canal exits from the lake to the south, passing close to the fortress and continuing down towards the hills beneath. The presence of another fortress has been recently reported near the village of Bdžni, but no further information is available at present.28 The remains of two further small fortresses were identified above the Hrazdan Gorge, near the village of Kaghsi.29 The first was built on a rocky spur protected on three sides by the Hrazdan Gorge. A large wall with buttresses defends the side facing the plateau; a small amount of Late Bronze–Early Iron Age pottery was found there along with some medieval sherds. The second fortress is located just 250m north-east of the first. Its 3.7m-thick ‘cyclopean’ walls are visible for only one course, and are built without buttresses or towers. The main entrance is visible on the north-west wall. Both fortresses have typical double-faced masonry with an internal fill of smaller stones and earth. West of Kaghsi, a few hundred metres from the modern Yerevan–Sevan highway, lies the fortified site of Solak, which has been investigated by the Italian-Armenian team.30 Although the structures date to the Middle Iron Age, a small amount of Late Bronze–Early Iron Age pottery was found on the surface. Next to the modern village of Lernanist there is another remarkable fortress, which is located on a mountain top overlooking the whole upper Hrazdan Valley and is known as Berdi Dar. It is trapezoidal in shape, with the longest side measuring about 50m (Fig. 9); the main defensive wall, around 3m thick and preserved up to 1m in height, is built with ‘cyclopean’ masonry. An entrance flanked by two towers is still clearly visible on the north-eastern side and several buttresses are present along the eastern side. Within the structure, traces of buildings can be seen. Surprisingly, no surface pottery was found at the site, so its identification as dating to the Late Bronze–Early Iron Age is somewhat uncertain. The presence of fortresses has been reported along the western part of the Pambak range and the Marmarik valleys, but these structures have yet to be studied in detail. At the point where the river Hrazdan exits from Lake Sevan lies one of the most important sites of the whole Armenian highlands for the Late Bronze–Early Iron Age, the fortress of Lčašen

28 29 30

Gharibyan et al. 2008. Petrosyan et al. 2015, 62. Petrosyan et al. 2015, 65‒67.

62

Manuel Castelluccia

Fig. 9   Aerial view of the fortress of Berdi Dar (Petrosyan et al. 2015, pl. 22.1)

Fig. 10   The fortress of Lčašen (Biscione – Parmegiani 2004, fig. 1).

The Archaeological Landscape of the Hrazdan River Basin during the Late Bronze–Early Iron Age

63

(Fig. 10). Lčašen is one of the largest sites of the pre-Urartian period in the Armenian highlands, with an area of over 35ha and extending along 15 ridges. Some soundings have been excavated at the site, but very few results have been published.31 At the foot of the settlement there is a large necropolis, 1.5km long and 200–300m wide; for a long time it has been submerged under the waters of the lake following a rise in the water level. Along with fortresses and the associated cemetery, isolated burial grounds are attested near the river, but none of them have been fully investigated and published. A large cemetery is located on the outskirts of Hrazdan in the locality known as Jrarat and was partially investigated by A. Mnatsakanyan.32 Unexcavated graves with cromlechs are still visible at the site. A small burial ground is also present next to the village of Kaghsi, in the river gorge. The presence of graves was discovered during construction work in Soviet times and some test excavations were carried out there by Mnatsakanyan.33 Other Late Bronze–Early Iron Age graves were investigated in the village of Bdžni, but these have been reported only in brief publications.34 The graves show the typical features of the period: tombs with cromlechs, low mounds and small stone cist-graves. The burials are mostly single and the deceased have usually been placed in a crouched position. Grave goods typically include some metalwork, weapons and various personal adornments, as well as a small amount of pottery. Conclusion In conclusion, the archaeological landscape of the Hrazdan River basin features a plurality of fortified settlements, mostly erected during the Late Bronze–Early Iron Age. Some fortresses are located just above the river gorge which offers a strong defensive position, whereas others are located on the foothills of the valley. All these sites are situated near economic resources, strategically controlling both the route passing along the river and the agricultural plains. Only the large fortress of Tghit is isolated on a mountain top. As mentioned at the beginning of this paper, the development of fortified settlements is a distinctive trait of a socio-political process involving the militarisation of society which characterised Transcaucasia and the neighbouring regions during the Late Bronze Age35. This socio-political process, whose origin can be dated to the end of the third millennium, led to the emergence of societies with a high degree of internal complexity and has been defined by the Russian scholar, B. M. Masson, as ‘Кавказский путь к цивилизации’ (‘the Caucasian way to civilisation’).36 This model is characterised by non-urban, non-state societies with strong social differentiation and an extremely unequal distribution of wealth, ruled by military aristocracies with a great capacity for accumulating wealth and organising labour and manpower, and with a hierarchy of large and smaller settlements.37 The model developed further during the Late Bronze–Early Iron Age towards a pre-state organisation characterised by the wider sharing of power and the increased size of the ruling military elites. The final development of the ‘Caucasian model’ proposed by Masson can be seen in the emergence of the Urartian kingdom, at the beginning of the ninth century BC, which subsequently brought the whole Armenian highlands under its control.

31 32 33 34 35 36 37

Mikaeljan 1968; Biscione – Parmegiani 2004. Mnatsakanyan ‒ Tiratsyan 1961, 69; Castelluccia et al. 2012, 30. Castelluccia et al. 2012, 30. Avetyan ‒ Biyagov 1977. Castelluccia 2017. Masson 1997. Masson 1997, 127‒132.

64

Manuel Castelluccia

References Aslanov et al. 1959 Г. М. Асланов ‒ Р. М. Ваидов ‒ Г. И. Ионе, Древний Мингечаур: эпоха энеолита и бронзы (Baku 1959). Avetjan ‒ Biyagov 1977 В. Аветян ‒ Л. Биягов, Новые находки бронзовых поясов, Вестник Общественных Наук Армении 12, 1977, 87‒89. Badalyan et al. 2009 R. Badalyan ‒ P. Avetisyan ‒ A. T. Smith, Periodization and chronology of southern Caucasia. From the Early Bronze Age through the Iron III period, in: Smith et al. 2009, 33‒94. Bajburtjan 1937 Е. А. Байбуртян, По поводу древней керамики из Шреш-Блура, Советская Археология 3, 1937, 209‒213. Belli ‒ Konyar 2003 O. Belli ‒ E. Konyar, Early Iron Age Fortresses and Necropolises in East Anatolia (Istanbul 2003). Biscione 2002 R. Biscione, The Iron Age settlement pattern. Pre- Urartian and Urartian periods, in: Biscione et al. 2002, 351‒370. Biscione 2009 R. Biscione, The distribution of pre- and protohistoric hill-forts in Iran, Studi Micenei ed Egeo-Anatolici 50, 2009, 123‒144. Biscione ‒ Dan 2011 R. Biscione ‒ R. Dan, Dimensional and geographical distribution of the Urartian fortifications in the Republic of Armenia, Armenian Journal of Near Eastern Studies (Aramazd) 6, 2, 2011, 104-120. Biscione ‒ Parmegiani 2004 R. Biscione ‒ N. Parmegiani, Armenian-Italian archaeological expedition. Field season 2004, Studi Micenei ed Egeo-Anatolici 46, 2, 2004, 284‒295. Biscione et al. 2002 R. Biscione ‒ S. Hmayakyan ‒ N. Parmegiani, The North-Eastern Frontier. Urartians and Non-Urartians in the Sevan Lake Basin. I. The Southern Shores (Rome 2002). Castelluccia 2017 M. Castelluccia, The militarization of a society. The example of Transcaucasia in the Early Iron Age. An archaeological overview, Ancient Civilizations from Scythia to Siberia 23, 2017, 91‒126. Castelluccia et al. 2012 M. Castelluccia ‒ R. Dan ‒ R. La Farina ‒ A. Petrosyan ‒ M. Raccidi, The first season of the Kotayk Survey Project. Preliminary report, Armenian Journal of Near Eastern Studies (Aramazd) 7, 2, 2012, 28‒35, pls. XII‒XVII, 172‒274. Chačatrjan 1975 Т. С. Хачатрян, Древняя культура Ширака. III–I тыс. до н.э. (Erevan 1975). Chanzadjan 1979 Э. В. Ханзадян, Элар-Дарани (Erevan 1979). Chanzadjan et al. 1973 Э. В. Ханзадян ‒ К. А. Мкртчян ‒ Э. С. Парсамян, Мецамор. Исследование по данным раскопок 1965‒1966 гг. (Erevan 1973). Devedžjan 1981 С. Г. Деведжян, Лори-Берд I (Erevan 1981). Džafarov 1984 Г. Ф. Джафаров, Связи Азербайджана со странами Передней Азии в эпоху поздней бронзы и раннего железа: по археологическим материалам Азербайджана (Baku 1984).

The Archaeological Landscape of the Hrazdan River Basin during the Late Bronze–Early Iron Age

65

Esajan 1969 С. И. Есаян, Ереван. Археологический очерк (Erevan 1969). Esajan 1976 С. И. Есаян, Древняя культура племен Северо-Востоной Армении (III‒I тыс. до н.э.) (Erevan 1976). Esayan 1984 S. A. Esayan, Gürtelbleche der älteren Eisenzeit in Armenien, Beiträge zur allgemeinen und vergleichenden Archäologie 6, 1984, 97‒198. Field ‒ Prostov 1937 H. Field ‒ E. Prostov, Archaeology in the Soviet Union, American Anthropologist 39, 3, part 1, 1937, 457‒490. Gharibyan et al. 2008 Ի. Ղարիբյան – Հ. Հակոբյան – Ա. Գնունի – Գ. Խաչատրյան – Տ. Վարդանեսովա / I. Gharibyan ‒ H. Hokobyan ‒ A. Gnuni ‒ G. Khachatryan ‒ T. Vartanesova, Հնագույն Բջնին. Նախնական դիտարկումներ / Hnaguyn Bjnin. Nakhnakan ditarkumner (The Ancient Bjni. Preliminary observations), Պատմա-բանասիրական հանդես / Patmabanasirakan handes 3, 2008, 185‒196. Hančar 1934 F. Hančar, Kaukasus-Luristan. Züge kultureller Verwandtschaft des prähistorischen Kaukasusgebietes mit dem Alten Orient, Eurasia Septentrionalis Antiqua 9, 1934, 47‒112. IAK 1909 А.С., Некоторые Закавказские могильники, Известия Императорской Археологической Комиссии 29, 1909, 1–17. Karapetjan 1972 Л. Л. Карапетян, Раскопки Кармирбердской (Тазакендской) крепости, Советская археология 2, 1972, 164‒170. Khanzadyan 1979 է: Վ. Խանզադյան / E. Khanzadyan, Էլառ-Դարանի / Elar-Darani (Erevan 1979). Kuftin 1943 Б. А. Куфтин, Урартский “колумбарий” у подошвы Арарата и куро-араксский энеолит (Tbilisi 1943). Kušnareva 1960 К. Х. Кушнарева. Тазакендский могильник в Армении, Советская археология 1, 1960, 137–147. Martirosjan 1964 А. А. Мартиросян, Армения в эпоху бронзы и раннего железа (Erevan 1964). Masson 1997 В. М. Массон, Кавказский путь к цивилизации: вопросы социокультурной интерпретации, Древние общества Кавказа в эпоху палеометалла. Ранние комплексные общества и вопросы культурной трансформации (Saint Petersburg 1997) 124‒133. Mikaeljan 1968 Г. А. Микаелян, Циклопические крепости Севанского бассейна (Erevan 1968). Minkevič-Mustafaeva 1963 Н. В. Минкевич-Мустафаева, О датировке и хронологических этапах некоторых памятников Азербайджана эпохи поздней бронзы и раннего железа, Материальная культура Азербайджана 4, 1962, 109‒139. Mnatsakanyan ‒ Tiratsyan 1961 А. О. Мнацаканян ‒ Г. А. Тирацян, Новые данные о материальной культуре древней Армении, Լրաբեր հասարակական գիտությունների / Lraber hasarakakan gitutyunneri 8, 1961, 69‒83. OAK 1904 Отдел I. Производство археологических исследований. 5. Тифлисская губерния, Отчет императорской археологической комиссии за 1902 год, 1904, 97–102.

66

Manuel Castelluccia

Petrosyan et al. 2015 A. Petrosyan ‒ R. Dan ‒ R. La Farina ‒ M. Raccidi ‒ M. Castelluccia ‒ B. Gasparyan ‒ A. Babajanyan, The Kotayk Survey Project (KSP). Preliminary report on 2014 fieldwork activity, Armenian Journal of Near Eastern Studies (Aramazd) 9, 1, 2015, 58‒68. Picchelauri 1979 К. Н. Пицхелаури, Восточная Грузия в конце бронзового века (Tbilisi 1979). Piotrovskij 1949 Б. Б. Пиотровский, Археология Закавказья (с древнейших времен до I тысячелетия до н. э.) (Leningrad 1949). Piotrovskij 1966 B. B. Piotrovskij, Il regno di Van. Urartu (Rome 1966). Pogrebova 2011 М. Н. Погребова, История Восточного Закавказья. Вторая половина II – начало I тыс. до н. (Moscow 2011). Sagona 2012 A. Sagona, Remarks on the East Anatolian Iron Age, in: A. Sagona ‒ A. Çilingiroğlu (eds.), Anatolian Iron Ages 7. The Proceedings of the Seventh Anatolian Iron Ages Colloquium held at Edirne, 19–24 April 2010 (Leuven 2012) 253‒267. Sanamyan 2002 H. Sanamyan, Architectural structure, defensive systems and building techniques of the fortifications, in: Biscione et al. 2002, 325‒350. Schachner 2001 A. Schachner, Azerbaycan. Eine terra incognita der Vorderasiatischen Archäologie, Mitteilungen der Deutschen Orientgesellschaft 133, 2001, 251‒332. Smith 1996 A. T. Smith, Imperial Archipelago. The making of the Urartian Landscape in Southern Transcaucasia (PhD Diss., University of Arizona, Tucson 1996). Smith ‒ Kafadarian 1996 A. T. Smith ‒ K. Kafadarian, New plans of Early Iron Age and Urartian fortresses in Armenia. A preliminary report on the ancient landscapes project, Iran 34, 23‒37. Smith et al. 2009 A. T. Smith ‒ R. S. Badaljan ‒ P. Avetisyan, The Archaeology and Geography of Ancient Transcaucasian Societies, Volume I. The Foundations of Research and Regional Survey in the Tsaghkahovit Plain, Armenia (Chicago 2009). Ter-Martirosov 2012 F. Ter-Martirosov, From the state of Urartu to the formation of the Armenian kingdom, in: S. Kroll ‒ C. Gruber ‒ U. Hellwag ‒ M. Roaf ‒ P. Zimansky, Biainili-Urartu. The Proceedings of the Symposium held in Munich 12‒14 October 2007 (Leuven 2012) 169‒176.

A Survey Project on the Borderlands of Turkey – Armenia – Nakhchivan – North-western Iran: Preliminary Report of the 2014–2016 Surveys on the Middle Araxes Basin Ayhan Yardimciel * – Mehmet Ali Özdemir ** – Mehmet Işıklı *** Abstract: The Araxes River basin is a significant geographical and cultural feature of South Caucasia that runs between the highlands of eastern Anatolia and across the Southern Caucasus. Geographically the Araxes Basin incorporates a large part of Armenia, the whole of Nakhchivan, a large part of north-western Iran (Iranian Azerbaijan) and part of eastern Turkey. This sub-region has distinctive geographical features, climate, environment and diverse natural resources. For these reasons, this area, as well as the larger Caucasus region, has been a centre of attraction throughout the ages. Archaeologically, the Middle Araxes Basin, particularly its Turkish parts, has not been sufficiently investigated; however, recent research and projects are addressing this situation. During the last three years a team from the archaeology departments of Kafkas University in Kars and Atatürk University in Erzurum has carried out an archaeological survey within a large region very close to the borderlands of Turkey – Armenia – Nakhchivan – Iran. The main aim of this paper is to present the preliminary results of the project’s work conducted between 2014 and 2016. Keywords: Extensive survey; Fortresses; Eastern Anatolia; Kars Plateau; Araxes River

Introduction Geographically, the Middle Araxes Basin spans a large part of Armenia, the whole of Nakhchivan, a large part of north-western Iran (Iranian Azerbaijan) and both the Kars and Iğdır provinces of eastern Turkey (Fig. 1a). Throughout the ages, this area – forming part of the Southern Caucasus – has been a centre of cultural significance by reason of its distinctive geographical features and climate, its strategic location and the diversity of its natural resources. Archaeologically, the Middle Araxes Basin has not been widely investigated, particularly its Turkish part; however, recent and current research projects are redressing this shortfall.1 During the last three years a team from the archaeology departments of Kafkas University in Kars and Atatürk University in Erzurum has carried out an archaeological survey within a large region very close to the borderlands of Turkey, Armenia, Nakhchivan and Iran.2 An area of almost 4000 square kilometres has been investigated as part of the first stage of this project (Fig. 1b). In the first three years of fieldwork, the survey has focused on gaining an overall recognition of the survey area’s archaeology which will be followed by a more local and intensive survey. The primary aim of this chapter is to present the preliminary results of the initial work, which was conducted between 2014 and 2016.

Tourism Faculty, Kafkas University, Kars, [email protected]. Protohistory and Near East Archaeology Department, Atatürk University, Erzurum, [email protected]. *** Protohistory and Near East Archaeology Department, Atatürk University, Erzurum, [email protected]. 1 Particularly during the last two decades, the archaeological investigations in the Armenian and Nakhchivan parts of the Middle Araxes Basin have been of vital importance for the regional archaeology. For more see: Badalyan et al. 2008; Bakhshaliyev – Marro 2009; Greene – Lindsay 2013; Hammer 2014; Lindsay et al. 2008; Lindsay et al. 2010; Lindsay – Greene 2013; Ristvet et al. 2012; Lindsay et al. 2014. 2 This survey project has been carried out by courtesy of the Cultural Ministry in Turkey and with the financial support of Kafkas University in Kars. This report is focused on survey materials from the survey project conducted by Dr Akın Bingöl between 2014 and 2015 and from the survey project conducted by Dr Ayhan Yardımcıel in 2016. *

**

68

Ayhan Yardimciel – Mehmet Ali Özdemir – Mehmet Işıklı

Fig. 1   a. Map showing survey areas; b. Archaeological sites in surveyed area: 1. Yenigazi Castle; 2. Kırankaya Castle; 3. Değirmentepe; 4. Azat Höyük; 5. Çalıgüney Castle; 6. Karakale; 7. Kuzeyhan Kurgans; 8. Kuzeyhan Castle; 9. Ziyarettepe Castle; 10. The necropolis of the city of Ani 11. Tuzluca Tuz Dağı; 12. Çiçekli Castle; 13. Musun I–II kurgans; 14. Melekli Höyük; 15. Bulakbaşı Castle; c. Profile view showing elevations of archaeological sites (graphics: M. A. Özdemir)

A Survey Project on the Borderlands of Turkey – Armenia – Nakhchivan – North-western Iran

69

The Survey Area and Survey Methodology The survey area is located in the Middle Araxes Basin and covers the lands of the Kars and Iğdır provinces and the Doğubeyazıt district in Ağrı province (Fig. 1b). The juncture of the Araxes and Arpaçay Rivers forms the central part of the survey region, which covers the western banks of the Arpaçay River and the southern and northern banks of the Araxes River, which spans an area of almost 4000km². This large region’s diverse geographical features include lowlands and river valleys, plateaus and mountainsides.3 The methodologies and strategies of the survey have been determined according to the geomorphology of the region; consequently the survey area has been divided into two main parts – plateaus and plains. These two main sections have then been divided into smaller areas taking account of principal geographical features such as hills, slopes, rivers and valleys, with the survey routes being specified according to these criteria. With the aim of the first years of fieldwork being to recognise and discover the region’s overall landscape, identification and analysis of key archaeological units such as mounds, castles and cemeteries have been prioritised. GPS (Global Positioning System) has been used to fix the location of each archaeological unit, and a topographical map of the survey area has been prepared with the determined units located on this map. The next step was to collect archaeological materials systematically from the surface and, lastly, the environmental status of each analysed archaeological unit has been observed. The surface of the site that is to be surveyed is divided into quarters and material collected accordingly. Materials are processed by being cleaned, drawn, photographed and analysed. All data generated from the surveyed site and its materials are then added to the project database using FileMaker Pro12 (Figs. 1b–c). As for the geography and borders of the survey area, the first sub-region is the Kars Plateau which consists of the territories on the western bank of the Arpaçay River and the northern bank of the Araxes River (Fig. 1b). The first substantial sub-area in this section is the Selim Plain – with a height of over 2000m above sea level (asl), it is one of the highest areas of the Kars Plateau and is bordered by Soğanlı Mountain and the Benliahmet Ridge. The eastern border of the plain is formed by the Kars River which flows through a deep valley. The second sub-area of the Kars Plateau is the Kars Plain. This large, level plain is bordered by the Yahni and Benliahmet ridges, as well as the Leylek, Diktepe and Çakmaktepe hills – which have heights approaching 3000m asl – and the Kars River. The Kars Plain, with a height between 2000 and 2500m asl, is one of the highest level plains in the Southern Caucasus. The last level area of the Kars Plateau in our survey region is the Şöregel Plain,4 which is bordered by Yahni Mountain and the Kars and Arpaçay Rivers. This plain has a height of between1400 and 1700m. The height of the Kars Plateau between the Aras and Arpaçay Rivers decreases from west to east, and this height difference creates the regional climate and vegetation zones within this large plateau. Indeed, this situation still affects farming patterns among communities living in this region in the present day. Those inhabiting the high Selim and Kars plains subsist on cattle farming, whereas the people inhabiting the lower Şöregel Plain subsist on sheep and goat breeding. The second main part of the survey area is the flat and wide Iğdır Plain, which is closer to the borderlands of Turkey – Armenia – Nakhchivan – Iran and lies low between the Doğubayazıt Plain and the Araxes River. This plain is another ecological zone which has distinctive climatic conditions in mountainous eastern Anatolia. An ecotone zone exists between the Kars Plateau and Iğdır Plain.5 For this reason this sub-area is very rich and varied with regard to plant and animal populations. This ecotone region lies between the juncture point of the Araxes and Arpaçay rivers, which

3 4 5

For more information about the geography of the Middle Araxes Valley see: Atalay – Mortan 2011, 447–530. It was suggested that Şöregel might be derived from the Armenian name Şirak (Shirak), see Kırzıoğlu 1953, 6. An ecotone zone is the intersection point between different ecological systems.

70

Ayhan Yardimciel – Mehmet Ali Özdemir – Mehmet Işıklı

is where the Turkey – Armenia – Nakhchivan – Iran borders intersect, and is known as Sürmeli Çukuru. The Iğdır Plain rises towards Ararat (Ağrı) Mountain and lava flows from this mountain have covered most of the plain. The last flat survey area is Doğubayazıt Plain, which has extensive irrigable agricultural lands and is located at the foot of Ararat (Ağrı) Mountain. The Doğubayazıt and Iğdır Plains are separated by the Pamuk mountain pass. Overview of the History of Research in the Survey Area The Araxes River valley is one of the natural routes crossing between hilly eastern Anatolia and the Southern Caucasus and because of its special location and distinctive ecological conditions it has been host to many communities throughout the ages. Researchers have been drawn to this particular region since the 19th century.6 The first archaeological expedition in the region was Georgian-born Nicholas Marr’s excavation in the ancient city of Ani in 1892. His excavations continued spasmodically between 1904 and 1917.7 In 1914 the Russian P. F. Petrov carried out one season of excavation on Melekli Höyük in the Iğdır Plain,8 and in the 1940s Kılıç Kökten’s survey was the first systematic survey project in the region. Kökten surveyed the Kars Plateau and the Iğdır Plain which are a significant part of the Araxes Basin. Within this framework he also carried out small-scale soundings at Azat Höyük and Ani Höyük.9 In the 1960s, similar works (soundings) were carried out at Ani Höyük and Melekli Höyük by Kemal Balkan.10 From then on, archaeological investigations in the region were interrupted until the 1990s and 2000s, when Semih Güneri and Alpaslan Ceylan respectively visited the region as part of their surveys including the entire hilly north-eastern Anatolia region.11 Between 2000 and 2011, Aynur Özfırat surveyed the Ağrı and Doğubayazıt provinces and carried out an excavation project at Bozkurt Kurgans Necropolis on the Doğubayazıt Plain.12 Akın Bingöl has maintained a survey project focused on the Kars and Iğdır provinces since 2014.13 Sites Selim Plain on the Kars Plateau Yenigazi Castle Yenigazi Castle, located in a village of the same name, is 10km from the Sarıkamış district. The castle is located on a rocky hill to the east of the village with the Kars River passing by parallel to it, to the south of the village and castle. The castle encompasses an area of 2700 square metres and its elevation is 1975m asl. The castle has a fortification wall which is made of medium-sized and unformed andesite blocks. The preserved length of the wall is 86m. The northern and eastern parts of the fortification wall, which has bastions, are better preserved than the rest. In addition, a few graves in the form of simple pits were discovered in the citadel, although most have been

6 7 8 9 10 11 12 13

Lindsay – Smith 2006; Bakhshaliyev – Marro 2009; Sagona 2010. Smith 2005, 241–242. The results of excavations on Melekli Höyük were partly published by R. D. Barnett: Barnett 1963. Kökten 1943; Kökten 1944, 667–671. Balkan – Sümer 1967. Güneri 1992; Ceylan 2008. Özfırat 2009; Özfırat 2015. Bingöl 2016.

A Survey Project on the Borderlands of Turkey – Armenia – Nakhchivan – North-western Iran

71

Fig. 2   Sites in the survey area: a1. Yenigazi Castle; a2. The fortification wall of Yenigazi Castle; a3. Pottery fragments from Yenigazi Castle; b1. Kırankaya Castle (fortification wall of the castle); b2. Material from Kırankaya Castle: pottery and obsidian fragments; c1. Azat Höyük; c2. Architectural remains at Azat; c3. Material from Azat: pottery sherds, fragments of obsidian and bone objects and spindle whorl; d1. Değirmentepe site; d2. Pottery fragments from Değirmentepe; e1. General view of the Karakale site; e2. Details from the Kuzeyhan Kurgans; e3. The fortification wall of Kuzeyhan Castle; f1. Çalıgüney Castle; f2. The fortification walls of Çalıgüney Castle; f3. Pottery fragments from Çalıgüney Castle; g1. The necropolis of the city of Ani, Ocaklı village and Harmanyeri (photos: A. Yardimciel; g1: Z. A. Yaşlı)

72

Ayhan Yardimciel – Mehmet Ali Özdemir – Mehmet Işıklı

destroyed by illegal excavators. Some pottery sherds have been collected in the citadel and around it, and they have the features of Iron Age pottery (Figs. 2a, 4l). Kırankaya Castle Kırankaya Castle is within the boundaries of the Sarıkamış district and is not far from Yenigazi Castle. The location of this castle is remarkable. The modern Erzurum–Kars road, a route which was probably also used in ancient times, passes close to this castle. It is located on a rocky hill which dominates this road and the valley through which it runs. There are also some streams, tributaries of the Kars River, around the castle. The shape of the castle is roughly triangular and its elevation above sea level is 1950m. Parts of the fortification wall of the castle have been well preserved. The wall, which is made of cyclopean (large-scaled, unshaped and unworked) rocks, was formed in two rows. There is a passage forming a gateway or entrance on the southern side of the citadel. There are also many illegal excavation pits in the citadel. Some sherds which date to the Bronze (Fig. 4b) and Iron Ages have been found around these pits and elsewhere in the citadel (Fig. 2b). Kars Plain on the Kars Plateau Azat Höyük Undoubtedly Azat Höyük is one of the best-known archaeological sites in the survey area. The mound is located in a village of the same name, 10km away from the city of Kars. The mound and village are situated in the Borluk Valley, which is one of the most important ecological zones of the Kars Plateau. The Borluk stream flows along the valley and creates this ecological zone, which is home to various animal and plant species. The earliest expedition at Azat Höyük and in the region was carried out by Kılıç Kökten in 1948. Unfortunately his brief and smallscale excavation was unable to obtain a good stratigraphic sequence relating to the mound.14 The shape of the mound is roughly round and it is almost 3ha in area. Its height above the valley floor is almost 20m. The northern side of the mound has been badly damaged, and the sections can be observed there. A wall of 5m length, constructed of unworked andesite blocks, can be seen. Some clear traces of the architectural remains can be seen on the surface of the mound too. While surveying on the surface, many sherds and fragments of andirons, dating to the Kura-Araxes Culture, were found around illegal excavation pits. Apart from these items, bone tools, obsidian fragments and tools, and a few spindle whorls were obtained. Overall, the materials collected show that the prominent cultural deposits on the mound are likely to be dated to the Kura-Araxes period (Figs. 4e, 4g, 4h), although some sherds belonging to the Iron Age have also been obtained (Fig. 2c). Değirmentepe Değirmentepe Höyük, which is a level settlement, is located on the outskirts of the modern city of Kars.15 There are freshwater lakes and fertile farmland around the mound. This land reaches as far as Azat Höyük, which is 8km away, and is extremely well suited to grain production. The elevation of the mound above sea level is 1785m; however, the mound has been badly damaged by modern occupation. Based on surface materials, the mound extended over an area of almost 3000 square metres. During the survey, a large quantity of Kura-Araxes pottery sherds were collected, as well as a few sherds belonging to the Middle Bronze Age (Figs. 2d, 4j).

14 15

Kökten 1944, 667–668. Bingöl – Yardimciel 2015.

A Survey Project on the Borderlands of Turkey – Armenia – Nakhchivan – North-western Iran

73

The Complex of Kuzeyhan The Kuzeyhan complex is comprised of a castle, a Kurgan necropolis and a mound, and is located in a valley.16 This large valley is 15km long and encompasses the Digor province and the villages of Hasancan, Oyuklu and Uzunkaya. A stream (Alem) flows through the valley. The Karakale settlement is at 2090m asl and is almost 5ha in area. The settlement lies along the western bank of the Alem stream. There are very clear traces of architectural remains on the surface. Some walls of the building still stand up to 3.5m. The site appears to be a settlement belonging to a civilian population. Some pottery fragments from the Late Bronze Age have been collected from the surface, principally from the stream bed near the mound. There is a necropolis on the south-east of the Karakale site, at a distance of c. 900m and 2170m asl. During the survey, 14 kurgans were recorded. The kurgans are roughly round in shape and their diameters vary between 2 and 5m. In particular, the kurgans with large diameters were surrounded by a double row of stones. Unfortunately no surface artefacts have been found in that area. There is a castle almost 1km away from the necropolis. There is a small valley formed by a little stream between the castle and necropolis area. Kuzeyhan Castle is at 2075m asl, and it is located on a rocky and rugged hill. The area of the castle is almost 11000 square metres. The fortification wall of the citadel has been partly preserved and the length of the wall has been measured at 175m. On the north-west side of the citadel, there is a two-stage gateway. On the north side of the citadel some traces of architectural remains can be easily seen on the surface. The plans of some buildings abutting the northern part of the fortification wall can be distinguished. Sherds dating to the Late Bronze Age have been collected in this area (Fig. 2e). Çalıgüney Castle Çalıgüney Castle is located to the south-east of Azat Höyük, 8km from Azat. This castle is situated at the highest altitude (2227m asl) of all the sites investigated during the survey. This large castle is positioned high on a rocky hill, and small streams flow on its north and south sides. The castle, which is roughly rectangular, dominates the extensive farmlands in the Borluk Valley. The fortification wall of the castle was constructed of partly worked stone blocks. There are some traces of a gateway to the citadel on the western side of the wall. The fortification wall on the eastern side is better preserved than comparable examples at other sites and is almost one metre thick, standing up to 2m in height. The pottery fragments which were collected show Middle Bronze Age (Figs. 4j, 4k) and Late Bronze Age features (Fig. 2f). Kars Plain The Necropolis in the City of Ani The famous medieval city of Ani is located on the western bank of the Arpaçay River, which forms the modern political border of Turkey–Armenia. This river flows within a deep and narrow mountainous valley and it divides this large plateau. The area of the city of Ani has been settled since the Early Bronze Age and the necropolis of this big city has expanded between the city of Ani and modern Ocaklı village along the Bayram stream, which is one of the tributaries of the Arpaçay River. Ani Necropolis is divided into two parts: the western part of the stream is known as Harmanyeri and the eastern part of the stream is known as Cirittepe. As mentioned above, N. Marr, İ. K. Kökten and K. Balkan have carried out excavations in this necropolis at different

16

Bingöl – Yardimciel 2015.

74

Ayhan Yardimciel – Mehmet Ali Özdemir – Mehmet Işıklı

Fig. 3   Sites in the survey area: a1. Ziyarettepe Castle; a2. The fortification walls of the castle; a3. Material from Ziyarettepe: pottery sherds, obsidian fragments and grinding stone; b1. Çiçekli Castle; b2. The fortification walls of the castle; b3. Pottery fragments from Çiçekli Castle; c1. Melekli Höyük; c2. Pottery fragments from Melekli; d1. Aerial view of Tuzluca Tuz Dağı; d2. Tuzluca Tuz Dağı site; d3. Pottery fragments from Tuzluca Tuz Dağı; e1. Bulakbaşı Castle; e2. The fortification walls of the castle; f1. General view of Kurgan I in Musun; f3. Pottery fragments from Kurgan I in Musun; f2. General view of Kurgan II in Musun; f4. Pottery fragments from Kurgan II in Musun, a general view and pottery (photos: A. Yardimciel )

A Survey Project on the Borderlands of Turkey – Armenia – Nakhchivan – North-western Iran

75

dates. During the present survey a large number of pottery fragments with Middle Bronze Age (Fig. 4l) and Early Iron Age properties were collected from this large necropolis area. Some remains of fortification walls have been discovered not far from Ani Necropolis on a plain known as Yevşan Düzü. This area is one of the largest flat expanses of land surrounding the city of Ani, and the medieval fortification walls of the city extend to here. Apart from this wall, the remains of another fortification wall can be observed in this area and it appears to be earlier. In particular, some sherds dating to the Early Iron Age have been found in this area (Fig. 2g). Ziyarettepe Castle Ziyarettepe Castle is located on the road from Kars to the city of Ani, and it is 35km away from Ani.17 The castle, at an altitude of 1800m asl, is positioned on a hill which dominates the flat lands around it. Also, the location of the castle is within easy reach of the borderland between Turkey and Armenia. The shape of the citadel is roughly circular and the remains of its fortifications can be observed clearly. The wall, which has bastions and towers, has been preserved to 2m of its original height. The two-staged gateway to the citadel is located on the south-eastern part of the citadel. During field work, a settlement (outer town?) on the southern slopes of the hill and a necropolis area in the south-west were identified. The pottery fragments collected from all these areas show properties of the Early Bronze Age (Kura-Araxes Culture) (Fig. 4a) and Early Iron Age periods (Figs. 3a, 4o). Iğdır Plain Çiçekli Castle Çiçekli Castle is located in the south of the Tuzluca province and is very close to the Turkey–Armenia–Nakhchivan borderlands. The castle, which is 2250m asl, is located on a highland known as Sinek Yaylası. The fortification wall of the castle has been preserved and can be seen outside the eastern side of the citadel. The eastern side of the citadel has steep and rocky slopes and this topography might have meant the fortification wall on this side may not have been constructed. The shape of the citadel is roughly rectangular. The gateway, which is located on the southern part of the citadel, has a width of 3.2m and from here a road with a slight ramp leads upslope to the citadel. There appears to have been a graveyard on the southern slope beneath the castle. While surveying that area, illegal excavation pits were observed. As a consequence of these illegal excavations, a few graves, which are cromlech-shaped, had been raised to the surface. In addition to these graves, a great number of pottery sherds dated to the Middle Bronze Age (Fig. 4m) and Late Bronze Age were found around and inside the castle (Fig. 3b). Melekli Höyük Melekli Höyük is located on the large, flat Iğdır Plain, where lava flows from Mt Ararat diminish. The mound is in a village of the same name and is not far from the city of Iğdır. The mound is also situated on the road to Nakhchivan. The elevation of the mound is quite low when compared with other sites which were visited during the survey. As mentioned above, the earliest investigation on the mound was carried out by P. F. Petrov in 1914. Petrov’s works were particularly focused on graveyards located south of the mound and the results of his excavation were later published in English by R. D. Barnett.18 Melekli Höyük is one of the biggest mounds in this region and its size is close to 10ha. The road between the Iğdır and Karakoyunlu provinces divided the mound into

17 18

Belli – Yardimciel 2016. Barnett 1963.

76

Ayhan Yardimciel – Mehmet Ali Özdemir – Mehmet Işıklı

two parts. It has been badly damaged, both by this road and by illegal excavations. Many typical Urartian pottery sherds have been collected from the surface of Melekli Höyük (Figs. 3c, 4p). Tuzluca Tuz Dağı This site is located in the Tuzluca province and its height is 1080m asl. One of the principal raw materials of this region is salt and it is likely that this site was founded in conjunction with the salt mine located here. Similar sites are known in Nakhchivan (Duzdağı) and north-western Iran (Khoy).19 This is a small site and surface materials are very scarce. There are no clear traces of architecture on the surface. It is hard to speculate about the site type and function because of the limited surface materials; however, Kura-Araxes pottery was found during the survey (Fig. 4d) as well as a few sherds dated to the Middle and Late Bronze Age (Fig. 3d). Bulakbaşı Castle Bulakbaşı Castle is located to the east of the city of Iğdır, just 19km away. As with Melekli mound, this castle is situated on a low hill which lies just beyond the limits of lava flows from Ağrı (Ararat) Mountain. The castle is surrounded by extensive farmlands and there are water resources thought to come from a small local river, the Karasu Çayı, on the northern part of the citadel, although a relationship between the river and citadel could not be determined from observation of the surface. The fortification walls of the castle have been constructed to follow the rise and fall of the land (Fig. 3e). Kurgans in Musun Two kurgans were recorded in Musun at the village of Suluçem in the Doğubayazıt province. The village is located 33km north-east of Doğubayazıt, where the historical Silk Road passed through the area. The kurgans, located on the flat plain surrounding Aras Güneyi Mountain, are 1800m asl. The shape of Kurgan I is elliptic and measures 38 × 41m, with a height of 2m. Kura-Araxes pottery sherds have been found around Kurgan I. Kurgan II, which is also elliptic, is located on the slope alongside the village of Suluçem, and its elevation is 1790m asl. The linear measurement of Kurgan II is 82 × 54m, and the height of the kurgan is almost 5m. Kurgan II has previously been analysed by A. Özfırat, who reported that painted second millennium pottery was found here.20 Unfortunately, both kurgans have been badly damaged by illegal excavations; however, these excavations brought a very few pottery sherds out into the open, identified as Kura-Araxes (Figs. 3f, 4c, 4f). Conclusions In regard to the ceramic evidence, as is mentioned above, much of the pottery from sites analysed during the survey shows Bronze and Iron Age properties; earlier and later examples are very scant (Fig. 4 and Tab. 1). Typical samples of Kura-Araxes Culture, which characterised the Early Bronze Age of the Southern Caucasus and eastern Anatolia, have been found on Azat Höyük, Değirmentepe kurgans in Musun, Tuzluca Tuz Dağı and Ziyarettepe in the survey area. Most of these fragments are non-diagnostic (body) sherds. The others are fragments belonging to jars and bowls. All of them are handmade. They have mostly slipped and burnished surfaces and inclusions are wide stone-grit tempered. They are mostly medium fired. Just one sample from Azat has a Naxkhcıvan-type handle. Most of the Kura-Araxian pottery from the survey has contrast

19 20

Marro et al. 2010. Özfırat 2011, 71.

A Survey Project on the Borderlands of Turkey – Armenia – Nakhchivan – North-western Iran

77

Fig. 4    Selected ceramic finds from the survey (see Tab. 1 for catalogue entries). Early Bronze Age sherds: a. Ziyarettepe Castle; b. Kırankaya Castle; c. Kurgan II in Musun; d. Tuzluca Tuz Dağı; e. Azat Höyük; f. Kurgan I in Musun; g. Azat Höyük; h. Azat Höyük. Middle Bronze and Iron Ages sherds; i. Değirmentepe; j. Çalıgüney Castle; k. Çalıgüney Castle (photo with permission of Kars Museum); l. The necropolis in the city of Ani; m. Çiçekli Castle; n. Yenigazi Castle; o. Ziyarettepe Castle; p. Melekli Höyük; q. Azat Höyük (photos: A. Yardimciel)

colour patterns (black exterior surface colour and red interior surface colour), and a few samples are brownish-coloured. Similar samples to these are known from many Kura-Araxes sites in the Southern Caucasus and eastern Anatolia.21 The samples show that those with Middle Bronze Age properties are fewer and they were found at Çiçekli Castle, Değirmentepe and Çalıgüneyi Castle. These are monochrome, mostly black in colour, and mostly jars. They are slipped and burnished, with predominantly stone and gritty fabric. Some of the fragments have grooved decorations; apart from this there is no decoration on their surfaces.22 Pottery fragments show that Late Bronze Age and Early Iron Age material is also rare. Diagnostic fragments are very scant and the samples are from jars and bowls. They are mostly handmade and dark coloured, slipped and burnished.23 One sample from the Ziyarettepe necropolis is a brownish-

21 22

23

Sagona 1984; Baldayan – Avetisyan 2007; Palumbi 2008; Bakhshaliyev – Marro 2009; Işıklı 2011. See for parallel samples and comparison: Baldayan – Avetisyan 2007; Bakhshaliyev – Marro 2009; Badalyan et al. 2010; Smith et al. 2009. See for comparison: Badalyan – Avetisyan 2007; Badalyan et al. 2010; Ristvet et al. 2012.

Diameter (cm)

Handmade

Undefined (body)

2.5Y 9/2 2.5Y 9/2 2.5Y 9/2 2.5Y 9/2 2.5Y 8/2 2.5Y 9/2 2.5Y 9/2 2.5Y 9/2 2.5Y 8/4 2.5Y5/4 7.5YR 7/10 2.5Y 9/2

Handmade Handmade Handmade Wheelmade Wheelmade Wheelmade Wheelmade Wheelmade Wheelmade Wheelmade Wheelmade Wheelmade

7.5YR 7/10

7.5YR 7/10

2.5Y 5/4

2.5Y 7/6

10Y 1/2

10R 4/6

10Y 1/2

2.5Y 8/4

10Y 1/2

10Y 1/2

2.5GY 1/2

10Y 1/2

2.5Y 9/2

Temper

Slipped, burnished

Slipped, burnished

Slipped, painted

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Slipped, burnished, painted Slipped, burnished

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Gritted, sanded

Slipped, burnished

Slipped

Slipped, burnished

Slipped, burnished

Slipped, burnished

Slipped, burnished

Slipped, burnished

Slipped, burnished

Slipped, burnished

Slipped, burnished

7.5YR 7/10 Slipped, burnished

2.5Y 5/4

2.5Y 8/4

2.5Y 9/2

2.5Y 9/2

2.5Y 9/2

2.5Y 8/2

2.5Y 9/2

2.5Y 9/2

2.5Y 9/2

2.5Y 9/2

2.5Y 9/2

2.5Y 8/2

10Y 1/2 2.5YR 2/8 7.5YR 7/12

2.5Y 9/2

5GY 7/2

10Y 1/2

2.5 GY 1/2

Surface treatment Slipped, burnished

Tab. 1   Catalogue of selected ceramic finds from the survey, illustrated in Fig. 4

2.5Y 9/2

2.5Y 8/2

2.5Y 9/2

5GY 7/2

Interior surface Exterior surface Paste colour colour colour 2.5Y 9/2 10Y 1/2 2.5Y 9/2

Handmade

Handmade

Handmade

Bowl

Undefined (body)

Handmade

Forming

Undefined (body)

Vessel type

Period: Early Bronze Age (Kura-Araxes Culture) Fig. 4e. Azat Höyük Jar Period: Early Bronze Age (Kura-Araxes Culture) Fig. 4f. Kurgan I in Musun Undefined (body) Period: Early Bronze Age (Kura-Araxes Culture) Fig. 4g. Azat Höyük 10 Bowl Period: Early Bronze Age (Kura-Araxes Culture) Fig. 4h. Azat Höyük 10 Bowl Period: Early Bronze Age (Kura-Araxes Culture) Fig. 4i. Değirmentepe 8 Bowl Period: Middle Bronze Age Fig. 4j. Çalıgüney Castle Undefined (body) Period: Middle Bronze Age Fig. 4k. Çalıgüney Castle Undefined (body) Period: Middle Bronze Age Fig. 4l. Necropolis in the city 14 Bowl of Ani (held in Kars Museum) Period: Middle Bronze Age (Araxes Painted Culture) Fig. 4m. Çiçekli Castle Undefined (body) Period: Middle Bronze Age (Araxes Painted Culture) Fig. 4n. Yenigazi Castle 8 Undefined (body) Period: Iron Age Fig. 4o. Ziyarettepe Castle Undefined (body) Period: Early Iron Age Fig. 4p. Melekli Höyük Bowl Period: Middle Iron Age Fig. 4q. Azat Höyük Undefined (body) Period: Late Iron Age

Fig. 4d. Tuzluca Tuz Dağı

Fig. 4a. Ziyarettepe Castle Period: Early Bronze Age (Kura-Araxes Culture) Fig. 4b. Kırankaya Castle 8.5 Period: Early Bronze Age (Kura-Araxes Culture) Fig. 4c. Kurgan II in Musun Period: Early Bronze Age (Kura-Araxes Culture)

Figure no.

Painted (7.5R 4/14)

Jagged

Incised

Painted (5YR 2.5/1)

Grooved

Grooved

Decoration

78 Ayhan Yardimciel – Mehmet Ali Özdemir – Mehmet Işıklı

A Survey Project on the Borderlands of Turkey – Armenia – Nakhchivan – North-western Iran

79

coloured fragment of a large bowl with handle. Identified Middle Iron Age sherds are mostly characteristic of Urartian pottery. The samples show typical Urartian ceramic features and just a few of them are of a colour which is close to claret red. These samples mostly come from Melekli Höyüğü where there is a large Urartian necropolis. Apart from the Urartian samples, there is only one sample found on Değirmentepe which is a painted fragment belonging to a cup. This fragment, which we regard as important, has buff-coloured fabric and bears light-red-coloured decoration. Lastly, there is a small group which dates to the Late Iron Age. They are from Azat Höyük and are mostly non-diagnostic fragments. They are painted (buff-coloured clay and stripes in light red). Apart from the ceramic evidence, an andiron or cup-stand fragment, a limited number of obsidian fragments and a number of bone tools have been obtained. These samples were also found on Azat Höyük.24 In conclusion, our extensive survey, performed over a large area in the Middle Araxes Valley, has been the result of three years of fieldwork. The main aim of this study has been to present primary results which will show the world of the Southern Caucasus more clearly, within the greater Caucasus, as a great cultural region, especially for archaeology.25 We believe the primary results of our survey project will promote this aim and that future local and intensive survey work will be started so that we may discover more information and greater detail regarding this largely understudied region. Acknowledgements: We would like to give thanks to the Coordinator of Scientific Research Projects in Kafkas University, Kars, for supporting our project financially. Also, we would like to thank Dr Akın Bingöl for his support during our field work. Lastly, we would like to express our gratitude to Ms Jan Bailey, BA (Hons), for editing our text. Her help is extremely valuable.

References Atalay – Mortan 2011 İ. Atalay – K. Mortan, Resimli ve Haritalı Türkiye Bölgesel Coğrafyası (Istanbul 2011). Badalyan – Avetisyan 2007 R. S. Badalyan – P. S. Avetisyan, Bronze and Early Iron Age Archaeological Sites in Armenia. I. Mt. Aragats and its Surrounding Region, BAR International Series 1697 (Oxford 2007). Badalyan et al. 2008 R. S. Badalyan – A. T. Smith – I. Lindsay – L. Khatchadourian – P. Avetisyan, Village, fortress, and town in Bronze and Iron Age southern Caucasia. A preliminary report on the 2003–2006 investigations of Project ArAGATS on the Tsaghkahovit Plain, Republic of Armenia, Archäologische Mitteilungen aus Iran und Turan 40, 2008, 45–105. Badalyan et al. 2010 R. S. Badalyan – A. T. Smith – L. Khatchadourian, Project ArAgats. 10 years of investigations into Bronze and Iron Age sites in the Tsaghkahovit Plain, Republic of Armenia / ArAgats projesi. Ermenistan Tsaghkahovit Ovası’nda 10 yıldır sürdürülen Tunç ve Demir çağ araştırmaları, Türkiye Bilimler Akademisi Arkeoloji Dergisi 13, 2010, 263–276. Bakhshaliyev – Marro 2009 V. Bakhshaliyev – C. Marro, The Archaeology of Nakhchivan. Ten Years of New Discoveries (Istanbul 2009). Balkan – Sümer 1967 K. Balkan – O. Sümer, 1965 Yılı Ani Kazıları Hakkında Kısa Rapor, Türk Arkeoloji Dergisi XIV, 1967, 103–118. Barnett 1963 R. D. Barnett, The Urartian cemetery at Igdyr, Anatolian Studies 13, 1963, 153–198.

24 25

Kalkan 2008. Kushnareva 1997.

80

Ayhan Yardimciel – Mehmet Ali Özdemir – Mehmet Işıklı

Belli – Yardimciel 2016 O. Belli – A. Yardimciel, Eski Çağ’da Kars- Ani Bölgesi’nin En Önemli Savunma Tesisi. Ziyaret Tepe Kalesi, TAÇ. Mimarlık, Arkeoloji Kültür Sanat Dergisi 7, sonbahar 2015 – kış 2016, 2016, 4–13. Bingöl 2016 A. Bingöl, 2014 Kars-Iğdır İlleri Yüzey Araştırmaları, Araştırma Sonuçları Toplantısı 33, 2/2015, 2016, 131–144. Bingöl – Yardimciel 2015 A. Bingöl – A. Yardimciel, 2014 ve 2015 Yılları Kars-Iğdır İl ve İlçeleri Yüzey Araştırmaları Işığında Güney Kafkasya Orta ve Son Tunç Çağı’na İlişkin Veriler, Yeni Türkiye 72, Kafkaslar Özel Sayısı II, Temmuz – Aralık 2015, 91–96. Ceylan 2008 A. Ceylan, Doğu Anadolu Araştırmaları. Erzurum, Erzincan, Kars, Iğdır (1998–2008) (Erzurum 2008). Greene – Lindsay 2013 A. Greene – I. Lindsay, Mobility, territorial commitments, and political organization among Late Bronze Age polities in southern Caucasia, Archeological Papers of the American Anthropological Association 22, 1, 2013, 54–71. Güneri 1992 S. Güneri, Doğu Anadolu’da Yeni Gözlemler, Türk Arkeoloji Dergisi XXX, 1992, 149–195. Hammer 2014 E. Hammer, Highland fortress-polities and their settlement systems in the southern Caucasus, Antiquity 88, 2014, 757–774. Işıklı 2011 M. Işıklı, Doğu Anadolu Erken Transkafkasya Kültürü. Çok Bileşenli Gelişkin Bir Kültürün Analizi (Istanbul 2011). Kalkan 2008 H. Kalkan, M. Ö. 6–4 Yüzyıllarda Doğu Anadolu. Arkeolojik Veriler Işığında Tarihsel ve Kültürel Değerlendirme (PhD Diss., Ege University, İzmir 2008). Kırzıoğlu 1953 M. F. Kırzıoğlu, Kars Tarihi I. Cilt (Istanbul 1953). Kökten 1943 İ. K. Kökten, Kars’ın tarih öncesi hakkında ilk kısa rapor, Belleten VI, 27, 1943, 601–613. Kökten 1944 İ. K. Kökten, Orta, Doğu ve Kuzey Anadolu’da yapılan Tarih Öncesi Araştırmalar, Belleten VIII, 32, 1944, 659–680. Kushnareva 1997 K. Kh. Kushnareva, The Southern Caucasus in Prehistory. Stages of Cultural and Socioeconomic Development from the Eighth to the Second Millennium B.C. (Philadelphia 1997). Lindsay – Greene 2013 I. Lindsay – A. Greene, Sovereignty, mobility, and political cartographies in Late Bronze Age Southern Caucasia, Journal of Anthropological Archaeology 32, 2013, 691–712. Lindsay – Smith 2006 I. Lindsay – A. T. Smith, A history of archaeology in the Republic of Armenia, Journal of Field Archaeology 31, 2, 2006, 165–184. Lindsay et al. 2008 I. Lindsay – L. Minc – C. Descantes – R. J. Speakman – M. D. Glascock, Exchange patterns, boundary formation, and sociopolitical change in Late Bronze Age Southern Caucasia. Preliminary results from a pottery provenance study in northwestern Armenia, Journal of Archaeological Science 35, 2008, 1673–1682. Lindsay et al. 2010 I. Lindsay – A. T. Smith – R. Badalyan, Magnetic survey in the investigation of sociopolitical change at a Late Bronze Age fortress settlement in northwestern Armenia, Archaeological Prospection 17, 2010, 15–27.

A Survey Project on the Borderlands of Turkey – Armenia – Nakhchivan – North-western Iran

81

Lindsay et al. 2014 I. Lindsay – J. Leon – A. T. Smith – C. Wiktorowicz, Geophysical survey at Late Bronze Age fortresses. Comparing methods in the diverse geological contexts of Armenia, Antıquıty 88, 2014, 578–595. Marro et al. 2010 C. Marro – V. Bakhshalıyev – S. Sanz, Archaeological investigations on the salt mine of Duzdağı (Nakhchıvan, Azerbaïdjan) / Duzdağı Tuz Madeninde Arkeolojik Çalışmalar (Nahçivan, Azerbeycan), Türkiye Bilimler Akademisi Arkeoloji Dergisi 13, 2010, 229–244. Özfırat 2009 A. Özfırat, Pre-Classical survey in eastern Anatolia. Sixth preliminary report, Archäologische Mitteilungen aus Iran und Turan 41, 2009, 211–231. Özfırat 2011 A. Özfırat, Doğu Anadolu Yayla Kültürleri (Istanbul 2011). Özfırat 2015 A. Özfırat, Bozkurt Kurgan Mezarlığı Kazısı (2007–2013), Kazı Sonuçları Toplantısı 36, 2/2014, 2015, 209–226. Palumbi 2008 G. Palumbi, The Red and Black. Social and Cultural Interaction Between the Upper Euphrates and the Southern Caucasus Communities in the Fourth and Third Millennium B.C., Studi di Preistoria Orientale 2 (Rome 2008). Ristvet et al. 2012 L. Ristvet – V. Bakhshaliyev – H. Gopnik – S. G. Ashurov, The origins of political complexity in Naxçıvan. Excavations and survey at Oğlanqala 2008–2010, in: A. Mehnert – G. Mehnert – S. Reinhold (eds.), Austausch und Kulturkontakt im Südkaukasus und seınen angrenzenden Regıonen in der Spätbronze-/Früheisenzeit, Schriften des Zentrums für Archäologie und Kulturgeschichte des Schwarzmeerraumes 22 (Langenweißbach 2012) 281–290. Sagona 1984 A. Sagona, The Caucasion Region in The Early Bronze Age. Part 1–3, BAR International Series 214 (iii) (Oxford 1984). Sagona 2010 A. Sagona, Past and present directions in the archaeology of the Transcaucasus / Transkafkasya Arkeolojisinde Geçmiş ve Güncel Yönelimler, Türkiye Bilimler Akademisi Arkeoloji Dergisi 13, 2010, 143–157. Smith 2005 A. T. Smith, Prometheus unbound. Southern Caucasia in prehistory, Journal of World Prehistory19, 4, 2005, 229–279. Smith et al. 2009 A. T. Smith ‒ R. S. Badalyan ‒ P. Avetisyan, The Foundations of Research and Regional Survey in the Tsaghkahovit Plain, Armenia, The Archaeology and Geography of Ancient Transcaucasian Societies 1 (Chicago 2009).

Prehistory from the Ploughsoil: Interpreting Artefact Distributions from Intensive Survey in the Highlands of Samtskhe-Javakheti, Southern Georgia William Anderson * – Michelle Negus Cleary ** Abstract: Intensive survey involves systematically recording surface artefacts whose distribution can provide insights into long-term aspects of settlement and land use. Current investigations in the hills of the upper Kura (Mtkvari) River basin in Samtskhe-Javakheti province, southern Georgia, show the potential for this method to enrich knowledge on highland environments as well as on plain and coastal districts, which is where intensive survey is more commonly practised. Preliminary results from four seasons of survey in the elevated valley around Chobareti and Zveli villages show patterns in the distribution of artefacts which reflect both primary discard and re-deposition through farming and waste disposal. The distribution of Bronze–Iron Age ceramics and lithics in relation to recorded features and landscape positions points to the past spatial organisation of the valley into zones of settlement, agriculture and pastoralism. The environs of Zveli, already known through excavations as a place of Bronze Age settlement and burial, have a higher presence of prehistoric artefacts than the rest of the valley. The extension of prehistoric artefacts in fields beyond previously recorded features may be the result of later farming practices which have dispersed material from the village but could alternatively be primary discard from prehistoric activity. Our results demonstrate the importance of accurately documented surface artefact scatters for interpreting landscapes which have undergone continual alterations into modern times. Keywords: Surface artefact survey; Survey methodology; Kura River valley; Samtskhe-Javakheti

Introduction Intensive ground survey, involving the systematic recording of surface artefacts across spatially defined areas of land, is a well-established method of archaeological landscape research. Using repeatable recording procedures, this ‘off-site’ approach maps artefact scatters whose distribution offers a diachronic view of human occupation and land use. While intensive survey has been widespread in Mediterranean and Near Eastern archaeology, particularly classical archaeology, for more than thirty years,1 it has rarely been deployed in southern Caucasia. Reasons for this omission may include a reliance on site-based surveys, as practised during the Soviet era, and the perceived unsuitability of mountainous and pastoral terrain for a method that requires consistent ground visibility afforded by cultivated fields. Yet, as shown by surveys in comparably varied environments, such as the rugged coast and hinterland of northern Anatolia, this method can still yield reliable and revealing information in such contexts.2 Investigations by the Landscape Archaeology in Georgia (LAG) Project in the highlands of Samtskhe-Javakheti province, southern Georgia, demonstrate how intensive survey can offer new insights into southern Caucasia’s archaeological landscapes. Initial results from the LAG Project’s intensive survey show patterns in the distribution of defined artefact types that represent primary discard as well as secondary re-deposition. Differences in the location of Early Bronze

Landscape Archaeology in Georgia Project, [email protected]. University of Melbourne, [email protected]. 1 Cherry 1983; Alcock 2000; Given 2013. 2 Doonan 2004; Matthews – Glatz 2009; Düring – Glatz 2015, 76–83. *

**

84

William Anderson – Michelle Negus Cleary

Fig. 1   1. Upland valley south of Zveli; 2. intensive survey of field south of Chobareti; 3. horse-drawn ploughing in the Chobareti Valley; 4. manure storage on the outskirts of Zveli (photos: W. Anderson)

Age to Early Iron Age artefacts compared with overall artefact distributions suggest that surface scatters of prehistoric artefacts are the residues of past habitation and farming practices. This has important methodological and interpretive implications, proving that intensive survey may be profitably deployed in highland environments to augment and enrich information gained from excavation, feature survey and remote sensing. The LAG Project’s Intensive Survey Since 2013, the LAG Project has investigated the high country between the towns of Akhaltsikhe and Aspindza.3 The study area spans the southern side of the Kura River valley, extending from the river plain southwards to the elevated valleys, and beyond to the highland pasture (yayla) of the Erusheti mountain range. A variety of methods and technologies are deployed to document this environmentally diverse district’s multi-period remains – among them Bronze Age monuments, Iron Age fortifications, medieval settlements and Soviet installations — and to address questions regarding settlement patterns, routes of movement and human interactions with the landscape. Intensive survey is a key part of the LAG Project’s methodology, focusing on the upland valley at Chobareti and Zveli villages (Fig. 1), at elevations of approximately 1450–1600m above sea level (asl). The valley sides have been extensively terraced into small, tiered fields where cereal and root crops, especially potato, are planted, tended and harvested, mainly using non-mecha-

3

Anderson et al. 2014.

Prehistory from the Ploughsoil

85

nised methods (Figs. 1.1–3). As the intensive survey relies on the ground visibility afforded by ploughed and cultivated land, it is confined to this zone and does not account for the surrounding forested slopes and grassy plains. However, its application within the particular ecology of the upland valley gives a consistent view of a micro-region where there is evidence for more than six millennia of human occupation. The intensive survey method involves selecting fields with sufficient ground visibility and with the aim of achieving coverage of the different topographic positions within the valley. Having chosen a field to survey, this is designated as a survey unit (SU); its location, ground conditions and geographic context are noted and its boundaries are recorded using a GPS and later drawn as polygons and added to the GIS. Field workers spaced at 10m intervals walk in straight lines, usually along ploughed furrows, recording and collecting all artefacts. Collected artefacts are washed, counted, photographed and analysed; unsystematic finds of diagnostic sherds and lithics are counted separately. The purpose of the intensive survey is to assess the distribution and density of surface artefact scatters and to gather a sample of artefacts which encompasses the range of material present. As the project progressed, specific and thematic questions emerged. First, at a basic level, we sought to find out whether intensive field-walking worked in this environment. Was the amount of material sufficient to provide meaningful quantitative information? Were there patterns of distribution shown by areas of artefact concentration and absence? Second, what does the overall distribution of artefacts indicate? Is it possible to distinguish between ‘primary discard’ and the consequences of intermediate land use, especially re-deposition through manuring practices? Third, can temporal patterns be distinguished from the distribution of identified artefacts, and how might these indicate defined practices of past settlement, burial, movement and dwelling in the landscape? The remainder of this article addresses these questions. Surface Artefact Scatters – Primary or Secondary Deposits? The LAG Project’s intensive survey has achieved coverage of fields across the upland valley at Chobareti and Zveli and in selected parts of the wider district. As of 2016, a total of 155 SUs, comprising a surface area of 51.3 hectares, have been intensively surveyed. This has resulted in the recording and collection of more than 2500 artefacts. The generally small and consistent size of SUs, which have an average surface area of 3312m² and a median area of 2876m², allows for artefact distributions to be mapped on a field-by-field basis (Fig. 2). Artefacts were recorded in all but 14 of the 155 SUs. Total systematic artefact counts range from 1–82. A basic calculation of density is reached by dividing the number of systematically gathered artefacts by the surface area (in hectares) of the SU. Average concentrations range from 2–590 artefacts per hectare (mean average: 60 per hectare), without weighting to account for the coverage of field-walking or levels of ground visibility. While this figure is low compared with the results of certain survey projects in Mediterranean and Near East regions, which have even reported ‘an almost unbroken carpet’ of pottery scatters,4 it does show a steady but variable presence of artefacts. Initial mapping of artefact densities shows distinctive spatial patterns (Fig. 2). Most obvious is the concentration of material in the immediate surroundings of Chobareti and Zveli villages: the ten SUs with the highest density are all within 300m of these inhabited settlements and the top three are adjacent fields to the direct east of Zveli. Also notable is the generally higher quantity of material on the north side of the valley compared to the south side. There are relatively sparse artefacts on the north-facing slopes, except very close to Chobareti village, while on the southfacing slopes moderate density artefact scatters continue some distance from any currently inhab-

4

Bintliff – Snodgrass 1988, 506.

86

William Anderson – Michelle Negus Cleary

Fig. 2   Map of intensive survey fields and showing quantities in Zveli-Chobareti district (map: M. Negus Cleary)

ited area. These general observations are important to consider when focusing on the distribution of identified artefact types. The majority of collected artefacts date from the medieval and post-medieval periods. In particular, ceramics dating from the 12th–13th centuries and onwards are strongly represented while earlier medieval, classical and prehistoric material is less common. The artefact scatters can be considered the result of what might be termed primary and secondary deposition. As noted, there is a relationship between artefact density and proximity to medieval and modern settlements, including both settlements that are inhabited into the present day and deserted settlements which have been unoccupied since the mid-twentieth century or earlier. This largely explains the prevalence of medieval, post-medieval and modern objects in the ploughsoil. Variation in the density of artefact scatters seems to represent both historical discard and intermediate processes. The manuring of fields, in which household waste is mixed with the fertiliser, may be the cause of material concentrating beside historical and contemporary settlements. The patterning of artefact scatters in the surroundings of settlements has been documented by survey projects of classical cities and their hinterlands. In Boeotia (central Greece), manuring in the classical period is seen to create a ‘halo effect’ of artefact scatters that occur as bands in fields that would have been cultivated from the nearby settlement.5 This appears to be happening in the surroundings of Chobareti and Zveli, where manure accumulated in villages is spread on fields, a practice that continues to this day (Fig. 1.4). Artefact concentrations near to current and abandoned settlements seem to result from farming practices in the recent and more distant past. In the close vicinity of abandoned settlements such as Chobareti Verana and Q’uri, they might also represent deposits from habitation and industry. This is especially likely in fields near to deserted settlements but distant from current

5

Bintliff et al. 2007; see also Alcock et al. 1994.

Prehistory from the Ploughsoil

87

population centres, for example, at Q’uri, where two fields beside the settlement yielded concentrations of late- and post-medieval ceramics.6 That these scatters are the locally re-deposited debris of past occupation is supported by the presence of recurring ceramic fabrics and shapes as well as wasters. While the clustering of artefact scatters near occupied and deserted settlements may be largely due to waste disposal and farming practices, there are distinctive trends in the distribution of certain defined artefact types. The relatively low numbers of Bronze–Iron Age artefacts make this material conspicuous alongside the mainly medieval and post-medieval ceramics. There is also a steady quantity of lithic artefacts, including formal tools. Focusing on these recognisable artefacts allows us to hear beyond the ‘background noise’ and the complex distribution of later historical material. The identified patterns of artefact discard both support and supplement information on larger-scale Bronze to Iron Age settlement remains, monuments and earthworks. Bronze and Iron Age Artefacts: Forms and Distribution Pottery Broad chronological categories based on local and regionally attested wares were used to classify pottery identified as ‘prehistoric’. The identifications were supported through comparison with material from the excavations at Chobareti, on the upper slopes of the peak known as Satikne that forms the northern crest of the valley.7 Identifications were also informed by finds from other sites surveyed by the LAG Project, particularly the multi-period (Chalcolithic – Iron Age and medieval) complex at Varneti,8 as well as by published information from other excavated sites in the immediate and wider region. Pottery assigned to the Early Bronze Age (EBA), Middle Bronze Age (MBA), Late Bronze Age (LBA) and Early Iron Age (EIA) was found in 34 (22%) of the SUs (Fig. 3, Fig. 5 and Table 1). Although the designation of ‘prehistoric pottery’ is broad, and intended simply to capture all material dating from before approximately 600 BC, it allows for assessment of a recognisable and relatively rare body of artefacts. Fewer than one hundred prehistoric sherds were identified in total, constituting less than 4% of systematically recorded artefacts from intensive survey. This arguably makes their spatial distribution all the more significant for interpreting what they might represent in terms of prehistoric settlement and land use. The attributes of EBA ceramics from the region and the immediate locality are well documented. Excavated sites dating from the fourth and third millennia BC include Amiranis Gora, near Akhaltsikhe,9 and Chobareti itself, the latter excavations having identified buildings, pits and burials dated to the last three centuries of the fourth millennium.10 The predominant EBA wares from Chobareti, including pale brown Kura-Araxes pottery and orange ware that is classed as a late Sioni (early-4th millennium BC) type, are represented among the survey finds (Fig. 3.1), with a lesser presence of chaff-faced ware; there are also sherds of finer Kura-Araxes pottery in mineral tempered and well-levigated fabric with red and black polished surfaces (Fig. 3.2).11 MBA material has not been reported from the Chobareti hilltop excavations, although from the lower parts of the valley, investigations of barrows in the surroundings of Zveli identified features

6 7

8 9 10 11

Negus Cleary et al. 2018. We are especially grateful to Claudia Sagona for sharing her knowledge and ideas with us about ceramics from the Chobareti excavations. Anderson et al. in press. Kushnareva 1997, 55 for summary and references. Kakhiani et al. 2013; Messager et al. 2015. Kakhiani et al. 2013, 27–29.

88

William Anderson – Michelle Negus Cleary

Fig. 3   Prehistoric pottery – see Tab. 1 for artefact details (photos: W. Anderson)

dated to the early- to mid-second millennium.12 Within Zveli, MBA artefacts have been recorded at the so-called Rabat, the mounded settlement site overlooking the Kura River valley (discussed further below).13 A new source of comparison comes from the LAG Project’s surveys at Varneti, where cultural deposits exposed in a landslip (‘Varneti Profile’) have been radiocarbon dated, on the basis of charcoal and bone samples, to the first half of the second millennium.14 Ceramics from intensive survey that resemble examples from the Varneti Profile have burnished, reddishbrown-coloured exterior surfaces, often with dark grey or black zones, in semi-fine, sandy fabric (Fig. 3.3–4). Also found at Varneti is an assortment of LBA–EIA ceramics. This eclectic material is identified on the basis of comparison with the many assemblages of these periods that have been excavated in southern Caucasia and eastern Anatolia. The variation in styles and wares of mid-second to mid-first millennium ceramics makes concise characterisation difficult. However, notable LBA wares found during the intensive survey, and also represented at Varneti, include examples in

12 13 14

Kakhiani et al. 2013, 5. Anderson et al. 2014, 19–20. Anderson et al. in press.

89

Prehistory from the Ploughsoil

Fig. 4   Lithic tools and debitage – see Tab. 1 for artefact details (photos: W. Anderson)

Figure number

Artefact number

Estimated age

Fig. 3.1

F14008.04

EBA

Fig. 3.2

F14050.16

EBA

Fig. 3.3

F14014.03

MBA

Fig. 3.4

F14008.02

MBA–LBA

Fig. 3.5

F15055.15

LBA–EIA

Fig. 3.6

F13014.01

EIA

Fig. 4.1

F14035

EBA

Fig. 4.2

F15049

EBA

Fig. 4.3

F15098

EBA–EIA

Fig. 4.4

F14035

EBA

Fig. 4.5

F14004

Post-medieval

Fig. 4.6

F14050

Unknown

Table 1   Register of illustrated artefacts

90

William Anderson – Michelle Negus Cleary

a light red and porous fabric which are coated with a pinkish-cream-coloured wash (Fig. 3.5); LBA–EIA ceramics include dense black wares and examples with incised decoration in a coarse, gritty fabric (Fig. 3.6).15 Lithics More than 100 lithic artefacts have been recorded during the intensive survey, constituting some 4% of all collections. We have divided these artefacts into ‘tools’ and ‘debitage’ to help distinguish those of likely prehistoric date. Tools are characterised by modification through retouch, use-wear and, less commonly, grinding and pecking. Among the lithic tools, most are scrapers and cutting implements that are fairly amorphous, although their shape and signs of working and usage indicate specific functions. There are also a number of finely worked and distinctive tools (Fig. 4.1–3). Of particular note is a tanged projectile point in transparent obsidian (Fig. 4.2) which can be assigned to the EBA on the basis of its form.16 Comparable points have been found in burial contexts, and an example from the Chobareti excavations, found in a pit along with intact vessels, animal bones and ash, may be a ‘deliberately structured’ deposit.17 Lithic tools were found in 17 (11%) of the SUs; notably, prehistoric pottery is also present in three-quarters of the SUs with lithics. All but a handful of lithic tools are obsidian; smaller quantities of flint and other materials, including basalt, are present. Although obsidian does not occur naturally within the immediate locality, the stone does appear to have been widely available. There is variety in its appearance, which ranges from grey- and brown-tinged, smoky and transparent to opaque and black. The small number of flint tools are made from fine-grained stone which ranges from red and maroon to dark-yellow-coloured and whose forms include blades and thumbnail scrapers (Fig. 4.4–5). Fine-grained and dark-grey-coloured andesite has also been fashioned into tools, most of which are identifiable as threshing stones: five examples have been found during intensive survey and several others have been located during unsystematic reconnaissance. These andesite flakes have been worked into elongated blades which would have been mounted onto wooden sledges and are conspicuous by their worn, rounded edges as the result of usage (Fig. 4.6).18 Traditional methods of hand threshing are known to have been used in the district until recent times. It is likely that most or all collected examples are post-medieval, although there is potential for some to be earlier. For the purpose of the present article, threshing stones are treated separately from the other lithic tools. Also of andesite, or fine-grained basalt, is a fistsized, sub-rounded implement with ground edges. This seems less likely to be of modern date, which is supported by the presence of comparable examples at Varneti, among a scatter of Chalcolithic–EBA artefacts. Waste flakes or debitage are present in 37 (24%) of the SUs. Of the SUs with debitage, lithic tools are also present in twelve and prehistoric ceramics are present in nine. The prevalence of obsidian debitage is noteworthy, but the status of this material – in terms of its origin, date and, above all, its archaeological meaning – is hard to determine (Fig. 4.7). Obsidian was certainly used into historical times, during and after the medieval period; in the Samtskhe-Javakheti region, obsidian flakes have been found in excavated, late medieval contexts, for example at Tiseli.19 The tendency for the stone to easily fracture and shatter might also partly explain the

15 16 17 18 19

Sagona 2012. Smith 2015, 142–143. Kakhiani et al. 2013, 12. Whallon 1978. Mindorashvili 2011.

Prehistory from the Ploughsoil

Fig. 5   Map of prehistoric pottery and lithic distributions (map: M. Negus Cleary)

Fig. 6   Map of prehistoric pottery and lithic distributions in the environs of Zveli (map: M. Negus Cleary)

91

92

William Anderson – Michelle Negus Cleary

presence of small fragments, which might have been reduced from once larger objects through repeated ploughing. Little reliable information can be gleaned from the widely dispersed obsidian waste flakes, but their distribution is notable in relation to lithic tools. Distribution The spatial distribution of prehistoric ceramics and lithics reveals distinct clusters of material, in particular surrounding Zveli village (Fig. 5). There are four zones of concentration: west of Zveli, on a slope leading up to a hill along the ridge overlooking the Kura River valley; south of Zveli, in fields at a lower elevation than the village; east of Zveli, along the ridge before it falls in elevation towards Orgora village; and in the centre of Zveli, close to the Rabat. Puzzlingly, prehistoric material is almost non-existent in fields below the Chobareti excavations. The largest and densest cluster of prehistoric artefacts occurs in fields to the west of Zveli (Fig. 6). These are located on a moderate slope that lies adjacent to an area of burial mounds (kurgans) and linear features, some of which were excavated during the 1970s as part of the MeskhetJavakheti expedition, led by Otar Gambashidze.20 The kurgans are located on the edge of the plateau where Zveli village is located, at 1510m asl. They are enclosed within several long, curving embankments that follow the land contours and from which there are excellent views to the west and south-west along Chobareti Valley. Almost all the SUs just to the north and west of the kurgan field contain prehistoric pottery and lithic artefacts. This concentration suggests primary discard associated with the kurgans rather than re-deposition; these fields are not those with the highest overall artefact densities, which are on the eastern side of Zveli. The concentration of prehistoric artefacts west of Zveli, together with the prominent shape and position of the landform, suggest that the original extent of the kurgan field extended further north into the area where there are now terraced fields. This shows that the many and intrusive changes that have occurred in the intervening period, including construction of terracing and more recent forest plantation, building of roads, and the gas pipeline which runs through this area, may have displaced built features but that the footprint of Bronze Age land use endures in the form of artefact scatters. Three other areas close to Zveli village have concentrations of prehistoric artefacts, which also appear to be representative of past discard. To the south of Zveli, in low-lying fields (around 1480m asl) where there is a moderate overall concentration of artefacts, half of the SUs contain prehistoric pottery and a third contain lithics; notably, fields in an equivalent topographic position to the south of Chobareti have yielded barely any prehistoric artefacts. On the immediate southern side of the village, cultivated land beside an open grassy field now used as a recreation ground has a particular density of EBA ceramics. On the western side of Zveli Rabat, the mounded settlement where the ground has been modified and built up through multiple occupation phases, fields have yielded EBA–EIA ceramics and lithics; on the mound itself, the LAG Project has recorded MBA pottery eroding from the surface, and excavations carried out in the 1950s and 1970s identified EBA and MBA deposits.21 The third area of concentration is on level ground close to the cliff edge to the east of Zveli, where there is a high count of lithic tools and debitage. The distribution of prehistoric artefacts can be argued to broadly represent primary deposition rather than later farming practices. This is suggested by two aspects of the data. First, the correspondence between the presence of prehistoric pottery and lithic tools. Of the 32 SUs with prehistoric pottery, 19 (59%) also have lithics: there are only seven SUs with lithics but no prehistoric pottery. Second, the presence of prehistoric artefacts does not necessarily relate with overall arte-

20 21

Kakhiani et al. 2013, 5. Djaparidze 1992, 195–197; Kakhiani et al. 2013, 5; Anderson et al. 2014, 19–20. The GAIA Project has recently commenced a new series of excavations at the Rabat.

Prehistory from the Ploughsoil

93

fact density. Of the SUs with the densest overall concentrations (the 28 SUs that have an average of more than 100 artefacts/hectare), prehistoric artefacts are found in less than half. This shows that there is not a simple correlation between high artefact density and the presence of prehistoric artefacts. On this basis, the distribution of prehistoric pottery and lithics appears to broadly reflect areas of past primary discard rather than later re-deposition. Discussion Intensive survey offers both opportunities and challenges of interpretation. The lack of contextual precision for artefacts recorded on the surface, the difficulty of distinguishing between ‘primary discard’ and subsequent re-deposition and problems of accurate dating make for coarse-grained results. As John Cherry has observed ‘excavation reveals a lot about a little of one site, survey can tell us a little about lots of sites’.22 Yet, it is the lack of focus on ‘sites’ that the systematic method encourages which is one of its strengths, demanding consideration of the landscape as a whole at a scale that excavations could not achieve. Results from the intensive survey at Chobareti and Zveli add to our knowledge on the longterm human occupation of the upper Kura River basin. They supplement and add new perspectives to other forms of information gained from feature recording, excavation and landscape research. Variation in the density of artefact scatters in different parts of the valley shows that there is sufficient material to give a reliable indication of distribution. The general concentration of artefacts beside current and former settlements seems to result from manuring, where household waste is mixed with animal dung that is spread on fields as a fertiliser. However, the distribution of Bronze–Iron Age pottery and lithics differs from the overall artefact densities, indicating distinctive discard patterns. Even through thousands of years of land modification and farming, artefact scatters retain spatial patterns that are suggestive of prehistoric deposition. What exactly these distributions represent in terms of past action is a matter of debate. The proximity of most material to Zveli village is perhaps unsurprising, given the presence of kurgans within and surrounding the current village, and Zveli Rabat, the tell-like settlement mound in the north of the village that overlooks the Kura River valley where occupation dates back to the EBA–MBA. The material might, then, have reached surrounding fields through modern dispersal of soil taken from within the village. Equally, artefact dispersal might have been the result of ancient rather than recent farming practices. Archaeobotanical analysis of organic material from excavated storage pits at Chobareti detected the strong presence of cereals, especially wheat, underlining the importance of cereal growing in Kura-Araxes agriculture.23 The analysis of seed isotopic values produced results that suggest the use of manuring by spreading animal dung on fields.24 Further afield, studies of agricultural terracing in the north Caucasus region have identified significant quantities of pottery buried within soil accumulated to form terraces as being the result of longterm manuring practices.25 This prospect complicates the equation of EBA artefacts with places of settlement or burial as the material may have been spread out from occupation areas through manuring. The idea that prehistoric artefact distributions result from agriculture that is roughly contemporary with the artefacts appears likely, and may allow us to go further in explaining the presence and absence of material in certain parts of the wider valley. The distribution patterns point to a landscape that was organised according to certain habitation and agrarian practices.

22 23 24 25

Cherry 1983, 387. Messager et al. 2015. Messager et al. 2015, 223. Korobov – Borisov 2013, 1089.

94

William Anderson – Michelle Negus Cleary

For example, the lack of material immediately below the hilltop settlement at Chobareti suggests that this land was not subject to cultivation during the settlement’s EBA occupancy, even though its proximity to settlement and south-facing aspect made it suitable for agriculture. As noted, prehistoric material is prevalent in the north-east of the valley and rare in the south. The presence of medieval artefacts but absence of prehistoric material in the south suggests that it was not until this later period that these slopes were terraced and farmed. However, Bronze Age features are prominent further uphill to the south, on the upper slopes that lead to the yayla. This may indicate these slopes as being a pastoral rather than arable farming zone, and indeed in the present day it continues to be the route by which herds are moved between villages in the valley and the upland pasture. Therefore, in the Bronze Age, the valley may have been spatially separated into settlement zones along the ridge to the north, cultivated fields on adjacent flat ground and gentle slopes, and pastoral herding on the southern half of the valley leading towards the yayla. On the basis of this preliminary analysis, we can draw conclusions both particular to the locality of the study area and with more general implications for Caucasian highland landscapes. The concentration of Bronze–Iron Age artefacts around Zveli underlines its long-term significance, and the spread of material may show that occupation here was more extensive than previously understood. These artefact scatters may be the remnants of areas where built features have since been eradicated by subsequent alterations to the land, particularly in the kurgan area on the western edge of Zveli where piles of stone and large boulders have been pushed to the margins of fields and tracks. On the other hand, if artefact scatters are taken as proxy indicators of agriculture through past manuring practices, the interpretation may be taken further to indicate the zoning of the landscape into settlement, mortuary, agricultural and pastoral quarters. Integrating the evidence of artefact scatters with the results of feature recording, environmental and landscape studies opens up possibilities for identifying diachronic land use patterns – showing how these changed and remained constant through time. Many factors have contributed to the transport of artefacts over the long term, but on the basis of the distributions presented here, we hypothesise that clusters of prehistoric artefacts are largely the result of deposition from past settlement, with manuring being a secondary factor. Although further information – especially through excavation at different points in the valley, and not only at settlement and burial sites – would strengthen this interpretation, we propose that on the basis of the intensive survey results, the clustering of material in certain parts of the valley indicates land where past occupation was concentrated. Aside from the complex issues raised by interpreting ploughsoil assemblages and the extent to which firm conclusions may be drawn from these assemblages, we can state with confidence that the intensive survey method does work in rugged highland terrain where there are areas of cultivated land. Furthermore, our project demonstrates that systematically gathered data which provide new insights into landscape history can be gained by a project team working within a limited timeframe and a small budget. There is little doubt that the further use of intensive survey will bring new perspectives on the prehistoric and historic landscapes of southern Caucasia. Acknowledgments: We thank our colleagues on the LAG Project team – Jessie Birkett-Rees (Monash University), Damjan Krsmanovic (University of Leicester) and Nikoloz Tskvitinidze (Ilia State University) – who were involved in all aspects of the intensive surveys. In 2015, our team was joined by Gia Chilingarashvili (Tbilisi State University), Natalie Langowski (University of Melbourne) and Madona Mshvildadze (Ilia State University) who assisted with the fieldwork and finds processing. The LAG Project operates alongside the Georgian Australian Investigations in Archaeology (GAIA), a research collaboration between the University of Melbourne and the Georgian National Museum. We are very grateful to GAIA’s co-directors, Antonio Sagona and Kakha Kakhiani, for their permission and support for our project, and to Claudia Sagona and Giorgi Bedianashvili. Funding for this research was supported by a Humanities Travelling Fellowship (HTF201600027) from the Australian Academy of the Humanities, the Dorothy Cameron Fellowship from the University of Sydney and the 2015 Australasian Women in Ancient World Studies Research Grant.

Prehistory from the Ploughsoil

95

References Alcock 2000 S. E. Alcock, Extracting meaning from ploughsoil assemblages. Assessments of the past, strategies for the future, in: R. Francovich – H. Patterson – G. Barker (eds.), Extracting Meaning from Ploughsoil Assemblages (Oxford 2000) 1–4. Alcock et al. 1994 S. E. Alcock – J. F. Cherry – J. L. Davis, Intensive survey, agricultural practice and the classical landscape of Greece, in: I. Morris (ed.), Classical Greece. Ancient Histories and Modern Archaeologies (Cambridge 1994) 137–70. Anderson et al. 2014 W. Anderson – J. Birkett-Rees – M. Negus Cleary – D. Krsmanovic – N. Tskvitinidze, Archaeological survey in the south Caucasus ( Samtskhe-Javakheti, Georgia). Approaches, methods and first results, Anatolia Antiqua 22, 2014, 11–33. Anderson et al. in press W. Anderson – M. Negus Cleary – J. Birkett-Rees – D. Krsmanovic – N. Tskvitinidze, Gateway to the yayla. Varneti Archaeological Complex in the Southern Caucasus Highlands (in press). Bintliff – Snodgrass 1988 J. L. Bintliff – A. Snodgrass, Off-site pottery distributions. A regional and interregional perspective, Current Anthropology 29, 1988, 506–513. Bintliff et al. 2007 J. L. Bintliff – P. Howard – A. Snodgrass, Testing the Hinterland. The Work of the Boeotia Survey (1989–1991) in the Southern Approaches to the City of Thespiai (Oxford 2007). Cherry 1983 J. F. Cherry, Frogs around the pond. Perspectives on current archaeological survey projects in the Mediterranean region, in: D. R. Keller – D. W. Rupp (eds.), Archaeological Survey in the Mediterranean Area (Oxford 1983) 375–416. Djaparidze 1992 ო. ჯაფარიძე / O. Djaparidze (ed.), საქართველოს არქეოლოგია ტ. II. ენეოლით-ადრე ბრინჯაოს ხანა / Sarkadvelos Arkelogia II. Eneolit–Adre Brinjoas Khana (Georgian Archaeology II. From the Eneolithic to the Early Bronze Age) (Tbilisi 1992). Doonan 2004 O. P. Doonan, Sinop Landscapes. Exploring Connection in a Black Sea Hinterland (Philadelphia 2004). Düring – Glatz 2015 B. S. Düring – C. Glatz, Kinetic Landscapes. The Cide Archaeological Project. Surveying the Turkish Western Black Sea Region (Warsaw, Berlin 2015). Given 2013 M. Given, Commotion, collaboration, conviviality. Mediterranean survey and the interpretation of landscape, Journal of Mediterranean Archaeology 26, 1, 2013, 3–26. Kakhiani et al. 2013 K. Kakhiani – A. Sagona – C. Sagona – E. Kvavadze – G. Bedianashvili – E. Massager – L. Martin – E. Herrscher – I. Martkoplishvili – J. Birkett-Rees – C. Longford, Archaeological investigations at Chobareti in southern Georgia, the Caucasus, Ancient Near Eastern Studies 50, 2013, 1–138. Korobov – Borisov 2013 D. S. Korobov – A. V. Borisov, The origins of terraced field agriculture in the Caucasus. New discoveries in the Kislovodsk basin, Antiquity 87, 2013, 1086–1103. Kushnareva 1997 K. Kh. Kushnareva, The Southern Caucasus in Prehistory. Stages of Cultural and Socioeconomic Development from the Eighth to the Second Millennium B.C. (Philadelphia 1997). Matthews – Glatz 2009 R. Matthews – C. Glatz (eds.), At Empires’ Edge. Project Paphlagonia. Regional Survey in North-Central Turkey (London 2009).

96

William Anderson – Michelle Negus Cleary

Messager et al. 2015 E. Messager – E. Herrscher – L. Martin – E. Kvavadze – I. Martkoplishvili – C. Delhon – K. Kakhiani – G. Bedianashvili – A. Sagona – L. Bitadze – M. Poulmarc’h – A. Guy – D. Lordkipanidze, Archaeobotanical and isotopic evidence of Early Bronze Age farming activities and diet in the mountainous environment of the South Caucasus. A pilot study of Chobareti site ( Samtskhe-Javakheti region), Journal of Archaeological Science 53, 2015, 214–226. Mindorashvili 2011 D. Mindorashvili, Tiseli settlement, in: P. M. Taylor (ed.), Pipelines to Cultural Heritage. Proceedings of an International Workshop (April 20–23, 2010, Baku, Azerbaijan) (Washington DC 2011) 502–525. Negus Cleary et al. 2018 M. Negus Cleary – N. Tskvitinidze – W. Anderson – D. Krsmanovic – J. Birkett-Rees, Mapping the Vilayet of Gurjistan. The southern Caucasus highlands in the late medieval – early Ottoman periods, in: A. Batmaz – G. Bedianashvili – A. Michalewicz – A. Robinson (eds.), Context and Connection. Essays on the Archaeology of the Ancient Near East in Honour of Antonio Sagona (Leuven 2018) 383–407. Sagona 2012 A. Sagona, Remarks on the east Anatolian Iron Age, in: A. Çilingiroğlu – A. Sagona (eds.), Anatolian Iron Ages 7. The Proceedings of the Seventh Anatolian Iron Ages Colloquium held at Edirne, 19–24 April 2010 (Leuven 2012) 253–267. Smith 2015 A. T. Smith, The Political Machine. Assembling Sovereignty in the Bronze Age Caucasus (Princeton, Oxford 2015). Whallon 1978 R.-J. Whallon, Threshing sledge flints. A distinctive pattern of wear, Paléorient 4, 1978, 319–325.

Phasis and its Landscape: Preliminary Report of the Archaeological Survey of the Lower Stream of the Rioni River Delta Davit Naskidashvili * Abstract: This paper reviews the historical and archaeological evidence for the location of ancient Phasis and outlines current landscape investigations being undertaken by the author in the Colchean Lowlands. These test the theory that, at least in its first stages, Phasis was a dispersed settlement which was adapted to the local landscape and has been mistakenly perceived as a classical urban city. Keywords: Greek colonisation; Phasis; Archaeological survey; Rioni Delta; Colcheans; Colchean culture; Rioni River; Eastern coastline of the Black Sea

Introduction The ancient city of Phasis (Phazisi) was founded after the arrival of Greeks in the Colchean Lowlands1. There is ongoing debate, however, about the exact date. Generally, scholars consider the city to have been founded in c. 7th/6th century BC.2 Pomponius Mela, a Roman author of the 1st century AD, provides us with specific information about the city. He states that, ‘Here are the Colcheans. The Phasis [Rioni] bursts into the sea here; here is the town colonised by Themistagoras the Milesian; here are the grove and temple of Phrixus, who is well known from the old legend of the Golden Fleece.’3 According to historical sources, the ancient city of Phazisi has a long history of occupation, approximately one millennium long, from the 6th century BC to the 6th century AD.4 The history of Phazisi spans several chronological stages: the Classical period, Hellenistic period, Roman period and Late Antique/Early Medieval period. Current research, however, has been focused more on determining the location of the city rather than on the periodisation or chronology. The Colchean Lowlands and the Location of Phazisi So far, there are 12 different proposed locations for the city of Phazisi. Most of these are based on interpretations of the historical sources.5 However, there have also been a number of different archaeological field campaigns led by Gela Gamkrelidze, Teimuraz Mikleadze and Guram Grigolia that have resulted in further possible locations. Fig. 1 illustrates the various possible locations that have been proposed for Phazisi as identified in Gela Gamkrelidze’s publication.6 All of the



*

1 2 3 4 5 6

Ivane Javakhishvili Tbilisi State University, [email protected], [email protected]. See Lordkipanidze 2000 for more information on the archaeology of the Colchean culture. Berdznishvili 1969, 34. Romer 1998, 65. See Gamkrelidze 2012, 38 for an extensive list of primary sources that mention Phazisi. For more information see Gamkrelidze 2012. Gamkrelidze 2012.

98

Davit Naskidashvili

Fig. 1 Proposed locations of Phazisi based on information from Gamkrelidze 2012. 1. F. Dubois de Montpéreux; 2. F. Brun; 3. N. Shafranov; 4. L. Elnitski; 5. M. Berdznishvili; 6. B. Kuftin; 7. N. Khoshtaria; 8. N. Lomouri; 9. G. Grigolia; 10. Dj. Djanelidze; 11. Ot. Lordkipanidze, T. Mikeladze; 12. G. Gamkrelidze (satellite image: © 2016 Google Earth; Cnes/Spot; DigitalGlobe; Data SIO, NOAA, U.S. Navy, NGA, GEBCO; TerraMetrics)

proposed locations are concentrated in the Colchean Lowlands, into which the region’s main river, the Rioni, flows from the north Caucasian mountains. The Rioni divides the lowlands into two parts and debouches into the Black Sea (Fig. 1).7 The Colchean Lowlands ecosystem is mostly comprised of marshland which is connected to the Rioni Delta. Therefore, the Rioni played an enormous role in the development of the local environment. Changes in the Black Sea level should also be taken into consideration, having had an important effect on the environment of the region. The Colchean Lowlands were flooded frequently which made them a risk zone for urban settlement.8 Given that, according to the historical sources, Phazisi was in existence for approximately a millennium, it seems that there is a high possibility that it changed location in different periods, possibly influenced by environmental conditions. A similar suggestion was put forward by Boris Kuftin.9 Furthermore, according to Gela Gamkrelidze, 3rd–7th century AD Phazisi should be identified with the ‘Natekhebi’ settlement, which is located on the western shore of Lake Paliastomi, though he supposes that through time the city changed location due to geomorphological changes in the landscape (Fig. 1).10 Regional Settlement Patterns in the 1st Millennium BC Before Greek colonisation in the Colchean Lowlands there were two types of settlement in this region, the first on dunes and the second on artificial mounds. Dune settlements are located along the eastern coastline of the Black Sea (Figs. 2 and 3). There is some debate as to when these natural dunes originally formed;11 however, we know that the

7 8 9 10 11

Gamkrelidze 2012, 56–58. Ewoldsen 2014, 3. Kuftin 1949/1950, 110–111. Gamkrelidze 2012, 58. Tavamaishvili 2012, 86–87.

Phasis and its Landscape

99

Fig. 2   Map of the Colchean Lowlands. Sites and palaeochannels mentioned in the text are indicated on the map (graphics: D. Naskidashvili)

Fig. 3   Dunes separating marshland from the Black Sea (photo: D. Naskidashvili)

100

Davit Naskidashvili

Fig. 4   Sachochuo mounds (satellite image: © 2018 Google Earth; Digital Globe)

settlements in this zone are mostly datable to the first half of the first millennium BC.12 The closest dune settlement to the Rioni’s estuary is Kulevi, which is 6km to the north (Fig. 2). By contrast, the Colchean Lowlands, inland from the area of dune settlement, are covered by mound settlements (Fig. 4). Most of them have satellite mounds and also contain occupation layers from the Early Bronze Age, indicating that they were inhabited over a long period of time13 (see Fig. 4 for an example of this type of settlement). It appears clear that the Colchean culture had specific settlement forms that were adapted to the local environment and landscape, that is, marshy conditions and frequent flooding. There is a possibility that Greek colonists in the Colchean Lowlands adopted the same traditions in the first stages of colonisation (6th–5th centuries BC). Supporting this argument is Hippocrates’ description of Phasis: ‘These are my opinions about the Longheads. Now let me turn to the dwellers on the Phasis. Their land is marshy, hot, wet, and wooded; copious violent rains fall there during every season. The inhabitants live in the marshes, and their dwellings are of woodland reeds, built in the water’.14 Nowadays, it is still popular to use old mounds for settlements (Fig. 5). The quote from Hippocrates, writing in the 5th century BC, confirms that the landscape was more or less identical to today in terms of the marshy conditions. Furthermore, he goes on to say that the inhabitants of Phasis ‘make little use of walking in the city and the harbor, but sail up and down in dug-outs made from a single log, for canals are numerous.’15 This information can be used to suggest that the ancient city Phazisi had at least two parts: a harbour, as well as a city centre.

12 13 14 15

Tavamaishvili 2012, 86–87. Jibladze 2007 for information on Colchean culture. Johns 1957, 112. Johns 1957, 112.

Phasis and its Landscape

101

Fig. 5   Gela Qorqia’s house is built on top of an ancient mound, which is 50 metres away from the Rioni River. According to the owner, the last time the area flooded was in 1987 when the entire nearby village of Chaladidi was flooded, but his house survived and stayed dry because of its elevated location (photo: D. Naskidashvili)

Evidence for settlement in the Colchean Lowlands from the same era as the first stage of Phazisi (6th–5th centuries BC) can be found at the settlement of Simagre. A mounded site, Simagre is located on the left bank of the current channel of the Rioni (Fig. 2). There are two dated occupation phases: 1. High Medieval period, 2. 6th–5th century BC; these layers are divided by a sterile horizon16. Morphologically, the Simagre mounds have the same form as the rest of the Colchean mounds. In the future, we suggest that investigations should focus on Simagre as it may be part of the ‘Country of Phazisi’. Current Research On review of the evidence presented above, I propose that in its first stages Phazisi was a dispersed settlement, which was adapted to the local landscape and has been mistakenly perceived as a classical urban city. In order to test this hypothesis, the current research project aims to record all of the archaeological features concentrated in the Rioni Delta region in order to better understand the archaeological landscape. So far, archaeological survey has been undertaken on the right bank of the river Rioni. Two methodologies have been used. First, we embarked on transect survey in cornfields within the survey area; this was done in the spring so as to take advantage of the good visibility provided by the freshly

16

Mikeladze 1978, 46–47.

102

Davit Naskidashvili

Fig. 6   The survey area of the current project. The agricultural fields pictured here were surveyed and small ceramic sherds were located mixed with river sediments. The survey was undertaken in the spring season when local farmers were ploughing the fields. Some of these farmers were interviewed to see if they had knowledge about ancient remains. Transect spacing inside the yellow rectangle was between 5 and 20m. Because of the minimal amount of material located and time constraints, the transect spacing was increased to between 20 and 40m in the area indicated by the red rectangle (satellite image: © 2018 Google Earth, Digital Globe)

Fig. 7   Transect survey being carried out by Shio Simsive, Ana Gabunia and Giorgi Uguzashvili (photo: D. Naskidashvili)

ploughed fields (Figs. 6 and 7). Figure 6 illustrates the coverage of the survey and the spacing of the transects within the survey area. Although ceramics were collected, none of the sherds were diagnostic. Furthermore, the sherds were all less than 1.5cm in diameter and very sparsely distributed. It is possible that ceramic finds from the area were transported to these fields by the Rioni River which flooded this territory, resulting in the deposition of ceramics along with river sediments. The second part of the investigation involves the survey and recording of palaeochannels (that is, ancient channels that generally no longer carry water) within the region of the artificial mound

Phasis and its Landscape

103

Fig. 8   Portion of Besarion Gogolishivli’s 1941 map of the Colchean Lowlands (Poti Museum of Colchian Culture)

settlements (see Fig. 2). The main aim is to determine if they were connected. This part of the research is intended to address whether Hippocrates’ description of Phasis as a dispersed settlement system is true. For example, a palaeochannel (approximately 1.5m wide) is visible surrounding the Sachochuo Mound, leading to the south (Fig. 4). This channel, located 2.5km north of the current channel of the Rioni River, appears from satellite and aerial photos to surround the mound and continue in a southerly direction (Fig. 2). There is a high possibility that the channel was, at some point, connected to the Rioni River. Unfortunately, only 1km of the channel is visible before it is cut by the Poti highway and its continuation is obscured. The next stage of the survey aims to detect the channels around the Sachochuo mounds and to see if they are connected. Such a channel could have functioned as a system of communication. This is important as these channels would have been the primary way in which people and goods would have been transported throughout the settlement area. Another aim of pedestrian survey is to ground-truth the archaeological map of the Colchean Lowlands made by Besarion Gogolishvili in 194117 (Fig. 8). Unfortunately, the map does not provide the precise coordinates for the sites, making it challenging to relocate them. Also, due to modern development in the region, the landscape has changed significantly from the time the original mapping was done. On Gogolishvili’s map, the blue circles with white dots in the middle indicate mounded settlements surrounded by channels; they are mostly concentrated along the banks of the Rioni (Fig. 8). These sites are the best preserved because they are located in the Kolkheti National Park, where the local ecosystem has been protected. Conclusion The current project, investigating the Colchean Lowlands, proves that mounds are located in the Rioni River delta. The Sachochuo mounded settlements are connected via a channel to the Rioni River. As such there is a high possibility that communication between mounds was facilitated through these channels, and that these channels were flowing into the Rioni’s delta. Also, we should consider the fact that lowlands in the vicinity of the Rioni are constantly flooding and only ancient artificial mounds stay dry. The next step of the investigation will be to

17

An archaeological map of the Colchean Lowlands is kept in the Poti Museum of Colchian Culture.

104

Davit Naskidashvili

undertake test pits on several of the artificial mounds mentioned above (see Fig. 2) to determine if evidence of occupation in the 6th–5th centuries BC is present. This should correspond with early occupation of Phazisi. Previous excavations already confirmed the evidence from the Simagre settlement, which was on an artificial mound and dates to the 6th–5th centuries BC. Theoretically, we can suggest that the Simagre mounds and the other artificial mounded settlements under investigation by this project, such as Sachochuo, were interconnected and represent the ‘country of Phazisi’ described by Hippocrates. Acknowledgments: We are grateful to the Shota Rustaveli National Scientific Foundation (Doctorate scholarship DO/173/2-105/14) which generously funded our fieldwork within the framework of the project ‘Archaeological survey of Rioni Delta using geographic information systems (for localisation of the ancient city Phazisi)’. Without the kind support of Professor Vakhtang Licheli, Professor Guram Grigolia, Professor Zviad Kvitsiani, Shio Simsive, Ana Gabunia, Elene Gabliani, Giorgi Uguzashvili, Zolini Gabelaia and Gela Qorqia, none of this could have been achieved.

References Berdznishvili 1969 მ. ბერძენიშვილი / M. Berdznishvili, ქალაქ ფაზისის ისტორიისათვის / Kalak Phazisis istoriisatvis (For the History of the Ancient City Phasis) (Tbilisi 1969). Ewoldsen 2014 H. Ewoldsen, Developing Climate Resilient Flood and Flash Flood Management Practices to Protect Vulnerable Communities in Georgia. Final report (Tbilisi 2014). Online (last accessed 16. Jan. 2018). Gamkrelidze 2012 G. Gamkrelidze, The land of Colchis and the city of Phasis. Towards a historical-archaeological study of western Georgia in the Classical period, in: G. Gamkrelidze, Researches in Iberia-Colchology (Tbilisi 2012) 38–74. Jibladze 2007 ლ. ჯიბლაძე / L. Jibladze, კოლხეთის დაბლობის ძვ. წ. III–II ათასწლეულების ნამოსახლარები / Kolkhetis dablobis dzv. c. III-II atastsleulebis namosakhlarebi (III–II millennium BC Settlements in Colchean Lowlands) (Tbilisi 2007). Johns 1957 W. H. S. Johns (trans.), Hippocrates (London 1957). Kuftin 1949/1950 Б. Куфтин, Материалы к археологии Колхиды (Tbilisi 1949/1950). Lordkipanidze 2000 Phasis. The River and City in Colchis, Geographica Historica 15 (Stuttgart 2000). Mikeladze 1978 თ. მიქელაძე / T. Mikeladze, არქეოლოგიური კვლევა-ძიება რიონის ქვემო წელზე / Arkeologiuri kvleva-dzieba rionis kvemo tselze (Archaeological Research on the Lower Stream of the Rioni) (Tbilisi 1978). Romer 1998 F. E. Romer, Pomponius Mela’s Description of the World (Ann Arbor 1998). Tavamaishvili 2012 გ. თავამაიშვილი / G. Tavamaishvili, ძველი სადგომები სამხრეთ-დასავლეთ საქართველოს ახალშავზღვურ ტერასაზე / Dzveli sadgomebi samkhret-dasavlet saqartvelos akhalshavzghvur terasaze (Ancient Settlements on the New Terrace of the Black Sea, South-West of Georgia) (Tbilisi 2012).

Samshvilde: Multidisciplinary Approaches to a Historical City of Central Transcaucasia David Berikashvili * Abstract: Samshvilde, a historical city of the Kvemo Kartli region in the southern part of Georgia, is a complex and multi-period archaeological site. The city occupies a strategic and defendable location on a long basalt plateau above ravines formed by the Khrami and Chivchava rivers. This distinctive landscape position, combined with environmental conditions that include a mild climate and an abundance of natural resources, has attracted human occupation for millennia. Samshvilde and its surroundings may have been inhabited since the Neolithic era, but the urban complex dates mainly to the medieval period, when it became the region’s principal fortress and political-economic centre. Proximity to the northern branch of the Silk Road further increased the site’s importance. Accordingly, Samshvilde was a place where various ethnic groups and cultures converged, which is reflected in the preserved archaeological remains. Despite the site’s importance and longevity, until recently there has been little concerted archaeological study of Samshvilde. In 2012 the Samshvilde Archaeological Expedition was initiated by the University of Georgia which has taken a multi-disciplinary approach to the site, using archaeological, historic, art-historical and geophysical methods. Future expansion of the project is envisaged through cooperation between the University of Georgia and research organisations and individual specialists who will bring new perspectives to the study of Samshvilde’s past. Keywords: Samshvilde; Samshvilde Archaeological Expedition; Geophysics; The University of Georgia; Interdisciplinary studies

Landscape and Historical Context Samshvilde is a historic city situated in the Kvemo Kartli province, in the southern part of Georgia. The area’s geology consists of Quaternary deposits, Cretaceous and Jurassic limestone, chalky clays, sandstones and volcanogenic formations. Volcanic plateaus have been cut by rivers to form deep ravines and basaltic plateaux that rise above incised valleys. Kvemo Kartli has abundant mineral resources, including copper, gold and iron deposits; stone was mined and quarried, including reserves of obsidian and the region’s ecology supports diverse flora and fauna.1 Since 2012, the Samshvilde Archaeological Expedition of the University of Georgia has been using a multi-disciplinary approach, involving archaeological, historic, art-historical and environmental science methods, to investigate the urban site at Samshvilde.2 Here, a selection of key archaeological features at Samshvilde are reviewed and recent fieldwork and preliminary results are summarised before future objectives for the project are proposed. Samshvilde offers both a challenge and a unique opportunity to investigate an archaeologically significant and multi-period landscape of central Transcaucasia. Georgian historic tradition associates the foundation of Samshvilde with the pre-Christian period. While our project has now detected proof of occupation dating back to the Neolithic period, it is recognised that Samshvilde was founded as an urban political-economic centre only in the early medieval period, in particular, during the 5th century.3



*

1 2 3

University of Georgia, [email protected]. Berdzenishvili 1979, 7. Berikashvili 2017a, 102. Chilashvili 1970, Vol. II, 118; Sanadze 2016, 256.

106

David Berikashvili

Fig. 1   Samshvilde citadel and the eastern tip of the promontory

As the Arabs appeared in Transcaucasia during the mid-8th century, a substantial part of eastern Georgia, including Samshvilde, was placed under the jurisdiction of the Arab Emir. This arrangement continued until the mid-9th century when the region fell under the influence of the Armenian royal Bagratuni dynasty of Shirak.4 In the 10th century, Samshvilde was the capital of the Armenian Kingdom of Tashir-Dzoraget which was a vassal of the Kingdom of Ani. The Geor-

Fig. 2   Samshvilde promontory, western and central parts of the city, facing east

4

Kutateladze 2001, 99.

Samshvilde

107

gian King Bagrat IV restored it within the borders of Georgia in 1064. During the second half of the 11th century Samshvilde was under the influence of the Seljuk Turks and this continued until 1110, when it was liberated by King David IV (the Builder) and placed under the jurisdiction of the Georgian State.5 Various Georgian feudal families controlled Samshvilde in the late medieval and post-medieval period. First the influence of the Orbeli was dominant and later the Baratashvili-Kaplanishvili, whose tenure continued up to the 17th–18th centuries.6 In the second half of the 18th century the ethnic situation in Samshvilde and Kvemo Kartli in general changed significantly. From the end of the same century, these areas were occupied by newly arrived groups of Armenians who settled in Samshvilde and nearby areas.7 From the beginning of the 19th century, Turkish-speaking populations, Germans (1818) and Greeks (1829) were settled in this region by the Russian Imperial government.8 Even today, the population in this part of Georgia continues to have a diversity of ethnic elements. Structures and Features at Samshvilde Samshvilde occupies a long basalt plateau aligned west–east that rises above the confluence of two important rivers, the Khrami and Chivchava9. The medieval city covered the entire length of the promontory and its layout was arranged according to the occupations and status of the population: the western part of the city, which may have been the residential area of the lower classes, was separated from the central part where nobles resided by a 4m-high and 2.5m-wide stone wall (Fig. 1). The central district was separated from the easternmost part of the city, where high-status structures were located, by a 12m-high and 7m-wide fortification wall, forming a citadel (Fig. 2). Such a heavily fortified defensive system has not been preserved at any other medieval site, not only in Georgia but also throughout the Southern Caucasus. The Water Supply System Despite the separation of the city’s three districts, its hydrological network was integrated so that water was supplied to residential, trade and artisanal areas as well as to the noble and royal districts. A well organised and serviceable water supply system was essential for cities of the medieval period. In designing this system, the builders of Samshvilde skilfully took advantage of the natural inclination of the promontory which is equal to 2cm per metre. The total difference between the highest and the lowest points of the city is 70m, which is sufficient to ensure the natural flow of water through the entire city. Our project dedicated a considerable amount of time to studying Samshvilde’s water supply system. It was established that the system starts from the Iraga River to the west of Samshvilde and that outside of the city it was arranged entirely underground. Thus, the system was concealed from the view of potential enemies, while open channels crossed the territory within the city. Two reservoirs or cisterns (diameter 20–25m, depth 2.5m) are cut into the basalt bedrock in the eastern part of the city. These were connected to the network and considerable reserves of water were stored in them, ensuring a ready supply as well as access to water during times of drought or prolonged siege. The ‘Royal Bath’, one of Samshvilde’s noteworthy standing structures, is also connected to the water supply system.

5 6 7 8 9

Kutateladze 2001, 137. Klimiashvili 1964. Berdzenishvili 1979, 114. Kaukhchishvili 1942, Vol, IV. Berikashvili – Alasania 2017, 15.

108

David Berikashvili

Fig. 3   The ‘Royal Bath’, a late- or post-medieval bathhouse

The ‘Royal Bath’ A bathhouse of the later medieval period is located in the eastern part of the city, in what is regarded as the ‘Royal District’, leading us to call this structure the ‘Royal Bath’ (Fig. 3). This unique structure has never been studied previously, making it difficult to talk about its chronology and past usage. While detailed survey and excavation are yet to be conducted, preliminary study through test trenches has established that much of the building is buried beneath soil and rubble. However, it may be said on the basis of its architectural details and by comparison with similar structures preserved in Georgia that the building dates to the 16th–17th centuries and may be regarded as a hamam of the Ottoman Turkish type. Its water supply and the furnace unit located underground require fuller study as does the purpose of walls that are directly connected to the main building. Excavations close to the Royal Bath were started in 2016 and will continue in future seasons. Palaces and Residential Buildings Residential structures at Samshvilde differed according to the social classes for which they were intended. For example, houses preserved in the district thought to have been inhabited by tradespeople and artisans were built from unprocessed stone and timber which contrasts with the mortared stone palaces in the supposed noble or royal districts. Although none of the palace buildings have yet been studied in detail, an impression of their construction and materials can be gained from walls preserved above ground. There are five ruined palaces preserved at Samshvilde. Three are located in the central part of the city and two are in the royal district. One of the royal district palaces was a two-storey building, judging from the preserved walls. It is assumed that this palace

Samshvilde

109

was intended for the king and his family. This structure has been identified as a target to excavate test trenches in future seasons. Churches There are seven recorded churches at Samshvilde that belong to different phases of the medieval period. According to the Georgian Chronicles, the earliest church was built here by Queen Sagdukht in the 5th century,10 but the whereabouts of this building is currently unknown. To help find its location, geophysical survey was used, together with the test excavation trenches; however, these did not establish the location definitively. Of the churches preserved above ground, Samshvilde Sioni is among the most outstanding (Fig. 4). Indeed, it ranks among the most significant monuments of medieval Georgian church architecture. The Old Georgian inscription mentioning the Byzantine emperors, Constantine V Kopronymos and Leo V the Khazar, is preserved on its eastern façade. According to the inscription, the church’s construction dates are defined precisely to the period of 756–777 AD.11 The art historian Niko Chubinashvili made small test trenches in the interior of Sioni Church in 1968,12 but otherwise the building has not been studied up until now. Samshvilde Basilica (10th century), located 150m to the west of Sioni Church, is a Monophysite church. The basilica was built and functioned when Sioni Church was already in ruins as a

Fig. 4   Sioni Church, aerial view facing east

10 11 12

Kaukhchishvili 1955, 142. Muskhelishvili 1943, 97. Chubinashvili 1969, 5.

110

David Berikashvili

result of an earthquake and at the time that Samshvilde was the capital of the Kingdom of TashirDzoraget.13 Restoration works at the basilica were carried out in the 1990s and today it is one of the active churches of the city. The so-called Palati Church, which belonged to the Kaplanishvili, the local feudal dynasty, is located near the basilica. Our surveys of Palati Church in 2014 established that members of the Kaplanishvili family were buried in the interior of this church. Ruins of the two-storey palace belonging to the same family are located nearby. Four other churches are known at Samshvilde: St George’s Church (10th–11th centuries), the Dormition Orthodox Church (10th–12th centuries), St Theodore’s Church (12th century) and the so-called Theogonida. These structures all require further investigation. The Cemetery of Samshvilde One of the main questions arising from our surveys at Samshvilde is the location of the city’s cemetery. After five seasons of fieldwork, using a combination of aerial mapping, pedestrian survey and geophysical methods, it has not yet been possible to locate any extensive cemeteries. One result was the discovery of graves dating from the 15th century that lie to the north of Sioni Church. However, these graveyards represent hastily arranged burials of people who likely died as a result of one of the invasions at this time.14 It may only be assumed that the main cemetery of the city was located far from the residential areas in the same manner as occurs at Dmanisi.15 Megaliths Megalithic features at Samshvilde are in the form of two menhirs and a ‘tetralith’, a monument consisting of three balanced boulders (Fig. 5). One of the menhirs, a 2.45m-high stone column, is located in the interior of the Dormition Church. It seems that it was erected in the pre-Christian era, possibly during the Bronze Age, and the church builders intentionally located it within the medieval building. The ‘tetralith’, which consists of three massive basalt boulders balanced on top of each other, is a unique feature with no known parallels in southern Georgia. At present it cannot be precisely dated because no archaeological excavations have yet been carried out at this location, but it may be presumed that this and the other megalithic monuments date from the Bronze Age. New Investigations at Samshvilde As may be seen from the above overview, Samshvilde is an outstanding archaeological complex. It is perhaps surprising, therefore, that this site has never previously been the subject of a fullscale archaeological investigation. Only small-scale fieldwork was carried out during the Soviet and post-Soviet period which did not provide detail on the site’s stratigraphy and chronology or the distribution of cultural features and monuments. In contrast, there has been fairly extensive archaeological investigation of the surrounding Kvemo Kartli region. This includes investigation of a burial mound of the early stage of the Kura-Araxes culture,16 the Tsopa Palaeolithic and Eneolithic site,17 Tetritskaro settlement,18 early agricultural cultures,19 Shulaveri, Arukhlo, Imiri and

13 14 15 16 17 18 19

Kutateladze 2001, 83. Berikashvili 2017b, 26. Chkhvimiani 2015, 35. Mirtskhulava 1975. Grigolia 1960; Grigolia 1963; Nebieridze 2010. Gobejishvili 1978. Chubinashvili – Nebieridze 1971.

Samshvilde

111

Fig. 5   Megalithic ‘tetralith’ monument

other Eneolithic hills,20 Khrami valley rock monuments21 and Dmanisi, the famous Palaeolithic site, where there are also remains of the Late Bronze Age and medieval period.22 The fact that a site with the importance of Samshvilde remains unstudied leaves important questions regarding Kvemo Kartli’s and the broader Caucasus region’s archaeology unanswered. The Samshvilde Archaeological Expedition of the University of Georgia, which has so far conducted five seasons of fieldwork from 2012 to 2016, is working to redress this situation. By adopting a variety of approaches to the varied archaeological remains at Samshvilde and its surroundings, we seek to generate new information on this multi-period complex. Of the investigated sites near Samshvilde, Dmanisi, located 22km to the southwest, is particularly relevant.23 This is because its geological and geomorphological situation is similar to that of Samshvilde. Similar to at Dmanisi, the promontory at Samshvilde is formed by deep valleys incised into basalt flows by two rivers. There are also archaeological parallels. At both sites, the upper stratigraphic layers are formed by thick layers dating from the high medieval period. The comparable landscape and terrain, geomorphology and nature of the medieval remains suggest that there is a possibility of early prehistoric remains existing at Samshvilde as well; however, locating such remains, if they do exist, would involve comprehensive surveys and specific geological and geomorphological conditions.

20 21 22 23

Gogelia – Tchelidze 1992. Bakhtadze 2007. Gabunia et al. 2000; Gabunia et al. 2003. Gabounia et al. 2002; Gabunia et al. 2003.

112

David Berikashvili

Such comprehensive surveys, involving archaeology, geology, geophysics, anthropology, palynology, remote sensing and osteology, are now being conducted in Samshvilde. The use of different methods is suited to the complicated nature of the site. Systematic and test excavations, geophysics, landscape and cave complex surveys and aerial mapping, as well as palynological, osteological and anthropological studies, provide complementary information to assess the varied archaeological conditions above and below ground.

Fig. 6   Obsidian scraper, arrowhead and amulet (above) and flint and obsidian sickle blades (below)

Samshvilde

113

Excavations Excavations have so far been conducted at two locations: the first is beside the main fortification wall of the city, within what is regarded as the citadel; the other is near the 8th-century Sioni Church.24 Excavations at the citadel have unearthed archaeological deposits of 1.3m depth belonging to the high-late medieval centuries (11th–13th centuries). Underlying and earlier archaeological layers have not yet been explored; the aim in future seasons is to determine the depth and age of the stratigraphy in the citadel area. Artefacts from the trenches are diverse and include ceramics, metal, stone, glass and bone items that date from the high and later medieval centuries.25 Excavations at the Sioni section, where two trenches were opened, have already produced important results. Remnants of a stone mortar structure and graves of the later medieval centuries were discovered to the north of the church. It is noteworthy that the depth of cultural layers, at 1m, is lower here compared to the citadel. The date of this layer is assigned to the 11th–13th centuries on the basis of finds. An initial interpretation of the graves is that they are later and belong to citizens killed during the invasion of the Turkmen leader, the king of Tabriz, Jahan Shah, in the 15th century.26 Alongside systematic excavations, test trenches measuring 1 × 1m, 2 × 1m and 3 × 2m have been excavated across areas of potential further investigation. Important results were gained from the test trenches in the Sioni section where obsidian and flint tools were discovered. Forms include scrapers, burins, points, arrow heads and notched sickle blades, attributed to the final stage of the Georgian Neolithic by Prof. G. Grigolia and chronologically placed within the Tsopa culture of Kvemo Kartli (Fig. 6).27 Until these discoveries, the oldest material from Samshvilde and adjacent areas was represented by artefacts of the Kura-Araxes culture.28 The discovery of Neolithic tools is a significant novelty and raises the prospect of identifying more extensive prehistoric deposits here. Geophysical Survey Concurrent with archaeological excavations, a geophysical survey is being carried out using ground penetrating radar (GPR).29 Three areas or ‘sections’ were surveyed using monostatic antennas of various capacities to the north, north-east and south of Sioni Church in 2015; features located at depths over 6m from the ground surface were observed. In the first section, north of Sioni Church, ten 15m-long passes were made. The results appeared to be very noteworthy as they established that the area to the north of the church is characterised by archaeological deposits at three levels. The first level was interpreted as graves and wall fragments located close to the ground surface; the second level (at 3m to 6m depth) was identified as fragments of the wall base and sinkholes; remnants of a 10m-long and 10m-wide structure of complicated composition with walls of different thicknesses were observed at the third level, below 5–6m.

Berikashvili 2017b, 23–25. Berikashvili 2017b. 26 Significant information about the invasion of Jahan Shah is provided by the 15th-century Armenian historian, Thomas of Metsoph: ‘On March 27, 1440, on the day of Easter … after a 50-day siege, they made the Samshvilde citizens suffering from hunger and thirst open the door of the fortress by deceit … thousands of people were killed, their property was seized. A minaret was built from 1664 cut heads near Samshvilde fortress by order of the Shah…’ Metsopeli 1937, 24–26. 27 Grigolia – Berikashvili in press. 28 Mirtskhulava 1970; Mirtskhulava 1975. 29 This survey utilises a certified GPR radar - Zond 12, 500, 300 MHz monostatic antennas and 150, 75, 38 MHz and 2 GHz bistatic antennas. 24 25

114

David Berikashvili

In the second section, to the north-east of the church, structural (wall) remains were also found at a depth of up to 6m in the shape of a ‘bow tie’. In the third section, to the south of the church and near the Khrami canyon, sinkholes were again located at various depths as well as fragmentary sections of walls. Analysis of the GPR images established that the area to the north of Sioni Church, where underground deposits are located at three levels, is the most promising from an archaeological point of view. The observed features located closest to the ground surface correspond to void spaces of anthropogenic origin, possibly graves. The excavation of graves, described above, confirmed the results of the GPR survey. Aerial Mapping and Cave Exploration Aerial mapping of the Samshvilde promontory and adjacent areas has obtained a precise view of the urban layout of the city-site, its main districts, the hydrological system and the location of architectural remains. In addition, aerial and video mapping is being conducted in the canyons of the Chivchava and Khrami rivers where there are multi-tiered cave complexes which are very difficult to reach. Using this method to define the exact location, number and size of these cave complexes is very effective. For example, as a result of aerial mapping of caves in the Chivchava canyon, carried out in 2015, access routes were defined, allowing our team to reach the caves and undertake test excavations there. Large amounts of ceramics and bone material were discovered in the test trenches as well as on the surface; obsidian artefacts were also discovered. Some caves have entrances that are disguised and built up with masonry. It is impossible to establish the nature of these disguised and almost inaccessible caves using aerial mapping alone. Accordingly, a collaboration with the Academy of Robotics and Engineering of the University of Georgia is currently developing an unmanned aerial vehicle which will be adapted for movement on the rocky terrain. Palynological, Archaeozoological and Anthropological Studies Palynological, osteological and anthropological analysis is being undertaken as part of the excavations at Samshvilde. Palynological results obtained from the medieval layers of the citadel provide information on the environmental conditions of Samshvilde during the 12th–13th centuries and the nature of household activities at this time. The obtained results suggest that pine as well as broad-leaf forest of beech, lime and elm must have grown around the city. Spores of forest fern are also seen in the palynological spectrum, particularly Asplenium trichomanes which grows only in broad-leaf humid forests. Lycopodium clavatum also grows in humid climatic conditions. As the pine tree grows in dry rocky areas, it may be concluded that these were the conditions at Samshvilde in the 12th–13th centuries. The plain near Samshvilde seems to have been used for seeding grains which points to welldeveloped agriculture. Grains of nut and hazelnut as well as grapes were discovered in the palynological samples which confirms viticulture and wine-making. There are also indications of the medicinal use of plants, which may have been stored in the Samshvilde citadel alongside cereals, flour and hazelnuts.30 Archaeozoological studies define the anatomic identity of excavated animal bones. Despite the fragmentary nature of the excavated osteological material, anatomic and taxonomic categories of 862 samples of vertebrate animals were defined. Fragments of longitudinal bones are dominant (52%) followed by ribs and vertebrae (20.5%); skull fragments (including teeth and mandibles) make up 11.5% of the material. Animals identified at this stage of the analysis are the following:

30

Kvavadze 2017.

Samshvilde

115

Bos taurus, Sus scrofa, Ovis aries and Gallus gallus domesticus; in addition, one sample belongs to a donkey or a small-size horse and one sample belongs to Silurus glanis. Anthropological studies were undertaken following the discovery of the first graves at Samshvilde during excavations in 2016. These studies are ongoing, and they are one of the project’s top priorities. At present it can be said that the bones from the excavated graves show clear traces of forceful mortification. This adds to the interpretation that the burials are of victims of a violent invasion of the city. Problems, Questions and Future Directions The results obtained by the Samshvilde Archaeological Expedition demonstrate the high significance of this unstudied site. The importance of the Samshvilde urban complex at a regional scale and at the scale of the entire Southern Caucasus requires a multi-disciplinary approach, using a variety of methods to obtain information on the varied archaeological remains. Due to the scale, complicated stratigraphy and large chronological range at the site, many issues remain problematic. However, the work completed so far has helped to define some key objectives. At the present stage of the survey, specific objectives of the expedition are as follows. We seek to define the hydrological system scheme in the royal district and the nature of the Royal Bath. Residential areas in the western section of the city appear to be the dwellings of the city’s lower classes and we will investigate the characteristics of this housing. The districts where nobles and the upper classes dwelt also require investigation, particularly the identified palace structures. Churches are prominent features at Samshvilde, and further research is required to understand the chronology and nature of these buildings. There is a particular question over the location of the church built in the 5th century by Queen Sagdukht which our project aims to locate using geophysics and test excavation. Sioni Church, a building of high significance in Georgian medieval church architecture, and its surrounding infrastructure will be the subject of further study. Mortuary practice at Samshvilde will be studied through the analysis of the graves excavated near Sioni Church, including osteological study of the recovered human skeletal remains. More generally, we seek to find the currently unknown location of the main cemetery and burial areas of the city. Future archaeological investigations will attempt to establish a stratigraphic sequence, especially in the citadel area. This will be assisted by the use of scientific dating methods. The potential for prehistoric occupation layers will be explored, both on the Samshvilde plateau and in the caves along the surrounding canyons. The discovery of Neolithic stone tools is a significant finding, which may point to a more extensive prehistoric occupation at the site. Advancing these objectives will require substantial time and material resources. Therefore, the University of Georgia and Samshvilde Archaeological Expedition seek to cooperate with scientific-research organisations and individual specialists to bring their expertise and knowledge to this unique and important archaeological complex. Acknowledgments: The Samshvilde Archaeological Expedition of the University of Georgia benefits from the contribution of specialists working in a number of disciplines: Dr David Odilavadze (Tbilisi M. Nodia Institute of Geophysics); Dr Maia Bukhsianidze (Zooarchaeologist, Georgian National Museum); George Gagoshidze (Art historian, National Agency for Cultural Monument Preservation of Georgia); Dr Eliso Kvavadze (Palynologist, Georgian National Museum); Dr Liana Bitadze (Institute of Anthropology of the Tbilisi Iv. Javakhishvili State University); Dr Nino Kebuladze (Restorer, Georgian National Museum); Dr Tengiz Gabunia (Architect, National Agency for Cultural Monument Preservation of Georgia); Prof Manana Sanadze (Historian, the University of Georgia); Prof. Sergo Tsiramua (Department of Engineering, the University of Georgia) and Professor Emeritus Guram Grigolia (Tbilisi Iv. Javakhishvili State University).

116

David Berikashvili

References Bakhtadze 2007 ნ. ბახტაძე / N. Bakhtadze, კლდის ხუროთმოძღვრების გენეზისი და განვითარების გზები საქართველოში / Kldis khurotmodzghvrebis genezisi da ganvitarebis gzebi sakartveloshi (The Genesis of Cave Architecture in Georgia) (Tbilisi 2007). Berdzenishvili 1979 დ. ბერძენიშვილი / D. Beridzenishvili, ნარკვევები საქართველოს ისტორიული გეოგრაფიიდან / Narkvevebi sakartvelos istoriuli geograpiidan (The Historical Geography of Georgia) (Tbilisi 1979). Berikashvili 2017a D. Berikashvili, Samshvilde. The project of the University of Georgia, არქეოლოგია / Archaeology. The Scientific Journal of the University of Georgia 1, 2017, 102–114. Berikashvili 2017b D. Berikashvili, არქეოლოგიური გათხრები სამშვილდეში (2015–2016 წლის მასალები) / Archaeological Excavations in Samshvilde (Materials 2015–2016) (Tbilisi 2017). Berikashvili – Alasania 2017 D. Berikashvili – G. Alasania, Historic city of Georgia – Samshvilde, in: M. Bouchenaki (ed.), Contributions to the General Assembly 2016. 18th International Assembly of the Experts of the Fondazione Romualdo Del Bianco (Florence 2017) 15–19. Chilashvili 1970 ლ. ჭილაშვილი / L. Chilashvili, ქალაქები ფეოდალურ საქართველოში / Kalakebi peodalur sakartveloshi (The Cities in Feudal Georgia) (Tbilisi 1970). Chkhvimiani 2015 ჯ. ჩხვიმიანი / J. Chkhvimiani, Gravestones with Georgian inscriptions from the high medieval period at Dmanisi, Georgia, კადმოსი / Kadmos 7, 2015, 66–105. Chubinashvili 1969 Н. Г. Чубинашвили, Самшвилдский Сион. Его место в развитии груз. архитектуры VIII-IX вв. ( Samshvilde Sioni and its Place in the Development of Georgian Architecture of VIII–IX Centuries) (Tbilisi 1969). [In Russian] Chubinashvili – Nebieridze 1971 ტ. ჩუბინაშვილი – ლ. ნებიერიძე / T. Chubinashvili – L. Nebieridze, 1969 წლის არქეოლოგიური კვლევის შედეგები ქვემო ქართლში / 1969 tslis arkeologiuri kvlevis shedegebi kvemo kartlshi (The Results of Archaeological Excavations in Kvemo Kartli in 1969) (Tbilisi 1971). Gabounia et al. 2002 L. Gabounia – M.-A. de Lumley – A. Vekua – D. Lordkipanidze – H. de Lumley, Découverte d’un nouvel hominidé à Dmanissi (Transcaucasie, Géorgie). Comptes Rendus Palevol 1, 4, 2002, 243–253. Gabunia et al. 2000 L. Gabunia – A. Vekua – D. Lordkipanidze – C. C. Swisher III – R. Ferring – A. Justus – M. Nioradze – M. Tvalchrelidze – S. C. Antón – G. Bosinski – O. Joris – M.-A. de Lumley – G. Majsuradze – A. Mouskhelishvili, Earliest Pleistocene hominid cranial remains from Dmanisi, Republic of Georgia. Taxonomy, geological setting, and age, Science 288, 2000, 1019–1025. Gabunia et al. 2003 ლ. გაბუნია – ა. ვეკუა – დ. ლორთქიფანიძე / L. Gabunia – A. Vekua – D. Lordkipanidze, ახალი ჰომო დმანისიდან. ქვედა პლეისტოცენი / Akhali homo dmanisidan. kveda pleistotseni (The new homo from Dmanisi (Georgia) Lower Pleistocene site), in: J. Kopaliani (ed.), Dmanisi IV (Tbilisi 2003) 40–52. Gobejishvili 1978 გ. გობეჯიშვილი / G. Gobejishvili, თეთრიწყაროს ნასოფლარი / Tetritskaros nasoplari (Tbilisi 1978). Gogelia – Tchelidze 1992 დ. გოგელია - ლ. ჭელიძე / D. Gogelia – L. Tchelidze, აღმოსავლეთ საქართველოს ენეოლითური ძეგლები / Aghmosavlet sakartvelos eneolituri dzeglebi (Chalcolithic archaeological sites from East Georgia), საქართველოს არქეოლოგია / Sakartvelos arqeologia (Archaeology of Georgia) 2, 1992, 17–35.

Samshvilde

117

Grigolia 1960 გ. გრიგოლია / G. Grigolia, ქვემო ქართლის უძველესი ძეგლები / Kvemo kartlis udzvelesi dzeglebi (The Ancient Sites in Kvemo Kartli) (Tbilisi 1960). Grigolia 1963 გ. გრიგოლია / G. Grigolia, ქვემო ქართლის ქვედა პალეოლითი (წოფი I) / kvemo kartlis kveda p’aleoliti (ts’opi I) (The Lower Paleolithic in Kvemo Kartli (Tsopi I)) (Tbilisi 1963). Grigolia – Berikashvili in press G. Grigolia – D. Berikashvili, The Neolithic Stone Industry from Samshvilde, Archaeology 2 (in press). Kaukhchishvili 1942 ს. ყაუხჩიშვილი / S. Kaukhchishvili, ბერძენ მოსახლეთა ისტორია საქართველოში / Berdzen mosakhleta ist’oria sakartveloshi (The History of Greek Residence in Georgia) (Tbilisi 1942). Kaukhchishvili 1955 ს. ყაუჩიშვილი / S. Kaukhchishvili, ქართლის ცხოვრება. ტომი I / Kartlis Tskhovreba. Tomi 1 (Life of Kartli. Vol. 1) (Tbilisi 1955). Klimiashvili 1964 ა. კლიმიაშვილი / A. Klimiashvili, მასალები ქართლ-კახეთის სამეფოების მე-15–18 საუკუნეების ისტორიისათვის / Masalebi kartl-kakhetis samepoebis me-15–18 saukuneebis istoriisatvis (Materials for the History of Kartli and Kakheti Principalities in the 15th–18th Centuries (Tbilisi 1964). Kutateladze 2001 ქ. ქუთათელაძე / K. Kutateladze, ქვემო ქართლის პოლიტიკური ისტორიის საკითხები / Kvemo kartlis politikuri istoriis sakitkhebi ( Kvemo Kartli – Issues for Political History) (Tbilisi 2001). Kvavadze 2017 ე. ყვავაძე / E. Kvavadze, 2017 წლის პალინოლოგიური კვლევების შედეგები სამშვილდედან / 2017 tslis palinologiuri kvlevebis shedegebi samshvildedan (The Results of Palinological Studies in Samshvilde in 2017) (Tbilisi 2017). Metsopeli 1937 თ. მეწოფელი / T. Metsopeli, ისტორია / istoria (History) (transl. from Armenian into Georgian by Melikset-Beg) (Tbilisi 1937). [translated from Armenian to Georgian] Mirtskhulava 1970 გ.მირცხულავა / G. Mirtskhulava, სამშვილდის სამაროვანი / Samshvildis samarovani ( Samshvilde cemetery), ძეგლის მეგობარი / Dzeglis Megobari 23, 1970, 47–53. Mirtskhulava 1975 გ. მირცხულავა / G. Mirtskhulava, სამშვილდე / Samshvilde (Tbilisi 1975). Muskhelishvili 1943 ლ. მუსხელიშვილი / L. Muskhelishvili, სიონის წარწერები / Sionis tsartserebi (Sioni Inscriptions. Moambe) (Tbilisi 1943). Nebieridze 2010 ლ. ნებიერიძე / L. Nebieridze, წოფის ენეოლოთური კულტურა / The Tsopi Chalcolithic Culture, Studies of the Society of Assyriologists, Biblical Studies and Caucasiologists 6 (Tbilisi 2010). Sanadze 2016 მ. სანაძე / M. Sanadze, ქართველთა ცხოვრება. ქართლის მეფეთა და პატრიკიოსთა ცხოვრების ქრონოლოგია / Kartvelta tskhovreba. Kartlis mepeta da patrikiosta tskhovrebis kronologia (The Life of Georgians. Chronology of Kings and Patricios of Kartli (from Parnavaz to Ashot Kurapalat)) (Tbilisi 2016).

Medieval Underground Shelters (Darnebi) of South-west Georgia Abby Robinson * – Giorgi Khaburzania ** Abstract: This chapter is concerned with a type of medieval architecture that seems almost entirely confined to the Samtskhe-Javakheti region of south-west Georgia: rock-cut underground shelters called darnebi (sing. darani). These remarkable structures, which occur beneath settlement remains, have certain characteristic features, including entrance tunnels and heavy stone doors with a distinctive pivot hinge. Up till now the darnebi have been little investigated – with one notable exception at Lebisi – and their dating has remained open to question. Here, we argue that their initial proliferation is likely to have occurred between the 8th and 11th centuries AD; this estimation is based on architectural parallels with several better-studied caves in the river valleys, as well as incidental mentions in medieval texts. The design of the darnebi suggests they were occupied temporarily but repeatedly, and we also discuss the most probable periods of re-use. If the chronology we propose for the darnebi is correct, this can cast light on the history of south-west Georgia as a frontier and buffer zone between powerful empires – Arab, Byzantine, Persian, Mongol and Turkish – that dominated the surrounding region in the medieval period. Most particularly, it can elucidate how the local inhabitants used their finely tuned knowledge of the geography, and skill in adapting it, to prevail against the onslaughts of multiple adversaries and hold their ground over centuries. Keywords: Darnebi; Underground shelters; Refuges; Caves; Samtskhe-Javakheti; Borders and frontiers

Introduction Between 2014 and 2017, an extensive field survey was carried out by the authors over an area of c. 1300km² centred on the Aspindza and Akhalkalaki districts of the south-west Georgian region of Samtskhe-Javakheti (Fig. 1). The Kura (Mtkvari) River forms a deep gorge through the middle of the survey region. On its eastern side is the Javakheti plateau, an extensive grassland plain dotted with lakes and wetlands and traversed north–south by the Abuli-Samsari mountain range. Opposite, to the west of the river, are the Erusheti highlands, part of Samtskhe, where lofty mountains enclose a number of smaller plateaus and basins. The second major river in the survey area, the Paravani, separates the north-east of the Javakheti plateau from the south-west. This chapter focuses on a feature which occurs often in this region but has rarely been observed elsewhere:1 a form of underground, rock-cut shelter known in Georgian as darnebi (sing. darani). To date, we have recorded more of these structures, with smaller distances between them, on the western side of the Kura than to the east. This circumstance may reflect geological conditions. The Javakheti area was significantly affected by volcanic activity in the Quaternary period and basalt lava flows were widespread, filling up the plateau and creating conditions unfavourable for excavating darnebi. The softer rock (tuffs, tuff breccias, ash tuffs) and lake sediments covering much of the Erusheti side were better suited to tunnelling underground. Nevertheless, darnebi do also occur on the Javakheti side – in pockets of softer material that are found mainly,



University of Melbourne, [email protected]. National Agency for Cultural Heritage Preservation of Georgia, [email protected]. 1 Pers. comm. Kakha Kakhiani (June 2017), who was one of the excavators of the Lebisi darani (see below). The only instance of darnebi he is aware of outside Samtskhe-Javakheti is in the Dmanisi area of Kvemo Kartli, to the east of the survey region. *

**

120

Abby Robinson – Giorgi Khaburzania

Fig. 1   Map of the survey area (map: A. Robinson, base map uses ASTER GDEM, a product of METI and NASA)

but not exclusively, on the edges of the plateau.2 In addition, the Javakheti plateau is today the most intensively settled part of the survey region, with many villages and agricultural fields, and it should be acknowledged that fewer darnebi may have been recorded there because their traces were destroyed by modern farming and construction works. Here, we focus on the darnebi as a tangible manifestation of disputed ground in this part of south-west Georgia during the medieval period. We will argue that similarities between architectural elements they incorporate and those found in the better-studied cave shelters in the river valleys, as well as supporting evidence from the ancient written sources, enable us to tentatively assign the proliferation of the underground shelters to the 8th–11th centuries AD. This represents more confident dating than has previously been attempted and may ultimately help to clarify the turbulent events of the medieval period and their outcomes in the study region. Regarding the overall themes of this volume, the broader questions at the heart of this investigation all concern landscape: how geography is incorporated into built defences; the qualities and significance of

2

We thank Rusiko Chagelishvili, geologist at the Georgian National Museum, for sharing her observations on these matters (pers. comm. March 2017).

Medieval Underground Shelters (Darnebi) of South-west Georgia

121

geographical borders; and how defensive strategies compare between highlands and lowlands. These are matters of the extent and distribution of human impact on the terrain and can only be explored through extensive field survey combined with fine-tuned analysis of individual sites in their immediate contexts where necessary. The darnebi were located by us, in the first instance, via descriptions in the publication Javakheti: Historical-Architectural Guidebook.3 In each case, Berdzenishvili and his co-authors provide the name of the closest modern-day settlement; we then cross-referenced this information against a 1:50,000-scale Soviet-era topographic map that forms part of the base cartography of the survey GIS. In such a way, it was possible to estimate geographic coordinates in preparation for ground-checking. Once in the field, additional examples were pointed out to us by residents of nearby villages. Finally, the survey was designed from the outset to be multi-stage, whereby the experience and results of each season contributed to shaping future expeditions.4 Accordingly, as we became increasingly familiar with the types of landscapes in which darnebi occurred and learned to recognise signs of them on the surface, we were able to incorporate investigating additional sites into our plans. Descriptions of the Darnebi While darnebi have to date been little investigated,5 there is an exception at Lebisi, located c. 20km south-south-west of the town of Aspindza, on the Niala plateau in the Erusheti highlands. This complex was excavated in 2007 under the direction of S. Burdiladze and K. Kakhiani, who cleaned it out so as to reveal ceilings, walls and floors, and determined that it was a one-layer site. They uncovered a system of 12 connected tunnels of varying lengths and diameters (see plan, Fig. 2), concealed under the site of an extensive ancient settlement on the plateau – the area is now deserted except for a scattering of farms. Most of the Lebisi complex’s walls and roofs are formed from the natural clay into which they are cut, reinforced by variously sized, but mainly large to very large stones. The floors are of packed clay, except in tunnel no. 7, where the ground is paved with stones; this tunnel is relatively short (4.8m) and narrow (0.6m) and it seems to be an entrance passage. There are three circular entrance/exit holes connecting the below-ground structures to the surface above in the vicinity of tunnels 5, 6 and 7. Off tunnel no. 8, which has niches for lamps along its walls, there is a small rectangular room (1.75 x 1.45m, and 1.2m Fig. 2   Plan of the Lebisi darani (after Burdiladze 2010, 79)

3 4

5

Berdzenishvili et al. 2000. See Cherry et al. 1991, 3–4, and Sagona – Birkett-Rees 2016, especially p. 88, for descriptions of other survey projects with multi-stage elements. Burdiladze 2010, 74.

122

Abby Robinson – Giorgi Khaburzania

Fig. 3   Stone door at the entrance to the Lebisi darani (photos: G. Khaburzania)

high) with a 0.5m² pit in the floor and evidence of smoke on its walls. Another room (c. 2.6m long, 1.1–1.5m high, and partially collapsed on one side) is located between the nearby tunnels 9 and 10. A damaged millstone was found at the intersection of nos. 10 and 11, the latter of which appears to be another entrance passage, connecting the darani to the plateau via an opening at its western end.6 Tunnel 2 at Lebisi – which is relatively narrow (0.6m) like no. 7 and also seemingly an entranceway – is closed off at the point where it joins the rest of the complex by a substantial stone door (Fig. 3). The door is formed of a single large stone slab, approximately 1m high, 0.7m wide and 12cm thick, contained within a frame built of four other large stones.7 It employs a distinctive type of pivot hinge, comprising a hemispherical knob on the door’s top edge that fits into a socket carved into the lintel. The door opens into the tunnel complex. A circular indentation, c. 12cm in diameter, carved into the middle of the door’s surface, appears to have acted as a kind of lock when a log was inserted into it and braced, at the other end, against a wall. A large stone rolled across the door once closed would also have prevented access from outside. During our survey, darnebi comparable to Lebisi were recorded on the Erusheti side of the Kura at Kuleti, Atskvita, Shalosheti and Toloshi; in addition, likely entrances were observed at Gundi, Dumeila and Giorgitsminda. On the Javakheti side, there are darnebi at Saro, Khizabavra, Vardistsikhe and Kulalisi. Further instances have been reported at Khando and Kotelia,8 but these could not be located by us; at Murjakheti we were shown the site of what villagers said was a darani that had been filled in due to the danger it posed to local children, but very little evidence now remained of the structure itself (Figs. 4–5). This list of confirmed and possible darnebi is not exhaustive.9 The entrances were made (presumably deliberately) unobtrusive in the terrain and may now in many cases be concealed still further by overgrown vegetation or damaged beyond recognition by time. Therefore, it can be assumed that some, if not many, of the structures have

6 7 8 9

Burdiladze 2010, 74–76. Burdiladze 2010, 74. Berdzenishvili et al. 2000, 60 (Khando), 62 (Kotelia). See also Tatishvili 1970, 18–22. The author identifies ‘tens’ of structures (p. 19) recorded in the area between Vardzia and Tmogvi (also within our survey region), called ‘darnebi’ by villagers and said by them to have been used in the past as hiding places during raids. Tatishvili seems, however, to be describing large kurgans (burial mounds), of the type later excavated by Otar Japaridze in the same region (see Japaridze et al. 1981); their entrances all face east, they are comprised of only single rooms, they are found in clusters or alone, they are sometimes surrounded by a circle of stones, they are somewhat at a distance from settlement remains, etc. Tatishvili himself concludes that the employment of these structures as hiding places was likely a secondary re-use, although he cannot be sure of their original purpose and differentiates them from what he knows as kurgans (as well as what he would usually classify as darnebi) (p. 22).

Medieval Underground Shelters (Darnebi) of South-west Georgia

123

Fig. 4   Distribution map of darnebi (map: G. Khaburzania)

a

b

d

c

Fig. 5   a. Stone door in unexcavated darani at Shalosheti; b. Interior of Kuleti tunnel; c. Entrance to darani at Saro; d. Roof of tunnel at Toloshi darani (photos: G. Khaburzania)

124

Abby Robinson – Giorgi Khaburzania

left no trace on the surface and are now undetectable, even if they are preserved underground. Of the darnebi that were explored and recorded by us, all shared architectural features with the complex at Lebisi: entranceways and passages leading down from the ground surface, tunnels, sometimes rooms, and the distinctive stone doors. The chronology of the darnebi has remained open to question. During the Lebisi excavations, pot sherds were recovered from tunnels 1, 2 and 11 and also near the intersection of nos. 1 and 8. They included mainly plain (dark grey or black) and a few green-glazed fragments, all identified by the excavators as late medieval types and dated to around the 16th–17th century AD on the basis of comparisons with materials recovered from nearby settlements.10 The excavators rightly note, however, that this is no sure guide to when the darnebi were built or their overall period of use.11 Their sturdy and relatively elaborate construction suggests they were intended for more than one-off use; thus the pottery at Lebisi may point only to a single, perhaps final, occupation among many. The scarcity of artefacts before the 16th–17th centuries may indicate that the facilities were cleaned out between visits. No pottery was found in any of the darnebi visited during the course of our survey. In its absence, however, another clue to the structures’ chronology is provided by comparable architectural features that appear in manmade caves in the walls of the Kura and Paravani valleys. Dating the Darnebi Several caves recorded during the survey are equipped with the same kind of entrances as the darnebi: narrow tunnels leading to heavy stone doors with the characteristic hinges and locking mechanisms (Fig. 6). Manmade or remodelled and hewn out of the cliffs overlooking the Kura and Paravani rivers, these caves are situated at Gelsunda, Akhali Navardzievi (1 and 2) and Khando. Further instances were reported by G. Gaprindashvili at Zeda Vardzia and Pia12 but have not as yet been located by us. As with darnebi, the list of caves cannot be exhaustive, due to difficulty of access, and the Zeda Vardzia and Pia caves may be among those that are no longer possible to see or to reach from the ground.13 In his typology of caves in the Kura Valley, Gaprindashvili dates Akhali Navardzievi 2 – and by association the other caves with the same doors and tunnels – to the 8th–11th centuries AD.14 The lower date (8th century) is based on resemblances he discerns between the shape of the rooms in Akhali Navardzievi 2 and the nearby Jolda caves, the latter having been assigned by him to the 8th–11th centuries based on parallels elsewhere.15 The upper date (11th century) is determined from an inscription that reportedly still exists inside Akhali Navardzievi 1; this records the sacking of Akhalkalaki, the administrative centre of the Javakheti region, by a force of Seljuk Turks under the command of Bars, son of the sultan Alp Arslan.16 The inscription’s author, one Bakhat, laments, ‘I have been in terrible trouble. [Bars] came to Javakheti and threatened all Christians and especially Akhalkalaki, which he conquered taking all the prisoners with him.’17 These events

10 11 12

13 14 15 16 17

Burdiladze 2010, 74 –76. Burdiladze 2010, 76. Gaprindashvili 1957, 35–36. In addition to consulting Gaprindashvili’s books (1957, 1959) and – as with the darnebi – being provided with information by villagers, we were assisted in locating caves through access to the records of a survey of the Kura River valley conducted by archaeologists from the Ministry of Culture and Monument Protection of Georgia in 2006 (unpublished inventory stored in the archives of the National Agency of Cultural Heritage Protection of Georgia). Gaprindashvili himself (1957, 12) notes that some caves are inaccessible. Gaprindashvili 1957, 35–36. Gaprindashvili 1957, 12–15. Gaprindashvili 1957, 16–17, 20. Gaprindashvili 2002/2003, 47–48. The translation from old Georgian is our own.

Medieval Underground Shelters (Darnebi) of South-west Georgia

a

c

b

d

125

Fig. 6   a. Entrance to cave shelter at Gelsunda; b. Pivot hinge on door at Gelsunda; c. Tunnel entrance at Akhali Navardzievi 2; d. Stone door lying on the ground at Akhali Navardzievi 2 (photos: G. Khaburzania)

took place in 1064, when the Seljuk Turks, the chronicles record, ‘took captive an incalculable number of Christians. They bore off treasure and many prisoners, and the river of [Akhalkalaki] was dyed with blood.’18 Given the architectural parallels, it is plausible to suggest that the caves and darnebi flourished at approximately the same time. A medieval chronicle can be used to further support the proposition, based initially on Gaprindashvili’s typology, that this was between the 8th and 11th centuries AD. The History of David, King of Kings, which describes the reign of David IV of Georgia (1089–1125), ‘the Builder’, and events leading up to it, tells how the residents of Samtskhe (among other regions in the south and west of Georgia) took shelter from the depredations visited upon them by the Seljuk Turks from the 1060s onwards. The Turks ‘began… to raid, to pillage and plunder, to burn, destroy, and to carry off captive us Christians.’ In response, the Georgians ‘escape[d]… to forests, cliffs,

18

Matthew of Edessa, 102; Thomson 1996a, 298–299. See also Salia 1983, 152.

126

Abby Robinson – Giorgi Khaburzania

caves, or holes [Old Georgian: ჴურელთა] in the ground.’19 There is nothing in the chronicler’s language to exclude that the holes he speaks of here are darnebi. To escape the 11th-century raids, the Georgians took refuge in both caves and underground shelters. In turn, this fits the pattern of the invaders’ tactics, which involved occupying upland pastures in the springtime, then relocating to river valleys in the winter.20 Sturdy but temporary defences were conceivably needed in both locations, to be employed as a response to the same – but seasonally shifting – threat. Thus, both the surviving architecture and the chronicle provide evidence for assigning the darnebi of south-west Georgia to the period between the 8th and 11th centuries AD. This proposition is further supported by a broad overview of historical events, which, in addition, suggests favouring the later part of that date range. The early-8th century saw damaging Arab incursions into the region, and periodic clashes continued for several decades more. However, the main intent of the Arabs of the early conquest era was subjection of local populations, then ongoing access to money and men: ‘[they] were concerned with nothing but tribute, taxes and military levies.’21 Indeed, it is arguable that frequent attacks on the lands from which they drew revenue and personnel would have been against the conquerors’ own interests. From the last quarter of the 8th century till the end of the 10th was a period of Georgian expansion and relative strength.22 The 11th century, however, was dominated by fighting at home. Hostilities with the Byzantine Empire broke out in the 1020s, and civil strife (in part brought about by imperial machinations) dominated the 1030s–1050s.23 But it was the arrival of the Seljuk Turks in the 1060s, as mentioned above, that instigated the kind of repeated, destructive raiding to which the rock-cut defences – caves and darnebi – seem particularly well suited as defences. Despite the fact that, at one time, ‘so great was [the Turks’] strength and multitude… no one could think at all… of expelling or harming them,’24 David IV did ultimately succeed in driving them out of Georgia. Even early in his reign people began to repopulate their lands.25 Although the struggle to secure the country’s borders continued, the peace held within them for almost a century; Georgia thrived right up to the end of the sovereignty of David’s much-admired granddaughter, Tamar (1184–1213), and then for some years more. It is conceivable that the rock-cut defences came back into use during the violent Mongol incursions that followed in the 13th century, although the Mongols, like the Arabs before them, tended more to treaties and taxation after an initial onslaught than to repeat attacks. The damaging assaults by Tamerlane’s forces in the 14th century are another possibility. So far, however, we have encountered no material or textual evidence to support either proposition. On the other hand, the late medieval pottery found in the Lebisi darani indicates that Lebisi and, by extension, other similar structures may have been used to shelter from the Safavid Persians and/or Ottoman Turks who periodically devastated south-west Georgia in the 16th and 17th centuries. Hasan Rumlu, chronicler of the early Safavid dynasty recounts that, in the face of a Persian attack in c. 1551, the Georgians in the south-west ‘took refuge in the hills, and in caves, and… in forts,’ but makes no mention of underground shelters in his account.26 A small quantity

19

20 21 22 23 24 25 26

Thomson 1996b, 310–311. The equivalent in modern Georgian of the word for ‘hole’ in this passage is ‘ხვრელი’, which can also have the meaning of ‘lair’ and plausibly denotes, in that sense at least, a place of concealment. Thomson 1996b, 323. Salia 1983, 124–129. See Salia 1983, 132–141, for an overview. See Salia 1983, 147–149, for an overview. Thomson 1996b, 323. Salia 1983, 154–156, 160; Thomson 1996b, 316. Hasan-i-Rumlu, 159. Although underground shelters are not part of his story, the author does in this context make much of the cave monastery at Vardzia (in his account called ‘Darzabad’). It should be noted that, while visible in the walls of the Kura Gorge by the time Hasan Rumlu was writing, this massive and intricate structure was, prior to an earthquake in 1283, supposedly completely secreted under the ground. See also Hasan-i-Rumlu, 169, for another reference to refuges, some three years later, and how the Georgians ‘escaped to the hills and caves and forests, and were besieged in forts.’

Medieval Underground Shelters (Darnebi) of South-west Georgia

127

of handmade and crudely formed medieval pottery was found by us in the cave known as Akhali Navardzieni 2, but as yet it has not been possible to refine the dating of it to the extent where it can cast any further light on the topic of the chronology of the darnebi.27 Another medieval text may provide one final, if tentative, pointer to the earliest date for the darnebi, and it again falls within the 8th to the 11th century range. A century before the arrival of the Seljuk Turks in Georgia, the Arab Muslim traveller and geographer Ibn Hawqal (fl. 943–969 AD), also linking caves and underground shelters, remarked that the Byzantine Empire (in his account extending, in the east, up to Armenia) was comprised of ‘mountains, citadels and fortresses, cave dwellings and villages dug out of the rock or buried under the earth.’28 Ibn Hawqal does not appear to have visited Georgia,29 and although his description evokes the landscape of SamtskheJavakheti, which is replete with all the features – manmade and natural – he describes, it also brings to mind the Cappadocia region, with its famous underground villages. Further research is required to clarify what relationship, if any, exists between the Georgian and Turkish structures, and it should be noted that there are significant architectural disparities; in particular, the doors in the Turkish tunnels, while also made from a single large stone slab, are disc-shaped and employ a different – sliding/rolling – mechanism to open and close. In addition, there can be several stone doors in each complex,30 while in the darnebi, in our experience, there is only one. Nevertheless, commentators agree that the Turkish structures were also used as temporary refuges, and that this was the case even as early as the 9th century.31 Conclusions and Future Research In summary, we argue that the darnebi first came into use as refuges between the 8th and 11th centuries, at approximately the same time as caves in the Kura and Paravani valleys that share some of their architectural features. The evidence from medieval texts and the broader historical record indicates that the tail end of this date range is more likely, but there is some support for earlier dates as well. We also propose that the most likely period of later re-use was the 16th–17th centuries, but that the 13th and 14th centuries are also possible. There are several promising directions for future research on the darnebi. First, and as mentioned above, while we are confident that our list of examples is broadly representative of the whole, it is not exhaustive. There would be much to be gained by using geophysical techniques to determine more precisely the extent of the darnebi in the region and to better understand the layouts of individual examples. Second, further investigation of possible regional parallels, including in Cappadocia, is warranted. Finally, while the darnebi are always associated with settlement remains, some of which have datable features, the physical links between the buildings on the surface and the structures underground are by now too disturbed to draw any conclusions about how they relate to each other in time. More excavations may in the future bring clarification. If the dates suggested here for the darnebi are correct, they have the potential to shed light on the status of south-west Georgia as a frontier and a buffer between the powerful empires – Arab, Byzantine, Persian, Mongol and Turkish – that dominated the centuries in question. So-called natural borders, especially mountains and rivers, seem seldom in the ancient world to have been

27

28 29

30 31

It should also be noted that it is possible the darnebi were from time to time used to store food supplies. Inside, there is a constant temperature that provides excellent conditions for such a purpose, although perhaps one would expect to find more remnants of storage vessels. Ibn Hawqal, 194–195. The chapter in which Ibn Hawqal discusses the Byzantine Empire deals specifically with nothing north-east of Lake Van, and his chapter on the Caucasus (‘L’Arménie, L’Azerbaidjan et L’Arran’) with nothing south-west of Tbilisi. Dawkins 1916, 15–16. Dawkins 1916, 16–17; Rodley 1985, 6–7.

128

Abby Robinson – Giorgi Khaburzania

perceived, in themselves, as impenetrable barriers, or to have acted as such.32 In fact, what we see in the darnebi and the caves is some continuity on both sides of the Kura River as well as between the highlands and the valleys. Where the terrain does form a genuine obstacle to the progress of invaders, it is because it has been shaped by its inhabitants to successfully meet their defensive needs. The darnebi are an instance of a phenomenon, also observed by us in other historical periods in this region,33 whereby the inhabitants used finely tuned knowledge of the local geography, and their skill in adapting it, to survive the onslaughts of multiple adversaries and hold their ground. As such, this chapter concerns landscape archaeology, the overarching theme of this volume, in a very fundamental way: the darnebi are evidence of terrain and people indivisibly linked through centuries in this much-contested space. Acknowledgements: We wish to express our sincere gratitude for the support and guidance of Tony Sagona, without whom our project – among so many others – would not have been possible; his loss is incalculable. We are also much indebted to Claudia Sagona and Kakha Kakhiani. We thank David Lordkipanidze of the Georgian National Museum and Zurab Makharadze of the Otar Lordkipanidze Centre for Archaeology in Tbilisi for permission to conduct fieldwork in Samtskhe-Javakheti. Our project was also made possible by a University of Melbourne Faculty of Arts fieldwork funding grant and Australian Government Research Training Program Scholarship.

References Berdzenishvili et al. 2000 დ. ბერძენიშვილი – ი. ელიზბარაშვილი – ნ. ვაჩეიშვილი – ც. ჩაჩხუნაშვილი / D. Berdzenishvili – I. Elizbarashvili – N. Vacheishvili – T. Chachkhunashvili, ჯავახეთი. ისტორიულ-ხუროთმოძღვრული გზამკვლევი / Javakheti. Historical-Architectural Guidebook (Tbilisi 2000). Burdiladze 2010 ს. ბურდილაძე / S. Burdiladze, ლეფისის დარანი / Lepisis Darani (Darani of Lepisi. Expedition report of 2007), ძიებანი / Dziebani 19, 2010, 74–82 [in Georgian with English summary]. Cherry et al. 1991 J. F. Cherry – J. L. Davis – E. Mantzourani, The problem orientation of the survey, in: J. F. Cherry – J. L. Davis – E. Mantzourani (eds.), Landscape Archaeology as Long-Term History. Northern Keos in the Cycladic Islands from Earliest Settlement until Modern Times (Los Angeles 1991) 3–12. Dawkins 1916 R. M. Dawkins, Modern Greek in Asia Minor. A Study of the Dialects of Sílli, Cappadocia and Phárasa with Grammar, Texts, Translation and Glossary (Cambridge 1916). Gaprindashvili 1957 გ. გაფრინდაშვილი / G. Gaprindashvili, ვარძიისა და მისი მიდამოების ქვაბები / Vardziisa da misi midamoebis qvabebi (Caves of Vardzia and its Surroundings) (Tbilisi 1957). Gaprindashvili 1959 გ. გაფრინდაშვილი / G. Gaprindashvili, კლდის სახლები ნასოფლარ ფიაში / Kldis sakhlebi nasoplar Piashi (Rock Houses at the Settlement of Pia) (Tbilisi 1959). Gaprindashvili 2002/2003 გ. გაფრინდაშვილი / G. Gaprindashvili, მირაშხნის გამოქვაბული და მისი წარწერა / Mirashkhnis gamoqvabuli da misi Tsartsera (A cave of Mirashkhani and its inscription), in: პ. გაფრინდაშვილი / P. Gaprindashvili (ed.), წელიწდეული / Tselitsdeuli (Yearbook) VI–VII, 2002/2003, 47–84.

32 33

Parker 2002, 373, 377–380; Robinson 2014. The current authors (Khaburzania – Robinson 2018) also observe this phenomenon in the ostensibly Late Bronze– Early Iron Age fortresses that dot the study area; their locations appear to have been deliberately chosen to maximise their defensive and strategic value.

Medieval Underground Shelters (Darnebi) of South-west Georgia

129

Hasan-i-Rumlu A Chronicle of the Early Ṣafawīs (Ahsanu’t-Tawarikh), Vol. 2, transl. C. N. Seddon (Baroda 1934). Ibn Hawqal Ibn Ḥawqal, Configuration de la Terre (Kitāb Sūrat al-Ard), transl. J. H. Kramers and G. Wiet (Beirut 1964). Japaridze et al. 1981 ო.ჯაფარიძე – ი.კიკვიძე – გ. ავალიშვილი – ა. წერეთელი / O. Japaridze – I. Kikvidze – G. Avalishvili – A. Tsereteli, მესხეთ-ჯავახეთის არქეოლოგიური ექსპედიციის მუშაობის შედეგები / Meskhet-Javakhetis arqeologiuri eqspediciis mushaobis shedegebi (Meskheti-Javakheti Archaeological Expedition Results) (Tbilisi 1981). Khaburzania – Robinson 2018 G. Khaburzania – A. Robinson, Nature and local knowledge combine in powerful strongholds. Preliminary findings of the Archaeological Survey in Samtskhe-Javakheti project, in: A. Batmaz – G. Bedianashvili – A. Michalewicz – A. Robinson (eds.), Context and Connection. Essays on the Archaeology of the Ancient Near East in Honour of Antonio Sagona (Leuven 2018) 303–324. Matthew of Edessa The Chronicle of Matthew of Edessa, transl. A. Dostourian (Ann Arbor, Michigan 1972). Parker 2002 B. J. Parker, At the edge of empire. Conceptualizing Assyria’s Anatolian frontier ca. 700 BC, Journal of Anthropological Archaeology 21, 2002, 371–395. Robinson 2014 A. J. Robinson, Contested and common ground. Geography and history at the limits of the early Islamic conquests, Ancient Near Eastern Studies 51, 2014, 317–340. Rodley 1985 L. Rodley, Cave Monasteries of Byzantine Cappadocia (Cambridge 1985). Sagona – Birkett-Rees 2016 A. Sagona – J. Birkett-Rees, Battlefield archaeology. Gallipoli, in: A. Sagona – M. Atabay – C. Mackie – R. Reid – I. McGibbon (eds.), The Anzac Battlefield. Gallipoli Landscape of War and Memory (Cambridge 2016) 83–97. Salia 1983 K. Salia, History of the Georgian Nation (Paris 1983). Tatishvili 1970 თ. ტატიშვილი / T. Tatishvili, მიწისქვეშა ნაგებობანი მტკვრის ხეობის ზემო წელზე / Mitsiskvesha nagebobani mtkvris kheobis zemo tselze (Underground shelters in the upper part of the Kura Gorge), ძეგლის მეგობარი / Dzeglis Megobari 21, 1970, 18–22. Thomson 1996a The Book of K’art’li, in: Rewriting Caucasian History. The Medieval Armenian Adaptation of the Georgian Chronicles. Translated with Introduction and Commentary by R. W. Thomson (Oxford 1996) 255–308. Thomson 1996b History of David, King of Kings, in: Rewriting Caucasian History. The Medieval Armenian Adaptation of the Georgian Chronicles. Translated with Introduction and Commentary by R. W. Thomson (Oxford 1996) 309–353.

Approaching Landscapes of Infrastructure: Methods and Results of the Vayots Dzor Silk Road Survey Kathryn J. Franklin * – Astghik Babajanyan ** Abstract: Within this chapter we will lay out a discussion of why landscape-scale archaeological research is so crucial to scholarship moving forward, particularly focussing on high and late medieval (12th–15th centuries AD) Silk Road heritage within the Republic of Armenia. We will provide a brief overview of how the methods and research priorities of the first seasons of the Vayots Dzor Silk Road Survey (VDSRS) emerged from historical data pertaining to that landscape, as well as perceived obligations to heritage management concerns at the local and institute level. Ultimately, this chapter will attempt a preliminary synthesis of the VDSRS data, with the aim in mind of (re)characterising the Vayots Dzor section of the Silk Road Corridor as an object of study both in terms of its particular history and also with an awareness of the contemporary relevance of archaeological research in this region. Keywords: Medieval landscape; Armenia; Silk Road Heritage; Infrastructure

Introduction This chapter presents a preliminary introduction to Armenia’s Silk Road Corridor, a complex palimpsest of archaeological landscapes running roughly east–west across the regions of Vayots Dzor and Syunik. For two seasons of research, the sections of the Silk Road Corridor located in Vayots Dzor (Vayots Canyon) have been the focus of the Vayots Dzor Silk Road Survey (VDSRS), a newly-formed collaborative project aimed particularly at researching the ways that the historical landscapes of Vayots Dzor mediated local participation in regional political interactions and long-distance trade and travel during the high and late medieval period1 (AD 1100–1400). In their local participation, Armenians in Vayots Dzor constituted a microcosm within the broader world of inter-Eurasian travel and trade, or what is commonly referred to in the current context as the medieval “Silk Road”. The research priorities of the VDSRS’ early seasons have been a synthesis of existing sources on the Vayots Dzor’s landscape and linking extant resources on sites and monuments to spatial data through the use of survey and GIS. The combination of non-systematic site based survey, surface collection and open source mapping provides the basis for a progressing collaborative project which entails data sharing between the project and scholars within Armenia (and elsewhere) through the use of a map form online database hosted by the Harvard WorldMap. Through the application of methodologies that are revolutionary to medieval archaeology in Armenia, the project has constructed a landscape approach, which acts as a framework for our ongoing excavations and our developing understanding of Vayots Dzor’s participation in the Silk Road cultural ecumene.



Oriental Institute, University of Chicago, [email protected]. National Academy of Sciences, Institute of Archaeology and Ethnography, Yerevan, [email protected]. 1 In Armenian history and historical archaeology, this period is designated the “developed medieval period” (զարգացած միջնադար (9th–14th centuries AD), dividing into two broad phases – the Bagratid kingdom (9th–11th centuries) and the “Zakarid period” (late 12th to the eve of the 14th century). For clarity and comparability with other regions, we will use the terms high (12th–14th centuries AD) and late (14th–15th centuries AD) in this chapter. *

**

132

Kathryn J. Franklin – Astghik Babajanyan

Survey and the Silk Road Applying survey as a methodology paired with historical research and targeted stratigraphic excavations is critically important for developing a more nuanced historical and material understanding of the complex cultural ecumene which is now referred to as the medieval Silk Road, within which cultural world localised communities participated with differing degrees of agency and reach. The VDSRS is particularly committed to challenging linear, ‘superhighway’ models of Silk Road complexity which feed back into the past from discourses in contemporary economics and development jargon as well as modernist takes on the anthropology of travel, and which focus on nodes of civilisation separated by long stretches of ‘non-place’. Of course this image of the Silk Road is, in part, a gift of the most famous late medieval and early modern narrators of travel in Eurasia: writers such as Marco Polo or Ibn Battuta created senses of time and distance in their writing by interspersing boilerplate passages describing the ‘desert wastes’ they passed through between populated cities.2 While this technique effectively conveys the real tedium of travelling by cart for months at a time, it also over-emphasises the nodal points of what was in actuality a much more distributed social landscape. Problematising the constrained point-A-to-point-B linearity of Silk Road landscapes is also of high relevance to the archaeology of the Armenian highlands, the historical polities of which have long been characterised as nodal or networked in direct comparison with the lowland default;3 the kingdoms of medieval (5th–15th centuries AD) Armenia were evoked by historians and archaeologists as places-in-between, galvanised and sustained by transit trade along roads connecting far-away cities in the Mediterranean, Near East and China. This conceptualisation holds true for the particular landscape under discussion in this chapter and the focus of our research: the canyons, valleys and mountains of southern Armenia. From an archaeological perspective, Vayots Dzor and neighbouring Syunik regions are studded with monumental sites from the medieval and early modern periods that trace a dotted line along the Arpa and Vorotan rivers, punctuated in the west by the monastery at Noravank and in the east by the equally picturesque Tatev monastery.4 Bridges and caravanserai (road inns) – including the caravanserai built in AD 1332 by the Orbelyan prince Česar at Selim5 and the ruins of a caravanserai built by another Orbelyan prince at Harjis in AD 1343, eastward almost as far as Goris6 – fill in the gaps in the line between monastic sites. This canalised, east–west roadway is, not coincidentally, both the Republic of Armenia’s Southern Heritage Corridor and also the route of the southernmost branch of the Armenian North–South Highway, planned for completion (according to one source) by 2020.7 A corridor concept of medieval heritage thus falls in line with the Republic of Armenia’s plans to revitalise the national trade economy through tariffs on transit trade between Iran and the Black Sea, recapturing the glory of the mountain country’s privileged nodal position on the medieval Silk Road.8 But development strategies focussed on the singular roadway neglect the richer and more extensive medieval landscape that extends through regions such as Vayots Dzor and thus also narrow the possibilities for narratives that can be told about Armenia’s past (and future) mediated by that archaeological heritage. A major impetus of the VDSRS long-term project

2 3 4

5 6

7 8

Gibb 1958; Polo 1904. Cf. Hammer 2014, 758. Of course, the strategic and social significance of Vayots Dzor extends back as far as the first millennium BC, when the region was marked by a closely-spaced net of fortresses and fortified settlements which includes parts of the present settlements of Guetap-Yeghegnadzor-Agarakadzor and the long-abandoned settlement of Aghasap; cf. Melkonyan 2003. Undoubtedly, this defensive system was also intended to monitor interregional routes of transit: see for instance Earley-Spadoni 2015; but also Hammer 2014, 769. Barxudaryan 1967, 177–178. Harut’yunyan 1960, 17–26, see also p. 7 for Harut’yunyan’s map of medieval trade routes; for the Harjis caravanserai, see Barxudaryan 1960, 82; Harut’yunyan 1960, 26–28. News.am 2015. See for instance World Bank 2015.

Approaching Landscapes of Infrastructure: Methods and Results of the Vayots Dzor Silk Road Survey

133

is to enrich our understanding of the high and late medieval landscape and thus similarly diversify possibilities for the stories it can help to tell.9 One story, which will be developed here, is of the intensive investments in the civic landscape which accompanied the florescence of trade through Vayots Dzor in the high medieval period – we thus propose a shift from a nodal to a network or infrastructural way of thinking about Silk Road heritage. The Region of Vayots Dzor and its History Vayots Dzor (also Vayoy Dzor, Vayo Dzor, Vayi Dzor, Vau Dzor) is the region (gavar) of the historical province (nahang) of Syunik. The region is surrounded by high mountains and watershed chains that act as natural dams, creating a geographical unity. According to medieval legend the name of Vayots Dzor derived from a combination of the words vay (cry, wail) and dzor (canyon)10 but the etymology may have deeper roots. According to 20th-century Armenian scholars, the region’s name arose from the name of the Urartian tribe vay, which settled in the territory of Vayots Dzor in the first millennium BC.11 In the periods of Persian and Russian domination of the Armenian highlands (17th–20th centuries), Vayots Dzor, by political division, became the region of Sharur-Daralagyaz or the south-eastern part of Yerevan province.12 In historical sources Vayots Dzor was first mentioned in the 5th century, in the context of the anti-Sasanian rebellions of Avarayr.13 From the 5th to the end of the 7th century, information related to the province is rare; however, the situation changes and political life becomes more historically visible from the 9th century onwards when the Vasakyan princes of Syunik transferred their residence from Tsghuk region to the Yeghegis burgh (gyułaqałaq) of Vayots Dzor. In the 9th–11th centuries a complicated political situation was sustained in Vayots Dzor, as the result of the first feudal conflicts and then Arab invasions. Similarly, in Vayots Dzor, as in the rest of Armenia, there is a dearth of historical sources about the 11th to late 12th century, when the region was under the rule of the Seljuks. The role of the Vayots Dzor region as the centre for social, political movements and the economic and cultural development of Syunik increased at the beginning of the 13th century (the onset of the developed medieval period) when the Proshyan and Orbelyan noble families appeared in the arena of history and settled in Vayots Dzor. The Orbelyans in particular played a significant role in political, religious and cultural life not only in Vayots Dzor but also in Armenia in the 13th–15th centuries. Unlike princely houses such as the Zakaryans/Mxargrcelis (Yerkaynabazuk), who ruled in Georgian-administered Aragatsotn and Shirak, the Orbelyans took a flexible, pragmatic political stance in their relationships with the Mongols (who ruled Armenia from the mid-13th century) and were thus able to maintain a degree of internal independence.14 The Orbelyans acquired authority as the Mongol Ilkhans and governors contributed to the reinforcement of the military and economic power of the principality of Syunik. The legal and political influence of the Orbelyans, especially under the princes Smbat and Tarsayich, extended from the mountains of Bargushat to

9

10 11 12 13 14

In applying a survey strategy to the landscape of southern Armenia we follow on the work of the Vorotan Project in the Syunik Region, cf. Alcock et al. 2006. We also draw on the work of colleagues studying the Paleolithic, Bronze and Iron Age landscapes of Vayots Dzor and Syunik: see for example Gasparyan – Arimura 2014; we also work just across the border from the Naxçivan Archaeological Project: see Ristvet et al 2011. Kaghankatuatc’i 1969, 252; Ganjakec’i 1982, 64; Orbelyan 1986, 70. Harut’yunyan 1967, 91–96. Lalayan 1904, 237. Yełiše 1958, 178. Significantly, the Orbelyans took advantage of inju law i.e. they were dependent directly on the Mongol Khan and were not vassals to local governors. Significantly, the Orbelyans took advantage of Ilkhanid inju status, according to which they were dependent directly on the Mongol Khan and were not vassals to local governors, specifically the Georgian Bagratids. See Bedrosian 1979, 184–185.

134

Kathryn J. Franklin – Astghik Babajanyan

Garni, Dvin and Bjni. Although the Orbelyan principality split into three branches (Orbelyan or Burtelyan, Liparityan and Jalalyan) after the death of Tarsayich, these ruling families managed to maintain an internal independent state up to the last quarter of the 14th century, which is illustrated by their ambitious and flamboyant construction projects. The Orbelyans built and reconstructed fortresses, civic buildings (residences, road-inns, markets), spiritual centres (churches, monasteries with attached schools and educational institutions), production facilities (mills, oil mills) and craft centres. They paid particular attention to roads and bridges, as well as to the irrigation systems and channels needed to bring new lands under cultivation. In the 13th–14th centuries the best educational centres of that time were created in Vayots Dzor: Gladzor and Hermon universities, as well as dozens of scholarly centres (in Tsaghatskar, Arates, Shativank, Verin Noravank and other monasteries). Thus the high medieval archaeological landscape of Vayots Dzor is to an enormous extent an imprint of the Orbelyans’ infrastructural legacy. Methods and Aims of the Vayots Dzor Silk Road Survey The VDSRS methodology is oriented towards unifying new with existing datasets in order to generate a fuller understanding of the Vayots Dzor Silk Road landscape, to support both future research at the landscape scale and more effective management of the archaeological heritage in this region. Our work in the 2015 and 2016 seasons focussed on directed survey of listed sites and their surrounding regions, and intensive surface collection and excavation at the site of Arpa, a central place in the late medieval history of Vayots Dzor and a tethering point for our material and spatial systematics. The Written Landscape: Textual Resources for Survey in Vayots Dzor A primary consideration in conceptualising the medieval archaeological landscape in Armenia, both as a subject of research and as a heritage resource, is that this landscape is overlaid with the historical imaginaries of centuries of textual references and descriptions from chroniclers, travellers and other literary voices. In the case of Vayots Dzor, the most prominent high medieval voice is that of the 13th-century historian Stepanos Orbelyan, whose History of the Province of Syunik describes the period from the second half of the 1st century to the 13th century AD.15 As an historical source, Orbelyan’s work is important because the author used not only documents whose originals are now lost, but also provides a (not always objective) record of inscriptions which have partially or entirely disappeared as the buildings on which they were carved decayed or were destroyed.16 Our second major body of sources consists of travellers’ accounts of Armenia and the surrounding regions both in the high medieval period and in the subsequent centuries. Sargis Jalalyants was the first known traveller who noted the monuments of Vayots Dzor and published a number of architectural inscriptions in his mid-19th-century work.17 Similarly, the 1892 Travelers Observations (Djanaparhordakan nkatołut’yunner) of Kajberuni is an inestimable resource for the study of the history of the villages and monasteries of Vayots Dzor as well as the lifeways, traditions and legends of Armenians.18 Drawing upon Armenian and foreign historians, Kajberuni confirmed (or speculated upon) the locations of numerous settlements and monuments and their historical names, which had inevitably been distorted over the 17th–19th centuries.

15 16

17 18

Orbelyan 1986. The epigraphic record of Vayots Dzor is itself also of course a major resource: see Barxudaryan 1960; Barxudaryan 1967. Jalalyanc’ 1858. Kajberuni 2003.

Approaching Landscapes of Infrastructure: Methods and Results of the Vayots Dzor Silk Road Survey

135

Finally, a primary resource used by the project is a list (the Hušarjanneri tełekatvakan štemaran or Monuments Database) of archaeological sites and monuments maintained by the HistoricalCultural Heritage Research Centre.19 This list has been made available online and is searchable by marz as well as by site type and ‘significance’ at the Ministry website. The database provides descriptive information and some locational information on settlements, cemeteries, fortresses, churches, burial fields and other forms of standing monuments (bridges, mills, etc.) as well as dating information. However, the locational information provided by the database is broadly narrative in style and in many cases is quite vague. Thus the database is primarily of use to those few members of the Ministry of Culture and Institute of Archaeology who have visited sites before and who know their locations. For archaeologists or other researchers – including heritage management officials working in Vayots Dzor – the database is not helpful in locating these sites in an effective way, and the list is only marginally useful to students trying to understand the state of research and knowledge of the archaeological landscape in Vayots Dzor (or elsewhere in Armenia). Another challenge to be faced when using the Monuments Database for archaeological research is the way that chronological information is provided in the list. While sites and monuments (such as xačkars or cross-stones) are dated – sometimes to the year – no source is provided for these dates.20 One of our primary aims in carrying out directed survey in Vayots Dzor is therefore to improve the accuracy and usefulness of available knowledge about listed archaeological heritage in the region. To accomplish this, we set out to augment the narrative descriptions of sites and monuments contained in the Database with coordinate data and to test the periodisation assignments of the list with surface collections where possible. Our survey sectors were laid out along the Arpa and Yeghegis rivers as well as some of the primary tributary canyons. We surveyed areas on the mountain slopes to the north and south of the Arpa River, and to the east and west of the Yeghegis and its tributaries. In addition to the historical sources mentioned above, the survey drew on local knowledge among the inhabitants of villages along the route, and on site and place name data recorded in Soviet 1:100,000 and 1:25,000 scale topographic maps of the survey region. Preliminary Results of the 2015–2016 VDSRS Survey Seasons The first two seasons of the VDSRS Project generated an array of spatial data providing information on the archaeological landscape of the Vayots Dzor region. This data forms the framework within which the Project will continue to organise and understand our material (artefacts), historical (inscriptions and primary sources) and other datasets (topographic survey, aerial and satellite imagery) related to archaeology in Vayots Dzor. While we maintain datasets related to our specific research questions (ceramic and botanical analyses, etc.), a version of our coordinate data and related chronological information is stored and displayed in a spatial database hosted by Harvard WorldMap. This open source platform supports the project’s commitment to providing open data from the beginning: our site locations and other forms of data are available to students in Armenia and other scholars who want to continue work in the region. Thus, while our data gathered so far is primarily connected to the medieval and early modern periods (roughly the 10th–18th centuries AD),

19

20

The Patmamšakut’ayin Zharangut’yan Gitahetazot’akan Kentron (Պատմամշակութային ժառանգության գիտահետազոտական կենտրոն), which is the research arm of the Republic of Armenia Ministry of Culture. More information at: http://www.armmonuments.am/about.html (last accessed 10 Jan. 2018). The lack of citations for dating information for sites is a challenge for all periods but is especially frustrating for the medieval period, wherein sites/buildings may be dated through a whole range of methods, from historical references, epigraphic evidence, architectural style or material collected through surface collections, chance finds or excavations. Each of these methods of dating is subject to its own particular standard, and the lack of citation is especially unproductive for research when the dating methods applied to one class of remain (engraved burial markers or xačkars) is tacitly used to date another class (a church, an entire settlement).

136

Kathryn J. Franklin – Astghik Babajanyan

Fig. 1   Map showing the distribution of sites, monuments and features recorded by the VDSRS. For a scalable, interactive map, site information and images, visit our open-source map database at (map: K. Franklin)

Approaching Landscapes of Infrastructure: Methods and Results of the Vayots Dzor Silk Road Survey

137

we are hopeful that as our work progresses, our colleagues will augment our collaborative map with data from other periods, filling in the spatiotemporal gaps. The full array of sites recorded in the last two seasons are shown in Fig. 1 and are available to explore on our online map database:21 in the following sections we will discuss a few key site types and patterns which we observed in the course of our work. Settlements Medieval settlement sites or ‘dwelling places’ (bnakateł) are located in river valleys, on canyon floors and high mountain slopes. While some of the settlement sites located by the project contain standing architecture or the visible depressions which indicate locations of rooms, a considerable number had been significantly destroyed by Soviet-era field amelioration (discussed below): these sites are detectable as curved berms of rubble at the edges of fields or of cliffs overlooking ravines, into which the bulk of the material from the village was dumped using tractors. According to the pattern of field clearance that we observed in Vayots Dzor, most frequently all remains of a village would be cleared except for stones or architecture apparently related to churches or cemeteries: these stones may, however, be pulled to the far margin of the cleared area. Of this latter type, key examples include the village sites at Ulgyur (on the slopes above Ałavnadzor) and Łešlał settlement (between Shatin and Hors). The church and cemetery at Ulgyur are well known as the location of the burial of the kin of the high medieval architect, Momik; otherwise the village exists only as rubble at the rim of the canyon. The location of Łešlał was marked as ruins on a Soviet survey map dated to 1964 but is visible currently as a line of rubble and a single, 1m × 1m stone plinth, carved with an image of the Holy Family. Other settlement sites are much better preserved, due to being located on slopes or locales where amelioration was not carried out. Prime examples include the high standing walls preserved at the settlement of Hostun on the Yeghegis River (Fig. 2), extensive settlement remains preserved at the site of Arpa (the medieval component of contemporary Areni), as well as the eroded but preserved terraced architecture at the site of Anapat, in the mountains above Agarakadzor. Ceramics collected from these sites will be discussed in future publications but demonstrate a remarkable regional consistency, especially manifested in a high-gloss red burnished ware, locally called ‘medieval-Urartu’ due to its resemblance to Iron Age fine wares from Vayots Dzor (see Fig. 3). The distribution of settlement sites demonstrates that during the high and late medieval periods, settlements in Vayots Dzor not only lined the major routes along river valleys but were also spread in a network through the mountains, connected by small trails and associated with water sources tucked back in the canyons. Fortresses Forts and fortresses (berd or amroc’) are distributed throughout the survey area, includ-

21

Fig. 2   High – late medieval Hostun settlement, on the Yeghegis River, photographed from the north-west (photo: VDSRS)

Visit our map and explore our data further at (last accessed 10 Jan. 2018).

138

Kathryn J. Franklin – Astghik Babajanyan

Fig. 3   A selection of the diagnostic high medieval ceramic material collected by the VDSRS. Note the pronounced carination and burnished red surface treatments of the ‘medieval-Urartian’ type found at contemporary sites throughout the survey area (graphics: K. Franklin and A. Babajanyan)

ing sites with clear architectural fortifications and locales that take advantage of the ‘natural fortifications’ of Vayots Dzor’s dramatic geomorphology to observe key points in the river routes. The most famous late medieval fortress in the survey region is the seat of the Orbelyans at Smbataberd (now rebuilt as a tourist attraction). Our results indicate a more dispersed landscape of surveillance and control (though control by what means remains to be further explored) in Vayots Dzor. We recorded architectural remains and collected developed medieval ceramics at a number of fortress sites perched atop pillars of natural canyon rock, including Ertich (Fig. 4) and Hrasekaberd, both along the Arpa River, and Berdakar, which overlooks the Yeghegis River just south of Shativank.

Fig. 4   A view of the high medieval fortress at Ertich, photographed from the south. Surface materials and architectural remains crown the natural outcropping and line the slopes below (photo: VDSRS)

Approaching Landscapes of Infrastructure: Methods and Results of the Vayots Dzor Silk Road Survey

139

Monastic Sites Monasteries (vank’) are defined as sites which incorporate settlement components but which were constructed primarily for the housing of a monastic population and which include a church context. As was the case elsewhere in the medieval period, monasteries in Armenia played a complex social role, being variously locales for withdrawal from worldly concerns but simultaneously loci for hospitality and economic production, endowed with lands, mills, workshops, oil presses and glassworks, etc. The array of monastic sites recorded by the VDSRS demonstrates this multivalence in the medieval monastery: for example, sites such as Aghvank (13th–14th century) and Shativank (10th–17th century), both in the Yeghegis River valley, are situated on steep cliff-sides, reached by winding paths, yet excavations in the last century at Shativank (Fig. 5) revealed several different workshop areas, indicating that the site was no remote hermitage cut off from society. Similarly, the church and associated settlement of Hostun (Vostink) are perched on the edge of the Yeghegis River north-east of the modern village of Shatin: recent excavations in preparation for the laying of a water pipeline through the core of the site generated a spectacular array of (unpublished) materials demonstrating the connection of this site to Fig. 5   A view of the high – late medieval monastery of Silk Road trade networks in the high and Shativank, from the south. Note the overgrown remains of late medieval periods.22 outbuildings around the central church (photo: VDSRS) Caravanserai and Guesthouses While the material record from excavations in the Republic of Armenia testifies to the active travel and trade through the highlands during the developed medieval period,23 the construction of caravanserai or road inns (karavanatun or ijevanatun) by aristocratic patrons demonstrates the infrastructural investment in Silk Road travel at an administrative level in Armenia during the periods of the Armenian-Georgian united kingdom and Ilkhanate. While the 14th-century Selim Pass caravanserai is already well known, the VDSRS recorded the location for a contemporary caravan inn at Harjis as well as one for a late-13th-century guest house associated with the monastery at Noravank along with its foundation inscription (now moved inside the partially-rebuilt monastic complex). Bridges A critical part of the landscape of long-term mobility through Vayots Dzor, bridges (kamurj) mark the locations of medieval river crossings, points of administrative intervention in travel and, in their differential states of preservation, the ways that travel directionality has shifted

22 23

F. Babayan, personal communication. For material discussions of medieval trade through and within Armenia, see: Arakelyan 1964, 31, 40–41, 46–53; Kalantaryan et al. 2009.

140

Kathryn J. Franklin – Astghik Babajanyan

over the last several centuries. For instance, while the VDSRS recorded late medieval bridges that are still in use, such as the 13th-century Dadali Bridge near Agarakadzor (Fig. 6), we also found and recorded ruined fragments of the 13th-century bridge at Arpa and the 17th-century (AD 1666) C’aturi Bridge south of Shatin. Bridges, especially those that are still in use, represent a major way that the medieval landscape continues to be significant for the inhabitants of Vayots Dzor, being places where the past and present literally cross paths. A great example of this is the late medieval bridge at Chubuk Kyorpi (Shatin), which is also the location of reerected xačkars documenting the significance of Shatin village in the 13th century.

Fig. 6   The 13th-century Dadali Bridge over the Arpa River, west of Agarakadzor (photo: VDSRS)

Maturs A final word must be said on a particular category of site found not only in Vayots Dzor, but all over the Republic of Armenia: small hand-built shrines, called matur, which contain one or several xačkar and which are often dedicated to the quasi-pagan cult of T’ux Manuk. Maturs represent the continuing significance of the (Christian) medieval built landscape in contemporary practice, as both the xačkar and stones within the walls of the shrines may have been carried from cemeteries, churches, or other medieval contexts. The devotional practices carried out within maturs – tying of votive rags or leaving of icons and other votive objects, lighting of candles and the performing of matał (animal sacrifice) – are also often performed at other medieval structures. Discussion: Challenges to Archaeological Survey – Landscapes of ‘Destruction’ and of Infrastructure As is apparent in the above discussion, the network of endowments and investments in travel made by the Orbelyans was not the only infrastructural landscape encountered by the VDSRS. Soviet utopian modernist24 projects in Vayots Dzor (and other regions of Armenia) shaped the landscape in dramatic ways that must be incorporated into – or rather grappled with by – survey strategies. In the 1960s and 70s, agriculture in the Soviet Union was progressively industrialised (modernised), with the intent to integrate agriculture with factory-based modes of production such as manufacturing.25 A central technique of agricultural modernisation of critical relevance to archaeological work in Armenia (and other post-Soviet landscapes) is soil amelioration or the improvement of natural (or perceived as natural) soil sediments for agricultural purposes by graded terracing of slopes; as we noted above, dealing with the taphonomic impact of these monumental technical processes was a daily problem-solving task on the VDSRS. Cherry discussed the danger of approaching the ameliorated valleys and mountain slopes of southern Armenia as ‘damaged,’ as this implies that prior to Soviet-era interventions the land-

24

25

A brief review of these projects as applied in particular to agricultural soil amelioration is provided by Lydolph 1992. Cf. Wädekin – Jacobs 1982, 240.

Approaching Landscapes of Infrastructure: Methods and Results of the Vayots Dzor Silk Road Survey

141

scape of places such as Vayots Dzor was somehow more pristine or more ‘true.’26 Perhaps more important to a landscape approach to archaeology in Armenia is understanding the ways in which this intensive Soviet landscape-shaping has been in many ways naturalised in the decades since its implementation, to the point that in many places Soviet and medieval infrastructural ruins vie for domination of the landscape of memory in places such as Vayots Dzor. A fascinating aspect of the mutual construction of space and power is that the marks of regimes in landscapes such as Vayots Dzor can cut so deep and yet work so subtly on affect and the telling of the past. Critically, a perception of the medieval landscape in rural Armenia – lonely monasteries perched atop grassy and apparently empty hillsides – is a product not only of medieval ideas about landscape, but also of the intensive production and later dissolution of the Soviet infrastructural landscape. In the contingent way that these produced spaces fade in and out of view, they operate in a way understood as part of the dynamism and also unpredictability of infrastructures as mediators of social and political life.27 The material contingency of Orbelyan and Soviet ruins brings home one of our major early observations, which we hope to develop further in the future, that global polities and large-scale phenomena such as ‘the Silk Road’ are constructed through infrastructural interventions at the local level – and these infra-structures persist long past the duration of the polities that constructed them (barely a century, in the case of both the Ilkhanids and the Soviet Union) and continue to constitute a complicated heritage landscape. Conclusion: Landscapes of Infrastructure, Routes of Movement The first seasons of the VDSRS have provided a promising basis not only for shifting the longstanding paradigms of medieval archaeology and approaches to Armenia, but also for solid ongoing research projects and collaborations within the intensely rich archaeological landscape of Vayots Dzor. In this and our future work we move towards replacing the idea of the high medieval Silk Road in Armenia as a ‘highway’ that runs along a narrow route to a landscape that is built over time and through the intersecting projects of local administrators, regional states and empires, and global shifts in trade and geopolitics. The survey data shows that Vayots Dzor was not merely a ‘place in-between’ but a place in itself, marked by princely building projects but also populated by people living out their lives along the roads and on the mountain slopes. It is interesting to contrast an ‘infrastructural’ or network approach to medieval life in Vayots Dzor with the ‘superhighway’ image conjured by contemporary development investment strategies. During the 13th and 14th centuries the Orbelyans as administrators for the Ilkhanids did ‘invest in infrastructure’ in ways that might seem quite modern and similar to the heavy investment of the current Armenian government and eastern banks. But if we take a landscape approach to medieval travel, we see that the bridges and caravanserai were part of a larger infrastructural landscape that also contained monastic centres, towns and villages. What this reveals is not that medieval politics was pragmatically concerned with infrastructure in the modern sense, but that we should incorporate the concerns of everyday life, comfort and maintenance which are hidden within infrastructure as a concept into research into the medieval Silk Road as a culture of practice. Acknowledgements: This research was carried out with generous support from a Collaborative Heritage Management Grant from the American Research Institute of the South Caucasus and from a Transregional Research Junior Scholar Fellowship (InterAsian Contexts and Connections) granted by the Social Sciences Research Council. Of the numerous individuals and organisations which helped us start this ongoing project, we would like to expressly thank Dr Pavel Avetisyan of the NAS Institute of Archaeology and Ethnography and Davit Davtyan who participated in the survey. Thanks also to Davit Simonyan, Levon Aghikyan and Dr Hannah Chazin.

26 27

Cherry 2014, 208–209. For a review of anthropological approaches to infrastructure, see Larkin 2013.

142

Kathryn J. Franklin – Astghik Babajanyan

References Alcock et al. 2006 S. Alcock – M. H. Zardaryan – A. V. Tonikyan – J. F. Cherry, The Vorotan Project, Armenia, 2005, in: S. Alcock – L. E. Talay (eds.), In the Field. The Archaeological Expeditions of the Kelsey Museum, Kelsey Museum Publications 4 (Ann Arbor 2006). Arakelyan 1964 Բ. Ն. Առաքելյան / B. N. Arakelyan, Քաղաքները և արհեստերը Հայաստանում IX–XIII դարերում 2 / Kałaknerə yev Arhestnerə Hayastanum IX–XIII dd. (Cities and Crafts in Armenia in the IX–XIII centuries) (Yerevan 1964). Barxudaryan 1960 Ս. Գ. Բարխուդարյան / S. G. Barxudaryan, Դիվան հայ վիմագրության II. Գորիսի, Սիսիանի և Ղափանիշրջաններ շրջան¬ներ / Divan Hay Vimagrut’yan II. Gorisi, Sisiani ev Łap’ani šrĵanner (Corpus Inscriptionum Armenicarum II. Goris, Sisian and Ghapan Regions) (Yerevan 1960). Barxudaryan 1967 Ս. Գ. Բարխուդարյան / S. G. Barxudaryan, Դիվան հայ վիմագրության III. Վայոց ձոր. Եղեգնաձորի և Ազիզբեկովի շրջան¬ներ / Divan Hay Vimagrut’yan III. Vayoc’ jor. Ełegnajori ev Azizbekovi šrĵanner (Corpus Inscriptionum Armenicarum III. Vayots Dzor. Yeghegnadzor and Azizbekov Regions) (Yerevan 1967). Bedrosian 1979 R. Bedrosian, The Turco-Mongol Invasions and the Lords of Armenia in the 13th–14th Centuries (PhD Diss., Columbia University, New York 1979). Cherry 2014 J. Cherry, Archaeological landscapes, pushed towards ruination, in: Ö. Harmanşah (ed.), Of Rocks and Water. Towards an Archaeology of Place (Oxford 2014). Earley-Spadoni 2015 T. Earley-Spadoni, Landscapes of warfare. Intervisibility analysis of Early Iron and Urartian fire beacon stations (Armenia), Journal of Archaeological Science Reports 3, 2015, 22–30. Ganjakec’i 1982 Կ. Գանձակեցի / K. Ganjakec’i, Հայոց պատմություն. Թարգմն. Վ. Առաքելյանի / Hayoc’ patmut’yun. Targmn. V. Arakelyani (The History of Armenia. Transl. V. Arakelyan) (Yerevan 1982). Gasparyan – Arimura 2014 B. Gasparyan – M. Arimura (eds.), Stone Age of Armenia. A Guidebook to the Stone Age Archaeology in the Republic of Armenia (Kanazawa 2014). Gibb 1958 H. A. R. Gibb, The Travels of Ibn Battuta, AD. 1325–1354 (London 1958). Hammer 2014 E. Hammer, Highland fortress-polities and their settlement systems in the southern Caucasus, Antiquity 88, 2014, 757–774. Harut’yunyan 1960 Վ. Հարությունյան / B. Harut’yunyan, Միջնադարյան Հայաստանի քարավանատներն ու կամուրջները / Miĵnadaryan Hayastani karavanat’nern u kamurjnerə (Caravanserais and Bridges of Medieval Armenia) (Yerevan 1960). Harut’yunyan 1967 Բ. Հարությունյան / B. Harut’yunyan, “Վայոց ձոր” տեղանվան ստուգաբանության հարցի շուրջ / “Vayoc’jor” tełanvan stugabanut’yan harc’i šurĵ (On the issue of the etymology of the toponymy of “ Vayots Dzor”), Պատմաբանասիրական հանդես / Patma-Banasirakan Handes 9, 1968, 85–96. Jalalyanc’ 1858 Ս. Ջալալյանց / S. Jalalyanc’, Ճանապարհորդութիւն ի Մեծն Հայաստան, մասն Բ / Čanaparhordut’yun i mec’n Hayastan, masn B (Travels to the Great Armenia, Part B) (Tbilisi 1858).

Approaching Landscapes of Infrastructure: Methods and Results of the Vayots Dzor Silk Road Survey

143

Kałankatuatc’i 1969 Մ. Կաղանկատվացի / M. Kałankatuatc’i, Պատմություն աղվանից աշխարհի. Թարգմն. Վ. Առաքելյանի / Patmut’yun ałvanic’ ašxari, Targmn. V. Arakelyani (History of Aghvank. Transl. V. Arakelyan) (Yerevan 1969). Kajberuni 2003 Գ. Տ.-Հ. Քաջբերունի / G. T.-H. Kajberuni. Ճանապարհորդական նկատողություններ / Čanaparhordakan nkatołut’yunner (Travel Observations) (Yerevan 2003). Kalantaryan et al. 2009 A. Kalantaryan – G. Karakhanyan – H. Melkonyan – H. Petrosyan – N. Hakobyan – F. Babayan – A. Zhamkochyan – K. Nawasardyah – A. Hayrapetyan, Armenia in the cultural context of east and west. Ceramics and glass (4th–14th centuries). Joint Research Project of the Swiss National Science Foundation (Yerevan 2009). Lalayan 1904 E. Lalayan, Azgagrakan Handes, Vol. XII (Tbilisi 1904). Ե. Լալայան / E. Lalayan, Շարուր-Դարալագեազի գաւառ / Šarur-Daralagyazigawaȓ (Sharur-Daralagyaz province), Ազգագրական հանդես / Azgagrakan Handes (Ethnographic Journal) XII, 1904, 235–293. Larkin 2013 B. Larkin, The politics and poetics of infrastructure, Annual Review of Anthropology 42, 2013, 327–343. Lydolph 1963 P. Lydolph, Schemes for the amelioration of soil and climate in the USSR, in: R. D. Laird (ed.), Soviet Agricultural and Peasant Affairs (Lawrence, KS 1963) 204–214. Melkonyan 2003 Հ. Մելքոնյան / H. Melkonyan, Ետավարայրյան իրադարձությունները Վայոց ձորում, մի զորաջոկատի նահանջի ուղին / Yetevarayryan iradardzut’yunnerə Vayoc’ jorum, mi zoraĵokati nahanĵi ułin (Events after Avarayr in Vayots Dzor, the path of a military retreat), in: Ավարայրի խորհուրդը / Avarayri xorhurdə (The Significance of Avarayr) (Yerevan 2003) 223–224. News.am 2015 Armenia’s North-South highway to be finished in 5 years, News.am, 8. Sept. 2015. Online (last accessed 10 Jan 2018). Orbelyan 1986 Ս. Օրբելյան / S. Orbelyan, Սյունիքի պատմություն. Թարգմանությունը, ներածությունը և ծանոթագրությունները Ա. Ա. Աբրահամյան / Syuniki patmut’yun. Targmanut’yun, neracut’yunə yev canot’agrut’yunnerə A. A. Abrahamyani (History of Syunik. Transl. A. Abrahamyan) (Yerevan 1986). Polo 1904 M. Polo, The Most Noble and Famous Travels of Marco Polo, One of the Nobility of the State of Venice, into the East Parts of the World, as Armenia, Persia, Arabia, Tartary, with Many Other Kingdoms and Provinces. The Translation of Marsden revised by Thomas Wright (New York 1904). Ristvet et al. 2011 L. Ristvet – V. Baxşəliyev – S. Aşurov, Settlement and society in Naxçivan. 2006 excavations and survey of the Naxçivan Archaeological Project, Iranica Antiqua 46, 2011, 1–53. Wädekin – Jacobs 1982 K. E. Wädekin – E. Jacobs, Agrarian Policies in Communist Europe. A Critical Introduction (The Hague 1982). World Bank 2015 World Bank, Armenia. South Corridor Tourism Development Strategy, World Bank Working Paper 103149 (Washington 2015). Online (last accessed 10 Jan 2018). Yełiše 1958 Եղիշեի / Yełiše, Վարդանանց պատմությունը. Թարգմ., ծանոթություններով պրոֆ. դ-ր Ե.Տեր-Մինասյանի /, Vardananc’ Patmut’yunə. Targmanut’yamb yev canotagrit’yamb Prof. Dr E. Ter-Minasyani (The story of Vardanants. Transl. E. Ter-Minasyan) (Yerevan 1958).

Approaches to Understanding Provincial Structure in the Early Islamic Caucasus: Historic Landscape Characterisation in the Kura Plain Paul Wordsworth * Abstract: The modern town of Bərdə at the heart of the Kura Plain in the Republic of Azerbaijan betrays almost no evidence of its historical political significance. This lowland area, a strategic frontier zone of the early Islamic world, has the potential to reveal a great deal about the structure of borderlands under the Abbasid caliphate (AD 750–1258) and its successors. In order to unravel the deep history of Bərdə, however, it is necessary to understand the long processes of change that have created the current landscape. This paper proposes the adaptation of techniques commonly used for landscape analysis in Western Europe to undertake a multi-scalar reconstruction of Bərdə (medieval Bardhaʿa) and its hinterland. Historic Landscape Characterisation (HLC) offers the opportunity to identify the traces of past spatial organisation through remote sensing and comprehensive aerial mapping. Furthermore, the method can be used to study towns, villages and rural structure in a comparable manner, rather than focussing solely on the configuration of extra-urban areas. While the intensive agricultural policies and urban redevelopment of the late Soviet Union present different challenges to those confronted by the numerous UK HLC projects, this case study demonstrates the value of this approach in understanding landscapes of the former USSR and in ecological zones where nomadic pastoralist economies played a significant role in the past. Keywords: Geographical Information Systems (GIS); Caucasus; Medieval archaeology; Historic Landscape Characterisation; Early Islamic world; Nomads

Comparing the British Inclosure Acts of the early 19th century and the agricultural reforms of the USSR might seem somewhat oxymoronic. The aims of the landholders dividing and administering common fields ostensibly stand at odds with the aims of Soviet collectivisation. Yet the physical processes of landscape change that took place during these movements bear striking similarities to one another, and their still extant traces represent a legible palimpsest of the stages of reorganisation. In the United Kingdom, archaeological practitioners are familiar with the need to see beyond the regularly parcelled pastures of the early modern period to trace the fragments of earlier systems. The science of delineating this evidence comes under the rubric of ‘Historic Landscape Characterisation’ (HLC). Vast tracts of the British countryside, and a smaller portion of other parts of Europe, have now been neatly digitised in Geographical Information Systems, allowing research projects to unpeel the layers of modern landscape transformation and scrutinise what remains from the past.1 For those working in the post-Soviet world, this should be an encouraging development, given the at least superficial similarities in the superimposition of modern field systems at a known point in time. Many archaeologists, however, maintain a fairly pessimistic view of what might be preserved of historical landscapes of countries formerly in the USSR, owing to the industrial scale of the terraforming that took place, and in many cases this view is substantiated. This paper, however, looks to explore the possibility of combining HLC techniques with more traditional regional survey in order to piece together fragments of a significantly changed landscape in modern Azerbaijan.



*

1

University of Oxford, [email protected]. Turner 2006; Nord 2009; Turner – Crow 2010.

146

Paul Wordsworth

The case study presented here is the city of Bərdə, synonymous with the medieval Caucasian city of Bardhaʿa – capital of the province Arrān that came nominally under the suzerainty of the Umayyad Caliphate in the mid-7th century AD.2 In spite of numerous (often anecdotal) references to the city in textual sources, very little is known of the structure of the town or its hinterland in the period before the Mongol conquest. The existence of a modern town essentially replacing the historical one renders the physical investigation of the urban layout problematic. The lack of substantial extant archaeological monuments might even lead to scepticism about the identification of this locale as early Bardhaʿa. Closer inspection of the open spaces and surrounding landscape, however, reveals a great deal of surface material culture from the medieval period. Findings from the limited amount of archaeological excavation that took place in the city in the late 20th century corroborate the existence of historical remains beneath the town dating from at least as early as the 10th century and apparently back to the early medieval period (3rd–4th century AD),3 but too little was recorded in detail to give an overview of city organisation and development.4 Placing Bardhaʿa Bardhaʿa lies on the plain of the Kura River as it snakes towards the Caspian Sea in between the Greater and Lesser Caucasus mountain ranges (Fig. 1). In spite of relatively low average annual rainfall,5 it is a fertile zone of silty alluvial deposits, made productive by the vast network of canals that crisscross the flat lands. The city is placed on a major tributary of the Kura, the Tərtərçay (Tərtər River), which, while fed by erratic meltwaters and winter storms, maintains at least some flow throughout the summer. Evidence that the course of this river was similar in the past can be seen in the city itself, where the remains of a pre-modern bridge (likely early medieval) can be seen in the middle of the river course. The alignment of the bridge, however, indicates dramatic changes in the immediate morphology of the channel as it runs almost in line with the current flow. As I will demonstrate below, this is but one of many considerable changes in the surface hydrology that took place during the 20th century, hand in hand with agricultural transformation. There is a certain degree of correlation between the modern city of Bərdə and the details given in early Arabic geographical accounts of Bardhaʿa. It is said that ‘a river flows through the town, and the River al-Kurr [Kura] is about two farsakhs6 away’.7 Although the distance is slightly short of a modern Cartesian measurement between Bərdə and the current course of the Kura (around 14 kilometres), this is the largest historical settlement in this region, and it indeed sits astride the Tərtərçay. That Bardhaʿa, at least in the 10th century, was a provincial capital, is not disputed. It is seemingly the administrative hub that appears in all political narratives of the region and in fact

2

3 4

5

6

7

Arrān gained lesser and greater degrees of independence throughout the period between the bulk of the early expansion of the Caliphate until the Mongol conquest as well as being conquered or partially occupied by other local dynasties. For an overview of some of the complexities of local governorship in the south-east Caucasus in the medieval period, see Minorsky 1953; Minorsky 1958. Nuriyev – Babayev 2001, 33–34. The study presented here is a small part of the ongoing research by the Archaeological Exploration of Bərdə Project (AEB), part of the Oxford Nizami Ganjavi Programme, a collaborative initiative between the University of Oxford and the Baku Branch of the Moscow State University. Average annual rainfall in Bərdə in the later part of the 20th century was around 370mm, which is well within the moderately arid category of land classification but not generally enough for productive agriculture that is solely rain-fed. An overview of the climate data for the region can be gained from the aggregated GIS layers of the WorldClim global climate data project (http://www.worldclim.org/), University of California, Berkeley. Values quoted in this chapter are based on these data. Two farsakhs is roughly equivalent to 12km, although the exact measurement varies depending on the author and the period concerned. Al-Muqaddasī, 305.

Approaches to Understanding Provincial Structure in the Early Islamic Caucasus

147

Fig. 1   Map of the south-eastern Caucasus region, showing the location of major archaeological sites with medieval remains (drawn by the author, using SRTM topographical data (courtesy of USGS)

it is described by al-Muqaddasī as the ‘Baghdād’ of this region.8 However, al-Muqaddasī and his contemporaries also hint that by the late 10th century, something is going badly wrong and the city is in decline. The most extreme example is given by Ibn Hawqal, who states that ‘I have learnt that today, its pitiful condition and the squandering of its resources are such that only five bread bakeries remain, where there had once been more than twelve hundred’.9 One of the reasons posited for the decline of the city is the well-known account of raids by the Rūs in the 10th century, who seemingly besieged the city for a year before eventually retreating.10 Setting aside this seemingly short-lived event, there is no indication in the sources of the reasons for a reduction in activity in Bardhaʿa, apart from general corruption and vice, mentioned anecdotally by Ibn Hawqal.11 The idea of Bardhaʿa’s decline through the course of the 10th and into the 11th century is central to modern narratives of the city, in which Ganja12 eclipsed the earlier capital and thus rendered it politically obsolete. There is strong evidence to suggest that as the centre from which the Shaddadids ruled, Ganja indeed played a pivotal role in the political geography of the Caucasus during the 11th–12th centuries.13 The ruin of Bardhaʿa, however, is less easily placed, and its validity has been questioned by our most recent archaeological investigations. For the purposes of the article

8 9 10

11 12 13

Al-Muqaddasī, 305. Ibn Hawqal, 330. Details of this episode, which is a complex affair in its own right, are most thoroughly regaled by Miskawayh (d. 1030) in his Tajārib al-Umam. See Margoliouth 1918 for a full description and translation. Ibn Hawqal, 330. Modern Gəncə, sometimes spelt in medieval Arabic sources as Janza. In spite of suffering considerable destruction during numerous raids and at least one devastating earthquake. Ibn al-Athīr, 354.

148

Paul Wordsworth

Fig. 2   The street plan of Bərdə in 1970, as seen from Corona satellite imagery. Zones that were subsequently built up between 1970–2015 are shown in black crosshatching (Corona basemap, courtesy of USGS)

here, however, it is important to emphasise that notwithstanding the ebb and flow of urban prosperity in the region, the toponym and indeed the settlement of Bardhaʿa has never completely died away. Coins continue to be minted under its name certainly into the 14th century under the Ilkhanids/Hulaguids.14 Even in the late 18th century, the town appears in the historical accounts of the campaigns of Nader Shah.15 Bringing us up to the present day, while the town of Bardhaʿa is conspicuously absent from accounts from the period relating to the Russian annexation of the Caucasus (1820s onwards), it appears on Russian and European maps throughout the 19th century.16 The only brief description of the town from this period is given as part of a general narrative in the Tārīkh-e Qarabāgh which states that ‘on the site of Bardaʿ there is an insignificant and small settlement, populated by the Burdalu tribe’.17 Historical Traces in the Modern Town Just as the history of rural landscapes can be read through the vestiges of former field boundaries and stream courses, it is unusual for a town with historical roots to lose all traces of its pre-modern urban fabric. Scrutinising even the most radically planned Soviet cities (for example, Tashkent), there are still, almost without exception, glimpses of earlier street layouts. The logical starting

14 15 16 17

Pakhomov 1966, 68, 73. Lockhart 1973 [1938], 248–249. Kolokolov 1836; Weller 1877. Bournoutian 2004, 39.

Approaches to Understanding Provincial Structure in the Early Islamic Caucasus

149

point for undertaking Historical Landscape Characterisation for Bardhaʿa is thus the city itself, which has little ostensible historical fabric on the ground but has not been dramatically demolished and replanned in the 20th century. Fig. 2 presents data taken from a snapshot of the town in 1970, provided by Corona satellite imagery.18 The urban area of Bərdə to the north of the Tərtərçay represents an area of some 268ha, less than half its present extent (546ha), owing to the rapid expansion of the town in the 1980s and 1990s. At the time of the aerial photograph, seemingly very little had been attempted in terms of formalising the urban layout and, while the streets conform roughly to a grid plan, the primary axes are not rigidly laid out, but they are present throughout the core of the town. While it is not possible to state with certainty at the present time how old this pattern of streets is, there is no reason to suppose that it does not relate to the pre-modern period. In support of such a hypothesis, the axes of the approximate grid run 167° and 255° from north.19 This slightly off-cardinal alignment precisely mirrors the layout of the only extant historical walls in the city, the enclosure known as Torpak Kala (examined below), and also the general alignment of the architecture exposed inside these walls during the recent archaeological excavations of the AEB. The latter buildings, probably domestic in nature, are dated by associated material culture to the 12th–13th century but may follow an earlier pattern. Whatever the reasons for and the exact date of the initial establishment of this particular grid, its persistence between the medieval and modern occupation of the city speaks to continuity of urban life and a high degree of survival of the early layout. In many cities with a substantial medieval Islamic population, orientation of buildings has often been linked to the establishment of religious structures on the line of the qibla. This is certainly not the case in Bardaʿa, as the qibla direction from the town is approximately 200°. The off-cardinal grid pattern also contrasts with very modern interventions in the town, including the establishment of the railway line running due north–south to the west of Bərdə. A similar juxtaposition can be seen with the cardinally orientated blocks in the northernmost part of the town and, from later satellite imagery, the structured polygonal-plan neighbourhoods constructed in the 1980s. It is also likely that the major east–west axis bisecting the city and linking it to the adjacent settlements was created after the main layout of the town. Not only does it sit at odds with the remainder of the urban street pattern (forming an awkward series of bends through the city), but it also appears to split several historical field systems in the immediate hinterland. Individual ‘neighbourhoods’ in the residential parts of the town of 1970 show a more irregular layout – their dendritic structure ending in a series of cul-de-sacs. Although it is tempting to delineate divisions of the housing units based on these patterns, there are potentially a high number of smaller informal streets and pathways invisible in the image, and the conclusions at this scale could be misleading. The investigation of persistent private land boundaries from earlier periods is a fascinating avenue of inquiry, which can only be glimpsed at this point but would benefit from further research on the ground. A further advantage of using imagery from the recent past is being able to evaluate the impact of development within the urban area. While in cities with better documentation a full historical map-regression exercise would be a preferable starting point for archaeological investigations, the resources for such an endeavour are unavailable for Bardhaʿa. It is possible, however, to identify some areas that have remained continuously ‘open ground’ from before the 1970s imagery was obtained and to investigate whether these zones may represent archaeological potential.

18

19

Decommissioned Corona high-resolution stereo panchromatic satellite imagery, courtesy of USGS. This imagery is now in common use by archaeologists across the world, but particularly in the Middle East, for regions where aerial photographs are not necessarily available and where it is useful to have a view of landscapes prior to the most recent modern development. 167°–347° South–North, 255°–75° West–East.

150

Paul Wordsworth

Fig. 3   1970 Corona imagery of the HLC study area (courtesy of USGS)

The application of HLC for cityscapes, while rapidly being adopted in the United Kingdom, is challenging in these circumstances, based on the resolution of the data available. Nevertheless, thinking about the historic fabric of Bərdə through this lens provides a basis for exploring alternative forms of information gathering. Two such examples in progress for the AEB project are buildings evaluation and public archaeology. It is anticipated that over the coming years it will be possible to gather spatial data from household interviews in the town and ground survey work to create a high-definition map of the modern town and the fragments of earlier periods currently hidden. Meanwhile, analysing the landscape of Bərdə at a broader scale can situate the historical city within its immediate hinterland. Beyond the City – HLC and the Hinterland of Bərdə When categorising the rural zone around the town with basic HLC classes (analogous to those systems developed in British examples), there are some zones that have direct parallels and can easily be delineated in a similar way. Both woodlands and urban areas (see above) are straightforwardly traceable on the aerial imagery and, on a subtler level, the satellite photography reveals the remains of current and relict water channels (natural, adapted and man-made) (Fig. 3). Woodland areas are easily categorised from the Corona imagery, appearing almost black with the dark green of the canopy. Although few patches of forest remain today, those stands that can be identified appear to be mature deciduous woodland, and several still exist at the present time. Documentary sources from the medieval period onwards suggest that a substantial portion of the region was previously wooded, far more than is represented by these isolated stands (see below for a further discussion of this). The remains of both relict and active water channels are somewhat easier to detect, appearing as dark linear features across the landscape – in both instances relating to increased

Approaches to Understanding Provincial Structure in the Early Islamic Caucasus

151

Fig. 4   HLC plot for the hinterland of modern Bərdə, showing land classes described in the text (drawn by the author)

localised vegetation growth. This particular snapshot from 197020 reveals the grid pattern of active irrigation canals (recently constructed at that time) superimposed upon an earlier stream network, particularly in the areas to the east and west of Bərdə. An equally dramatic change can be seen in the construction of the Upper Karabakh Canal: part of a wider re-engineering of the Kura River flow that was carried out in the 1950s which included several new dams and canals across the Kura Basin. The extent of these hydrological changes is attested by a considerable number of inactive east–west running streams, visible in the Corona images from 1963 and 1970. Reading the images further it is apparent that for the majority of the area these early streams are deeply incised into the plain, cutting into a slightly rolling landscape. The prominent meanders suggest that these are natural stream courses (or at least, were only partially canalised) and are not intended for field irrigation. Their depth would also have rendered redistribution of water resources difficult and is undoubtedly the reason why in areas, particularly to the west of the town, a new orthogonal grid of shallower canals has been superimposed to transform this land into consistent arable land.21 The combination of the semi-dryland signatures and the deeply incised streams suggests that prior to the recent period, most of these watercourses irrigated specific, small zones around the town and individual villages, rather than supporting widespread intensive agriculture.

20

21

The metadata from the image reveals that it was captured on the 30th May 1970. A similar medium-resolution photograph was also consulted that had been captured on the 27th June 1963 and provides a slightly different perspective, both in terms of the soil conditions and also that further landscape development had taken place in the intervening period. While some of the land in the Kura Basin can be managed through limited rainwater catchment, intensive agriculture is reliant on stable irrigation, given the rainfall/enviro-transpiration ratio. It was anecdotally confirmed by local community members that these interventions were mostly carried out from 1955 onwards, but it has not yet been possible to corroborate this with detailed documentary evidence.

152

Paul Wordsworth

The presence of so many streams in the immediate vicinity of the town also gives us cause to revaluate the current appearance of the urban space. Several of the abandoned channels run towards the town, disappearing from the satellite image as they meet the built area. While it is not currently possible to trace their original courses through the modern houses and streets, in the earlier part of the 20th century these may have been a crucial part of urban organisation in either open or closed conduits (Fig. 4). Turning to the land use, it is apparent from the discussion above that defining a ‘system of fields’, the basis for much of the analogous work in UK HLC, requires a more locally applicable labelling, as the visible palimpsest does not suggest large open-area cultivation. On the basis of the Corona image, it is possible to divide the open ground into two categories, simply on the basis of reflected tone in the greyscale photograph. The perceived difference is one of vegetation density, as the ‘greener’ areas reflect a darker signature in the image, broadly speaking. Compare, for example, the dense stands of trees, noted above, which appear almost black, and the dry flood beds of the Tərtərçay. These differences are evident too in the enclosed ‘fields’ which appear to be in a process of transition during the period 1963–1970. Away from the town, the bulk of the land can be characterised as comparatively dry and non-irrigated ground, with rolling, slightly higher elevation, such as would be suitable for pasturelands. It is clear that at the point the image was captured a regular grid of irrigation ditches had recently been superimposed although was not yet fully functional. On imagery from the last few years this zone has been completely transformed into uniformly flat arable fields and this can be corroborated on the ground. That a substantial proportion of the land around Bardhaʿa should at one point have been devoted to pasturelands rather than crop cultivation is unsurprising, given the dramatic remodelling of the countryside undertaken under Soviet administration and its persistent legacy. To gain a more detailed impression of the extent of the historical pastoral economy, it is necessary to consult documentary sources, which, although patchy, do shed some light on the division of the economy. The most enlightening in this respect are administrative papers relating to the annexation of the Caucasian Khanates at the beginning of the 19th century, which prompted substantial investigation and documentation of the region by the new Russian leadership. One result of this, The Description of Karabakh Province Composed in 1823,22 is not widely available but has been transcribed in full into English by Bournoutian.23 This document reveals that as well as the villages administered previously under the Khanate, the majority of the rural zone is specifically designated as nomadic pasturelands. Nomadic groups were registered and taxed by local landlords on their use of pastures and expected to give payment both in gold and in livestock products. There are also several instances in the census where nomadic groups are periodically renting ploughs to undertake agricultural activities. If more evidence were needed for the often complex nature of nomadic and semi-nomadic economies, the land system around Bardhaʿa is an excellent case in point. Although Bərdə falls within the province of Karabakh at this time, there is unfortunately no mention of the town within the surviving document, and it is unclear what its role is as a settlement in its mahal (district). Furthermore, while historically Bardhaʿa was considered as part of Javanshir24 mahal, these lands are subdivided into several smaller units in the census and it is not clear which the town would belong to. Examining the taxation record for the area designated as Javanshir mahal as an example of land use, however, 98% of the families surveyed are designated ‘nomadic’.25 A slightly more balanced figure is given for other parts of the plains adjacent to the Kura, such as the mahal of Otuz Iki, whose nomadic families represent 72% of those registered. Nevertheless, it is clear that one should expect a considerable proportion of the land around Bərdə

22 23 24

25

Opisanie Karabakhskoǐ provintsii sostavlennoe v 1823 godu. Bournoutian 2011. Javanshir is originally a tribal/family name, but it came to denote a region of the Karabagh Steppe, to the south-west of the Kura River. Bournoutian 2011, 234–248. This figure includes both taxed and tax-exempt families.

Approaches to Understanding Provincial Structure in the Early Islamic Caucasus

153

to be devoted to livestock raising, as opposed to agriculture. Returning to the Historic Landscape Characterisation plot, these unirrigated areas visible on the satellite imagery are seemingly the relics of this system, the substantially pastoralist economy, prior to widespread field creation. Projecting further back, it is reasonable to expect that a high proportion of the landscape may have supported these activities, although in order to reconstruct the palaeoenvironment in greater detail it will be necessary to undertake sampling from test pits across the study area. A different pattern of land use can be seen in some limited zones to the north-east of the town, where small relict linear field boundaries are visible as pale lines in an area that is darker overall than the zone described above. It would appear then that irrigated agriculture can be confirmed in the immediate vicinity of the urban area, and indeed similar signatures can be found around the boundaries of several of the nearby villages but are small in number and poorly defined. This is not to suggest that additional portions of the countryside may not have been cultivated in this manner in the past, but only small areas are traceable by the mid-20th century, and the fields appear to adhere strongly to the edges of settlements. One model which may represent an extant and functioning parallel can be seen in the small gardens unusually preserved around the historic village of Lenberan, 23km to the south-east of Bərdə. Unlike most rural settlements surrounded by the uniform modern field systems, the patchwork of small plots around Lenberan gives a glimpse of a structure similar in size to those seen in the satellite imagery. Both examples reasonably represent a mode of land management consistent with that of the early- and pre-modern period. Each cellular unit is surrounded by mature hedges and trees, while small ditches distribute water from the slightly higher ground at the centre of the village outwards to irrigate the whole area (notwithstanding the general average gradient in this region from west to east, towards the Kura). Some plots are further subdivided to cultivate a range of crops rather than homogenous cultivation. The AEB project is in the process of mapping this system around Lenberan to understand its historical origins and to ascertain further details about how it functions in practice. Historical Maps In terms of identifying individual classes of land, it is important to realise the limitations of the perspectives available for Bərdə, particularly given that there are far fewer maps than for comparative areas of Western Europe. For instance, one of the earliest detailed examples available for the province is preserved as a 1941 German transcript of an earlier (and possibly more detailed) Soviet chart.26 While this map is drawn at 1:200,000 and thus not sufficient to trace land classes for small areas, the area around Bərdə is generally designated an area of woodland (Wald). Given the relatively low accuracy of this map, however, the idea that the whole immediate area around Bərdə was forest prior to the mid-20th century should be approached with caution, most clearly as there is not sufficient detail on the map to distinguish between individual patches of woods and a single contiguous zone. Nevertheless, it is striking that many of the general medieval descriptions of Bardhaʿa also mention its trees. When added to the fact that in the HLC study several stands of mature trees were identified, it is possible to state that a considerable portion of the rural zone around Bardhaʿa would have been maintained as forest in the past. Some of these trees may also pertain to sparse gardens, as a similar symbol is used for the gardens which surround the village of Lenberan (see above). The same map data from the first half of the 20th century also corroborates that some of the land is designated pastureland. Small structures, labelled Nomadenlager (nomad camps) pepper the spaces in between named villages.27 These not only occur in areas designated as grasslands,

26 27

Generalstab des Heeres 1941. Generalstab des Heeres 1941.

154

Paul Wordsworth

but also in the vicinity of wooded areas. Given that several of the camps are named, it may eventually be possible to match some with the census records cited above, but linking these two sources is troublesome as many of the names are fairly generic, and they are liable to change as families move or simply over time. It is curious, however, that to the south of Bərdə around the town of Ağcabədi, much of the land designated as steppe is also liable to flooding (Überschwemmungsgebiet) and there are few nomadic camps in this area. The picture of the wider Kura Basin that emerges from these maps is therefore one that contrasts completely with the modern view of the landscape. Historically, this region is highly varied, with local land-management, shifting river courses and micro-topography. While an impression of this diversity has been gained through the methods applied above, further investigation is now required to establish the deeper history of discreet land classes and how their properties were managed. Most of the wide-coverage late Soviet cartography from the second half of the 20th century is now publically available,28 but although charts at 1:50,000 scale and larger were almost certainly made, the only format that has been retrieved is at 1:100,000. There are several interesting features included on these more detailed, topographic maps including basic land categories, and water channels, still active when the chart was produced. Given that it was printed in 1985, however, its main usefulness is in broadly confirming features identified on the Corona imagery in the period of the expansion of the irrigation system. The results of the map analysis have thus been incorporated into our regional HLC, but no particular aspect warrants further consideration here. While earlier map resources may eventually surface, the paucity of data for the early 20th century highlights the primary importance of the early satellite imagery in the case of Bardhaʿa. Fieldwalking In order to further define the character of the land types being observed remotely and investigate the chronology of their occupation, it has been possible to undertake a limited amount of fieldwalking in the region around the town (Fig. 5). On a very general level, the ground-truthing exercise confirmed that the extensive terraforming of the Soviet period has almost entirely obliterated the distinction between the two landscape types. The deeply ploughed fields that have replaced the pasturelands appear flat and featureless, and on the ground it is impossible to identify any relict landforms in current agricultural areas. The ploughing, however, has turned over material culture that had previously been buried under the cultivable layer or raised to the surface residual ceramics from ‘tell’29 sites now bulldozed. As the material is very mixed, only a rough approximation of an occupation history can be made, and even this requires checking. Particularly prominent among the material collected in the immediate environs of the city were sherds pertaining to the 11th–13th century AD. Notwithstanding the limitations and caveats mentioned above, the distribution map suggests that while historical sources indicate a decline in activity in Bardhaʿa from the 11th century onwards, occupation from this period is in fact represented over an area even wider than the current extent of the town. In order to verify the occupation sequence of the city of Bardhaʿa, small-scale excavation is currently being undertaken in the centre of the town, as well as at a few key rural sites. While this contributes a more orthodox approach to understanding urban development, it also informs our view at the landscape level, providing comparative data for specific periods in time and refining

28

29

It was previously mostly under strict state control. See the general overview of the status of public and state maps in the Soviet Union given in Postnik 2002. Like much of the Middle East and Central Asia, the predominance of unfired mudbrick structures in the Southern Caucasus leads to the creation of settlement ‘tells’ – hills formed from successive layers of destroyed, eroded and rebuilt earth. While in many landscapes, the high frequency of these man-made hills attests the widespread presence of historical settlements, in areas being prepared for agriculture, the smaller of these are routinely flattened, and their survival in the lower Kura Basin is thus rare.

Approaches to Understanding Provincial Structure in the Early Islamic Caucasus

155

Fig. 5   Map of fieldwalked areas around Bərdə indicating zones with ceramics and those without (basemap 1970 Corona image, courtesy of USGS)

the chronology of local ceramics. Targeted excavation has also been carried out at a handful of rural sites with seemingly exceptional preservation in order to understand the stratigraphy of their development in comparison with that of Bardhaʿa. Regional Survey While it has not yet been possible to characterise a large region with the techniques described above, the AEB Project has been carrying out a limited amount of more traditional site prospection and survey across Bardhaʿa rayon.30 Completing a comprehensive regional study is an ongoing aim of the project but for the methodological discussion it is useful to highlight what the preliminary results can reveal alongside HLC. Remote sensing for archaeological remains using satellite imagery, and Corona imagery in particular, is now a well-documented practice among archaeologists working in the Near/Middle East31 and needs no further explanation here. Owing to the impact of agricultural reorganisation, however, comparatively few sites can be traced with certainty from satellite imagery and all require ground-truthing to determine whether or not

30

31

Rayon is the regional administrative unit in Azerbaijan, comparable to a district. The modern boundaries of Bərdə rayon encircle an area of approximately 1000km². Philip et al. 2002; Casana – Cothren 2008.

156

Paul Wordsworth

Fig. 6   Map of sites surveyed from satellite imagery and field investigation, indicating those with confirmed archaeological remains, those with null results and those yet to be ground-truthed (basemap 1970 Corona image, courtesy of USGS).

archaeological material is present. Fig. 6 summarises sites identified remotely from both the 1970s’ and recent images.32 The results of the survey so far have revealed a widespread distribution of archaeological remains, perhaps surprising when considering the degree of transformation that has taken place. Some of the mounded sites traced on the Corona satellite imagery, typically as lighter, regularly shaped features (in the case of mounds), are no longer extant, but the remains of surface ceramics verified that these had once been settlements. Among the features recorded, a high proportion were circular in plan and small in diameter and were interpreted as burial mounds or kurgans. Beyond a general categorisation, it is difficult to determine a specific size threshold to distinguish kurgans from settlement sites, but the former also tend to appear as very regular, steep-sided, sub-conical mounds and often form clusters of two or three, rather than forts and villages which are generally isolated, less regular and are mostly flatter on top. The distinction between these two categories of site is further complicated by the infrequent practice of later reuse of existing burial mounds for settlement or other activities. Typically, however, very little material culture is found on the surface of burial mounds, and they are consequently difficult to date accurately without excavation.33

32

33

While a handful of mounds or tepes were identified from the historical map sources, the scale of the charts rendered them difficult to trace on the ground. In some instances, marks on the maps corroborated the presence of individual or clusters of mounds viewed on the aerial imagery. On the basis of precedent, it might be possible to state that the majority of burial mounds in Azerbaijan are demonstrably pre-Islamic, but this has not been firmly proven and further research is required to ascertain how far this practice may have been continued across the Caucasus after the 8th century AD.

Approaches to Understanding Provincial Structure in the Early Islamic Caucasus

157

Analysing the general settlement pattern of the survey data, sites are clustered in two rough groups: those around Bərdə and the Tərtərçay and those which follow a roughly linear distribution east–west, and appear to follow another now relict tributary of the Kura River.34 The clustering of sites along natural river courses may reflect a realistic pattern of occupation, given the observation above that in the past, irrigated agriculture was concentrated in the areas immediately bordering settlements. This practice would naturally encourage more substantial occupation around the river course fed by the floodwaters of the lesser Caucasus, with bands of steppe land in between. Kurgans, meanwhile, are spread more or less evenly throughout the study area, although those in the western half are better preserved. Identifying archaeological signatures as tells or tepes, however, risks overlooking a considerable portion of the archaeological remains, particularly from the first millennium AD. As demonstrated by Bardhaʿa itself, there is a precedent of persistent occupation of settlement sites from the medieval period to the present day, and many of the more substantial settlements may be masked under modern development. Without a more detailed Digital Elevation Model (or LiDAR data for built areas) it is difficult to determine which of the smaller villages may conceal earlier remains. Beyond settlements, the HLC approach outlined here has also demonstrated that there is a substantial amount to be learned through the study of landscape management. Furthermore, through understanding current and relict land forms, from remote sensing and on the ground, the relationship between natural features and settlement patterns can be further explored, including the tentative correlations described for the survey data. By creating a comprehensive characterisation of this area, then, it will be possible to overcome some of the limitations of surveying a landscape of seemingly poor archaeological preservation. Conclusions While the practice of Historic Landscape Characterisation is firmly founded on the tradition of English landscape studies, it offers a useful window to understanding the stratigraphy of dramatic rural change. In the absence of the stalwart tools relied upon in British archaeology – a standardised series of detailed modern maps and extensive documentation of early-modern field division – archaeologists in the Caucasus, and the former Soviet world, must think laterally about combining cartographic and documentary sources with aerial imagery and pedestrian survey. Using these materials, however, it is possible to apply a similar methodological framework to that used in HLC, albeit with appropriately different categories of land use. In this region of the world, one such category will often be nomadic pasturelands. Although examples were identified in the brief case study presented here, further work is required in order to develop tools to concretely identify zones used as pastures, preferably in conjunction with palaeoenvironmental studies. The question of how to date landscape features is also a persistent concern, notwithstanding the usefulness of relative dating by visible stratigraphy and juxtaposition with archaeological sites. Once more the ideal solution is to carry out extensive test-pitting or coring over a large sample area to reconstruct a cross section of development. It is anticipated that a trial study of this type across Bərdə rayon will allow the project to refine the land classification already undertaken and gain an insight into the deeper history of rural development in this region. Finally, it is important to emphasise the value of HLC as a means of studying landscapes as a continuous narrative, from the present back into the past. There has been an overindulgence of studies ostensibly covering the longue durée. These studies, while often groundbreaking, are sometimes hampered by trying to cover too much in little detail or are too narrowly focussed to consider multiple vectors of change. I would like to avoid these issues by stressing another advan-

34

Only sections of this channel are visible as crop-marks indicated desiccated meanders and a handful of low lying gullies, and further work is needed to reconstruct the palaeohydrology of the region.

158

Paul Wordsworth

tage illustrated above which is that the long-term perspective provides an opportunity to combine the threads of textual and archaeological evidence into a single account of landscape change. A more optimistic view of archaeology in landscapes of industrial-scale destruction, then, seeks innovative ways to understand those processes and thus uncover what still remains of the past. The South Caucasus region is an excellent case in point, and this study hopefully demonstrates that a fruitful way to take research forward in the region is to devote greater attention to studies of landscape development, away from a focus on the few remaining islands of archaeological sites. References Bournoutian 2004 G. Bournoutian, Two Chronicles on the History of Karabagh. Mirza Jamal Javanshir’s Tarikh-e Karabagh and Mirza Adigözal Beg’s Karabagh-name (Costa Mesa, CA 2004). Bournoutian 2011 G. Bournoutian, The 1823 Russian Survey of the Karabagh Province. A Primary Source on the Demography and Economy of the Karabagh in the Early 19th Century (Costa Mesa, CA 2011). Casana – Cothren 2008 J. Casana – J. Cothren, Stereo analysis, DEM extraction and orthorectification of Corona satellite imagery. Aarchaeological applications from the Near East, Antiquity 82, 2008, 732–749. Generalstab des Heeres 1941 Generalstab des Heeres, Rußland, 1:200 000 (Berlin 1941). Ibn al-Athīr Ibn al-Athīr, The Chronicle of Ibn al-Athir for the Crusading Period from al-Kamil fiʿl-Taʿrikh, transl. D. S. Richards (Ashgate, Surrey 2005). Ibn Hawqal Ibn Ḥawqal, Configuration de la Terre (Kitab Surat al-Ard), transl. G. Wiet and J. H. Kramers (Beirut 1964). Kolokolov 1836 П.Ф. Колоколов, Карта российских владений за Кавказом Составлена корпусным топографом, поручиком колоквым (Saint Petersburg 1836). Lockhart 1973 [1938] L. Lockhart, Nadir Shah. A Critical Study Based Mainly Upon Contemporary Sources (Lahore 1973 [1938]). Margoliouth 1918 D. S. Margoliouth, The Russian seizure of Bardha’ah in 943 A.D., Bulletin of the School of Oriental and African Studies 1, 2, 1918, 82–95. Minorsky 1953 V. Minorsky, Studies in Caucasian History (London 1953). Minorsky 1958 V. Minorsky, A History of Sharvan and Darband in the 10th–11th Centuries (Cambridge 1958). al-Muqaddasī al-Muqaddasī, The Best Divisions for Knowledge of the Regions. Aḥsan Al-Taqāsīm Fī Maʿrifat Al-Aqālīm, transl. B. Collins (Reading 2001). Nuriyev – Babayev 2001 A. Nuriyev – Ə. Babayev, Bərdə Şəhərinin Tarixi-arxeoloji Ocerki (Antik və Orta-əsrlərdə) (Baku 2001). Nord 2009 J. Nord, Changing Landscapes and Persistent Places. An Exploration of the Bjäre Peninsula (PhD Diss., Lund University, Lund 2009).

Approaches to Understanding Provincial Structure in the Early Islamic Caucasus

159

Pakhomov 1966 Е.А. Пахомов, Монетные клады Азербайджана и других респубпик, краев и областей Кавказа, Выр. 9 (Baku 1966). Philip et al. 2002 G. Philip – D. Donoghue – A. Beck – N. Galliatsatos, Corona satellite photography. An archaeological application from the Middle East, Antiquity 76, 2002, 109–118. Postnik 2002 A. V. Postnik, Maps for ordinary consumers versus maps for the military. Double standards of map accuracy in Soviet cartography, 1917–1991, Cartography and Geographic Information Science 29, 3, 2002, 243–260. Turner 2006 S. Turner, Historic Landscape Characterisation. A landscape archaeology for research, management and planning, Landscape Research 31, 4, 2006, 385–398. Turner – Crow 2010 S. Turner – J. Crow, Unlocking historic landscapes in the Eastern Mediterranean. Two pilot studies using Historic Landscape Characterisation, Antiquity 84, 323, 2010, 216–229. Weller 1877 E. Weller, The Caucasus (Circassia, Georgia, etc.) and Armenia (London 1877).

Index A

illegal excavation 72, 75, 76

Ani 68, 70, 71, 73, 75, 77, 78, 106

extensive field survey see survey

Aragats 29, 55

F

Ararat (Ağrı) 55, 70, 75, 76 Aras/Araxes River see river

fortification 40, 59, 84, 107, 113

borders and frontiers 119, 120–121, 127–128, 145

fortress (castle) 45, 47, 48, 49, 53, 56–63, 67, 69–78, 105, 113, 137, 138 Berdi Dar 61, 62 Kaghsi 54, 61, 63 Karashamb 54, 59 Karmir Berd 54, 57, 58, 59 Karmir Blur 55, 57, 59 Lčašen 54, 61, 62, 63 Muchannat-tapa 54, 56, 57 Solak 61 Tghit 53, 54, 59, 60, 63 Tsitsernakaberd 54, 56

bridge 24, 29, 132, 134, 135, 139–140, 141, 146

‘friction of terrain’ 24, 26

Bronze Age 37, 39, 55, 57, 59, 72, 73, 75, 76, 77, 78, 83, 84, 87, 92, 93, 94, 100, 110 Late Bronze Age 15, 39, 41–49, 53–63, 73, 75, 76, 77, 87, 111

G

Armenia 12, 14, 23, 28, 30, 39, 45, 53–56, 67, 69, 70, 73, 75, 106–107, 131–141 Arpaçay River see river artefact distribution 7, 13, 83 Azerbaijan 12, 14, 53, 67, 145, 155, 156 B Berdi Dar

see fortress

Black Sea 23, 24, 67, 29, 39, 97, 98, 99, 132

C caravanserai 132, 139, 141

geophysical survey see survey Georgia 12, 14, 15, 16, 37–39, 44–45, 53, 83, 105–110, 119–120, 125–128 GIS 11, 13, 15, 16, 23, 24, 26–30, 85, 121, 131, 145, 146

cave 112, 114, 115, 120, 124–128

graves 37, 39, 44, 53, 57–58, 63, 70, 75, 87, 110, 113–114, 115 cist-graves 53, 58, 63

Chobareti 83–87, 90–94

Greek colonisation 97, 98

church 109–110, 113–114, 115, 134, 135, 137, 139, 140

ground-truthing 154, 155, 156

Colchis/Colcheans 37, 97

H

Caucasus Mountains 13, 23, 146

copper 37–39, 44–49, 105 cumulative cost path (CCP) 26–30 D Dariali Gorge 29 darnebi/underground shelters 119–128 Debeda River see river deposit alluvial 146 ore 15, 37–39, 45, 46–49 secondary deposit 85–86 Dmanisi 49, 110, 111, 119 E Early Iron Age see Iron Age Eastern Anatolia 53, 67, 69, 70, 76–77, 88 excavation 45, 49, 56, 57–58, 63, 70, 72, 73–75, 83–84, 87– 88, 90, 92, 93–94, 104, 108, 109, 113–115, 121–122, 124, 131, 132, 134, 139, 146, 149, 154–155, 156

Hellenistic 55, 57, 97 heritage cultural heritage management 134–135 Silk Road heritage 131, 133

14, 16–17, 132–133,

hilltop sites 39, 40–49 Hrazdan River see river I Iğdır Plain

see plain

illegal excavation see excavation intensive survey

see survey

iron 37–38, 39, 44–49, 53, 57, 105 Iron Age 37, 39, 57, 61, 72, 76, 77, 78, 79, 83, 84, 87, 93, 94, 133, 137 Early Iron Age 15, 37, 38, 39, 41, 42, 44, 45, 47, 49, 53, 54, 56, 58, 59, 61, 63, 75, 77, 78, 84, 87, 128 Islamic 11, 145, 149, 156

162

Index

K

Phasis 97–104

Kaghsi see fortress

plain Iğdır Plain 69–70, 75–76 Kars plain 69, 72, 73 Kura plain 30, 39, 49, 145 Selim plain 69–72

Karabakh 53–54, 151, 152 Karashamb

see fortress

Karmir Berd

see fortress

Kars plain

see plain

Kars plateau 67, 69, 70, 72 Kars River

see river

Kura-Araxes 72, 75–78, 87, 93, 110, 113 Kura River

see river

kurgans 53, 68, 73, 76, 92, 93, 94, 122, 156–157 Kvemo Bolnisi 38, 39, 44, 48, 49 Kvemo Kartli see region L landscape 11, 13–15, 23, 24–29, 37–39, 53–55, 63, 69, 83, 84, 85, 93–94, 97–103, 105, 111, 112, 120–121, 127, 128, 131–135, 138–141, 145–146, 150, 151, 153, 154, 157, 158 ‘landscapes of movement’ 15 landscape survey see survey

ploughsoil assemblages 94 Poladauri Valley

see valley

pottery 39, 40–48, 53, 56–63, 71–79, 83, 85–93, 102, 113, 114, 124, 126–127, 137–138, 154–156 primary discard 83, 85, 92, 93 R region Kvemo Kartli 15, 37–39, 49, 105–107, 110–113 Samtskhe-Javakheti 12, 13, 14, 16, 83–94, 119–128 Vayots Dzor 14, 15, 16, 131–141 remote sensing 11, 13, 14–16, 38–39, 55, 84, 112, 145, 155–157 Rioni River

see river

manuring 14, 85–86, 93–94

river Aras/Araxes River 13, 29–30, 53, 54, 55, 59, 67, 69, 70 Arpaçay River 69, 73 Debeda River 39, 45 Hrazdan River 16, 53, 54–59, 63 Kars River 69, 70, 72 Kura River 13, 29, 39, 55, 83, 84, 88, 92, 93, 119, 122, 124, 126, 127, 128, 146, 151, 152, 153, 154, 157 Rioni River 97, 101–103 Tərtər River 146

Mashavera Valley see valley

Roman 25, 28, 30, 97

medieval 13, 16, 23, 41, 47, 59, 61, 73, 75, 84, 86, 87, 90, 94, 97, 101, 105–115, 119, 120, 124–127, 131– 141, 145, 146, 147, 149, 150, 153, 157 medieval landscape 132, 133, 140, 141 metallurgy 37, 38, 45, 49 metal production 37, 38, 40, 44, 45, 47, 48, 49 mining 37, 41, 45, 46, 47, 48, 49 monastic sites 132, 139

route systems 15, 29–30

Late Bronze Age

see Bronze Age

Lčašen see fortress least cost path (LCP) 23, 26–29 lithic 40, 83, 85, 87, 89, 90–93 M

Mtsvane Gora 41, 44, 45 Muchannat-tapa

see fortress

S Samshvilde 13, 16, 105–115 Samtskhe-Javakheti Selim plain

see region

see plain

settlement 11, 14, 15, 37–41, 44, 45, 47–49, 53, 55, 57, 59, 63, 72, 73, 75, 83–88, 92–94, 97–104, 110, 119, 121, 122, 124, 127, 132, 134, 135, 137, 139, 146, 148, 149, 152, 153, 154, 156, 157 Silk Road 14, 15, 23, 76, 105, 131–134, 139, 141

N

Sioni Church 109–110, 113, 114, 115

Neolithic 13, 105, 113, 115

Solak see fortress

networks 23–24, 26–29, 37, 48, 107, 132–133, 137–141

South Caucasus 23, 24–30, 37, 158

osteology 112

survey 13–16, 37–41, 49, 67–70, 97, 101–104, 131–141, 155–158 extensive field survey 15, 67, 119, 121 geophysical survey 109, 113 intensive survey 15, 67, 79, 83–94

P

T

palace 108–109

terminology 12

palynology 112, 114

Tərtər River

Paravani River see river

Tghit see fortress

Peutinger Table 30

Tsitsernakaberd

O obsidian 47–48, 55, 71, 72, 74, 79, 90–92, 105, 112, 113, 114

see river see fortress

Index

U

W

Urartian 24, 39, 55–57, 63, 76, 79, 133, 138

water supply 107, 108

V

Y

valley 39, 40, 48, 49, 53, 63, 69, 72, 73, 83, 84, 85, 87, 93, 94, 105, 111, 124, 128, 132, 137, 140 Araxes Valley 53, 54, 59, 69, 70, 79 Borluk Valley 73 Chobareti Valley 84, 92 Hrazdan Valley 59, 61 Kars Valley 69 Khrami Valley 111 Kura Valley 83, 84, 88, 92, 93, 124 Marmarik Valley 61 Mashavera Valley 47, 48 Poladauri Valley 48 Yeghegis Valley 139

Yerevan 54–57

Vayots Dzor

see region

Z Zveli 83–88, 91–94

163

Authors William Anderson is an archaeologist specialising in the material culture of late antique and medieval societies. His research interests include pottery consumption, pilgrimage cults and using artefacts to interpret historic landscapes. He holds degrees from the University of East Anglia, Leiden University and the University of Melbourne. He has undertaken surveys and excavations in Turkey and Georgia, and is co-director of the Landscape Archaeology in Georgia Project. He has worked for several years in commercial archaeology in Australia and the UK. He is based in Galway, Ireland. Email: [email protected] Astghik Babajanyan received her MA in Art History and Theory from Yerevan State University in 2006 and her PhD in Archaeology with a thesis on The Ceramics of Armenia of the 14th–17th Centuries, in 2015. She is employed at the Institute of Archaeology and Ethnography of the Armenian National Academy of Sciences, as archaeologist and senior researcher; from 2017 she has also been the president of Young Scientists Council. She is author of more than ten articles related to the history and archaeology of medieval Armenia. From 2015 she started a collaborative project with Dr Kathryn Franklin (University of Chicago, Oriental Institute) on the archaeological survey of the medieval Silk Road in Vayots Dzor Region, Armenia (VDSRS). Her current research is focused on the material culture of the Aragatsotn region in the Middle Ages. Her interests include also issues of habitation and the urban system in the Ararat plain in the 9th–17th centuries. Email: [email protected] David Berikashvili is head of the Department of Archaeology, Anthropology and Art at the University of Georgia (Tbilisi). Since 2012 he has led the Samshvilde Archaeological Expedition of the University in Georgia. He has authored many scientific articles in Georgian and international journals. His interests cover the Samshvilde historical settlement (central Caucasus), Bronze Age megaliths and cave complexes in the Near East and cultural contacts between the Caucasus and Iranian world in prehistoric and medieval periods. Email: [email protected] Manuel Castelluccia studied at the University of Udine and earned a PhD in Turkish, Iranian and Central Asian Studies at the University of Naples L’Orientale. He has taken part in archaeological projects in Italy, Syria, Oman, Georgia and Armenia, and has lectured in Near Eastern and Iranian Archaeology at the University of Udine. He is also Visiting Professor at the Moscow State University. Email: [email protected] Nathaniel Erb-Satullo has been exploring landscapes of metal production in the Caucasus since 2011. His research interests include technological innovation, the organisation of craft production, and the application of natural science techniques to archaeological research. He received his PhD in Anthropology from Harvard University in 2016, and is currently a lecturer in archaeomaterials at the University of Oxford. Email: [email protected] Lara Fabian is a post-doctoral researcher in the Seminar für Alte Geschichte at Albert-Ludwigs-Universität Freiburg, where she works on the European Research Council-funded project Beyond the Silk Road: Economic Development, Frontier Zones and Inter-Imperiality in the Afro-Eurasian World Region, 300 BCE to 300 CE. She is an archaeologist researching Rome’s eastern borderlands and the spaces beyond imperial control, focusing on the Caucasus, Circum-

166

Authors

pontic, and Caspian zones. Her research looks at the relationships between local communities, the Roman and Arsacid empires, and the ‘Sarmatian’ (mobile pastoralist) world. She also explores the intellectual history of archaeology in the Russian Empire, the Soviet Union, and the contemporary Eurasian space. She is the co-director of the collaborative Azerbaijani-American Lerik in Antiquity Archaeological Project, which has since 2016 explored a highland landscape in the southern mountains of Azerbaijan. Email: [email protected] Kathryn Franklin is an anthropological archaeologist focused on material and spatial mediation of distance and difference in the medieval period. She has worked at multiple medieval sites in the Republic of Armenia for almost a decade, and received her doctorate from the Department of Anthropology at the University of Chicago in 2014. Kathryn has taught Armenian archaeology and histories of Silk Road travel at the University of Chicago, and taught Anthropology at the School of the Art Institute of Chicago; she is currently Project Manager for the Afghan Heritage Mapping Partnership, a collaborative remote geospatial project focused on the archaeological landscapes of Afghanistan, housed at the Center for Ancient Middle Eastern Landscapes at the University of Chicago Oriental Institute. For the last three years, Kathryn has been co-director (with Astghik Babajanyan) of the Vayots Dzor Silk Road Survey, a multiscalar project directed at the research and conservation of medieval Silk Road landscapes centred in Vayots Dzor, Armenia. Kathryn has published on Silk Road caravanserai in Armenia and their role in medieval social life, on the politics of assemblage related to medieval political performativity, on village spaces as articulated within social networks and along trade routes, and on the question of perceived and conceived frontiers in the medieval Caucasus.  Email: [email protected] Kristen Hopper has been exploring the landscapes of the frontiers of the Sasanian Empire in Southern Caucasia, Iran and Southeast Arabia since 2013, first as a PhD student and then as a Research Associate with the ERC-funded Persia and its Neighbours Project. She has co-directed the landscape survey of the Dariali Pass in northern Georgia, and the Gorgan Plain in northeast Iran, and has worked on fieldwork projects in Oman, Lebanon, Kuwait, Syria and Jordan. Her research interests include GIS and remote sensing applications in archaeology, landscapes of empire in the Near East and Central Asia (1st millennium BC – 1st millennium AD), and the archaeology of mobile pastoralism. She is currently a postdoctoral research associate at Durham University. Email: [email protected] Mehmet Işıklı was born in Izmir in 1968. He graduated from Ege University Faculty of Letters, Department of Archeology and Art History in 1992 and completed his Masters and PhD at Ege University, Institute of Social Sciences, between 1994 and 2005. In 1998 he worked as a Research Assistant in Ataturk University Department of Archaeology, Protohistory and Near Eastern Archaeology. He was appointed Assistant Professor at Ataturk University in 2007, became Associate Professor in 2011 and in 2015 was appointed Professor. His main area of expertise is the archaeology of eastern Anatolia, and he has participated in many excavations and surveys in the eastern and southeastern Anatolia region. He has been involved with the Van Ayanis Castle excavations since 1989, and has been the director of this excavation project since 2014. Email: [email protected] Giorgi Khaburzania is an archaeologist and data management specialist at the National Agency for Cultural Heritage Preservation of Georgia, and a graduate student at Ilia State University in Tbilisi. He is a contributor to several international archaeological and cultural heritage projects based in Georgia and Europe. His main areas of expertise are Geographic Information Systems, databases, mapping, survey methodologies, the analysis of aerial imagery, and project management. Email: [email protected] Davit Naskidashvili is currently a doctoral student at Tbilisi State University (TSU), where he also lectures. His PhD research focuses on the Greek colonisation of the Black Sea coast in the Antique Period and has been funded by a Shota Rustaveli national scholarship. He has excavated on a variety of international field projects; most recently he has been working with Oxford University at the medieval city of Bərdə in Azerbaijan and with the Persia and its Neighbours

Authors

167

Project in collaboration with the Universities of Edinburgh and Durham, for which he has been involved in fieldwork in Iran, Oman and Dariali in northern Georgia. Email: [email protected] Michelle Negus Cleary is an archaeologist and researcher at the University of Melbourne, with expertise in archaeological field surveys, GIS, excavation and settlement analyses in Central Asia, Turkey, the Caucasus and Australia. She recently received her doctorate from the University of Sydney examining settlement patterns and urbanism in ancient and late antique western Central Asia. Her research interests include landscape archaeology, settlement patterns and Eurasian mobile societies. She has published internationally in many journals and edited books. Her most recent article is titled ‘Social Complexity and Political Capitals in Ancient Eurasia’ in Oxford Handbooks Online, 2017. Email: [email protected] Mehmet Ali Özdemir was born in Kayseri in 1983 and graduated from Erzurum Ataturk University, Department of Archaeology in 2008 before commencing postgraduate studies in the Protohistory and Near Eastern Archaeology Department at the Social Science Institute in 2009, where he is currently a PhD student. He has participated in many excavation and survey projects in eastern and southeastern Anatolia, including working at the Van Ayanis Castle excavations since 2013. Email: [email protected] Abby Robinson is a PhD candidate at the University of Melbourne, and also works on the Ancient Near Eastern Studies journal and supplement series. She has taken part in archaeological fieldwork in Turkey and Georgia for more than a decade and her PhD research is based on an extensive survey in Samtskhe-Javakheti, southwest Georgia, conducted with Giorgi Khaburazania. Email: [email protected] Paul Wordsworth is a Research Fellow at the Faculty of Oriental Studies, University of Oxford and a Junior Golding Fellow Brasenose College, Oxford. His research focuses on the northern and northeast regions of the medieval Islamic world, looking at what landscape archaeology can reveal about the maintenance of routes and frontiers. He currently directs the Archaeological Exploration of Bərdə Project (AEB), exploring a provincial capital in the southeast Caucasus through urban excavation and landscape survey. He has also conducted archaeological research in Northern Afghanistan and Turkmenistan. His forthcoming book, Moving in the Margins, examines in detail the practice of travel and trade in medieval Central Asia based on data from field survey in the Karakum Desert, Turkmenistan. Email: [email protected] Ayhan Yardimiciel graduated from Istanbul University Faculty of Letters, Department of History in 2007. He completed his MA in Ancient History at the Graduate School of Social Sciences of the University of Kafkas in 2010 and obtained his PhD from the same department in 2015. Since 2016, he has been Assistant Professor at the Tourism Faculty of Kars Kafkas University. His research focuses on second millennium BC culture, and since 2016 he has been conducting archaeological surveys on the Middle Bronze Age Aras Painted Culture in Ağrı and Iğdır districts of the South Kafkas. Email: [email protected]