Nomads in archaeology
NEW STUDIES
IN
ARCHAEOLOGY
Series editors Colin Renfrew, University of Cambridge Jeremy Sablo...
274 downloads
2771 Views
18MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Nomads in archaeology
NEW STUDIES
IN
ARCHAEOLOGY
Series editors Colin Renfrew, University of Cambridge Jeremy Sabloff, University of Pittsburgh
Other titles in the series include Graham Connah ; Three Thousand Years in Africa Richard Gould, Living Archaeology Ian Hodder and Clive Orton, Spatial Analysis in Archaeology Keith Muckelroy, Maritime Archaeology Stephen Plog, Stylistic Variation in Prehistoric Ceramics Peter Wells, Culture Contact and Culture Change Ian Hodder, Symbols in Action Geoffrey Conrad and Arthur Demarest, Religion and Empire: the Dynamics of Aztec and Inca Expansionism Patrick Kirch, Evolution of the Polynesian Chiefdoms Dean Arnold, Ceramic Theory and Cultural Process Graeme Barker, Prehistoric Farming in Europe Daniel Miller, Artefacts as Categories Robin Torrence, Production and Exchange of Stone Tools Rosalind Hunter-Anderson, Prehistoric Adaptation in the American Southwest Michael Shanks and Christopher Tilley 7 Re-Constructing Archaeology Bo Graslund, The Birth of Prehistoric Chronology Ian Morris, Burial and Ancient Society: The Rise of the Greek City-State John Fox, Maya Postclassic State Formation Joseph A. Tainter, The Collapse of Complex Societies Peter Bogucki, Forest Farmers and Stock Herders Alasdair Whittle, Problems in Neolithic Archaeology Olivier de Montmillon, The Archaeology of Political Structure Steven J. Mithen, Thoughtful Foragers: A Study of Prehistoric Decision Making Robert Chapman, Emerging Complexity: The Later Prehistory of South-east Spain, Iberia and the West Mediterranean
Nomads in archaeology
Roger Cribb Central Land Council, Alice Springs
The right of the University of Cambridge to print and sell all manner of books was granted by Henry VIII in 1534. The University has printed and published continuously since 1584.
CAMBRIDGE UNIVERSITY PRESS CAMBRIDGE NEW YORK PORT CHESTER MELBOURNE SYDNEY
PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE The Pitt Building, Trumpington Street, Cambridge, United Kingdom CAMBRIDGE UNIVERSITY PRESS The Edinburgh Building, Cambridge CB2 2RU, UK 40 West 20th Street, New York NY 10011-4211, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia Ruiz de Alarcon 13,28014 Madrid, Spain Dock House, The Waterfront, Cape Town 8001, South Africa http ://www. Cambridge. org © Cambridge University Press 1991 This book is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 1991 Reprinted 1993 First paperback edition 2004 National Library of Australia cataloguing in publication data Cribb, Roger. 1948Nomads in archaeology. Bibliography. Includes index. ISBN 0 521 32881 0 hardback 1. Archaeology — Methodology. 2. Nomads. 3. Turkey — Antiquities. 4. Iran—Antiquities. I. Title. (Series: New studies in archaeology). 930.1 A catalogue recordfor this book is available from the British Library Library of Congress cataloguing in publication data Cribb, Roger, 1948Nomads in archaeology. (New studies in archaeology) Includes bibliographical references. 1. Nomads — Turkey. 2. Nomads — Iran. 3. Turkey — Antiquities. 4. Iran—Antiquities. I. Title. DR434.075 1990 935 89-25172 ISBN 0 521 32881 0 hardback ISBN 0 521 54579 X paperback
To my son, Thomas Hasan, and the memory of my father Thomas Henry
And so we moved over the earth; sometimes free, sometimes captive, conquering and conquered . . . To Anatolia we came, and before us rose Kayseri Mountain, Ararat, Siiphan, Nemrut, Binboga and Jilo mountains . They have driven us onto the dusty roads, They have cast us up into the snowcapped mountains In every province we abandoned a part of us, In every clime, in every tract of land, Discarded tents, forgotten, left to rot away . . .
(from The Legend of the Thousand Bulls
by Yashar Kemal)
CONTENTS List of illustrations List of tables Preface 1
Introduction
ix xii xiii 1
Research objectives Ethnoarchaeology Geographical scope and environment 2
Origins and definitions Historical evidence Theories of nomadic pastoral origins The time frame Nomadic pastoralism defined Hunter-gatherers and nomads compared
9
3
N o m a d pastoral e c o n o m y Nomadism and the integration of pastoralism and agriculture Factors of production: the pastoral herd The pastoral mode of subsistence
4
Residence, descent and territory 44 Nomad social organization The role of the tribe The tribe as a territorial system Nomadism as a cyclical phenomenon
5
N o m a d s - the invisible culture? Nomadism as an archaeological problem The organization of nomad material culture The role of pottery in nomad material culture Stone artifacts Stratigraphy and architectural remains The issue of 'household impedimenta' Conclusion
65
6
N o m a d architecture and domestic space What is a tent? Types of nomad tents Fixtures Composite dwellings Tents and houses The spatial organization of the tent Indications of wealth and status Towards a typology of dwelling forms Variation in dwelling types over the seasonal cycle Long-term processes influencing dwelling forms
84
7
Ali ; s camp: a n o m a d household campsite 113 The region Past nomad populations A case study Instability and opportunism in southern Anatolian nomadism Ali's camp: site structure at the household level Conclusion
23
Contents
8
The structure and location of nomad settlements
133
Nomad settlement systems The internal structure of nomad campsites The overlap of nomad and sedentary settlement forms Camps and villages Conclusion 9
10
SanaydinYayla The camp and its inhabitants The domestic complex Quantitative analysis of discard Some implications for archeological investigation The domestic complex in context
The lost world of Nemrut Dag
162
185
The setting Initial observations Interpretation of campsite ND-1 The Alikanli: an ethnographic account ND-1 in retrospect: an informed reassessment Conclusion
11
12
212 N o m a d archaeology: a n assessment Tepe Tula'i Other evidence ofpossible tentsites The Early Transcaucasian Culture: a test case Future prospects T o w a r d s a model of unstable settlement systems Theoretical perspectives Comparative and diachronic approaches The future
225
References
230
Glossary
245
Index
248
Vlll
ILLUSTRATIONS 1.1
General map of Middle East with historical and archaeological sites and modern nomad groups 2.1 Interaction between mobility and mode of subsistence 3.1 Internal dynamics of the pastoral and agricultural modes of subsistence in response to favourable and unfavourable conditions 3.2 Computer simulation of pastoral flocks showing the effect of high kill-off of immature males 3.3 Computer simulation of different herd compositions subjected to a sequence of adverse seasons 3.4 Nomadic migration in eastern Turkey 3.5 Distribution of household livestock ownership for Yoriiks and Yomut Turkmen 4.1 Schematic diagram of levels in a nomad residence hierarchy 5.1 Sherds found in campsites in western and eastern Turkey 5.2 A large ceramic churn for yoghurt manufacture; from western Iran in the last century 5.3 Grindstone and slabs from a campsite in eastern Turkey 5.4 Stratigraphic context revealed in a bread-oven pit, eastern Turkey 5.5 A long-abandoned campsite in central Turkey 6.1 Ali;s tent, southern Turkey 6.2 A topak ev tent form from central Turkey 6.3 Black tents, eastern Turkey 6.4 Frame of an alacik from the Taurus Mountains, Turkey 6.5 Diagram of an alacik showing construction and internal organization 6.6 Plastered stone hearth inside an alacik, southern Turkey 6.7 Stone-built fireplaces from tentsites in southern and eastern Turkey 6.8 Plan of a tentsite with internal and external hearth, southern Turkey 6.9 A long-abandoned tentsite with stone platform, southern Turkey 6.10 Floorplans of tentsites and house employed by villagers in western Iran 6.11 Floorplans and elevations of adjacent tentsite and hut from Baluchistan 6.12 Variability in dwelling forms among the Beritanli, eastern Turkey ix
6 17 24 30 32 37 43 46 TJ 78 79 81 81 85 87 87 89 90 92 93 94 95 98 99 102
Illustrations
6.13 Seasonal variation in Yoriik tent plan and facilities 6.14 A typology of dwelling forms: tents, composite dwellings and houses 6.15 Seasonal variation in dwelling form among villagers in western Iran 6.16 Schematic diagram of changes in dwelling form 7.1 Regional map of the southern coast of Turkey around Silifke 7.2 Plans of abandoned campsites around Karakegi, southern Turkey 7.3 Historical Yoriik migration routes in southern Turkey 7.4 Ali's winter campsite 7.5 Interior views of Ali;s dwelling 7.6 Distribution of items in surface recording of Ali's camp 7.7 Principal components analysis of spatial distributions at Ali;s camp 7.8 Spatial cluster analysis of distributions at Ali's camp 7.9 Spatial patterning of distributions in relation to structures and features of Ali's camp 8.1 Abandoned Luri campsites, Zagros Mountains, Iran 8.2 Campsite locations on mountain shelves, Turkey 8.3 Principles of campsite spatial organization 8.4 Variability in campsite organization among Bakhtiari nomads 8.5 Schematic diagram of the relationship between pasture lots and the size of campsites 8.6 A Kurdish camp in eastern Anatolia photographed in the 1930s showing stratified organization and linear structure 8.7 Beritanh chiefly camp in eastern Anatolia 8.8 Campsites located in association with ancient ruins, western Iran and eastern Anatolia 8.9 Plans of Luri encampments near archaeological mounds, western Iran 8.10 Village plans from Syria and Iraq 9.1 General plan of Sanaydin Yayla 9.2 An alacik at Sanaydin Yayla 9.3 Floorplans of tents from Sanaydin Yayla {alacik) and a nearby camp (black tents) 9.4 General view of Sanaydin Yayla looking across Area A 9.5 The domestic complex featuring structures and activities 9.6 Plan of Area A, Sanaydin Yayla 9.7 Distribution of items over recording zone A, Sanaydin Yayla 9.8 Results of principal components and spatial clustering for recording zone A 9.9 Results of principal components and spatial clustering for recording zone B 9.10 Contoured distributions of items in Area A by 5 metre quadrats 9.11 Distribution of faunal remains in Area A, Sanaydin Yayla 10.1 Map of Nemrut Dag 10.2 Abandoned campsite ND-1 from the north 10.3 Plan of campsite ND-1, Nemrut Dag, together with floorplan of tentsite
104 106 109 111 113 116 121 122 123 125 127 129 131 136 138 140 142 143 148 149 150 152 159 163 167 168 171 172 173 174 177 179 180 181 186 189 190
Illustrations
10.4 10.5 10.6 10.7 10.8 10.9 11.1 11.2 11.3 11.4 11.5
Estimation of household composition and social structure at ND-1 Nomadic migration routes in the Lake Van region Changes in Alikanli tribal structure and residence hierarchy Plan of campsite ND-47 Nemrut Dag General view of ND-4, Nemrut Dag Members of the Alikan tribe outside the headman's tent, Nemrut Dag Stone platforms and tent footings, Tepe Tula7i; Khuzistan Suspected ancient tentsites from Khuzistan A suspected nomad tentsite from the late Byzantine levels of Han Ibrahim §ah, eastern Anatolia Tent-like floorplans from the Early Transcaucasian Culture Changes in settlement structures, Early Transcaucasian Culture
XI
194 197 204 206 206 211 213 217 219 222 223
TABLES 3.1 4.1 5.1 7.1 7.2 7.3 7.4 7.5 9.1 9.2 9.3 9.4 9.5 9.6 10.1 10.2 10.3
Comparative sizes of small stockholdings of households in Near Eastern communities Functional properties of different levels of segmentation for nine Near Eastern tribal groups Inventories of Near Eastern nomad and village material culture Classification of materials recorded at Ali's camp Nearest neighbour and segregation statistics for selected materials at Ali;s camp Results of principal components analysis of Ali's camp Chi-square tests for significant difference between cluster groups at Ali;s camp Faunal elements from recording zone at Ali's camp Nearest neighbour and segregation statistics for distributions in recording zone A; Sanaydin Yayla Principal components analysis of recording zone A7 Sanaydin Yayla Principal components analysis of recording zone B, Sanaydin Yayla Comparisons between recording zones A and B Faunal elements in Area A, Sanaydin Yayla by anatomical classes Chi-square tests for significant difference between domestic complexes and special areas, in terms of faunal element classes Estimated demographic composition of ND-1 Demography and animal holdings for five wealth classes, Alikanli Distribution of households (tents) over size classes of units, Nemrut Dag
xn
35 50 70 124 126 127 129 131 175 175 178 179 181 182 192 204 208
PREFACE The convergence of three long-standing interests - a fascination with Anatolia, the phenomenon of nomadism and commitment to the progress of analytical archaeology - have come together in producing this book. Scattered observations during travels in Iran and Turkey during the mid-1970s had convinced me that pastoral campsites provided an ideal laboratory in which to pursue certain strands of emerging middle range theory, particularly with respect to occupation floor models and site structure. What might have been a detailed empirical study, carried out under controlled conditions in carefully selected sites, changed direction in the field. Confronted with the realities of ethnoarchaeological research in the modern Near East, I was not only forced to be more eclectic in terms of the campsites studied, but was also encouraged to diversify, taking an interest in Seljuk history and local history, ruined caravanserais and the accounts of nineteenth-century travellers, together with neo-Marxist anthropology and quantitative techniques of spatial analysis. While I had planned for a more in-depth coverage of a single region or group of nomads, there were compensations in the form of greater geographical range and diversity in the nomadic and semi-nomadic groups visited. The beginning of my ethnoarchaeological fieldwork coincided to the day with the October 12 military coup in 1980 which put certain constraints on my activities. Yet compared with other countries in the region, conditions in Turkey were quite favourable. Iran, in the midst of an Islamic revolution, was closed to foreign fieldworkers and the regions of Khuzistan and Luristan, where so much recent nomad ethnoarchaeology has been carried out, are now a battlefield. Afghanistan is likewise in the grip of war, and conditions on the peripheries of neighbouring states have been affected by this instability. It occurred to me that this is no accident for, as I shall argue later, nomads have been in part a response to - and also frequent contributors to - such turbulent conditions since at least the beginnings of civilization in the region. I began to realize that an archaeological perspective on nomadism would have to come to terms with the sociopolitical factors which have influenced the extent and nature of nomadism through time and that these same factors could be detected, albeit in static form, in numerous ethnographic accounts of nomads. Many individuals have played a part in the genesis of this volume. Colin Renfrew has always been a source of inspiration and was never slow to respond to my pleas for advice and direction from Ankara. Clive xm
Preface
xiv
Gamble, my supervisor at Southampton, was always on hand to get me over the rough patches. I am also grateful to the relentless Peter Ucko, not only for a heavy hand during the dissertation years but also for extorting funds from various sources on my behalf. Steve Shennan has helped to bring the study into focus and has assisted with sundry statistical and computer problems. Thank you also to the other post-graduate students at the Department of Archaeology, University of Southampton, particularly Sarah Colley and Siriol Mynors who shared my ups and downs in the last frantic year of the doctoral dissertation. Without support from the University of Southampton Studentship the study could not have continued. I recall with gratitude my colleagues from the Social Science Department at the Middle East Technical University in Ankara where I lectured for two years, and early discussions with a number of prominent Turkish scholars on this and related topics. Unpublished material and moral support was provided by Meral Ozbek, Giiven Guliiksoz and I§in Aricanli. Qok sagol to the villagers and nomads of Anatolia whose generosity and sense of humour made the experience of field work a rewarding one. I also owe a great debt of gratitude to Giilgin Cribb who helped with the study in many ways, large and small, and shared with me the rigours of Anatolian roads. I am indebted to Tim Ingold for his comments on earlier drafts of this work. Robert Whallon and Frank Hole made available unpublished manuscripts which greatly influenced my earlier approaches to the subject. Antonio Sagona of the University of Melbourne has generously provided much background material and feedback on the Early Transcaucasian Culture. Peter Andrews provided information on nomad architecture and also supplied some of the photographs. I also benefited greatly from the detailed comments of the late Professor Emrys Peters of Manchester, my external dissertation examiner. Facilities and office resources for the preparation of this volume were provided by the Department of Anthropology and Sociology at the University of Queensland.
1 INTRODUCTION Persian, I have never yet run from any man in fear and I am not doing so from you now. There is, for me, nothing unusual in what I have been doing: it is precisely the kind of life I always lead, even in times of peace. If you want to know why I will not fight, I will tell you. In our country there are no towns and no cultivated land; fear of losing which, or seeing it ravaged, might indeed provoke us to hasty battle. If however you are determined upon bloodshed . . . one thing there is for which we will fight - the tombs of our forefathers. Find those tombs and try to wreck them, and you will soon know whether or not we are willing to stand up to you. (Herodotus, The Histories, Book IV, 125)
The defiant reply of the Scythian chieftain Idanthyrsus to Darius, King of Persia, contains a number of points of interest to any student of nomadic pastoralism as an anthropological and historical problem. The ceaseless cycle of mobility, the lack of fixed assets and the military advantages which flow from both - all are common themes in the anthropology of Near Eastern nomadism. Historians receive the occasional glimpse of nomadic cultures and life ways, such as the above. But, like Darius, archaeologists have had great difficulty in pinning down ancient nomads. Indeed for many years it seemed that the only archaeological traces of nomadic cultures were likely to remain those mysterious tombs or kurgans to which Idanthyrsus referred. If, like Darius, we intend bringing nomads to boot, then we would do well to concentrate not on the obvious manifestations of nomadic material culture - the tombs, the weapons, the works of art - but instead attempt to understand nomadism as a phenomenon, as a system with its own underlying dynamics and its correlates in space and time. Had Darius possessed such an understanding then he would perhaps not have been led such a merry chase; nor would his departure from the field have been quite as undignified. Likewise the uninformed archaeologist will be at the mercy of uncontrolled observations, chance encounters and tantalizing glimpses from afar, remaining blissfully unaware of the larger pattern. He will most probably depart the field in equal disarray.
Research objectives High in the Taurus Mountains of southern Turkey, looking down on an encampment of black tents drawn up in a valley 150 metres below, I began to realize that the aim of my study was not the discovery of ancient
Nomads in archaeology
nomad cultures but something equally challenging. From my vantage point I could see that there was an underlying order in the location and orientation of tentsites. The layout of each tentsite and its associated features - forecourts, hearths, ash dumps, midden deposits - conformed to a single blueprint. I had seen and visited many such camps before and had studied published photographs and plans of campsites from Morocco to Afghanistan and Siberia to the Sudan, but perhaps not from quite the same angle. I was of course familiar with Yellen's (1977) ring model of !Kung settlements and Binford;s (1978a) Nunamiut hunting stand, but three things struck me here. First, most of the structures and features I could observe were recoverable archaeologically. Second, there were at least three distinct levels in the organization of the campsite: one set of rules governing the layout of workspace and features with respect to the tent itself, another set controlling the location of tents vis-a-vis each other as well as the terrain, and another set of factors which dictated that the campsite should be located here as opposed to a dozen other possible locations. Third, I was struck by how very different was the algebra underlying this pattern from any of the other settlement systems documented by archaeologists and by the impression - only a hypothesis at this stage - that the entire pattern was, in principle, deducible from what is known about the social and economic organization of pastoral peoples. The settlement pattern I was observing contained three interesting qualities: redundancy, 'grain7 and 'texture'. There was a high degree of redundancy in the organization of household space. Each tent, together with its work areas, features and facilities, was laid out apparently according to a single blueprint. The 'grain; was apparent in linear patterns in the location and orientation of tentsites forming parallel lines where possible but responding also to aspect and the lie of the land. The 'texture' of the settlement was a regular alternation of living floors and open spaces. There appeared to be a simple grammar underlying the spatial organization of the campsite, an understanding of which would enable a researcher to both recognize and predict the layout of other campsites. Perhaps more importantly I began to perceive that this grammar was a manifestation of the wider organization of nomadic and pastoral life. A method of series of methods, perhaps a whole framework, would be needed to explain why pastoral campsites should be distributed and organized in such a characteristic way. In common with other volumes in this series (Gould 1980; Torrence 1986) this book is about the building of middle range theory. Middle range theory involves the identification of key variables or indicators, the construction of valid instruments for the analysis of archaeological data (Binford 1983, p.129), the definition of units of analysis and the perception of appropriate scales of resolution for the solution of different problems. Middle range theory seeks to establish sets of variables which are known to interact in a certain way for a wide range of problems, although the exact nature and extent of the interaction will vary with individual applications. These do not constitute 'covering laws7 or 'laws of culture process'. A law implies a certain relationship between components E always equals me - whereas the most we can hope for in a subject like
Introduction
archaeology is some general statement that questions about energy will involve mass and some exponent of velocity. Middle range theories are not just 'middle sized' theories or small theories upon which larger theories may be built, but theories which occupy a mid-point in the research process. If we wish to propose general theories about raw materials procurement and exchange in cultures at different levels of complexity, then we may first need to recognize that the key variables involved include control over supply and production, efficiency, sophistication of technology, simplification, standardization and specialization (Torrence 1986, pp.40-5). If an example of a 'law' is the formula for relating floor size or settlement size to population (Narroll 1962; Cook and Heizer 1968) then a corresponding instance of middle range theory is the recognition that settlement size and population size or density interact with each other though the nature of this interaction varies at different population sizes and also in relation to external factors such as the mode of production (Fletcher 1981, p.98). Likewise the identification of domains and modular units in David Clarke's (1972b) study of Iron Age Glastonbury is a case of middle range theory employed in the wider task of understanding the principles of Iron Age settlement systems. Essentially, middle range theory is not about laws, but about constructs and model building. Since the inception of behavioural archaeology (Schiffer 1976) archaeologists have begun to look at sites in a new way. While Schiffer's approach was mainly confined to the reconstruction of behavioural events and distinguishing behavioural from post-depositional processes, the behaviourist perspective does not exhaust the range of possibilities opened up by the new approach. Other studies have shifted the emphasis from patterning in 'primary' refuse to regularities in the distribution of 'secondary' refuse (Schiffer 1972; Murray 1980), with a corresponding attention to the general 'maintenance of life-space' (Binford 1981) as opposed to the performance of specific tasks. If works as disparate as Yellen's (1977) account of !Kung campsites, Binford's (1978a) Eskimo hunting stand, Flannery's (1976) Mesoamerican village, Kramer's (1979) Iranian villages, Hodder's (1982) accounts of East African settlements and David Clarke's (1972b) study of Glastonbury have anything in common, it is the central notion of site structure defined most recently as 'the spatial distribution of artifacts, features and fauna on archaeological sites' (Binford 1983, p.144). This definition is inadequate in that it is not so much the distributions themselves that are important as the spatial relationships between items and/or features and relationships between the factors or variables that account for these visual patterns. Any site may be envisaged as a set of items, artifacts or debris, organized with respect to a set of features (Binford's 'framework'). The underlying control variables may be divided into: those relating to 'body mechanics' (ibid., p. 145) or kinetics which may be presumed fairly constant across space and time; environmental variables involving terrain, aspect, shelter, etc., which will also be fairly constant depending on the priorities of the site's inhabitants; and social or economic factors such as kinship distance, wealth and status which may be expected to vary between populations. Also present will be sets of variables which might be
Nomads in archaeology
termed 'cultural', and it is here that the greatest scope for variation occurs, particularly as the size and complexity of settlements increases. I am therefore less sanguine than Binford (ibid., p. 146) about being able to decipher all the information 'coded into the organization of site structure7. Cultural variables in particular are liable to produce variations in form and scale which will prove resistant to the application of 'laws ;. In any particular case it will therefore be necessary to understand the unique patterns introduced by cultural factors in addition to those accountable by kinetic, environmental and social factors. This approach to site structure implies a form of 'structuralism', which should however be distinguished from the 'structuralisme7 (Gellner 1982) of the so-called cognitive school. We are concerned with pattern and redundancy in spatial form and relationships without necessarily reading into this any notions of essentialism, that such forms 'have deep natures or constitutions or inner essences7 (ibid., p.99). Of course the kind of middle range theory advocated here could just as well be pressed into the service of 'structuraliste' or cognitive theories about human spatial behaviour, as attempted in numerous works by Hodder (e.g. 1982). The concept of site structure can be used with equal effect to support arguments about on-site behaviour and the organization of activity space on the one hand, and theories of spatial patterning based on conceptual or ritual systems on the other. A concern with site structure need not commit the archaeologist to any stance at the 'grand theory' level. However it does presuppose the use of a range of techniques for the detection and analysis of spatial patterning, which in turn place certain demands on the conduct of excavation and the manner in which data are defined, recorded, stored and processed (see Winter 1976; Bogucki and Grygiel 1981; Ammerman et al. 1978). If the New Archaeology of the late 60s and early 70s taught us anything it was that data do not exist as a priori facts but are the result of constructs which researchers use in approaching their raw material, and reflect not only the researcher's aims and presuppositions but also the general paradigm within which he or she operates. Currently a range of statistical techniques is being developed for the analysis of intra-site and regional spatial patterning (Hodder and Orton 1976; Orton 1980, 1982; Kintigh and Ammerman 1982; Whallon 1973, 1974, 1984). Attempts are even being made to address the difficult question of quantifying relationships between features and item distributions (Simek 1984). At the same time revolutions are occurring in the computer storage of spatial information. Unfortunately excavation technology and organization generally lag behind these trends, and much of the current effort in the development of site structure models has been directed into ethnoarchaeological work.
Ethnoarchaeology Like many other students of site structure I have been forced to confront the issue of the role of ethnoarchaeology. Our discipline recently went through a decade of an almost indecent concern with the use of ethnographic analogy (Binford 1972a; 1972b; 1972c; Chang 1967). While
Introduction
recourse to ad hoc ethnographic parallels will no doubt continue to inform and misinform archaeologists for years to come (Ucko 1969), arguments by ethnographic analogy have no place in middle range theory. The close association of middle range theory and ethnoarchaeology derives not so much from the use of ethnographic analogy to predict the past (Binford 1967; Durrenberger and Morrison 1977; Orme 1973; Allen and Richardson 1971; Gummerman and Phillips 1978) but from the fact that ethnographic settings provide an optimal environment in which to identify and study the behaviour of key variables, refine measuring instruments and isolate appropriate analytical units. The ethnographic setting may be likened to a laboratory in which phenomena are studied under controlled conditions. The application of constructs, instruments and models so derived to 'real7 archaeological problems and data has nothing to do with ethnographic analogy. Nor is there any necessity that they be derived from ethnographic contexts. One of the most elegant middle range models, Clarke's (1972b) 'modular unit7, was derived almost entirely from archaeological data, with a little help from Roman ethnohistory. Yellen's (1977) !Kung campsites were archaeological in the sense that most of them had no recorded systemic context, and this was true also of some of my abandoned nomad camps. Binford's (numerous publications) view of ethnoarchaeology as a laboratory for generating middle range theory by studying the relationship between archaeological 'statics7 and the 'dynamics' of a living system is a useful one. But it should not be forgotten that dynamic relationships can also be studied through archaeological data, and that these data, as opposed to the physical remains in the archaeological record, are never really 'static7. There is no reason why the study of dynamics should be confined to the ethnographic domain, although it is certainly true that understanding the operations of dynamic systems in the present can assist greatly in imparting form and meaning to archaeological residues. In documenting the spatial organization of contemporary nomads I am not attempting to map this organizational structure onto ancient nomads. I am suggesting that in prospecting for nomad sites we need to design surveys with regard to certain factors: that, for example, temporal and spatial variations in both the permanency of dwellings and the physical layout of settlements have a bearing on the detection of regional tendencies towards nomadization or sedentarization; that the investigation of suspected pastoral or nomadic sites will be more informative if certain kinds of study units and sampling schemes are employed and particular forms of spatial analysis applied. Investigations along these lines may well point to a system organized in a similar fashion to those observed today analogy, or to something quite different, anomaly (Gould 1980) - but whatever the outcome it is the tools forged in the ethnographic setting which
are imported into the past, not that setting itself. As a laboratory for the development of middle range theory pastoral campsites present certain obvious advantages. Campsites that are abandoned, but still active may be studied without the impediments imposed by continuous occupation. Campsites may be studied while occupied as well as in various stages of abandonment - i.e. they may be observed in both systemic and archaeological context, enabling us to move from the
Nomads in archaeology
realm of statics to dynamics and back again with relative ease. Although my own observations were confined to a few years, the long-term study of single campsite locations offers the prospect of detecting diachronic variations in discard patterns, architecture, settlement plans and intensity of use. As Binford (1983, p.397) has recently observed, while site structure has been approached in terms of seasonal and functional variations, There has however been essentially no discussion of long term patterns of change in the disposition of a complete system in space7.
Geographical scope and environment Fig. 1.1 General Map of Middle East with historical and archaeological sites and modern nomad groups. Modern campsites mentioned in connection with this study are also shown, distinguished according to winter or summer camps.
Before moving on to an account of contemporary nomadic pastoral systems it is necessary to define the geographical scope of the study. While evidence is drawn from the entire area of the Near East and Central Asia, from the Atlas Mountains to the Altai, the core area, for which the conclusions reached in the course of this study are held to be valid, is the mountain arc of the Taurus and Zagros, covering much of modern Turkey and Iran, together with associated low and high altitude steppelands (see Fig. 1.1).
Introduction
During one of my sojourns in a summer campsite some 1,500 metres up in the Taurus Mountains, a small shepherd boy brought me a piece of rock, explaining that it contained eski Musiilman sanat (ancient Islamic artworks). The delicate patterns inscribed in the stone were indeed reminiscent of the kind of intricate inscriptions and miniatures found in mosques, but the whole composition could be identified as a starfish, a fossil from an ancient Miocene sea. Other marine fossils of seashells and molluscs were common in rocks all over the campsite. Some millions of years ago, during the Miocene era, the tectonic plates supporting Africa and Eurasia began to come into collision, with the Arabian plate caught in between. The tremendous forces involved have raised the great mountain chains of the European Alps, the Taurus and the Zagros, across to the Hindu Kush and Himalayas, where similar processes are under way involving the Indian Subcontinent. Along much of the length of the Taurus Mountains a cross-section through the ranges follows the following sequence. Fairly precipitous limestone hills or karst formations, trending in an east-west direction, rise directly out of the sea or the coastal plain, cut by the deep gorges of streams whose catchments lie in sheltered valleys to their north. Behind this initial barrier rises an even higher escarpment with the limestone
KEY summer fr camp spring/ o D autumn camp winter camp x modern village
THE TAURUS - ZAGROS ARC LOCATION OF NOMAD CAMPSITES AND OTHER SITES 0
50
100
200
300 km
Gulf Bartholomews 1: 4 000 000
tribal migration routes town, city
archaeological site A ancient campsite
Nomads in archaeology
giving way to granite and more gently rounded summits, falling on their northern slopes to low foothills and the vast sweep of the Anatolian Plateau. Rising directly out of these steppelands are the great volcanic cones from Emir and Hasan Dag in the west, to Erciyas, Nemrut, Siiphan and Agri Dag in the east, thrown up by disturbances deep in the subduction zone where the African plate plunges beneath the Eurasian one. To the east, where the mountain chains begin to sweep around to a northwest-southeast direction, rises the limestone massif of the Jilo Mountains, a tangled maze of saw-toothed ridges and deep abysses. While a certain amount of shearing and slippage occurs along the Taurus section of the arc, in western Iran the tectonic plates meet head on, producing downwarping in Mesopotamia and the Persian Gulf and setting up a series of parallel ripples running from northwest to southeast, each a little higher than the previous one, culminating in the snow-capped peaks of the Zagros. Again the relatively open terrain of the higher altitudes, or thrust zone (Gilbert 1983), contrasts with the convoluted limestone ridges and deep, sunless gorges at the lower altitudes (Sunderland 1968). The implications for human settlement of such a landscape have been considerable. From very early times until the present, population has been concentrated on the plateau or the coastal plain and Mesopotamia. Moreover the mountain chains have served as significant cultural barriers, though by no means impenetrable ones. In sharp contrast to the situation in Arabia and Central Asia, the major access routes through the ranges, by which trade and communications are carried on, also serve as the migration routes of nomads, bringing them into continual contact with sedentary society. It is the character of this interaction, as much as the great environmental and seasonal contrasts, which has influenced nomad social structures and settlement systems.
2 ORIGINS AND DEFINITIONS The weapon is his companion . . . Who knows no submission, Who eats uncooked flesh, Who has no house in his lifetime. (from a Sumerian text, in Chiera 1924, p.20)
While the quest for the origins of nomadic pastoralism is not the central theme of this work, the issue must be raised if only because it has formed the focus of so much archaeological work. Yet the origins of nomadism may not be of quite the same significance as other origins - such as the origins of food production and of urban life. In the latter instances we are dealing with major systems transformations which require extremely complex explanations and models and which are part of a general trend towards organizational complexity and increased energy capture. The advent of nomadic pastoralism has its place in all this but if; as suggested earlier, it is peripheral to the main line of cultural evolution, then it may not require the same level of explanation. Perhaps nomadic pastoralism cannot be said to have had a precise origin, or perhaps it has had many origins in many places and at different times. If, as argued by some (Khazanov 1984; Gilbert 1983; and also in this work), nomadism is an unstable and fluctuating phenomenon involving a heavy commitment to pastoralism, then we would be far better off studying those sets of conditions under which it emerges or declines than in searching for origins or building causal models. A good starting point would be to define the necessary conditions for the existence of nomadic pastoralism. These include the presence of suitable animal species, a suitable physical environment for the maintenance of these species and the appropriate technology and social relations for domestication and continuous herding. The period in cultural evolution at which all these conditions were fulfilled may be taken as the predicted baseline for the subsequent appearance and reappearance of nomadic pastoralism. Yet existing theories of nomadic origins have tended to concentrate around two periods. It has been argued, most frequently by social anthropologists and historians, that nomadism is a consequence of the specialization and integration of pastoralism and cultivation which was
Nomads in archaeology
10
only possible with the development of complex (i.e. urban) societies (Lees and Bates 1974; Khazanov 1984, etc.). Others, predominantly archaeologists (Hole 1978; Gilbert 1983), have pushed the search for nomadic origins far back into the neolithic.
Historical evidence Although accounts of ancient Near Eastern history often seem to assume the presence of a fluid, peripheral and mobile pastoral population, they are rarely explicit on this point. To the extent that recognizably nomadic groups appear at all it is generally within the context of the wider political designs of neighbouring states whose sources tend to stress administrative categories rather than economic or social ones. Apart from terms loosely translatable as Sandbewohner, Wilden, Zeltwohnern, Wandern and Raubern (Vardiman 1977, p.22), there seems to be no ancient term which conveys the exact sense of 'nomad'. Even today, terms which initially conveyed the meaning 'nomad7 have taken on an increasingly ethnic flavour. For instance, the term Baluchi was originally employed by settled populations to refer to all nomads, but as many nomads settled the term remained as an ethnic designation (Spooner 1969, p. 150). The Turkish word ydrtik (wanderer) has also assumed a quasi-ethnic character. It is possible that a purely economic distinction between nomad and sedentary, pastoralist and cultivator, was not considered relevant in the past at least to those who kept the records. The earliest documentation of something resembling nomadism in the Near East concerns groups coming under the blanket term Amorites. These appear to have been a highly mobile, semitic-speaking population occupying the margins of the Fertile Crescent in the first half of the second millennium BC (Oppenheim 1977; Girshman 1964; Albright 1940, p.109) (see Fig. 2.1). They are described as keeping large herds of cattle, sheep, goats and asses (ibid., p.122) which confirms a pastoral emphasis, but not necessarily nomadism. Mural paintings from the 1,900 BC site of Beni Hasan portrayed a tribal chieftain named Absha with his retinue of kinsmen armed with throwing sticks and javelins and a baggage train of asses (Patai 1951, p.404). This could well be a nomadic group in migration, but there is no corresponding evidence for large flocks in attendance. It has been suggested (Roux 1966, p. 138) that the Amorites - or at least large sections of them - practised a form of sheep nomadism analogous to that pursued by modern Bedouin. Valuable insights into this period (circa 1,750 BC) have been obtained from archives discovered at Mari on the Upper Euphrates. Interpretations of these sources (Luke 1965; Rowton 1973, 1974) point to a small, vigorous kingdom deriving its income from river trade, intensive riverine agriculture and extensive pastoralism. Surrounding peoples, of an evidently tribal character - the Haneans, Suteans and Jamenites - appear to have been under varying degrees of political control. The Haneans, who seem to have formed the tribal power base for the ruling dynasty, displayed a 'semi-nomadic; pattern of residence in riverside villages, cultivating during the summer months and migrating into the steppe and northwards into the Khabur Valley during winter. The
Origins and definitions
11
other groups appear to have had more extensive migrations, taking them beyond the authority of the King of Mari for much of the year (Luke 1965). Extensive surveys and excavations in the Khabur Valley (Mallowan 1977; Oates 1976) have however failed to reveal any traces of pastoral campsites, though none were explicitly sought. Dramatic shifts in the zone of settlement within this area suggested to Oates (1976, p.235) that agriculture and large-scale settlement was only possible when state authority was able to resist nomad incursions. The period covered by the Mari archives was apparently a low point for settlement in the Khabur basin. Similarly the advent of dark ages has led historians and historical archaeologists to put forward nomad invasions as an explanation. This approach is epitomized in a volume entitled Dark Ages and Nomads, 1000 BC (Mellink (ed.) 1964) based largely on chronological arguments and exotic artworks. The theoretical poverty of this approach has led one reviewer (Barnett 1968) to remark: 'Yet a nagging doubt remains. Dark ages certainly, but nomads? Was the question precisely enough formulated, if at all?7 A useful critique of this and other migrationist themes is contained in a review article by Adams et al. (1978). Apart from lists of spoils, including large numbers of livestock taken on Assyrian or Hittite raiding expeditions, little is known of the subsistence base of the peoples inhabiting the Zagros and Taurus Mountains where most of the contemporary nomads appearing in this study are located. These people of the hills, the Gutians, Kassites, Hurrians, Mannai and Lullubi, make fleeting appearances in Mesopotamian and Anatolian history as maurauders, founders of dynasties and victims of Assyrian razzias. Archaeological evidence from settled sites in the region confirms an emphasis on pastoral production (Levine and Cuyler-Young (eds.) 1977) but beyond that we cannot go. People of known nomadic affinities, the Scythians, Cimmerians and Sarmatians, erupt fleetingly onto the scene. Historical references to them are common, and there are some indications of economic and domestic life. For instance Strabo describes the Sarmatians as '(partly) nomads and others tent-dwellers and farmers7 (Khazanov 1984), while Herodotus (IV, p. 105) reveals considerable variability in the extent of pastoralism practised by various Scythian tribes. But apart from a few arrowheads and isolated hoards (Phillips 1972; Mellink (ed.) 1964) all attempts to pin them down to recognizable assemblages or settlements have been inconclusive. They are still best known from the royal tombs to the north of our study area, although recent excavations in the Soviet Union have revealed what appear to be the remains of open campsites (Shilov, unpublished paper, World Archaeology Congress, 1986). A form of Bedouin nomadism in the piedmont zone at the foot of the Taurus Mountains was documented in Roman times (Peters 1978) during which the north-south migratory cycle occasionally interfered with the major trade routes running east-west. This is in the same region as the kingdom of Mari, arguing for a continuity of pastoral and possibly nomadic tendencies from very early historical times. Historical accounts during the Islamic Period (AD 1,000 to 1,900) provide secure evidence for a substantial nomadic presence in large areas of
Nomads in archaeology
12
the Near East, particularly during the Turkic and Mongol incursions during the eleventh, twelfth and thirteenth centuries AD (Cahen 1968; Minorsky 1978; Turan 1971; Siimer 1980). At the time of the initial Seljuk migrations into Iran and northern Mesopotamia there is some evidence of indigenous migratory pastoralism in the Zagros and Taurus foothills and conflict over the control of pasture between the new immigrants and original occupants (Lambton 1973; Bos worth 1973). The question of the presence of ancient nomadic groups, however these are defined, in the mountain regions of the Near East is left unresolved by the available historical evidence. There is enough evidence to encourage further investigation and to point to archaeology as the principal medium of inquiry.
Theories of nomadic pastoral origins Lees and Bates (1974) propose a model according to which an initial pattern of mixed dry farming and subsistence pastoralism is gradually replaced by a more specialized form of cultivation based on canal irrigation. The new system supports an increased population and, as it is labour-intensive, problems occur in the scheduling of agricultural and pastoral activities. As a consequence pastoralism becomes confined to the more marginal steppe and mountain regions where seasonal migrations of greater amplitude become necessary, but is assured of a future through the demand for specialized pastoral products from the heavily populated areas. Increasingly asymmetric power relations develop between the distinct populations associated with each mode of subsistence and, by dint of their military capabilities, nomads are sometimes able to extend political control over settled communities while the latter tend to dominate the economic and cultural spheres. A slightly different scheme involving nomadization by default has been put forward independently by Adams (1974), Wright (1977, p.388), Flannery (1972) and Spooner (1972, p.126). With the intensification of irrigation agriculture, those who were unable to cope successfully with the change might be tempted to fall back on their pastoral resources - either temporarily until new agricultural opportunities emerged or permanently as specialized pastoralism became a viable strategy. This model may be partly subsumed by that of Lees and Bates. A further variant of the Lees and Bates hypothesis has been proposed by Irons (1975), and supported by Gilbert (1983), in which the growth of urban centres in the Zagros highlands created an increased demand for pastoral products, leading to a section of the population engaging in transhumant pastoralism and ultimately nomadism. Interestingly Gilbert (1983) argues that it is the reduction in the pastoral sector, relatively speaking, which has the effect of creating scarcity, raising the relative price of animal products and opening up a niche for nomadic entrepreneurs. Some empirical support for these views comes from Hole's (1977) study of prehistoric sites in the Deh Luran Plain. He argues that an intrusive population of irrigation agriculturalists originating from northern Mesopotamia moved down the Assyrian Steppe in a steady wave of advance
Origins and definitions
13
(see Ammerman and Cavalli-Sforza 1973). By about 6,000 BC the indigenous mixed dry farming and herding subsistence pattern was displaced into areas unsuitable for irrigation: Once the higher and more dependable yields of irrigation agriculture were recognized it would be a simple step to remove most of the herding to the margins of the plain and the significantly better pastures in the mountains. Such a physical separation would have benefitted farming and herding by Keeping animals out of the fields and by allowing them to graze in cooler, lusher pastures. It is notable how much larger the animals kept by nomads today are than those kept year round in Deh Luran. (Hole 1977, p.14) The increased productivity of irrigated land then led to a greater all-round dependency on cereal resources, not only by sedentary villagers but also by pastoralists who now began to concentrate on renewable products such as milk and wool as selection for these physiological capacities occurred. These products could then be exchanged for cereals to augment the small quantities of winter wheat grown by the pastoralists. The combination of pastoralism and limited dry farming is a subsistence strategy commonly employed by nomads today. Schaffer (1972) also proposes a model of dual specialization for the Sistan region of eastern Iran. This developmental model begins with irrigation agriculture supporting village and incipient urban settlements along limited strips of alluvial land. With increased growth, the surplus population begins to occupy the arid periphery by exploiting thinly spread pastoral resources and adopting a mobile lifestyle. These ideas are reminiscent of the conclusions reached by Masson (1972) for the nearby region of Soviet Turkmenia. However Schaffer adds a further twist. Subsequent population increase is accommodated by agricultural intensification and urbanization which are both facilitated by the ready-made regional communication network established by the nomads. Nomads have even been credited with playing a vital role in the rise of civilization and state formation. During the Uruk period, according to Johnson (1973, p.159): coordination of local production and seasonal exchange with herders may thus . . . have increased decision-making requirements to the point that an expanded decision-making organization [the incipient state] was necessary. In a similar vein, Wright (1977, p.338) speculates: Could an increase in the number of pastoralists appearing every winter in the lowlands put pressure on the regulatory institutions of the settled enclaves? Ancient trade has long been linked with the activities of nomads. Ever since Perrot (1962) noticed that the distribution of traded goods in the neolithic Levant and Syria closely followed modern Bedouin migration routes, speculations along these lines have abounded. There is some historical evidence (Klengel 1977, p.164; Khazanov 1984, p.209) that pastoralists were already involved in the caravan trade of southwest Asia by the Middle and Late Bronze Age (second millennium BC) though chiefly
Nomads in archaeology
14
as guides and escorts but; as Khazanov (ibid., p.211) points out; they 'neither created this trade nor did they determine its development and its fortune7. There are certainly contemporary examples of nomadic pastoralists-cum-traders (Ferdinand 1962), and most nomads have a keen interest in trade as consumers. But attempts by archaeologists to invoke nomads as active agents in the long-distance movement of commodities and dissemination of 'information7 (Wright 1969; Schaffer 1972), or to detect the presence of ancient nomads by following the pattern of traded goods (Crawford 1978), are far from convincing. Underlying most of these nomad-trader schemes is the assumption of an underlying commerce in 'ideas7 (e.g. Hole et al. 1969, p.354). However there is little in contemporary accounts to point to nomads as an ideal medium for such exchanges. Attempts to identify nomadic cultures with certain material complexes have rarely been successful. The widespread dissemination of ceramic styles over the Sistan region, attributed to nomad entrepreneurs using the vessels as containers (Schaffer 1972), could easily be accounted for by other mechanisms. The rapid spread of black burnished pottery, together with other elements of the Early Transcaucasian Culture in the third millennium BC, has also been linked to nomadic migrations (Burney 1961; Whallon and Kantman 1969, p.130). However the Early Transcaucasian Culture is now known to contain a full range of settlement types from villages to cities and, while indicators of a mobile lifestyle are present, this is hardly a nomadic culture (see Chapter 11). All of the above scenarios have much in common. All take a holistic or systems approach to the question of origins. All emphasize specialization and integration, with only Lees and Bates raising the question of asymmetries in power. Interestingly all seem to associate the appearance and enhancement of nomadic tendencies with the growth of sedentary populations and the expansion of urban centres. While it is certainly difficult to imagine nomadic pastoralists existing entirely independently of settled and urban communities, as Khazanov (1984) and many others have suggested, I believe that the above scenarios are to some extent based on a misunderstanding of how pastoral nomad and sedentary systems interact. This in turn may stem from the mistaken concern with the search for origins mentioned above. If nomadism is to be understood as an alternative specialization which arises under certain economic and sociopolitical conditions rather than a discrete system which somehow evolved, then we would expect these conditions to follow from the above explanations. We would expect tendencies towards the consolidation of state power, economic integration and the spread of cultivation to be associated with an increase in nomadic activity and in the degree of specialization in pastoralism. While this may have been true of the initial establishment of nomadic systems in previously unoccupied marginal environments, such a correlation does not seem to hold for long-term interactions between the sedentary and nomadic sectors. In Chapters 4 and 5 I will attempt to outline why this is so, drawing on ethnohistorical sources as well as a host of ethnographic observations which, though synchronic in character, nevertheless suggest the presence of underlying diachronic tendencies.
Origins and definitions
15
The time frame It is intriguing that, while the models dealt with above apply to time periods ranging over four millennia, much the same systemic factors are mentioned in each case. In fact for Luristan and the middle Zagros Mountains nomadism appears to have had at least two origins, the first in the seventh millennium BC (Hole 1977; Goff 1966, 1981) and the second in the third millennium (Gilbert 1983). Moreover a third origin has been proposed for the region in the first millennium BC, involving the immigration of 'real7 nomads from further north to replace existing seminomads (Khazanov 1984, p.7). Much of this confusion arises from the difficulties of adequately defining nomadism as well as the nature of the data involved. Those authors who draw on archaeological and subsistence data tend to place the point of origin earlier, while those concerned with historical and ethnohistorical data prefer later dates. In particular many students of Central Asian nomadism tend to identify its origins with the first historical sources either from southwest Asia or China (Lattimore 1951; de Planhol 1959) and to see the origins of southwest Asian nomadism in terms of migration from Central Asian heartlands: 'In the first centuries of the last 1000 years BC the pastoralists and semi-nomads of the previous thousand years BC began to be replaced in certain cultural areas of western Asia by real nomads who were often migrants from other areas7 (Khazanov 1984, p.7). Where 'real7 nomads are defined in terms of complex sets of political and socioeconomic criteria, the problems of detection for the prehistorian are considerable. Evidence for developed pastoralism and seasonal habitation are insufficient to meet such definitions. In my view we should cease asking 'Were prehistoric mobile pastoralists real nomads?7 in terms of ethnographically derived models, and instead begin to ask to what extent can we evaluate certain key dimensions in production, mobility and settlement in prehistoric pastoral and agro-pastoral cultures? The parallel development and interplay of cultivation and pastoralism (Reed 1959; Sherratt 1981) in the Near Eastern food-producing revolution suggest that these key dimensions were present from a very early date.
Nomadic pastoralism defined The time at which nomadism is perceived as originating depends on one's definition of nomadism. Any such definition must involve the key factors of pastoralism and mobility, and it is the weighting given to each of these and the way in which they are measured that give rise to the many different approaches. Khazanov (1984, p.7) defines nomadism as: 'a distinct form of food-producing economy in which extensive mobile pastoralism is the predominant activity and in which the majority of the population is drawn into periodic pastoral migrations7. His approach approximates that employed here in that he gives due weighting to 'the ratio of pastoralism and agriculture7 (ibid., p.9) rather than the extent of mobility per se.
Nomads in archaeology
16
However difficulties with the scheme employed by Khazanov arise when he attempts to circumscribe and classify nomadism. An exacting set of requirements needs to be met before any group is allowed as 'real nomads7. Pastoral specialists in mixed farming are not regarded as nomadic because this represents a division of labour within the same society while nomadism requires a division of labour between different societies (ibid., p.18). 'Pure7 nomads are marked by a complete absence of agriculture, even in a supporting role; and are thereby confined to pockets in the Eurasian inner highlands and steppes, Arabia and the Sahara. Most of the peoples described as nomadic by Near Eastern ethnographers are classified as 'semi-nomadic pastoralists' characterized by extensive pastoralism in conjunction with supplementary agriculture (ibid., p. 19). 'Semisedentary pastoralism' or herdsman husbandry is village-based, while yaylag pastoralism is a specialized mountain form of herdsman husbandry (ibid., p. 23). Forms of 'household husbandry7 with or without intensive fodder feeding are fully integrated at the household level (ibid., p.24). However 'semi-nomads' are further classified into those who are specialized at the family level and those specialized at the 'sub-society7 level (ibid., p.20). To my mind such complicated and rigid typologies could be easily replaced by a single dimension specifying the level - household, village, tribal section, region - at which specialization occurs. The search for a fully nomadic society should be abandoned in favour of an approach which recognizes nomadic tendencies manifested in varying degrees in a wide range of societies and communities. Nomadic pastoralism is a dual concept comprising two logically independent dimensions - nomadism and pastoralism. Within each of these dimensions dualisms such as nomadic/sedentary, agricultural/pastoral, the desert and the sown, perpetrate gross distortions of our ability to understand the relationship between the two. Each dimension may be viewed as a continuum, and the relationship between them is best represented in terms of a probability space (see Fig. 2.1) in which groups or individuals are uniquely located with respect to each axis. While the pastoral/agricultural and nomadic/sedentary dimensions must be distinguished conceptually, this need not imply that groups or individuals are necessarily randomly located within this probability space. In fact, as indicated by the location of the groups shown in Fig. 2.1, there appears to be a strong linear relationship between the two. The greater the degree of pastoralism, the stronger the tendency towards nomadism. The association between pastoralism and nomadism is recognized implicitly in the statements of some of my informants in Anatolia: 'We have only 20 goats so we stay around the village, but cross-eyed Ahmet has 100 sheep and goats so he goes each summer to Sultan Dag. The more animals you have, the further you have to move/ I do not hesitate to discuss village pastoralism and nomadic pastoralism as part of a single phenomenon for, while the two represent very different solutions, they are nevertheless solutions in terms of the same system of constraints and opportunities. In tackling the problem of Near Eastern nomadism, the village herd may not be such a bad place to begin.
Origins and definitions
Fig. 2.7 Interaction between mobility and mode of subsistence. Schematic diagram showing a degree of correlation between the two variables. The approximate locations of a number of communities mentioned in the text are indicated with respect to both axes.
17
Fully nomajijc^lfkanli • ) Karakeqi;;
s'' Lurs'' Sanaydin MODE OF / SUBSISTENCE Agriculture
#
Mixed
Shirdasht /
/
Pastoralism
Hasanabad Shahabad Fully sedentary MOBILITY zone of high interaction •
communities mentioned in text
Pastoralism Pastoralists have been defined as 'those who are dependent chiefly on their herds of domestic stock for subsistence' (Krader 1959, p.499). This definition implies ownership and should also include an element of mutualism as the animals involved are also dependent on humans (Ingold 1980). Pastoralism is only one of the possible modes of exploiting herd animals - hunting and ranching are others (ibid., 1980) - characteristically involving protection of the herd and systematic consumption of its renewable products. However while some (e.g. Bates 1971, p.112, 1972, p.52; Vincze 1980, p.392; Goldschmidt 1979, p.16; Ingold 1980, pp.2, 25) refer to pastoralism as a 'mode of production', I prefer, in view of the particular connotations given the term by Marxist theory, to substitute the term 'mode of subsistence7. By 'mode of subsistence7 is meant: 'some configuration of productive forces and relations permitting the exploitation of natural resources and reproduction of the social groups involved' (Cribb 1984b, p.4). It does not extend to the relations of production above the reproduction of the community itself, and does not involve the 'mode of appropriating surplus labour7 so essential to a Marxist analysis (see Hindness and Hirst 1975; Friedman 1974; Godelier 1977; Parsons (ed.) 1978). Thus a
Nomads in archaeology
18
pastoral mode of subsistence may coexist with an agricultural mode of subsistence - perhaps involving the same social group or specialized sectors within a single community - under a wider mode of production. By the same token it would be possible to consider two forms of pastoralism operating under different modes of production - for example, pastoralism during the Seljuk period and that of the twentieth century - as part of a single mode of subsistence. There are certainly plenty of examples of adjacent nomads and peasants, or of pastoralists and cultivators within the same village (Sweet 1975), employing different tools and techniques, distributing land and animals according to different systems of allocation, yet who are integrated into a wider regional economy and subject to appropriation by the same ruling class whether a landed, pastoral or urban elite (see Lambton 1953; Rosenfeld 1965; W. Swidler 1972). Groups (or individuals) may therefore be ranked along an agriculturalpastoral continuum according to different criteria such as the overall level of pastoral production or the extent to which their livelihood is met through pastoral as opposed to other modes of subsistence. Small numbers of livestock may be kept by cultivators without necessarily bringing into play all the elements of a pastoral mode of subsistence. However increasing emphasis on pastoral production, accompanied by a rise in the number of animals herded, will require specialized forms of animal management and changes in household organization. The full expression of such a tendency towards pastoral accumulation will ultimately involve a household in the kinds of logistics and social networks characterized as 'nomadic'.
Nomadism While nomadism need not imply a lack of patterned movement or fixed migration cycle as suggested by some (Gulliver 1958), neither does it exactly connote a 'fixed round of movement . . . a cyclical or rhythmic movement' as suggested by Krader (1959, p.499). The presence of a regular, seasonal cycle of movement - universally attested to in the Near East and also Central Asia (see Vainshtein 1980, pp.96-7) - should be regarded as a necessary but by no means a sufficient condition for the full expression of nomadism. Viewed in the short term it is understandable why so much weight has been placed on the seasonal cycle and regular migration tracks. Such regularities loom large within the scope of the ethnographic time frame. Yet in study after study there emerge instances of departures from regular migration routes or situations in which access to these routes is only tenuously maintained (e.g. Bates 1972; Irons 1975). While the political ramifications of such events have been discussed (Bates 1972), the implications for patterns of mobility have not been explored. Viewed over the long term it becomes clear that small perturbations in seasonal migration tracks could eventually build into wholesale shifts in the distribution of migratory groups. It is this latter tendency which I would regard as being characteristically 'nomadic'. A useful distinction might be drawn between forms of tied or tethered nomadism (Ingold 1987, p. 184), fixed to a well-worn migration track, and untied nomadism, in which shifts into new migration tracks are frequent.
Origins and definitions
19
Again, however, it is not difficult to see, over an extended time frame, how the former condition might be transformed into the latter. Much of the confusion in definitions of nomadism may be traced to the use and abuse of the term 'transhumance7. It has been variously used to describe the exploitation of seasonal pastures within the same valley system by village flocks under the care of shepherds (Monteil 1959, p.579; Krader 1959, p.500), the migration of whole village communities with their flocks for a season (Johnson 1969, p.18; Swidler 1973, p.32) and the seasonal movement of nomadic pastoralists (Stauffer 1965, p.290; Stenning 1957, pp.58-9; Tapper 1974, 1979c). It has frequently been employed - implicitly and sometimes explicitly - to describe an intermediate state between sedentism and nomadism. Again the difference between the terms nomadism and transhumance is logical rather than categorical - the two operate on entirely different scales. I would suggest that the term 'transhumance' be confined to a form of livestock management making use of seasonal variations in the availability of pasture. As such it may be either vertical or horizontal, and could include all three of the patterns referred to above. Given this definition, typologies according to which groups are categorized as sedentary, nomadic or transhumant (e.g. Swidler 1973; Krader 1959) become meaningless. 'Alpenwirtschaft7 is a form of transhumance (Miiller 1938; Vincze 1980) but has nothing to do with nomadism. The transhumance of village flocks may or may not involve an element of nomadism, depending on whether the population moves with the herds. The degree of human involvement is therefore one vital index of the presence of nomadism. The other is the degree of potential mobility of the transhumance track itself. Only where a transhumance pattern involves substantial numbers of people engaged in tending and exploiting the herd for subsistence, and is potentially subject to lateral shifts into new zones at either end of the cycle, are we dealing with the upper end of the sedentary-nomadic continuum. Like 'transhumance7, 'semi-nomadism' is a much abused term. For example it has been used by Patai (1951, p.405) to cover forms of pastoralism based on sheep and goats, as opposed to 'pure7 nomads herding camels (Middle East) or horses (Central Asia), and by Bacon (1954) to denote pastoralists who import grain. Both usages confuse the separate dimensions of pastoralism and nomadism. Nomadism then reduces to the organization of migratory communities, and nomadic pastoralism focuses on those communities whose migrations are organized around the transhumant migration of livestock (see Spooner 1972, p.130). At the local level, and viewed over a short time scale, transhumance emerges as the salient characteristic of nomadic groups. It is only at the regional level and over an extended time scale that the full implications of nomadism, as distinct from transhumance, emerge with full clarity. And it is at just such a temporal and regional scale that archaeology, as opposed to social anthropology, is best equipped to approach the problem of nomadism.As recognized some time ago by Dyson-Hudson (1972), nomadism cannot be reduced to a back-and-forth movement according to the seasons. Annual removal to widely dispersed seasonal encampments is of
Nomads in archaeology
20
less interest in itself than the possibility of frequent changes in residential association which it opens up. This constant circulation and instability in the composition of residence groups has long attracted social anthropologists, although until very recently (Tapper 1979a) the reasons for this attraction were not explicitly discussed. It is this very quality which likewise affects the physical formation of nomad campsites and settlement systems and affects the character and texture of the archaeological traces into which they are transformed. It may lie at the very heart of the problem of detection and recognition of nomad sites. Four features of nomadism are therefore pertinent to this inquiry: (1) The presence of transhumant migrations of livestock accompanied by human communities - vertical or horizontal, long or short distance. (2) The possibility of fluidity in the formation of residential associations, including some turnover in group membership and both seasonal and long-term changes in the size of co-resident communities. (3) A high degree of household autonomy at the operational level of herding logistics and camp group formation. (4) The maintenance of a flexible system of rights of access to territory in accordance with frequent changes in residence and wholesale shifts in the pattern of migration tracks and demands on grazing land. More than the overt pattern of geographical movement, the overall organization of a community needs to be stressed. Essentially nomadism involves the regular migration of a community together with much of its productive base within a single ecological niche. This may occur between different environmental zones or within a single zone. According to this definition neither hunter-gatherers nor shifting cultivators are nomadic. The former may be highly mobile, but exploit a variety of resources and niches which constitute the productive base. The latter may well exploit a single niche - perhaps through swiddening - but again they do not move their productive base into the new location.
Hunter-gatherers and nomads compared In view of the potential for confusion between the archaeological traces of hunter-gatherers and pastoral nomads it is necessary to spell out the differences between their respective productive strategies and corresponding territorial systems. As Khazanov (1984, p. 15) has pointed out, the 'nomadism' involved in a system based on food procurement is quite different from that of a food-producing system. There are important differences not only in the motives which govern the pattern of movement but also in the pattern of movement itself and, above all, in the principles of site generation which apply to each. Enough is now known about the territorial behaviour of existing and recent hunter-gatherers (Lee and DeVore (eds.) 1968, 1976; Lee 1972; Woodburn 1968, 1972; Binford 1978b, 1980, 1983; Yellen 1977; Gould 1980) to permit some broad generalizations. In spite of major differences in technological complexity and range of movement (Binford 1979), the underlying themes that emerge are consistent. The texture of a huntergatherer territorial system is dominated by seasonal base camps centred on water resources and located so as to provide access to major resources.
Origins and definitions
21
To these must be added special purpose sites such as ceremonial sites, killsites and hunting-blinds. Rights in territory are acquired mainly through descent or marriage, though the actual composition of residential groups fluctuates according to seasonal resource density following lines of kinship and affiliation to estate owners. Some archaeologists, such as Foley (1981) in East Africa and Hiscock (1984) in Australia, see such territorial systems in terms of relative densities of item distributions over a continuous surface with nodes of high density rather than as a series of discrete sites. With the exception of base camps for the exploitation of abundant (e.g. marine) resources or scarce resources (e.g. dry season camps around waterholes), changes of residence are frequent and usually take place over short distances, attuned to the progression of a constantly changing resource-use schedule. While violent conflict is by no means absent from hunter-gatherer life, and while it may well arise from considerations of territory, the element of territorial aggrandizement appears to be absent. This is not to say that displacement of some groups by others does not occur, but it may be difficult to determine just what constitutes a 'group'. As Ingold (1987) has recently suggested, the fundamental difference between hunter-gatherers and nomads is that, whereas in the former case the relations of production are tied to territories or significant sites, nomadic productive relations are largely independent of place, being vested in mobile pastoral capital. While it is not true to say that nomadic pastoralists lack a concept of territory, their territorial system is geared to the acquisition of pasture rather than establishing vital symbolic links between the community and a particular locale. Accordingly, the logistics of nomadic and hunter-gatherer migration differ fundamentally. While the hunter-gatherer, whether forager or collector (Binford 1980), schedules his movements with regard to consumption, the pastoral nomad is tied to a schedule of pastoral production which is in turn dependent upon the consumption patterns of pastoral stock. Whereas the hunter-gatherer exploitation strategy is an eclectic one which incorporates a wide range of available and preferred plant and animal species, that of the pastoralist is driven by a single overriding preoccupation - the search for pasture for his flocks. This means that the hunter-gatherer;s territorial system is far more complex than that of the pastoralist - attuned as it is to the competing demands of a wide range of resources. While the hunter-gatherer will be familiar with the entire spectrum of resources and features in his landscape, the pastoralist will be acquainted with only a narrow band comprising the pastoral niche itself. In this sense he is more like an agriculturalist except that, instead of an intensively farmed plot of land, his range consists of a continuous strip of real estate over much of which he may lay only temporary claim. The pastoral nomad's mode of migration is therefore direct and certain. This is not to say that nomads are not logistically organized or incapable of following 'embedded7 strategies (Binford 1978a) nor that forms of 'multiresource' nomadism (Salzman 1971, 1972) are impossible - only that the search for pasture provides the main impetus. Given the possibilities of modern technology, the advent of a nomad commuting between campsites by truck or motorbike represents no special paradox (Bates 1980; Chatty 1980).
Nomads in archaeology
22
The fundamental differences from which many other contrasting features flow are these: (1) Hunter-gatherer migration is centred on procurement and consumption, while pastoral migration is motivated by production which is in turn tied to the consumption patterns of livestock. (2) Whereas the territorial system of the hunter-gatherer is based on moving himself towards his resources, or moving resources to people, for purposes of consumption, that of the nomad involves moving an entire infrastructure of production and consumption - including accommodation and staple foods - with little regard to the regional availability of micro-resources apart from the basics of water and firewood. (3) Whereas hunter-gatherers characteristically move in order to vary their procurement strategy to exploit a wide variety of different resources in different places and different seasons, nomadic pastoralists move in order to exploit the same basic resource - pasture - in different seasons. In a sense the main motivation behind pastoral migration is to maintain access to a single environmental niche by following its seasonal relocation from one geographical location to another. (4) For hunter-gatherers the activities of procurement and consumption are spatially and temporarily associated (i.e. consumption follows close on procurement, usually at the same place or nearby, except where long-term storage is available). For nomadic pastoralists the activities of pastoral production and human consumption are continuous and largely independent of each other (see Hesse 1984). (5) Hunter-gatherer migrations follow a complex spatial pattern covering the greater part of a well-defined territory or range, while nomad migrations are simpler, following a migration track which may cover only a small part of a number of territories. (6) Hunter-gatherer movement tends to be governed by the principle of risk-minimization, whereas the migration of pastoralists is motivated as much by a desire to optimize conditions for pastoral production as to minimize risks for the herd. (7) Along with the less complex territorial system of nomadic pastoralists, we would expect their archaeological remains to be more spatially constrained and to exhibit less functional variability than in the case of hunter-gatherer sites. The model of a base camp surrounded by special purpose sites connected to different procurement strategies holds little relevance for the archaeological study of nomadic pastoralists. (8) Finally, whereas hunter-gatherer territorial patterns and migratory behaviour tend to be fairly stable and localized, those of nomadic pastoralists incline towards marked instability and dramatic shifts in migration tracks (Stenning 1957; Ingold 1987). Displacements of large residential and political units by others are extremely common. This stems from the fact that territorial aggrandisement is a key feature of nomad life. This non-seasonal component in nomadic migration, arising from the very forces of production central to specialized pastoralism, marks the nomadic system as inherently expansionist. To fully understand the pattern of nomadic migration requires a large tapestry indeed.
3 NOMAD PASTORAL ECONOMY It should be known that differences of condition among people are the result of the different ways in which they make their living . . . Some people live by agriculture, the cultivation of vegetables and grains; others by animal husbandry, the use of sheep, cattle, goats .. . Those who live by . . . animal husbandry cannot avoid the call of the desert. (ibn Khaldun, AD 1332-1406, The Muqaddimah)
Nomadism and the integration of pastoralism and agriculture There has been much discussion concerning the differences between pastoralism and agriculture, the desert and the sown (Nelson 1973); with much stress being laid on either the complementarity of pastoral and agricultural products or the conflicts generated through competition for resources. Much of this misses the point. Pastoralism and agriculture differ not only in the things produced but more fundamentally in the nature of the productive process itself. While both modes of subsistence in a Near Eastern context are subject to large fluctuations in productivity (Adams 197A), the manner in which each system responds to these fluctuations is quite different.
The organization of pastoral and agricultural production
Whereas fluctuations in the level of agricultural production tend to be buffered and absorbed by the stable structure of a sedentary community, similar fluctuations in pastoral production are amplified throughout the organization of a pastoral community. The essential difference stems from the fact that, whereas farming productivity is influenced directly by environmental factors such a temperature, amount and timing of rainfall or overexploitation of soil, pastoral productivity is affected via the mediation of herd animals. As already indicated, pastoral success tends to result in an increase rather than a slackening of labour requirements, not so much in herding itself but in extractive activities such as milking, processing and shearing. Conversely failure and decline of pastoral capital will lead to reduced labour requirements in both herding and the processing of pastoral products. The nature of the contrast is indicated schematically in Fig. 3.1. Given a sequence of favourable years and disaster years the two 23
Nomads in archaeology
KEY Level of production _ Productive capital _ Amount of land „ Labour required ... Consumption f r. outside sources _
Fig. 3.1 Internal dynamics of the pastoral and agricultural modes of subsistence in response to favourable and unfavourable conditions. (a) An agricultural mode of subsistence tends to minimize fluctuations in productive factors. (b) Pastoralism tends to enhance the extremes in variability. Moreover the extremes tend to run together in a strong boom and bust cycle.
24
HIGH
AGRICULTURE LOW Temporary rise in m*Qt productivity due to animal losses
TIME HIGH
PASTORALISM LOW
BAD
GOOD BAD SEQUENCE OF YEARS
GOOD
systems respond as follows. As capital losses drive down productivity, the pastoralist will require less pasture land while the farmer's land holdings and requirements need not be diminished. At the same time labour requirements are greatly reduced for the pastoralist, but will remain at much the same level for the cultivator. By the same token, during a sequence of favourable years, existing levels of agricultural land and capital investment (e.g. plough animals, equipment) will tend to hold down production. For the pastoralist, all the factors of production will work together to initiate an upward spiral in productivity (see Cribb 1984c). In summary, the agricultural system emerges as a stable one which tends to absorb fluctuations in productivity through the relative constancy of land, labour and capital requirements. The pastoral system, by contrast, emerges as a highly unstable one in which requirements of land and labour tend to closely follow fluctuations in the level of capital holdings. Under favourable conditions the rapid build-up of productive capital will lead land and labour requirements upwards, while a succession of reverses will result in the under-use of pasture and the off-loading of herding labour into seasonal work. The pastoral mode of subsistence therefore tends to oscillate between upward spirals of accumulation and a vicious circle of decline (Spooner 1972, 1973). These cycles of engagement and disengagement of the factors of production are linked to highly flexible institutions for the distribution of grazing rights, the circulation of livestock and the concentration or dispersal of labour resources through impermanent residence associations (Cribb 1984c). It is this that has led some to describe pastoralism as 'fluid, marginal, transitional and unstable7 (Spooner 1972, p. 130). The terms 'marginal' and 'transitional7 are quite inappropriate, and 'instability7 should be confined to the cycle of production itself, not the social forms which help to absorb this. No assumptions should be made about the complementarity of the two modes of subsistence - they do not necessarily coexist in order to ensure a
Nomad pastoral economy
25
balanced diet or to even out ecological imbalances as has been argued by some participants in the domestication debate (e.g. Flannery 1972). Nor is there much support for the view that pastoralism is a form of resourcebanking against the uncertainty of agricultural production (Hole and Flannery 1962; Adams 1974). On the contrary, the evidence strongly suggests that pastoralists and cultivators suffer simultaneously from environmental calamities, though political instability tends to favour the pastoralist. For instance livestock numbers in the Zagros village community studied by Watson (1979) were much reduced at the same time as a sequence of crop failures had brought great hardship to the villagers. Moreover the forms of scheduling required by the two systems likewise complement at some points and contradict at others. Each represents a distinct set of options with its own rules, potentialities and constraints. Of course one or the other or both, in various measures, may be pursued by the same group or individual - the variety of agro-pastoral and specialized agricultural or pastoral systems in the Near East is limited only by the range of available environments and market niches. Pastoral and agricultural production may be integrated at a number of levels, the most basic being that of the household itself. This generally occurs where cultivation is dominant and limited numbers of livestock are kept. Although some nomadic pastoralists also practise cultivation, this is usually seasonal only and of a very primitive kind. The two modes of subsistence are often more effectively integrated at the community level and this may occur in two ways. One is to split the community into specialist sections. Tapper (1977) describes an arrangement among Pashtuns in Afghanistan whereby some families remain in the village to attend to cultivation while others - not necessarily the same ones each year specialize in pastoralism throughout any given year and migrate in summer to distant mountain pasture. Coon (1965, p.213) describes something similar in the Atlas Mountains of Morocco. Some Qashqa'i beyleh (minimal camp groups) split into two sections in summer, one of which remains to grow crops and fodder in winter quarters while the other moves with the flocks to summer pastures (Marsden 1976, p. 15). The second kind of arrangement is for the entire community to alternate between pastoral and agricultural pursuits in the course of a single annual cycle. This is a common pattern throughout the mountain regions of the Near East (see Chapter 9; also see Watson 1979, ch.8; Edmonds 1957; Edelberg 1966-7; W. Swidler 1972, 1973) and not simply a stage on the way to sedentarization. It has been widely reported among Kurdish groups in the nineteenth and early twentieth centuries (Millingen 1870; Geary 1878; Maunsell 1901; Sykes 1908; Dickson 1910; Hay 1921). Partition into agricultural and pastoral sectors subjects one part of the community to pressure towards nomadization. It is interesting that this may last for no more than a year and that the 'nomadic' and 'sedentary7 sectors of the community are actually interchangeable!
Integration at higher levels The integration of pastoralism and cultivation at higher levels allows nomadic tendencies to be more fully realized than is the case with any of the above examples. The presence of nomadic and sedentary sections of
Nomads in archaeology
26
the same tribe (see Sykes 1908) reflects a high degree of specialization and interdependence within a single political and territorial unit. Such integrated' tribes (Rowton 1974) have often been confused with a nomadic tribe in the process of sedentarization (i.e. a transient form) but do in fact represent a persistent pattern, leading Rowton to suggest that 'it is in the integrated tribe that the symbiosis between nomad and sedentary finds its fullest expression' (1974, p.8). Many tribes of the Bakhtiari; Khamseh and Qashqa'i confederacies in Iran, and the larger Kurdish tribes of eastern Turkey, are organized along these lines. Division into nomadic and sedentary sections is a persistent theme in a catalogue of Kurdish tribes drawn up at the turn of the century (Sykes 1908), although the precise levels of segmentation involved are somewhat unclear. Still higher levels of integration are possible. The terms 'dimorphic chiefdom' and 'dimorphic state7 (Rowton 1973) have been used to describe a political system where a chiefly family or dynasty installed in a town or city holds sway over a mixed population of tribal nomads (forming its power base), and non-tribal peasantry (forming its productive base). While Rowton (1973, 1974) presents cogent arguments and evidence to support the presence of such structures in the recent and ancient historical past (see also Rosenfeld 1965; Cole 1973; Woods 1976), ethnographic examples are understandably rare. Nevertheless the manner in which such an entity might evolve is readily discernible in the situation of the tribal khans of Iran: The power of the khans derives from two sources: on the one hand, from the tribe of which they are the leaders, and, on the other hand, from the land which they own. In their capacity as landowners they collect their share of the produce of the land or their rents and dues as other landowners, while in their capacity as tribal leaders they collect certain levies from their followers. (Lambton 1953, p.283) While much exchange between nomads and the sedentary sector occurs locally, and is directed primarily at the reproduction of household units (Oberling 1974, p. 18; Swift 1977), a large pastoral surplus can be an important component of regional and even national economies. For example the Qashqa'i supply Fars Province in Iran with the bulk of its mutton, leather and milk products in addition to carpets and other textiles (Oberling 1974, pp. 16-17). The extent of the nomad contribution to the Iranian economy was underlined during the period of enforced sedentarization under Reza Shah during the 1930s which resulted in severe shortages of meat and milk throughout the country which could not be made up by sedentary pastoralism (Stauffer 1965, p.285; Katouzian 1978, p.366). Historical sources suggest that exchanges at this level are not a recent phenomenon. During the first part of the nineteenth century the Lurs, Bakhtiari and Qashqa'i were supplying urban centres in western Iran with meat, milk products and textiles during the summer months (Rawlinson 1839, p.104; de Bode 1845, p.90). A number of generalizations, vital to what follows, may be drawn on the basis of the above discussion: (1) The articulation of the agricultural and pastoral modes of subsistence does not necessarily occur in a complementary fashion. The two
Nomad pastoral economy
(2)
(3) (4)
(5)
27
modes represent separate options in the pursuit of productive strategies which are responsive to both market pressures and conditions of political stability. Some support is lent to the suggestion (Lees and Bates 1974) that the origins of nomadism (i.e. the nomadic tribe) are to be sought at a time when specialization in pastoral and agricultural production became a possibility. This does not preclude the possibility of a degree of nomadism at lower levels of integration (household, village) prior to such specialization. There are indications that optimal conditions for the efficient deployment and accumulation of livestock are to be found where pastoralism is specialized above the level of the community. The emergence of nomadic features seems to relate to the level at which pastoral and agricultural production are integrated into a functional economic and ecological unit. Only at the supra-community level, where specialized agriculture and pastoralism are pursued separately by different communities within the same tribal unit or in a tribal-peasant combination, are both conditions of nomadism fulfilled - an entire community in constant association with transhumant herds and one which is free to cut its ties with a particular locality or transhumance track. Finally, the process of nomadization itself may be conceived of in terms of the integration of pastoral and agricultural subsistence modes at progressively higher levels in the sociopolitical and residence hierarchy. By the same token, sedentarization occurs when the two modes of subsistence are combined at progressively lower levels within this hierarchy.
Factors of production: the pastoral herd Large-scale pastoralism involves an essentially unstable interaction between three basic elements - herds, people and pasture (Sweet 1974; Swidler 1973, p.31; Swift 1977, p.461). These three elements may be considered at a number of analytical levels, but for the time being we are concerned with the physical properties of pasture land, the speciesspecific characteristics of the animals herded and the labour processes involved.
Pasture Statements to the effect that nomads create their own desert and steppelands wherever they move may hold a degree of truth. Historically the distribution of steppelands in the Near East and Eurasia appears to have fluctuated according to the balance of power between states and empires based on sedentary agriculture and those whose power base resided in shifting populations of herders and mixed farmers (Hyames 1972, pp.1739). Pasture is never 'natural', in the sense of being in a pristine condition. While pastoralists may make little conscious attempt at pasture improvement or modification, continuous grazing exerts selective pressure on the distribution of plant species. Overexploitation of pasture often results in a
Nomads in archaeology
28
climax vegetation unfavourable to pastoralism, characterized by minimal species diversity and selection in favour of highly toxic species (see Nyerges 1980). Pasture areas in the Near East represent a retardation of the post-Pleistocene ecological succession of vegetation communities (see Dimbleby 1976; Wright 1968). Palaeobotanical evidence for the modification of vegetation in favour of pastoral species occurs as early as the seventh millennium BC (Wright 1968; Dimbleby 1976). It is the marked seasonality in the development of pastures in different climatic zones that gives rise to regular flows of livestock over a network of pastures, and a tendency for intensively used pastures to become exhausted is one of the factors behind lateral shifts in migration routes (Stenning 1957).
Characteristics of herd animals: sheep and goats
Near Eastern conditions of semi-aridity and strong seasonality in rainfall make this region best suited for the herding of small stock, particularly sheep and goats. Complementary patterns of selective grazing by sheep and indiscriminant browsing by goats (Johnson 1969; p.9; Spooner 1972, p. 122; Nyerges 1980) make this combination of species a useful one both in terms of productivity and stock management. The species-specific tendency towards flocking among sheep (Garrard 1984; Swidler 1973, p.31) is exploited by man to concentrate animals in manageable units, while the higher mobility of goats makes them useful as flock leaders (Dahl and Hjort 1976, p.250; Nyerges 1980). A common sight along roadsides throughout the region is a tightly packed column of sheep; flanked by a phalanx of black goats strutting out ahead, with a shepherd and a dog or two bringing up the rear. Rudimentary time and motion studies indicate that the optimal size of aflockcapable of being managed by a single team of shepherd and dog falls somewhere between 250 and 400 head (Barth 1961, p.22; Dahl and Hjort 1976; W. Swidler 1972, p.74; 1973, p.28) and it has been suggested that the rate of ingestion of grass tends to be higher among flocks of this size range due to bio-social factors (Tribe 1950). The choice of sheep or goats, or some combination of the two, may also be guided by economic factors. Goats, with their higher milk yields and shorter life cycle, are more suited to household subsistence needs, while sheep and sheep products tend to command a higher market price (Spooner 1972, p.123; Stauffer 1965, p.292). This question has recently been investigated in some depth by Redding (1984). The choice of breed may also take account of productivity (Swidler 1973, p.26), but nomadic breeds are generally adapted to their cycle of movement. They tend to be larger and more robust than village breeds and good all-rounders, but less tolerant of climatic extremes (Bates 1972, p.51; Barth 1961, p.6; W. Swidler 1972, p.70). A certain amount of gene flow between nomadic and sedentary breeds occurs through mutual rustling activities (Bates 1972, p.52).
Flock demography Natural tendencies of fertility and mortality in aflockare to a large extent under cultural control. Annual fertility rates of of 0.7 to 0.9 births per adult
Nomad pastoral economy
29
female per year are common in Near Eastern flocks (see Demiruren 1974; Bates 1973, p.149; Stober 1978, p.84; Ehmann 1975, p.101) and may be even higher for goats due to the tendency to produce litters (Dahl and Hjort 1976, pp.91-3; Stauffer 1965, p.292). High natural losses of up to 30 per cent are common among neonatals (Stober 1978, p.84; Dahl and Hjort 1976, p.95) and the survivors, particularly males, are usually heavily culled in order to reduce the ratio of unproductive to productive animals. Maximum age usually corresponds with the end of reproductive life, around five or six years for sheep with a median age of three (Dahl and Hjort 19767 p.95; Bates 1973, pp.147-9). Particularly among sheep, whose breeding tends to be more under human control than that of goats, breeding is strongly seasonal, occurring between July and September, so that lambs are dropped in late winter, between January and March (Barth 1961, p.7; Bates 1973, p.147; Sweet 1974, p.93; Stober 1978, p.84). The seasonally of the breeding cycle leads to the presence of annually spaced age cohorts which, together with the effects of high fertility and the annual cull, result in herds of a distinctive age-sex structure which varies seasonally (see Hesse 1984). This structure is largely a product of human intervention designed to secure the flow of selected pastoral products in certain seasons. Flock numbers will be at their peak immediately following the lambing season. Spring and early summer are the high points in the preparation of milk products as lamb numbers are systematically reduced and the milk supply diverted to human use. At other times of the year flocks will shrink through the sale of yearlings and natural attrition, and activities of herd maintenance and reproduction predominate. Thus different seasonal settlements will contain flocks of varying demographic structures.
Fluctuations in the size of flocks Not only in the course of the seasonal cycle but also in the longer term, flocks of small stock are given to large and sudden fluctuations in size and changes in age-sex composition (see Bates 1973, pp. 150-3, and figs. 12, 13). Compared with larger stock, such as cattle or camels, they contain a higher risk factor (Dahl and Hjort 1976, p.254) balanced by a capacity for rapid recovery. The explosive growth potential of a herd of goats or flock of sheep is illustrated in Fig. 3.2 which uses computer techniques to trace the growth of a simulated flock set to moderate fertility and mortality rates, 30 per cent loss of neonatals and an annual cull of male yearlings. The simulation demonstrates that even a heavy take-off of animals for meat consumption or sale need not jeopardize reproductive success. The age structure of the flock and of the death population are controlled by the prescribed fertility and mortality rates which also provide data for the computation of rough productivity indices for meat, milk and wool (see also Payne 1973). A flock with this age-sex structure turns out to be a highly efficient productive unit (Davis 1984) and one which is capable of prodigious expansion (see Cribb 1982, 1984b, 1985, 1987). If left unchecked, such a system would quickly lead to the imposition of Malthusian population controls. The tendency towards herd expansion is a fundamental feature of the
30
Nomads in archaeology
Fig. 3.Z Computer simulation of pastoral flocks showing the effect of high kill-off of immature males. (a) A typical Near Eastern herding scenario.
FLOCKS tttfttftittt
RUN NO. 1. Low Ittature Cull. Adult Hort: 12 I Young Hort: 30 I
l u a t Nort: 10 I Birth Rate: 44 I
GROWTH YR 0 20 40 I 0I 1 I 2 I1MII 3 HIM 4 HIM 11 5 HUM 11 6 Hlflfllll 7 IIIIIflIM
P R O F I L E
NO, OF LIVESTOCK 80 100 120 240 140 ISO 200
lIliMHUM
9 10 11 12 13 14 15
iiiiiiiinuiiE iiHiinimitn jiiiiimiiiimii ifiiiinfiiiiiutiii 11II1I1UIllfi 111111111 J 0
1 + 4. + 1 1 + 1 + + 20 40 40 80 100 120 240 160 180 200 Annual growth Rate = 11.157 1 p.a.
A6€ PROFILE 0-1 I| 1-2 II
Ib.tt
2* umiiiiiiffiiiiniHiiiiiiiiiif
MORTALITY PROFILE o-i ifimiiiiiiiiifiiiiiii
1-2 inn
44,91 8.21
2+ I I
0
10 20 30 40 50 40 70 80 90 Percent 1
100
Productivity Indites 1 HEAT = 0.141 HILK = 0.182 WOOL = 0.318
pastoral mode of subsistence (Herskovits 1926; Dahl and Hjort 1976; p. 17). Viewed over a regional perspective in terms of the interests of pastoralists as a whole, it is a recipe for ecological disaster. But when considered at the level of the individual unit of production - the household - its underlying logic is difficult to refute. First, large numbers of animals, although stretched to the limit of endurance during the worst grazing season, will repay the investment of time and energy under favourable conditions. Second, since there is an element of chance in the manner in which disasters befall individual flocks, at least some of the contributors to a pastoral population explosion can hope to escape its consequences. Finally, the more livestock are maintained, the more will be likely to survive any given setback (Barth 1961, p.101; Johnson 1969, p.5). In reality the exponential expansion of flocks is subject to ecological controls which come into play long before the absolute carrying capacity
Nomad pastoral economy
Fig. 3.Z (b) Increased kill-off of immature males not only fails to slow the growth rate but also enhances efficiency.
31
FLOCKS ftimmtt RUN NO. 2. High Iwature Hale Cull.
Adult Hort: 12 Z Young Hort: 30 Z
Iitat Hort: 50 Z Birth Rate: 80 Z
GROWTH YR
HO. OF LIVESTOCK 20 40 60 80 100 120 240 160 180 200
0 0 1 2 3
PROFILE
i I 1 I
4 limiinii 5 IIIIIIIIIMI
6 7 8 9 10
HUH III! !H UlIIIIIIIBfl imiiiiiiimiE iinmiifiifiiiii iiiiiniiffiiisnju!
n iuniiffiiiiittmiMi
12 iiniiitiiiiiiiieiniEi 13 1llllfIllilllliilllH!ItIIH 14 IIIilllllllllfillliflHIl I 15 IIIIIilifISIIttllllllilH11111II11 0
20 40 60 80 100 120 240 160 180 200 Annual Growth Rate - 11.029 Z p.a.
A6E PROFILE
0-1 II t-2 If ?* iiifiiifiiiiiiiiiimsgfifii HORTftLITY PROFILE 0-1 lUllllllllllllfllll 12 I11IIIIIIII!tlIIIlllt
24 jBimnn I—
0
3 1 . IX 13. n 55.0Z
37.91 40.81
2i.«
k — t — + — + —I—t—i—f—t—+
10 20 30 40 50 60 70 80 90 Percent I
100
Productivity Indites : HEAT = 0.276 H1LK = 0,2B5 HODL = 0.275
of an environment is approached. Epidemics tend to hit hardest when livestock are in poor condition as a result of grazing shortages. Frosts and unseasonal snowfall can inflict high losses on flocks which are being advanced or retarded along the normal migration schedule to take advantage of marginal grazing (Bates 1972, p.51). Disasters such as drought or epidemic - often in combination - or a particularly severe winter tend to affect herds generally, while others such as attack by predators, theft and local weather conditions take a more selective toll. According to Hole (1978, p. 142), two out of every five years in Khuzistan are likely to be disaster years and may result in the loss of more than half of livestock holdings. This general pattern is borne out by other ethnographers (Bates 1973, pp.150-3; Barth 1961, p.103; Irons 1975, p.156; Shahrani 1979, p.91). As the Mongolian proverb has it: 'A herd belongs to the first adverse season or powerful enemy7 (from Khazanov 1984).
Nomads in archaeology
Fig. 3.3 Computer simulation of different herd compositions subjected to a sequence of adverse seasons. The herding scenarios are run over the same sequence of good and bad years. (a) Herding scenario set to the same parameters as Fig. 3.2b. (b) Increase in the kill-off of breeders, compensated by rising birthrate: the herd begins to decline. 3
32
The profound effect that such factors can have on individual livestock holdings among the Yoriiks of southern Turkey has been demonstrated by Bates (1973). Over a three-year period the fluctuations were dramatic, with virtually all households being affected by the severe winter of 1968, but with large herd owners weathering these setbacks more easily than small herd owners. The growth of a herding system consisting of herds whose age-sex structure is under human control has been modelled through computer simulation. While the composition and rate of growth of a flock are determined by factors endogenous to the system and under the control of man, limitations on flock growth of the kind discussed above are exogenous, deriving from the physical environment. Fig. 3.3
FLOCKS
FLOCKS
tifftftftili
ftttftttttf
RUN NO. 1. Parameters as for Fig. 3.2b
RUN NO. 4 . High Adult H o r t a l i t y , Reducing Young M o r t a l i t y .
Adult Hort: 12 Z Young Hort: 30 Z
Adult Hort: 20 Z Young Hort: 10 Z
liiat Hort: 50 Z Birth Rate: 80 Z
B ft 0 N T H
PROFILE
GROWTH
NO. OF LIVESTOCK 100 120 240140
180 200
-•-
0 1 I 2 I inmiiii 3 IIIlllllilSlifUJM 4 IIIIIIIIIfHH 5 lIlillflillflMM 6 1 Hill
7 iniiiuiiifiim
9 Illflllllttll
10 lummmiii M IffillllifMlttl
12 IHHIIIIUIHIIII
n inmiinmii M ininnnimiu
is i6 t7 IB 19 20
i minimi iinnmn imnmi iHiiinn IHJmilll
0
i M a t Hort: 50 Z B i r t h Rate: 80 Z PROFILE
YR
HO, OF LIVESTOCK 60 80 100 120 240 160 1B0 200 0 20 ._-+ * + 4. * 1 1 1 I 0I I 2 inillHIIMHIilt 3 ItlHItflllliiIMM 4 IIIIIIIIIIBMM 5 lllllllllllBMIt 6 lllllllllfllBMU 7 iniiiiiiinimiiM 8 unuiflii 9 ininimi 10 I nun 11 iiiimiHtn 12 illinium it 13 inniimt 14 liniiiiiiii IS llffllllltil 16 I 17
20 40 60 80 100 120 240 160 180 200 Annual Growth Rate = 12.903 I p.a.
0
20 40 60 80 100 120 240 160 180 200 Annual Growth Rate - 9.804 Z p,a.
AGE PROFILE
31.41 14.3Z 54.31 HORTALITY PROFILE
o-i imiiniiniiiiiin
38.21
41.21 20.61
1-2 l l l l l l l l l l l l l l l l l l l f l l
2* iwmaam I 0
-+ ¥ f + + 1. 60 10 20 30 40 50 Percent I
•
f
f
70 80 90
f
Productivity Indices : HEAT = 0.242 HILK = 0,281 WOOL « 0.271
16.11 48,21
HORTALITY PROFILE 0-1 U t t H I 1-2 UllltttfIllllllllltltlllll
2*
11.81 52.91
••Illlllllinill
I
100
35.71
0-1 I i-2 iiniiui 2+ I
0
+
4.
+
+
1
35.3Z
4.
4.
+
+
10 20 30 40 50 iO 70 80 90 Percent Z
1
100
Productivity Indices : HEAT = 0.28A KILK = 0.214 W 0 L = 0.241
33
Nomad pastoral economy
Fig. 3.3 (c) Increased adult kill-off balanced by further enhanced birth rate and reductions in mortality of young and immatures: even so the flock falls well below the level of viability, (d) Optimal herding strategy, conserving breeders, killing off immature males.
shows how differently constituted flocks fare over the same sequence of good and bad years. The pattern of exponential growth punctuated by periodic setbacks illustrates a number of interesting properties of an intensive ova-caprine herding system. First, any herd with a moderate natural growth rate can adequately cope with a series of setbacks. Second, disasters tend to be most debilitating when they occur closely together before the regeneration process is fully under way. Finally, those flocks which contain the highest proportion of breeding stock are best able to withstand a series of setbacks, vindicating conventional pastoral wisdom. Over the long term, and on a regional scale, animal populations nomadic or otherwise - are constrained by severe and periodic exogenous FLOCKS
FLOCKS
tttmttftftt
iittttfiiti
RUN NO. 3. Increased Adult Mortality, Higher Birthrate,
RUN NO. 2. Decreased Adult Mortality.
Adult Mort: 20 Z Young Hort: 30 Z
Adult Mort: 10 I Young Mort: 30 Z
I Mat Mort: 50 Z Birth Rate: 90 Z
GROWTH YR 0
20 4 0
Uiat Mort: 50 Z Birth Rate: 80 Z
GROWTH
PROFILE
NO. OF LIVESTOCK 60 BO 100 120 240 160 180 200
I 0
PROFILE
NO. OF LIVESTOCK 20 40 60 80 100 120 240 160 180 200
0I
0 I 1 I 2 llllllllltilltll 3 1111111111111111
i ii nun uiiiiiiuimmi
IIIIIIIIIIIIIRHI!
imimiiimin
i illinium i 5 minium i
liiiiiiiuiiimm iflllllllllllHMIIH i if f iiiiniHf ssn 111111 HIM iiiimiiimni niiiiiiiiiuiui IllllllllllllfllHIII IIIIIIIIIIIIIIIBUIIII lUIHIIIIIIflllHI imiiiimiiimiiiii I HIM III 111 i l l i n i u m ii i IllfllllllHUII IIIIIIMHUI timuuimiii
6 iminium it 7 illinium ii ii
ilium niiini Hum it lima 11
n
UIIIMIIIHIIIII
0
ftSE PROFILE o-i iifiiiimififiiif
33.31 15.41 51.31
1-2 innIHI
2* i iiiiiiifiKitHliiiuiiiii MORTALITY PROFILE
33,31 40.0Z 26.72
01 Illlllflllllllllll 1-2 Illlfllllttllllllilll 2+ 0
0
20 40 60 80 100 120 240 160 180 200 Annual Growth Rate - 8,333 I p.a.
10 20 30 40 50 60 70 Percent I
80
100
Productivity Indices : HEAT = 0.321 HILK = 0.295 WOOL « 0.256
AGE 01 1-2 2*
20 40 60 80100 120 240 160 ISO 200 Annual Growth Rate 13.953 t p . a .
PROFILE tlllflfllllllfffl Illlllli IlllllflllllllllllllllHIHIl
31.31 13.61 55. It
MORTALITY PROFILE 0-1 IllllllllflllllUlfU 1-2 U11IIiIIIIlll 11UIIII 24 IHIIini 0
40.4Z 42.6Z 17.0Z
10 20 30 40 50 60 70 80 90 Percent I
100
Productivity Indices : MEAT « 0.255 MLK • 0.286 WOOL = 0.276
Nomads in archaeology
34
controls. However the high reproductive potential of small stock permits a rapid recovery to previous levels. This alternating 'boom and bust7 cycle (see Stauffer 1965, p.292), the corollary of a high accumulation and highrisk strategy, is fundamental to any understanding of a nomadic pastoral system based on the husbandry of small stock. Its effects ramify outwards into other areas of nomadic life, including the structure of co-resident communities and tenure of grazing land.
The pastoral mode of subsistence Animals as capital The singular fact about animals under a pastoral mode of subsistence is that they are owned, and for most Near Eastern pastoral systems it is the individual family that constitutes both the unit of ownership and the unit of production. Digard (1975) seems to be alone in regarding the basic unit of production as the cooperative herding unit/ which is perhaps better understood as a management unit. This is not to deny the presence of systems of pastoral tenancy whereby animals are lent, leased and placed in the care of shepherds or contract herders - we shall encounter such systems presently. Nor does it necessarily imply egalitarianism in nomad political structures. What it does mean is that in understanding pastoral herds as a form of capital we need to be aware of those limiting properties which constrain the scope of pastoral accumulation and transactions. For this reason I would regard pastoral flocks as a form of restricted capital. While the term 'pastoral capital' is not to be taken in the same sense as market capitalism, I would defend its use to represent a factor of production alongside land and labour, based on the definition of capital as 'a resource in respect of which one controls its reproductive value7 (Paine 1971, p.158). But while herds may be described as 'fields on the hoof7 (Ekvall 1968), there are major differences between pastoral and agricultural capital. The above account of herd demography has made it clear that herd animals - particularly small stock - constitute an extremely volatile form of capital. However pastoral capital is not subject to diminution through inheritance as are, for example, peasant land holdings (see Spooner 1964, p.65, 1973). Given the necessary conditions for pastoral accumulation, each division of a herd among sons can be offset through a 'topping up7 process - so that a herd of 100 animals divided among three sons could be expected to grow into three herds of 100 each. If the analogy of a finite cake cut into ever smaller segments with each generation is appropriate to an agricultural system, then its pastoral counterpart would be the process of biological cell division by which each division results in new units similar in size and structure to the original. While the numbers of animals held per household vary greatly both within and between groups of Near Eastern pastoralists, it would appear that between 50 and 100 sheep and/or goats is a normal complement for the contemporary household. Information on average herd/flock numbers gleaned from a range of Near Eastern sources is presented in Table 3.1. It
Nomad pastoral economy
35
should be borne in mind that; due to variations in household size and composition between nomadic groups, and between nomads and villagers, the average number of animals per family is not always a reliable indicator of relative livestock wealth. Correlations between wealth (in terms of livestock numbers) and household size; such as those observed by Bates (1973, pp.182-3) and Irons (1975, pp.163-3) should be reconsidered in this light. For this reason Be§ikqi (1969) has employed an index of average livestock holdings per person as a measure of household wealth and as a basis for comparison with smaller village households. Nevertheless Table 3.1 indicates that the average numbers held by nomadic groups are consistently a few orders of magnitude larger than those held by villagers in the same regions, although the numbers per village household in one region may approach - though rarely exceed - those held by nomads in another. The ratio of nomad flocks to village flocks for most regions appears to hover around 5:1. There is a sizeable literature addressed to the question of optimal household herd size (Stenning 1958; W. Swidler 1972; Bates 1973; Dahl and Hjort 1976; Redding 1984; Rubel 1969) all of which stresses the crucial role of the size and consistency of the labour supply and the division of household labour along age-sex lines. But while the household may be the unit of ownership and the key institution allocating roles and responsibility in pastoral labour, it would Table 3.1a Comparative sizes of small stockholdings (sheep and goats) of households in Near Eastern communities (nomadic communities) Group Afshar Alikanli Baharvand Baharvand Bakhtiari Basseri Bedouin Boyr Ahmad
Jaf
Kirghiz Kirghiz Qashqa'i Qashqa'i
Sagikara Yomut Yoriiks Yoriiks Unspecified Unspecified Unspecified Unspecified
Region Kerman, Iran E. Anatolia Luristan Luristan W. Iran Fars, Iran N. Syria W. Iran N.E. Iraq Pamirs Afghan Pamirs Fars, Iran (using 'hired' shepherds) Central Anatolia Gurgan, Iran Ala Dag, Anatolia Ala Dag, Anatolia
Average no. per household
Date of observation
Source
100 (S) 160.4 (S) 200 (SG) 40 (SG) 50 (SG) 80-160 (S) 100 (S) 45 (SG) 21 (S) 95 (S) 200-1200 (S) 140 (S)
1970s 1965 early 20th C 1970s 1970s 1950s 1960s 1970 1922 1890s 1970s 1970s
Stober (1978: 84) Besjkgi (1969:124) Baharvand (1975:59) Hole (1979:146) Ehmann (1975:101,121) Barth (1961:13) Sweet (1974:100) Fazel (1973:132,148) Edmonds (1957:148) Hedin (18981:177) Shahrani (1975: 89) Beck (1981:349)
235 (S) 268 (S) 94 (S) 80 (SG) min. 40, max. 600-700 (SG) 50 (SG)
1970s 1969 1970s 1930s
Beck (1981:349) Bates (1973:148) Irons (1975: Appendix 1) Frodin (1943-44: 243)
1930s 1960s
Frodin (1943-44: 243) Smith (1971:132)
1815-16 1815-16 1890s
Morier (1837:25) Morier (1837:25) Hedin (18981: 272)
Iran Persia (middle herd owner) 100 (S) Persia (wealthy) 1000 (S) 1200 (SG) Turkestan (rich)
S = sheep; G = goats; SG = sheep and goats
Nomads in archaeology
Table 3.1b
36
Comparative sizes of small stockholdings (sheep and goats) of households in Near Eastern communities (village communities).
Village
Area
Average no. per household
Ahlat Asadabad Batman Burkhaar Qizre Hazro Giingiis Hasanabad Haydari Hizan Idil Joushaqan Kozluk Mountain villages Mukti Pervan Qom-Yazd
E. Anatolia Iran E. Anatolia Iran (wealthy) E. Anatolia E. Anatolia E. Anatolia W. Iran Iran E. Anatolia E. Anatolia Iran E. Anatolia Central Anatolia E. Anatolia E. Anatolia Iran
29.9 (SG) 4 - 6 (SG) 11.3 (SG) 1000-1200 (SG) 20.8 8.3 (SG) 5.2 (SG) 21.0 (SG) 5-10 (G) 14.9 (SG) 25.7 (SG) 50-60 (G) 8.1 (SG) 25 (SG) 12.6 (SG) 13.3 (SG) a few goats
Saggiz Shirdasht* Silopi Silvan Yassihoyiik
Iran W. Iran E. Anatolia E. Anatolia W. Anatolia
10-20 (SG) 52 (SG) 21.5 (SG) 7.0 (SG) 31 (S)
Date of observation
Source
1965 1950s 1965 1950s 1965 1965 1965 1960s 1950s 1965 1965 1950s 1965 1969 1965 1965 1950s
Besjkcj (1969:124) Lambton (1953:353) Besjkgi (1969:124) Lambton (1953:353) Be§ikcj (1969:124) Besjkcj (1969:124) Besjkcj (1969:124) Watson (1979: table 4.1) Lambton (1953:353) Be§ikqi (1969:124) Besjkcj (1969:124) Lambton (1953:353) Besjkcj (1969:124) Bates (1973:20) Besjkcj (1969:124) Besjkcj (1969:124) Lambton (1953:353)
1950s 1960s 1965 1965 1965
Lambton (1953:353) Watson (1979:249) Be§ikgi (1969:124) Besjkcj (1969:124) METU (1965:82)
/ 1 i n *.n 9fifl\ ^Up 10 L\J\J)
*Migrating in summer. S = sheep; G = goats; SG = sheep and goats
be misleading to envisage each household as an isolated unit. Although economic self-sufficiency is a positive virtue, household flocks are frequently subsidized far in excess of the resources of an individual household. Kin support and patron-client relationships are often crucial factors in household viability. Furthermore the very mobilization of labour resources for the management of flocks and the exploitation and control of territory involve households in wider cooperative groupings. It is necessary here to draw an important distinction between what I have termed the productive unit or unit of ownership in pastoral economy (i.e. the household) and units of livestock management or rangeland management. The flock whose straggling passage across the AnkaraAdana highway interrupts your Near Eastern odyssey is almost certainly not the holdings of a single family but a composite of sections drafted from a number of household flocks, usually by age-sex criteria, in order to achieve a certain size and mix of old and young, male and female or sheep and goat (Fig.3.4). The shepherds and dogs will, in all probability, come from different households from the animals themselves. This will be so whether the flock originates from a settled village or a tent encampment, though is somewhat less likely in the latter case. Some authors, such as Nyerges (1980) and Vincze (1980), have concentrated on the actual mechanics of herding, the tempo of flock movements, grazing
Nomad pastoral economy
37
Fig. 3.4 Kurdish nomads passing through a village on the autumn migration trail, eastern Turkey. I was struck by the similarities in dress and material culture between nomads and villagers inhabiting the same region.
patterns and work shifts of shepherds (see also Watson 1979; Middle East Technical University 1965). The basic techniques and practices seem to vary little from camp to village (see Sweet 1974). Studies of the formation of camp groups such as that of W. Swidler (1972) have laid great stress on logistical factors such as the need to consolidate small family holdings into a single herd and the formation of effective shepherding teams, pointing to features of nomad residence groups which facilitate these. It should be stressed that such organization occurs in order to exercise management options and that it may tell us little about the driving forces behind nomad economies. A distinction has been drawn between the dual roles of herder and husbandman (Paine 1971; Vincze 1980), and Tapper (1979a) has even suggested employing the term 'flock' for units of livestock owned by a household, reserving 'herd7 for the kind of management unit described above. These distinctions reflect a more fundamental one between the level of technique, logistics and labour management on the one hand - the 'forces' of production in Marxist terms - and that of long-term strategic factors, access to resources and accumulation - relations of production - on the other.
Characteristics of pastoral labour In the first instance, the accumulation of pastoral capital is limited by the availability of herding labour, though this need not always be drawn directly from the nuclear family. Pastoral labour has three distinctive properties. First, as noted by Barth (1973a, p. 15), labour is directed primarily at capital maintenance, the success of which incurs more rather than less effort. Second, pastoral labour can be intensified only in so far as it serves to concentrate the ownership of livestock within certain households. It does not permit any given region to support a greater number of animals (Bates and Lees 1977, p.836). Third, pastoral labour is highly
Nomads in archaeology
38
competitive. Herding is a zero sum game in which one herder's gain is usually another's loss, except under conditions of herd expansion during recovery from an extensive drought or epidemic. Central to an understanding of the economics of pastoral nomadism is the concept of household viability (Stenning 1958). The achievement of a satisfactory balance between labour and capital, personnel and flocks, emerges as the primary constituent in household viability. Where household composition is such as to provide an adequate and well-balanced labour pool, capital may be adjusted to labour through flock expansion or direct acquisition of livestock. Alternatively labour may be adjusted to capital as where, for example, an increase in flock numbers provides the necessary wealth to hire additional herding labour. The effectiveness of the primary labour pool, that provided by the household itself, depends on both the number and sex of children born and on the stage occupied by the household in terms of its developmental cycle (Goody (ed.) 1958). According to Barhavand (1975, pp.60-1), a moderately well-off Luri household with 200 sheep requires a minimum of one elder son as herder, one younger son to look after the lambs, an adult male to manage cows and pack animals, two or three women or girls for milking and domestic tasks and one or two men for cultivation. Clearly a perfect balance is possible only at a certain stage in the developmental cycle of any given nuclear family. An optimal strategy aimed at ironing out imbalances as a family moves through its domestic cycle would be, in effect, to run the developmental cycles of successive generations into one another by delaying the economic independence of new nuclear family units (i.e. maintaining an extended family structure). Bates (1973, pp. 17789) presents data from Turkish Yoriik nomads which suggest that the optimal ratio of workers to consumers and optimal productivity per unit labour are achieved in his third wealth quartile (those owning 250-349 sheep), which also contains a higher than normal proportion of extended families. Note also that such flocks lie within the range (250-400) for efficient herding in a single household-based management unit. Under a strongly virilocal system involving payments of brideprice, the flow of women in marriage is accompanied by a counter flow of livestock, often resulting in temporary imbalances between livestock and labour. Each marriage of a son causes a depletion in pastoral capital through brideprice payment but brings in a new unit of female labour. A family with many sons will thus experience a series of losses of livestock but will retain a core of pastoral labour - together with additional female labour - capable of taking up the slack as flocks regenerate. However a household with only daughters will experience a series of windfalls in livestock but will lack the herdsmen to manage them. This point has been well illustrated by Irons (1975, pp. 164-7) in a comparison between the life cycle of two Yomut families, one with five sons and a daughter, the other with five daughters and two sons. Again options centre on the scheduling of marriages and the rate of fissioning of new nuclear units. These problems, common to all Near Eastern pastoral nomads, have invited a range of responses. According to Irons (1975), the marriages of sons among the Yomut Turkmen are delayed for as long as possible, thereby retaining a pool of herding labour capable of amassing a parental flock large enough
Nomad pastoral economy
39
to sustain a series of brideprice payments. Subsequently, delayed cohabitation of spouses and co-residence in the parental household continue to enhance the concentration of herding labour. While such a strategy may well serve as a form of population control, it also paradoxically tends to maximize the personnel within each extended household, thereby 'minimizing the number of households that fall below the the viable limit in terms of their labour resources' (Irons 1975, p. 169). By way of contrast, Barth (1961, pp.18-29) describes the Basseri as facilitating early marriage and any early splitting off of new households - a practice which seems to be associated with a higher rate of population increase, a higher proportion of non-viable households and marked trends towards sedentarization among such households (see also Irons 1975, p.99). However no theory of household viability can be adequate which confines itself to the internal resources of each household. While certainly autonomous in terms of decision-making and the disposal of resources, nomad households do not stand alone. Each participates in an extensive support network based on kinship, affinality, patronage and voluntary association. Even severe shortfalls in capital and labour need not result in a loss of viability where the support of kinsmen can be relied on. An anecdote from the Qashqa'i will serve as illustration. Ahmet Bey's herds were devastated by a freak snowstorm in one of the treacherous passes of the Zagros. Within a few days each household in the beyleh (camp group) had contributed one or more sheep, according to means, to make up a new herd equal to the original. Other resources, particularly seasonal labour, may serve as a backup for nomad families fallen on hard times - in recent years the petroleum industry has played this role in southwestern Iran (Katouzian 1978). Salzman (1972) even talks of a form of 'multiple resource nomadism7 in which no particular resource base is predominant. Where pastures are owned, as among the Shahsevan (Tapper 1979b), speculation in pasture land can preserve viability where other resources fail. Only where all three major components - herds, kin support and land - are lacking does loss of viability and sedentarization become inevitable among the Shahsevan. The most important options in ironing out imbalances in labour and capital may be summarized as follows: (1) Internal adjustments such as delays in marriage and/or household segmentation or adoption. (2) The formation of cooperative herding units to facilitate mutually beneficial exchanges of labour and deployments of livestock. (3) Hiring of shepherding labour from outside the household unit. (4) Contract herding involving the temporary consignment of flocks to other households, together with usufruct rights over perishable products. (5) Material or labour support from more affluent kinsmen. A related process, usually involving non-kin or distant kin, is the formation of patron-client relationships. (6) Recourse to alternative resources such as land, seasonal work, military adventurism or commercial enterprises. The range of responses of ethnographically documented pastoralists to imbalances of household labour and capital appears to reflect underlying differences in economic and social organization which may vary not only
Nomads in archaeology
40
geographically but also temporally in response to changing political and market conditions. Furthermore the involvement of households in associations beyond the extended family and herding group appear to correlate with degrees of inequality in the distribution of wealth and status.
The egalitarian response: the Basseri The Basseri, as described by Barth (1961), epitomize a form of egalitarianism sometimes mistaken for the norm among pastoralists but indicative of a certain solution to the balancing of the key factors of production labour, capital and land. Organized into self-sufficient household units, mostly nuclear families, the Basseri are dependent on labour from within the household, maintained by a high birthrate but limited by early marriage and fissioning of new nuclear units. Household viability is pegged to a minimum capital requirement of around 60 sheep, and labour and capital are adjusted through the formation of cooperative herding units consolidating the flocks of two to five families into an optimal herd of 200 to 400. According to Barth, those whose flocks exceed this size may be forced to rely on hired shepherds or contract herders, the inherent inefficiency of which imposes limitations on the further accumulation of livestock. The pastoral economy is therefore seen as being held within limits defined minimally by the capital necessary to remain viable, and maximally at that point where labour efficiency and herd productivity are subject to diminishing returns (but see Huntington 1972). At the lower end of the range households are forced out of the nomadic sector into the lower levels of village society, at the upper end a few large herd owners are persuaded to convert their pastoral holdings into a more secure investment in the form of agricultural land, entering settled society as landlords. Through attrition at both ends of the wealth spectrum a certain stable ratio of humans to animals and of both to land is maintained. We might expect a characteristic negatively skewed distribution curve of animals among households - a preponderance of small herd owners tapering off to a few big-time herders. According to Barth (1961, p.126) the human and animal populations are regulated through Malthusian controls acting directly on the flocks via over-grazing and epidemic, and indirectly on the human population via enforced sedentarization as flock numbers fall below the minimum threshold for economic viability.
The stratified response: the Qashqa'i More recent work among southwest Asian pastoralists (Irons 1975; Beck 1980; Bates 1973; Shahrani 1979) suggests that other responses are possible. Under the right conditions nomadic herding can be big business. The claim made by Barth (1961, p. 103) that the 'net productivity rate for the owner declines as the size of the herd increases7 can no longer be accepted as an inherent property of pastoral capital. The initiation of a spiral of accumulation seems to be related to the development of elaborate systems of shepherding, contract herding and pastoral tenure involving the alienation of livestock, paralleled by the emergence of large
Nomad pastoral economy
41
differentials in wealth and the rudiments of social stratification (Black 1972; Garthwaite 1978; Digard 1979; Beck 1980; Shahrani 1979). Elements of this type of response are to be found among the Qashqa'i, neighbours of the Basseri. Whether the contrasts stem from real 'cultural7 differences, or the particular window which the respective ethnographers had on each society, or historical changes between the 1950s and the 1970s, is difficult to assess. According to Beck (1980), Qashqa'i households mostly comprise nuclear families, as among the Basseri. But the Qashqa'i make extensive use of shepherd labour to make up for labour shortages within the household. Household flock sizes tend to be relatively high (see Table 3.1), approaching the range for optimal herding efficiency, and the banding together of households into herding units is not much in evidence. Moreover there seems to be no minimal capital threshold below which households are forced into sedentarization - kin support is available, or shepherding contracts may be entered into with wealthy families. But shepherds are not necessarily prospective herd owners going through the initial stages of flock accumulation as occurs in many other pastoral societies (Yoriiks, Basseri). Many remain as shepherds all their lives, so that shepherd households are recognized as a distinct social category and the status of shepherd may be permanent, if not ascribed. Shepherds, together with landless labourers, represent the fourth in a fivetiered system of stratification ranging from tribal khans down to servants and craftsmen (Beck 1980, p.330). Moreover they occupy a special position with respect to the tribal organization: Hired shepherds are not only poor in animals . . . they also lack personal and tribal access to pastures and they have few affiliations with tribal leaders. Poor integration in the tribal system is the key element. Most shepherds are members of poorly integrated sub-tribes that
have weak leadership systems . . . The minor subtribes, although structurally equivalent to the major ones, are a source of shepherds to the major
ones . . . Because they usually lack pasture rights and adequate wealth in animals, they cannot easily quit shepherding . . . They tend to have weak agnatic and affinal links that provide them with little economic or political assistance. (Beck 1980, pp.334-5; emphasis added) More is involved here than the logistics of levelling out imbalances between labour and capital at the household level. The system allows wealthy households with a number of flocks to stay within the pastoral sector 'where they create sources of employment for the poor' (Beck 1980, p.335), and permits a greater range of economic diversity. This 'stratified7 response should be marked by a bimodal distribution of livestock with a large bulge of households with few or no animals and a smaller peak of large herd owners. The adjustment of human to animal population is likely to be such that a higher proportion of humans can be maintained as nomads, relative to numbers of livestock, than in more egalitarian forms. The effects of Malthusian controls on livestock numbers will be to set up a flow of personnel into dependent or auxiliary statuses within the nomadic sector rather than into the sedentary sector. In short, the maximum concentration of individuals will be achieved that can be supported by nomadic pastoralism within a given region.
Nomads in archaeology
42
Pastoral 'feudalism': Central Asia and the Kirghiz
This process has been taken to its quintessence in Central Asia through a kind of pastoral 'feudalism', an elaboration of contract herding whereby wealthy households place sections of their flocks with poorer families, known as the saun system (Lattimore 1951, p.96; Shahrani 1979, p.179; Vainshtein 1980, p.104). Like shepherding, this system can serve to balance out discrepancies in labour and capital between equivalent and often related households. But it can also lead to a system of pastoral tenancy in which the huge livestock holdings of tribal elites are farmed out to 'tenant' households who have only temporary usufruct rights over milk and wool on the understanding that any capital increase must be restored to the owner. Under such a system the physical distribution of animals among households will bear little relation to the distribution of ownership. For instance, among the Kirghiz of the Afghan Pamirs, the khan alone owned over 16,000 sheep and goats of which some 14,000 were dispersed among 130 households (Shahrani 1979, pp. 165,182). The annual redistribution of livestock by a politically powerful elite ensured the supply of sufficient animals for the subsistence needs of each household, thereby retaining personnel within the nomadic sector. The ownership or control of pasture land may also be an important factor here. The buying and selling of grazing rights or pasture lots has been reported among the Shahsevan in northwest Iran and the Pashtuns of Afghanistan (Tapper 1979b), and among other nomads the domination of large tracts of land by tribal leaders appears to form the basis for the annual allocation of grazing rights among tribal members (Shahrani 1979; Beck 1980; Oberling 1974, p.15; Fazel 1973). Such forms of pastoral tenancy may be viewed at the sociopolitical level in terms of patron-client relationships. Clients may well form an expendable labour pool which can be dumped in hard times (Peters 1966; Swift 1977; Marx 1977, p.354) but, where supported by kin or tribal ties, the elevation of an elite can also serve as a major resource during difficult times (e.g. Spooner 1964, p.63).
The commercial response: the Yoriiks of Southern Turkey Where nomads are heavily incorporated into a system of capitalist market forces the costs incurred through household consumption and herd maintenance often tend to outrun livestock market prices. Such appears to be the case among the nomadic Yoriiks where such fluctuations condition the flow of personnel into or out of the nomadic sector. Under such conditions the minimum threshold of animals required for household viability will be raised, resulting in fewer nomads with larger flocks though not necessarily greater disposable assets. Since individual flocks are likely to fall within the optimal range (250-400) for efficient herding, cooperative herding units are not formed and shepherding labour tends to be drawn from within the household. Since the household is both the unit of production and the unit of management, any severe imbalances in the matching of labour and capital resources may lead to sedentarization. The pattern of livestock ownership will tend towards a normal distribution with a majority of middle herd owners. In ecological terms, the number of animals pastured within a given region is likely to remain constant, but
Nomad pastoral economy
Fig. 5.5 Distribution of household livestock ownership for Yoriiks and Yomut Turkmen. The Yoriik example (a) conforms closer to a normal distribution than the Turkmen (b). (After Bates 1973,fig.14; and Irons 1975, table 12.)
43
30
No. of t«nts
20
YOROKS
"
M
.I. I1 I Herd size
No. of tents
15-
n...
YOMUT
10-
1 1
10
20
$0
100
ZOO
400
600
700
Herd size
they will be concentrated in fewer hands. Fig. 3.5 shows the distribution of livestock ownership among the Yoriiks (Bates 1973, p. 163, fig. 14) and the Yomut (Irons 1975, table 12). The Yoriiks display a roughly normal distribution, slightly negatively skewed, with a mean of 268.35 and median of 245, compared with the more nearly bimodal distribution of the Yomut.
4 RESIDENCE, DESCENT AND TERRITORY The social significance of nomadic mobility is that it gives the opportunity for continual choice and change in residential association, within a wide but limited and relatively homogeneous social field, an opportunity inherently denied to settled people. (Tapper 1979a, p.46)
Nomad camps have been likened to 'mobile villages7 (Barth 1961, p.25; Stauffer 1965, p.285). The analogy holds up to a point. Camps, like villages, are composed of individual dwellings, each containing a household unit. Tents are generally clustered together for reasons of sociability and security. The residents of both settlement types form a face-to-face community. However as our attention shifts to residential patterns, community organization, property relations and territory, the analogy begins to break down. Failing major catastrophe, there is a certain inertia in the physical layout and membership of a village, which is absent in a nomad camp. Other things being equal, the villager will tend to remain where he is and any change in residence will involve a major decision. For the nomad such decisions are routine and the composition of a nomad camp can persist even over short timespans only through constant reaffirmation (Barth 1961, p.25). The size and composition of nomad residence groups is therefore highly responsive to decisions taken at the household level.
Nomad social organization Central to this residential flexibility is the nomad facility for reinforcing social relationships through physical distance. A village plan may well reflect lineage structures or status distinctions (Kramer 1979; Watson 1979), but any changes in these structures will not necessarily result in people moving house en masse to fit the new situation. By way of contrast, each new nomad encampment is laid out according to a subtle grammar of kinship networks, status and herding logistics (see Chapter 8). A key factor in any current camping arrangement is the state of conflict between households. The principle is beautifully illustrated by Barth (1961) who observed two families, who had clashed over the occupation of a choice 44
Residence, descent and territory
45
tentsite, dragging their goods and chattels off in opposite directions to set up their tents at each extremity of the camp. Many ethnographers particularly those who have worked in both migratory and settled village communities (W. Swidler 1972; Watson 1979) - have remarked on the generally low levels of interpersonal tension, if not overt conflict, among nomads as opposed to villagers. This may be due in part to the ability to defuse conflict situations by reinforcing social distance with physical distance. These residential arrangements are spin-offs rather than causal factors. The constant need for decisions affecting residential location and associations is underlied by the chronic instability and changes in fortune which have emerged as the hallmark of the pastoral mode of subsistence. Nomadic mobility provides the framework through which such decisions may be implemented. Changes in the size of household flocks can upset labour-sharing arrangements within herding units leading to the formation of new cooperatives each season. Where pasture is allocated in fixed lots, annual variation in quality together with fluctuations in flock sizes may result in families being allocated different lots each year. Conflict is a frequently cited reason for the fissioning of camp groups, but such imbalances between labour, capital and pastureland are frequently at the heart of conflicts (see Peters 1960). It is well known that small cooperatives have a low tolerance threshold for conflict which, short of the intervention of some higher authority, will normally lead to the hiving off of individual households or the splitting of camps.
Descent groups and residence groups Given the regular modular layout of most nomad camps it is tempting to superimpose kinship structures on these plans, particularly for archaeologists raised in the archaeology-as-anthropology tradition. Let it be said from the start that a tendency for camps to form on the basis of agnatic descent groups is virtually universal among Near Eastern nomads. But this is a tendency only, and a wide range of options is normally held open: 'Fully effective exploitation of nomadic mobility . . . is possible only if each household and each lineage is free to move over a wide area and, if necessary, to take up permanent residence with people with whom they share no close ties of kinship7 (Irons 1975, p.39). The most intensive studies of camp group formation (Bates 1973; Barth 1961; Tapper 1979a; W. Swidler 1972, 1973; Peters 1960) have invariably found that cognatic and affinal ties are as statistically common as those of patrilineal descent although, given the widespread practice of parallel cousin marriage (Barth 1973b; Murphy and Kasdan 1967; Khuri 1970), these different strands are often difficult to unravel. But more important than empirical arguments based on actual camp group membership are conceptual differences between the processes of lineage segmentation and residence group formation. Lineage segments proliferate through time in a regular manner forming structurally equivalent units in a nested hierarchy. The segmentary lineage model as summarized by Sahlins (1961) envisages all units, however high or low in the hierarchy, as fulfilling much the same function - the
Nomads in archaeology
46
pursuit of the blood feud. The residence hierarchy, by contrast, is one in which each level is formed in accordance with a different set of principles and for a different purpose. While certain levels of agnatic segmentation may sometimes be correlated with a particular level in the residence hierarchy (e.g. a camp group composed predominantly of members from a single maximal lineage) not all levels of segmentation correspond with any level in the residence hierarchy. In short, as pointed out long ago by Evans-Pritchard (1940, pp.203-7), the physical distribution of households and residence units can never correspond exactly to the structure of agnatic descent, although the latter may often serve as a rough characterization of the former. Moreover, among different groups of nomads, different levels of segmentation materialize at the core of functioning communities. Yet in a comparative study of a number of nomad ethnographies from the Near East and North Africa Tapper (1979a) discovered a remarkable degree of consistency in the size of residence groups at each level in the residence hierarchy. Five or six levels of group formation emerge overall, each characterized by a discrete range of lowest to highest numbers. He points out that: 'the communities . . . vary widely in size within particular societies, while average sizes are remarkably uniform7 (ibid., p.61). The results of his comparisons have been graphed in Fig. 4.1. Tapper employs these data to support the hypothesis that these regularities are a reflection of species-specific size thresholds (Forge 1972; Johnson 1978) somehow Fig. 4- i Schematic diagram of levels in a nomad residence hierarchy based on estimates from 14 ethnographically recorded nomadic societies in the Near East and North Africa. The numbers refer to households or 'tents'. The dispersion of average figures for sheep herding groups and mixed herding groups are shown at right. (Compiled from data presented by Tapper 1979a.) 1: Herding unit, a: 'A' type or primary community. b-e: 'B; type or secondary community levels.
Mixed herding (camel, cattle,sheep)
5cde
i.abcde
i.abcde
i.abcde
i.abcde
i.abcde
•
Residence, descent and territory
47
'wired into' all human societies but achieving their full geographical expression only among nomads and other mobile groups: it will be among nomads rather than among settled peoples that we can expect the emergence and evidence of any inherent social dynamic processes generating interactional communities of a certain size and character . . . The same factors may operate in settled societies but it is sooner in mobile and flexible than in stationary and fixed societies that we can expect them to express themselves in social groupings. (Tapper 1979a; p.61) While I am attracted to the kind of theory being advanced here, I must admit to being highly suspicious of the data employed in this instance, and on two grounds. The first involves the source of the data - ethnographers attempting to convey an overall view of a society on the basis of reports from a limited number of informants are prone to cite rounded figures such as 5 to 10, an average of 100, etc., rather than observed ranges of 4 to 13 or averages of 93.5. Second, when dealing with a hierarchy, estimates of higher level units, even less accessible to direct observation, are apt to become multiples of lower order units. Such values are liable to fit a log scale, such as that employed in Fig. 4.1, by way of their formal mathematical properties rather than any actual relation to the phenomenon under study. Such amplified imprecisions may well convey an impression of discrete ranges and hierarchy where none in fact exists. Sheep
herding
20.000 I 1000
M.ON 5OO0
ii I
,
)ioo
ioo<
TOO
200
80 80
30 I
*f
35
J0
Z
4
labcde CO Q
on :*: —>
cr \— X
CD
labcde
QL
Ld CO CO
OQ
i.abcde
i.abcde
i a bede
i.abcde
— •
a
x o 3
CO
X
X
C5
CO
CO
i.abcde"
CO
o
a.
m
'•
Nomads in archaeology
48
More substantial and relevant to our present aims is the identification of a unit at the lower end of the hierarchy dubbed the A Type or primary community composed of somewhere between 20 and 60 households: 'based on a dominant lineage, with a greater or lesser proportion of temporarily resident affines, multilateral kin, unrelated clients or herdsmen7 (Tapper 1979a; p.60). The primary community, whose defining feature is that it permits the direct interaction between all household heads, tends to form around pastoral resources such as tracts of grazing or water sources. They vary in year-to-year stability according to whether grazing rights are individually owned or jointly held (maximal stability) or subject to frequent redistribution, with the strongest tendencies towards the formation of agnatically based residence groups occurring in the former case. Within its structure camp groups of various sizes may be constituted in response to different seasonal demands and direct adjustments between personnel and resources effected. It is commonly a pasture-holding estate controlling - at least for a season - a particular pasture lot allocated on the basis of tribal membership, within which members of the community share equal rights to the available pasture. It may coincide with a lineage segment, but is more commonly composed of an agnatic core together with collateral, affinal and unrelated households. Within the structure of the primary community it is possible to identify camp groups and herding units formed for logistical herd management purposes. Within the primary community no permanent authority structures emerge, though a headman or elder is usually 'elected7 each year. Fission or fusion of these groups occurs in response to imbalances between population and resources. But, according to Tapper (1979a, pp.59-61), ecological factors and the matching of human, capital and pastoral resources cannot account for the regular formation of primary communities of a certain size range cross-culturally, and he points to aggregations of a similar order (regional bands) among hunter-gatherers. The implication is that size-related factors involved in conflict resolution and corporate decision-making define the range within which solutions to ecological and economic imbalances can be arrived at. While the need for concensus on issues such as grazing territories and commitment to joint herding arrangements is important, the ability to arrive at such a concensus - in fact any concensus - is lost once the number of household heads has exceeded a certain threshold, not because of lack of agreement per se but through a breakdown in the processes by which agreement may be reached. This model may be subsumed under a broad body of organizational theory, used to account for the formation of chiefdoms and the origin of the state, which ultimately falls back on social psychological explanations (Lee 1979; Johnson 1978; Renfrew et at. 1982). It tends to stress the role of variables such as scale and the size of information networks as against the constant of human information processing capacity. As such it does not necessarily go against the theory of modes of subsistence outlined above but raises another set of factors which must be built into the model at the relations of production level. Above the level of the primary community the character of the groupings and their reasons for formation undergo important changes. The order of groupings fall within the ranges 80-1,000, 1,000-5,000 and 7,500-25,000 households, and are characterized by Tapper (1979a) as
Residence, descent and territory
49
B Type or secondary communities. These are no longer residence units but levels in a settlement hierarchy above the face-to-face level. Agnatic descent may be emphasized as a primary criterion for membership in pasture-holding territorial groups, and membership in these units is far more stable than that at the primary community level. Outsiders must be incorporated as opposed to being simply associated. At some level above that of the primary community we find a unit - oba, oymak, oulad, etc. associated with a particular territory or migration schedule, though within this group there will be considerable fluidity in the allocation of pasture estates and the membership of camp groups (primary communities) from year to year. Although face-to-face contact within such communities may be rare, they act as important reference groups of a sub-ethnic character within which there is broad agreement on social norms and customs. Furthermore, secondary communities at some level - usually 'tribe' or 'sub-tribe7 - constitute a marriage isolate or endogamous unit, though a statistically high level of endogamy may also be found within the primary community. It should be emphasized that each segment within the nomad residence hierarchy forms not as a result of any inherent segmentary principle, but through the effect of different sets of factors operating at each level. The household is the primary unit of labour supply and ownership. At the level of the herding unit, logistical and herd management considerations are paramount. The primary community forms around the operational units of the pastoral economy - actual physical camps of varying seasonal composition. At a still higher level blocks of territory are associated with stable descent groups which are endogamous and act as cultural reference groups. The dichotomy between primary and secondary communities marks that point below which some degree of flexibility in economic and residential association is essential, and above which the emphasis begins to shift to membership in territorially linked groups defined in terms of either agnatic descent or political allegiance (see Table 4.1). The Basseri provide a good illustration of this kind of structure. Households form into herding units (2-5 tents) on the basis of bilateral or affinal ties, camping together throughout the annual cycle. Camps (10-40 tents) are the closest thing to a primary community, forming 'a very clearly bounded social group, their relations to each other as continuing neighbours are relatively constant while all other contacts are passing, ephemeral and governed by chance7 (Barth 1961, p.25). But camps constitute a single settlement unit only during migration. While an agnatically related core forms the nucleus of most camps, the presence of cognatic kin and 60 per cent intra-camp endogamy make it in essence a 'bilateral, nearly self-perpetuating kin group7 (ibid., p.41), although there is some fluidity in membership. Herding units do not represent segments of the camp since they are differently constituted each year. Nor does the camp represent a lower-order segment of the next level in the hierarchy, the oulad. The oulad, of around 100 tents, is an agnatically defined descent group with a formally recognized leader, the katkhoda, who is directly responsible to the tribal chief. This is the pasture-holding estate within which all oulad members enjoy equal grazing rights. It has its own migration schedule
50
Nomads in archaeology
Table 4.1
Group Alikanli
Bakhtiari
Baluch
Basseri
Jaf
Functional properties of different levels of segmentation for nine Near Eastern tribal groups (from information contained in Tapper 1919a)
Level of segmentation asiret
No. of households
Pasture estate
Migration Contiguous Political territory unit group
? ?
kabile oymak
50-60
zoma
20-40
//
?
tayfeh tireh tash
?
35
oulad
11
camps
?
tribe section subsection khel
Reference group
Marriage isolate
Descent group
X
X
X
X
X
(partial)
X
x X
?
X
X
X
X
X
X
10,000 ?
500 150
camp
20?
herding units
3-4
confederacy tribe (it) tireh oulad
20,000 3,000 200 100
camps herding units
10-40 4-5
asiret
?
tribe
800
X X
(agnatic core)
X
X X
X
X
X X
X
X
(2/3 within camp) ?
[headman raiz)
hoz (max. lineage) khel herding units Pashtun
confederacy tribe
X
X
?
khel
20-50
X
bonkuh beyleh
X
(lineage segment)
X?
tayfeh tireh
X
x (elder)
?
tayfa
//
X
20-30
1,000 ?
herding units Qashqa'i
?
X
x {malik) x{khan)
X
(headman )
?
25,000 3,000 80 30-40 ?
{katkhoda) X
X
(elder)
X
X ?
Residence, descent and territory
Group Shahsevan
Level of segmentation
5,000 500-1,000 200 100
// tribe
tayfa section
Yomut
No. of households
tireh
20-30
/7 tribe
5,000 1,000
oba
25-1,000
Reference group
Pasture estate
51
Migration Contiguous Political territory group unit
Marriage isolate
Descent group
X X
X
(90%)
X
(chief) X
(40%)
X
x x
(agnatic core)
?
( pokesmar
• )
(agnatic core)
Dotted line shows boundary between primary •nary and secondary communities
Key:
a$iret beyleh bonkuh hoz il habile katkhoda khel malik
tribe minimal camp group camp group tribal section tribal confederacy maximal descent group headman tribal section/camp group chief
oba oulad oymak raiz tash tayfa tayfeh tireh zotna
large migratory unit migratory group migratory unit chief descent group tribal section tribe tribal section camp group
and is therefore the key administrative and political division within the tribe. Above the oulad stands the tribal section; or tireh (over 200 tents); a stable agnatic descent group which has no political role or leader and seems not to be important in pasture allocation, though it normally covers a territory of contiguous estates. The tireh seems to function as the primary reference group and marriage isolate. The tribe itself (around 3,000 tents at the time of the study) is a political creation consisting of tribal sections brought together under the domination of a khan, forming a territorial and migratory unit but lacking any notion of common agnatic descent. The Basseri constitute a Persian-speaking ethnic unit within the Khamseh Confederacy (10,000-20,000 tents), a loose association of five tribes - the others being Turkic- or Arabic-speaking - administratively created as a counterweight to the powerful Qashqa'i confederacy to the northwest. The Ilkhan (paramount chief) of the confederacy was drawn from the Qavam, a family of wealthy Shiraz merchants (Barth 1961; Marsden 1978, p.7). One wonders whether the importance attached by Barth to the oulad as an organizational unit may have been partly due to the phase in the annual cycle at which the Basseri were observed. Had they been studied in summer pastures, possibly some other unit may well have emerged in the allocation of pasture. For instance Tapper (1979b, p. 100) suspects that some form of individuated access to pasture lots may well have been present among the Basseri, as among the Qashqa'i (Beck 1980) and the Lurs (Black 1972).
Nomads in archaeology
52
The role of the tribe Nowhere is the 'problem of the tribe7 (Helm (ed.) 1968) more problematic than among Near Eastern pastoral nomads. Traditionally the term has been used loosely to describe politically recognizable units of unspecified size and function. Recently however it has come under close scrutiny and given a range of connotations. While no definitive solution to the 'problem' will be offered here; a discussion of the concept may help to draw together some of the lines of argument followed above. Having described pastoral nomadism in terms of a Marxist-inspired theory of modes of subsistence, it might appear tempting to dismiss the notion of tribe as a piece of superstructural baggage, as indeed some Marxists have tended to do (Fried 1968; Marsden 1978; Bruinessen 1979). However a number of problems have emerged from this analysis, particularly the matter of access to far-flung pastoral resources, which have not so far been adequately accounted for. I believe that an account of tribal structures and processes will serve to clarify these questions. First I will dispose of some recent misuses of the term. Any attempt to fit tribally organized Near Eastern nomads into a scheme of sociopolitical evolution moving from band through tribe to chiefdom and state (Service 1966; Sahlins 1961) is bound to weave a web of contradictions. 'Tribes7 often appear to be organized in a manner more appropriate to that of chiefdoms, or even states, and may be found incorporated into state structures. Such 'tribal chiefdoms7 seem to have existed alongside states for some 5,000 years, in close interaction, without the former having evolved into the latter in any general sense, though of course particular states have been founded or conquered by tribal elites (Rosenfeld 1965; Spooner 1969; Woods 1976). Objections to the evolutionary view by Marxists are founded on a different view of the genesis of tribal structures. Fried (1968, pp.15-16) and Bruinessen (1979, pp. 148-9) see tribes as a secondary phenomenon, an organizational response to the pressures of adjacent states. According to Marsden (1978, p.10) the tribe - specifically the Qashqa7i taifeh emerged as a unit of organization and a means of classifying people only in the twentieth century. Prior to this the operational units tended to be sub-tribal and more numerous, the size of the units varying with the power of the state (Marsden 1976, p.12). By contrast, Bruinessen (1979, p.228) sees a kind of devolution occurring in nineteenth century Kurdistan as more complex organizational forms, Emirates collapsed, allowing 'simpler forms7, tribes of varying degrees of complexity, to emerge. These suggestions emanate from a primarily political conception of the tribe as 'a political means of defining populations7 (Marsden 1976, p.9). This view denies that there are any economic and social processes at work which might be elucidated by such a concept as 'tribe7, and focuses on a particular level of organization - the interface between tribesmen and state administration. Closer to my own view is that expounded by Emmanual Marx (1977) who sees the tribe as an autonomous territorial unit (see also Dole 1968); though rather than a bounded unit I prefer to see it as a territorial system,
Residence, descent and territory
53
an organization for the control of territory. This should not be confused with Sahlin's (1961) model of the segmentary lineage system, useful for envisaging the operation of the blood feud and territorial conflict in some societies. The tribe too is a segmentary system, but its segments have little in common with those of the segmentary lineage. While Fried (1968, p.5) is correct in defining the tribe in terms of a series of aggregates, it should be stressed (as above) that these aggregates are neither structurally nor functionally equivalent but formed on the basis of different criteria, for different purposes and for the solution of different problems at each level. Certain southwest Asian anthropologists (e.g. Irons 1968, 1975; Salzman 1978) have correctly identified many of the features of the segmentary lineage system among nomadic pastoralists, but they are mistaken in using such a model to account for nomadic territorial systems. Whereas within primary communities criteria of common residence and economic cooperation predominate, at the secondary level greater weight is given to membership in lineages which serve as marriage isolates and cultural reference groups. A third element is usually involved in the form of authority structures centring on chiefs which penetrate the secondary level and are capable of holding together large lineage-based units which recognize no common descent. The process by which tribal units crystallize around a central lineage or powerful individual seems to have been historically common. Such units possess a central core organized in lineages, together with a fluctuating element which may shift allegiance to rival leaders (Millingen 1870, p.28; Edmonds 1922). Such accretions could also occur, according to Bruinessen (1979, pp.136-7), not only through recruitment from other tribes, but also by attracting nontribal peasantry. Barth (1961, pp.132-3) recognizes two patterns of tribal evolution: the first by way of population growth, segmentation and incorporation; the second occurring much more rapidly through the aggregation of tribally - and even ethnically - diverse elements around a strong leader. The impression of stable lineage-based structures often conveyed as a result of synchronic modes of ethnographic inquiry may conceal a system in a constant state of flux. This perceived lineage structure is frequently a product of conscious rationalizations by tribal leaders or provincial administrators, an edifice of kinship erected on the basis of a common political or military purpose (see Barth 1961, p.55). This was recognized long ago by ibn Khaldun (1967, p. 100): It is clear that a person of a certain descent may become attached to people of another descent . . . Such a person comes to be known as having the same descent as those to whom he is attached and is counted as one of them . . . Family lines in this manner continually changed from one tribal group to another . . . Such may have been the origins of the Marri tribe, characterized by Pastner (1978, p.262) as originally a 'military and looting organization' only later welded into a structure of lineages. A case of administrative creation is found in the Khamseh Confederacy - of which the Basseri are members - referred to above, founded by wealthy Shiraz merchants under state sponsorship. A closely related phenomenon is the creation of tribes or confederacies
Nomads in archaeology
54
'by decree' around a nucleus of freshly recruited military units. Mercenaries drawn from diverse sources, together with their households and retainers, may be offered 'tribal7 status as a reward for loyal service. The Shahsevan, literally 'those who love the Shah' (Tapper 1979c), and the Alikanli (Be§ikgi 1969) seem to have had such an origin, and historical accounts of warfare from Seljuk to Safavid times employ the convention of equating cavalry strength with household heads of nomadic tribes or 'tents' (Minorsky 1978; Masson-Smith 1978). Certainly a case can be argued for a partial rethinking of 'tribal' processes in terms of the accretion of heterogeneous units in response to largely political pressures. The tendency is greatly facilitated by the fluidity in composition of units at the primary community (camp and migration group) level, and soon rationalized into an overarching structure through the grammar of kinship and descent and the ramifying effects of affinal ties transformed into kinship networks with the passing generations. But the spontaneous generation of tribes and confederacies in this manner is no fleeting phenomenon, for the achievement or conferment of tribal status invariably involves, at the very least, identity with a territory and/or migration cycle. It may well involve substantial control of a territory.
The tribe as a territorial system There are in the Near East many tribal groupings which involve nomadism only marginally, if at all, ranging from sedentary Kurdish mountain villagers (Barth 1953; Bruinessen 1979) to Berber citrus cultivators (Hart 1970) to Marsh Arabs (Thesiger 1964). In all cases, however, the common denominator appears to be a fluid territorial system and intense competition for scarce land or water resources. It is becoming a little clearer now why pastoralism, nomadism and tribalism should form such a coherent package. The inherent instability of a pastoral mode of subsistence, accompanied by constant changes of residence and fluctuations in the size and composition of co-resident groups, both demands and facilitates a territorial system of great complexity and maximumflexibility.Just such a system has been envisaged by Spooner (1971, p.205): 'security of access to a particular set of resources is less important than the maintenance of a network of relationships which would secure options to switch from one set of resources to another as conditions required'. The sociopolitical structure which secures such options is what I understand by the term 'tribe'. Salzman (1978) has suggested that the segmentaiy lineage acts in this way, providing a stable system of organization for those with no stable territorial framework. Granted that the structure of segmented lineages and the nested descent segments involved in the blood feud may be observed among Near Eastern pastoralists (Irons 1975), they serve merely to trace out the lines of current territorial disputes rather than provide the basis for territorial control. This must frequently be sought in structures of authority, whether generated from within the tribal system or externally imposed. The 'tribe' simply refers to a territorial system in which control is not vested in the state apparatus, or at least
Residence, descent and territory
55
where certain areas of control are relinquished to local interests and collectives. This is illustrated by the frequently stated adage that tribesmen do not pay taxes, only dues to tribal notables (Rowton 1973, 1974; Bruinessen 1979). The tribe constitutes the operational unit through which units of population are matched to units of territory and in which hegemony over, or rights within a territory or migration track, are vested (Cribb 1982, 1984c). Among nomads it is the larger divisions at the tribal section level that are most clearly identified with a particular tract of territory.
Tribal allocation of territory The allocation of territory within a tribal framework can occur in various ways. A common pattern observed among the Qashqa'i, Basseri, and Boyr Ahmad in southwest Iran is the seasonal allocation of pastures owned by the khan or held as common territory to sections or estates depending on relative need, livestock numbers and bargaining power within the tribal framework (Oberling 1974, p.15; Barth 1959, p.10, 1961; Fazel 1973). For the most part, however, the khans hold this power in reserve, exercising it only in cases of conflict or during times of major reorganization such as in the turmoil after World War II (Barth 1961, p.76) and as recently as the overthrow of the Shah among the Qashqa'i (Beck 1980, p.349). Something similar was found among the Bakhtiari: The khans had little control over the pastoral base of the Bakhtiyari [sic] which was controlled by family or extended family units . . . Apparently the khans could assign or reassign pasture land as reward or punishment, but seldom did so' (Garthwaite 1978, p. 185). In Turkey, where pastures are usually rented, tribal elders are heavily involved in the formation of rent collectives by means of which household livestock resources are matched to available tracts of grazing (Be§ikgi 1969). The correlation of tribal organization with systems of territorial control is perhaps best illustrated by instances where access to pasture is achieved in other ways. Where this occurs, tribal organization is either weak or non-existent. Anatolian Yoriiks, though fully nomadic, are forced to rent pasture on a piecemeal basis such that there is little consistency in their territorial pattern from one year to the next. Pasture rental is an individual concern, and tribal identity appears correspondingly weak (Bates 1973). The Kurdish pastoral village studied by Watson (1979) made use of pastures owned by the village landlord and lacked any suggestion of tribal organization. Village-based groups in Southern Turkey migrating to summer pastures owned corporately by the village (Tuncdilek 1962, 1963; Johanson 1966) are devoid of tribal organization. It was observed that among Kurdish villagers in northern Iraq and southeastern Turkey tribal affiliation tended to be weak, only being made explicit in the course of feuds or conflicts, while among nomads tribal identities were continually reinforced through migration and the division of pastures (Bruinessen 1979, p.51). Finally, Tapper (1979b, pp. 102,109) appears to regard the loss of pasture rights as coterminous with the loss of Shahsevan tribal membership. Even though pastures are individually owned and alienable, it is highly significant that property transactions are confined to members of the same tribal section (Tapper 1979b, p.111).
Nomads in archaeology
56
The regional variation in modes of controlling and allocating pasture may well be matched by temporal variations. This may be true even in the course of a single seasonal cycle. I have a suspicion that if Barth had stayed with the Basseri through an entire annual cycle, instead of just the spring migration, the egalitarian image that emerged might have been different. Authoritarian elements in pasture allocation are more likely to come into play where large tracts of summer pasture are concerned than for the patches of roadside grazing exploited in the course of migration. I have a further suspicion that an ethnographer returning to the same nomadic group once every decade may well find vastly different modes of pasture allocation in operation each time in the wake of war ; revolution, shifts in market forces and political policies. The effects of the current war in Afghanistan on nomadic lifeways and systems of pastoral tenure would prove very illuminating. A tantalizing glimpse of the rapidity with which such changes can occur is offered by Beck's (1980) observations of the Qashqa'i in post-revolutionary Iran. An increasingly individuated system of pasture acquisition and rental which prevailed during the last decade of the Pahlevis (1970s) was replaced, it seems almost overnight, with one in which tribal systems of collective tenure and allocation by khans were reasserted (Beck 1980, p.332). The system of individuated pasture 'ownership' described by Tapper (1979b, 1979c) among the Shahsevan may well have been a product of temporary stability and state ascendancy, as with current pasture rental systems in Turkey. It remains an open question to what extent the differences between ethnographic accounts of nomads (e.g. Irons 1971; Tapper 1979a) are due to unique and stable features within each system or to the varying political and economic conditions which prevail during different periods in which fieldwork is carried out.
Tribalism and territorial instability The tribal system can only be fully understood when it is set in motion. The inherent tendencies towards the expansion of pastoral herds do not automatically result in pastoral expansion but, given an adequate labour supply and means of acquiring pasture, a well-organized group can embark on a cumulative spiral which will place increasing strain on any static organization of territory. The manner in which this occurs is suggested by the following observations made by Barth in the course of the Basseri migrations: If the Kurdshuli population and flocks grow, sections of that tribe will tend to hold back one or two marches, thereby encroaching on the rights of the Basseri . . . Thus, through competition and continual pressure to utilize alternatives, the pattern of division of pastures by allocation of different il rahs [migratory cycles] is continually kept in rough relation to the needs and powers of different tribes. (Barth 1959-60, p.8) Similar mechanisms of pasture reallocation by khans, speculation in seasonal pasture lots (Tapper 1979a) or the pooling of tribal resources for collective pasture rental (Be§ikgi 1969) exert an inexorable bias towards those households or sections who are both sufficiently rich in livestock to
Residence, descent and territory
57
generate demands on pasture, and strong enough in personnel or influence to back up these demands. Progressive displacements of this kind, often accompanied by territorial conflict, form a dominant theme in tribal history (Hole 1979, p. 157; Baharvand 1975; Sykes 1908; Be§ikgi 1969) and have been vividly described by nineteenth-century observers of Bakhtiari and Kurdish nomads (de Bode 1845, p.87; Maunsell 1901, p.123; Sawyer 1894; Wilson 1926). Such readjustments of personnel and flocks to pastoral resources cannot easily be effected through the kinds of voluntary shifts in camp group membership that occur at the primary community level, but frequently involve direct competition and conflict with large groups of strangers. Moreover territorial extension generally involves the segmentation of territory-holding groups. The evolution of the Baharvand tribe in Central Luristan during the early twentieth century occurred through a series of episodes of growth and fission (Baharvand 1975, pp.47-9). Just as the promotion of ascendent groups on the hierarchy of named segments proceeds with their territorial expansion, so territorially weak groups merge and decline. In this way the formal structure of tribal organization is preserved against a backdrop of the continual reshuffling of segments. Where authoritarian leadership is an important component in tribal organization, as in the Zagros region, the security and prosperity offered by a strong leader can exert a kind of gravitational pull on the peripheral elements of neighbouring tribes, causing them to shift their allegiance (Barth 1961, pp.85-6). This process, aptly described by Edmonds (1922, p.348) as a 'perpetual intrigue to undermine the position of a rival group by detaching its doubtful adherents', is part of the pattern of tribal accretion described above and can occur on a large scale. It is recorded that during a severe famine in 1871 no fewer than 4,000 Qashqa'i families shifted their allegiance to the Bakhtiari (Oberling 1974, p.218). It is not difficult to envisage the initial build-up of large tribal empires such as the Seljuk (Cahen 1968; Sumer 1980) and the Akkoyunlu (Woods 1976) through precisely these mechanisms (see Krader 1979), although it should be noted that the growth of tribal entities in this way need not necessarily entail any territorial displacement of constituent units. It is possible that the relatively stable nomadic territorial systems observed by ethnographers in the course of the 1960s and 70s are largely the result of increasingly effective control by modern nation states at that time (see Stauffer 1965; Fazel 1973). While invariably desired by earlier rulers (Bates 1971), such control was not always possible: The Persian government is ever jealous of the migrations of these tribes, and they cannot remove from one province to another without first having obtained the Shah's permission. In times of trouble, such as the death of the king, frequently, if they be strong enough to encounter opposition, they pass from their old haunts to better places. (Morier 1837, pp.237-8) ; being so powerful, the Kashka i [sic] are able to select their own pasture ground. The Kashka;i thus range over a great extent of country, doing great injury in their movements. (Sheil 1856, pp.398-9)
Nomads in archaeology
58
Observations in Luristan early this century reveal a confused patchwork of constantly shifting allegiances, the rise and decline of tribal segments of various orders and the continual displacement of some groups by others, among which a number of the processes discussed above may be discerned: The Yusufvand are now in the process of absorption by the Hasanvand. . . . According to the Lurs; the Qalarvand and Baharvand [see Baharvand 1975] were not among the nine original tribes of the Dirakvand but came in comparatively recently and absorbed most of them . . . The former summer quarters of the Sagvand in Hurud are now occupied by the Bairanvana. Owing to the bitter blood feud the Sagvand Ali Khani no longer go south while the Rahim Khani stay summer and winter in Araoistan. (Edmonds 1922, pp.345-7)
Nomadism as a cyclical phenomenon The great thirteenth-century Arab thinker ibn Khaldun (1967) saw nomadism not only as a cyclical phenomenon but also as one closely interwoven with the dynastic cycle of civilizations with which nomads interacted. Is there any justification in this view? Having sketched the main economic and sociopolitical variables involved in the phenomenon of nomadism, it remains to put the system in motion. What factors or conditions tend to be associated with the expansion or retraction of nomadic activity?
Phases of territorial stability and instability
Within any given matrix of pastures and migration tracks patterns of mobility may remain relatively stable for a time. However changes in climate or environmental stress, as well as general changes in economic and political conditions, can lead to radical shifts in migration tracks (Cribb 1984c). Such 'migratory drift7 (Stenning 19577 pp.58-9) appears to be most frequent under Near Eastern conditions when a breakdown of central authority allows full rein to the territorial aggrandizement and opportunism inherent in pastoralism. Such displacements, prompted by an intensification of the kinds of territorial mechanisms described above, will usually occur at right-angles to the direction of seasonal migration and along the major ecological divide, whether mountain-lowland or steppe-arable. The net effect of such displacements, by which tribal units are conveyed rapidly from one region to another, may well emerge historically as a major migratory episode or 'invasion7 such as the Turkic incursions of the eleventh and twelfth centuries (Bosworth 1973; Lambton 1973; Cahen 1968; Minorsky 1978). It must be said here that, while I fully endorse recent criticisms of 'migrationist7 themes in archaeology (see Adams et al. 1978), particularly where nomad invasions are held responsible for cultural lacunae or change, migratory episodes such as those described above are not illusions but real events which are likely to have archaeological correlates -
Residence, descent and territory
59
though perhaps not of the kind that have in the past sprung all too readily to hand. In terms of our model, such an episode represents 'untied7 nomadism. Pure untied nomadism involves continual migration into new territories. According to Vainshtein (1980, p.97) something like this occurred on the Central Asian steppe only briefly during the twelfth and thirteenth centuries when conditions of chronic warfare caused the formation of large, compact migratory groups consisting of hundreds or even thousands of tents. The huge concentrations of livestock that resulted led to the rapid exhaustion of pastures, hence the need for constant migration into new areas. Historical sources record the presence of large camps of kibitkas, a kind of tent on wheels or covered wagon, during these periods (see Vainshtein 1980, pp.99-100) and passages from Herodotus (IV, p.119) describing 'waggons which served as houses for the women and children7 indicate that something similar may have been present on the Eurasian steppes two millennia earlier. Constant incursions by militarily strong groups seem to have had a chain reaction effect on other pastoralists, forcing them to adopt similar strategies of rapid deployment and concentration. It was the outer ripples of this cumulative process that impacted on the periphery of the Near East during the latter part of the eleventh century and throughout the twelfth. Such an intrusive episode is likely to be followed by one in which the nomadic groups gradually settle into relatively fixed migration tracks, exhibiting a 'tied7 rather than an 'untied7 pattern. Nomadism then operates in what de Planhol (1979) has called a 'saturated7 milieu in which it must come to terms with competing interests from a wider sociopolitical spectrum. Following the Turkic, Mongol and Timurid episodes, nomadic groups in southwestern Asia seem to have become progressively tied to stable territorial systems (Reid 1978, pp. 120-1; Wood 1976) with periodic recurrences of instability such as that during the nineteenth century (Morier 1837; de Bode 1845; Sheil 1856; Millingen 1870).
The nomadization-sedentarization cycle
The continuum between nomadism and sedentism may be viewed synchronically in terms of a range of productive strategies from settled agriculture, through various forms of mixed farming and forms of 'seminomadism7, to tied and, ultimately, untied nomadism. But the nature of this continuum cannot be fully appreciated unless it is viewed diachronically and unless explanations are offered for the movement of individuals or groups from one polar extreme to the other. Ultimately it is necessary to consider nomadization and sedentarization as systemic processes connected with changes in land-use patterns and sociopolitical trends over broad regions (Barth 1973a) or, in the words of Khazanov (1984, p. 12) 'to return to diachrony its proper place in the study of nomadism7. Alternating phases of nomadization and sedentarization observed in Iran in the present century appear to have been closely linked with the evolution of the modern state, changes in government policy and variations in the effectiveness of central and local administration. While there are many unprecedented factors guiding official action in this case, the
Nomads in archaeology
60
responses of nomadic groups to state pressure seem to exemplify certain features of nomadic, and tribal, organization. The early decades of the twentieth century seem to have been a period of nomadization. It was at this time that the Baharvand reached their apogee of power and established a wide-ranging migratory system over much of southern Luristan (Baharvand 1975, pp.48-9). The establishment of the Pahlavi Dynasty under Reza Shah and the implementation of modernization policies during the 1920s and 30s (Pahlavi 1961, p.220) led to a programme of enforced sedentarization whose motives were primarily political - to break the basis of tribal power and remove what was considered to be an anachronism incompatible with the development of a modern state (Marsden 1978; Stark 1933; Baharvand 1975, pp.51-2; Oberling 1974, p. 153). These were difficult times for nomads. The enforced curtailment of pastoral migration led to massive losses of livestock which in turn affected supplies of pastoral products to regional markets (Stauffer 1965). But Reza Shah was mistaken in the belief that tribalism could be obliterated by bringing an end to nomadism while an essentially pastoral orientation and strong pressures towards transhumant movements remained. As it happened, the removal of coercive pressures led to the resumption of nomadism as before. With the collapse of Reza Shah's government during World War II and the imposition of a joint Anglo-Soviet military administration, there was a massive resurgence of nomadism over much of Iran. The Yomut Turkmen, who had been settled a generation earlier, 'destroyed the permanent dwellings they had been forced to build and reverted to year-round residence in tents7 (Irons 1971, p.147; 1975, p.12). A similar pattern of renomadization occurred among the Basseri (Barth 1961, p.3), the Boyr Ahmad (Fazel 1973, p. 135), the Shahsevan (Tapper 1979a and c) and the Qashqa'i (Garrod 1946a, p.35). With the re-establishment of the Pahlavis in the post-war period, the trend was again toward sedentarization, this time accompanied by major economic changes favouring agricultural intensification and a programme of land reform - the Shah's White Revolution (Pahlavi 1961). By the 1970s nomadism seemed little more than an embarrassing echo from the medieval past, and it must be said that most academic writers of this period tend to convey this impression (Oberling 1974, ch.XVII; Tapper 1979b; Beck 1980; Stauffer 1965). The history of the Iranian Revolution of 1979 has yet to be written. But on current indications one of the side-effects of this event was yet another resurgence of nomadism in those areas which began to slip from the direct control of Teheran. Full details have not yet emerged (Tapper pers. comm.; Beck 1980, pp.349-50). The Bakhtiaris, staunch supporters of the Shah's regime (the empress Soraya, the former head of the secret police (SAVAK), Timur Bakhtiar, and the Shah's short-lived successor, Shahpur Bakhtiar, were all Bakhtiaris), presumably have no love for the present regime. The Kurds and the Brahui made a strong bid for independence. The situation of the Shahsevan and Yomut is unknown at present. The most interesting developments occurred among those masters of international intrigue, the Qashqa'i, located not far from the Iran-Iraq war zone. Shortly after the abdication of the Shah, Nasr Khan, the former head of the Qashqa'i Dynasty who had been exiled in California,
Residence, descent and territory
61
returned to the Qashqa;i country and it was not long before migrations resumed in earnest. The Qashqa'i were back in business. This will be a fascinating area for future research, although by the time political conditions are sufficiently stable to permit ethnographic research, we will, in all probability, be well into another phase of sedentarization. The lesson for students of the history of nomadism is that policies of enforced sedentarization are likely to be ineffective while the underlying factors giving rise to nomadism remain unaltered. Even radical restructuring of systems of land tenure and agricultural policy are likely to be effective as a means of sedentarization only so long as such systems remain in operation. In any given nomad population - particularly when tied to stable migration routes - both nomadizing and sedentarizing tendencies will be present. The degree of nomadism practiced can be seen as a balance struck between these opposing tendencies (Barth 1961, p.118). Short-term cycles or recurrent nomadism (Monteil 1959, p.581) are deducible from the model of pastoral and agricultural modes of subsistence outlined above. Depending on the current state of livestock v. agricultural markets or the level of security, households may disengage temporarily from pastoral production only to return at a later date. For example impoverished Yoriik households may cease migration during difficult times, only to return when sufficient livestock have been amassed or prices for pastoral products have improved (Bates 1973, p.219). Indeed some households have been through this process more than once. Edmonds (1957, p. 146) stresses that there is 'nothing irrevocable7 about settlement among Jaf nomads in Kurdistan and that nomadism may be resumed once a household's viability has been restored. Cumulative tendencies towards either nomadization or sedentarization on a regional basis will involve a complete restructuring of the system of rewards and constraints which govern the interplay of the agricultural and pastoral modes of subsistence. Considerably more is at stake here than the economic profitability of one or the other subsistence strategy. Ultimately the choice of strategy is linked to broad patterns of land use, demographic factors and political stability (Bates 1971).
The process of sedentarization It has been noted (Hole 1978, p. 157) that concurrent tendencies towards high population growth among nomads as well as settled groups and the expansion of agricultural production seem to be inimical to the continuation of nomadic forms of pastoralism. The persistence of such trends, usually under conditions of political stability and security, can tip the balance against an unstable pastoral adaptation. The way in which this can occur - often against the wishes of the producers themselves - is illustrated by the case of a predominantly pastoral Kurdish community in western Iran (Watson 1979, ch.8) whose absentee landlords were forcing them into cultivation because of the lure of increased profits. Even where pastoralism remains dominant, changing systems of rewards can lead to a situation such as that described by Black in Luristan during the height of the late Shah's agricultural reforms:
Nomads in archaeology
62
For the last ten years it has been becoming 'old fashioned' to be a nomad, and only the lure of unlimited speculation in sheep leading to the potential acquisition of great wealth has prevented tne development of an overtly anti-pastoral ideology . . . [The] aim [is] to be rich in animals without pursuing nomadic skills. (Black 1972, p.620) Imbalances between human and livestock populations may also be associated with trends towards sedentarization. Hole (1978, p.148) noted that a rise in the nomad population, together with a decline in livestock holdings per household - too many people, too few animals - was making the nomadic option less viable. Conversely, among Yoriiks in Turkey, a rise in minimal flock numbers and the wasting away of nomad populations as an increasing number of households become unviable, seems to be part of a cumulative trend towards sedentarization (Bates 1973). Though acting in opposite directions in each case, a state of imbalance is common to both. According to our model, phases of sedentarization should be accompanied by a decline in specialization between pastoralism and cultivation, the merging of nomadic and sedentary sectors within tribal units, the weakening of tribal organization - though not necessarily its disappearance - and the pursuit of pastoralism largely within the framework of a mixed agro-pastoral system. Within a regional context we would expect to find a rise in the ratio of total human population to total animal population, though overall increases in human population may leave the absolute numbers of livestock still at a high level. In human terms, the carrying capacity of a given region will be enhanced. Once sedentarization is under way it can produce major changes in patterns of village and even urban settlement as well as shifts in the ethnic composition of settled populations (Coon 1965; Eberhard 1953a and b). Spooner (1969, pp.147-9) has outlined a schedule of stages in the sedentarization of Baluchi nomad tribes. The mobility inherent in nomadism means that any particular group will rarely undergo the full cycle of nomadization to sedentarization within the same locality. The wholesale voluntary settlement of one large Kurdish tribe in northern Syria is said to have resulted, within a matter of a few years, in the establishment of 29 farms, 48 hamlets, 28 villages and one town (Bruinessen 1979, p.103). The socioeconomic changes accompanying sedentarization include an ossification of the current state of inequality to produce a more class-like form of economic stratification, stabilization of the territorial system and changes in the function of lineage and marriage systems (Bates 1973, ch.VII; Baharvand 1975, ch.9; Swidler 1973; CM. Pastner 1978; Salzman 1980).
The process of nomadization Due to the conditions of state ascendency and modernization under which most ethnographic work on nomadism has been carried out, the process of nomadization is poorly understood (Glatzer 1982, p.61). It is nevertheless possible to make a number of deductions on the basis of our mode of subsistence model, supported by some ethnographic and ethnohistorical observations.
Residence, descent and territory
63
Nomadization occurs in response to conditions of environmental and political insecurity (the two are frequently related) through the exercise of options inherent in an unstable subsistence strategy based on pastoral accumulation. It proceeds not simply by an all-round increase in emphasis on pastoral capital, but also through the concentration of livestock in the hands of certain households and communities. The process is considerably more complex than the scenario outlined by Adams (1974) whereby failed cultivators move off with their livestock as a new mobile resource base. Small-scale mixed farming may continue, but increasingly large numbers of animals will be tied up in the hands of specialized pastoralists. The disparity in animal numbers per household between village and nomadic pastoralists referred to earlier (see Table 4.1) indicates the presence of certain thresholds beyond which nomadic options, involving the familiar systems of flexible access to pasture regulated through tribal institutions, become increasingly advantageous. Agricultural and pastoral production will become specialized above the community level as nomadic and sedentary sections crystallize within tribal units and whole tribes or groups of tribes become predominantly identified with pastoralism and nomadism. Isolated cases of nomadization have been observed in the Near East, though none is as well documented as that observed in the Sudan by Haaland (1969). Coon (1965, p.215) mentions a gradual transition from village agriculture to sheep nomadism in parts of the Near East, but gives no specific examples. In Iraqi Kurdistan during the 1920s Edmonds (1957) observed the replacement of a nomadic group by a non-nomadic section of the Jaf tribe and noted that: 'As the size of their flocks increased the newcomers began to follow their predecessors northwards on the summer migration7 (ibid., p. 143). An interesting case has been reported from the Wakhan Corridor in the northeastern corner of Afghanistan shared by Kirghiz nomads and Wakhi agro-pastoralists (Shahrani 1979). Among the latter, two pastoral strategies were observed. On the one hand the poorer villagers 'keep a few cattle and sheep and goats for transport, milk, fleece, manure . . . or cash7, while the other form is confined to 'a small number of land-rich Wakhi households . . . that own large herds of sheep, goats [at least 100 head], cattle or yak7 (Shahrani 1979, p.63). The latter appear to be undergoing the initial stages of nomadization in emulation of, and in competition with, their Kirghiz neighbours (Shahrani 1979, p.191). The best-documented case of nomadization is that of the Pashtuns in western Afghanistan among whom it arises as a solution to tensions between the agricultural and pastoral economies attendant on growth in the numbers of livestock, which is itself a response to fluctuating ecological conditions and economic demand (Glatzer 1982). While nomadization and renomadization are relatively easy within the Pashtun ethnic group, movement in both directions between nomad and sedentary sectors becomes more difficult where major ethnic differences occur, as in the Jawand region of Afghanistan where the nomads are recent Pashtun immigrants alongside the Persian-speaking Firuzkuhi sedentary majority (ibid., p.76). The reasons for an increased emphasis on livestock as a basis for subsistence are to be found, not in dietary requirements nor market exchange, but in positive selection for an unstable productive system which
Nomads in archaeology
64
is able to accomodate rapid fluctuations of fortune and is responsive to any opportunities afforded by unsettled conditions. Associated with a trend towards nomadization will be an overall decline in the ratio of human population to that of livestock within a given region. The reversion of marginal cultivated land to pasture would effectively lower the carrying capacity of a given region.
5 NOMADS - THE INVISIBLE CULTURE? Pastoralists are not likely to leave many vestiges by which the archaeologist could recognize their presence. They tend to use vessels of leather and basketry instead of pots, to live in tents instead of excavated shelters or huts supported by stout timber posts or walls of stone or brick. Leather vessels and baskets have as a rule no chance of surviving. Tents need not even leave deep postholes to mark where they once stood. (Childe 1936, p.81)
As the title suggests, it is the purpose of this chapter to question the assumption that nomads and their material remains are inaccessible to archaeologists. Nomads need not be archaeologically invisible. At the same time they do not constitute an archaeological 'culture7 in the Childean sense. Paradoxically therefore, although the artifacts and campsites used by ancient nomads need not lie beyond the capacity of modern archaeology, there may be no simple means by which these are distinguishable from the productions of more settled communities. The traveller in the more isolated parts of the Near East should not be surprised to see groups of migrating nomads passing through villages whose inhabitants dress in the same manner as the nomads, speak the same dialect, employ the same range of household utensils, possess the same species of domestic animals and, in some cases, claim the same tribal affiliation. Differences there certainly are, but these are often ideological, organizational and economic rather than 'cultural'.
Nomadism as an archaeological problem It has frequently been asserted that nomads are light travellers and therefore leave little archaeological imprint in the way of recoverable materials or modifications to their environment. Ever since Childe (1936) pointed out the futility of searching for archaeological traces of nomads this proposition seems to have been accepted by generations of Near Eastern archaeologists (but see Hole 1974, 1978) and is still reiterated by some (Gilbert 1983, pp.107-8; Evans 1983, p.77). Such arguments, based on ethnographic observations, have been applied to other kinds of material culture assemblages as well (Heider 1967; Cranstone 1971; Ascher 1968; Deetz 1968). Ethnoarchaeological investigations of pastoral sites have also 65
Nomads in archaeology
66
tended to stress the fragility of nomad material culture (Gifford 1977, 1978; Robbins 1973; Murray and Chang 1980). Yet generations of prehistorians, who face recovery problems of similar magnitude, have not been prevented from building up an impressive profile of hunter-gatherer material culture, subsistence and organization. In comparison with hunter-gatherers, the prospects for nomad archaeology appear quite favourable. Many contemporary pastoralists use ceramic pots and storage jars and it is highly likely that in the past, before the widespread availability of metal, glass and plastic, this was even more common (see Hole 1978, p.150; de Schauensee 1979, p.37; Solecki 1979, p.323). Tentcloth may be perishable, and tentpoles leave only shallow depressions, but these portable structures are often accompanied by more substantial features or fixtures. These include raised or levelled floors, mud or stone-lined hearths, stone storage platforms, alcoves, tent foundations and sometimes substantial walls. Solid stone-built corrals are also common features in nomad camps. Nor does Childe mention the presence of animal droppings which can, under some circumstances, form quite thick deposits. If we accept the assumption of nomad invisibility then we might expect archaeologists to consider the matter closed or to suspend speculation on matters which can neither be proved nor supported by archaeological evidence. The fact that they have not done so underlines the paradox that if a proposition cannot be proved, neither can it be disproved. Childe himself was aware of this: The failure to recognize prehistoric settlement sites as belonging to pure pastoralists is not any proof that such did not exist. (Childe 1936, p.81) The presumed absence of evidence for ancient nomadic pastoralists has effectively removed the burden of proof and left a vacuum into which has rushed a great deal of hot air. The topic abounds in ill-founded speculations, untestable theories and magnificent flights of fancy. The sudden appearance of a new culture in an area may be attributed to the arrival of a new wave of nomads. Anomalies in settlement evidence are all too easily attributable to the presence of nomads who, of course, left no record. In this way the presence of elaborate public architecture unaccompanied by settlement evidence is easily explained: We must imagine the plain of Pasargadae full of tents, under which passed a good deal of the daily life. . . . such a place cannot be called exactly a town. It looks more like the first settlement of nomads. (Herzfeld 1935, p.28) This suggestion was later rejected by Stronach (1963, p.19) on the equally tenuous grounds that the presence of pottery around the site proves that the inhabitants could not have been nomads. Thus much of the evidence for ancient nomads is negative - the disappearance of populations or cultures, the absence of summer resources, the lack of settlement evidence. Concrete evidence for nomad occupation sites is meagre indeed and has led to suggestions that evidence for ancient nomads must be sought elsewhere:
Nomads - the invisible culture?
67
We must look at other means of testing the theory [of the presence of ancient nomads] other than the identification of nomad sites in conjunction with settled sites. (Crawford 1978, p.130) While some attempts have been made to infer seasonal occupation of sites on the basis of the age-sex structure of faunal remains (Bokonyi 1972; Wheeler Pires-Ferriera 1977), these face problems of partial or biased sampling and can serve as guidelines only (see Cribb 1984b). The task of finding and interpreting settlement evidence of early nomad occupations is one that cannot be evaded. Others have implicitly acknowledged the need for settlement evidence of this kind and see the absence of such data as inhibiting the progress of research: Archaeologically there is virtually no data with which to test such a hypothesis [about nomads and decision hierarchies] . . . Thus far no sites which can be identified as herding camps have been discovered on the Susiana Plain itself. (Johnson 1973, p.159) While no one has completely overcome the formidable problems of locating nomad sites and assessing their relative density . . . no researcher has claimed increased evidence of pastoralists early in the fourth millennium. (Wright 1977, p.348) After proposing the nomadization of the entire Late Bronze Age population of Mesopotamia, Adams admits: But like virtually everyone else, I havn't [sic] yet really faced the underlying problem or systematically sought for data that would clarify it. (1974, p.9) Not only have nomad sites failed to materialize, but there appears to be no firm indication of what such a site may look like or how it might differ from village or hunter-gatherer sites (see Smith 1983, p.85). As Crawford (1978, p. 131) admits, 'no criteria exist at present for identifying the remains of nomadic settlements'. This in turn raises the possibility that nomad sites have been discovered but not recognized as such: If a nomadic site is located it may well be misinterpreted and attributed to an earlier period because of the paucity and simplicity of the material remains. (Crawford 1978, p. 130) This observation is important in view of the presence of a small body of settlement evidence which may have a bearing on the issue of early nomadism - if only the means of evaluating it were at hand. A number of caves and rockshelters in mountainous areas of the Near East show signs of having served as makeshift corrals or temporary campsites (Hole and Flannery 1962; Solecki 1964, 1979; Dupree 1963b). Flat settlements such as Zawi Chemi Shanidar in northern Iraq (Solecki 1981) and the basal levels of mounds in the Zagros Mountains, such as Tepe Sarab (Braidwood 1961) and Tepe Guran (Melgaard etal. 1963; Mortensen 1964), have
Nomads in archaeology
68
evidence of pastoralism and temporary, seasonal settlement but they occur very early in the development of food production, well before the period suggested by most theories of nomad origins. Surveys conducted in areas of southwest Iran where nomadism is conspicuous today have generally failed to detect sites that can be identified unambiguously as nomadic (Hole 1962, 1968; Thrane 1964; Mortensen 1974). Even where such sites have been explicitly sought the result has usually been the discovery of still more village mounds (Hole 1979; Zagerell 1975). To date there is only one site, Tepe Tula;i in Khuzistan, that could be identified by the excavator (Hole 1974) as a seasonal pastoral campsite containing what appear to be tentsites. Significantly the site was discovered quite by chance in the course of a salvage excavation of a Sassanian complex. The 8,000-year antiquity claimed for Tula'i (Hole 1974, p.237) is still in dispute as the stratigraphy and ceramic associations are far from clear (see Chapter 11). The dimensions of our problem are now becoming clearer. It is not a simple matter of nomad sites being present or absent. On a practical level there is a need to develop appropriate techniques for the location of such sites. This in turn presupposes the existence of clear criteria for the identification of a site as 'nomadic' or 'pastoral'. Finally, such criteria need to be grounded in explicit definitions of concepts such as 'nomadism' and 'pastoralism', such as those discussed in the previous two chapters, together with operational definitions of 'nomadic campsite7, 'pastoral camp;.
The organization of nomad material culture Nomad material culture is best conceived in terms of three key dimensions. The first of these concerns the degree to which items of material culture exist permanently on a site - whether occupied or unoccupied - or are circulated from one site to the next. This may be dichotomized into fixtures on the one hand and portables on the other, the main consideration being the function, size and weight of the item. The concept of fixtures used here corresponds closely with Binford's (1978a) understanding of 'site furniture'. The second dimension of material culture ranges from durables on the one hand to perishables on the other, according to the likelihood that items will deteriorate through time. This affects the archaeologist's perception of material culture in that highly perishable items are likely to be continually renewed rather than circulated, but are less likely to materialize in the archaeological record. A third dimension concerns the value of items, measured in terms of the difficulty or cost of acquiring or replacing them (i.e. valuables as opposed to expendables). The former are more likely to be curated or preserved in systemic context and transported from site to site. The latter are more likely to be discarded on the site and, if durable, preserved in archaeological context. The position of any item with respect to the three axes will reflect both its role in the organization of material culture and its chance of surviving
Nomads - the invisible culture?
69
into archaeological context. For instance items which are portable, perishable and valuable (like carpets) will have virtually zero archaeological visibility. Fixtures which are both durable and expendable and of low intrinsic value (such as stone hearths) will be site-specific and have a much higher visibility quotient. Intuitively one would expect nomad sites to contain a very high proportion of fixtures as opposed to portables and expendables as opposed to valuables. In what follows an attempt will be made, using these dimensions as guidelines, to build up a profile of nomad material culture with a view to assessing its visibility and identifying those variables most pertinent to archaeological investigation.
Nomad and village material culture compared While there may be a bias towards lighter, more portable materials, the range of objects that may be observed in a nomad camp is comparable to that observable in Near Eastern villages. A comparison of material inventories from five different communities - both nomadic and sedentary (Table 5.1) - bears this out. Hasanabad and Shirdasht (Watson 1966, 1979) are both in the Kermanshah Valley in western Iran, within a few kilometres of each other though separated on economic and ethnic criteria. The Hasanabadis are Lakispeakers (a dialect somewhere between Farsi and Kurdish), while the Shirdashtis are Kurds, the former share-cropping cultivators with a few animals, the latter predominantly pastoral. According to Watson (1966, p. 19) 'Most of the household equipment used in the Kurdish village is very similar to that at Hasanabad7. This is in spite of the fact that the inventory shown in Table 5.1 was recorded at the Shirdashtis7 spring tent camp. Apart from the use of modern plastic items, the inventory for Sanaydin Yayla in southern Turkey does not differ in its essentials from these Iranian examples in spite of the great geographical distance and ethnic differences. Also notable is the relative absence from the inventories of items specifically connected with herding. We are reminded that herding as an activity requires virtually no tools (Hole 1978, p. 166). The physical nomad household, like its village counterpart, is primarily concerned with reproduction and consumption rather than production, though animal products are sometimes processed near tents. The most obvious material evidence of pastoral activity on most nomad camps I visited - apart from stone corrals, bones and deposits of animal droppings - was perishable wool clippings, bits of cloth (for tying the legs during shearing) and small bottles which had contained veterinary preparations (cf. Murray and Chang 1980). Self-sufficiency in material paraphernalia appears to be the rule, with minimal lending and borrowing of common household items even among close relatives (Barth 1961, p.3), the only exception being for newly established households (Be§ikqi 1969). In contrast to one of the earlier myths about pastoral society, property and domestic goods are individually owned by each household and not communally.
Ethnohistorical accounts of nomad material culture
Although often fragmentary and impressionistic, ethnohistorical descriptions of nomad material culture suggest the same basic inventory. Most of
70
Nomads in archaeology
Table 5.1 Inventories of items of material culture, Near Eastern nomadic and village communities Sanaydm Yayla Summer Camp, South Central Anatolia. (Fieldwork, September and October 1981.) Metal: Shield for baking bread (sag) Cooking tripod Poker Aluminium pots & bowls Heavy, engraved brass serving tray Teapot Cutlery, knives, teaspoons Shovel Sheep shears Shotgun Axe Kerosene lamp Pocket knives - wood & metal Gas cooker & cannister Wood: Large tripod for churning cheese & yoghurt Loom - 3 heavy beams Woven bread basket Bread board & rolling pin Wooden spoons Mortar & pestle Sieve for flour - together with wire netting Large wooden chest containing clothing Board supporting chest
Ceramics: 2 or 3 small pots teapot & tea saucers Glass: Tea glasses Medicine bottles Plastic: Bowls Large basins, buckets Water cannisters Clear polythene plastic - multiple uses Textile: Mats Kilims or rugs Storage sacks Donkey bags Bedding Leather: Skin container for churning Shoes
Alikanh Nomads, Eastern Anatolia (Besjkcj 1969:112, table 27). (Observed in summer camp in Nemrut Dag.) Metal: Iron plate or convex shield - baking bread (sac) Iron tripod to support plate (sag ayak) Large dish - mixing dough Large cauldron - iron Small cauldron - iron Crockery and cutlery - copper Tray for serving meals (sini) Tin container for water Ewer- vessel with handle & spout Copper bowl (tas) Tea pot Tray for serving tea Lamps and lanterns Shovel Sheep shears
Ceramics: Tea pot
Wood: Large boxes - storing belongings Comb for wool
Other: Broom
Loom (spi) Spindle (tefi) Mallet - for driving in tent pegs
Glass: Tea glasses Bottle for keeping paraffin Textile: Mattress Quilt Cushions Felt mats (kilim) Goat-hair bags Leather: Skin bags - churning Skin bags - all-purpose containers Suitcases
Comments: Besjkcj does not mention the use of large ceramic cooking pots, such as those sunk into pits to bake bread observed at ND-1. However there is no doubt that these are in regular use.
Nomads - the invisible culture?
71
Qashqd'i Nomads, South West Iran (after Marsden, 1976:33,55). (Note: not confined to household items; excludes jewellery, horse trappings, etc.) Metal: Cooking tripod Kettle & tripod - copper, iron Tongs - iron Exterior of open box Cooking pots (4) & lids (2) - copper Mallet for pegs - wood & metal Trays - tin & brass Cooking pot & lid - copper Kitchen knife Ladle - aluminium Bread turner or shield - iron Cooking tray-iron Plates (2) - aluminium Water jug - brass Teaspoons Sugar hatchet - steel Club-headed stick Pipe - metal & wood
Wood: Tent lattice - willow cane Tent pins Tent buckles Tent poles Tripod - for churning yoghurt Looms Pestle & mortar Bread board & rolling pin Samovar box Baby shade, frames Flutes Ceramic:
Drum-aluminium Branding iron Sheep shears Sheep/goat bells
Water u ) § Textile: Goat-hair cloth Tent decoration Storage bags (2) Bed covers (3) Cushions (2) Prayer mat
Tongs-brass
Inside of o p e n b o x of fired clay
Porcelain teapot (repaired with wire) Porcelain saucers (made in Japan) P i p e ( p j u s t i n & p l a s t i c components) Opium pipe . Dl
fastlc:.
Shirdasht Spring camp of Shirdasht village at Duzaray, W. Iran (after Watson 1979: 263). Metal: Iron plate for cooking bread (saj) Fire tongs Iron pincers - to support vessel over fire Iranian pots (4) - tinned copper Long-handled frying pan - tinned copper Small brass tray Small, heavy metal bowls (3) Light aluminium bowls Large shallow pan - for mixing dough Tin sieve - flour Tin tea kettle Tin samovar Tin can-tea container Brass bowl - to hold tea glasses & pot Short-handled adze (sugar hatchet) Kerosene lantern WoodLarge wooden spoons (2) Wooden sugar bowl Twig broom
Ceramic: Cloth-covered pot lids (2) Small ceramic stew pot with perforated loop handle Saucers for tea cups (3) TextileCarpet C A s or rugs (2) Quilts (2) B o l s t e r s (2) W o o j b l a n k e t s (2) B u n d l e of e x t r a b e d d i n g L
?ther: Basketry trays (3) Comments: The interior of the tent is kept very clean with all bedding put away during the daytime and all unused utensils in their place' (Watson 1979: 263).
72
Nomads in archaeology
Table 5.1 continued Hasanabad Sedentary village (after Watson 1966:1814). Metal: Iron plate for cooking bread Metal support for fire Tongs; poker Pots; pans Small metal bowls (1 or 2) Metal tablespoon Tea kettle Teapot Sugar hatchet Steel awls, needles Axe or adze Kerosene lamp Pocket knives Tin container Bowls & pots - tinned copper Wood: Bread board Flour sifter Large wooden spoons Small tray Sugar bowl Meat beater
Spindles (2 or 3) Matches Mortar & pestle for salt Simple vertical loom Ceramic: Imported jars Plain bottle-like jars rUcc. Vofldbb. Tea glasses (2 or 3) & saucers Textile: Woven rugs (1 or 2) Felt mat Quilts (2 or 3) Bolsters (2 or 3) Woven bag to hold salt Leather: Goatskin bag Basket: Wicker work pot lids Basket tray - for draining rice Stone: Rotary quern
Yoruks Nomads, Western Anatolia (Bent 1891:271-5). (Observed within occupied camp.) Camel pack saddles - textile & leather Skin churns Churning tripod - wood Wooden stirrer Wooden mortar - for pounding coffee
Wooden dishes Wooden bowls Copper utensils Copper heirlooms
Bakhtiari Nomads, Central Zagros, Iran (Thesiger 1979:123). (Observed during migration.) Tents Bedding Pots - metal Pans - metal Sacks of grain Baharvand Lurs Central Zagros, Luristan, Iran (Baharvand 1975:57). Bags Bedding Pots (not specified whether metal or ceramic) Other cooking utensils Pack bags - for grain Carpet bags Goatskins - for water Lamb & kid skins - for butter Blankets - placed under the loads carried by animals Ropes
Nomads - the invisible culture?
73
Lurs Nomads, Central Zagros, Iran (Stark 1933:250). (Observed in camp.) Convex metal shield - for bread baking A few copper vessels A rug or two Some goatskins - for carrying water Comments: 'Poverty had simplified their household furnishings' (Stark 1933: 250).
Basseri Southern Zagros, Fars, Iran (Barth 1961:11,123). (Observed during migration period.) Rugs Blankets Pails - milk Skins - milk Cooking pots Pack bags Other equipment
Kirghiz Pamir Mountains (Hedin 18981:423). Metal: Large iron cooking pot Iron or copper cans with handles or lids Potstand & poker - iron Hatchets-iron
Ceramic: China basins Other jent Saddies
Wood: Flat wooden dishes Loom Kneading trough
Bedclothes Carpets Sacks of grain & flour Goat bags
Cor
Comments: 'Household goods are not many... accommodated b y 2 o r 3 y a k s ' (Hedin 18981:423).
» sieve j,, Fiddle & guitar
these observations have been made in the course of migration when nomad material culture is 'on display', and the impression is usually one of structured chaos and highly eclectic modes of transportation (e.g. Mounsey 1872, pp.247-8). One early military observer provides the following vivid account of the Kurdish Herki tribe on migration where the borders of modern Iran, Iraq and Turkey intersect: A long, straggling string of laden travellers, mules, horses and cattle . . . They travel in a leisurely fashion, two or three families apparently travelling together, their impedimenta of sacks of corn, tents, rugs, felt mats, clothes etc. loaded onto horses and cattle while here and there a sick sheep or calf will be tied onto the top of a load. The women plod along in rags, carrying on their backs an odd assortment of cradles, pots, piles, trays, butter-making skins, gourds, samovars, gear for weaving and spinning and the inevitable baby. (Dickson 1910, p. 13)
Nomads in archaeology
74
Other Kurdish groups on the march in the early 1920s are described as: driving along thousands of sheep and goats, their tents, cooking pots, sacks of grain and household impedimenta piled up on the backs or ponies, cattle and women. (Edmonds 1957, p. 13) oxen and donkeys laden with black tents, poles, reed screens, huge cooking pots, skins of rhon (yoghurt), cradles and babies. (Edmonds 1957, p.366) Casual observations made in tents suggest that nomad material culture may have been somewhat simpler in the recent past than at present, though the economy of space used in stacking items and concealing them under rugs can be deceptive to all but the most dedicated observer. Freya Stark;s descriptions of Luri tents in the 1930s suggest a relatively impoverished material culture: 'Poverty simplified the household furniture. A few copper vessels, a rug or two and some goat skins for water was all I could see' (Stark 1933, p.251). So too in the Elburz Mountains near the Caspian Sea she talks of only the 'bare essentials' in summer campsites (Stark 1934, p.423). It should perhaps be borne in mind that this was the period of Reza Shah's sedentarization campaign and not a time of great prosperity for nomads in general. However earlier observations (e.g. Hedin 1898 I, p.423) seem to confirm a paucity of material culture. For instance the Hamawand Kurds were described as: 'not much better than armed gipsies. Their tents are filthy and they have very little of the commodities which pass for riches in that country' (Dickson 1910, p.376). It should also be noted, however, that comparable observations in villages of the period have also yielded an impression of equally spartan furnishings (e.g. Brandenberg 1905; Mordtmann 1925). Although any assessment is bound to be tenuous, the ethnohistorical evidence suggests that, questions of relative wealth and poverty aside, the range of items in the average nomad household inventory is comparable with that observed today.
Changes in inventories Most items of nomad material culture are portable, though heavy items such as tents, poles, shovels and even axes may be cached. Many items are of metal and therefore potentially durable - these are also valuable and will be curated and repaired and will have a low rate of entry into archaeological context. Even today small groups of tradesmen make a viable living travelling from campsite to campsite fixing metal pots, repairing shoes, etc., as well as performing circumcisions (e.g. Barth 1961). The high proportion of metal recorded among debris on some of my southern Anatolian campsites was due to the ready availability of castoffs from nearby modern settlements - petrol cans, jam tins, metal strips, nails from packing cases, and wire. Prior to the advent of relatively cheap tin and aluminium containers it is possible that functionally equivalent items were made of more perishable materials such as wood or skins. Among the Baharvand in southwest Iran metal trays and dishes have to an extent replaced wooden precursors (Baharvand 1975, p.57; Hole 1978, p. 149). A similar process seems to have
Nomads - the invisible culture?
75
occurred in Anatolia; if early travellers; reports are to be believed. Among Yoriiks on the Mediterranean coast wooden artifacts such as mortars for pounding coffee; dishes and bowls were said to be 'most generally in use', although 'each tent has its heirlooms of copper utensils which are mended with great care and handed down for generations7 (Bent 1891, p.31). In recent years plastic has begun to replace metal and wood, particularly for large bowls, basins and buckets. It has the advantage of being light and portable, but wears out rapidly. In the course of numerous visits to campsites I found that plastic cannisters and buckets were among the most appreciated gifts. Plastic scrap formed a high proportion of debris on the campsites where surface recording was carried out. Plastic containers were used for carrying and storing water, though drinking water was kept cool in aluminium jugs. Skins were still used for churning yoghurt and cheese. One view of the impact of plastic on nomad material culture might be that it represents a modern corruption of a traditional culture. However it serves as ample illustration of the priorities according to which nomad material culture is organized. The rapid adoption of a material which fits the key criteria of portability, both in terms of its lightness and robustness (resistance to breakage), has important implications when we come to study the role of pottery, functionally equivalent but possessed of the opposite physical properties.
The role of pottery in nomad material culture The possibility that pottery may have been more extensively used by nomads in the past than it is today has recently been raised by Hole (1978, pp.150-1). Assuming that high-quality metal containers were in short supply in the distant past, wooden functional equivalents might have been employed though - significantly - not in operations involving heating, cooking or boiling. With the availability of new materials such as enamel or aluminium there has been a trend away from the use of pottery for these purposes so that the relative scarcity of sherds on modern nomad sites may well be deceptive in terms of our assumptions about older sites. Observations of abandoned Bedouin campsites in the Syrian Desert seem to confirm this: 'If the encampment is an old one, a sherd or two may also be found, but if recent there will be none as modern Bedouin generally use enamel or aluminium vessels' (de Schauensee 1968, p.36). However a recent impoverishment in ceramic technology and decline in usage has also been noted in settled communities. Hole (1978, p. 149) notes that, except for ceramic jugs to cool water, the household inventory of settled Iranian villagers includes very few ceramics, although the archaeological record attests to the fact that this was not always so in the past. Watson (1966, 1979) also found that her sedentary and pastoral communities differed little in the extent of their use of pottery.
Some preliminary hypotheses While the large cooking pots described in the passages quoted above were almost certainly of metal, pottery use among Near Eastern nomads is
Nomads in archaeology
76
widely known. Ceramics occupy an interesting position in nomad material culture since, in terms of our model, intact vessels are highly 'perishable7 in the sense that they are easily broken. This property also places them low on the scale of portability - one would not expect to find large ceramic pots on nomad baggage trains. However the sherds produced by the breakage of a pot are themselves highly durable. While potentially portable, they are of such low value that they become, in effect, site fixtures. Two types of ceramics might therefore be expected to occur on nomad sites. Small portable items such as teapots, cups, jars or clay pipes - most of which would be regarded as valuables - will be circulated from campsite to campsite, only occasionally entering archaeological context through breakage or loss. Large items such as cooking pots, storage jars or bowls might be expected to remain on particular sites as fixtures and to sustain a much higher rate of breakage or abandonment and incorporation into archaeological context. Ceramic cooking pots are known to have a relatively short use-life (de Boer and Lathrap 1979) and would be discarded on occupied campsites at a relatively constant rate. It is now possible to put forward a number of preliminary hypotheses concerning the analysis of pottery on nomad sites: (1) Pottery use is unlikely to be as intensive as for settled sites sharing the same material culture complex. (2) The range of sizes and types of vessel represented is likely to be much narrower than is the case for sedentary sites sharing the same cultural complex. (3) The distribution of sherds in terms of vessel size should be bimodal, with many sherds from large vessels (fixtures) and a smaller concentration at the lower end of the size spectrum. (4) Depending on the extent of use of pots as fixtures, we might find a size distribution skewed towards the upper or lower end of the range. These hypotheses suggest the kinds of variables that might be of interest to an archaeologist wishing to distinguish between nomad and sedentary sites on the basis of ceramics scatters alone. The data with which they might be tested would require a much more detailed recording and collection of sites, ancient and modern, than I was able to attempt. However scattered observations of contemporary campsites by archaeologists suggest that they may be pointing in the right direction.
Evaluation of the hypotheses Many campsites observed by myself and other archaeologists appear to contain very few sherds, widely scattered and of small size. Watson (1979, p.267) noticed 'a piece of tea saucer' lying in the Medan campsite above Shirdasht village, and describes a small ceramic stewing pot with perforated handles (ibid., fig. 10.4) in use by one tent-dwelling family. Many of the southern Anatolian campsites studied contained just a few scattered sherds from small vessels such as cups or teapots. China teapots and a few medium-sized ceramic jars were used at Sanaydin Yayla and some of the sherds recorded in the midden areas were from such vessels. Glass sherds from tea glasses were also common in and around tentsites. At one summer camp on Sultan Dag in western Anatolia, abandoned for
Nomads - the invisible culture?
77
around six months, a concentration of sherds from a single thin-walled jar was found in situ immediately in front of the kitchen area of a tentsite where someone had dropped it (Fig. 5.1a). Sherds were also found thinly scattered around other tentsites at Sultan Dag. In sharp contrast, the abandoned summer campsites recorded at Nemrut Dag in eastern Anatolia contained an abundance of coarse, blackened sherds from large cooking pots or bread ovens which are sunk into the ground outside the tents (Fig. 5.1b). The vessels had obviously been Fig. 5-1 Examples of ceramics on nomadic campsites. (a) Sherds from a single vessel at a campsite in western Turkey. The vessel was painted with a white geometric design. (b) Coarse sherds from cooking pots at a summer campsite at Nemrut Dag in eastern Turkey. The sherds are lying in a stone-walled tent enclosure near a recessed hearth. A small grinder manufactured from local low-grade obsidian is also shown.
Nomads in archaeology
78
imported into the area which contained no suitable clays, but had been installed as site-specific fixtures rather than being moved from site to site. These examples are highly suggestive in terms of the fourth of the hypotheses outlined above. The western Anatolian sites contained small quantities of mainly fine ware and glass sherds falling at the lower end of the range of vessel size; while those in eastern Anatolia contained large quantities of coarse sherds from very large vessels. Paradoxically the Kurdish tribes responsible for the latter assemblages are 'nomadic7 to a far greater degree than are the Yoriiks of western Anatolia, for the most part living under the tent for the entire annual cycle. The different assemblages represent alternative strategies in the organization of material culture constant circulation of valued, portable items in contrast to the installation of more common and cumbersome vessels as fixtures in particular campsites, to be used seasonally. These modern observations are borne out by ethnohistorical examples. Among some nomadic groups large ceramic vessels were used as churns in the production of cheese, though a goatskin container swung from a tripod is the more usual method (e.g. Bent 1891, p.31; Ferdinand 1959-60, p.36). One of the more intrepid and observant nineteenth-century travellers, Lady Ella Sykes (1898, p.155), displays one such churn in use near Kerman in western Iran (Fig. 5.2). The large pot, resting on a piece of Fig. 5-2 A large ceramic churn for yoghurt manufacture from Kerman, western Iran, in the last century. Note the construction of the tent in the background, of alacik type (see Chapter 6). (Reproduced from Sykes 1898, plate opposite p. 155.)
Nomads - the invisible culture?
79
Fig. 5.3 Grindstone and slabs from a campsite at Nemrut Dag in eastern Turkey.
cloth; is being rocked backwards and forwards by hand. Lynch (1901, p.323) observed a similar device in use for the same purpose, though apparently suspended from a tripod, not far from Nemrut Dag: 'A young woman . . . is engaged in swinging to and fro a large vessel of earthenware, which they use for making cheese'. Some of the surface sherds on the abandoned campsites at Nemrut Dag may have come from such a vessel though I did not observe any in use, nor are they mentioned by Be§ikgi (1969).
Stone artifacts Only at Nemrut Dag did I observe stone grinders and slabs. These were found inside the abandoned dwellings, apparently in their location of use and represent items of site furniture or fixtures (see Fig. 5.3). Hole (1978, p. 154) also notes that metates and manos are not found on Luri campsites, cereals being obtained in ground form from mills, though acorn smashers are frequently found near procurement locations. Projecting backwards to when lithic technology was widely used, it is possible that this would tend to raise the overall visibility of nomad sites
Nomads in archaeology
80
where lithics were used or manufactured. However, in view of the difficulties involved in distinguishing nomad from sedentary inventories, the same difficulty would arise in attempting to distinguish between lithic industries. The kinds of indicators suggested above in relation to ceramics would be less applicable here. Similarly, with nomad and hunter-gatherer sites, any such size/portability function would be equally applicable to both. Indeed as Hole (ibid., p.166) acknowledges, it may well be impossible on the basis of lithics alone to distinguish the campsites of nomadic pastoralists from those of hunter-gatherers.
Stratigraphy and architectural remains Although the shallow stratigraphy of nomad sites has been stressed, observations in some of the larger summer campsites in eastern Anatolia, which are occupied regularly and for long periods of time, indicate that well-defined stratigraphic sequences may develop, though these may be confined to external hearth and midden areas. The bread-oven pit shown in Fig. 5.4 contained not only sherds from the cooking pot which had once occupied it but also, observable in its walls, a complex stratigraphic record of earlier occupations. The assertion by Childe (1936) that nomad architecture is perishable and insubstantial is questionable, as we have seen above. The frequency with which stone tent footings or walls are employed by contemporary nomads and the ready availability of stone suggests that such practices may also have existed in the past. In the course of fieldwork not only were long-abandoned tentsites frequently encountered but also whole campsites. The campsite shown in Fig. 5.5 consisted of more than a dozen highly disturbed stone arrangements whose dimensions suggested that they were the remains of tentsites. The fact that some were almost submerged beneath layers of alluvium, together with an almost total lack of surface artifacts, indicated a respectable antiquity.
The issue of 'household impedimenta7 We have already seen how nomad material culture tends to reflect that employed by related sedentary groups and how at least some of the problems of maintaining an adequate level of material welfare can be solved by caching strategies and by fitting out campsites with fixed facilities and site furniture. However this does not alter the fact that a very substantial proportion of essential items needs to be transported, particularly heavy and cumbersome perishables such as tents, rugs, bedding and foodstuffs for day-to-day subsistence. The use of pack animals such as donkeys, horses, mules, camels and even cattle has provided the solution for many generations of nomads. But what of that critical period in the evolution of nomadism when these species remained undomesticated or, in the case of the mule, non-existent? To propose alternatives such as the use of sheep or goats to carry small loads as reported for Tibet (Ekvall 1961) or women as beasts of burden is to evade the central issue and deflect attention from this key logistical problem - the main limiting condition by which the nomadic option is circumscribed.
Nomads - the invisible culture?
81
Fig. 54 Micro-stratigraphy revealed in a bread-oven pit outside a tentsite at Nemrut Dag in eastern Turkey. Sherds from large cooking pots are scattered around the edges of the pit. The remains of a broken cooking pot are embedded in the pit.
Fig. 5-5 A long-abandoned campsite at Ala Dag in the highlands of central Turkey. The concentrations of stone represent old tentsites now partly submerged under the alluvium.
In arguing for an early origin of nomadic pastoralism in western Iran; Hole (1978, 1984) claims that pack animals are in principle not essential to a nomadic system, that their presence merely 'encourages the nomads to build inventories of equipment. . . and makes longer migrations possible7 (Hole 1978, p.150). This implies that household impedimenta are somehow peripheral to the essential nomadic way of life and can be shrunk down to the bare minimum to be carried on the backs of the people themselves. First to go is the tent, to be replaced by a rough shelter, or kula, constructed entirely from local materials at each campsite. Next to
Nomads in archaeology
82
go is bedding, through careful scheduling of migrations to avoid extremes of cold. Then the sacks of grain or flour, noted in nearly all migration trains, are jettisoned and replaced by a diet of wild acorns which today serve as a back-up resource. With respect to the last point, the use of acorns to make a kind of bread has been reported widely elsewhere in Iran (Stark 1933, p.251) and in Anatolia (Bent 1891, p.273) but, significantly, Stark also notes that sustained reliance on this diet is 'fatal to many of the children'. So it is with the other solutions as well. A scenario such as that proposed by Hole is untenable due to a basic misunderstanding of the role of material culture in the lives of Near Eastern nomads. Certainly the average nomad is a hardy customer used to long spells of night grazing with nothing but the hard stones and a rough blanket for a bed. But the blankets and eiderdowns, the cushions, kilims and embroidered rugs - indeed the very tent itself - are much more than just home comforts. In a sense they come as a complete cultural package, a transformation of the same standards and models of household organization that apply in sedentary village society. In much the same way as that outlined for East African pastoralists (Hodder 1982), they offer a medium for the affirmation of tribal identity as well as the expression of subtle shades of status, rank and wealth. It is impossible to imagine a pastoral nomad aspiring to the level of simplicity idealized in, for instance, studies of the !Kung bushmen (Lee 1979) or Australian Aborigines (Gould 1980). Unlike the bushman he cannot give away all his material possessions in return for the esteem of his fellows. The ethic of generosity characteristic of hunter-gatherer society stands in sharp contrast to the code of hospitality which applies in nomad society, which implies asymmetric reciprocity and the expression of dominance and rank. Far from being incidental to nomad life such an ethic, together with the material means for its expression, is fundamental to the organization and mobilization of a nomadic community, and a relatively high profile of material culture must be seen as an ideal, though perhaps not always realized, which is built into Near Eastern nomadism. A nomad lacking the wherewithall to entertain a 'guest of Allah' is not worthy of the name. Consequently, even if a group of herders could scrape by for a season or two in the manner proposed by Hole, it is doubtful whether this would represent a durable and stable adaptation. Pastoral production, combined with a pattern of consumption based on wild resources, would inevitably result in major strains and conflicts between the schedule for exploiting wild resources and that necessary for successful herding - except in the happy circumstance that both kinds of resources are abundant in the same localities at the same time. In an economy centred around the singleminded pursuit of pasture there are obvious advantages in a pattern of consumption which is independent of the resources available in any particular area through which one is passing. In short, there exist powerful constraints and imperatives which would, in the course of time, transform Hole's 'herder foragers7 into either full-blown nomadic pastoralists with an independent food supply or hunter-gatherers who just happen to keep a few goats.
Nomads - the invisible culture?
83
Conclusion Thus although we are now able to move beyond the simple assumption that 'nomads leave no traces', it is equally clear that the material dimension is an important, and in many ways a limiting, factor in any nomadic system. We are prompted to inquire: If nomad material culture is as developed as has been suggested, then why are nomad sites so difficult to detect? The answers lie in a better understanding of the processes by which material remains are distributed over the landscape and the factors which control this distribution. It is necessary to consider in some detail the organization of particular nomadic pastoral systems in time and space.
6 NOMAD ARCHITECTURE AND DOMESTIC SPACE
As I approached Ali's winter camp all illusions about the simplicity of my task began to crumble. The tent - if it could be called that - was fashioned out of a mixture of rough limestone boulders and dressed blocks from a nearby Byzantine complex (Fig. 6.1). Its walls were nearly 1.5 metres high, and it was entered via a short passageway and a rickety wooden gate held by lengths of twisted wire. The structure was surmounted by wooden beams and a mixture of tentcloth; straw and polythene plastic. The interior of the dwelling contained the now-familiar array of rug-covered platforms though its proportions were less elongated than the norm. In the front wall, adjacent to the doorway, was a recessed fireplace beautifully constructed out of Byzantine building rubble and capped by a large limestone block covered by floral bas-reliefs. In summer Ali would be up in the Taurus Mountains living in a black tent with only the barest of stone footings. The following year at about the same time (early September) I found him camped some 200 metres away in a small beige-coloured tent whose interior was almost identical to that observed in the previous winter, but which was devoid of any foundation other than a cleared earth floor and a small hearth consisting of three stones. The term nomad architecture may appear to contain an inherent contradiction. In fact there is a sizeable literature on the subject (Fagre 1979; Andrews 1981; Ferdinand 1959-60; Feilberg 1944; Edelberg 1966-67). Much of this is confined to a discussion of fabrics and supports, those portable elements which are least likely to survive in archaeological context. The distinction drawn in Chapter 5 between portables andfixturesis highly relevant to any discussion of nomad architecture. Much of the present discussion will focus on fixtures and the more durable elements of tent architecture. However, while tentcloth and tentpoles are perishable, the very size and shape of nomad tentsites have been influenced by the inherent potentials and constraints imposed by woven fabrics and supports. 84
Nomad architecture
85
Fig. 6.1 Ali's winter tentsite, southern Turkey. The stone-walled tentsite was surmounted by a temporary structure of straw, tentcloth and polythene plastic.
What is a tent? In view of the difficulty of defining precisely where tent forms end and house forms begin I have frequently had recourse to the neutral term 'dwelling7. Bedouin tents have been defined as 'an eminently practical, completely portable house demonstrating all the principal elements of permanent architecture7 (de Schauensee 1968, p.32). Most other definitions also rely on analogies with permanent architecture, though that used by Hole (1984), which permits the term 'tent' regardless of 'whether or not the space . . . was actually covered with a woven fabric', approaches closer to an archaeologist's understanding. For our purposes a tent may be defined as a prefabricated structure consisting of a flexible covering and structural supports temporarily brought together to form an integrated architec-
tural unit. As indicated above, tents are probably best regarded as covering a certain range on a continuum from extremely temporary arrangements to more permanent dwelling forms. As archaeologists, rather than architects, what interests us most are the dimensions, features and configurations of household space which arise where the tent form is employed. These in turn may depend on the type of portable structure employed.
Types of nomad tents Nomad tents may be divided into a number of classes, each of which contains its own implications and limitations in terms of floorplans and internal organization. The Near East is the meeting place of two great
Nomads in archaeology
86
tentage traditions of entirely independent origins and principles of construction, the one based on the beehive-shaped Central Asian yurt or kibitka, the other based on the locally evolved black tent. There has been a certain amount of hybridization, but for our purposes the following categories may be discerned.
Central Asian types Tents known as topak ev (literally, round house), or alacik in Anatolia, consist of a rigid wooden frame over which is draped an inert covering of felt (see Fig. 6.2; also Andrews 1981; Duly 1979, pl.88; Rappoport 1969, p.110). The frame is usually a hemispherical structure of curved ribs plugged into a central wheel at the top (see Land 1966, fig.3). The wheel is supported by a central pole and secured by a pair of ropes twisted together at high tension and fastened to a stout wooden stake driven into the centre of the floor to prevent wind uplift. The walls may consist of ribs only, as in the Shahsevan tent (Land 1966; Tapper 1979) or the ribs may be shorter, fastened to the top of a collapsible, cylindrical lattice to form the 'trellis7 tent (Andrews 1981; Ozkan 1978). A wickerwork or reed screen is then placed loosely around the felt-covered walls. The term alacik was initially used to describe this screen and retains that meaning among the Qashqa'i (Marsden 1976, p.26). More commonly the term stands for the tent as a whole, as among the Shahsevan (Land 1966) and throughout Anatolia. This is perhaps the most functionally specialized of all tent types, the product of a long evolution on the Central Asian steppes, perfected some time before the Turkic migrations into the Near East in the eleventh and twelfth centuries (see Andrews 1981). Such is the skill involved in its manufacture that it was often custombuilt by specialists. One old man living in a topak ev in central Turkey told me that all the tentframes in the camp had been made by a single carpenter in the provincial capital. This kind of structure is designed for protection against cold and strong winds rather than rain as the thick felt covering is slow to dry out. Its distribution is confined to Central Asia, Mongolia and the northern region of the Near East among Turkic-speaking groups such as the Yomut of the Gurgan Plain (Irons 1975), the Shahsevan of Azerbaijan (Tapper 1979c), Uzbeks and Kirghiz in northern Afghanistan (Shahrani 1979; and Michaud 1977) and scattered groups of relatively recent immigrants to Anatolia (Dominian 1915).
The black tent The Middle Eastern black tent, kara qadir (Turkish) or siyah chadour (Persian), by way of contrast, is a tensile structure consisting of tentcloth, usually woven goat-hair, supported by vertical poles and anchored by guy ropes (see Fig.6.3). Its principle of construction differs from the Central Asian type in that the cloth itself plays a role in holding up the structure (Evans 1983). The tentcloth consists of a series of standardized strips fastened together by wooden pegs (see Baharvand 1975, p.58; Watson 1979, figs. 10.1 and 10.2). Due to the properties of this material and the nature of construction, the black tent is well designed for protection
Nomad architecture
87
Fig. 6.2 A topak ev or alaak tent form from central Turkey. These frames are made in a nearby town by specialist craftsmen.
Fig. 6.3 Black tents, eastern Turkey. The black tent is able to accommodate changes in the size and composition of households.
against both heat and rain. In hot ; dry weather the loosely woven fabric permits the circulation of air within the tent but ; when wet, the goat-hair fibre expands to form a fairly waterproof shield (Baharvand 1975, p.58). Alternative tent forms are easily achieved by varying the numbers and arrangement of poles and strips of tentcloth to produce anything from the small tig direkli gadir (three-poled tent) used by Yoriiks in southern Anatolia (Giirdal 1976; Johansen 1965; Kademoglu 1974; Johnson 1969; Bates 1973)
Nomads in archaeology
to the huge ten- or twelve-pole structures of Kurdish tribes in eastern Anatolia (Ozbek-Koroglu 1980; Be§ikgi 1969) and the Qashqa'i of southwestern Iran (Marsden 1976). Such an architectural form is highly responsive to the variable requirements of seasonal settlements, fluctuating household size and the demands of hospitality, so that variations in size and form may occur not only between ethnic groups but also within individual encampments. This structural generalization means that the black tent does not lend itself so readily to classification by type as does the Central Asian form. Though more adaptable than the latter, the black tent is also less selfcontained and may require the addition of stone or mud walls to make it habitable during lengthy stopovers. The precise historical origins of the black multi-poled tent are uncertain (Daker 1984). It may well reach back into prehistoric times, but its evolution is most probably linked with that of village architecture - the tentcloth providing a convenient roofing material for sets of pre-existing walls. It is also the prototype for a much more elaborate form of 'royal tent' used by rulers in the Near East from early historical times (Evans 1983). Examples of such royal encampments have been described by Andrews (1981).
Barrel-vaulted tents This elliptical or semicylindrical form shares elements of those described above - more or less rectangular in plan but semicircular in section. On structural criteria, barrel-vaulted tents may be divided into two types. The first is a variant of the Central Asian yurt (Andrews pers. comm.). Instead of a central wheel, a straight ridge-piece with a row of sockets on each side runs down the spine of the roof, supported by a pole at each end. Ribs, similar to those used in the alaak or topak ev are fitted into this to produce something analogous to the inverted hull of a boat. Many of the components are interchangeable with the hemispherical form (see Rappoport 1969,fig.5.12),the same structural elements being used to produce tents with either rectangular or circular plans. The second form is the 'bender7 (Andrews pers. comm.) or gatma ev, similar in external appearance to that described above, but differing fundamentally in its mode of construction. There is no central ridge-piece or vertical supports. Instead curved hoops, standing parallel to each other, are fastened to horizontal stays and other curved pieces arranged concentrically at each end (Figs. 6.4 and 6.5). Such tents, also known as alaak, have been described in the central Taurus Mountains: The tents of this oymak (Kulfalli tribe) are made of felt . . . Their shape is round and long like a tunnel. The inner frame of the tent is criss-crossed with long sticks and these are covered with felt7 (Yalman 1977, p.214, my translation). Other reports mention both yurts and black tents in a Central Anatolian village (Naumann 1893). It is often difficult to discern whether gatma ev or Central Asian forms are being described. For instance, Mordtmann (1925, p.257) states 'Die Zelte der Jliriiken sehen wie Frachtwagen aus;, and photographs by Brandenberg (1905, p.12) are ambiguous on this point. Libby and Hoskins (1905) have described a single campsite in northern
89
Nomad architecture
Fig. 6.4 Frame of a qatma ev
or alaak at Sanaydin Yayla, Taurus Mountains, Turkey. The frame was left standing for the winter. Note the horizontal stays, curved end pieces and wickerwork screen.
rr
s
Syria apparently containing both circular and rectangular forms which are definitely of qatma ev type: These [Turkmen] shelters differ from all other tents in Syria in having no centre pole and no ropes. Five or six huge ribs; shaped like wagon bows, are stuck firmly into the ground, sometimes crossing each other, thus making a round tent, but more often the ribs are placed parallel to each other, forming an oval tent. A screen or wall of papyrus, about three feet high, is planted like a picket all around the bows with an opening left for the door. The black hair covering is then thrown over the bows and tied firmly down to them and the papyrus picket. (ibid., p.105) Similar tents are also found in northwestern Iran (d'Allemagne 1911, pi. onp.100). According to Andrews (pers. comm.) the 'bender7 represents a local development within the Near East, not a simplification of the Central Asian form. It may be considerably more ancient than the latter. In Turkish-speaking areas this form is known as alaak or qatma ev (throwntogether or improvised house) and has a circular counterpart in the form of the 'rib-strut tent7 (Andrews pers. comm.), a crude arrangement of concentric ribs, with or without a central pole. The barrel-vaulted bender is based on the same principle of construction as a variety of barrel-vaulted huts found in oases and marshlands in the Near East (Duly 1979, pl.86; Thesiger 1964). Either palm ribs (Ferdinand 1959-60, p.44) or reeds (Nissen 1968, p.109) are tied together in bundles and bent to form arches, then lashed together with horizontal stays of similar construction. The ends are rounded off with a semicircular arrangement of ribs tied to the frame. This frame is often covered with palm or reed mats and may also be plastered with mud. Dwellings of this kind are known from at least Sumerian times and are still employed today.
90
Nomads in archaeology
Fig. 6.5 Diagram of a gatma
ev or alaak, southern Turkey. (a) Frame and floorplan. (b) Cutaway view of a typical occupied alaak. Note the juxtaposition of public (lounge) and private (kitchen) domains.
poles tied stone platform (yiikluk)
h
wicker screen
plastered
^entrance . - platform
-r-
bedding; platform ' .
ALACIK CONSTRUCTION, ~~SV-1
felt
open for smoke
goat hair tent I cloth
\
bread
basket 3m
ALACIK INTERIOR, SV- 1
Ridge-pole tents The ridge-pole tent; common in the Taurus ranges of Antolia and east to Baluchistan (Ferdinand 1959-60); consists of two uprights over which a pole is placed. Over this is draped the same goat-hair tentcloth employed in the black tent.
Qardaks or Kulas
A 'bower7 of vertical poles topped with branches is commonly employed throughout the Near East by both nomads and villagers during the hottest
Nomad architecture
91
months (Edelberg 1966-67; Hansen 1961, p.22). Known as a gardak (Turkish) or kula (Persian), it is often used instead of a cloth tent in the summer camps of eastern Anatolia (Frodin 1943-4, p.285; Sykes 1908, p.455; Ainsworth 1842 II, p.238) and western Iran (Hole n.d.; Edelberg 1966-67; Stark 1933, pp.247-8; Loffler and Friedl 1967). Similar 'huts made of branches7 have been reported from as far east as Nuristan (Edelberg and Jones 1979, p. 100). This form is of particular interest as it is often used as a functional equivalent of the black tent or bender, may possess a similar floorplan and is frequently surrounded by the same wickerwork screen (Stark 1933, p.248; Loffler and Friedl 1967, pl.6). It is ideally suited to hot, dry conditions. Its open sides permit free circulation of air, while the roof is a more effective sun-screen than tentcloth, though not as impervious to rain.
Implications of variability in tent form The implications of this short excursion into tent typology are, first, that principles of construction help to determine the shape, floorplan and range of possible variation in floor size. Second, we are nevertheless reminded that different principles of construction can result in similar floorplans. There is ample evidence that tent form is linked to ethnic affiliation and culture complexes (Patai 1951) and that the distribution of tent forms can throw light on ethnic migration patterns. However there are also instances of groups adopting new tent forms as they enter new areas. This has occurred on a large scale in the course of Turkic migrations into Iran and Anatolia. Some of the more southerly groups of Turkic-speakers, such as the Qashqa'i and Yoriiks, at some time switched from the Central Asian form to the black tent or bender. Other equally dispersed Turkic groups have retained the Central Asian form. Environmental factors may have played a part here. Those retaining the Central Asian form are to be found permanently on the high plateau or the margins of the Inner Asian steppe, while those whose migration track includes the Mediterranean littoral or Mesopotamia (i.e. those subject to warmer, wetter winter conditions) are more likely to have adopted the black tent. However within any single region it is not uncommon to find two or more tent types in use by different ethnic or tribal groups (Johnson 1969, p. 158). For instance Pashtuns in Afghanistan using goat-hair tents rub shoulders with Uzbeks who employ Central Asian yurts (N. Tapper 1977, p.165). Ferdinand (1959-60) recorded a variety of tent forms employed by different ethnic groups in the Sistan region of Baluchistan. Moreover different tent forms may coexist within a single ethnic or tribal unit. In one of my fieldwork areas in southern Turkey I found camps of black tents and alaak tents within a few kilometres of each other, both belonging to Yoriiks, and in one case both types were present within a single campsite. Andrews (pers. comm.) also found black tents and bender or alaak tents in the same campsites in the Elburz Mountains of northern Iran. Once again the point is driven home that the connection between ethnic categories and material indicators is at best problematic.
Nomads in archaeology
92
Fixtures The kinds of fixtures employed in tent architecture are subject to less variation than for portable elements, though there are degrees of size or elaboration depending on how long the tentsite is occupied.
Hearths: chaleh (Persian), ocak (Turkish) Even in the most ephemeral transit camp some kind of hearth is bound to be present. Minimally this will consist of two or three hearth-stones or a shallow firepit (Barth 1961, p.12; Loffler and Friedl 1967, p.112). In longerterm seasonal campsites hearths are frequently surrounded by a protective wall of stones which may be plastered, while other fixed campsites contain quite elaborate recessed hearths with chimneys (see Figs. 6.6 and 6.7). The hearth is frequently the most conspicuous surface feature on an abandoned tentsite. Two hearths - interior and external - are commonly employed in summer camps, the smaller internal one for brewing tea (e.g. Hole 1968, p. 151), the larger external one for cooking and boiling. A small internal and larger external hearth are also common in Qashqa'i (Marsden 1976), Yoriik (Kademoglu 1974) and Bedouin (de Schauensee 1968, p.36) tentsites. Alternatively, subterranean ceramic cooking pots may be employed for baking bread, a widespread practice in Kurdistan (Lynch 1901, pp.301-4) and also among nomads in Morocco and Baluchistan (Andrews 1971, p.141).
Stone storage or bedding platforms: chul (Persian), yiikliik (Turkish) A low platform of stones or mud generally runs along the rear wall of the tent to act as a dry foundation for storage sacks or bedding materials (see Fig. 6.6 A plastered stone hearth inside an alaak at Sanaydin Yayla in southern Turkey. Yufka bread is being turned on a metal plate.
Nomad architecture
Fig. 6.7 Stone-built fireplaces from tentsites in southern and eastern Turkey. (a) External hearth structure from a summer camp at Sultan Dag in western Turkey. (b) Recessed hearth with vent inside a tentsite at Nemrut Dag, eastern Turkey.
93
Nomads in archaeology
94
Figs. 6.8 and 6.9). An additional stone platform may be found along the front wall (Hole 1978, pp.151-2), or the platform may encircle the tent on three sides. Platforms are sometimes plastered or entirely of mud construction (Audouze and Jarrige 1981). Stone platforms are common even in short-term transit camps; though sometimes platforms of twigs and branches are used during short stopovers (Barth 1961, p. 12), possibly where field stones are absent. According to one observer in the Syrian desert: These [stone] platforms are usually the chief evidence of deserted Bedouin encampments, sometimes associated with the ruins of a sheep enclosure7 (de Schauensee 1968, p.36).
Fig. 6.8 Plan of a tentsite with internal and external hearth, Sanaydin Yayla, southern Turkey. This is a typical floorplan for tentsites in the region. The internal hearth is employed for baking bread, brewing tea and smoking meat, while the external one is used for heavy cooking, boiling water in cauldrons,
oo rral
Nomad architecture
95
Fig. 6.9 A long-abandoned tentsite with levelled floor and disturbed stone platform, in the Taurus Mountains, southern Turkey. Nearby villagers could not recall nomads having camped there in the last 20 years.
Levelled floors Particularly where tents have been pitched on a slope and occupied for any length of time, the floor area is levelled out by cutting and filling (Ehmann 1975; Digard 1975, p.120) and lined with stone margins and baulks. According to Hole (1984) this was not practised by the Baharvand, possibly because the tentsites observed were transit camps (also Barth 1961, pp.15-16). Additional platforms or terraces may be constructed in front of the tents, also through cutting and filling. A drainage ditch is sometimes placed along the uphill margin of the site. If occupation is lengthy and regular, floors are smoothed and compacted or even plastered. Even tents of the Central Asian variety can leave traces of this kind. The Shahsevan dig a shallow trench, which contains a stone hearth, from the entrance to the central pole, (Land 1966, p.241) and, when on sloping ground, construct circular cut and fill terraces, in some cases lined with kerb stones (Andrews pers. comm.).
Walls The use of substantial stone walls to enclose tentsites is so common that it must be considered a standard feature in the repertoire of tent architecture, and ethnohistorical accounts indicate some antiquity for this practice. Stone-walled tentsites have been reported among Turkmen tribes in southern Anatolia from the early nineteenth century (Aswad 1971). A black tent pitched over 4-foot-high stone walls is described by Burkhardt (1822, p.636) and 2-foot-high walls surrounding a Kurdish tent near
Nomads in archaeology
96
Erzurum in eastern Anatolia were observed in the mid-nineteenth century: 'bis zwei Fuss iiber dem boden sich erhebende die Zeltwande bildet7 (Wagner 1852, p.267). At the bare minimum, the perimeter may be outlined with kerb stones which can also serve to secure a wickerwork screen. Stone walls are used in winter camps or high-altitude summer camps to screen out cold winds (e.g. Imhof etai, 1970, pl.68; Frodin 194344, p.258). Where stone is unavailable, mud brick or pis is often substituted: 'I could see the mud walls which surround the winter tent sites of the Tarhanis on the farther side [of the river] built up on salty limestone foothills' (Stark 1933, p.253). The Lurs are recorded as plastering mat walls or wickerwork screens with mud to make them windproof (Fagre 1979, p.46; Edmonds 1922, p.448). Substantial mud tent walls occur in arid areas of Afghanistan (Ferdinand 1962, p.131; Jentsch 1973, pp.226, 229) and Baluchistan (Audouze and Jarrige 1978, 1981). Mud walls may also enclose kulas or qardaks (e.g. Baharvand 1975, p.387; Stark 1933, p.249).
Storage structures Apart from the stone bedding platforms already mentioned, specialized storage structures in the vicinity of the tent are not uncommon. Edelberg (1966-67, p.388) describes grain bins and pots adjacent to tentsites while Hole (1978, p. 153) reports grain storage bins 'close to fields and threshing floors rather than alongside tents'. Stark (1933, p.248) describes 'mounds of earth which cover up the winter fodder or straw7 immediately outside Luri tents. Although the sites left by Central Asian tent types may be ephemeral, these are often accompanied by stone storage huts for storing fodder (Michaud and Michaud 1977, pi. 14) and this practice is also known from the ethnohistorical record: They build houses which are used as stables and granaries, but for dwelling places they prefer always the raised tent in the courtyard7 (Vambery 1868, p.303). During Reza Shah's sedentarization campaign in western Luristan in the 1930s, Freya Stark (1933, p.258) noticed that Luri nomads stored grain in the small huts 'which the government has made [them] build but which [they] cannot be induced to live in7.
Corrals Stone-built corrals are common features on nomad sites and also have counterparts of mud in parts of Iran and Afghanistan (Jentsch 1973, p.229). Some of these may be actually roofed over with a large black tent of the same kind used for dwellings (see Fig. 6.13c) so that it may be difficult to distinguish corral walls from tent footings apart from the absence of hearths and platforms and the presence of animal droppings. Brush or thornbush is often used for corrals in short-term transit camps.
Composite dwellings Tents accompanied by substantial fixtures - referred to henceforth as composite dwellings - represent a trade-off between the demands of
Nomad architecture
97
mobility and flexibility on the one hand, and the need for a degree of security and comfort on the other. The portable and perishable superstructure may be either carried from one campsite to another or cached nearby. Tentpoles are often cached on or adjacent to a campsite. The entire frames of topak ev structures may be left at summer and winter campsites (i.e. two sets of complete frames are kept by each family) (Ozkan 1978, p.30). The more durable fixtures form the substructure and may be expected to remain more or less in situ whether or not the campsite is' re-occupied. The composite dwelling exists only when a set of portables - tentcloth, ropes, poles, screens - is placed over a set of fixtures, or when a set of fixtures is constructed to fit the assembled portable elements. In terms of our formation process model, while portables exist more or less permanently in systemic context, but enter a site only seasonally, fixtures reside permanently on the site but are activated (i.e. enter systemic context) only when the site is occupied. Therefore fixtures alternate between systemic and archaeological context according to the ebb and flow of nomad migration. The use of composite dwellings is not confined to tent-using nomads. An interesting example occurs among some Kurdish villagers where the valuable roof timbers were carefully conserved while mud walls were considered expendable. The villagers frequently moved roofing material onto existing substructures or built new ones: 'If a Kurd does not like the site of his village, he simply pulls off the roof, loads the wood on his donkey and moves elsewhere7 (Hay 1921, p.45).
Tents and houses It has already been noted that tents of different types may display similar floorplans and fixtures. Nor should it now surprise us that in some regions the organization of space within both tents and village houses is very similar. Although permanent dwellings often consist of multiple rooms, most of these are for storage or livestock, while much domestic activity takes place within a single living room (see Watson 1978; Kramer 1979; Jacobs 1979). While construction materials differ radically, the similarity in patterning of functional space to that found in nomad tents is often striking and has led one observer (Watson 1979, p.280) to draw close parallels between the two. In part the similarity may be due to the persistence of nomad traits among sedentarized village populations. However cases have been reported of newly settled nomads imitating the design and layout of existing village houses (Sweet 1974; Bates 1973, p.193). Where seasonal dwellings of the same group of people are concerned, the conventions governing the use of space and deployment of household goods will generally apply equally to an enclosed tentsite or a permanently roofed dwelling - this was observed for both tents and huts at Sanaydin Yayla (see Chapter 9). The organization of living space observed in the Hulailan Valley is remarkably consistent, although the structures involved - zemga, tent and kula - are of markedly different construction
Nomads in archaeology
Fig. 6.10 Floorplans of tentsites and houses employed by villagers in western Iran. (a) Stone-walled, singleroomed village house with a markedly 'tentlike' floorplan. (b) Floorplan of a tent set up in the courtyard adjacent to the house in anticipation of the spring migration. (c) Tentsite in a nearby summer campsite. (After Watson 1979, fig.10.)
98
pans
Wa
~^^ bedding platform
hearth .
3
n
stone
platform j
/
wicker screen
^ bedding platform
* x
stacked pots & wool
hearth samovar wooden chest
mud storage chest hearth wall niches
V//////A m
(Edelberg 1966-67), (see Fig. 6.16). The same is true of seasonally used house and tent plans of a single family in Luristan (see Fig. 6.10; also Watson 1979) and of a tent floor and house plan located side by side in the same enclosure from Baluchistan (see Fig. 6.11; also Audouze and Jarrige 1978, 1981). But there are exceptions. Examples of tents and huts employed by the Beritanli tribe in eastern Anatolia display variations in their internal organization, though the range of variation between tents is also considerable (see Fig. 6.12). Is it possible to distinguish between tents and houses on the basis of floorplans alone, independently of materials used or mode of construction? Hole (n.d.) has suggested that nomad tentsites are, on average, larger
Nomad architecture
99
Fig. 6.11 Floorplans and elevations of adjacent tentsite and hut from Baluchistan. Note the similarity in dimensions, orientation and internal organization. The only archaeologically visible distinction would be three depressions for tentpoles as opposed to a single posthole in the case of the hut. (After Audouze and Jarrige 1981, fig.5.)
than village living rooms. At the same time the amount of living space per person is quite small. Variability in tent size was found to be due mainly to variation in length rather than width, perhaps because of the relative ease of adding more tentpoles as opposed to increasing the span of tentcloth. However the small size of many of the tents encountered in the course of this study argues for caution in the use of such formulae for distinguishing tent from house plans. Furthermore the extremely large size of some tents in eastern Anatolia is often accompanied by the internal division of living space, suggesting that the 'indivisibility7 of space within a tent (Hole n.d.) is a local rather than a universal feature.
Nomads in archaeology
100
The spatial organization of the tent As the tent generally lacks the cellular structure of the typical village house in which rooms may be added to accomodate specialist activities or new social units, the organization of space within it needs to be carefully regulated. Kinetic studies of spatial organization in unplanned settlements (e.g. Fletcher 1977) have found that each community or culture develops a kind of conceptual map of the dimensions and scale of architectural features and that cultural preferences, guided by social and practical considerations, tend to develop strong regularities through time. The compromise involved in achieving a satisfactory 'fit' between portable and fixed architectural elements imposes obvious constraints on the use of internal space. As Fagre (1979, p.7) points out: The space within the ordinary nomad tent is not large and so must be carefully organized. This organization is always a reflection of social organization and determines where people are seated and where possessions are kept\ Within tents of rectangular or elliptical plan, the internal arrangement of household possessions and activities is remarkably consistent. There is invariably a pile of storage sacks and household items running down the rear wall. One end of the tent may carry an additionai storage platform, while kitchen utensils and a hearth occupy the other end. One side of the tent is generally reserved for more formal social activity such as meals and entertainment, while the other side - the kitchen end - is given over to food preparation. It is possible to infer the location of these zones from stone platforms, hearths and other surface traces on almost any abandoned campsite and, by extension, any occupation floor which might emerge on an excavated campsite. Fagre (1979, p.7) maintains that tents are always divided along the lines of the division of labour between men and women, and Qashqa'i tents are also said to be divided into two parts: 'the left or west side of the tent is the woman's domain, where domestic and family activities take place, while on the right and east side the men congregate and entertain visitors' (Marsden 1976, p.25). The Baharvand maintain the same pattern (Hole n.d.; Baharvand 1975, p.58), and also the Bakhtiari, though apparently in the reverse order (Digard 1975, p.128). In addition to men's and women's quarters there is often a separate section within the tent for animals Qaubert 1921, p.82; de Morgan 1894-1901 II, p.32). These sections are often divided by reed screens (Lynch 1901, p.323; Ozbek-Koroglu 1980) though the use of stone partitions is not unknown (Figs.6.12, 6.13) and appears to have some ethnohistorical depth: Tinterieur, une cloison transversale de pierres divise la maison [sic] en une grande piece, a gauche pour les femmes, et une plus petite, a droite, pour les homines, devant laquelle it peut y avoir encore une petite piece pour une cheval' (Burckhardt 1822, p.636). Whatever the internal divisions, the tent is rarely a self-contained spatial unit. Activities and their associated paraphernalia tend to spill outside, particularly in the case of open-fronted summer tents. It is convenient to envisage the division of space in and around a tent not in terms of fixed zones but rather as adjustable and overlapping domains. The 'men's sidewomen's side' division is but one of a number of possible arrangements,
Nomad architecture
101
and domains may be reconstituted to serve different social occasions through the rearrangement of materials or the installation of temporary screens or hangings. Domain boundaries can be shifted at will and it is possible for two domains to successively occupy the same physical space - guest area during the day, domestic area at night. Some, such as kitchen and work domains, may extend wholly or partially outside the tent itself. Two or more domains may be integrated with each other such as where stacks of stored belongings serve as back-rests on a kind of divan made up of mats and cushions.
Indications of wealth and status As a general rule it may be accepted that all nuclear families within a certain cultural context require roughly the same amount of space for domestic concerns. Larger tents become necessary either to house coresiding members of large extended families or to meet the demands of extra entertaining and hospitality. Wealth or status and household size appear to be correlated due to the fact that a large labour force is necessary in order to generate wealth. At the same time, as seen in Chapter 4, wealth of a household head may tend to delay the separation of new nuclear units (Bates 1973; Be§ikgi 1969; Irons 1975). Casual observers of tent architecture have long been aware of the correlation between wealth or status and tent size. Around the turn of the century it was observed that among the Moukri Kurds: La grandeur des tentes varie avec l'importance de la familie qui l'habite. Cette tente de chef 13 m. de long, est la plus grande qu'ait. (de Morgan 1894-1901 II, p.32) This is well within the size range of the larger Alikanli tentsites observed by the author in eastern Anatolia. And more recently: The size of these tents [Hakkan, southeast Turkey] depends on the status and income of the family within the a$iret. Some of the tents are large enough only for 3-5 people, while others could accommodate [seat] 40-50 people. (Hakkan II Yilligi 1972, p.93; my translation) [Among the Qashqa'i] tent size varies slightly according to income, indicated by the number of poles in front. (Marsden 1976, p.25) Wealth or status may be indicated in other ways. The larger Beritanli tents possess a separate area for entertaining guests and, in some cases, a special sleeping area for guests, though an alternative arrangement is often found in the form of white, circular guest tents situated close to the tents of tribal notables. A connection between tent size, and therefore the diversity of domains within it, and the prominence of the occupying household is suggested by Baharvand;s (1975, p.58) observation that larger tents and guest tents among the well-to-do were the result of the demands of hospitality rather than wealth per se. He goes on to say that only the wealthy could afford to amass the numbers of pack animals and manpower needed to transport larger tents.
102
Nomads in archaeology
Fig. 6.12 Variability in dwelling forms among the Beritanli, eastern Turkey. (a) Stone-walled tentsite and hut. (b) Tentsite enclosing both living areas and corral. (c) Winter corral tent: there are no internal features. (Based on drawings provided by Meral OzbekKoroglu, METU (not to scale).)
firewood store barley storage area
entrance
hand washing
storage platform
large cauldron for boiling water sleeping area
entertaining guests and living area
s[eeping a r e a
straw and cushions"
walls 180 cm high 50 cm thick
Nomad architecture
103
bowl for hand washing
yukluK
(storage platform)
dung storage-
entrance
reed screen partitions male
entrance ••••
forsheep b
1 m high stone walls
corral section
entrance
VtaaoaBOBCOBCB 1 m high stone walls cloth
rope
Nomads in archaeology
104
Fig. 6.13 Seasonal variation in Yoriik tig direkli gadir
(three-poled) black tent plan and facilities, southern Turkey. (a) Winter tent plan: note enclosing stone footings. (b) Summer tent plan with storage platforms only. (c) End elevation of a winter tent. (After Kademoglu 1974 (not to scale).)
stone walls
I did not find such great variations in tent size among the Yoriiks in western Anatolia. This may be partly due to the nature of their construction particularly in the case of the alaak, although the same lack of size variation was observed among users of the tig direkli gdir, or three-poled black tent, which is potentially expandable to four or more poles. Neither did there appear to be as great a specialization in domains. Both of these features may reflect the relatively low level of ostentatious hospitality among the Yoriiks, which in turn may be connected to a decline in the institution of chieftainship and reduction of pan-tribal structures. Similarly the camps of smaller Kurdish tribes to the west of the Beritanli lacked any marked variations in tent sizes as did those of the non-tribal Kurds of Shirdasht in western Iran (Watson 1979,ch.8).
Nomad architecture
105
Tent size varies both according to culture or ethnic affiliation and ; within cultural/ethnic groups; according to differences in household size, wealth, status and political office. All of these latter variables appear to be inter-correlated so that it is not possible to state categorically that larger tents are due to larger households or, alternatively, to greater wealth or power. Large tent size, the partitioning of space within tents and the appearance of specialized units such as guest tents, appear to derive from an increase in the variety of activities, causing domains to become fixed and identified with functionally specific areas. The area under a tent is likely to reflect cultural standards or household size as much as status. A more accurate measure of household wealth and status might be the variety of domains and the presence of a range of specialized domains and/ or structures within or around the tentsite.
Towards a typology of dwelling forms Continuous variation in dwelling forms may be defined in terms of our three primary dimensions - portable/fixed, perishable/durable and valuable/expendable - the most ephemeral occurring where conditions selecting for portable, perishable and expendable elements are involved, the most substantial involving fixtures, durables and valuables. It is also possible to define arbitrary types of dwelling according to where they lie along this continuum. Bear in mind that we are concerned here with actual building components rather than floorplans (see Fig. 6.14). Dwelling type 7 The most basic dwelling form is the tent unaccompanied by any surface features or fixtures apart from a few peg holes and depressions left by tentpoles. This is rare, as even in the most transitory of camps hearth stones and concentrations of ash are also present. Dwelling type Z Here the tentsite may be marked by a storage platform and a stone hearth or firepit. This is commonly the case even in transit camps. Dwelling type 3 This would consist of a levelled terrace outlined with stones, perhaps surrounded by a drainage ditch, and storage platforms. Two hearths, internal and external, may be present, together with special storage structures and stone-built or wickerwork corrals. Dwelling type 4 — the composite dwelling Substantial walls of mud or stone enclose the tentsite, leaving a narrow entrance. Storage platforms and hearths are built into these walls. The tent itself serves as a temporary roof over this structure and side flaps or wickerwork screens may be dispensed with. In plan the structure will closely resemble a permanent dwelling.
106
Nomads in archaeology
Fig. 6.14 A typology of dwelling forms: tents, composite dwellings and houses. The elaboration of architectural features varies according to seasonality and sedentism.
TYPE 3
TYPE 4
TYPE 5
TYPE 6 1V1 Mi V I I v S3
TYPE 7
IF
Dwelling type 5 The fixtures are comparable with type 4, the main difference being that a portable superstructure is now directly supported by the walls. The actual covering may be either of tentcloth, matting, thatch or, in contemporary times, polythene plastic. A type 5 dwelling will contain no tentpoles, though one or more central posts may be used to prop up the roof. Dwelling type 6 Permanent walls extend to full height and now bear a fixed superstructure, for example, a series of beams crossed by rafters and topped by thatch or mud. The dwelling may still be composite to an extent, as in the case cited above of the Kurdish villager who carts his roof poles from one set of house walls to the next. Dwelling type 7 We now have a more or less permanent village house. The roof is almost as durable as the walls and the structure may have been extended by
Nomad architecture
107
adding more rooms. Storage and other specialized structures begin to proliferate. This typology is not intended to be exhaustive, nor should it be taken as an evolutionary scheme of development from simple to more complex architectural forms. It is a classification which is hopefully sensitive to our key variables of portability, durability and expendability. Considerably more than environmental conditions, such as the availability of building materials and cultural preference, appear to be involved in the choice of dwelling type (see Hole n.d.; Cook and Heizer 1968). The nomad case is of particular interest in that it involves the operation of our key variables on two different time scales - the seasonal cycle on the one hand and long-term changes in the degree of mobility on the other. It was stressed in Chapter 4 that characteristic nomad flexibility in organization and group formation operates on two levels and over two time scales. On a local scale, and within the range of a single annual cycle, seasonal changes in residence, environment and community organization are brought into focus. On a regional scale, and over a much longer timespan, our perspective shifts to progressive changes in migratory orbits and the degree to which a group is tied or untied in its commitment to a particular region. The architectural typology outlined above is but a particular case in which these two analytical levels may be applied.
Variation in dwelling types over the seasonal cycle Seasonal change in residence is frequently accompanied by a change in dwelling form. While this practice is by no means limited to nomadic pastoralists (see Orme 1981, ch.2), it is certainly one of their distinctive traits. While the change may involve no more than slight alterations in portable components, such as the deletion of side cloths while on migration (Barth 1961, p. 11), quite radical changes in the extent of fixtures and the use of portable items such as tentcloth are more common. Among the Qashqa'i three types of tents are employed. During migration a simple lean-to consisting of a strip of tentcloth and a few poles is set up. In summer an open-fronted summer tent is used, while in winter the living space is enclosed by stone walls except for a small entrance at the side and the tentcloth is sunk into the mud at the base of the walls to keep out draughts (Marsden 1976, p.25). Alternatively some families move into ordinary village huts or houses for part of the winter. The Baharvand spend three months of the year 'in tentsites enclosed on four sides' (Hole n.d.), and similar winter quarters are known among the Bakhtiari (Digard 1975; Cooper 1928) and other Lurs (Feilberg 1944,fig.12) who may also move into houses. During summer these groups occupy type 2 or type 3 tents. The Alikanli in eastern Anatolia - of whom we shall hear more presently - employ two completely different tents in winter and summer (Be§ikgi 1969). The winter tent is left behind in the care of village neighbours and a different and lighter tent taken to the summer quarters, though in both seasons substantial stone foundations are employed. The same tent fabric is employed by Yoriiks in western Anatolia, but there are
Nomads in archaeology
108
differences in the extent of elaboration and manner of arrangement of stone fixtures (see Fig. 6.14; also Kademoglu 1974). Such seasonal variations have some historical depth and consistency through time as indicated by ethnohistorical accounts: These people [eastern Anatolia] live in ordinary Kurdish tents in summer on the Taurus slopes, in winter they proceed to fixed camps about 30 miles to the west of Killis. The fixed camps have stone walls for the tents. (Sykes 1908, p.481) From here it is but a short step to the use of permanently roofed structures or fixed villages in winter (e.g. Chantre 1924, p.12). As indicated in Chapters 3 and 4, this in itself need not serve as an automatic disqualification from nomad status. In many cases the use of permanently roofed structures as opposed to walled tentsites does not accompany any essential difference in livelihood, settlement pattern or the pattern of migration. The Lurs in the central Zagros region illustrate well some of these ambiguities (Edelberg 1966-67) (see Fig. 6.15). Although characterized as 'farmers', these people also keep moderately sized herds of goats. Their migration cycle, all within the same valley, follows four stages, and their architectural repertoire includes three distinct dwelling forms. Spring and autumn are spent in black tents, but in the hottest summer months tents are abandoned in favour of kulas which may be set up close by. The third type of dwelling, known as a zemga, is employed in winter and is an archetypal example of a type 5 composite dwelling (Edelberg 1966-67, fig. 112). While certain sections - stable and fodder storage rooms - are roofed, the actual living quarters are covered with tentcloth and resemble in their layout the plan of a typical tent, or kula. In spring the tent is removed and set up nearby. A somewhat similar cycle occurred among the Baban Kurds in the early part of this century. From October to February they dwelt in villages in Upper Mesopotamia, moving into tents from May until early June (spring), whereupon 'they lay up their tents and construct bowers of green wood [kulas or qardaks] in which they live until the autumn nights grow chilly, when they return to their villages' (Sykes 1908, p.454). The practice of shifting between houses in winter and tents in summer is a long-standing one in the Near East. It is often taken as an indication of 'transhumance' or 'semi-nomadism7 and a by-product of recent sedentarization. While the latter may be true in some cases (e.g. N. Tapper 1977), such a pattern can also be stable and enduring (Evans 1983). Groups of people who make seasonal shifts in residence can be characterized according to the combination of dwelling types employed by them in different seasons. In archaeological terms, the presence of different dwelling forms in contemporaneous settlements occurring in different environmental zones within the same region may indicate seasonal shifts in residence rather than the presence of two adjacent 'cultures'. Archaeologists have been aware of this for some time (e.g. Hole 1978, n.d.; Orme 1981) but have been less receptive to the possibility that the very combinations of dwellings employed may change over time in response to much the same set of factors.
109
Nomad architecture
Fig. 6.15 Seasonal variation in dwelling form among villagers in western Iran. (a) Zemga : a composite dwelling extended to storage rooms and courtyard, employed in winter. (b) Kula, employed in summer. (c) black tent used in spring and autumn. (After Edelberg 1966-67.)
(a) Zemga 1. Stone wall, plastered 2. Hearth in kitchen 3. Manger for oxen 4. Forked main posts for roof 5. T-shaped tent posts 6. Goat skins for water in a niche in the wall 7. Containers of clay for skin filled with ghee 8. Containers of clay for floor or grain 9.
Box
10. Stand for bedding and carpets 11. Store for fodder 12. Room for livestock 13. Open yard for goats and sheep 14. Rush mats 15. Fence of dry oak-branches the leaves of which will rustle when touched by approaching wolf or thief (the wall around 13 is furnished with oak-branches too) (b) Kula 1. Sack for wheat and horse bag 2. Wrought iron tripod, yoke and wooden jar for xu/ 3. Rolled tent cloth and string bag for carrying straw 4. Hearth in kitchen 5. Iron griddle for baking bread 6. Fuel consisting of droppings from goats
7. 8.
Stand for bedding Stone platform with four water skins 9. Stone platform for milk products surrounded by rush matting 10. Stone platform for flour sack and bread plate 11. Carpets 12. Felt covering hearth lajd out for the reception of goods 13. Cushions 14. One of the forked posts 15. Entrance (c) Siah cador 1. Stone recess for broody hen 2. Stone platform for waterskin 3. Stone bed for carpets and bedding 4. Stone platform for flour sack and skin containing ghee 5. Hearth in kitchen 6. Hearth in guest room 7. Droppings from lambs and kids 8. Pile of stones indicating where children have been playing 9. Peg used to tether oxen 10. Peg used to tether calf 11. Sticks for attaching rush mats 12. Manure from oxen 13. Stone recess for dog 14. Entrance 15. Furrow for draining rain-water from tent cloth ,_ , „ . _ . , ._. (Edelberg (1966-67)
Nomads in archaeology
110
Long-term processes influencing dwelling forms An important factor governing the degree of elaboration of fixtures has to do with the likelihood of being able to re-use these fixtures in successive years. This in turn depends on factors such as fluctuations in the availability of pasture from year to year, the stability of the migration track and, above all, on the security of tenure in seasonal pastures. Where access to a given set of pastures may be denied, there is little point in investing much effort in the construction of facilities in those camps from which the pastures are exploited. However, where access is assured for the forseeable future, there will be fewer constraints on the tendency to make one's habitation as congenial as possible. Under modern conditions in the Near East access to pastures may be secured in three ways. The pastures may be owned by a corporate group, as is often the case where village communities have joint rights in distant summer pastures. This is very common throughout Anatolia where villages, or sections of villages, migrate as a unit to traditional summer yaylas which often contain permanent, but seasonally abandoned, houses or huts (Bruinessen 1979, pp.45-7; Tuncdilek 1962, 1963). Systems whereby tracts of pasture are individually owned are also known (Tapper 1979b). A second means of gaining secure access to pastures is through the establishment of a tribal territory, usually held in the name of the tribal leader and, in theory at least, distributed on an annual basis among tribal members (Digard 1979; Fazel 1973). In practice the pattern of allocation tends to remain fairly stable from year to year. Finally, where pastures are obtained on a rental basis the establishment of a stable, contractual relationship with the landlord or community owning the pasture is in the pastoralist's best interests. The undesirable consequences of being forced to rent pasture on a piecemeal basis have been described by Bates (1971, 1973) and will be illustrated by a case study in Chapter 7. Among the Beritan tribe in eastern Anatolia there exist great variations between camps in the degree of elaboration of fixtures (Ozbek-Koroglu 1980). The Kosan Kabile, which owns its summer pastures, has constructed fixed summer campsites with high stone walls and corrals. Other tribal sections, which rely on rented pasture, construct only the bare minimum of storage platforms and hearths though all groups spend much the same length of time in the same place. Although relying on rented pasture, the Alikanli summering on Nemrut Dag also build sturdy stone tentsites. Here, as described in Chapter 10, there is considerable fluidity in the pattern of camp group formation from one year to the next. Nevertheless the yaylas used by the tribe are constant so that any facilities built in the course of one summer will be used by other Alikanli the following year (Be§ikgi 1969). Notwithstanding a moderate turnover in camp membership it follows that, given assured access to regular seasonal campsites, a 'settling in ; process may commence. Whether or not such a process leads in the long run to a seasonally occupied village or complete sedentarization depends on other factors - in general, the criteria for selecting a site for permanent
Nomad architecture
Fig. 6.16 Schematic diagram of long-term changes in dwelling form in response to the interplay between sedentism and nomadization. Nomadic groups mentioned in the text are characterized according to their seasonal variations in dwelling types. The dashed line represents hypothetical changes in seasonal dwellings such as might be experienced by a group undergoing nomadization followed by sedentarization.
111
settlement differ fundamentally from those involved in the selection of campsites. But whether or not seasonal campsites are slowly transformed into villages, the net result of an extended settling-in process would be a regional shift towards a more extensive use of fixtures and durables. Just such a process may be occurring at Sanaydin Yayla (see Chapter 9), a summer camp in the Taurus Mountains, where some families are abandoning tents and moving into stone huts (i.e. from type 3 to type 5 dwellings). The retention of the zemga type of composite dwelling in Hulailan could indicate a lack of commitment to a permanent base, particularly in view of the unstable environmental and political conditions in the region. The combination of dwelling types employed is influenced not only by the seasonal cycle, but also by the relative position in a much longeramplitude cycle of nomadization and sedentarization. Seasonal variations in dwelling type employed by ten nomadic groups in Anatolia and Iran are shown in Fig. 6.16. These groups may be seen as occupying locations in a succession from more sedentary to nomadic patterns, shifting up or down the spectrum of dwelling types in response to changes in extent of mobility and security of tenure. For example a group may begin by alternating between a type 6 hut and a type 4 walled tentsite. With declining security of access to both winter and summer pasture, a shift may occur to a type 2 tentsite in summer and a type 4 tentsite in winter. Such a pattern might well be maintained during a period of shifting migration tracks and inter-tribal rivalries of the kind outlined in Chapter 4. As stable conditions return, successively more permanent forms are adopted until
CO
t
o
^
^
t
m
m
s
5
m
c
J
r
^
to
used In transit t2 s
kula
Juila t 3
sp* t4
11
sp-a
\w
w
w W
JL 11
TYPE 7 H
ill
w
w w
w w
t4
Nomads in archaeology
112
finally the group settles into a village of type 7 houses, with some families employing a type 4 tent in summer. Few insights into these processes are likely to be forthcoming from ethnographic studies of contemporary nomads as the processes operate over too long a time scale to be detected within the span of a single generation. Greater insight can be obtained by comparative studies of historical; ethnohistorical and ethnographic accounts in varying environmental and sociopolitical contexts. In the final analysis it is the archaeological study of settlements over successive occupations that will probably provide the best documentation of such processes over a suitable time scale (see Plog 1974; p.44).
7 ALI'S CAMP: A NOMAD HOUSEHOLD CAMPSITE
Along Turkey's southern Mediterranean coast, in the area known to classicists as Rough Cilicia (Fig. 7.1), small numbers of nomadic pastoralists continue to set up camp for the winter. During two successive autumns 1980 and 1981-1 was able to observe the comings and goings of a single household and carry out a detailed analysis of their abandoned 1980-81 campsite. Fig. 7-1 Regional map of the southern coast of Turkey around Silifke with locations of sites mapped or visited.
SILIFKE REGION: SOUTHERN
A campsite • village — highway — secondary road track
ANATOLIA
0
113
Nomads in archaeology
114
The region The small village of Karakegi, which served as a base for much of my fieldwork, lies strewn among citrus groves, overshadowed by the battlements of a Crusader castle and the ruins of what must have been a sizeable city in Hellenistic, Roman and Byzantine times. Limestone strata rising abruptly from the sea climb tier upon tier towards the Taurus Mountains in a karst landscape heavily dissected by seasonal watercourses, faulted by massive sink holes up to 200 metres deep and drained by subterranean systems of swallow holes. Unable to retain surface water, this stony landscape has little agricultural potential except where small patches of alluvium or swampland occur at the mouths of streams. Straggly stands of winter wheat and heavily irrigated citrus orchards clothe the lower slopes. Precipitation occurs between December and May along with moderate winter temperatures, and snowfalls are almost unknown, while summer temperatures hover around 35°C. In the Hellenistic period the region lay at the boundary of the kingdom of Olba (see Mackay 1968) whose tall stone-built watchtowers can still be seen in the hills overlooking the sea, spaced at roughly 2 kilometre intervals. The capital of Olba, now Uzuncaburg, was connected to the coast by a paved road running into the hills behind Karakegi and this road, now used as a migration track by nomads, continues on to the Anatolian Plateau. In Roman times the city and port of Korycus served the Anatolian hinterland where large estates produced wool and grain for export (see Mitchell 1974, p.242). That the region once enjoyed a far higher population density is confirmed by the ubiquitous ruins of the Classical, Byzantine and Crusader periods. For approximately 20 kilometres along the coast building foundations and tottering walls, stone quarries, rock-cut tombs and limestone sarcophagi crowd the narrow modern highway. A highly sophisticated aqueduct system, drawing on permanent mountain streams some 30 kilometres to the northeast, terminates in a massively constructed stone cistern within whose interior boundary lines and goal posts have recently been set up. From the sheer walls of a nearby valley about a dozen basrelief figures of local notables, carved into the solid rock and each about 4 metres high, survey the current desolation. It was near here that Mark Anthony and Cleopatra sojourned prior to the battle of Actium, and the apostle Paul was born. The castle, originally established by the kings of Armenia, was refurbished as a major stronghold of Frederick Barbarossa (Bulucu etal 1977). With the decline of the Byzantine Empire and the establishment of Turkic emirates on the Anatolian Plateau much of the region reverted to malarial swamp, dangerous to human health in summer, and inhabited only by wintering Turkmen nomads whose herds of goats were able to survive off the rough grazing. According to Siimer (1980, pp.153, 175) the Taurus foothills from Silifki to Adana were then dominated by a tribal confederacy known as Varsaklar, whose name survives to this day among smaller tribal segments to the west. Permanent settlement along the coast became possible during the twentieth century with the eradication of malaria, and the present population consists largely of sedentarized
A nomad household campsite
115
nomads (Aswad 1971, p.32; Bates 1973, ch.VTI; Eberhard 1953a, pp.40-1). Even today many villagers with strong pastoral traditions continue to migrate to yaylas or summer quarters (Tuncdilek 1962).
Past nomad populations Scattered among the ruins can be found the remains of numerous longabandoned campsites, readily identified by the villagers as eski Ttirkmenler (ancient Turkmen). Nineteenth-century travellers have described such camps in other parts of the southern Anatolian coast: Many of them [nomads] adopt a semi-sedentary life, and dwell in huts built of rough stone, with walls three or four feet high, and a round thatch roof without central support. (Bent 1891, p.270) During this [winter] season many of these nomads adopt a semisedentary life among ancient Greek ruins, in huts built of reeds or wicker work or in their own black tents. (Garnettn.d.;p.l99) These foundations were remarkably standardized, with stone rubble walls built to a height of about a metre and approximately 4 metres on a side. A low stone platform (yukluk) ran around two or three sides, though in many cases this was concealed beneath earth and rubble. The entrance was usually located in the southwest corner with a recessed hearth (ocak) to the east and a further recess for food storage in the southeast corner. A stout wooden pole up to 4 metres in length, used as a roof support, was frequently found lying nearby (Fig. 7.2). The ancient ruins act as a magnet for such temporary settlements. The rubble provides ample building material, and dressed limestone blocks sometimes containing inscriptions or motifs - have found their way into the walls or been used as door jambs. A few well-placed clumps of thornbush could easily convert a Byzantine courtyard into a corral. According to informants some of the Byzantine water cisterns remained in use until well into the present century when careless villagers began throwing rocks in, destroying the sealed bottoms. In the mountainous regions inland from Karakeqi I observed tentsites closely conforming to the above pattern though lacking the encircling stone walls. Those which were still occupied were surmounted by tents supported by heavy cross-poles resting on two upright posts, and in one case the tent was entered through a rickety wooden gate. These were the summer and transit camps of people who occupied the kind of dwelling described above during winter. Around Karakeqi the dwellings occurred in groups of from three to seven. Other observers of southern Anatolian nomads (Bates 1973; Johnson 1969) have noted that winter camps tend to be smaller and more dispersed than in summer and Bates cites the scarcity of winter grazing as the main contributory factor. Altogether I located 34 abandoned dwellings, grouped in eight main clusters, within an area of approximately one square kilometre though, as this was not the result of a systematic survey,
116
Nomads in archaeology
Fig. 7-2 Plans of abandoned campsites around Karakegi, southern Turkey. The tentsites were often found in association with Roman or Byzantine ruins, rubble from which had been incorporated into the walls. A large wooden pole for supporting tentcloth or temporary roofing was frequently found nearby.
3
0zBS ancitnt --;/iarzophaQM
KEY ancient wall
m htarth
dressed block rubble wall 0
1
2
3m
A nomad household campsite
117
many more may have been missed. Any estimate of settlement density is however confounded by the probability that not all of these structures were in use at any one time. Two tribes - or tribal segments - are reported as wintering in the region of Karakeqi during the 1920s and 30s. The Aya§, consisting of around 100 tents with a total population of 750, clustered around the Silifki-Mersin road, dwelling half in tents, half in huts, and practising pastoralism, agriculture and bee-keeping (Yalman 1977, pp.194-8). They migrated to Bulgar Dag in the Taurus Mountains in summer. Their livestock holdings are put at 3,800, giving an average of 38 per household, far fewer than would be expected of fully nomadic pastoralists. Indeed the contemporary village of Aya§, just up the coast from Karakegi, is almost certainly the result of a settling-in process. The Boynuinceli, consisting of 200 to 300 tents, wintered to the west of Karakeqi and moved to summer quarters around Magara (see Fig. 7.1). They are reported as owning 21,130 livestock, yielding a mean of about 70 per household, which is far more consistent with a pastoral and nomadic bias (ibid., pp.206-11).
A case study Ali was born in a mountain village about 50 kilometres west of Karakegi. The family, consisting of Ali, his wife and seven children (out of ten births) had been nomadic for nearly ten years. The eldest son, in his early twenties, had completed military service and returned to take over the main task of herding. The youngest child, a boy, was born in 1980. Those children of school age attended the local school during the autumn and winter months and this was itself a factor to be considered in the scheduling of migrations. Nomadism for Ali was a rational choice made in response to a critical lack of grazing in the vicinity of his home village and was undertaken with the explicit intention of improving his economic position. Nomadism was seen as a means to an end, the strategy being to accumulate as much pastoral capital as possible and sell out at the top of the market in a year when animal prices were high. At the time of fieldwork animal prices had been falling steadily for two successive years, enabling pastoralists to buy up additional stock in the hope of better times ahead. Three distinct herds were maintained. The 'commercial herd ; consisted of 20 sheep and an equal number of goats, mainly angora goats raised for their hair. Most of these had been recently purchased and were being brought up to prime condition before being resold. The second, a subsistence herd, contained goats only and was divided into a core of around 100 breeders and 5 studs, together with 90 young or juveniles destined for the meat market. About 80 young were born in February each year and, according to Ali, all female kids were sold immediately and the males kept for a year. These were joined by a further 50 males purchased from other pastoralists to make up a juvenile herd or toklu of around 90 animals, all sold on the meat market as yearlings in February. Such a system of herd splitting appears to be common among nomads in southern Anatolia (see Bates 1973, ch.VI; Roux 1970).
Nomads in archaeology
118
Although not able to read or write, Ali displayed a keen and wellinformed interest in both local and international market trends and the micro-economics of herding. In response to my queries about the annual turnover of stock, births, sales and prices per head, Ali reached into an elaborately embroidered camel bag and produced a small electronic calculator on which he proceeded to knock out some rough estimates. Other animals owned included one milch cow, two donkeys, a dog, a cat and a few chickens. The household practised no cultivation but made use of wheat stubble in the rented fields for fodder. Summer is spent under the tent in the yayla at Azen, high in the Taurus Mountains on the ancient route to Karaman (see Fig. 7.1). During the autumn migration Ali follows this ancient trail to the village of Magara, a traditional crossroads of pastoral migrations, and then on to the coast at Karakegi. The herds are supervised by the older sons who move them slowly over roadside grazing, while the wife and younger children are transported by hired tractor and trailer together with household belongings. Camels are still employed by some nomads, but the majority have converted to motorized transport which has had significant social effects, particularly in enhancing the entrepreneurial role of the household head (see Chatty 1980; Bates 1980). Ali rides ahead on an ancient motorbike to enter rental negotiations before the arrival of the herds. In October 1980 I found Ali installed at campsite KK-1. The site and pasture were rented from a local landowner for 25,000 TL (about 100 pounds sterling). Ali had reached an agreement with two of the local landowners by which he would rent pasture from each in alternate years. This suited the landowners, who not only benefited from the income but also from the deposition of animal droppings on their stony fields. The dwelling had been substantially renovated and a stout corral constructed adjacent to it. In addition, part of a large nearby Byzantine complex was used as a makeshift animal pen and a smaller corral was being put together from building rubble. In the course of this an existing tentsite, part of an abandoned camp of six tents, was almost completely obliterated. Ali was well pleased with this location and outlined plans to build another corral nearby for use in this and subsequent seasons. He was joined by a relative from a nearby town who was accommodated for a week in a small canvas army tent kept by Ali for use during migration. In June of 1981 I again visited the area and made a pilot study of this campsite while the family was absent. During August and September I carried out a thorough surface recording of the 1980 winter campsite, and in the first week of October, while visiting a tent about 6 kilometres inland along the ancient road, I encountered Ali's eldest son who was grazing the herd nearby. Ali himself turned up in the teahouse at Karakeqi on the same day and set up a temporary camp not far from the previous year's campsite. The family slept in the small canvas tent, though all their belongings were arranged under a nearby tree in much the same manner as that observed in a typical tent or hut. The main subsistence herd and toklu were still being kept in the hills above Karakegi under the supervision of the eldest sons, and Ali explained that he had yet to conclude negotiations with the landowners. He complained that the landowner
A nomad household campsite
119
whose fields he had hoped to rent was asking too high a price and that he was considering staying on the present site and building a new set of stone foundations from scratch. This situation had not altered when I left Karakegi a few days later. Some weeks later, during a brief stopover, I learned that Ali had shifted to a new campsite a few kilometres inland where grazing fees were not so high. Apparently negotiations with the landowners in Karakegi had broken down.
Instability and opportunism in southern Anatolian nomadism This case illustrates a number of points about nomadism already touched on in Chapter 4. First, insecurity in access to grazing results in a highly unstable settlement system with little consistency from one year to the next, though this need not necessarily prevent considerable investment in the construction of fixtures, given that the camps are used continuously for five to six months of the year. The large number of stone foundations may be explained if they are considered as fixtures, redeemable for a season and discarded in the following year depending on the circumstances. Like the relatively unmodified tentsite platforms encountered in summer camps, they are not owned by any particular household. However it is possible that in previous centuries, when sedentary occupation was much sparser on the coast, wintering nomads had greater security of tenure over pastures. Second, the case study highlights the highly opportunistic nature of nomadism in this region, both in terms of short-term or tactical responses to variable access to resources and longer-term economic strategies. A number of settlers around Karakeqi owe their current affluence to a spell of nomadism undertaken to raise sufficient capital to make a start in citrus orchards or small business. A move from pastoralism into commerce is quite a common career pattern among Yoriiks (Bates 1973, p.26). More seems to be involved here than the recent impact of capitalist market forces for this kind of attitude is hardly new. Nomadism in the nineteenth century was by no means a pristine condition in which independent tribal units engaged in pastoral production for subsistence as has been claimed by some (e.g. Hole 1978). Evidence from nineteenth-century observers suggests a complex chain of transactions from the producer, through the middle man or agent, to distant markets, as well as the presence of forms of pastoral tenure and contract herding: Often a well-to-do citizen of some town furnishes a body of Yourouks with a flock by contract, the Yourouk to provide so many okes of milk, cheese and butter etc., whilst the tribes get what milk is over, the hair etc. and the contractor agrees also to keep up theflockif by cnance it diminishes. In this way the Yourouks often amass flocks of their own, and in time pay off tne lender. (Bent 1891, p.272)
Nomads in archaeology
120
Finally, Ali;s continuing affiliation with the village in which he was born in no way contradicts his nomadic status. Although contacts with the village are maintained (it is not far west of his annual migration route) he has not resided there for a decade and his younger children have never seen the village. However his summer camp is located within the yayla of this home village. This was found to be a common pattern among tent dwellers in the region who also claimed village affiliations but denied residence there. Those referred to were invariably mountain villages with strong pastoral traditions and an established pattern of migration to summer yay las (Tuncdilek 1962, 1963). The shift to a migratory cycle involving absence from the village in both winter and summer appears to be the outcome of a local shortage of winter grazing combined with economic conditions favouring the accumulation of livestock. In accordance with the model of alternating cycles of nomadization and sedentarization outlined in Chapter 4, it is suggested that under favourable conditions migratory groups tend to aggregate in villages at some point on their migration track. Seasonal migrations continue, albeit tied to the village base. Increased populations and pressure on grazing land, combined with market forces, will frequently induce the adoption of more intensive agriculture at the expense of pastoral production. But, by the same token, economic incentives favouring the expansion of pastoral production will induce sections of the village population to resume an untied migratory orbit, with the option of moving into relatively unexploited areas. Specific instances of re-nomadization are known. A group to the east of Karakegi, the Kulfalli Oymagi, are reported to have settled in 1869 but returned to nomadism soon after (Yalman 1977, p.215). This line of reasoning leads to the apparent paradox that the village may constitute a vital component in long-term migratory trends as well as short-term schedules. Such long-term processes may not be obvious to the people involved nor easily discerned using standard ethnographic techniques (e.g. Eren 1980). Nevertheless the gradual eastward drift of Yoriiks along the mountains and coastline of southern Anatolia in the course of the last century (Aswad 1971; Bates 1973, pp.36-7, 197) does seem to have left a partial record in the distribution of tribal and village names. The processes of tribal segmentation and territorial expansion discussed earlier (Chapter 4) are clearly discernable: The various places through which the Yoruk spread later will be shown to coincide with the formation of named descent groups' (Bates 1973, p.36). Moreover such tribal designations are also frequently applied to the villages in which such named descent groups settle. Tribal names in Turkey are generally composed of a root followed by the suffix -// which implies 'of or 'belonging to ; (Dogru 1978). Thus a tribe may aquire the name Haciahmetli by way of a founding father named Haci Ahmet. Alternatively some characteristic or nickname such as sari kegi ('yellow goat') may be reified into the tribal name Sankegili. The fact that many village names contain precisely the same structure simply reflects the fact that tribal and village designations are frequently interchangeable. When asked about their tribal affiliation, Yoriiks will frequently show puzzlement but may eventually say, 'Well, we are from the village of Deli Ahmet so we are Deliahmetli'. Nor is it
A nomad household campsite
Fig. 7-3 Historical Yoriik migration routes in southern Turkey. The arrows show the directions of spring/summer migrations of nomadic and semi-nomadic groups. The dotted lines indicate possible connections on the basis of similar tribal names (not to scale).
121
inconceivable that pastoral sections of a home village bearing such a name might, on migrating into new areas, adopt that village name as a tribal designation. The same tribal names often occur at widely separated points and, similarly, villages bearing identical tribal names are often found in widely separated locations. Such a pattern could well have been produced through a sequence of nomadization, segmentation, settlement of some tribal segments in permanent villages, followed by subsequent renomadization of sections of the village population and a recurrence of the cycle. There is certainly some evidence for an eastward trend in migration routes. One nomadic group to the east of Karakegi, the Bah§i§ Yuriikler, have a tradition of having migrated from the region of Ermenek, some 100 kilometres to the west, in 1773 (Yalman 1977, p. 221). Further examples of this kind are suggested by Fig. 7.3, derived from a comprehensive map of the migration routes of Yoriik tribes over the last century. The most common pattern of sedentarization throughout Anatolia is for closely related families, often a tribal section or camp group, to establish a new village to which the name of the tribal unit is then applied. Most villages in Anatolia have been established relatively recently and few can trace their origins back further than 300 years (see Eberhard 1953b). A high turnover in the establishment and abandonment of villages, together with frequent changes of name, complicate the investigation of village and regional history, but may also throw light on past population movements in so far as village names tend to reflect tribal affiliations. Other criteria are also important in the genesis of village names and frequently new tribal segments are named according to the region or village with which they are associated (de Planhol 1959; Dogru 1978). MIGRATION ROUTES & MIGRATORY D R I F T : SOUTHERN ANATOLIA
A
NORTH
spring migration to yayla connections: tribal/village names town
SILIFKE
Nomads in archaeology
122
Under such a system of nomadism, punctuated by phases of settlement, it follows that splinter groups from the original tribal body or villagers returning to a nomadic lifestyle may have carried the name into new areas. The processes investigated here raise important questions about the interaction between nomadism and sedentism. Nomadism may perhaps be viewed as a mechanism for facilitating the spread of sedentary settlement through segmentation and colonization (Tuncdilek 1962; Coon 1965; Eberhard 1953a) under conditions of low population density.
Ali;s camp: site structure at the household level Having placed the Karakeqi camps in their regional context we now turn to the question of their archaeological profile. Ali's camp was a good candidate for such an ethnoarchaeological study due to the presence of substantial features, a detailed surface collection, some observation of on-site activities in the previous season and a good deal of contextual information (see Fig. 7.4). The title of this chapter recalls 'Millie's Camp7, an experimental study carried out in the first flush of the new archaeology (Bonnichsen 1973). The differences in the conclusions drawn between that and similar studies (Longacre and Ayres 1968) and the present one reflect not only different subject matter but also changes in the aspirations of archaeologists as well as the techniques they employ. The objectives of the former were to assess the likelihood of recovering information about the activities carried out on the site, the numbers of individuals present and their Fig. 7-4 Ali's winter campsite (August 1981). The walls are constructed out of loose field stones and rubble from ancient Byzantine ruins. The recording zone is shown to the left of the dwellingcorral complex. The site of a small visitor's tent is indicated at upper left.
A nomad household campsite
123
demographic structure. My own objectives, while including the above, were primarily to search for some underlying pattern of spatial organization on the site. Any such pattern, if present, should be detectable in archaeological contexts of any age, independent of arguments by ethnographic analogy. The dwelling itself was built mostly of rough limestone blocks, together with shaped stone taken from the ruins, with walls about 1.2 metres high, the east wall resting on the foundations of a Byzantine structure (see Fig. 7.4). The hearth, deeply recessed into the south wall, was surmounted by a kind of chimney. At the entrance was a wooden gate loosely hinged with wire. The roof rested on a stout pole crossed by wooden joists and covered by polythene plastic and straw. The interior of the occupied dwelling was organized according to a familiar pattern. Cooking and other kitchen utensils were stacked around the hearth while the rear wall and sides contained the family's belongings stored in camel bags draped with kilims (Fig. 7.5). In addition to the more Fig. 7.3 Interior views of Ali's dwelling when occupied (October 1980). (a) Kitchen section showing recessed hearth. Note the careful fitting together of stone blocks. (b) Rear section with covered storage sacks filling the stone platform on three sides. Note the light joists of the roof construction.
Nomads in archaeology
124
traditional items of material culture such as metal cooking pots and convex shield for baking bread there was a wide variety of modern items: cutlery, frying pans, a gas oven and metal gas cannister, a kerosene lamp, aluminium and plastic bowls and dishes, glass bottles, tea glasses and a china teapot. Plastic containers were used for storing water and metal gasoline cans for carrying water to livestock. Wooden packing cases scavenged from the village and tourist resort were used as kindling wood, leaving behind concentrations of wood chips and nails. A summary of these materials is contained in Table 7.1. Table 7.1 Classification of materials recorded in recording zones ofAli's camp (KK-1) Material
Code
Durability
Animal bone
BONE
durable
Pottery
POTT
durable
Glass
GLAS
durable
Hard plastic Polythene plastic Metal
PLAS POLY META
durable perishable durable
Nails Wire
NAIL WIRE
durable durable
Rubber Leather Textile
RUBB LEAT TE)Cr
perishable perishable perishable
Binding Wood
BIND WOOD
perishable perishable
Charcoal, ash
CHAR
durable
Definition All discrete occurrences, complete or fragmentary. With one single exception (teapot handle) all ceramics derive from Roman/Byzantine occupation. Glass sherds scored as 1. Whole vessels scored as 5. All articles and fragments ofrigidplastic. All polythene sheeting and thin, flexible plastic All metal fragments, batteries, metal strip (scored as 1) and whole vessels (scored as 5). Loose nails or those embedded in wood. Wire of various thicknesses. Scored twice if extending over adjacent quadrats. All types. All animal hide and skin products. All discrete occurrences of cloth, fragments of clothing, etc. • All cotton, string, cord and rope. Woodchips, twigs and sticks, recorded individually where possible; major concentrations indicated. Single occurrences and patches recorded.
In terms of behaviour, two main activity sets were identified following the division between domestic and herding concerns. Food preparation and consumption, cleaning, and domestic refuse disposal were confined to the dwelling itself or immediately outside. Activities associated with herding were more extensive and may be divided roughly into three concentric zones - inner corrals, butchery zones and outer corrals. Table 7.5, summarizing the faunal elements recorded in the butchery area, indicates the presence of a small number of individuals in the assemblage. The surface collection was carried out in 1 metre quadrats across the discard and butchery areas to the west of the dwelling (see Fig. 7.4). Apart from the ubiquitious Roman/Byzantine potsherds, the only ceramic item recovered was a small teapot handle (cf. Watson's 1979 Duzary campsite). The functionally equivalent material, glass, was well represented and strongly clustered in the form of sherd scatters or as whole bottles. A glance at Fig. 7.6 is sufficient to show that, while there is differential
A nomad household campsite
Fig. 7-6 Distribution of items in surface recording of Ali's camp, (a) Stone; (b) Potsherds; (c) Bone; (d) Glass sherds; (e) Plastic scrap; (f) Loose nails; (g) Metal scrap; (h) Wire; (i) Polythene scrap; (j) Leather scrap; (k) Rubber scrap; (1) Textile scraps; (m) Binding (string, cloth); (n) Wood and charcoal.
m
125
Nomads in archaeology
126
patterning in the item distributions; the patterns are generally not strong enough to permit the spatial characterization of the area by visual inspection. Nearest neighbour tests on each distribution and bivariate tests of association between pairs of distributions (Table 7.2) in general failed to add significantly to this picture. Such techniques; depending on a null hypothesis of randomness or random association (see Dacey 1973; Hietala and Stevens 1977; Hodder and Orton 1976; Orton 1982) may inform about the degree of clustering or randomness of a distribution but cannot tell us about the shape of a pattern. For example, it is all very well to know that polythene has a highly clustered distribution (with an r of 0.520), but this does not inform us that there are two main clusters located in the upper and lower part of the upper study zone, information that can be obtained by a glance at the raw distribution in Fig. 7.6i. In view of the difficulty of envisaging more than two spatial patterns at any one time, further analysis of spatial organization on the site required the use of multivariate techniques which are capable of extracting summaries of the multi-way associations between distributions. For this purpose the items were represented as quadrat counts within a 2 metre grid. In conducting such analysis we are looking for two things. First, we Table 7.2a Nearest neighbour statistics for selected materials at Ali's camp, indicating the degree of clustering Distribution Bone Plastic Glass sherds Metal scrap Nails Wire Rubber Polythene scrap Textile scrap
r-Value
Number
0.486 0.613 0.487 0.712 0.383 0.666 0.706 0.520 0.679
89 71 152 111
65 40 33 62
87
The r statistic ranges from 0.0 (maximum clustering) through 1.0 (random distribution) to 1.4 (systematic arrangement)
Table 7.2b Test for segregation between pairs of item distributions at Ali's camp. Scores for the most segregated pairs of items are emphasized. They indicate that bone is segregated from all other distributions Bone Bone Plastic Glass sherds Metal scrap Nails Wire Rubber Polythene scrap Textile scrap
Plastic 0.747 0.430 0.788 0.960 0.928 0.811 0.918 0.920
Glass sherds Metal scrap
0.492 0.230 0.617 0.436 0.288 0.621 0.244
0.433 0.662 0.497 0.456 0.790 0.471
Nails
Wire
Rubber
0.333 0.071 0.219 0.584 0.251
0.293 0.397 0.637 0.595
0.447 0.608 0.260
Polythene scrap Textile scrap
0.509 0.317
The S-statistic ranges from 1.0 (maximum segregation) through 0.0 (random association) to -1.0 (complete association)
0.646
127
A nomad household campsite
Fig. 7- 7 Principal components analysis of spatial distributions at Ali's camp for all materials. A degree of spatial coherence may be detected. (a) Component 1: a general mix of concentrations of materials. (b) Component 2: highlights a marked concentration of polythene plastic scrap in the upper part of the study area, indicating a specialized work area. (c) Comparative test result using randomized quadrats.
need to know the composition of the abstract summaries obtained and to be able to infer their meaning. Second, we are interested in the degree of spatial coherence displayed by our units when grouped according to these dimensions. One such method is that of principal components. Principal component and factor analysis have been widely used in archaeology to identify groups of variables forming tool kits (Binford and Binford 1966), ceramic styles (Frankel 1974) or funerary complexes (Rowlett and Pollnac 1971). The statistical results of principal components analysis for the above distributions are shown in Table 7.3. Where a spatial locus can be assigned to each case, as is common with archaeological data, a data summary can be arrived at by contouring the distribution of component or factor scores for each case. Fig. 77 tells us that Principal Component 1, involving textiles,
component. 1
Table 7.3a Results of principal components analysis ofAli's camp (KK-1) - correlation coefficients for durables and perishables Bone Pottery Glass Plastic Polythene Metal Nail Wire Rubber Leather Textile Binding
0.1910
0.0045 -0.1695 -0.1376
-0.2349 -0.1963
-0.2529 -0.0471 -0.1481
-0.3000 -0.0894
Pottery
Glass
0.0488 -0.0119 -0.1222 -0.0099 -0.0732 0.2081 -0.0050 -0.0058 0.0149 0.1585
-0.0047 -0.1420 -0.0130 0.0603 0.0317 -0.0670 -0.0755 0.441 -0.0613
Higher values are emphasized in bold text.
Plastic Polythene
Metal
Nails
0.4068 0.3189
0.3917
Wire
Rubber
Leather
Textile
0.0687 0.0624 0.2966 0.5013
0.3213 0.1738 0.1507
0.2840 0.1518
0.5095
0.0567
0.3224 -0.0500
0.0191 0.0090 0.0324 0.0120 0.1190 0.0199 0.0444 -0.0917 0.2439 0.0667 0.2332 -0.1074
0.0981 0.1836 0.1833 0.3024
0.1621 0.2769 0.0935 0.0849
Nomads in archaeology
128
Table 7.3b Results of principal components analysis ofAli's camp (KK-1) - component loadings on variables Durables (6 variables)
All recent materials (11 variables)
All materials (12 variables)
Principal components
Principal components
1
2
Principal components
1
2
3
4
3
1
2
Bone Pottery Glass Plastic Polythene Metal Nails Wire Rubber Leather Textile Binding
-0.227 0.004 0.000 0.163 0.006 0.404 0.277 0.432 0.129 0.205 0.434 0.458
0.171 0.579 0.095 -0.015 -0.128 -0.003 -0.028 0.137 -0.017 -0.054 0.072 0.116
-0.025 -0.002 -0.029 0.116 0.353 -0.024 -0.253 0.000 -0.110 -0.215 0.067 0.122
0.112 0.047 -0.148 0.009 0.000 -0.069 -0.126 -0.080 0.351 0.191 0.027 0.039
-0.234
-0.017
-0.003
-0.309
-0.024
0.000 0.165 0.007 0.406 0.281 0.426 0.134 0.208 0.436 0.456
-0.015 0.106 0.309 -0.023 -0.059 0.002 0.092 -0.243 0.071 0.141
-0.339 0.022 0.074 -0.034 -0.048 -0.156 -0.343 0.198 0.002 0.000
0.003 0.144
0.178 -0.584
0.580 0.492 0.462
-0.041 0.147 0.126
Eigenvalue % variance
20.74 220.83
10.38 110.48
10.28 100.67
10.20 90.97
20.74 240.89
10.28 110.64
10.22 110.11
10.99 330.16
10.10 180.30
Binding Textile Wire Metal
Pottery
Polythene
Rubber
Binding Textile Wire Metal
Polythene
Rubber Glass
Metal Nails Wire
Plastic
Higher values are emphasized in bold text.
binding, wire and metal, is distributed as an arc occupying the upper study zone. While the distribution is spatially coherent, the functional significance of such a complex of materials is far from obvious. Component 2, dominated by ceramics - almost entirely Roman-Byzantine may be disregarded. Component 3, loading positively on polythene and negatively on leather and nails, makes more intuitive sense. Essentially it tells us that where the former occurs the latter tends to be absent, and vice versa. However principal components analysis has not enlightened us a great deal about the general spatial organization of the site. While we are not now trying to assimilate 12 distributions at the same time, we are still dealing with three separate components one at a time. Cluster analysis (Hodson 1970; Blashfield and Aldenderfer 1978) has the advantage over principal components analysis of producing a single solution and more informative spatial plots, though the meaning of the summaries or clusters may be less clear. In a variation of a technique pioneered by Robert Whallon (1984), the 2 metre quadrats were subjected to cluster analysis by Wards method and average linkage (see Cribb 1983, 1984a). It should be pointed out that there is no perfect clustering method and that, unlike principal components analysis, different results may be obtained via different techniques. The results of the cluster analysis are shown in Fig. 7.8. Its details do not concern us here. What is important is that, for a number of different clustering techniques as well as for principal components analysis, the same basic patterns emerged. There occurred a central zone of general discard located immediately to the west of the dwelling, flanked by
A nomad household campsite
129
Fig. 7-8 Spatial cluster analysis of distributions at Ali's camp by 2 metre quadrats. (a) Spatial representation of these results. The most noticeable features to emerge are linear patterns or arcs across the central study zone. (b) A similar cluster map for durables only. Linear patterns are also prominent.
5
5
T\5 all recent materials
5 5 \ 1
S
2
5) 1 1
durables
peripheral zones in which one material predominated. Even when perishable items were deleted from the analysis the trend was not significantly altered. This juxtaposition of generalized and specialized zones recalls Yellen's (1977) analysis of !Kung bushman camps in the Kalahari. But in addition to the distinction between 'nuclear7 and 'special activity' areas there emerged clear evidence of zonation within the general discard area itself. Moreover the zonation appeared to follow a distinct pattern of concentric arcs located around the external hearth (Fig. 7.8c). The arcs, corresponding to the four strongest clusters, were quite distinct from one another in terms of their proportions of key items (Table 7.4) and may be summarized as follows. Group 7 Located at the northern rim of the zone of general discard, this group was dominated by metal, together with considerable quantities of plastic and Table 7.4a Chi-sauare tests for significant difference between cluster groups on major constituent materials at Ali's camp (KK-1) - groups 1 and 4
Group 1
% Group 4 (Exp) (Obs)
Plastic
Metal
Total
16 20 7.8 26
64 80
80
i = 53.1 1 degree of freedom Ho of no significant difference rejected at the 0.001 level
31.2
13
39
Nomads in archaeology
Table 7.4b
130
Chi-sauare tests for significant difference between cluster groups on major constituent materials at Ali's camp (KK-1) - groups 3 and 4 Plastic
Metal
Total
75
19
16
110
68.2 30.0
17.3
14.5
7.6 26
6.4 13
Glass Group 3
% Group 4 (Exp) (Obs)
5
44
1 = 72.1 2 degrees of freedom Ho of no significant difference rejected at the .001 level
small amounts of other materials. The nails and metal strip were the byproduct of the reduction of packing cases for kindling wood. The distribution of Group 1 quadrats exhibited a high degree of spatial coherence. Group 2 This group also contained a high proportion of metal, particularly nails and wire; together with glass sherds. It was less spatially coherent than any of the other three groups, but the four constituent quadrats formed an almost perfect arc around the external hearth. Group 3 This was an almost continuous strip composed of a mixture of glass sherds, plastic, metal and cloth, less spatially coherent towards the eastern end of the discard zone and with an outlier in the lower recording zone. Group 4 This was another continuous strip, contiguous with Group 3, and characterized by a mixture of plastic, metal and cloth, differing from Group 1 in the relative proportions of each. A number of hypotheses were advanced to account for these patterns. In view of the relatively small number of quadrats involved and the confined space within which they occur, it is possible that much of this apparent organization was the result of random processes. A major disadvantage of the unconstrained clustering technique is the difficulty of testing the significance of whatever spatial patterns emerge, a problem stemming directly from its main advantage - the lack of a priori assumptions (Whallon 1984). The development of tests for spatial autocorrelation in the distribution of clusters, components or discriminant functions should be high on the list of priorities in spatial archaeology. A second hypothesis, based on the behavioural model of Schiffer (1976), was that the bands or arcs may be the result of direct discard of materials deposited in situ. A third possibility is that they represent 'toss zones' (Binford 1978a) from a central area of activity around the hearth. There may be elements of both present but, whatever the events giving rise to such observable patterning, we are led to consider a most intriguing proposition: that human occupation of such impermanent sites normally results in a patterned distribution of debris which, though difficult
A nomad household campsite
Fig. 7- 9 Spatial patterning of distributions, as summarized by spatial cluster analysis, in relation to structures and features of Ali's camp. The linear zones of concentrated material detected in the cluster analysis form concentric arcs around the small external hearth within a roughly 10 metre by 20 metre activity zone which includes the habitation. This is flanked by special activity zones (polythene processing, butchery).
cutting rubber cutting up ^ polythene sheeting
131
O/"\ ,'*y \ bone refuse \y*s /
*
K
/\
KEY ash e charcoal
<&& wood scrap ,** activities progressive reduction of kindling wood limits of nuclear area
anaent sarcophagus
chicken coop
£
butchery area (threshing 10m
ACTIVITIES
to interpret in functional terms, reflects the structural organization of a household unit. While the effects of pre-depositional and postdepositional events may be impossible to unravel, their interaction bears the unmistakeable imprint of organized life-space. This impression was reinforced when the study zone was considered as part of a wider context. The distribution of metal and glass containers continued as arcs outside the immediate study zone. A final summary of Ali's camp appears in Fig. 7.9, illustrating the area of general discard surrounded by specialized activity zones involving butchery and discard of animal bone (see Table 7.5) and processing of polythene sheeting. The central zone of discard covered an area of approximately 10 metres by 10 metres which, together with the dwelling, formed a rectangular unit measuring roughly 20 metres by 10 metres. This domestic complex, equivalent in some ways to Yellen's (1977) nuclear area, represents the core area of household organization and day-to-day consumption, surrounded by Table 7.5 Faunal elements (capra) from recording zone at Ali's camp (KK-1) Element Mandible Humerus Radius Ulna Metacarpal Phalange Femur Tibia Fibula
Frequency 1 (+ fragments)
2 1 1 2 0
I(dist.),2(pro42(frag.) 1 (dist.), 2 (piox.), 1 (frag.)
0
Element Metatarsal Calcaneum Astragalus Scapula
Rib Vertebra Pelvis Longbone fragments Unidentified
Frequency
0 2
1
3 + 4 fragments
8 33 (mostly fragments)
2 15 14
Nomads in archaeology
132
specialized areas of productive activity. The domestic complex will be analysed in more detail in Chapter 9 where observations were made across a number of household units.
Conclusion The detailed study of a single household unit has given us some idea of the factors and constraints governing the formation of tentsites and their associated domestic space. Both here and in the previous chapter certain regularities have been observed in the size; orientation and division of internal space as well as the dimensions of living zones, discard zones and the location of specialized work zones. These findings support the approach taken by Yellen (1977) in defining organizational units such as the 'nuclear area', which tell us not so much about the nature of activities as about the rules and constraints which circumscribe human activity in a particular social setting. A similar set of constraints appears to operate at the next level in the nomad residence hierarchy, the campsite, to which we now turn.
8 THE STRUCTURE AND LOCATION OF NOMAD SETTLEMENTS The tribe was camping in Hassan Agha's estates near a clump of mulberry trees. They had pitched their long black tents in a neat, orderly file, and this did not escape Hassan Agha's notice. Just as if they were founding a new village, he said to himself with growing anxiety. (From The Legend of the Thousand Bulls by Yashar Kemal)
We have seen that the low archaeological profile of nomad sites cannot be attributed to the absence of material remains alone. But while nomad material culture and architecture may be comparable in variety and durability to that of more settled communities, the distribution of these remains occurs over a much wider range and in greatly reduced densities. Moreover the manner in which fluid patterns of residence group formation and fortuitous shifts in migration tracks are translated into archaeological assemblages will vary according to time and place. It is possible nevertheless to make some broad generalizations about what sets a nomad settlement system apart from a sedentary one and how nomad campsites differ fundamentally in their internal structure from sedentary communities.
Nomad settlement systems The term 'settlement system7 has been carefully chosen as against 'settlement pattern7 in order to distinguish a system of controlling variables (Flannery 1976b, pp.162-3) from a physical pattern of settlement (Chang (ed.) 1968; Trigger 1968). The latter approach applied to nomad camps would yield a picture of segmented settlement units located in relation to each other according to patterns of kinship, tribal affiliation and economic cooperation. What we are interested in is not so much the form itself as the whole system of cultural preferences and logistical constraints which go into its making - the set of rules which governs the location and formation of campsites. The same set of rules may well produce different patterns in different environments. The system consists of variables which are ranked in a kind of hierarchy - some may override others, 133
Nomads in archaeology
134
some may come into play only if some other factor is present. The major factors affecting the location of nomad campsites are now discussed and an attempt will be made to assess their relative importance.
Altitude In mountainous areas of the Near East summer pastures are sustained by the melt-water from heavy snow drifts and rarely occur below l ;500 metres. There appear to be no definable upper limits short of the permanent snowline. In Anatolia the major plains on the central plateau (1,000-1,500m) are rarely used by nomads; though short-range village transhumance may occur (Miiller 1938, p.47; Alagoz 1938; Frodin 194344, pp.232-4). Well-known and named yayla areas (Tuncdilek 1962,1963) follow the parallel chains of the Taurus Mountains and the belt of lofty volcanic cones lying to their immediate north. Yoriik nomads in western Anatolia are found in summer at altitudes of from 1,500 to 1,800 metres (Johnson 1969, p.29) and slightly higher further east, between 1,500 and 2,000 metres (Bates 1973, fig.2) or 1,700 and 3,000 metres (Frodin 194344) depending on the heights of the main local ranges. Minimum altitudes increase towards the east, following a rise in the elevation of the plateau, until east of Malatya where yaylas are rarely below 2,000 metres and even the upper slopes of Nemrut Dag, Siiphan Dag and Agri Dag (Mt Ararat) are grazed in late summer (Frodin 1943-44; Be§ikgi 1969; Imhof et al. 1970). Altitudes of summer pastures (sarhad) in Iran also tend to be high, particularly towards the south where decreasing latitude has the effect of raising the minimum altitude (see Barth 1959, 1961; Sunderland 1968). Winter pastures (Turkish ki$lak, Persian garmsir) are generally located at the lowest altitude possible along the coast of western Anatolia, the Mesopotamian steppe in eastern Anatolia and Iraq, and the Zagros foothills in southwest Iran. The altitude of Basseri winter quarters is given as 2,000 to 3,000 feet (Barth 1961, p.4) or 700 to 1,000 metres. Some nomads wintering at higher altitudes, in order to avoid lengthy and dangerous migrations to winter pastures or as a result of conflicts, construct huts or houses or some kind of composite dwelling as refuge from harsh winter conditions (see Watson 1979, ch.lO; Zagerell 1975; Edelberg 1966-67). Such a pattern, also reported among Kurdish groups at the turn of the century (Sykes 1908), need not imply sedentarization and appears to be a long-standing practice. The stone-built settlement of Shirdasht (Watson 1979, ch.8) is a good example, regarded by its inhabitants as little more than a winter refuge. Although it is sometimes possible for winter and summer pastures to occur within a short distance of each other within a single valley system (Watson 1979; Edelberg 1966-67; Fazel 1973), the overall effect of altitude on site location is to segregate winter and summer camps into two parallel and mutually exclusive bands of territory - discontinuous in places following the major mountain chains and corresponding to the zones of greatest climatic extremes (Hole 1978, p. 165). Here will be found the most prolonged occupations while spring, autumn and transit camps established in the course of migration between these zones are generally of shorter duration and are less substantial.
Structure and location of settlements
135
Access to grazing
Within these broad seasonal zones the major consideration is access to specific tracts of grazing land. The density of campsites will depend on the overall amount and the pattern of distribution of pasture (whether in continuous or dispersed tracts); its quality and the ratio of humans to animals. Winter camps of the Shahsevan in northern Azerbaijan ideally lie at the centre of an 'estate; of about 8 kilometres radius (i.e. 16 kilometres apart) (Tapper 1979b; p.107). Due to the higher quality of summer grazing on the slopes of Mt Savalan in summer quarters, summer camps are much closer together. Among the Basseri and Qashqa'i, summer camps are also closer together than in winter, facilitating a higher level of interaction (Barth 1961, p.25; Marsden 1976, p.14). The location of Alikanli summer camps is highly responsive to variations in the amount and quality of pasture throughout the season - some of the higher pastures may not be used at all if those lower down are not exhausted by September (Be§ikgi 1969; Johnson 1969, p.7). For most nomads access to winter pasture is more of a problem and may impose an upper limit on livestock numbers or increase the dispersal of nomad population. This may result in the under-use of summer pastures as animal numbers become pegged to the carrying capacity during the worst season.
The effects of agriculture and sedentary settlement
As frequently stressed (Bates 1971, 1973; Beck 1980), the adjustment of nomadic pastoralism to systems of sedentary agriculture - at whatever level of integration, whether household, lineage, village, tribe or region is not a simple ecological balance but an unstable compromise in which the apparatus of political power may favour one or the other party. However most observations of nomad settlement patterns have been made under conditions of population growth, agricultural expansion and state ascendency during which the burden of compromise has fallen most heavily on the nomad party. Under such conditions campsites may be confined to areas of nonarable land use, particularly valley slopes or valley bottoms above the spring line and the upper limits of irrigation. This is apparent even where pastoralism and cultivation are closely integrated within the same community, as at Sanaydin Yayla, where the camp itself and pastoral facilities and grazing are located above the spring and irrigation system serving cultivated plots occupying the valley floor. Campsites located on the high margins of cultivated fields are extremely common. A typical location of this kind is described by Stark (1933, p.253) in Luristan 'on barren limestone foothills just above the plowed river lands'. Aerial photographs of the same region (Schmidt 1940) show many campsites in such locations. Fig. 8.1, adapted from one such photograph, is typical. About a dozen abandoned campsites ranging from five to 20 tentsites occupy ancient agricultural terraces, possibly dating from the second millennium BC, above the present zone of cultivation surrounding an adjacent village. As the shots were taken during the 1930s at the height of Reza Shah's sedentarization programme, these sites may well have been permanently rather than seasonally abandoned and the village itself may have been a product
Nomads in archaeology
DESERTED CAMPSITES: LURISTAN 0
50 m (approx)
136
J:
A I
slope
Fig. 8.1 Abandoned Luri campsites in the Pusht-iKuh, Zagros Mountains, Iran: 11 campsites are visible. The site plans resemble the zemga (stonewalled winter dwellings partly covered by tentcloth) employed by some Luri pastoralists and cultivators. Note the rectangular layout of campsites nos. 3 and 8, implying the presence of herding units. (Based on aerial photographs taken in 1937 by Schmidt 1940, pl.117.)
of sedentarization. This settlement pattern clearly monitors the decline of agricultural activity in the valley since ancient times. Pastoral pursuits are somewhat less likely to clash with cultivation in high altitude summer pastures. However, during winter, which is often spent in close proximity to intensive agriculture, the situation is more ambivalent. On the one hand, grazing land is scarce and the threat of crop damage is a source of much tension between nomad and cultivator (Bates 1973; Barth 1961). Crop damage can also be a problem for nomads who engage in seasonal cultivation (Spooner 1964, p.63). On the other hand, the interests of both cultivators and pastoralists are often served through the practice of pasturing flocks on fallow stubble fields which are fertilized by their droppings. Particularly in autumn and early winter, tent
Structure and location of settlements
137
camps may be found on the newly harvested fields themselves. Winter campsites are therefore subject to both maximum dispersal and a tendency to cluster in the vicinity of fields or settlements (Schumann 1962, p.30). Both tendencies are especially marked in the case of the Alikanli where small groups of four or five tents regularly attach themselves to villages each winter (Be§ikqi 1969).
Terrain Within broadly defined pastoral zones, and subject to the constraints imposed by other forms of subsistence, specific campsite locations tend to favour human needs or, as Hole (1978, p. 165) suggests, 'most camps are sited to accommodate the needs of humans who are herding animals'. Preferred locations are generally those combining a stretch of relatively flat and unbroken land with shelter in the form of steep adjacent slopes or rock outcrops (Marsden 1976, p.14; Stark 1933, p.255; Hole n.d.). Ravines and dry stream beds are also popular, though the effects of a sudden downpour upstream can be disastrous. In limestone country sinkholes provide shelter (Johnson 1969), while in volcanic areas the broken terrain and precipitous margins of lava flows afford ideal locations for campsites (Lynch 1901, pp.324-5). In the course of fieldwork I frequently came upon tents in steep-sided ravines, in sheltered valleys, or on flat ground in the lee of a steep hillside (see Fig. 8.2). Larger enclosed plains and alluvial valleys in eastern Anatolia frequently contained substantial camps. Another favoured type of location is a flat terrace on a steep hillside. Just such a location was recognized as typical by a nineteenth-century observer in the Taurus ranges south of Malatya: This [Kurdish] encampment occupied one of the usual peculiar sites selected by the mountaineers: a mere shelf on the acclivity of the hill receding far enough to allow the dusky tents being pitched in a line in its remotest part, with only the customary cross poles in front of each, on which the women churn their butter, ana without even a fragment of the camp being visible from below. (Ainsworth 1842 I, p.249) This could easily have been the campsite shown in Figure 8.2a. Travellers7 accounts also describe tent camps located at the head of high, open valleys (Sarre 1896, p. 116; Ains worth 1842 I, p. 116). In fact, although data are not available in quantifiable form, it is clear from these and other travellers7 accounts (e.g. Sheil 1856; Morier 1837; de Bode 1845; Rawlinson 1839) that the kinds of locations favoured as campsites by nomads have not changed appreciably over the last 150 years.
Water and fuel Hole (1978, p.151,1984) plays down the importance of water sources as a factor in campsite location, and Edmonds (1957, p.373) noted 'I was surprised to find how far from water Jaf camps were7. Nor does Barth (1961, p.53) stress the importance of access to water in the location of Basseri camps. However many of these observations were made in the course of migration when plentiful supplies of drinking water are not crucial. While a lake or soak at a distance of a few kilometres may serve as an adequate
Nomads in archaeology
138
Fig. 8.2 Typical campsite locations on narrow shelves of flat land. (a) Two black tents on the edge of an escarpment sheltered by large boulders, Taurus Mountains, southern Turkey. (b) A small campsite on a shelf with nearby spring, Ala Dag, central Turkey.
^
•
•
:
•
•
•
watering point for livestock, campsites which endure for any length of time require a reliable source of drinking water, usually a spring or running stream, close at hand (e.g. Marsden 1976, p. 14). Sources of firewood within easy walking distance of a campsite are required not only to keep the domestic hearths burning but also in the preparation of milk products (Martin 1980, p.24). This can pose problems in heavily deforested areas occupied for lengthy periods. While animal dung (tezek) may sometimes serve as an alternative, nomads are normally too mobile to build up stocks of this material which requires careful stacking and a lengthy drying process.
The interaction of locational factors Rather than ranking locational factors in a strict order of precedence - for instance Hole (1978, p. 151) lists terrain, pasturage and water in descending order - 1 would argue that different factors come into play at different
Structure and location of settlements
139
scales of resolution: regional, local and site-specific. Sets of factors form a nested series in which broad topographic and climatic constraints impose limits within which access to specific grazing tracts further narrows down the range of choice. Of the areas left, those under cultivation will most probably be given a wide berth unless some trade-off of dung for stubble grazing can be worked out with the local peasantry. Given that a group of nomads is able to exploit a particular area, they will then set about siting their camp or camps according to terrain and local resources. In mountainous areas they will make straight for an expanse of flat land beneath an overhanging slope, having first checked that the nearest spring holds a sufficient supply of water. In all likelihood they will find the foundations of earlier tent camps neatly laid out there. In more open country a grove of trees, a ruined building, an archaeological tell or relic of Roman, Byzantine, Sassanian or Seljuk architecture may provide at least the semblance of shelter. On the basis of my observations in Anatolia, the ideal summer campsite would be located at about 2,000 metres on the edge of an extensive tract of grazing land, well above the limits of cultivation, in a sheltered side valley or in the lee to the south of a steep slope or rocky outcrop. It would be within half a kilometre of a spring and within moderate distance of a watering point for animals. While foliage is most probably absent in the immediate vicinity — due to previous camping episodes? - there would be a wooded slope or ravine nearby. While not all of the above factors need come into play in every case, the advantage of such a nested model is that, when linked to a particular piece of archaeological territory, it could form the basis for a survey design geared to detecting surface indications of ancient campsites in optimal locations. Of course the order of precedence of these factors may change in response to external constraints, for example an expansion of cultivation may restrict access to both pastures and campsites. It is fairly certain that while such middle-level factors may vary greatly from one historical period to the next, the lower order criteria of terrain, shelter and aspect are likely to have remained constant.
The internal structure of nomad campsites No one familiar with Near Eastern tent encampments can fail to be impressed by their overall visual impact. In some of the Kurdish camps of eastern Turkey the long black tents drawn up in close formation are reminiscent of battleships deployed in line astern. There is a pronounced military flavour about many campsites and, whatever the condition of the tents themselves and the mixture of apparent chaos and lethargy in camp life, one has a distinct feeling that nothing is arbitrary - no tent has been placed without careful consideration of its relation to other tents, no activity occurs outside a regular and finely tuned schedule. Many features of campsite organization may be attributed to environmental factors, others are more likely to be the result of social factors. Due to the flexibility and vulnerability of the tent as a dwelling form, the environmental constraints of any given campsite are immediately
Nomads in archaeology
140
Hilhil Yatak
F/g. 5.3 Principles of campsite spatial organization. Tents tend to be aligned parallel to each other and to form up in rows. (Not to scale.) (a) A summer camp of the Beritan A§iret; eastern Anatolia. Note the alignment and grouping of tents. (After OzbekKoroglu 1980, fig.6.) (b) Spring camp at Duzaray, in the Zagros Mountains of western Iran. (After Watson 1979, fig.9.1.) SKETCH
MAP OF SHWDASHT CAMP
O Tenis
A
North
reflected in the physical arrangement of tents. Tents are frequently oriented along an east-west axis so that the sun passes over from one end to the other. They may consistently face in the same direction (aspect), to the south in Baharvand and Qashqa;i summer camps (Hole n.d.; Marsden 1976, p.26). Wind direction may interfere with this preferred alignment, and prevailing southerly winds seem to be the main factor causing Beritanli and Alikanli summer tents to face north (see Fig. 8.3). Sometimes
Structure and location of settlements
141
Fig. 8.3 (c) Plan of a settlement in Baluchistan consisting of pise foundations roofed with either palm leaves or black tents. Note the tendency towards clustering and consistent east-west orientation. (After Audouze and Jarrige 1981, fig.2.)
topography will tend to produce such a pattern of consistent orientation as tentsites are often terraced onto slopes through cutting and filling, in which case they will all face downhill and follow the contours of a gentle slope. The effect of slope will normally prevail over other tendencies towards alignment, bending rows of tents around in a sweeping curve as the direction of slope changes. It has been noted that the formal arrangement of Qashqa'i tents in straight rows applies only 'unless the terrain prohibits this7 (Marsden 1976, p.26). Tents are often aligned in parallel lines, sometimes constrained by restricted space such as in valley bottoms (Evans 1983), but also on relatively flat and open terrain. A common pattern is a row or rows of tents aligned on level or slightly sloping ground along the base of a steep slope. As Barth (1961, p.42) notes, locations often restrict the possible distribution of tentsites, particularly in the case of small patches of level ground in otherwise hilly terrain. Other examples, such as the four successive seasonal Bakhtiari campsites recorded by Digard (1975; see Fig. 8.4), display no discernible pattern of orientation or alignment. It is not possible to offer any set of hard and fast rules applying to all campsites and there are insufficient quantitative data to determine whether or not particular patterns vary according to tribal affiliation or seasonality. For instance it is reported that the Boyr Ahmad, neighbours of the Qashqa'i, may choose a variety of patterns - in rows at the foot of a steep slope, in circular formation or in apparent disorder (Loffler and Friedl 1967, p. 103). All that is certain is that settlement patterns within campsites are highly responsive to simple mechanical factors such as slope, shelter, aspect and wind direction. All may act together in such a way as to produce a strong pattern of regular alignment in the distribution of tents. Conversely, where such factors tend to counteract each other, we would expect greater variations in tent alignment. It is interesting that camps of circular yurts are often arranged in a
Nomads in archaeology
Fig. 8.4 Variability in campsite organization among Bakhtiari nomads. Plans of four successive Bakhtiari campsites. Note the turnover in camp membership and variations in the location and orientation of individual tents. Households are numbered 1 to 16. Household no.12, that of the camp headman, occupies two tents. (a) Spring camp, mid-May to mid-July. (b) Summer camp, mid-July to mid-September. (c) Autumn camp, early October to end of December. (d) Winter residence, end of December to end of March. Winter dwellings have substantial walls and some are roofed. (After Digard 1975, figs 2-5.)
142
J3
0
19
20 m
y{ Funnel shaped structure for milking«'
Enclosure for horses (autumn)
O
Grain storage (autumn)
Corral for milking
0 S t o n e walls enclosing front of C.jtent (summer)
E n d o s u r e for smal| ani
ft
circular formation. A monastic encampment observed in western Mongolia (Ambolt 1939; p.38) consisted of around 100 yurts in a perfect circle around a large; centrally placed black tent. A similar example is found in Jentsch (1973, p.221). Such an arrangement, though superficially similar to that encountered in some African settlements, cannot be attributed to household organization as has been argued by Flannery (1972) in a discussion of rectangular and circular house plans and also suggested by Hodges (1972).
Structure and location of settlements
143
The formation of camp groups In a pastoral system comprising pastures, herds and people the number of households (tents) that will congregate in a particular camp depends on two primary factors: the numbers of animals that can be carried by grazing tracts within reach of the campsite, and the average number of animals per household. The way in which these factors influence camp group formation is illustrated in Fig. 8.5. The size and placement of campsites will depend on whether grazing is distributed in large continuous tracts or in dispersed patches. Given a particular distribution, the number of tents per camp will vary according to the livestock holdings of the households involved - the more animals per household, the fewer tents. The camp group is therefore a highly flexible and transient unit able to mediate between the formal tribal or corporate structures and the constraints of a particular environment (Marsden 1976, p.13; Bates 1973, ch.5). Camp groups may also fluctuate in size on a seasonal basis due to regular changes in the composition of flocks and the size of cooperative Fig. 8.5 Schematic diagram of the relationship between the size and distribution of pasture lots and the size of campsites. (a) Continuous grazing tract; 50 sheep per household. (b) Continuous grazing tract; 100 sheep per household. (c) Dispersed grazing tracts; 50 sheep per household. (d) Dispersed grazing tracts; 100 sheep per household.
.
.
.
'
•
\
.
"
i
•
/ /
,
%
1
\
*
'
*
'
*
I
1
1 , \
>
•
/» •
.
/
1'
\
0.
.
1
/
.•
'
.
•
.
•
t
\
1
V . t
*
ty» '
»•
•
>
V ^^w *
*
r
*r>- -) •« #
tent
'y.y
grazing
access to pastures
.
X
*
;
\
Nomads in archaeology
144
groups needed to manage them (Hole 1978, p.163; Tapper 1979b). Larger concentrations of tents tend to occur under conditions in which defence and security are paramount, although this necessarily involves a trade-off between length of stay and availability of exploitable pasture. The largest camps commonly occur during migration when nomad-villager and nomad-nomad tensions are heightened, and when short sojourns in any one area place fewer demands on pasture. Where herds and people are massed in constricted and heavily populated valleys during migration, shortage of pasture and lack of security will force them to move rapidly, clustering together in large overnight camps. Summer camps of between four and 20 tents are common under present conditions, though large camps still occur in eastern Anatolia where strong tribal organization permits the efficient exploitation of large tracts of pasture. For the Kosan Kabile of the Beritanli tribe in eastern Turkey the large size of the camp is due to the fact that much of the land is owned by the kabile chief (Ozbek-Koroglu 1980) - as the pasture area covers a single large block and access presents no problem, much of the kabile is able to congregate in a single large encampment (see Fig. 8.3a). Shortage of grazing and competition with village agriculture in winter quarters is often responsible for a dispersed pattern of small, widely separated clusters of three or four tents (e.g. Basseri, Yoriiks, Alikanli) while restricted summer pastures, under conditions of relative security, will produce the same result in summer camps (Yoriiks). Although I observed a number of isolated tents in western Anatolia strung out along migration routes or in coastal winter quarters, single-tent camps are reportedly uncommon in other parts of the Near East and any tract of grazing land sufficient only for the needs of a single household may be passed over (Barth 1973a, p.121). Camp groups may be built up in modular fashion from herding units or may coincide with a single herding unit. In fact it is often assumed that the one stands for the other (Barth 1961; Swidler 1973). However if the herding unit is defined as a cooperative group brought together for the optimal management of household flocks in a single composite herd (i.e. an economic rather than a residential unit) then it is clear that herding units are often absent. Moreover, even if present as cooperative groups of households, herding units need not emerge as discrete spatial units within a campsite plan. Assuming the minimal herd size for efficient management to be around 250 sheep and goats (W. Swidler 1972, 1973), households whose flocks fall far short of this may be inclined to coalesce into herding units of four or five tents. However, among groups such as the Yoriik (Bates 1973), where household flocks are commonly well within the range of the optimal grazing unit, tents still cluster together in summer camp groups of minimally three to four, suggesting that the mechanism involved is not so much economic as social. Where extremely large herds of up to 1,000 are owned by individual households it is not uncommon to find camps of a similar size or larger, composed of the households of hired shepherds and clients, as among the Qashqa'i (see Marsden 1976; Beck 1980). Where this occurs campsite organization will tend to occur along hierarchical lines, with the smaller client tents being placed relative to that of the dominant household, rather than in modular units.
Structure and location of settlements
145
Whether or not herding units are involved, the physical arrangement of four to five tents in a quadrangular formation is an extremely common one in the Zagros region. That this is not a new development is indicated by early-nineteenth-century descriptions of groups of tents 'pitched in a quadrangular form7 (Ker Porter 1821, p.474) and an illustration in Grass, the epic account of the Bakhtiari migration, showing four tents in winter quarters pitched in a close square formation (Cooper 1928, plate opposite p.83). Larger campsites may be built up through the proliferation of such modules as is superbly illustrated in Watson's aerial shot of the Shirdashtis7 spring camp (see Watson 1979; also Fig.8.3b, this volume). Evans (1983, pp.75-6) also noted a tendency towards 'clustering7 in some Kurdish campsites in western Iran, alongside others organized in parallel rows.
Settlement density within campsites The density of settlement within a camp is the result of a compromise between opposing tendencies towards nucleation on the one hand and dispersal on the other. Requirements of sociability and security will serve to draw tents together (Hole 1978, p.151; Bates 1973, pp.122-4). Tendencies towards dispersal arise through the need for room between tents to manoeuvre herds of animals and basic requirements of activity space for domestic chores and trash disposal. It has already been observed in Chapter 7 that an area of approximately 10 by 20 metres operates as the optimal spatial unit within which a household can function efficiently. Camps organized in modular fashion through the accretion of herding units are the outcome of both nucleative and dispersive tendencies operating at different levels. Within the herding unit itself the joint management of livestock typically draws households into tight clusters, often around a central animal corral. Outside the herding unit factors leading to the dispersal of units - but also related to animal management - begin to take over: 'it is convenient to keep the tents of different herding units at a minimal distance of 20-30 meters from each other to avoid the mixing of herds7 (Barth 1961, p.42). In short, tents are located so as to facilitate interpersonal contact and economic cooperation while maintaining household autonomy in the domestic sphere. Digard (1975, pp.122-3) observed that Bakhtiari households tended to be spaced at a distance from one another suitable for carrying on a normal conversation. This concurs with Hole's (1984) estimate that Baharvand tents are seldom more than 10 metres apart and with my own observations of the location of domestic complexes in western Turkey. While the huge size of Alikanli tents in eastern Turkey creates the impression of higher settlement density, actual measurements reveal a similar modal distance between tents.
Patterns of residential association within camps
Traditional settlement patterns are usually not without kinship connections (Fraser 1969, p.47), and nomad camps are no exception. Camp groups are generally based on a core of agnatically or affinally related kin. Among the Yoriiks it is a generally recognized - though not rigidly adhered to - rule that sons, fathers and brothers should camp as close as possible to each other (Bates 1973, p.84) and it is said that 'only a hired
Nomads in archaeology
146
shepherd . . . would normally camp with non-kin7 (ibid., p. 11). However not all kinsmen in any given segment are expected to be present in the one camp and the layout of campsites is rarely a faithful reflection of agnatic structure. The co-residence of cognates and affines has been widely documented (Peters 1960; N. Swidler 1972; Barth 1961; Bates 1973), and although Watson (1979, ch.8) provides no systematic data on kinship connections in the Kurdish spring camp she visited, she does mention that her host's tent was pitched next to that of his mother-in-law. Although nomads themselves often assert that group cohesion is preserved in the course of migration from one campsite to another, in fact variability seems to be the rule (Barth 1961, p.42; Hole 1978, p.151) and both Bates (1973, p.202) and Be§ikgi (1969) observed that no camp group endures intact throughout a complete annual cycle. Four successive Bakhtiari campsites recorded by Digard (1975, fig. 150) show a considerable turnover in membership around a stable core of households and great variations in the actual spatial configurations of tents (see Fig. 8.4). More important in understanding the processes by which campsites are created as archaeological entities is the degree of consistency with which the same tentsites are re-occupied over successive annual cycles. This will vary with the stability of migration tracks and security of tenure in seasonal pastures. At one extreme, as among the Sagikara Yorliks, known campsites may not be re-occupied each year or may be occupied by different groups: There is great fluctuation in the amount and distribution of grazing tracts available each year. No segment of the population - however defined - returns year after year to the same pasture tract in summer or winter areas . . . [there is] continual movement of tents among fluid camp groups as they adjust to the grasslands available each season. (Bates 1973, p.83) Where relatively certain access to the same pastures and campsites is assured, greater consistency in the composition of campsites occurs, as was observed for the Beritanli and Alikanli. Among their respective Luri and Kurdish groups Edelberg (1966-67, p.388) and Watson (1966, 1979) noted that families occupied the same tentsites year after year. Both cases, like Sanaydin Yayla (Chapter 9), involve village-based groups. It has already been observed that sedentary life encourages an ossification of spatial associations within a community. It might be hypothesized that campsites occupied by groups involved in a more fluid system of pastoralism are subject to different rules. Where groups of different size and composition occupy the same locality only sporadically, recognizable tentsites will change hands repeatedly and many, if not all, will be out of use during any particular season. Long-term archaeological deposits on such sites will be built up irregularly in spatial terms and spasmodically through time. It is unlikely that the number of tentsites present would reflect the average size of the communities that had occupied the site in the past.
Socioeconomic factors in campsite formation
It has been argued (Chapter 4) that the problem of adjusting interdependent populations of animals and humans in a given area, within the frame-
Structure and location of settlements
147
work of a pastoral mode of subsistence, gives rise to three characteristic solutions - egalitarian, stratified and commercial. It is tempting to extrapolate these socioeconomic forms to the physical organization of campsites. For instance an 'egalitarian' community might be expected to display a modular spatial structure as herding units of three or four tents coalesce into larger, more open campsites. Just such a structure is documented for the Basseri on migration: Even on an open, level steppe the tents of herding units tend to form small semicircles within a larger semicircle, formed by the whole camp. This group migrates as a unit and its tents are pitched close together in a more or less standard pattern. (Barth 1961, p.42) A very similar modular pattern is found in the Shirdashti spring camp Fig. 8.3b) and it is worth noting that the average number of animals per household is given as 50, not far different from the Basseri or, for that matter, the Baharvand (Hole 1978, 1979), among whom a similar camp format, lack of stratification and tendencies towards sedentarization were also found. A stratified nomad community should exhibit hierarchical rather than modular organization, and again this is borne out by the arrangement of tents within a Qashqa'i camp consisting of a few large herd-owning households, together with their hired shepherds and servants: Tent groupings are usually based on kinship or economic dependence and the senior member of a group will choose the place for his tent first while the positions of the rest will be dependent on the relation of the owners to him, with his sons and brothers nearest them, at a distance of perhaps 100 m. or so, affinal kin and furthest away hired shepherds and labourers. (Marsden 1976, p.25) According to Beck (1980, p.336) the position of the shepherding family within a camp is ambivalent and its relations with the family of the herd owner highly asymmetrical. The women of the shepherd family may be required to assist in the domestic tasks of the employer household to a far greater extent than would be acceptable between independent households. Shepherd households are not considered full members of the residence group and this is 'indicated by their placement within encampments' (Beck 1980, p.340). A similar grammar of spatial organization operates in Alikanli camps where relatives place their tents with respect to the dominant household (normally occupying the largest tent) while hired shepherds and clients are located at a distance (Be§ikgi 1969, p.76). If Alikanli camps resemble battleships in line astern, then the chief's tent is the flagship at the head of the formation. The pattern is epitomized in a large Alikanli camp photographed in 1939 by Frodin (1943—44,fig.l4) which is shown in Fig. 8.6. In a rare series of plans of Beritanli campsites in eastern Turkey made by a Turkish architecture student (Ozbek-Koroglu 1980), the reflection of a strongly hierarchical tribal structure in the spatial stratification of the camp is clearly illustrated (see Fig. 8.7). The chief of the Kosan Kabile has his own camp of seven tents a good few hundred metres from the much
Nomads in archaeology
148
Fig. 8.6 A Kurdish camp of the Alikan A§iret in eastern Anatolia showing stratified organization and linear structure. (a) As photographed in the 1930s. (Frodin 1943-44, fig.14.) (b) An approximated plan of the campsite (not to scale). Note the focal location of the large sixpole black tent, probably that of the headman.
tent showing no. of poles
lake
broad valley
larger camp (around 70 tents) of the rank and file (see Fig. 8.3b). In contrast to the fairly homogenous and slightly modular layout of the latter, the chief's camp is highly structured, with careful placement of the tents of relatives, guests and shepherds. Present also are specialized structures such as a bath-house. While such examples are rare in the ethnographic record they no doubt persist in those tribal heartlands seldom penetrated by western ethnographers, and were probably common in the past. Offering as it does the prospect of re-creating tribal structures through settlement plans, this line of inquiry should appeal to those archaeologists who approach ethnography armed with a shovel. A note of caution is necessary. Even should it be possible to expose sufficiently large areas of a site and isolate single settlement events from the stratigraphic record, tent size may be related to other factors than social status, such as the number
Structure and location of settlements
Fig. 8.7 Beritanli chiefly camp in eastern Anatolia. This campsite, occupied by the chief of the Kosan Kabile of the Beritan A§iret, was located some few hundred metres from the camp of the rank and file shown in Fig. 8.3a. Note the spatial organization of the camp with tents of relatives, clients and guests arranged around that of the chief; also the rare luxury of a special bath tent at upper left. (After Ozbek-Koroglu 1980, fig.6, not to scale.)
149
access path
, G - t e n t ^ f
III//'•':•-'..:•'>
•:. Kabile chief
' / , 's '/\\l/mf%^ chiefs brother
of individuals in a household, and the logic governing the placement of tents is not self-evident in the spatial distributions themselves. The removal of all labels from Fig. 8.7 would leave us with few clues about the significance of the settlement plan, apart from a suggestion of hierarchy.
The overlap of nomadic and sedentary settlement forms The annual appearance of tent camps belonging to nomads of the Khamseh Confederacy in the vicinity of the ruins of Persepolis in Fars Province of southwest Iran (Field 1939, p.549) may be more than just fortuitious. The plans and mode of construction of the palaces themselves bear more than a passing similarity to the royal tent enclosures of Central Asian and Near Eastern tradition (see Andrews 1977), and at least one writer (Herzfeld 1935) has drawn attention to this. Could Persepolis itself, with its open architectural plan and serried ranks of pillars like rows of tentpoles, have been derived from a tent-like blueprint? Are the Persian-speaking Basseri of the Khamseh Confederacy descendents of original Persian nomad stock before the time of Darius and Xerxes? Indeed the form of the palace set atop an extensive stone platform has a modern analogue in some Qashqa;i chiefly tents. These speculations lie outside the scope of the present inquiry and, besides, one could just as easily argue on stylistic grounds for a nomad connection in the Parthenon or the Chinese pagoda.
Nomads in archaeology
150
But the image of Persepolis surrounded by black-tent encampments is a useful metaphorical vehicle in another sense. There have been many instances reported of tent camps in close association with ancient ruins and archaeological sites throughout the Near East (Ainsworth 1842; Bent 1891; Bishop 1891; Field 1939; Forbes 1831; Garnett n.d.; Hole 1979; Schmidt 1940) and my own field observations bear this out. An encounter with Shammar Bedouin tribesmen occupying the ruins of Hatra in northern Iraq is not atypical: Here we perceived the tents of the Bedwins extending far and wide within the ruins and without the walls to the south west. The finding of Arabs here is indeed what may generally be expected of any traveller in search of these ruins. Tne number of halting places which present what is actually necessary for the Arabs - water and grass - is not so great . . . For the same reason that cities were built on these oases in the wilderness, the wandering Arab now resorts to situations where there are waters, and with them pasturage. Hence the traveller may be almost as sure of meeting Arabs at Palmyra as at Al Hadhr. (Ainsworth 1842 II, p. 164) My first contact with the Alikanli occurred in just such a location beneath the walls of a ruined caravanserai of Seljuk vintage where a dozen or so tents were drawn together in a pre-migration autumn camp (Fig. 8.8a). The resemblance of this to one of Schmidt's (1940) air photographs of abandoned tentsites and corrals clustered around a ruined Sassanian palace is uncanny (Fig. 8.8b), and descriptions by nineteenth-century travellers such as that of Ainsworth (1842 II, p. 339) of a ruined fortress towering prominently above an encampment of tents like an ancient feudal castle7 have a similar flavour. We have already come across references to Yoriiks wintering in classical Greek, Roman and Byzantine ruins on the Mediterranean (Garnett n.d., p.199; Bent 1891, p.271) and seen a modern example at Karakegi (Chapter 7). Examples from other areas of the Near East are not difficult to find (e.g. Hole 1978, p.153; Dickson 1910, p.368). Fig. 8.8 Campsites located in association with ancient ruins, western Iran and eastern Anatolia, (a) Alikanli autumn migration camp adjacent to a ruined caravanserai of the Seljuk period. Tents have been placed over existing stone walls.
-' A,
Structure and location of settlements
151
Fig. 8.8 (b) Abandoned stone tentsites and corrals adjoining the ruins of a Sassanian palace, Firuzabad, western Iran. (Reproduced from Schmidt 1940, pi. 17.)
In terms of the locational factors examined above; it is not difficult to see why nomads should be drawn to ruined structures. Not only are these likely to be located close to important resources or access routes, but they also provide shelter, and rubble furnishes building materials for tent foundations, corrals and hearths. Not only extensive monumental ruins but also mounds and tells of modest proportions attract nomad encampments and for much the same reasons. A good example is the small Luri campsite in the lee of an archaeological mound and two campsites extending from the foot of Tepe Guran in Luristan (Fig. 8.9). In respect of the latter example it is perhaps worth noting that the basal layers of Tepe Guran were considered by the excavator (Mortensen 1964) to represent an early neolithic seasonal camp of herders. Among the records of nineteenthcentury travellers in the Near East, the association of tent camps with tells or ruined villages is extremely common (e.g. Ainsworth 1842 II, p.120).
Camps and villages Just as fixtures such as tent foundations may be specially constructed by nomads, often out of the rubble of ruined structures, so existing walls in abandoned settlements may be pressed into service, as has already been noted in Chapter 7. There are scattered historical references to recently abandoned villages being occupied by nomad tents (Mordtmann 1925, p.531; Frodin 1943-44, p.268; Tchalenko 1953, p.104). Sixteenth-century settlements on the Amuq Plain, abandoned until well into the twentieth century, were apparently used in this fashion by wintering Turkmen and Kurdish nomads: 'there were no longer villages in the plains but only ruined walls which the nomads often used, in combination with their tents, as temporary winter encampments' (Aswad 1971, p.18). Abu Marri,
152
Nomads in archaeology
Fig. 8.9 Nomad campsites in association with archaeological mounds. (a) Plan of a Luri encampment near Tepe Guran, western Iran. Note the presence of three rows of tents and close correspondence of middens with tentsites. (After Melgaard ef */. 1963, fig.4, not to scale.) (b) Campsite in the lee of an archaeological tell, Deh Luran, western Iran. (After Hole n.d., fig.4.)
zjfi 'i '• ) I \ \ \ /
-.Tepe • / ' >, I Guran. '.\-\
.
TepeGuran
TEPE GURAN
&• ash& midden
corral enclosed by brush fence about 30x40 metres
LEGEND
a chul (for bedding) » stones for bags * cooking fire .#. dung am ash -^tent opening — ditch o
archaeological site, DL-61 4-5 metres high
10
a settlement in the Mosul district described by Forbes (1851) as a 'ruined village7, was observed a few years later to be 'now inhabited, but only by occupiers of tents7 (Ainsworth 1842 II, p.123). An alternative possibility is that the 'abandoned village7 represented a fixed stone-walled campsite, but the earlier account by Forbes suggests that it was a permanent village. The implications of such behaviour for nomad archaeology are twofold. First, in interpreting the settlement histories of particular sites, makeshift reconfigurations of architectural features and the reconstruction of
Structure and location of settlements
153
crude shelters out of the rubble of more sophisticated dwellings may well have been the work of nomads. This does not mean that every site exhibiting such features was a nomad camp - the criteria for determining this are by no means clear cut, but the possibility is ever present (Burney 1961; Voigt 1977). Second, a persistent regional pattern of contemporaneous abandonment or destruction followed by ephemeral settlement may be evidence of a phase of nomadization. In view of the long history of migrationist and invasionist arguments, established conventions in the naming of stratigraphic levels and construction of chronologies, this should come as no great revelation (Mellaart 1960; Mellinck (ed.) 1964). Before I am accused of harbouring migrationist tendencies, let me point out that few migrationists have bothered to address the question of how variations in architectural forms, site density and accumulation rates can be used as evidence for regional settlement and population trends. The connection is simply assumed (e.g. Stronach 1963; Oates 1976). Invasion, or a breakdown in state power opening the way for invasion, is assumed to lead to the demise of one culture and population as they are overrun by another. The latter, being 'nomadic', leaves few material remains. This is not what happened on the Syrian steppe in the last century, nor in western Iran during World War II and only a few years ago after the fall of the Shah. What would a migrationist make of the dozens of abandoned and burned villages? What of the nomad camps that replaced them? Yet in both cases - certainly in the latter - the departed villagers and the newly arrived nomads were in fact the same people! The collapse of state authority, instead of opening the way for a nomad 'invasion7, simply permitted large numbers of people to resume a preferred migratory lifestyle more consistent with their unstable mode of subsistence. The widespread desolation in northern Mesopotamia recorded by nineteenth-century travellers, with numerous deserted villages interspersed or overlapping with tent camps or Villages of tents' (Ainsworth 1842; Geary 1878; Ubicini 1970), is consistent with the known facts of weak Ottoman rule and mounting disorder. But there is nothing to say that the nomads were not themselves former inhabitants of these villages, or others nearby, who had simply become too mobile to maintain a permanent village base. Certainly the entire northern Mesopotamian steppe witnessed major shifts in tribal affiliation and spheres of influence at this time (Aswad 1971; Sweet 1974; Eberhard 1953b, Tchalenko 1953). There is also some suggestion of the nucleation of settled population in fewer and larger villages and towns - generally walled - with large tracts of intervening land remaining uncultivated (Geary 1878 II, p. 12). In Iran, where conditions were equally unsettled, it has been estimated (Hambly 1964, pp.69-70) that over half the population was nomadic. A decline in the state of repair of irrigation systems and under-cultivation attended this phase of nomadization and, according to Hambly (ibid., p.70), 'all observers describe in some detail the deserted villages and abandoned agricultural land formerly under cultivation'. However while nomadization was very much in evidence, and a certain amount of territorial instability was present, these events could in no way be likened to an invasion or even a continuous in-migration. It is curious that the period during which major nomad invasions are
Nomads in archaeology
154
known to have occurred throughout the Near East - the eleventh and early twelfth century - was one of general prosperity and flourishing trade with little hint in the historical or archaeological settlement record of major upheavals or depopulation (Cahen 1968; Turan 1971; Minorsky 1978). Prior to the Seljuk Empire most of this area was Greek- or Persianspeaking, and afterwards mostly Turkish-speaking, showing how complex are the interwoven strands of nomad migration, ethnic affiliation and regional economy.
Composite settlements: tents in a sedentary context
It was noted earlier (Chapter 6) that tent-like spatial organization may be preserved in the floorplans of village houses established by sedentarizing nomads, but there is also evidence for the actual use of tents within sedentary settlements either as temporary dwellings or as appendages to more permanent structures. Instances abound of newly settled nomads preferring their traditional tents pitched in the courtyard beside newly constructed houses, the latter being used mainly for storage (Stark 1933, p.258; Schuyler 1966, p.192; Vambery 1868, p.303). Prior to the spring migration, the Shirdashtis (Watson 1979, p.243) moved into tents pitched in the courtyards of their stone-built houses and one family resided permanently in a tent. A few families in the nearby village of Hasanabad also moved into tents on the outskirts of the village during summer (Watson 1979). Fully sedentary Kurdish villagers were accustomed to pitching tents in their villages to avoid the heat and infestation of vermin (Hay 1921, p.46). At Tell Toqaan, in northern Syria, two families in the early stages of sedentarization continued to live permanently in tents on the village outskirts, not to be confused with the large seasonal increment of nomad tents occupying the village common during the summer, autumn and early winter (Sweet 1974, p. 115). The arrangement by which the Alikanli and other eastern Anatolian nomads pitch their tents in lowland villages to become full members of the community for the duration of the winter has already been described (Chapter 6), and there are indications that similar practices occur in other areas of the Near East. We are again in that fuzzy area, already encountered in respect of tents and houses, where we are dealing with neither one thing nor the other. The ambiguity is underlined by the use of the term Village' by Audouze and Jarrige (1981) to describe Baluchi campsites consisting of both roofed mud huts and tents and the frequent use of the expression Villages of tents7 by travellers and casual observers. The composite dwelling has its counterpart in the composite settlement. It is certain that in past centuries composite settlements consisting of both tents and houses - perhaps accompanied by a fortress or public buildings - were a regular feature in many parts of the Near East (see Ainsworth 1842; Jarring 1939, pp.60-1). Historical accounts of the Aqqoyunlu Empire (fifteenth to sixteenth century) mention the establishment of 'fortified campsites' at those points in eastern Anatolia where the east-west trade routes and the north-south pastoral migration routes intersected (Woods 1976). Pastner (1978) also suggests that a settlement type consisting of a central fortress, permanent dwellings and - at least seasonally - surrounding tent 'suburbs' was common in pre-modern Baluchistan.
Structure and location of settlements
155
Even today complexes of yurts surrounded by substantial mud walls and forming an orderly arrangement of blocks and streets have been reported from Mongolia (Hodges 1972, p.525). During the last century summer camps of the more powerful Bakhtiari khans were located around a central stone fortress (Bishop 1891 I). Such a composite settlement form is the logical corollary of the integrated tribe and achieves its fullest expression in the fortified camps of dimorphic chiefdoms (Rowton 1973, 1977) and the royal tent enclosures attached to Central Asian cities (Andrews 1977). The possibility raised by Adams (1974, p.7), that certain early Mesopotamian cities consisted of an administrative core surrounded by a large floating population of tent-dwelling tribesmen, should be reconsidered in this light.
Settlement density Although camps and villages may be found in association and although they display superficial similarities, the rules governing their formation are profoundly different. As stressed in Chapter 4, the high fluidity in camp group membership, together with the fact that each camp is a fresh settlement event, allows full scope for the expression of kinship, social, political and economic relations between households in spatial terms. The study of nomad camps can add an important dimension to settlement archaeology, a field represented so far by studies of hunter-gatherer campsites (Weissner 1982; Yellen 1977; O ;Connell 1979; Whitelaw 1983), village communities (Kramer (ed.) 1979a; Jacobs 1979; Watson 1979; Sumner 1979) and formative cities (Adams 1974; Johnson 1973). Recent comparative studies (Whitelaw 1983; Fletcher 1981) suggest the operation of certain tolerance thresholds relating to crowding and communication networks, the spatial equivalents of Tapper's (1979b) hierarchy of community types. Whitelaw (1983), using an array of multivariate techniques, demonstrates that the tolerance threshold of crowding in hunter-gatherer communities depends very much on kinship criteria. In other words households are not disturbed by crowding per se, but only crowding by strangers, a further demonstration of the direct relationship between social distance and physical distance in such settlements. Quantitative studies of the fit between patterns of kinship and affinal relationships between households in a campsite and the spatial relationships of the corresponding tents are urgently needed to throw light on this question. Given the constraints under which fieldwork was carried out, I was unable to undertake such an investigation. Be§ikgi (1969) collected detailed information about the kin and affinal ties between household heads in one Alikanli campsite, but unfortunately the exact locations of the tents with respect to one another were not recorded. Nevertheless, on the basis of numerous observations and examples from existing literature, it appears that tent camps respond in a fundamentally different way from fixed settlements to increases in population. Whereas the density of a sedentary settlement might be expected to increase or remain fairly constant with the addition of new households, the same increase in a nomad camp will result in an increase in area and a decrease in density per unit area. It is therefore of some interest that the only extant density studies of
Nomads in archaeology
156
nomad camps, this time without corresponding kinship data, carried out by Hole (n.d.) revealed precisely the opposite trend. Hole concluded that intra-site density in Luri nomad camps was comparable to ; perhaps even in excess of, that in sedentary villages in the same region. Moreover there was a tendency for the smallest campsites to display the highest densities. This latter finding at least is in line with the hypothesis outlined above, but perhaps for the wrong reasons. I believe that these results are due to a fundamental error in the measurement of occupation density. Most of the nomad camps measured by Hole (n.d., table 2) were small, consisting of only four or five tents (see Fig. 8.9b). They were apparently herding units laid out in quadrangular form around a corral, and most were migration transit camps. In settlement units of this size there are problems of area definition which would not arise if a much larger number of dwellings was present. For instance an area occupied by four tents arranged in a square formation, measured from the outer corner of each tent, would yield a far higher density that that for 16 tents, the same distance apart, and measured in the same way. But 16 tents are unlikely to be arranged the same distance apart as are four. The modular organization of tent camps referred to above - four clusters of four tents each - would result in a far larger area and a much lower density. By allowing the results to be unduly influenced by distorting effects at the lower end of the size range, the relationship between area and number of units has been obscured. Well might the author remark that 'the camps are compact, although because of their small size they seem relatively open7 (Hole n.d.). The camps are in fact quite open, and because of the small numbers of tents involved they give the statistical illusion of being compact. If consolidated into a larger camp group the tents would not be distributed evenly over the surface but clustered into herding units, or strung out in parallel lines, with large spaces between each unit or line. This would have the effect of substantially lowering the overall settlement density. The same variables appear to work in the opposite direction in the case of village settlements - the smaller the village, the lower the population density, and vice versa. But in order to rigorously test such hypotheses, settlement data from nomad camps of a much higher quality than exist at present are necessary.
Sedentarization and settlement Regardless of whether or not fixtures are present, a campsite is reestablished anew with each annual occupation. Each occupation is a fresh event, to a large extent independent of previous events. By contrast, a permanently occupied village, or even one that is seasonally abandoned, has a history. Each new alteration or addition builds on a sequence of existing structures. Moreover a sedentary settlement normally passes through phases of occupation - establishment, growth, saturation, decline - and its position within the sequence is reflected in its physical appearance, layout and density (see Fraser 1969). The typical nomad tent camp, like the camps of hunter-gatherers, has no such history but remains permanently retarded in the initial stages of the normal developmental cycle. But nomad camps differ from those of hunter-gatherers in that there is
Structure and location of settlements
157
some interaction with sedentary settlement. Campsites may develop into villages or composite settlements. Sedentarizing nomads may reproduce in their newly established settlement the same kind of open plan and alignments of dwellings that are found in their campsites. A survey of village settlements in Fars Province, Iran (Sumner 1979), distinguished three types of village plan: square or rectangular settlements of high density surrounded by a wall; unwalled and irregular but densely built-up settlements; and a third type, representing only about 5 per cent of the sample, characterized by a low settlement density and open plan: There can be no doubt that cultural preference plays a large role in determining the plan and layout of villages which in turn affects the density of occupation. In this case it is known that a number of the low density villages, composed of scattered houses, are inhabited by recently settled pastoral nomads. It is probable that most of the very low density villages are of this type. (Sumner 1979, p.168) I am a little sceptical about the role of 'cultural preference7. As Sumner himself points out (1979, p. 167), it is the smaller villages which are the least densely settled and both variables, size and density, may be dependent on the length of occupation and the stage which a community occupies in the cycle of settlement development. But the identification of open-plan villages with recently settled nomad communities is interesting. Many of the aerial photographs taken by Schmidt (1940) show small villages, consisting of parallel strings of houses arranged end to end, almost indistinguishable from campsites such as those depicted in Figs. 8.1 or 8.3. While Schmidt also shot numerous deserted campsites with lines of stone foundations, there are few photographs of camps actually occupied by black tents. It will be recalled that the 1930s, when these reconnaissance flights were made, was a period of enforced sedentarization, and it is highly probable that most of these villages were occupied by recently settled nomads. Indeed the author describes a number of landings at Bakhtiari villages where he was the recipient of a full measure of nomad hospitality. Had he flown this way ten years later, he would most probably have found the same settlements burned and deserted, their former occupants having returned to the previously abandoned campsites. Observations in the Syrian village of Tell Toqaan (Sweet 1974), which drew much of its population from nomadic tribes in the vicinity, throw further light on the organizational changes which occur with sedentarization. At the time of study part of Tell Toqaan was regularly used as a camping ground by nomads who had tribal and kinship affiliation with the Tell Toqaanis and two families had only recently made the changeover from tents to mud-brick, domed houses. The village was still a focus of sedentarization and can be characterized as a composite settlement. According to the ethnographer: At first view the settlement of Tell Toqaan appears open and to have grown up by random accretion wherever there was space to build . . . Distribution of the dwellings and compounds of the village is not as compact as that in much older and more populous places in the area
Nomads in archaeology
158
. . . There are many instances of open spaces between blocks of several compounds wnose walls are shared, a number of compounds and dwelling units stand free. At Zammar and Ras i Ain; both tribal villages of the Bu Layl, the houses and compounds are detached and do not share walls with their neighbours. (Sweet 1974, pp.48-51; see also Fig.8.10a) An open settlement plan was also reported for a new Yoriik settlement on the Black Sea coast of Anatolia in which the 'tradition of scattered living' is said to derive from a recent nomadic past (Donmez 1964, p.167). The sedentarization of Yoriiks on the outskirts of an existing village has been described by Bates (1973), though he does not trace any subsequent changes in settlement plan, and Peters (1976) gives an account of the architectural changes accompanying sedentarization in an east Anatolian Kurdish community. While these examples are highly suggestive of the kinds of processes at work in the course of sedentarization, a full account of the settlement cycle is not possible on the basis of synchronic observations of settlement plans in different phases of the sequence. While no such sequence has been described for Iran or Turkey, changes in settlement pattern following sedentarization most probably occurred in a manner similar to that documented in Algeria: These new villages differ from the traditional ksour not only in density but also in the type of house which is eventually built. The process is gradual ana begins with the tent and its enclosed corral, spaced at the usual distance from the neighbouring tent. As materials become available, individual rooms are built inside the corral and soon the tent is folded for the last time. (Etherton 1971, p.182) Such a pattern of accretion appears to have been responsible for the compounds and dwelling clusters at Tell Toqaan and other recorded villages, evolving from a terminal campsite whose tent groups ossified into the present form (see Jarno 1984). The intensification of settlement in an already established village through the accretion of new dwelling units is nicely illustrated by an . abandoned village in southern Iraq (Nissen 1968). The village, established around 1902, appears to have had a strong pastoral component and may have been established by settling nomads (Fig. 8.10b). It was abandoned after a shift in the course of an irrigation canal along which the settlement had grown up. The dwellings, analogous to some tent forms, have been described in Chapter 6. Variations in settlement density between the compact southern and dispersed northern parts suggest different stages in the settlement's evolution (ibid., p.108). The sequence of construction of walls and other features indicates that courtyards and ancillary structures were added to the dwellings and that the crowding together of units in the southern portion is due to the expansion of courtyards to fill in the open spaces between dwellings, combined with the establishment of new dwellings within these spaces. This expansion must be largely attributed to the accomodation of livestock in the courtyards, recalling Summer's (1979, p.172) observation that the more dispersed character of sedentarized nomad villages may be influenced by the need for corral
Structure and location of settlements
159
space. The layout of the dwellings alone, minus the courtyards, is not unlike a typical nomad camp. The presence of a strong pastoral component in the subsistence base and of composite dwellings with mud walls and reed superstructures suggests a highly unstable settlement system. At the same time the joining up of courtyards into complexes in the southern section is reminiscent of the settlement pattern described above for Tell Toqaan. Through a process of continual infilling and the joining up of courtyards a densely packed, agglomerative village may eventually arise, examples of which can be found in Summer's (1977, 1979) and Jacobs;s (1979) walled villages and the communities studied by Kramer ((ed.) 1979a) and Watson (1979). Such concentrated settlements appear to be preferred among peasant cultivators in Iran and Turkey for the sense of security they confer, as well as conserving precious agricultural land. Where villages are surrounded by arable land there are naturally strong pressures to locate new dwellings within the settlement rather than occupying potentially cultivable land. Once saturation point has been reached - and this may vary culturally - any further population increase is likely to spill outwards as lowdensity, extra-mural settlement where the process of accretion begins anew (Jacobs 1979). An insight into village spatial organization comes from a resettlement scheme after a major earthquake in Qazvin, north-
Fig. 8.10 Two plans of villages arising out of initial impermanent settlement, (a) Plan of Tell Toqaan, northern Syria. Note the open but clustered settlement pattern. Two tents of sedentarizing nomad families are present in the southeast, adjacent to a camping ground seasonally occupied by migrating Bedouin. The village itself appears to have been formed through the accretion of sedentarizing households. (Reproduced from Sweet 1974, fig.4.) Fig. 8.10 (b) overpage
Nomads in archaeology
160
Fig. 8.10 (b) An abandoned village in Iraq. Note the variations in settlement density between the northern and southern parts. (From Nissen 1968, pl.l.)
KHANAZIRIYA t
4
M
10
!•»
n
JO " C u t l»7
GQflZH
IJ4
DWELLING ROOMS GUEST ROOMS ROOMS FOR
ANIMALS
UNIT ENCLOSURE WALLS INTERIOR
DIVIOIKIG WALLS
SLEEPING PLATFORMS OPEN SILOS OPEN
FOR
SILOS FOR
FOOOER
AND
SET ON WALLS
STRAW FOOD
WATERTROUGHS
TENUR.
Plate 1
west Iran. It was found that villagers had definite ideas about the construction and spacing of houses and many new government-built settlements remained unoccupied because they did not conform to those norms: The older villages were built with little space between the houses; streets were wide enough to allow the passage of animals and carts. The closer the dwellings, the more secure the villagers felt: animals had to be kept closely under the owner's eye; (Bina and Dyas 1975-76, pp. 174-5). It is clear that changes in dwelling type from tent to composite dwelling to permanently roofed structure are strongly associated with changes in
Structure and location of settlements
161
settlement structure and density. Any attempt to evaluate archaeological material in terms of house form alone, or settlement plan unrelated to house form; will be sacrificing half the evidence. The sedentarization process should be archaeologically visible in terms of a system of variables relating to both the dwelling unit and the overall settlement plan; as well as trends observable at a regional level.
Conclusion In developing the above settlement sequence I am not suggesting that all villages in the Near East began as nomad camps. Rather the principles of nomad settlements may help us understand the early stages in the establishment and growth of settlements generally. In view of the similarity of some kinds of locations favoured by both nomads and villagers (e.g. on the edge of mounds or tells, as in the case of Tell Toqaan) there is reason to suspect that the transformation of nomad camp into permanent settlement may have been a frequent one. It is in just such locations that archaeological indications of the interplay between sedentism and nomadism should be sought. In terms of intra-site spatial organization, the highly generalized living space around each individual tent will be partitioned and enclosed within walled compounds as sedentarization proceeds. The pattern of discard changes from household-specific middens to specialized communal midden zones. Once defined by permanent stuctures, the organization of living space is better understood in relation to an overall settlement plan rather than in relation to each individual dwelling. In other words the idea of a domestic complex begins to lose its utility at that point where a settlement begins to acquire a history and to pass through a developmental sequence.
9 SARIAYDIN YAYLA Beyond Aladag Mountain is a long valley, densely wooded, with hundreds of springs bubbling forth everywhere, bright, cool, pebbly springs bedded in mint and heather. It is light that flows from these springs, not water, but a tinkling brightness. Since time out of mind this valley has been the summer pasture of the wandering Turcomans, the Yoriiks .. . (From the Legend of the Thousand Bulls by Yashar Kemal)
From the modern highway crossing the Taurus Mountains, Sanaydin Yayla appears as a collection of small beehive-shaped tents in a narrow hanging valley. Closer investigation will reveal stone and thornbush corrals scattered amongst the tents, a number of mud-roofed huts and well-beaten paths (see Fig. 9.1). What at first appears to be a stone-built, roofed corral roughly in the centre of the camp soon reveals itself as a ruined Seljuk caravanserai (Erdmann 1961; Ozergin 1965), and out of the stone rubble lining the banks of a seasonal watercourse the practised eye will discern the remains of a paved road and arched stone bridge. The paved track winding its way from the valley floor to the Taurus watershed, at around 1,600 metres, is part of an ancient trade route running from Silifki on the coast to Karaman, Konya and ultimately Ankara on the Anatolian Plateau, mentioned in sources of the early Seljuk Dynasty in the thirteenth century (Turan 1971, pp.346, 512). That this area also served as a camping ground or migration route for nomadic groups of the period is strongly suggested by what can be pieced together from the historical records. Late in the twelfth century, we are told, tribes of Turkmen nomads, newly arrived on the Cilician Plain, began to shift their migration tracks westward into the Goksu Valley, following a pattern very similar to that outlined in Chapter 4: Finally, due to population density and the pressures of migration, the [Turkmen] nomads entered Cilicia and reached Kozan by AD 1187. According to certain historical calendars it seems that Kilif Arslan conquered Silifke . . . These migrations from Central Asia increased the [population] density at the Anatolian frontiers and resulted in expansion and conquest at the expense of the Byzantines. (Turan 1971, p.216; my translation) 162
Sanaydin Yayla
163
^- v / W &.£r
NK.
F/g. £ 1 General plan of Sanaydin Yayla, a summer camp of Yoriiks based in a village in the Goksu Valley, southern Turkey. Note the ancient road cut by a stream bed and the ruined caravanserai, now used as a pen for cows and donkeys.
V^,::^^
The conquest of Silifke at the coastal end of the valley no doubt opened up the Goksu to regular tribal migrations, and Sanaydin Yayla, occupying the upper extremity of the valley, would have served as an ideal transit camp for large groups of nomads who followed in the wake of the Turkmen forces and constituted both their manpower reserve and productive base. Whether tentsites of Seljuk vintage are to be found beneath the stamped earth floors of the contemporary tentsites is an open question. Whether there is any direct historical link between the Seljuk tribesmen and the current inhabitants is also uncertain. The subsequent history of tribal movements and settlement in the area is imperfectly understood. The continual process of subdivision, extinction, realignment and movement of similarly constituted tribal units does not yield easily to the historian's tools of trade.
Nomads in archaeology
164
The camp and its inhabitants Climate
The appearance of Sanaydin Yayla varies dramatically with the seasons. Winter will find only the tops of the stone-walled huts and hearths protruding above a heavy snow cover. With spring comes a raging torrent of melt-water down the stream and springs bursting unpredictably out of the valley slopes, though deep snow drifts persist on sheltered northwardfacing slopes. By May, when the pastoralists arrive, the last snows have ceased and the valley floor is carpeted with grass and wild flowers. Summers are mild and dry except for a few early and late thunderstorms. During autumn temperatures remain around the 25-30°C mark until midOctober when they begin to drop sharply and the frequency of showers and storms increases. One by one the families depart until, by the time the first snows lightly carpet the valley, the campsite is once again deserted.
The people The inhabitants of Sanaydin Yayla claimed no tribal affinity, but identified themselves as Beyirli, indicating an origin in the winter village of Beyir, 1,500 metres below in the Goksu Valley. On a map of tribal migration routes in southern Anatolia (see Fig. 7.4) are mentioned two tribal groups bearing this name, one 300 kilometres west of the Goksu, the other about half that distance. One might therefore suggest a western Anatolian origin, making these part of the recent eastward movement of Yoriiks along the southern Anatolian coast (Bates 1973). Each year up to 50 individuals may reside in the camp, the majority originating in the village, although a few families were said to migrate from the area around Tarsus 150 kilometres to the east. The campsite itself was owned by the village cooperative, as is commonly the case with village communities migrating annually to yaylas (Tuncdilek 1963, p.64). Many of the families were related in some way - or a number of ways. All the families in the central part of the camp were described by my main informant, Mustapha Bey, as akraba (consanguilial kin), while those on the southern edge of the camp were described as kom$u (neighbours). Two themes relating to marriage, commented on by ethnographers working with Yoriik populations, were very much in evidence at Sanaydin. The first of these, kiz kaqirma (literally 'kidnapping the girl; or elopement), has the effect of obviating or substantially reducing brideprice payments (Bates 1973, ch.III; Kudat 1974). The second is the institution of parallel cousin marriage, common throughout the Near East (Barth 1973; Khuri 1970; Murphy and Kasdan 1967). Most of my informants claimed to have been married through kagrma and the topic seemed capable of generating a great deal of ribald merriment among both men and women. According to one man, it was his wife who did the kidnapping, seating him on the donkey and urging haste lest her outraged relatives catch up with them. Another couple claimed to have made kagrma on the Ankara intercity bus. (My own claim to have made kagrma by
Sanaydin Yayla
165
aeroplane almost brought the tent down.) The presence of cousin marriage was strongly suggested by the fact that kin ties frequently appeared to be reinforced by affinal ties. The combination of kiz kaqirma and parallel cousin marriage; it has been suggested (Bates 1973,ch.III), gives rise to a settlement pattern characteristic of nomads. Since many kaqirma marriages occur between close relatives - the boy, considering that he has first right of refusal over his cousin, may kidnap her to prevent an arranged marriage with a non-relative - the resultant animosity causes a splitting of agnatic camp groups. I have no evidence with which to support or contradict this claim, however one marriage practice observed at Sanaydin is worth noting. There was evidence of a pattern of marrying into villages adjacent to the yayla, which may be part of a long-term strategy to build up ties of affinity - and later kinship with villagers in the yayla area. Another rather intriguing possibility is that the pastoralists have always maintained closer links with the mountain villages than with those in the region of their winter village, that in fact the affinal links have some historical depth, and that close ties of kinship exist between the two communities - ties which may reflect a common tribal origin. If this is correct then existing notions about the relationship of so-called transhumant villagers to their summer camping grounds (Tuncdilek 1963) may need to be drastically revised. For the people of Sanaydin Yayla the summer migration did not seem to represent a break with on-going community ties with which one tends to associate the idea of sedentism, but a re-establishment of those ties. Viewed such, the winter village or ki§lak may be seen as a place to sit out the winter rather than the main focus of kinship and community relationships.
Subsistence Like many contemporary nomads and transhumant pastoralists, the inhabitants of Sanaydin Yayla pursued a mixed productive system based on both pastoralism and agriculture, although certain families tended to specialize more in one or the other. Gardens and orchards were concentrated at the lower end of the valley where an improvised irrigation system permitted the exploitation of pockets of fertile alluvium to grow apples, tomatoes, corn, beans and chick peas, most of which were consumed locally. Any surplus not consumed during summer was bagged for transport to the village. Of the four species of large animals kept, sheep were the primary commercial focus. A few families kept a cow for regular milk supply, and donkeys were used to transport both people and loads, particularly firewood. Goats contributed milk, hair for weaving kilims and tentcloth and also served as flock leaders. While wool and milk products were used for subsistence purposes, sheep pastoralism was geared primarily to the meat market. Pastoralism was largely a male preserve, and within the pastoral sector the division of labour between husbandman and herdsman lay along generational lines. This was illustrated in the family with whom I spent most time at Sariaydin. At first glance there seemed to be some support for the view that Mustapha Bey's life consisted of a ceaseless round of excursions to villages and other camps and fireside chats with kinsmen or neighbours. A
Nomads in archaeology
166
second glance revealed that much of this social activity had as its object matters such as the negotiation of prices for animals and animal products, the buying of provisions and arranging transport for the migration or the sale of livestock. He was concerned with the 'foreign affairs7 side and with securing the necessary conditions and markets for a viable herding system. It was the eldest son, Salih, who carried out the bulk of the physically demanding herding activity, operating at the tactical level of exploiting pasture, protecting the stock and managing short-range movements. The flock was taken out every evening at around 5 p.m. and grazed all night in upland pastures. The herdsman returned with them at about 10 a.m. and had a meal and an afternoon sleep - having taken only cat naps during the night watch - before the gruelling cycle began again in the evening.
Architecture Many of the processes underlying changes in dwelling form discussed in Chapter 6 were strongly in evidence at Sanaydin. Both tents and stonewalled huts were present, the former outnumbering the latter by about two to one, but there was evidence of a trend away from the use of tents and towards the construction of huts. The location of Sanaydin with respect to the sedentism-nomadism continuum might be described as intermediate or transitional, with a currently discernible tendency towards sedentism. Under conditions of relative peace and security, where the land is owned or leased, and where market forces do not favour the expansion of pastoral activity one might expect, (a) a mixture of temporary and permanent dwelling forms, and (b) a tendency towards the predominance of fixed as opposed to portable architectural elements. This is precisely what seems to be occurring at Sanaydin Yayla. The tents employed at Sanaydin, of a kind common in southern Anatolia, are known variously as alaak or gatma ev, and may be represented geometrically by a half cylinder rounded at both ends. Examples of this type have already been referred to in Chapter 6. Four or more semicircular hoops, each consisting of two curved sticks lashed together, are placed in a row parallel to each other and horizontal struts are tied at intervals, parallel to the ground. At each end curved sticks are placed in an arc and tied together at the top. A reed screen is placed around the perimeter and thick felt slabs or goat-hair tentcloth draped over the top of the framework thus formed, held in place by encircling ropes weighted with stones (Figs. 6.4, 6.5 and 9.2b). Tents of essentially similar construction have been described in adjacent areas of the Goksu Valley and the coast between Mersin and Silifke by early-twentieth-century observers (Yalman 1977, p.241; Mordtmann 1925; Brandenberg 1905; Wenzel 1937). Though structurally and genetically the Sanaydin tents bear no similarity to the tig direkli (four-poled) black tents also employed in the same region, the ground plans and arrangement of fixtures within and around the dwelling site were indistinguishable from those found in black-tent encampments. Observation of an abandoned tentsite would provide few clues as to the superstructure, aside from the presence or absence of peripheral tent-peg holes and depressions or stone supports where poles
Sanaydin Yayla
167
Fig. 9.2 Views of an alaak at Sanaydin Yayla, abandoned and occupied. (a) Frame of an abandoned alaak. The wooden frame is only partly intact and the plastered stone hearth and storage structures have been somewhat disturbed. (b) An occupied alaak showing forecourt and work areas. External storage structures have been restored. Note also the work areas in front of the tent.
had been placed. The tentsites from Sanaydin shown in Fig. 9.3a do not differ markedly in dimension, form or feature from those in Fig. 9.3b, the remnants of a black-tent encampment some 10 kilometres to the southeast. Strong regularities in the internal organization of tents of radically different form extended to the stone huts employed by a number of families at Sanaydin Yayla. These were of sturdy construction with rough stone walls about 1.5 metres high and roofed with beams and crosspieces
Nomads in archaeology
168
plastered
3m
reeds 0
1
i......... i
horn core—* * goat skull
«i
3m
169
Sanaydin Yayla
/7/ h U\J A
Fig. 9-3 Hoojplans of tents from sites within a few kilometres of each other at the head of the Goksu Valley. Note the similarity in floorplans although the tent superstructures are based on entirely different principles. (a) Plans of alaaks at Sanaydin Yayla. (b) A nearby camp of black tents with floorplans of similar dimensions and comparable features.
C
KEY hearth
c
chicken box
xxx thornbush i^mJ midden /fy
cache of tentpoles
topped with straw or reeds and compacted earth. In a few cases a central log support was employed and a wooden door attached by wire to the entrance. Yet in terms of the overall dimensions and the arrangement of fixtures, features and furniture both inside and out, the interior of a hut differed not at all from that of an alaak. Like the houses and tents observed by Watson (1979) in western Iran and the tent footings and mud huts observed in Baluchistan (Audouze and Jarrige 1978, 1981), the removal of above-ground structures would leave dwelling sites indistinguishable in size, shape and overall layout, regardless of the type of tent or walled structure that had surmounted them.
Nomads in archaeology
170
The domestic complex Looking down on the campsite from the highway above I was immediately struck not only by the regularities in layout of each tent complex but also by consistencies in orientation, alignment and spacing (Fig. 9.4). This high degree of redundancy observed in the construction and layout of the dwellings themselves and associated living and work areas may be formalized in terms of a spatial model. Each individual dwelling and its surrounding living space may be schematically representated by a domestic complex such as that illustrated in Fig. 9.5a. The entire unit occupies a rectangle or rhomboid approximately 20 metres by 10 metres divisible into two roughly equal subunits - a living zone and a discard zone. The living zone may be further divided structurally into the dwelling, forecourt and features, and functionally into a number of domains as illustrated in Fig. 9.5b. The discard zone, it turns out, is similarly structured, its features bearing a consistent relationship to the complex of tent and forecourt. A major channel of activity occurs along the axis of the complex, from tent entrance to external hearth and beyond to the ash dump and wood-chop area, with kindling wood going into the living zone and charcoal, ash, food scraps and domestic refuse coming out. The entire domestic complex may be represented in terms of two intersecting axes, the short a axis defining the orientation of the tent or dwelling and the longer b axis defining the extent and orientation of the complex as a whole. Ideally the axes will intersect in the general vicinity of the internal hearth and be approximately perpendicular. The actual orientation of the axes and their angle with respect to one another will depend largely on the subtle interaction of slope and terrain. Normally the b axis, defining the domestic complex, will be located in the direction of slope, and the a axis along the contour line. However, when freed from the constraints of terrain, the ideal configuration will be an east-west orientation for the a axis (i.e. the dwelling) and a north-south alignment along the b axis, with the living zone located in the northern sector and the discard zone in the southern. The actual alignment and angle of intersection of the axes will depend on the interaction of factors of slope, terrain and aspect, with the last of these being the main one under ideal conditions, but otherwise largely subordinated to constraints imposed by the first two. Before proceeding to a consideration of the implications of these ground rules for the overall structure of the campsite it is necessary to turn to the internal organization of the domestic complex. If each axis is envisaged in terms of a section through the domestic complex, then it is apparent that much of the architectural and locational information within the unit occurs along these axes. In other words an archaeologist wishing to excavate such a unit would be just as well served by a couple of wellplaced 1 metre trenches following the two axes as by a full exposure of the entire area. But, bearing in mind that there may be no a priori grounds for determining the precise location of either axis, the initial problem becomes one of being able to locate oneself accurately in terms of the domestic complex model, whatever piece of the complex may be initially encountered.
171
Sanaydm Yayla
Fig. 9-4 General view of Sanaydm Yayla looking across Area A. Note the consistency in orientation of tents and organization of living and work space.
•*•« • ''it
AL
•'
*-jM
vr.
In view of the archaeological objectives of the investigation it is necessary to ask ourselves the following questions: (a) to what extent is spatial patterning in the living zone of a particular domestic complex related to spatial patterning in the corresponding discard zone ; and (b) how might such patterns be detected archaeologically?
Nomads in archaeology
Fig. 9-5 The domestic complex model. The a-a1 axis defines the orientation of the tent while the b-b1 axis defines the main axis of activity space. (a) Formal composition of a stylized domestic complex. Domains include: A: dwelling B: forecourt C: external hearth D: external storage E: wood-chopping area F- ash dump G: midden. Structures include: i: walls ii: hearths iii: storage structures iv: plastered floors v: stamped earth floors. (b) Activity areas superimposed over the domestic complex. (M=men; W=women; MW=men and women.) Activities include: 1: sleeping (MW) 2: eating (MW) 3: cooking (W) 4: weaving 5": general domestic chores (W) 6: churning (W) 7: chopping wood (MW) 8: milking cow (W).
>— — —
———-——
r
——-i-.
172
_ _ _^
n//////Ai'////}/A^%
Wk t . . fi
/III/j
ai LIVING ZONE
a
LIVING ZONE
»»* ,' %•''":" V • ; *.
DISCARD ZONE
t:•&*£ h v i n g :%*•••*•.• f | o o r
DISCARD ZONE midden
|
H ^ l ash ..v.-.-.v.
U1
deposit
Y////{s\orage $&%: w o o d ? I
3.v£*3 f i v i n g •.%•*••*•.• f | o o r
? ? 1 5 M £ storage
Quantitative analysis of discard Conceived in terms of a site structure model; with the domestic complex as the operational unit, the archaeological problems involved in detecting and analysing an open settlement such as Sariaydin Yayla are thrown into a new light. It soon became clear that detailed observations of daily life and activities, such as those prompted by a functionalist perspective (Bonnichsen 1973; Binford 1978a; Longacre and Ayres 1968; Watson 1979; Murray and Chang 1980); were of limited potential in answering the kind of questions I wished to address. In carrying out a surface recording of selected areas of the site I was not so much concerned with establishing a link between observed artifact distributions and past activities as in determining the scale, density and texture of artifact and refuse scatters and their location vis-a-vis the dwellings and features. My aim was to explore the possibility that the distribution of artifacts on such a site might bear a close relation to the pattern of features and structures summarized in the domestic complex.
Quantitative spatial analysis: Area A Surface recording using a 5 metre grid over a 50 by 50 metre area of the site, designated Area A, revealed that, while there was a thin background
Sanaydin Yayla
Fig. 9.6 Plan of Area A showing domestic complexes A and B and recording zone A.
173
AREA A & RECORDING ZONE A
SV-1
#?•
ash dump
**
wood pile
#
tracks
<£~> stoney area c
hearth
scatter of items over the entire area; all item classes were strongly clustered, the highest concentrations occurring within the discard zones of the two major domestic complexes contained within the area (Fig. 9.6). The density of items in the living zones was generally no higher, and often lower, than in the areas outside domestic complexes. Of the four dwelling sites within Area A, only two (nos. 2 and 3) possessed well-defined discard zones. Dwelling site no. 6 appeared not to have been used for some considerable time, while no. 8, a hut site, was located at the junction of two arms of the stream so that refuse was dumped directly into the gullies to be washed away with the first downpour. Tentsites nos. 2 and 3 had been occupied during the summer of 1980, but remained abandoned throughout 1981, ensuring the survival of much surface material which had had a chance to settle after the last winter snow cover and spring deluge. From a practical viewpoint, the area offered advantages unavailable in those areas of the camp which were still occupied. Although the daily passage of Mustapha Bey's 200 strong flock in and out of the adjacent corral precluded the laying out of a full grid system, with the aid of a few cunningly-placed datum pegs and a collapsible 1 metre recording frame, I was able to pursue my eccentricities without intruding unduly into day-to-day household activities.
Nomads in archaeology
174
• .. •
J . • ** • ** • *• .** *
*
5
•
a bone 1
1 •
f nihhprl.
/
1
. . ••
• •
.*. •• •• • ••. *
h
b pottery 1 • •
...
J ••
c a lass |. -
J • ••• ••
i
•
g charcoal L wood
• • * •
.••
• • * • • > • "
h fextilel / , -'
:.
r1 •
•
•
s
•
4
• <
•
•
<
• • •
• d plastic 1 •
e metal 1 DURABLES
Fig. 9.7 Distribution of items over recording zone A, Sanaydin Yayla. (a) animal bone; (b) pot sherds; (c) glass sherds; (d) plastic scrap; (e) metal scrap; (f) rubber scrap; (g) charcoal and wood scrap; (h) textile; (i) binding (string, cloth); a; b, c, d and e are classed as durables, f, g, h and i as perishables.
bindina [ • PERISHABLES
Within Area A the discard zone associated with tentsite no. 2 was selected for more detailed study. All surface material was recorded by location and item class within a 1 metre grid system. The main line of investigation here was the extent to which the distribution of materials in the recording zone - roughly coinciding with a discard zone - was structured both internally and in relation to the features contained in the living zone. Very little information can be gleaned from the raw distributions of item classes (Fig. 9.7) other than that some materials, notably glass sherds, rubber and plastic, tended to concentrate in the upper part of the recording zone nearest the tentsite, while others, particularly textile and binding materials, were scattered in a more even and 'random' pattern throughout the area. Nearest neighbour analysis of each distribution confirms this observation (Table 9.1) with glass sherds exhibiting the strongest tendency towards clustering. However certain distributions appeared to be closely associated with each other. The underlying trends in spatial patterning were most effectively brought out by means of multivariate analysis. Principal components
Sanaydin Yayla
175
Table 9.1a Nearest neighbour statistics for distributions of items, recording zone A, Sanaydin Yayla, indicating the degree of clustering. Only glass shows a significant degree of clustering Distribution
r-value
Number
Bone Potsherds Plastic Glass sherds Metal scrap Rubber Textile scrap
0.853 1.173 0.869 0.671 1.117 0.831 0.818
118 8 83 31 18 34 113
The r statistic ranges from 0.0 (maximum clustering) through 1.0 (random distribution) to 1.4 (systematic arrangement) Table 9.1b
Bone Pottery Plastic Glass sherds Metal scrap Rubber Textile scrap
Test for segregation between pairs of item distributions, Area A, Sanaydin Yayla. Most pairs of distributions display a random association only Bone
Pottery
Plastic
0.066 0.002 0.177 -0.075 -0.059 0.219
-0.089 -0.024 0.224 0.189 0.091
0.226 0.024 -0.069 0.220
Glass sherds Metal scrap
0.040 0.176 -0.073
-0.054 -0.073
Rubber
Textile scrap
0.197
The S-statistic ranges from 1.0 (maximum segregation) through 0.0 (random association) to -1.0 (complete association)
analysis revealed two main complexes of items, one dominated by plastic and rubber together with moderate loadings in glass sherds, metal and animal bone, and the other dominated by cloth and binding material (see Table 9.2). The former may be regarded as general discard from tentsite no. 3 ; while the cloth; wool and string are probably the result of shearing, an activity observed earlier in another part of the site. Cloth or string is used to tie the animals' legs during the shearing operation. The distribution of the components displayed a high degree of spatial coherence, with Component 1 located in the upper discard zone and Component 2 in the
Table 9.2a Principal components analysis of recording zone A, Sanaydin Yayla - correlation coefficients for durables and perishables Perishables
Durables Pottery Glass Plastic Metal Rubber Textile Binding
Bone -0.0694 0.4331 0.4907 0.1738 0.5826 0.1423 0.1226
Pottery
Glass
Plastic
Metal
Rubber
Textile
0.1549 -0.1213 -0.0894 0.1730 -0.1020 -0.0246
0.4020 0.3763 0.4217 0.1767 0.1329
0.6192 0.7239 0.2354 0.0803
0.3483 0.1590 0.0243
0.0961 0.1071
0.4131
Nomads in archaeology
176
Table 9.2b Principal components analysis of recording zone A, Sanaydm Yayla - compound loadings on variables
Bone Pottery Glass Plastic Metal Rubber Textile Binding Eigenvalue % variance
All materials Principal components 1 2 0.494 0.006 0.000 0.145 0.458 0.009 0.751 0.007 0.409 0.005 0.664 0.068 0.130 0.525 0.063 0.531 2.97 1.35 37.14 16.85 Binding Plastic Rubber Textile Bone
Durables Principal components 1 2 0.004 0.476 0.865 0.007 0.510 0.174 0.737 0.027 0.538 0.028
2.26 45.30 Plastic Metal Glass
1.10 22.05 Pottery
lower area (Fig. 9.8a). A similar result was obtained through cluster analysis (Fig. 9.8b). With the exclusion of perishable items Component 2 and cluster group C disappeared, leaving the distribution of the remaining units relatively unchanged and indicating the selective effects of differential preservation. As with Ali's camp, the overall spatial pattern that emerged may be summarized as a series of broad arcs around a focal area coinciding with the ash dump. More than any particular spatial pattern or association between items, the analysis revealed the presence of spatial structure within the midden deposit, a structure which is highly suggestive of the general domestic complex configuration of which it is a part. Assuming no knowledge of features outside the study zone, the discovery of such a pattern might serve to alert an investigator to the possible presence of a living zone and tentsite in the area peripheral to the focal zone of concentration of items.
Spatial analysis of two complexes: Area B
Having established that it is possible to detect the presence and structure of a discard zone, the next question is: given an undifferentiated area of midden deposit, is it possible to distinguish between two or more discard zones emanating from separate dwellings? Area B, located in the southern part of the campsite, contained two adjacent tentsites and a swathe of refuse along the slope below. While the discard zones of the two tents could be easily distinguished by the separate distributions of wood scrap and ash, no such pattern was immediately apparent in the distribution of other items. None of the distributions displayed marked tendencies towards clustering and most pairs of items exhibited no more than a random association, indicating a comparative lack of structuring in the data. The application of multivariate analysis again confirmed its ability to
Sanaydm Yayla
177
Fig. 9-8 Results of principal components and spatial clustering for recording zone A, Sanaydin Yayla. (a),(b) and (c) Plots of scores on principal components 1, 2 and 3 over recording zone A. Components 1 and 2 show a high degree of spatial coherence. (d) Spatial clustering of 8 distributions across recording zone A (Ward's method on raw data). (e) Spatial clustering on durables only.
+ Scoras
- SCOTM
0 -1--2 0-2-3 (H) o-i
1 bone pottery glass plastic metal
rubber
textiles binding
''3 3
b^ 2j 4
:8=8r
3 { 4
3 ,'
I, • i l l
fS:::S:\
n mm
•5*::: 5*:.
d
3)
• A—a-
^^
bone pottery glass plastic metal
m
1
'2
: 5 : : '.6'-"5::''7.". 2)
3
3/ 1 ! 3
::5:
-.6:: :7:::5--'7:- p :5:
e GROUP
:5 :*::*5:::5V: :::*: •.*.•.".:\v3
|§]C
extract intuitively meaningful and spatially coherent patterns from very low-resolution data (Table 9.3). In each case; for both all materials and durables alone, spatial plots of both principal components and clusters revealed clear segregation into two discrete complexes, each of which could be linked to an associated dwelling structure (Fig. 9.9).
Comparative spatial analysis of Areas A and B
Recording zones A and B differed substantially in their respective proportions of animal bone as opposed to other remains and also in the relative
Nomads in archaeology
178
Table 9.3a Principal components analysis for item distributions in Area B, Sanaydm Yayla - correlation coefficients for durables and perishables Durables
Bone Pottery Glass Plastic Metal Rubber Textile Binding
Perishables
Bone
Pottery
Glass
Plastic
Metal
Rubber
Textile
0.1437 -0.0572 0.3397 0.1579 0.1226 -0.0605 0.0221
0.2160 -0.0519 -0.2230 0.1326 -0.2127 -0.1391
0.0229 0.0292 -0.0267 0.0112 0.1914
-0.0140 0.2869 0.2497 0.0564
0.2001 0.1825 0.0004
0.3230 0.1812
0.0239
Table 9.3b Principal components analysis for item distributions in Area B, Sanaydm Yayla - component loadings on variables All materials (8 variables)
Durables (5 variables)
1
Principal components 2
3
Bone Pottery Glass Plastic Metal Rubber Textile Binding
0.176 0.281 0.032 0.324 0.212 0.452 0.345 0.096
-0.172 0.912 0.491 0.192 0.007 -0.023 -0.000 0.211
-0.506 0.001 0.089 -0.130 0.156 0.023 0.267 -0.157
0.496 0.371 0.059 0.307 0.257
-0.201 0.256 0.373 -0.293 0.097
Eigenvalue % variance
1.92 24.00 Rubber Textile Plastic
1.29 16.12 Glass
1.11 13.90 Bone
1.49 29.71 Bone Pottery Plastic
1.22 24.31 Glass Plastic
Principal components 1 2
proportion of durables to perishables (Table 9.4). While the higher proportion of perishables in Area B may be a result of its being only recently abandoned as compared with tentsite no. 3 which had been unoccupied for a year, the disparity in faunal remains may well reflect an economic reality. According to Mustapha Bey the households occupying Area B maintained only small flocks for domestic use and which were integrated with the flocks of more pastorally oriented households for herding purposes.
Bone and artifact distributions: Area A Although multivariate analysis was not attempted for the whole of Area A; the locations of the discard zones of tentsites nos. 2 and 3 were broadly consistent with the density contours for each material (Fig. 9.10). However the distribution of faunal elements in Area A was complex. As areas
179
Sanaydm Yayla
Fig. 9- 9 Results of principal components and spatial clustering for recording zone B. Note the replication of both component score contours and cluster groups in two parts of the study zone. (a) and (b) Plots of scores on principal components 1 and 2 over recording zone B. (c) Spatial clustering of 8 distributions across recording zone A (Ward's method on raw data).
0:1
m 1-2
2-3
a all materials scores 0to-1
-1to-2
b durables
(f'i'l ?___*_ Iw^^^ c all materials
Table 9.4a
Comparisons between recording zones A and B in terms of proportions of items Recording zone A
Material
Bone Pottery Glass Plastic Metal Rubber Textile Binding
Recording zone B
(A/)
(%)
(A/)
(%)
140 8 35 88 20 38 135 52
27.13 1.55 6.78 17.05 3.87 7.36 26.16 10.08
27 10 8 40 15 11 125 52
9.37 3.47 2.78 13.89 5.21 3.82 43.40 18.05
516
100.22
288
99.99
Nomads in archaeology
180
Table 9.4b Phi-square tests of significant difference between recording zones A and B for item classes Durables 291 56.4
Recording zone A Recording zone B (Exp) % (Obs)
£=55
162.4 100 0 2 = 0.1910 Bone
Recording zone A % Recording zone B (Exp) % (Obs)
J
I
H
6
hZ 3-5
125.6 180
288
Non-bone
Total
376 72.9 209.9 261
516
140 27.1 78.0 27
288
0 = 0.1589
K
F
6-10 1-5
J
I
H
6
F
E
-I
2. nuclear area 2 3- nuclear area 3
0
Total 516
2
= 45.78
K
Perishables 225 43.6
-2
0
C
16-20 Zh
Items per 5m1 quadrat
BONE Fig. 9.10 Contoured distributions of items in Area A, Sanaydin Yayla, by 5 metre quadrats. (a) Density of animal bone in terms of counts per 5 square metres. (b) Combined density of all recorded artifacts by counts per 5 square metres.
ALL ARTIFACTS (metal, pottery, tentc loth, glass)
of high density occurred well outside the two discard zones the faunal elements were grouped according to anatomical classes in order to detect any further patterning (see Fig. 9.11). Apart from the discard zones of the two domestic complexes; three major concentrations could be identified, two to the south of domestic complex 3 and one to the north, near the entrance to the corral. These have been designated 'special areas7. The counts and percentages for each area are shown in Table 9.5, and it appears that the special areas, particularly area b, differ in containing a
181
Sanaydm Yayla
Fig. 9-11 Distribution of faunal remains in Area A, Sariaydin Yayla. Five major concentrations are detectable, two of these coinciding with discard zones 2 and 3. The others appear to be special activity zones (a, b and c). Zone (a) was a recent butchery site where two goats were prepared for the kurban bayram feast.
a
cd
SV-1.6/* . 1 .. C
/
^
A
tro •
("• •'''
*"• xx:
•
* • '•
B
D Head (skull,horn, horn cores) o Mandible/maxilla x Loose teeth * Body (vertebra,scapula, pelvis,ribs)
i Upper leg (humerus,rad, ulna,fem, tab tib.fib) • Lower leg (metapod., astrag7calc.,phal.) • Longbone fragment • Bone scrap
A
a o . ** Sr
Table 9.5 Faunal elements (ova-caprine, bos, eauid) in Area A, Sanaydm Yayla - frequencies and percentages of anatomical classes in different zones of Area A (see Fig.94)
Domestic complex
(AO (%) Head Body Upper leg Lower leg Lone bone Scrap
Total
16 20.5 2 2.6 2 2.6 2 2.6 9 11.5 47 60.2 78 100.0
Domestic complex (AO (%) 29 22.6 5 3.9 13 10.2 10 7.8 14 10.9 57 44.5 128 99.9
Total
Special Area a
Special Areab
(A0 (%) (AO (%) (AO (%) 45 21.8 10 27.0 14 46.7 7 3.4 6 16.2 2 6.7 15 7.3 3 8.1 0 0.0 12 5.8 2 5.4 0 0.0 23 11.2 5 13.5 1 3.3 104 50.5 11 29.7 13 43.3 206 100.0 37 99.9 30 100.0
Head (skull, horn, horn cores, mandible, maxilla, loose teeth) Body (vertebra, scapula, pelvis, ribs) Upper leg (humerus, radius, ulna, femur, tibia, fibula, patella)
Special Total Other Areac areas (AO (%) (AO (%) (AO (%) 9 60.0 33 40.2 19 26.0 0 0.0 8 9.7 6 8.2 4 26.6 7 8.6 1 1.3 0 0.0 2 2.4 2 2.7 0 0.0 6 7.3 3 4.1 2 13.3 26 31.7 42 57.5 15 99.9 82 99.9 73 99.8
Grand total (AO 97 21 23 16 32 172 361
Lower leg (metapodials, astragalus, calcaneum, phalange, etc.) Longbone (longbone fragment) Scrap (bone scrap - unidentifiable)
(%) 26.9 5.8 6.3 4.4 8.9 47.6
99.9
Nomads in archaeology
182
higher proportion of head elements (mandible, maxilla, horn cores) than in the discard zones. The discard zones contained a significantly higher proportion of bone scrap (see Table 9.6). There is some indication that the so-called special areas represent butchery events. This was known to have been the case for special area a, where a few goats had reportedly been slaughtered for Kurban Bayram, marking the end of the holy month of Ramazan, a few weeks earlier. Mustapha Bey explained that animals were always slaughtered away from the living areas. Special area c, with the highest proportion of head elements and the lowest proportion of scrap, may represent an earlier butchery event. Table 9.6 Chi-scjuare tests for significant difference between domestic complexes and special areas, in terms of faunal element classes a. Identified bone v. bone scrap, domestic complexes and special areas Identified
Strap
Total
Domestic complexes %
102 49.5
104 50.5
206
Special areas (Exp) (Obs)
(40.6) 56
(41.4) 26
82
Scrap
Total
57 44.5
128
2
X = 11.57 If. = 1 Significant at 0.001 level b. Identified bone v. bone scrap, domestic complexes 2 and 3 Identified 71 55.5 Domestic complex 2 (Exp) (43.3) (Obs) 31 X2 = 7.85 d.f = \ Not significant at 0.001 level
Domestic complex 3 %
(34.7) 47
78
c. Head and lower leg, v. body and upper leg, domestic complexes and special areas Head and lower leg
Body and upper leg
Total
Domestic complexes %
57 72.2
22 27.8
79
Special areas (Exp) (Obs)
(36.1) 35
(13.9) 15
50
= 0A2 d.f = \ Not significant at 0.05 level
Some implications for archaeological investigation Suppose an archaeologist happened upon a site such as Sanaydin Yayla. Assuming that features other than the ruined caravanserai and bridge were noticed, the most likely result would probably be the recording of a thin surface scatter and stone foundations and an immediate loss of interest in favour of other sites with more interesting stratigraphic contexts.
Sanaydin Yayla
183
But just suppose that the archaeologist is intrigued by the site or is vaguely aware of the possibility that he or she is dealing with a pastoral site. How might the archaeologist go about investigating such a site? What kind of excavation or surface recording strategy might be pursued? Techniques for the investigation of extensive shallow sites have been discussed at a general level in The Early Meso-American Village (Flannery (ed.) 1976), but a more pertinent example is to hand in the form of Tepe Tula'i in southwest Iran, which may be regarded as the first test case of nomad, or at least pastoral, archaeology in the Near East (Hole 1974-75, n.d.). Where little or nothing is known in advance about the nature of a site, and resources of time and money are insufficient to undertake a more leisurely investigation, the archaeologist's knee-jerk reaction is generally to put in test pits or 'telephone boxes' in order to see what turns up. Such was the case at Tula'i, where constraints of time left little leeway for detailed planning. Out of a series of telephone boxes, two hit what appear to have been midden zones, while three others were placed within what appeared to be the stone foundations of tentsites showing up on the bulldozed surface (see Chapter 11). It will be clear from the above discussion that, if indeed Tepe Tula;i was a nomad or pastoral campsite, as was suspected by the excavator, the last place one would locate test pits in order to locate ceramics and household debris would be in the tentsite itself. By the same token the location of tentsites linked to the suspected midden zones uncovered in another part of the site might have been deducible on the basis of trends in the orientation and associated features of the tentsites already excavated. What might at first appear as a series of discrete and unrelated occurrences may well come together as a coherent pattern whose details differ slightly from the particular instance of the domestic complex outlined above, but whose general dimensions and structure can be fitted to the model. As a practical device for the detection and elucidation of nomad campsites, the domestic complex model may prove useful. But in view of the uncertain future of nomad archaeology in a Near Eastern context, its more general applicability needs to be considered.
The domestic complex in context Given that the primary analytical units that have emerged in relation to the domestic complex are structural rather than behavioural - in the strict sense (Schiffer 1976) - we are led to inquire as to the role of the domestic complex in the overall structure of the site. From the form of the domestic complex itself two principles may be derived. First, domestic complexes are discrete spatial units and tend to be well spaced in relation to one another. In nomad camps, one suspects, residential units are so placed that each household is able to maintain sufficient space not only for a full range of household activities to be carried out but also to preserve a sense of spatial integrity with regard to other units. Second, given the complex and asymmetrical nature of the unit - in contrast to those derived by Yellen (1977) in the study of the !Kung - it is clear that, above a certain density of settlement, the juxtaposition of units in relation
Nomads in archaeology
184
to one another must be highly constrained. The most economical settlement pattern might be envisaged as a series of rows of tentsites arranged end to end in parallel lines, the lines of tents being separated by at least 20 metres and the tentsites in each line separated by 5 to 10 metre intervals. This tendency has been remarked on by a number of observers (Barth 1961; Digard 1975; Hole n.d.). On a perfectly flat and featureless plain - of the kind not often encountered by nomads observed in the course of this study - one might expect campsites to build up through accretion in such a manner, the initial orientation of rows being a random matter. But campsite structure is seldom free of other environmentally imposed constraints. Indeed, as indicated in Chapter 7, campsites are initially selected according to a stable set of favoured environmental criteria. One such factor which is relatively free of location is that of aspect. As can be observed from a glance at the Sanaydin Yayla camp (see Fig. 9.1), other things being equal, the preferred orientation of tentsites appears to be consistently along an east-west axis. That is, where other factors such as slope do not intervene, such an orientation occurs. By extension the orientation of domestic complexes under such conditions will be roughly north-south. Depending on the desirability of reducing or enhancing absorption of the sun's rays by tentcloth or stone foundations, the direction of such preferred orientation may be expected to vary both seasonally and with altitude (Hole 1978; Be§ikgi 1969; Baharvand 1975). In the case of Sanaydin Yayla, entrances and forecourts occurred generally on the south side of the tent in order to take advantage of the sunlight which, in these latitudes, always comes from the south. The preference for sheltered locations on gently sloping ground in the lee of steep hillslopes brings into play the factors of intensity and direction of slope and contours. Although a particular orientation may be preferred, this will nearly always be overridden by factors linked to terrain. On slightly or moderately sloping ground the tentsites are located along the line of contour, with the front or forecourt on the downhill side. Only where slope is negligible, as in the cases of tentsites nos. 3, 4, 5 and 9, does the tent begin to rotate away from the line of contour towards the preferred east-west axis. Direction of slope becomes an increasingly overriding factor as the steepness of slope increases, while an extremely steep slope will tend to preclude the establishment of a tentsite altogether. Thus a large part of the spatial structure of Sanaydin Yayla may be deduced from the interaction of the environmental features and settlement constraints discussed above.
10 THE LOST WORLD OF NEMRUT DAG Yet not a single Kurdish tent, no shepherd, no wayfarer can we descry in the wide landscape of the volcanic basin. We observe paved holes in the ground where it is evident that bread has recently been baked. There are stone enclosures for penning cattle. More and more clearly we realize that the crater must be inhabited and that this floating population have decamped at the approach of the soldiers. (Lynch 1901, pp.303-4)
The setting Having climbed the 2,000 metres or so from the shores of Lake Van to the rim of the crater of Nemrut Dag, the traveller is treated to an awesome sight. Below stretches a huge basin surrounded on all sides by precipitous walls. The interior is a tumbled chaos of conical hills, lava flows, depressions and jagged outcroppings of rhyolitic rock and obsidian, its western half drowned by the icy waters of a large semicircular lake. Nemrut Dag is a collapsed caldera, the remnant of a lofty volcano which emerged from the floor of the plateau to dam up the waters of Lake Van to the east. It is one of a chain of volcanic eminences lying to the north of the folded ridges of the Taurus Mountains - Ararat and Siiphan to the east, Bingol, Erciyas and Hasan Dag further west. On descending into this lost world, the encircling mountain rim closes off the outside world leaving only the barren moonscape of stony ridges, scree slopes and flat internal drainage basins of alluvial ash occupied by shallow lakes. The floor of the crater is an almost perfect circle nearly 8 kilometres in diameter, lying at 601 metres above the surface of Lake Van (see Fig. 10.1). Near the centre of the caldera is a conical hill truncated by a deep crater and surrounded by lava flows thrown up in the last recorded eruption during the fourteenth century (Yalginlar 1972-73). Much of the rubble strewn around the floor of the caldera is a form of low-grade obsidian or volcanic glass, mottled with white flecks. High in the surrounding walls, where jets of lava were spewed out and supercooled when the volcano collapsed inwards upon itself, are to be found deposits of high-grade obsidian of a pure, lustrous black or dark emerald green (Renfrew et al. 1968; Wright 1969). From Palaeolithic through Neolithic and Early Bronze times this precious material, used to manufacture stone tools with a fine cutting edge, was quarried and traded as far afield as the early Neolithic site of Ali Kosh in southwest Iran (Hole et al. 1969). 185
Nomads in archaeology
Fig. 10.1 Interior of the volcanic caldera of Nemrut Dag, eastern Anatolia. The locations of campsites studied are indicated, together with the possible location of the campsite at which I§mael Be§ikc,i resided in the 1960s. (After Yalgmlar 1972-73, fig.14.)
186
A
KEY Campsites mentioned in text Approximate location of Be^ikci's zoma, 1965
Lava flow
Two rough tracks, fit only for four-wheel-drive vehicles, cross the eastern rim of the caldera where it dips to its lowest point, and thefloorof the crater is criss-crossed with tracks and pathways. Earlier travellers were greeted by a similar scene. In the 1890s a British geologist, interested in the vulcanology of Nemrut Dag, mounted an expedition into the crater under heavy escort provided by the Ottoman Sultan: But what is the meaning of these many paths which seam the interior, arguing a considerable traffic to an fro. Are there villages in the crater? We have never heard of any, we are assured that none exist. Not a fire, no light is anywhere visible, but the tracks are broad and have all the appearance of being regularly used. We feel surprise and express it to the Kaimakam. He answers us naively that Kurds come here now and then. (Lynch 1901, p.301) Contemporary writers such as Williams (1972, pp.124-5) also report that the crater is uninhabited except for a few 'Y6riiks; (sic) who graze their flocks there in summer. Nemrut Dag is indeed an inhospitable place for human settlement. Summer temperatures fall between 20 and 25° while winter temperatures never exceed 0°. Precipitation of 300-400 mm per annum on the plateau and 1,000-1,500 mm on the higher slopes occurs chiefly in winter as snowfall which covers the interior of the crater to a depth of 1 to 1.5 metres for a period of up to 130 days. As early as the end of September the passes into the crater may be blocked by heavy snowfall (ibid., p.124), and the more concealed crannies in the southern wall contain snowdrifts
NemrutDag
187
which never melt. The highly acidic volcanic soils are unsuited to agriculture as is the rocky nature of the terrain. Only during spring and early summer do the crater and its external slopes experience a brief explosion of pasture growth fed by unlimited supplies of melt-water. It is the prospect of a few months7 rich grazing that attracts the nomads to Nemrut Dag every summer. Nemrut Dag holds special significance not only because it contains such a large concentration of campsites within a well-defined area; but because it was here that a young social anthropologist named I§mail Be§ikqi carried out what is perhaps the most intensive study of a nomadic community in Anatolia. As a young army conscript posted to the wilds of eastern Anatolia, Be§ikgi became acquainted with other conscripts who, it turned out; came from local Kurdish nomad tribes. He visited their campsites, extended his network of acquaintances, and gradually conceived the idea of carrying out a systematic study. As a graduate student at Erzurum's Atatlirk University, Be§ikgi returned to the field armed with notebook and questionnaires, and set up camp in Nemrut Dag with a large camp group of the Alikan A§iret (tribe). The social survey methods and tabulated results employed in the study are not every anthropologists idea of the ideal ethnographic method, but the range of information and detail of observations are exemplary. Unfortunately the work is little known - it is in Turkish. My own researches in Nemrut Dag were restricted by the extreme difficulty of access to the place. My efforts were not made any easier by the attentions of the huge, shaggy mastiffs with iron-studded collars which the nomads keep as watch dogs and for protection against wolves. More than one foray into a nomad camp had to be hastily aborted under their onslaughts. On other occasions the shepherds, having assured me that the dogs had been trained not to attack humans, were able to explain the purpose of various features and structures in the abandoned campsites while visits to the few occupied campsites provided glimpses of the same features in systemic context and were also memorable for displays of nomad hospitality. In the following account I shall attempt to link my own observations and uninformed interpretations of abandoned campsites to the dynamic perspective provided by Be§ikgi's (1969) account of Alikanli economic and social organization.
Initial observations In the course of my initial visit to the crater in late July 1978 my attention was caught by what appeared from a distance to be a deserted, stone-built village. The settlement occupied a small plateau on the edge of a rugged lava flow perched above a small briny lake. The following account is taken from fieldnotes made at the time: On a rise to the south of this [lake] is located what I will call an 'abandoned village7 - very recently abandoned at that. The walls are of stone and consist of foundations only, about 1 metre to 1.5 metres high, both circular and rectangular. They show obvious signs of very
Nomads in archaeology
188
recent human habitation - pieces of cloth, pieces of fresh goat skin, animal bones with traces of flesh; empty bottles . . . pieces of rope; string etc., and many broken shoes. As mere is not the slightest trace of any superstructure or roofing material, I assume that these are the 7 foundations for a 'black tent encampment of nomadic Kurds . . . Each 'house7 consists of two or three 'rooms7 and each has a large fireplace in the main room. Some fireplaces consist of besser blox transported especially for that purpose and there are many lids of tin cannisters that may have served as frying pans or fire guards. Many 'houses7 have large storage jars [sic] sunk into the ground just outside the front door. The general impression is one of comfort and some degree of permanence, or at least regular occupation. Obviously the inhabitants nave departed with their flocks for greener pastures a few weeks ago which makes their occupation late June or mid July. (Fieldnotes 26/7/1978) That such a settlement type has a respectable history in the caldera of Nemrut Dag is confirmed by the remarks of the geologist and explorer Lynch (1901, p.304) who was also puzzled by the abundant evidence for human habitation in such a wilderness. A second concentration of stone-built structures (ND-2, see Fig. 10.1), in a much worse state of disrepair than those described above, was discovered not far from here in a dry stream bed between two lava flows where the small lake mentioned above discharged into Nemrut Lake. The remains of at least 44 structures could be discerned, partly submerged in alluvial deposits. Large blocks of white, mottled obsidian and rhyolite were used in the construction. Surface material consisted of animal bones, particularly ova-caprine mandibles, and a few large, coarse ceramic sherds, some with interior glazing. A third site (ND-3), which appeared to consist of corral structures only, but may have contained disturbed tentsites as well, was observed in a small valley near the centre of the caldera. A fourth site (ND-4), observed in the distance at the base of a lava flow in the northwest of the crater, was still occupied. A single large black tent pitched over stone walls was visible and flocks of sheep were being grazed on the surrounding lava flows. To its west, abandoned tentsites (ND-5) were strewn at the base of the lava flow.
Interpretation of campsite ND-1 In late September 1981, when I returned briefly to Nemrut Dag, the crater was dry and desiccated, grazing having removed most of the grass, and the alluvial plains and shallow lakes shimmered in the heat. High in the crater walls outcrops of obsidian glinted in the sunlight. From the central cone the site designated ND-4 was seen to be occupied by a number of black tents. The flocks, strung out in long lines like an army on the march, raised clouds of yellow dust high into the air. ND-1, as before, was abandoned (Fig. 10.2).
Recording of ND-1 According to a passing shepherd the campsite had been constructed six years earlier but had not been used - at least by his own group - for the
Nemrut Dag
Fig. 10.2 Abandoned Alikanli campsite ND-1 from the north. Note the location of the campsite on a terrace at the edge of a lava flow overlooking a small lake. The lake level has been lowered during the summer months through evaporation and the watering of livestock. Fragmentary tentsites were observed on the small plain at upper right during July 1978.
189
-
• " .
~
:
:
•
• . > • —
:
~ '
•
-
:
;
"
^
;
;
last four years. Since in the short time available I was unable to glean much information about the demographic composition of previous inhabitants; the interpretation that follows is based entirely on the material remains themselves and will be checked later against information from Be§ikqi's account of the Alikanli. Although other tentsites lacking high stone walls, but with stone outlines, were to be found among the hillocks of the lava flow to the south of ND-1, I shall be concerned here only with the structures of the main camp shown in Fig. 10.3.
Dwelling structures Walls, constructed from low-grade obsidian and porous rhyolite boulders lightly held together with mud mortar, ranged from 1 metre to 1.5 metres in height, and were 50 to 60cm in thickness (see Fig. 10.3). Dwellings consisted of variable numbers and sizes of rooms together with partly enclosed forecourts, entrances and annexes. Most rooms possessed stone storage platforms running along one or more walls. Internal features included hearths, usually a simple structure of three rectangular stones, sometimes set into stone platforms. Some hearths were more elaborate, such as that recorded in dwelling 3 (see Figs. 6.8b and 10.3). This was a recessed hearth with a vent conveying the smoke up inside the wall itself, similar to the one in Ali's camp. The floors were of levelled and stamped earth and on some were to be found small stone circles or depressions marking the position of tent poles. Some rooms contained large rectangular grinding slabs, though a large grindstone present in 1978 had since disappeared. Large, coarse rim, base and body sherds were scattered about the floors of some dwellings.
External features Outside the dwellings were to be found more hearths, similar to those described above, in forecourts or kitchen annexes. The subterranean 'storage vessels7 are in fact bread ovens, and most dwellings had at least one, either in the forecourt or in a separate kitchen enclosure. In some cases a collapsed pit marked the place where a bread oven had previously been
190
Nomads in archaeology
situated, the vessel having either been removed or the broken sherds buried. Kitchen middens, wood scrap and ash dumps were found in association with all of the dwellings, but not with the corrals on the southern margin of the site. Other structures included small annexes for storing dung, and even smaller stone structures which seem to have served as chicken coops.
Fig. 10.3 Plan of campsite ND-1, NemrutDag, together with floorplan of tentsite no. 11 (inset). Note the location of middens relative to tentsites, the line of corrals to the south and the clustering of tentsites into contiguous groups. The camp has been divided up into zones and demographic units corresponding to households.
'"""' drain\
/
\
'
'
-
P
•
($£#
'
\ *
bovine I equid
('""'}
3g$ ALIKANLI CAMPSITE ND-1 NEMRUT DAG
>m
0
10
20
30m
e
hearth
©o
bread oven/pit
0
stone slab
i;V^->
midden deposit
?<<£. (V;
dung deposit site for guest tent
NemrutDag
191
Adjacent to some of the dwellings were circular patches of stamped earth with signs of a central posthole. These were the sites of small, white conical 'guest tents', in which guests or hired shepherds are accommodated. I had seen one of these freshly abandoned in 1978 with the green sapling tent pegs still in place. Among the most intriguing features were two latrines, solidly constructed out of stones to waist height, and situated strategically at the edge of the plateau. There was no cesspit - instead, an opening at the base of the rear wall was designed to convey waste matter down the slope and away from the settlement. The considerable effort that must have gone into these sanitary arrangements suggests that the site was intended to be used for long periods at a stretch.
Stratigraphy This is an unlikely term to be using in connection with nomad sites! But although no excavation was carried out, the walls of one of the breadoven pits displayed about half a dozen clearly stratified layers in section, some of them heavily carbonized (see Fig. 5.6). Since the pit occurred in a kitchen area, it is likely that successive years - perhaps generations - of activity could have built up sizeable deposits from the dumping of ash cleaned out of the bread oven or from the hearths and from food scraps, sherds and other debris mixed with the fine volcanic dust that blows across the site in summer. Geological processes were also operating on ND-1. Behind the line of sheep corrals at the southern margin of the campsite, deposits of alluvium, either washed down from the slopes above or carried by snow movement, had piled up against the rear walls to a depth of nearly 0.5 m. Eventually such deposits might be expected to cover the site to a considerable depth, leaving only the remnants of the stone walls above the surface. Such a site would be not only visible to the trained eye, but also stratified (both geologically and archaeologically), and should contain recognizable floorplans and diagnostic artifacts such as ceramics and bone.
The spatial analysis of ND-1 As already suggested in Chapter 9, the layout of nomad sites in modular units or domestic complexes should facilitate the analysis of demographic and social structure. From the siteplan shown in Fig. 10.3, ND-1 appears ideally suited to an experiment of this kind. On the basis of the location of structures and other features, the area was partitioned into seven zones (Fig. 10.3), one of which (zone G) contained only corrals. Each of the other six zones (A-F) was defined in terms of a spatially discrete building complex, for the most part containing separate but associated dwellings. A dwelling is defined as a continuously walled structure containing domestic features such as hearths or platforms. Altogether there were 11 dwellings distributed among six complexes. Most consisted of a number of rooms - not all of which contained hearths - together with annexes, passages and forecourts and associated features such as external kitchen facilities, guest tents and middens. This
Nomads in archaeology
192
Table 10.1 Estimated demographic composition ofND-1 based on the number of rooms and presence/absence of domestic facilities
Rooms
E
1 2 3 4 5 6 7 8 9 10
1 3 3 3 4 2 1 4 2 2
F
11
3 27
c D
Total
Ext. hearths
2 2 2 2
2 2 2 1
CSJ
Complex Dwelling
Int. hearths
2 0 1
2 1 4 3 2 3 25
0 1 3 0 14
Estimated no. of nuclear Estimated Middens Guest tents Corrals families population**
Slabs
Platforms*
Bread ovens
3 4 0 0 1
0 0 1 1 1 2 0 1 1 1
1 3 1 1 2 1 0 2 1 0
1 0 0 0 0 0 0 3 0 0
0 0 0 0 1
0 0
3 2 1 5 3 3 2 6 3 3
1 11
3 34
1 9
2 14
2 6
2 0 1
2 1 4 1 2
1 3 2 2 2 2 1 4 2 2
5 11 8 8 8 8 5 14 8 8
1 12
3 24
11 94
Average population per unit = 8.6 * Number of sides ** Population estimates based on 5 for the first nuclear family and 3 for additional nuclear families
information has been summarized in Table 10.1. Of the 27 rooms - not counting corrals - 24 contained hearths and/or other domestic facilities and may be considered as residential units.
Reconstructing the demographic composition of the community at ND-1 A major difficulty in making demographic inferences from archaeological data - particularly in the case of nomad sites - is that the site may have been successively occupied by groups of different sizes and varying composition. That this factor seems to have been important at ND-1 is evidenced by the changes in function undergone by a number of structural units. The corral structure G.17 had an external hearth and kitchen annex ; suggesting that it once served as a dwelling unit. Similarly the small corral annex attached to B.5 contained a few hearth stones. Dwelling unit A.3 had recently been used as a fodder store, presumably while empty during the latest occupation. In order to avoid such ambiguities, the demographic and social reconstruction of ND-1 will be based on the following assumptions: (1) The interpretation will be based only on information that is recoverable archaeologically, for example walls, hearths, bread ovens or pits, floors, middens, dung deposits. (2) The site will be regarded as representing a single and final occupation phase which has subsequently been undisturbed. As our concern is with population composition and social organization rather than with population size per se, estimates of population from floor area (Narroll 1962; Cook and Heizer 1968; Sumner 1979) are of doubtful utility. Another approach in which the presence of domestic and cooking
NemrutDag
193
facilities is used to infer the number of domestic units involved in an architectural complex (Kramer 1979a; Jacobs 1979; Watson 1979), seems to accord better with the purpose in hand. Controlled ethnoarchaeological studies in Iranian villages have shown that the number of hearths and other facilities can serve as a rough indicator of household composition: Built-in features such as bins, hearths, and ovens, while not reflecting variations in wealth in either their building materials or numbers in a compound, do reflect in a rough way the number of coresiding nuclear families, rather than the number of individuals within compounds. (Kramer 1979, p. 159) Hearths for baking bread are the only permanent fixtures used in cooking. Any redundancy of these [cooking] tools within compound walls woula imply more than one family unit. Indeed any more would imply a multi-family compound. (Jacobs 1979, p.188) The criteria by which an individual nuclear family is recognized in the present instance are as follows: (1) The presence of an internal hearth in a room indicates that the room contained a separate nuclear family. (2) Failing this, the presence of other facilities such as grinding stones or storage platforms may indicate that a room served as the abode of a nuclear family, provided that the room has access to a hearth elsewhere. The resulting classification, on the basis of dwelling units, is shown in Fig.10.4a.
Social organization Altogether 23 nuclear family units were identified according to the above criteria. Most of these could be further grouped into larger units corresponding to a whole detached dwelling, of which there were 11. It was hypothesized that these larger units represented joint or extended households. On the basis of the distribution of kitchen facilities, cooking appeared to be shared within the 'extended family7 and to occur within the forecourt areas in front of most of the dwellings. These contained the external hearths and ceramic bread ovens that were presumably used in common by the women of each extended household. The pattern of refuse disposal revealed by the location of midden deposits was specific to each dwelling unit, though some had more than one midden, while other middens seem to have been shared between households. The total of 14 external hearths, 11 grinding slabs, 9 bread ovens and 14 middens, although not evenly distributed between dwellings, is nevertheless consistent with the estimated presence of 11 units of consumption or households. Assuming the presence of kinship connections between households and that spatial distance between dwellings reflects the closeness of the relationship, it might be surmised that each of the six zones or dwelling complexes shown in Fig. 10.3 and Table 10.1 represents a minimal lineage segment of co-residing agnates. Additionally, or alternatively, it could be
Nomads in archaeology
Fig. 10.4 Hypothesized and actual social organization of Alikanli campsites. (a) Hypothetical reconstruction of population structure at ND-1 arrived at on the basis of numbers of 'household' units and spatial arrangement of dwellings. (b) Agnatic and affinal ties between household heads within a 37-tent Alikanli zoma at Nemrut Dag in 1965. A high level of agnation is indicated together with affinal links between agnatically based clusters. (After Be§ikc,i 1969, fig.4.)
194
LINEAGE
^7 MINIMAL
LINEAGE?
HOUSEHOLD
A
1 2
3
4
SEGMENT
^
X
B/'
C
A
A
A
X
7
8
9
5
6
A
\D
I
""E--
A
A
10
F A 11
NUCLEAR FAMILY
Mar^jff<( vvomon 0/
*% \
Ns T / Xx # X A'
-A KABILES OF THE ALIKAN AflRET • • D •
• O o A
Cudigan Cengovan Birivan Mehman
Neciman Ozigan ^ehidan ^ehevan
RELATIONSHIP
— brother^ ... father-son fa. br. - br. so. fa.br. sons at fines
hypothesized that each zone corresponds to a cooperative herding unit. Arguments against this second hypothesis are, first, that transient herding units are hardly compatible with fixed, subdivided stone dwellings of given size, and, second, apart from a few small corral annexes there is no indication of facilities for the penning of animals within or adjacent to each complex. Instead we find a localized concentration of corrals (zone G) which apparently served the whole settlement. However if the hypothesis of organization into herding units is accepted, the circular floorplans of 'guest tents7 attached to some of the dwellings can be explained as quarters for hired shepherds. The large dwelling complex in zone D, consisting of a single dwelling
NemrutDag
195
but containing at least four nuclear units, stands out as the largest household. Its central location and the presence of three 'guest tents' and four small animal pens; indicate that this must have been the most prominent household in the settlement, perhaps that of the headman or chief. The most economical interpretation of the social structure at ND-1 is that it represents a small hamlet consisting of 11 households, most of them extended or joint households, and 24 nuclear families (see Fig. 10.4a). The households or units of consumption are clustered in groups or complexes, each representing a minimal lineage segment. The whole settlement may represent a higher order lineage segment or a core of agnates together with associated cognates and affines. A rough figure for the total population of the settlement may be arrived at by assuming that: (a) the first nuclear family of each household contains five members, and (b) each additional nuclear family contains three members. This estimate yields a population figure of some 94 individuals and a mean household size of 8.4 individuals (see Table 10.1). There is a suggestion of some form of rudimentary ranking on the basis of relative household size and the number of facilities and ancillary structures.
A naive archaeological assessment of ND-1
The above interpretation, based on the evidence of fixtures and durable elements, makes no presumptions about the nature of the settlement or its economy. From the point of view of some future archaeologist, who having excavated ND-1 then attempts to interpret its significance, there would be little - beyond its location in such a harsh environment and its open ground plan - to indicate that it was anything other than a sedentary village. If deposits of animal dung had survived it would be possible to identify some of the structures as corrals, but the amount of corral space need not suggest that unusually large numbers of livestock were kept. At most the economy might be seen as having a pastoral bias. By the same token, the presence of grinding slabs, on the basis of conventional wisdom, would suggest the practice of agriculture. Although it is known that the structures of ND-1 were roofed with black goat-hair tents, it is difficult to envisage how this might have been achieved on the basis of the ground plans alone. These do not always conform to the regular, rectangular plan of the large tents used by the Kurdish tribes of this region. Each tent should correspond to an individual dwelling, but in the case of some of the more irregular and larger dwelling units shown in Fig. 10.3 the location of the portable superstructure can only be guessed at. In the absence of any trace of superstructure, and assuming that only wall stubs and rubble remain, the dwellings might easily be interpreted as permanently roofed stone-walled houses. Again the presence of fixtures such as subterranean bread ovens and coarse potsherds would reinforce this impression. In terms of their size, shape and mode of construction, the dwellings themselves would not be out of place in any east Anatolian mountain village. The presence of stratigraphic layers would also suggest a sequence of permanent occupations. In short an archaeologist chancing upon such a site as a surface manifestation or excavated occupation floor, and lacking any circumstantial evidence to the contrary, might easily be
Nomads in archaeology
196
led to identify it as a permanent - or at least seasonally occupied - agropastoral village settlement. Only the dispersed nature of the settlement plan, the absence of any organic structure and the pattern of refuse disposal in close association with individual dwellings might alert the achaeologist to the fact that he was dealing with an unusual kind of village.
The Alikanli: an ethnographic account It has been the practice among ethnoarchaeologists to attempt to confirm their interpretations of habitation sites by checking them ex post facto against information provided by the ex-inhabitants of the site, or first consulting with the inhabitants and later trying to confirm their account through observation or excavation (Bonnichsen 1973; Longacre and Ayres 1968). The present exercise offers an interesting variation on this procedure. We have a site and an interpretation of it. But it would have been difficult to locate and interview those who had used the site recently and, even if this were possible, memory of the occupation, who was present, their kinship affiliations, etc., would most likely have been hazy. What will be attempted here is to fit the organizational features of the Alikan tribe, as described in the ethnographic account, like a template against the demographic structure inferred from observations of ND-1.
Historical background The area around Lake Van has historically been one of political instability, a condition which may be partly attributed to its status as a frontier region caught between major political entities. After the Turkic invasions of the eleventh century AD there was a brief unity of Anatolia and Iran under the Seljuks, during which the caravanserai shown in Fig. 8.8a was constructed at the junction of the major east-west and north-south trade routes. Following the break up of the Seljuk Empire the region became a part of the Akkoyunlu Empire, itself a nomad dynasty (Woods 1976). Subsequently it formed an unstable buffer zone between the Ottoman Empire and successive Persian dynasties, becoming progressively depopulated during the conflicts of the Safavid period. Although migratory Kurdish tribes were undoubtedly present in earlier times, the unstable conditions of the period following AD 1,600 brought about the strengthening of tribal affiliations and an expansion of nomadic activity. The legacy of these turbulent times is still very much in evidence today in the area, which retains a 'Wild West' quality reflected in contemporary epic literature and cinema. The historical origins of the Alikan A§iret are uncertain, though the name appears in connection with the formation of tribal militias, the Hamidiye regiments, by the Ottomans in 1891 and a reorganization of tribal groups as mobile support units. This is reminiscent of the creation of tribes by government decree (Marsden 1978; Bruinessen 1979). According to Be§ikqi, the Alikanli were already in existence. In any case the creation of new tribal entities involved political alignments of existing
Nemrut Dag
Fig. 10.5 Nomadic migration routes in the Lake Van region. The migration route of the Alikanli appears not to have changed very much since early in the present century from which time a close association with the Duderan has also been maintained. (After Be§ikqi 1969, map 2.)
197
tribal segments and not necessarily the setting up of new groups of nomads. Some of the nomadic groups in and around Nemrut Dag reported by Lynch (1901) may well have been Alikanli, although it is far from certain what meaning this designation may have had at the time. An Alikan tribe is included in a list of tribes of the Ottoman Empire (Tiirkay 1979, p. 199) in the region south of Lake Van but, in line with the official denial of Kurdish ethnicity, is described as a 'Turkmen7 tribe. An inventory of Kurdish tribes compiled early in the present century (Sykes 1908, p.460) contains the following cryptic reference: 'Alikanli. 150 families. Nomads. Very insignificant. Probably a subtribe, but could not discover whose7. They are placed to the south of Lake Van, though their migration routes are not indicated. By this time, according to Be§ikgi (1969), the Alikanli should have been established as a tribe or a$iret in its own right, having branched off from its parent tribe, the Miran, late in the nineteenth century. They appear as a migratory unit in studies by Hiitteroth (1959) and Frodin (1943-44, fig. 13). The latter describes an Alikanli campsite similar in appearance to those I observed in and around Nemrut Dag. Today the Alikanli migrate between winter quarters in the region of Silvan, Bi§iri, Garzan, Siirt and Qzre and summer quarters to the south and west of Lake Van (see Fig. 10.5).
\ _ _ SIPKAN
BERCUNU
) KEY
— * " • •
ALIKAN Duderan Mehmediy2n 1 '• Ostukan —•-> Garisan ••«••> Soran Danudiya •> Batuyan ^ Kican = = * Tayan > Ispirti '•'> Zivikan Kitan
summer camp visited by BesjQkixt A summer camps D
winter quarters
o town
• MAP 9. Migration routes, S.E.Anatolia.
n
'
S /
/ //
0
(after Bes^ki 1969: map 2)
10
20
30km
Nomads in archaeology
198
Migratory cycle During the winter months the tribe is dispersed in small units of two to five tents pitched within or on the outskirts of villages to the south of the Taurus Mountains. Although the villagers have no tribal or kinship connection with the Alikan A§iret, the wintering nomads become a temporary part of the village community, drawing on its services and land resources and coming under the authority and protection of the village aga or kaymakam. The main consideration guiding the choice of location in winter is security from theft and violence and, to this end, longstanding relationships are established between particular nomad families and certain village communities to which the nomads habitually return. Since only limited grazing is available in the vicinity of each village, only a small number of nomad families can be supported and land is rented from the villages. Special winter tents are employed which are larger and heavier than those used in summer, enclosing both living space and sheep corrals. Other Kurdish tribes such as the Beritanli also employ such tents together with special tents to accommodate all the livestock of a herding unit. There is no mention of the construction of stone walls, though this seems highly probable by analogy with other southeast Anatolian Kurdish tribes and in view of the relatively stable location of winter campsites. Winter tents are left in the care of village neighbours during the migration to summer quarters. Winter is a time of subdued social activity for the nomads during which the main focus is on the maintenance of the flocks and gearing up for the spring migration. The change in season brings about not only a return to full productivity and more intense social activity, but also major changes in social organization. As previously scattered families begin to converge on the migration trail, lineage ties come into play and migration groups grow through accretion until the tribe is moving in large units - subtribes or oymak strung out along the route through Bitlis Pass. Here it comes into contact with sections of other tribes sharing the same migration route, heightening tension and strengthening the sense of tribal identity (see Siirt II Yilligi 1973, p. 148). In a very real sense the tribal structure, after lying dormant and in storage through the winter months - like the summer tents is reassembled from its constituents and revitalized as a functioning institution. During migration the family belongings are carried by mule - eight or nine mules per household - while the herds are driven on ahead of the baggage train. They are kept off the main roads and circle around towns and villages, while the tribespeople themselves pass through the settlements buying and selling in the local bazaars, visiting mosques, shrines, etc. Stops are made at transit camps on the way for a maximum of five or six days. Having passed the bottleneck of Bitlis Pass, the a§iret begins to disperse to widely scattered, rented summer pastures. The large migratory oymak units disintegrate into smaller camp groups comprising up to 20 tents, considerably larger than those formed in winter. They do not necessarily occupy the same yayla, or summer pasture, for the entire summer, but
NemrutDag
199
may move on to higher areas or break up into smaller units as grazing resources are depleted. While the membership of individual camp groups is constantly changing, kinship ties and tribal categories are continually invoked in respect of camp group membership. In the return migration during autumn, the remaining fragments of camp groups begin to move down from the high mountains in stages, concentrating on the plain above Bitlis Pass until the a§iret is again moving in large consolidated units. After descending through the Taurus passes, the tribal structure begins to disintegrate as groups of families branch off and disperse to their winter villages.
The a§iret: tribal and lineage organization
Ethnographic accounts of nomads invariably convey the impression that tribal organization is something continuous, existing over and above the changes in residence groupings that occur seasonally. The above account raises some doubts about this view. If tribal categories are relevant only seasonally, if people think in terms of tribal concepts for only six or seven months of the year, then the 'tribe7 may be seen as something which is put together and taken apart in response to certain temporary but recurrent organizational needs. While people may retain a sense of belonging to a particular tribe, the organizational structure itself is activated or deactivated as required. In principle it is easily seen how, if the constituent units of a nomadic tribe were physically prevented from joining up each year into migratory groups and camp groups, such a tribal organization would lapse and the tribe itself cease to be. It is noteworthy that this principle has been invoked time and time again by rulers wishing to control, reorganize or disband nomad tribes! It is also evident that the appearance of tribal structure varies according to the season in which the observations are made. The following account of Alikanli tribal structure will apply largely to the a$iret as it appears during the summer months. Superficially the Alikanli are organized according to the following hierarchy: Confederacy A§iret - tribe Kabile - maximal lineage Zoma - camp group Hane - unit or tent Such a scheme however conceals many complexities. Hane (unit or tent) Individual nuclear families normally exist only as a part of extended households, but the term hane does not quite coincide with the extended family: A unit [hane] can be formed by several families, with or without any blood ties, getting together (e.g. hired shepherds could stay in the same tent) or one person (e.g. a widow) can also form a unit . . . So every family coula form a unit, but every unit is not necessarily a famiIy<
(Be§ikgi 1969, p.98; my translation)
Nomads in archaeology
200
The defining feature of the unit is that it occupies a single tent. Since the tents are large, adjustable and internally divisible, the tent will be smaller or larger depending on the number of people - or the number of nuclear families - housed under it. The other main criterion is that the unit constitutes a unit of production and consumption: A unit [hane] is based on unity of earning [income, livelihood])... It is a community formed by one or more persons and families formed by mother, fatner, married son, daughter-in-law, grandchildren, etc., whether there are any blood ties or not, but they eat together and their livelihood has the same source. (ibid., p.98; my translation) The hane is the unit of livestock ownership though apparently it does not always form a viable management unit. Among the group of 37 tents recorded by Be§ikgi, 37.5 per cent of the tents consisted of single-family units, 32.5 per cent were two-family units, three-family units comprised 19 per cent and four-family units 11 per cent. The mean numbers of individuals were 6.0, 7.5, 11.3 and 14.5, respectively. The unit is consistent in its membership over time, unlike the more ephemeral herding units such as those encountered among the Basseri (Barth 1961).
Zoma (summer camp group)
This is the summer settlement unit or camp of 20 to 40 tents. It does not correspond to any lineage or descent group, nor even does it necessarily form around a core of lineage members. Its membership is constantly changing from one year to the next and even in the course of a single season, but in a way that seems to involve more than the addition or subtraction of a few tents. Any one zoma is likely to contain tents from all or most of the eight maximal lineages or kabiles of which the tribe is composed, with agnatic ties forming the basis of association among tents from a single kabile and marriage ties linking households from different kabiles as indicated in Fig. 10.4b. In order to understand this it is necessary to consider the mechanism by which access to pasture is acquired by the Alikanli. All pasture, in both summer and winter quarters, is rented from village cooperatives or from large landowners or agas (Bruinessen 1979; Ozbek-Koroglu 1980). But unlike the Yoruks in western Anatolia (Bates 1973) who rent pasture on a piecemeal basis, the Alikanli practise a system of collective rent. The rent for each yayla is the subject of negotiation between the landowners and leaders of the asiret, in the course of which there is much to-ing and froing between winter and summer quarters. Once a price is agreed, individual tents make the decision about which zoma to join. The rental fee is shared by all household heads, not equally, but on the basis of the number of sheep owned by each, since yaylas are rented according to the amount and quality of grazing available in each. An additional fee is paid for each tentsite. Factors influencing the choice of zoma include the following: belonging to the same kabile, being closely related, to be near wealthy or powerful households, to be close to those who understand Turkish or to avoid conflicts connected with blood feuds. A chief or headman is elected to represent the zoma each year. This is not a formal office and incumbency usually changes from year to year.
NemrutDag
201
Towards the end of the summer season; as grazing becomes depleted, zomas gradually break up into smaller camp groups which then disperse to different yaylas, often joining with other small groups on the way: I never met the same group for too long. Economic and human factors meant that zomas and obas were changing in size all the time . . . .1 met a 37 tent oba and started on these. But once they ran out of grass, 9 of the 37 stayed at Nemrut and others scattered into all of the other yaylas in the region . . . .1 followed a large group to Sute Yayla and continued work in a zoma of 25 tents - 1/tents were already there. (Be§ikgi 1969, p.38; my translation) However Be§ikgi is less than clear on how such apparently ad hoc arrangements fit in with the system of collective pasture rental described above. In summary then the zoma forms a shifting, flexible and temporary primary community. Kabile (maximal lineage) The kabile consists of all those tracing agnatic descent from an apical ancestor. Although it is not identified with any particular territory and never forms a physical community or settlement, it acts as an important reference group. Genealogies are continually invoked to substantiate membership in a kabile and it appears that membership in an Alikanli kabile is a prerequisite for the right to join any zoma within the pastures rented by the Alikan A§iret. There is a high - though unspecified - rate of endogamy within the kabile, though marriage between members of different kabiles is statistically more frequent. There are eight kabiles within the Alikan A§iret - Cudigan, (^engovan, Ozigan, Neciman, Mehman, Birivan, §ehedan and §ehevan - each with a recognized and more or less permanent chief or reis. Closely related members of the same kabile tend to form blocks or segments within the zoma. A§iret (tribe) The a§iret consists of all those considered to be descended from a single founding ancestor, Alikan (Ali Khan), though this cannot necessarily be established through genealogies. The a$iret is the effective marriage isolate, only two marriages having been known to occur outside it - each with members of the closely associated Duderan A§iret. The a§iret is essentially an association of kabiles under the leadership or 'management' - to use Be§ikgi's (ibid., p.69) term - of the chief of the largest and strongest of the kabiles. It constitutes a stable political and administrative unit: The a§iret is the total of all the family relations which are formed by way of marriage . . . The number of kabiles increases or decreases according to the size of the a$iret. . . There is a continuous friendship between the kabiles of an a$iret. The asiret chief is responsible for this friendship and harmony. (ibid., p.69; my translation) Confederacy Although the Alikanli do not combine with other a$irets in any wider political organization, an affinity is recognized with adjacent a$irets such
Nomads in archaeology
202
as the Duderan, Soran, Garzan, Kiqan, Mehmediye and Tayyan, all of which trace a common origin from the Miran from which the Alikanli split in the late nineteenth century. In the past this association seems to have amounted to a kind of confederacy under a religious leader known as sheikh or uluki$i. Little is known of the Miran; the parent body from which the Alikanli and other tribes emerged in the late nineteenth century. They are described as a tribe by Sykes: Miran. 1000 families. Low tribe of shepherds migrating from Jeziret ibn Omar to Lake Van in spring and returning in autumn. This tribe has an atrocious reputation for all kinds of villainy . . . They move up within about 15 miles of Lake Van annually. (1908, p.460) This may represent what was left of the Miran after it split into new tribal units. Of the adjacent asirets mentioned by Be§ikqi, three - the Kichian (Kiqan), Duderi (Duderan or De-dere in Frodin 1943-44) and the Tiyan (Tayyan) - are also listed by Sykes (1908, pp.460-1). The Tayyan were recorded 70 years later by Bruinessen (1979, p.44).
Organizational changes On the basis of what can be pieced together from Be§ikgi and the historical sources, it appears that the Miran was originally a large, loose organization of kabiles, many of which became independent in the late nineteenth century in the course of administrative reform, warfare with the Armenians and Russians, and military reorganization. It is far from clear whether certain kabiles evolved into tribes or whether groups of kabiles joined together and split off from the Miran. The Alikan A§iret may have come into being through the following process: Asirets and kabiles are usually named after the chief. In time, when one or these grows, expands and increases its population and power to the detriment of others, a new a$iret can come into being. But the growth of an a$iret takes place completely within itself. Members From other asirets can never join the a$iret or the kabile but, as a result of armed conflict between asirets, as in past wars, some asirets can become subservient to other larger ones. (Be§ikqi 1969, p.70; my translation) Nevertheless the continuity of tribal names and the constancy of their location argue for a high degree of territorial stability in the region over the last century or so. The growth of the Alikan A§iret, according to the normal evolutionary development of a tribal entity, was complicated by major structural changes in the tribal system itself having to do with the nature of acquisition of pasture rights. Under the Ottoman Empire nomad tribes were able to control tribal territories or were granted grazing rights in return for military services. This situation changed with the establishment of a Republic as the new government did not recognize tribal rights to land. Since nomadism, by definition, involves the seasonal use and subsequent abandonment of territory, the spread of village settlement backed by government policy has resulted in the progressive alienation of land from tribal nomads. This in turn has brought about changes in the function of tribal institutions and, to an extent, in social organization itself:
NemrutDag
203
Firstly, in the old (pre-Republican) arrangement, there was no level called zoma. Social dynamics created such a mechanism during transition into the new order. Whereas in the past kabiles used to form several separate units for stopping, today members of many different kabiles come together and form a unit for stopping together. The reason for this is that solidarity within the kabile in the past was adequate to provide security. But after the Republic was formed, as a result of the power of the central government, asirets became weaker and weaker, and solidarity within the kabile became inadequate and it was necessary to have solidarity within the asiret. (ibid., p.70; my translation) What this implies is that, in the past, each kabile constituted a distinct territorial unit, most probably splitting into sub-units based on lesser lineage segments to form camp groups which were probably larger than at present - the oba mentioned by Be§ikgi. The tribe or asiret was simply a loose association of adjacent kabiles. With the loss of hereditary pasture rights the kabile lost its territorial integrity and no longer constituted a settlement unit, its members being forced to join the pasture-exploiting collectives formed annually in response to whatever tracts of pasture could be assembled through renting. Thus the zoma came into being as a unit of settlement and social organization cutting across the descent structure of the kabile. This in turn reinforced the sense of belonging to an asiret, which now represented a sufficiently large population pool from which to draw members for each zoma. The latter now became the primary territorial and residential unit - at least in the summer months while the kabile retained meaning as a lineage system conferring membership within the asiret (see Fig. 10.6). The evolution of the zoma, of which our campsite ND-1 is an example, therefore closely parallels the changing function of the asiret from an association of lineage-based territorial sections to an organization for the acquisition of pasture rights through the formation of rent collectives.
Wealth, unit size and the organization of herding Wealth among the Alikanli is measured in terms of livestock holdings, particularly sheep which form the major commercial focus. Animals are sold via agents and may be exported as far afield as Jordan, Iran and Saudi Arabia. As stated earlier, animals are owned collectively by the unit or tent but there are great variations in the numbers owned. Of the 37 units in the camp studied by Be§ikgi, three owned fewer than 5 sheep and two owned more than 500, the modal group owning between 50 and 250 (see Table 10.2). There is some evidence for a direct relationship between herd size and the number of nuclear families per unit. The three units with fewer than 5 sheep were all single-family units, while the two with over 500 comprised four or five nuclear families. Since sheep are taken out to graze in groups of around 500, accompanied by two or three shepherds and two dogs, small herd owners are obliged to cooperate in making up the livestock numbers and manpower. However Be§ikqi (1969) does not mention the formation of herding units comprising a number of tents maintaining a close association and habitually camping together. Large owners are able to make up a single herd - or
Nomads in archaeology
Fig. 10.6 Changes in Alikanli tribal structure and residence hierarchy from late nineteenth century to recent times, according to arguments pursued by Be§ikcj (1969). (a) Ottoman period. Kabiles (maximal lineages) are selfsufficient units for for the formation of camp groups. The tribe is a loose association of kabiles. (b) Republican period (post1923). Members from different kabiles combine to form rent collectives (zomas). Tribal identity is accentuated.
204
Asiret
KABILE
KABILE
KABILE
DDDC CAMPS
kabile
kabfle
kabile
i \>< >><7
a" trt5 ZOMAS
Table 10.2 Demography and animal holdings for five wealth classes, Alikanh summer camp (afterftesjkci,1969) Wealth classes (no. of sheep)
No. of units
1(3-4) 2 (5-50) 3 (51-250) 4 (251-500) 5 (500+)
3 9 18 5 2
Total
37
(%)
8.3 24.3 48.8 13.5 5.5 100.0
Average nuclear Total no. of families per unit sheep owned
1.0 1.4 1.9 3.6 4.5
9 330 2497 1676 1403 5935
(%)
Average no. of sheep per unit
0.1 5.5 42.0 29.4 23.0
3.0 34.6 120.2 366.6 662.5
100.0
160.4
even multiple herds - for grazing purposes and to maintain the necessary shepherding labour or afford the services of hired shepherds. Herding tactics are left largely to the shepherds themselves who divide an area of rented pasture up into lots of roughly equal size. At the time of my September 1981 visit, three large flocks of at least 500 each were being grazed within the crater. All were made up of sheep owned by the 15 tents camped at ND-4, the only campsite occupied at the time, giving an average of around 100 sheep per unit. One flock was being grazed in the area to the southeast of Nemrut Lake, in the region of the abandoned campsite ND-1, another along the eastern shore and a third on the extensive lava flows to the north and west of ND-4. The flocks were taken out
NemrutDag
205
in the evening, pastured throughout the night, and returned to camp by about 10 a.m. The bulk of the flocks were not corralled but kept at a distance from the camp (cf. Nyerges 1980; p.39).
Tents and household status
The large, multi-poled black tent employed by the Alikanli and other Kurdish tribes in southeastern Anatolia is a highly flexible dwelling form which is easily adjustable to accommodate changes in the size or composition of units (cf Evans 1983). It lends itself readily to the implementation of quick decisions about abandoning a particular location, a routine feature of nomad life. The tent is woven from the hair of black goats kept specifically for this purpose. It consists of a number of cloths stitched or pegged together and is supported by two, sometimes three, rows of poles running along its length. It is normally surrounded by a wickerwork screen and drainage ditches, though sometimes stone walls entirely surround the tent. The floor is levelled off and made firm prior to occupation. Be§ikcj gives few details of the physical factors governing the location and orientation of tents, though he does mention that 'a tent is put up against the wind direction7 so that it strikes the rear. Unlike the tents of tribes such as the Beritanli to the west, which have long guy ropes and pegs, the tents of the Alikanli are secured by means of stout wooden stakes driven into the ground just outside the walls and attached to the tent by a rope only a few inches long. Another distinctive feature is that the tops of the poles poke through the tent, forming a series of steeply inclined peaks along the crest (see Evans 1983). The size of a tent is determined by the size of the unit, the number of nuclear families and the status of the household head. The households of important men tend in any case to be larger, but an influential man will require a disproportionate amount of space to accommodate guests and large social gatherings. The status of a household is approximately indicated by the number of tentpoles along the middle row. The tent of a low-status household will have a minimum of two poles. Most have four or five, while the tent of a chief or influential man may have as many as six or seven poles. An 'average7 tent is around 4.5 metres in width and 8 or 9 metres in length.
Camp layout; status and ranking
Social status is indicated not only by the size of the tent but also by its position with respect to that of the zoma chief or headman: The position of the tent . . . the distance or proximity7 of the tent to the chief's tent could also determine status. The chief s tent is nearly always at the front. The tents of his relatives and those close to him are placed around his tent. Later on those who are important and others have their tents. (Be§ikgi 1969, p.76; my translation) A camp structure of this kind can be observed at ND-4 where tents are strung out in two lines behind that of the zoma chief (Figs. 10.7 and 10.8).
206
Nomads in archaeology
Fig. 10.7 Plan of occupied late summer campsite ND-4, Nemrut Dag. The campsite is situated on a flat plain at the base of an extensive lava flow. Note how the tents follow the base of the lava flow.
lava flow
disused
^"
\ ^
s
fj% M
9
\
K 0
0
Fig. 10.8 Panoramic view of Alikanli summer camp ND-4; Nemrut Dag, from the the lava flow to the west. Note the parallel lines of tents and the location of the two largest tents at the entrance to the camp. Unused stone foundations can be seen in the background.
»
u ?
t
\
%
W
\
30
to
— .--occupied LL^J tentsite ( ) tentsite <W £2 « c
woodpile midden ash dump pens etc.
.
Nemrut Dag
207
A remarkably similar pattern can be seen in an Alikanli camp photographed in 1939 by Frodin (1943-44) and reproduced in Fig. 8.6. A large, multi-poled black tent - probably that of the chief - is surrounded by a cluster of smaller tents (his relatives?) and forms the focal point of three roughly parallel rows of tents stretching along the edge of the plain. The social status of a household is also reflected in the size of the ash dump near the tent, a large ash dump indicating that the family is noted for its hospitality - its hearths are always kept burning in anticipation of guests.
ND-1 in retrospect: an informed reassessment It should be emphasized at this point that the interpretation of ND-1 from its observed remains was carried out prior to my having read Be§ikgi's account of the Alikanli, and was guided only by what few observations I had been able to make of occupied camps in the crater, plus a little common sense. In view of the fact that a high level of accuracy was not anticipated, the degree of fit between the initial interpretation and the ethnographic account - in some respects at least - was surprising. In comparing the two we are trying to determine how closely the interpretation conforms to the principles of demography and social organization derived from the ethnographic study. It should be stressed that the main
Nomads in archaeology
208
object is not to determine whether the conclusions reached were correct or incorrect, but rather to assess the information value of this kind of reconstruction, of a type frequently attempted by archaeologists.
Demography The division of living space into rooms containing nuclear families and dwellings containing extended families is broadly consistent with the ethnographic account. The identification of the dwelling as the primary unit of consumption accords well with Be§ikgi's account of the hane, or unit, except that the unit may occasionally include non-family members. The estimated average number of occupants per dwelling of 8.6 persons (see Table 10.1) matches the 8.4 calculated by Be§ikgi (Table 10.3a). When the distribution of the population among units of different sizes is considered, however, some anomalies emerge. In Table 10.3 the number of units or extended households falling within each size class - from a single nuclear family to four or five nuclear families - in the community studied by Be§ikgi are compared with those estimated for ND-1. Clearly there are fewer single family units at ND-1 than would be expected if Be§ikgi's account is representative of all Alikanli campsites. Two explanations may be offered. Either the average unit size at ND-1 was larger than normal or the estimation of the number of nuclear families, one per room, is overgenerous. In the latter instance, it may be that not all rooms in a dwelling need be occupied. With fluctuations in the size and composition of camp groups from one year to the next it is probable that stone fixtures will be underused, but unlikely that they will be over-used, given minimum requirements of privacy and living space. Spare rooms could easily be pressed into service as fodder stores or lamb pens. The presence of such a built-in Table 10.3a Distribution of households (tents) over size classes of units - camp studied by Be$ikci, Nemrut Dag, 1965 (after Be$ikci 1969, table 21) No. of families per unit 1 2 3 4-5
No. of tents
(%)
Population
(%)
Mean population per unit
14 12 7 4
37.5 32.5 19.0 11.0
84 90 74 58
27.0 28.0 26.0 19.0
6.0 7.5 11.3 14.5
37
100.0
311
100.0
8.4
Table 10.3b Distribution of households (tents) over size classes of units - estimates for ND-1 Nemrut Dag, 1978 No. of families per unit 1 2 3 4-5
No. of tents
(%)
Population
(%)
Mean population per unit
2 6 2 1
18.2 54.5 18.2 9.1
10 48 22 14
10.6 51.1 23.4 14.9
5 8 11 14
11
99.9
94
100.0
8.6
NemrutDag
209
bias would affect all demographic estimates based on structural remains in settlements with a high population turnover.
Social organization The question of population turnover also affects the interpretation of higher order social structures. The equation of clusters of dwellings with 'lineages' is valid only in a sense. Assuming that closely related members of the same kabile camp in close proximity to one another, as stated in the above ethnographic account, each zone or cluster could represent a contingent from six of the eight kabiles combining to form a zoma. But these clusters do not correspond to the lower-order segments of a single lineage, as was originally suggested. The spatial organization of dwellings reflects the structure not of a lineage system but of residence groups. There is nothing in the settlement plan itself to suggest either of these alternatives, though in making inferences from archaeological data it might be prudent to confine oneself to residence rather than kinship categories.
Wealth, ranking and the organization of herding
The designation of the large complex D.8 (see Fig. 10.3) as that of the camp headman of course cannot be confirmed, though its location where two of the paths leading into the camp converge recalls Be§ikgi;s observation that the tent of the headman is located at the 'front7 of the camp. The correlation of unit size with both wealth and status suggested by Table 10.2 points to D.8 and F.ll as the residences of the most prominent households. Note also that it is with these large and isolated residences, but also A.I, that the circular 'guest7 tents are associated. The tendency noted by Be§ikgi for important households to be located close to the tent of the chief is also consistent with the location of D.8 and F.ll in the same area of the site. In terms of the distribution of wealth in livestock, ND-1 could be expected, on the basis of Table 10.2, to contain one large herd-owning unit in wealth class 5 (over 500 sheep), one or two units in wealth class 4 (251-500), around six in class 3 (51-250) and three or four in classes 1 and 2. This would result in a total of some 1,800 sheep for the 11 units estimated to be present at ND-1. This is clearly too many to be accommodated by the walled corral space available at ND-1. However, due to the practice of night grazing, most animals, apart from the sick and the young, are not penned at night. Some further inferences are possible. On the basis of the herding practices reported by Be§ikgi, 1,800 sheep would be divided into at least three herds. Note that ND-4, consisting of 15 tents, also maintained three herds (see p.210). Three herds mean three contingents of shepherds, either hired or drawn from the young men of the herd-owning units. Two would be required for duty at any one time and another one or two as replacements. Assuming that the circular 'guest7 tents were used to accommodate hired shepherds, it is just possible that the distribution of these features in three areas of the camp - attached to D.8, F.ll and A.I reflects the presence of three livestock managements units. Since wealthy
Nomads in archaeology
210
families are more likely to be able to employ hired shepherds, D.8 (presumably a large herd owner) could form a single herding management unit with a few hired shepherds. F.ll, together with E.9 and E.10 to make up the numbers, could constitute a second group, while small herd owners occupying the remaining tents in zones A7 B and C make up a third herding cooperative based on A.I. This arrangement would also conform to the location of tents being determined by the wealth and status of the household head and his connections with the zoma headman.
Comparative observations on ND-4 In the course of my 1981 visit the camp in the northeastern part of the caldera, currently occupied by members of the Alikan A§iret, was briefly visited. The camp consisted of 26 tentsites, 15 of which were surmounted by tents (Fig. 10.9). The camp was located on the margin of a flat, barren plain of volcanic ash below the over-hanging slopes of a lava flow which served as protection from the westerly winds that prevail during summer and autumn. A shallow brackish lake occupied the plain about half a kilometre from the camp, but a small spring of drinking water emerged from the lava flow behind the camp. The tents were enclosed by high stone walls and surrounded by reed screens. Most of the dwellings were internally divided although the largest one, the headman;s tent in which I was received, had no internal divisions. This tent, 16 metres in length and 9 in width (see Fig. 10.8), was able to accommodate a large proportion of the camp's population who had crowded in to welcome Allah's guest. Many tents had well-defined forecourts and kitchen alcoves in which cooking and domestic activities were being carried out by the women. Large woodpiles were stacked in front of many tents and some also had piles of unprocessed wool for use by the women, many of whom were engaged in beating, spinning and dyeing the yarn inside the tents. Middens and ash dumps were located directly in front of or between tents. There were only a few small corrals and at least one latrine. As shown in Figs. 10.7 and 10.8, the tents were arranged in two rows following the curving line of the lava flow and the tendency towards clustering was less marked than at ND-1. The large headman's tent, with four central poles, was located at the entrance to the camp, where a rough vehicle track terminated. In front of it was the only circular, white 'guest7 tent in the camp. Another large four-pole tent was positioned directly behind it, again confirming Be§ikgi's observation that the most important households - with the largest tents - have their tents close to that of the headman. ND-4 illustrates well the extent of variability in tent size. Apart from the two large four-pole tents, most of the others had three poles and a few had only two. Due perhaps to the formation of rows, the density of settlement at ND-4 was considerably greater than at ND-1. The differences in layout between ND-1 and ND-4, together with the fact that on both visits the former was deserted and the latter occupied, have prompted some questions about possible differences in function between the two campsites. The location of ND-1 in the south of the caldera, facing north, and ND-4 in the north, facing east, suggest the
211
Nemrut Dag
Fig. 10.9 Members of the Alikan A§iret outside the headman's tent at ND-4. Note the large size of the tent and the two rows of four poles.
• ' -
:
-
*
\
'
-
interplay of seasonality and aspect. ND-1, located on the cooler, northfacing slopes would be a good place to camp in mid-summer. By autumn the long shadows cast by the southern rim of the caldera would make a northerly location more attractive. In view of the frequent changes of campsite during the spring-summer-autumn cycle described by Be§ikgi, this interpretation would be consistent with Alikanli spatial behaviour.
Conclusion The 'lost world7 of Nemrut Dag provided an ideal laboratory in which a number of phenomena of nomadism and nomad archaeology could be studied within a clearly demarcated area. Not only was it possible to observe nomad material culture at the upper range of its visibility, but also the meaning of intra-site settlement structure could be explored in full. Above all the existence of an independent ethnographic account permitted an indirect comparison of ethnographic and archaeologically derived analytical categories. The above exercise illustrates the potential value of analysing a settlement pattern in relation to a relevant model of social organization and economic structure, rather than attempting on an ad hoc basis to match surface features to a particular occupational event in the recent or distant past.
11 NOMAD ARCHAEOLOGY: AN ASSESSMENT
This study has been built around the proposition that nomadic sites constitute an analytically distinct settlement category requiring different methods of survey, recording and excavation from those commonly employed by archaeological fieldworkers in the Near East. Having arrived at an understanding not only of the structure, location and contents of nomad campsites, but also the kinds of methodological and theoretical problems that are best pursued in relation to this kind of site, let us now reconsider what archaeological evidence is available.
Tepe Tula;i It is significant that the only site securely identified as a nomadic - or at least pastoral - campsite was detected not through intentional survey but in the course of rescue operations on a small Sassanian site prior to an agricultural development scheme in Khuzistan (Hole 1974-75, 1975). Tepe Tula'i lies near the northern edge of the Khuzistan Plain about 4 kilometres east of the Karkheh River and 17 kilometres west of Dezful, placing it on a possible seasonal migration route from Khuzistan to summer pastures around Khorammabad in the Zagros Mountains. The site first appeared as a low mound dissected by old irrigation canals, the surface of which was covered by loose, rocky fill containing Sassanian sherds. These were partially mixed with prehistoric sherds from a layer of hard-packed fill, disturbed by burrows, which extended well beyond the confines of the mound itself. A 5 metre square test pit cut into the mound encountered, at a depth of 2 metres, a layer of undisturbed clay containing a fragment of a pebble floor on which lay a Khazineh red pot. Below this were alternating strata of sterile clay, sand and river gravel. Mechanical levelling of the area to the north of the mound, removing anything from a few centimetres to 2 metres of earth, 212
213
Nomad archaeology
revealed a series of concentrations of ash and stones on the graded surface that intersected the original gently undulating modem surface. Two thick and extensive areas of ash revealed no architectural traces. However exposed lines of stones lying 50 to 100 metres to the north of the mound and in association with prehistoric sherds were readily identified by the local Luri workmen as the remains of tentsites (see Fig. 11.1). While the stratigraphy of the deposits within which the stone platforms were embedded remains unclear, it appeared that they lay either slightly above or directly on top of the sterile layers of sand, clay and gravels which formed the substrate over the entire site. Fig. 11.1 Stone platforms and tent footings, Tepe Tula'i, Khuzistan, from contexts dated to approximately 6,200 BC. (After Hole n.d., fig.7.)
mortar fragaents 4 scattered stones
ash
ash disposal metres front of tent
approximate "tent outline
OOi Q1
Chronology of Tepe Tula;i
On the basis of the ceramics found in association with the stone platforms, and the sequence uncovered from the main test trench (Area A) in the mound, the prehistoric occupation of Tepe Tula;i has been dated to the late Mohammad Jafar and early Sefid phases of the Deh Luran sequence, estimated at around 6,200 BC (see Hole etal. 1969). The ceramics found in the same area as the stone platforms consisted solely of 'grey ware which, according to the excavator, places these features at the earlier phases of the sequence. However the absence of clear stratigraphic connections between the culture-bearing deposits of this northern part of
Nomads in archaeology
214
the site (Area D) and the main mound (Area A) must cast some doubt on the chronology. The sherds in Area D were apparently not found in direct association with the stone platforms but in lower deposits. Unfortunately no charcoal samples suitable for carbon-14 dating were recovered. It is possible that the stone features may not be quite as ancient as has been claimed, though their depth indicates that neither are they of modern origins.
Was Tula'i a campsite? If the chronology is correct then the implications for theories about the course of Near Eastern food production and pastoralism are far-reaching. Of course the presence of a campsite of such an age does not per se indicate nomadism - it could just as well have been a seasonal settlement of village-based pastoralists or agro-pastoralists (see Wheeler PiresFerreira 1977). It nevertheless suggests some degree of mobility and a certain level of pastoral specialization. Whatever the age of Tula'i, its identification as a campsite seems more secure, particularly in relation to the kinds of household organization and settlement systems discussed previously (Chapters 4 and 8). Tula'i represents an interesting application of direct ethnographic analogy. The Luri workmen employed on the excavation were themselves recently settled nomads and were able to provide insights into the location and interpretation of features as the excavation was in progress. The stone features were readily identified as the chul, or bedding platform, found in contemporary tents or kulas. The orientation of the platforms and the fact that some appeared to be facing north while others faced south suggested, on the basis of current practices, that the site was a spring or autumn camp. This accords well with the location of the site close to a migration route near the ecotone between lowland and upland zones. The precise locations of hearths and ash dumps could also be predicted on the basis of the nomadic experience of the workmen. For example, in the case of structure D4, which appeared to face south: 'When we cleared the rocks, I asked a workman where the fireplace would be. He glanced around and quickly took two paces to the south. We dug there and found the ash' (Hole 1975, p.67). Another stone outline was identified by the workman as that of a 'headman' on the basis of its greater size. We have here an example of some of the rules of campsite organization discussed above, albeit applied in a piecemeal rather than a systematic fashion. No conscious attempt was made to apply the kind of excavation strategy appropriate to such a site - and perhaps could not have been, given the circumstances under which the excavation was carried out. To begin with, the use of test pits placed directly inside the structures themselves is unlikely to be a productive approach in campsites. Even in terms of the limited objectives of obtaining ceramic evidence for dating, we would expect most of the sherds to occur in the discard zone at a distance from each tentsite. In Chapter 9 it was shown how the location of middens could be predicted from architectural features, and vice versa. Short
Nomad archaeology
215
of exposing large areas equivalent to a whole domestic complex; it might have been possible to place trenches along the axis of a domestic complex perpendicular to the orientation of the tentsite. The spacing of structures at Tula'i also recalls the familiar open settlement pattern found in most campsites. But in the absence of some kind of sampling scheme we cannot be certain that all dwelling sites have been recovered and have no idea of the extent of the site. In terms of site structure, the following observations by Hole do not match up with what is now known of campsite organization: The mound portion of the site was clearly a focal point of activity in the sense that sherds, flints, figurines and animal bones were left there. However it is worth noting that grinding stones were found only in the campsite area. Pending further excavation I suggest that the mound is a midden area used DV the people who were camping nearby . . . a similar midden lies under campsite Dl where we placed another test pit. (1975, p.71) By their very nature, campsites do not have 'focal points', at least not in the form of communal middens. Either the interpretation is incorrect or we have a genuine case of 'anomaly' in which unexpected findings may suggest that things were done differently in the past. The discovery of a midden layer directly under one of the tentsites suggests that the function of this particular zone changed with time from a discard to a living zone.
The faunal evidence The faunal remains from Tepe Tula'i not only confirm a pastoral emphasis but also fit the domestic complex model, since animal bones were found predominantly in the middens rather than in the vicinity of the tentsites. A study of the faunal remains (Wheeler Pires-Ferreira 1977) shows a predominance of ova-caprines. A high proportion of fused as opposed to unfused elements suggested to the author, on the basis of modern practices, that the faunal population of Tula'i represented a fallow herd of unproductive adults split off from a nearby village herd. This interpretation, if correct, need not threaten the identity of Tepe Tula'i as a pastoral campsite, though it certainly casts doubt on the extent to which it can be regarded as 'nomadic'. As recently stressed (Cribb 1985) it is very difficult to draw conclusions about herding strategies - particularly those involving 'transhumance' or migration - on the basis of age criteria alone. Whether or not it was appended to a nearby village, Tepe Tula'i was a seasonal campsite of some kind. It points to a degree of separation between pastoral and agricultural subsistence strategies, though just how far this had progressed and the extent of the migratory orbit involved remains unclear. Nevertheless the discovery of such a site, conforming in most respects to what would be expected in a contemporary nomadic camp, and at such an early date - assuming this to be correct - puts the lie to the argument that ancient nomad camps are somehow archaeologically invisible.
Nomads in archaeology
216
Other evidence of possible tentsites In the light of the present study and field work carried out by Hole (1978, 1979), other Near Eastern sites are worth reassessing as possible nomad camps. In other cases floorplans found in sedentary settlements may turn out to be tentsites or composite dwellings.
Khuzistan Sections of wall footing found in the Sabz phase at Ali Kosh (Hole et al. 1969, p.59) are reminiscent of tentsites (Fig. 11.2). A complete floorplan exposed on the surface of Tepe Ashrafabad (ibid., p.60, see Fig. 11.2b) suggested in retrospect to Hole (n.d.) that this might be the storage platform and outline of a tentsite. The large size and lack of internal subdivisions, the structure of the platform, lined with cobbles and filled with smaller pebbles, and the narrow stone outlines at the sides are all inconsistent with the foundations having acted as a load-bearing structure for mud-brick walls and a permanent roof. The presence of a pivot stone for a door might suggest a more permanent structure but, as observed in Chapters 5 and 6, doors or gates are commonly employed in modern tents. Hole (n.d.) has since suggested that Tepe Ashrafabad, a low and extensive mound, may have been a nomadic campsite. Excavations at Tepe Farukabad (Wright 1981) also revealed a number of wall footings from the Uruk, Jemdet Nasr and early dynastic levels that suggested to the excavator similarities with modern Luri tentsites. One of the dwelling plans (Fig. 11.2b) consisted of two rows of ovoid limestone cobbles measuring 5.0 by 2.9 metres, with no trace of mud-brick walls or any kind of superstructure. Two successive floors were present, each bearing the imprint of a central post or pole: 'During excavation, the posthole and the absence of brick walls were taken as evidence that this footing was for a tent similar to the stone tent footings used by Luri herders today; (Wright 1981, p.78). We have seen that successive floors, prepared and re-plastered with each occupation, are regular features in some tentsites. The presence of a further floor, adjacent to the structure, is reminiscent of the forecourts which accompany most tentsites. In the light of our discussion of 'composite settlements7 (Chapter 8), the presence of such a structure in a built-up, mounded settlement should occasion no great surprise.
The Zagros Mountains The surface features identified as tentsites by Edelberg (1966-67) fail to match most of the criteria outlined in this study. Hole (n.d.) is quite correct in dismissing them as 'houses of the dead7 (Hole n.d.), ritual tentshaped structures to commemorate deceased tribal leaders. The absence of ash or artifacts and the conspicuous lack of a storage platform, the highly stylized layout and the location of the structures on a featureless and very stony plain, make them unlike any tentsite encountered in the course of this study. Certain sites in the Zagros Mountains - Tepe Sarab, Ganj-i-Dareh,
217
Nomad archaeology
Fig. 11.2 Suspected ancient tentsites from Khuzistan. (a) Dwelling foundations at Tepe Ashrafabad, Deh Luran Plain. (After Hole etal. 1969, fig.17.) (b) A tent-like floorplan from the Uruk phase, Tepe Farukabad, Deh Luran. The structure reminded the excavator of the layout of a modern Luri tent. (After Wright 1981, fig.36a.) (c) This example is most probably not a tentsite but a ritual structure associated with funerary rites and of recent origin. (After Edelberg 1966-67, fig.21.)
ooooo o ooooooc,
entrance
hearth
hearth
j (
b corner stone
y 5m
0 n
J
Nomads in archaeology
218
basal Tepe Guran (Mortensen 1972) and early Neolithic Shanidar (Perkins 1964) - have long been suspected as pastoral campsites (see Hole 1978, p. 137). The difficulty involved in positively identifying them as such lies not so much in the lack of archaeological data but on the absence hitherto of any firm defining criteria. Recently Smith and Mortensen (1980) have drawn attention to large numbers of early sites located in sheltered side valleys of the Zagros away from the main areas of cultivation: The clustering of sites in these small and nearly inaccessible valleys shows that these are more promising locales for investigating early neolithic sites in the Zagros than are the broad intermontane plains7 (Smith and Mortensen 1980; p.512). It is clear from Chapter 8 that this is precisely the kind of location in which nomad pastoral campsites are to be expected, and the article by Smith and Mortensen (1980) actually contains a photograph of a small modern camp of four tents in such a location. Although a sedentary settlement, the seventh and sixth millennium BC site of Haji Firuz Tepe in the Solduz Valley near Lake Urmia illustrates certain features of spatial organization that have been discussed in relation to nomad camps. On the basis of the excavated area, the site appears to have had an open settlement plan with refuse disposal occurring in the spaces between the dwellings: 'Nearly all of the exterior areas excavated in 1968 were littered with trash, including ash, burnt earth and broken pottery; it seems likely that a significant quantity of organic refuse, including animal bone, was also disposed of in these areas7 (Voigt 1977, pp.31819). The absence of open area excavations on low mounds or flat settlements makes it difficult to assess how common such dispersed settlement patterns might be, but they certainly indicate that not all food-producing settlements were densely settled and that the domestic complex model, in such cases, may provide a useful framework for analysis.
Eastern Anatolia Archaeological surveys in the Near East have generally been either concentrated in areas of high ancient population density or concerned primarily with conspicuous mounded settlements. An exception has been the Upper Euphrates survey carried out prior to the Karababa dam project in eastern Turkey. Although the survey was confined largely to riverlands due to be inundated by the new dam, the coverage was thorough, and many small and flat settlements were recorded (see Ozdogan 1977). Two such settlements were tentatively identified as possible campsites. Both were located on alluvium, close to springs and adjacent to the first river terrace: Alikan Mevki (flat settlement) . . . about 2 km north west of Biriman, in a field . . . west of the Alikan Gozesi [spring]. There is a heavy concentration of sherds over an area of approximately 100 by 60 m on the ridge and the gentle eastward slope, next to the spring which has running water all year round . . . The surface material is variated but not very dense. The site may have been a campsite at various times. (ibid., pp. 122-3) §uruk Mevki (flat settlement). . . about 2 km north of Biriman, on the west bank of the Alikan Deresi. A slope settlement on a low ridge.
Nomad archaeology
219
Traces of stone foundations appear on the surface. There is a small spring at the foot of the slope. Sherds are scattered over an area of approximately 70 by 80 m; there are also a few flint tools . . . Very small settlements or campsites were set up here next to the spring during various periods. (ibid., p.124) Sherds recovered from the sites indicate occupation during the Iron Age, Classical and post-Classical, with a Late Bronze occupation for Alikan Mevki. While such a location is not typical of nomad campsites the slope and shelter provided by river terraces, together with proximity to a spring, are consistent with the locational model outlined in Chapter 8. It is unlikely that either site will be investigated further due to the tendency to concentrate resources on the larger mounds. Numerous other 'small unimportant settlements' in the survey area, some showing surface traces of stone foundations, may be worth investigating as possible pastoral camps. A few sites excavated in the course of the Keban project, another predamming salvage operation in eastern Turkey, exhibit some of the features to be expected in nomad camps or unstable settlements. One site in particular, Han Ibrahim §ah, is a shallow mound, the highest level of which contains circular hearths and pits and one rectangular structure (Ertem 1972, pl.53.1 and fig.47). The central feature, consisting of parallel lines of stone footings, is within the range of a typical black tent, approximately 7 by 3 metres (Fig. 11.3). The location of pits and stone hearths and features to the south and west of this floorplan recalls the forecourt and work areas of contemporary tentsites discussed in Chapter 7. The southwest-northeast orientation is unusual, but fits well with the terrain of the mound, whose ridge line follows this direction. The location atop a mound is also atypical for a campsite, though by no means unknown. Fig. 11.3 Plan of features in Level Al at Han Ibrahim §ah, eastern Anatolia. The lines of stones and hearth features are reminiscent of the layout of modern black tent encampments. (After Ertem 1972, fig.47.)
hearth ptts
HAN IBRAHIM $ AH, 1970
5m
Nomads in archaeology
220
This level, designated Al and only 40-50 centimetres deep7 was dated through ceramics and coins contained in the stratum immediately beneath to the late Byzantine period (eleventh or early twelfth centuries) (ibid., pp.70-1). At this time Oghuz (Turkmen) tribes would have been moving into the area; though it is possible that they would have still been using yurts of Central Asian type which would not fit these foundations. However local groups of Kurdish affinity may even then have lived in tent encampments during some seasons. The excavator adds that The fact that there were no traces of finds or of fire strongly suggests that the last settlement floor of the mound was evacuated7 (ibid., p.73). Investigations of a nearby rockshelter also showed traces of a stonelined tentsite or corral on the surface, though these may well be modern (Kokten 1976, pl.6).
Eastern Iran and Afghanistan Schaffer (1972, pp. 177-8) has pointed to the discontinuous nature of the stratigraphy in numerous rockshelters in the Sistan region of eastern Iran as possible evidence for seasonal, pastoral occupation. The interleaving of culture-bearing deposits with sterile strata is suggestive of cycles of alternating occupation and abandonment. The mound site of Deh Morasi Ghundai (Dupree 1963a; Fairservis 1961) exhibits a similar stratigraphy and has likewise been tentatively designated a nomadic or semi-sedentary occupation. Stratigraphic sequences of this kind may well represent an archaeological signature for the abandonment and re-occupation of sites over a long-term cycle such as that discussed in Chapter 8. Again numerous rockshelter sites in Afghanistan show traces of seasonal occupation or use by herders throughout lengthy periods of history and prehistory (Dupree 1963b). One particular site, Dara-i-Kur, occupied sporadically from 2,180 to 1,800 BC, contains traces of what appears to be the reed screens commonly placed around tents, though these could also have been used to enclose animal corrals. Lines of around 80 postmoulds 2 to A centimetres in diameter appeared in one occupation level, just under the drip line of the shelter (Schaffer 1978, p.82). This is strongly reminiscent of the seasonal nomadic occupation of Shanidar Cave observed by Solecki (1979).
The Early Transcaucasian Culture: a test case Covering the period 3,300-2,200 BC, the Early Transcaucasian Culture or Kura Araxes complex spread rapidly from what is now Soviet Georgia and Armenia across the highland valleys of northwestern Iran and eastern Anatolia as far as Syria and Palestine (Sagona 1984). Its hallmark was a distinctive form of black-burnished ceramic ware which appears to have been derived from a metal prototype and whose attached perforated lugs and handles suggest that portability was a prime consideration in its manufacture. Portable hearths and potstands, also of the same burnished ware, add weight to the impression of a high degree of mobility and the hearths at least find a modern equivalent in the small metal tripods commonly used and transported by contemporary nomads.
Nomad archaeology
221
The architectural features found in sites of the Early Transcaucasian Culture from Soviet Georgia to eastern Anatolia also suggest a high degree of impermanence and mobility. While it is unlikely that tentcloth was used as a superstructure, the layout of the dwellings, with storage platforms and wattle-and-daub walls, have parallels in some of the kula structures of Iran (Stark 1933) and Afghanistan (Ferdinand 1962). Such architecture appeared episodically in eastern Anatolia along with fresh ceramic traditions from the Transcaucasian homeland (Sagona 1984, p.104). Such transportable material culture, together with the rapid spread of the culture in highland areas in which modern nomads are found, has led to suggestions that the Early Transcaucasian Culture may have had a nomadic base or origin (Burney 1961, 1971; Whallon and Kantman 1969). We now know that the Early Transcaucasian Culture exhibits the full range of settlement forms from tiny hamlets to urban or semi-urban settlements. Arslantepe, one of the larger settlements, contains what appear to be public buildings (Palmieri 1981). However some excavated settlements contain dwelling forms of the wattle-and-daub type which may be 'composite7, and variations in the stability of settlement and relative permanence of architecture have been detected (Burney 1961; Esin 1974; Hauptmann 1974; Ko§ay 1976; Ertem 1972; Sagona 1984). Wattle-and-daub dwellings, which appeared in the Upper Euphrates region of eastern Anatolia early in phase two of the Kura Araxes sequence, were followed by rectangular mud-brick houses, which in turn were replaced on some sites by further light wattle-and-daub structures early in the third phase (Sagona 1984, pp.99, 102, 104). While there appear to be no grounds for regarding the Early Transcaucasian Culture as a nomadic 'culture7, there is good reason to consider whether it may have had a fluctuating nomadic component as revealed in the variations in ceramic forms and architecture. Dwellings on many of the Early Transcaucasian Culture sites appear to be of types 4 or 5, i.e. light-walled enclosures containing a central hearth and platforms around one or more walls and a forecourt, with little suggestion of superstructure other than a central post or pole that could have supported either a light thatched roof or a tent of some description. Some of the floorplans in Fig. 11.4 have clay platforms and rows of postmoulds such as might be left by a reed screen or wattle-and-daub structure. Another (Fig. 11.4c) floorplan from Soviet Georgia, together with its reconstruction (Dzhavakhisvili 1973), displays an almost uncanny resemblance to the mud foundations found in some contemporary Baluchi campsites (Audouze and Jarrige 1978, 1981). The Baluchi examples may have either a black tent or a light roof of poles and reeds as superstructure. The spatial organization of many Early Transcaucasian Culture settlements is camp-like in the consistent orientation of dwellings, although their high density suggests otherwise. The settlement shown in Fig. 11.5 appears to be the result of accretion, with the earlier phase displaying a more open plan with a single cluster of dwellings in a rectangular formation (Dzhavakhisvili 1973). While the Early Transcaucasian Culture cannot be regarded as a nomadic culture - we have seen that such a concept is inapplicable in the Near East generally - there are certainly suggestions of nomadic traits in
222
Nomads in archaeology
Fig. 11.4 Tent-like floorplans from the Early Transcaucasian Culture, third millennium BC; in eastern Anatolia and Transcaucasia. (a) Plan of type 2 dwelling at Nor§un Tepe, eastern Anatolia, described as yapilan direk evler (houses made of poles). (After Hauptmann 1974, fig.40.) (b) A solid-walled dwelling of similar internal organization, Tepecik, eastern Anatolia. (After Esinl974, pl.100, not to scale.) (c) Dwelling plan of the Early Transcaucasian Culture from Soviet Georgia. (After Dzhavakhisvili 1973, fig.12.) (d) This reconstruction of the superstructure of (c) resembles mud-walled composite dwellings in modern Baluchistan (see Fig. 6.13). (After Dzhavakhisvili 1973, fig-13.)
—«
entrance
0
entrance
1
2
3
4
5m
Nomad archaeology
Fig. 11.5 Successive settlement structures, Early Transcaucasian Culture, Soviet Georgia. Note the expansion of the settlement in the second phase through a process of accretion, similar to that observed in modern camps undergoing sedentarization. (a) Early phase. (b) Late phase. (After Dzhavakhisvili 1973, fig. 11, not to scale.)
223
the form of composite dwellings and settlement patterns, with some evidence of phases of permanent architecture alternating with less permanent and substantial forms. The archaeological evidence accords well with the model outlined in Chapter 7 of a form of periodic nomadization alternating with lengthy phases of settlement.
Future prospects A full assessment of the existing settlement evidence from the Near East is beyond the scope of this book. I hope that the broad principles of nomad architecture; domestic organization and settlement systems outlined earlier will serve as guidelines for the many excavators troubled by the suspicion that they may be dealing with the remains of nomadic settlements or a nomadic component on a sedentary site (composite settlements). The open structure of most campsites, repeated occupation, re-use and re-arrangement of fixtures or site furniture, together with the generally preferred use of stone rather than mud as a building material, leads to the formation of relatively shallow occupation levels that do not build into conspicuous archaeological mounds. In the high-energy geomorphic environment of the Near East the survival of such deposits in an undisturbed state may be rare. As we have seen (Chapter 8) nomad camps are frequently established in sheltered locations in upland valleys and on alluvial deposits immediately below abrupt slopes. Such locations are prone to either degradation or disturbance and concealment by hill-wash. However the concealment and preservation of deposits by such processes will tend to conserve archaeological information, while substantial stone footings or walls provide ample surface indications, as at ND-1 (Chapter 10).
Nomads in archaeology
224
An understanding of the regularities in nomad spatial behaviour offers the prospect of being able to pursue the study of nomadism and of regional systems with a nomad component far back into the past. While no simple formulae for the location of ancient campsites have been discovered, certain weaknesses in existing survey designs might be suggested. In an ideal world surveys would be based on proportional or stratified sampling of all types of terrain. However the vast bulk of archaeological effort over the last two decades in Turkey and Syria has been concentrated in areas earmarked for flooding by large dam projects (1974, 1976; Middle East Technical University Ozdogan 1977), i.e. in the kinds of locations where nomad remains would be both rare and difficult to distinguish from those of more settled people. Few surveys have been carried out in the high mountain pastures or steppes where, on the evidence of this study, nomadic settlements take on their most characteristic forms. Such areas of highest intensity of nomad occupation have been purposefully located away from major population centres and areas of arable land. Surveys geared to the detection of large population centres are preprogrammed to miss the optimal locations for finding nomadic sites. Winter campsites are more likely to fall into the archaeological net, but here we are faced with the problems alluded to in Chapters 6, 7 and 8 of distinguishing substantial nomadic tentsites and campsites from sedentary architecture. Paradoxically the very features of nomad architecture that enhance its archaeological visibility only serve to confound its distinction from sedentary forms. We are therefore forced to rely not so much on the architectural elements themselves as on factors such as size, consistent orientation, organization of domestic space and discard and settlement pattern. Even so, recognizable tent and campsite features shade into the range of sedentary forms, and it is only when many lines of evidence reinforce each other that the identification of a site as 'nomadic' is likely to be secure.
12 TOWARDS A MODEL OF UNSTABLE SETTLEMENT SYSTEMS
The future of nomad archaeology in the Near East remains uncertain. With large regions closed to archaeological research through war and political instability or heavily restricted, opportunities for the kind of surveys and excavations necessary to uncover nomad sites on a large scale may not be forthcoming. But neither is this desirable. After all nomad sites should form but one component of a wider regional perspective along with many other kinds of sites. The principles outlined above should provide guidelines in designing representative surveys, recognizing the distinctive features and organization of pastoral settlements and detecting the presence of nomadic or composite dwelling forms.
Theoretical perspectives This study, I hope, has gone some way towards providing, at least for a certain class of sites, what Binford has referred to as: a descriptive and analytical procedure which attempts to define the site framework in terms of features and which is followed by a study of the relationships between this skeletal framework and the dispersion of items. (1983, p. 147) Following recent advances in intra-site spatial analysis to decipher the meaning of item distributions (Whallon 1984; Hodder 1976; Orton 1982), efforts are now being made to tackle the problem of relating such patterns to features. In Chapter 7 of In Pursuit of the Past Binford explores this interaction between site 'frameworks7 and item patterning, but his observations remain tied to particular activity spaces such as hearths, specialist work areas, etc. Although his approach is much more sophisticated than a simple search for activity areas, and due weight is given to life-space 225
Nomads in archaeology
226
maintenance, there remains a preoccupation with functionalist goals as revealed in the following extract: I am convinced that fundamental clues to the character of activities, the labor organization employed in their execution and the anticipated use of a location in terms of the overall subsistence-settlement system are coded into the organization of site structure. (Binford 1983, pp.145-6; emphasis added) This is reminiscent of the old Binford of the late 1960s. The implicit hope that archaeological data can lead directly to the understanding of whole systems is more in tune with the goals of New Perspectives in Archaeology (Binford and Binford (eds.) 1968) than the more sober objectives of the 1980s. This is curious as Binford himself (1983, p.147) is critical of Leroi-Gourhan for being overly optimistic in advancing a very modest construct of household spatial organization and discard (see LeroiGourhan and Brezillon 1966; p.254). The point is worth taking up as similar accusations have been levelled at Yellen;s (1977) nuclear area and ring model (see Binford 1978) and might equally apply to the 'domestic complex' model employed in this study. In relation to Leroi-Gourhan, Binford (1983, p. 147) suggests that archaeological data are simply not like that. Patterns of artifact density on a site cannot be used to predict the presence of other structures. Moreover the model will be totally misleading if a house is not in fact present. On the other hand, Yellen has been criticized for being particularistic (Binford 1978a). The ring model is seen as static and inflexible, its relevance bounded by the particular circumstances of Yellen;s fieldwork. Yet the ring model follows a geometric logic which is independent of any particular case. The generalization that settlement area depends on circumference, which in turn depends on the number of households, could be applied to a wide range of settlement types in many parts of the world. The essential difference between Binford and Yellen lies not in the universalism of the former as opposed to the particularism of the latter, but in Yellen's (1976) concern with broad organizational principles which apply to whole sites as against Binford's (1983) concern with the elements of which sites are composed and the sets of factors that bring them together as a functional unit of settlement. In opting for the former approach I am not implying that all nomad camps are organized in exactly the same pattern as described in Chapter 8, much less that this pattern can be applied to all kinds of sites. Instead, and in common with Yellen (1977), Clarke (1972b) and to some extent Leroi-Gourhan and Brezillon (1966) and Flannery (1976), I recognize the need for some spatial frame of reference when looking for or analysing archaeological sites. Such a frame cannot be imposed - the site itself will soon dictate which model is appropriate. Once a working model has been adopted factors of terrain, aspect, scale, orientation, 'grain7 and redundancy can be used to discern patterns on the site. This is true whether one is making on-site decisions about where to excavate next, or piecing together a settlement system from sixty-year-old site plans (Clarke 1972b) or taking a purely academic interest in the way contemporary people organize their settlements.
Towards a model
227
Comparative and diachronic approaches In a series of recent articles Hole (1974-75, 1978, 1979, n.d.) has argued that the study of ancient nomadism through archaeology is both necessary and feasible. His case, supported by both ethnographic and archaeological data from a single region of Luristan, fulfils one of the key requirements of the use of ethnographic analogy - a common regional focus. However this approach does not permit full documentation of the range of variability exhibited by nomadism either in its modern or ancient forms. The purpose of the present study has been to greatly extend the spatial range of ethnographic documentation both through original fieldwork and the assembly of existing material from various sources. The object has been not to extrapolate contemporary nomad material culture and spatial behaviour back into the past, but to try to discover the factors, and systems of factors, that give rise to the observed regularities and variability. More is involved here than the adoption of uniformitarian assumptions and arguments by analogy. If the case presented above rested solely on the assumption that nomads in the past behaved similarly to nomads in the present then it would be flimsy indeed. I have tried to show that methods can be developed which do not necessitate recourse to such assumptions. Wherever possible, contemporary observations have been supported by ethnohistorical data extending back as far as the early 1800s. While this is probably insufficient to monitor long-term cycles or evolutionary change it nonetheless provides guidelines for investigating the remote past. Variations in dwelling form, settlement pattern and discard behaviour may be related not only to seasonally but also to security of access to pasture and the degree of commitment to a nomadic pattern of migration. The study of nomadism is a well-established sub-field of social anthropology. The elements of nomadism that have so attracted ethnographers - seasonal migration, fluidity in residence associations, the formation of small cooperatives and the distinctive segmentary or tribal organizations - have been shown to derive from an essentially unstable mode of subsistence (Chapter 3). This unstable productive system has its temporal dimensions which the discipline of archaeology is uniquely equipped to pursue. But the diachronic study of nomadism will not be achieved by trying to impose ethnographically derived categories on the archaeological record. The limits of this approach have been explored in Chapter 10 of this volume. On this issue I am in full agreement with Plog. If the archaeological record is anything it is a record of long term change. And, to trie extent that arcnaeologists have failed to make use of this record, our failure can be traced to an overwillingness to borrow the largely synchronic models of change that the ethnologist employs. (1974, p.ll) This is not to say that ethnography is irrelevant to the archaeological study of change. Observations within an ethnographic context can throw important light on variations which can serve as indicators in the study of
Nomads in archaeology
228
change through archaeological data. Thus variations in dwelling form (Chapter 6) and settlement pattern (Chapter 8), hitherto discussed mainly in relation to seasonality; emerge as indicators of longer-term processes of nomadization or sedentarization. Seasonal migration and portable dwellings presuppose the capacity to readily change migration routes should the necessity or opportunity arise. The question of ancient nomadism is closely bound up with spatial and temporal variations in population distribution and density, ethnicity, architectural forms, intra-site settlement structure and changes in subsistence, to name but a few. Among nomads, the changeable size and composition of groups occupying particular localities underlines the problem of defining residence groups. It soon became clear that the categories normally employed in ethnographic analysis are inadequate for archaeological analysis. The operational construct which most closely monitors the archaeological traces left by a fluid and ephemeral settlement system was the 'domestic complex7. This may be considered the basic analytical unit in the study of nomadic campsites. Discard patterns and activities connected with the maintence of living space were found to influence not only the location of refuse and features but also the spacing and orientation of tentsites, giving rise to a symmetrical and open settlement structure. Each campsite is a single occupation event, or a succession of discrete events, by independent domestic units of fluid composition. Sites such as this invite the search for horizontal structures rather than stratigraphic sequences.
The future Like Hole (n.d.) I am confident that nomadic campsites will continue to emerge as a minor component of the archaeological record of the Near East. I am also confident that the problems of distinguishing nomad camps from those of hunter-gatherers and seasonal settlements from sedentary ones are not insuperable. The essential issue does not lie here. I remain unconvinced, however, that the mere accumulation of nomad sites, of itself, will repay the necessary effort. While the accumulation of such evidence will no doubt advance the study of pastoralism and ancient economic systems in general, the major issues that have arisen in the course of this study are of a different nature. How do nomad sites differ in their internal organization and distribution from sedentary sites? What kind of variability exists between nomad sites cross-culturally, in space and through time? What kinds of methods are required to exploit the information potential contained in this type of site, and what kind of theoretical questions are best pursued in relation to such evidence? The important issue is not whether nomadic pastoralism, however defined, evolved as a distinct specialization in 1,500 BC or 5,000 years earlier, nor whether floorplans can be identified as tents or houses. The significant finding which emerged from this research is that nomad campsites are structured in a distinctive way that bears the imprint of an inherently unstable mode of subsistence. This impression was reinforced at every level of analysis, from architectural features (Chapter 6) to the
Towards a model
229
positioning of tentsites within a camp (Chapter 8) to the location of campsites within the landscape (Chapter 10). The typical nomad camp emerges as a cluster of spatially discrete units, each containing patterned residues of a household, in which the organization of discard was specific to each household. Such a structure is not necessarily unique to nomadic pastoralists but does suggest certain facts about the community which produced it. Furthermore this kind of shallow and extensive site places heavy demands on archaeological technique. Intra-site sampling and excavation methods presuppose some predictive model of site structure an arrangement of discrete units (domestic complexes) aligned and distributed according to simple criteria of aspect, slope, optimal living space and social distance. There is a need for the further refinement of techniques of spatial analysis appropriate for the interpretation of such patterns in the archaeological record. It also emerged that variations in dwelling form and settlement pattern appeared sensitive to long-term cyclical changes related to the stability of residential groups, access to territory and subsistence strategies. This suggestion contains the seeds of many hypotheses which promise to enliven the archaeology of small settlements both within and outside the Near East for many years to come.
REFERENCES
Adams, R.M.C. 1974. The Mesopotamian social landscape: a view from the frontier. In B. Moore (ed.), Reconstructing Complex Societies. American Schools of Oriental Research No.20: 1-11. Adams, R.M.C. and Nissen, H.J. 1972. The Urule Countryside. University of Chicago Press. Adams, W.Y., Van Gerven, D.P. and Levy, R.S. 1978. The retreat from migrationism. Annual Review of Anthropology 6:483-532. Ainsworth, W.F. 1842. Travels and Research in Asia Minor, Mesopotamia, Chaldea and Armenia (2 vols). J.W. Parker, London. Alagoz, C.A. 1938. Sur la transhumance en Anatolie, particulierement nord d'Ankara. In Comptes Rondus de Congres International de Geographie. Amsterdam, pp. 5-10. Albright, W.F. 1940. From the Stone Age to Christianity, Academy of Science, New York. Allen, W.I. and Richardson, J.R. 1971. The reconstruction of kinship from archaeological data: the concept, methods and the feasibility. American Antiquity 36:41-53. Ambolt, N. 1939. Karavan: Travels in Eastern Turkestan. Blackie and Sons Ltd, London and Glasgow. Ammerman, A.J. and Cavalli-Sforza, L.L. 1973. A population model for the diffusion of early farming in Europe. In A.C. Renfrew (ed.), The Explanation of Culture Change. Duckworth, London, pp.343-8. Ammerman, A.J., Gifford, D.P. and Voorrips, A. 1978. Towards an evaluation of sampling strategies: simulated excavations of a Kenyan pastoral site. In LA. Hodder (ed.), Simulation Studies in Archaeology. Cambridge University Press, Cambridge, pp. 123-32. Andrews, P.A. 1971. Tents of the Tekna, south west Morocco. In P. Oliver (ed.), Shelter in Africa. Barrie and Jenkins, London, pp. 124-41. 1977. The Tents of Timur: an examination of reports on the Quiritay at Samarkand, 1404. In D.P. Denwood (ed.), Arts of the Eurasian Steppelands. University of London Press, London, pp. 143-69. 1981. Das Tiirkmenische Zelt. Hali 4:109-17. Angel, J.L. 1972. Ecology and population in the eastern Mediterranean. World Archaeology 4:88-105. Ascher, R. 1968. Time's arrow and the archaeology of a contemporary community. In K.C. Chang (ed.), Settlement Archaeology. National Press, Palo Alto, California, pp.43-52. Aswad, B. 1971. Property Control and Social Strategies in Settlers in a Middle Eastern Plain. Museum of Anthropology, University of Michigan, Ann Arbor, No.44. Audouze, F. and Jarrige, C. 1978. Perspectives et limites de Pinterpretation anthropologique des habitats en archeologie, un example contemporain. UArcheologie de LTraq, pp.361-81. 1981. Vestiges au sol et mode de vie: essai d'interpretation archeologique des ruines camp Pathan au Baluchistan. Estratto de Annali dellTnstituto Orientate de Napoli 4:99-117. Bacon, E.E. 1954. Types of pastoral nomadism in Central and South West Asia. South West Journal of Anthropology X: 44-68. Bagherzadeh, F. 1975. Proceedings of the Second Annual Symposium on Archaeological Research. Muzeh-e Iran-e Bastan, Teheran.
230
References
231
Baharvand, S.A. 1975. The Baharvand, Former Pastoralists of Iran. Unpublished PhD dissertation, Rice University, New York. Barnett, R.D. 1968. Review of 'Dark Ages and Nomads'. American Journal of Archaeology 72:178-9. Barth, F. 1953. Principles of Social Organization in Southern Kurdistan. Universitets Etnografiske Museum Bulletin No.7, Oslo. 1959—60. The land use pattern of migratory tribes of South Persia. Norsk Geografiska Tidskrift XVII: 1-11. 1961. Nomads of South Persia: the Basseri Tribe of the Khamseh Confederacy. Little Brown and Co., Boston. 1973a. A general perspective on nomad—sedentary relations in the Middle East. In C. Nelson (ed.), The Desert and the Sown: Nomads in the Wider Society. University of California, Berkley, pp. 11-22. 1973b. Descent and marriage reconsidered. In J. Goody (ed.), The Character of Kinship. Cambridge University Press, Cambridge, pp.3-20. Bates, D.G. 1971. The role of the state in peasant-nomad mutualism. Anthropological Quarterly 44:109-31. 1972. Differential access to pasture in a nomadic society: the Ydriik of South Eastern Turkey. In W. Irons and N. Dyson-Hudson (eds.), Perspectives on Nomadism. Brill, Leiden, pp.48-59. 1973. Nomads and Farmers: A Study of the Ydriik of South East Turkey. Museum of Anthropology, University of Michigan, Ann Arbor, No.52. 1980. Yoriik settlement in southeast Turkey. In P.C. Salzman (ed.), When Nomads Settle. Praeger, New York, pp.124-39. Bates, D.G. and Lees, S.H. 1977. The role of exchange in productive specialization. American Anthropologist 79:824-41. Beck, L. 1980. Herd owners and their hired shepherds: the Qashqa'i of Iran. Ethnology XIX:327-52. Bent, T. 1891. The Yourouks of Asia Minor. Journal of the Royal Anthropological Institute xx.269-76. Be§ikgi, I. 1969. Doguda Degisim ve Yapisal Sorunlar: Gocebe Alikan Asireti. Sevinq Maatbasi, Ankara. Bina, A. and Dyas, M. 1975-76. Earthquake in Qazvin, North West Iran, 1st September 1962. Persica VII: 173-8. Binford, L.R. 1967. Smudge pits and hide smoking: the use of analogy in archaeological research. American Antiquity 32:1-12. 1972a. Archaeological reasoning and smudge pits revisited. In L.R. Binford (ed.), An Archaeological Perspective. Seminar Press, New York, pp.52—8. 1972b. Archaeology and anthropology. In L.R. Binford (ed.), An Archaeological Perspective. Seminar Press, New York, pp.217-25. 1972c. Methodological considerations in the archaeological use of ethnographic data. In L.R. Binford (ed.), An Archaeological Perspective. Seminar Press, New York, pp.60-6. 1978a. Dimensional analysis of behaviour and site structure: learning from an Eskimo hunting stand. American Antiquity 43:330-61. 1978b. Nunamiut Ethnoarchaeology. Academic Press, New York. 1979. Organization and formation processes: looking at curated technologies. Journal of Anthropological Research 35:255-72. 1980. Willow smoke and dogs' tails: hunter-gatherer settlement systems and archaeological site formation. American Antiquity 45:4-20. 1981. Behavioural archaeology and the Pompeii premise. Journal of Anthropological Research 37:195-208. 1983. In Pursuit of the Past: Decoding the Archaeological Record. Thames and Hudson, London. Binford, L.R. and Binford, S.R. 1966. A preliminary analysis of functional variability in the Mousterian of Levallois facies. American Anthropologist 68:238-95. 1968 (eds.). New Perspectives in Archaeology. Aldine, New York. Bishop, I.B. 1891. Journeys in Persia and Kurdistan (2 vols). John Murray, London. Black, J. 1972. Tyranny as a strategy for survival in an 'egalitarian' society: Luri facts vs. an anthropological mystique. Man 4:614-34.
References
232
Black-Michaud, J. 1974. An ethnographic and ecological survey of Luristan, western Persia: modernization in a nomadic pastoral society. Middle East Studies 10:210-28. Blashfield, R.K. and Aldenderfer, M.S. 1978. The literature on cluster analysis. Multivariate Behavioural Research 13:271-95. Bogucki, P.I. and Grygiel, R. 1981. The household cluster at Brezesc Kujanski 3: small site methodology in the Polish lowlands. World Archaeology 13:59-72. Bokonyi, S. 1972. Zoological evidence for seasonal or permanent occupation of prehistoric settlements. In P.J. Ucko, R. Tringham and G. Dimbleby (eds.), Man, Settlement and Urbanism. Duckworth, London, pp. 121-6. Bonnichsen, R. 1973. Millie's camp: an experiment in archaeology. World Archaeology 4:277—91. Bos worth, C.E. 1973. Barbarian incursions: the coming of the Turks into the Islamic world. In D. Richards (ed.), Islamic Civilization, 950-1150. Cassirer, Oxford, pp. 1-16. Braidwood, R.J. 1961. The Iranian prehistoric project. Iranica Antiqua 1:3—7. Brandenburg, H.E. 1905. Kysylbasch und Jiiriikendorfer in der Gegend des Tiirkmendag. Zeitschrift Fur Ethnologie XXXVII: 188-98. Bruinessen, M. 1979. Agha, Shaikh and State: On the social and political organization of Kurdistan. Private publication, Rijswick. Bulucu, T., Goksoy, B. and Z. and Tacettin, M.T. 1977. Restoration Project of a Historic Castle Site: Castle of Corycus, Kizkalesi and Corycus Settlement. Joint dissertation for Master of Architecture, Middle East Technical University, Ankara. Burckhardt, J.L. 1822. Travels in Syria and the Holy Land. John Murray, London. Burney, C.A. 1961. Circular buildings found at Yanik Tepe in North West Iran. Antiquity XXV:237-40. 1971. The Peoples of the Hills: Ancient Ararat and the Caucasus. Weidenfeld and Nicolson, London. Cahen, C. 1968. Pre-Ottoman Turkey. Sidgewick and Jackson, London. Chang, K.C. 1967. Major aspects of the interrelationship of archaeology and ethnology. Current Anthropology 8:227-34. 1968 (ed.). Settlement Archaeology. National Press, Palo Alto, California. Chantre, E. 1924. Les Kurdes. [s.n.], Lyon. Chatty, D. 1980. The pastoral family and the truck. In P. Salzman (ed.), When Nomads Settle: Processes of Sedentarization as Adaptation and Response. Praeger, New York, pp.80-94. Chiera, E. 1924. Sumerian Religious Texts, Vol.1. Crozier Theological Seminary, Babylonian Press, Upland, Pa. Childe, V.G. 1936. Man Makes Himself. Watts and Co., London. Clarke, D.L. 1968. Analytical Archaeology. Methuen, London. 1972a. Models and paradigms in archaeology. In D.L. Clarke (ed.), Models in Archaeology. Methuen, London, pp. 1-60. 1972b. A provisional model of an Iron Age society and its settlement system. In D.L. Clarke (ed.), Models in Archaeology. Methuen, London, pp.801-70. 1977 (ed.). Spatial Archaeology. Academic Press, New York. Cole, D.P. 1973. The enmeshment of nomads in Saudi Arabian society: the case of the Al Murrah. In C. Nelson (ed.), The Desert and the Sown: Nomads in the Wider Society. University of California Press, Berkley, pp.113-28. Cook, S.R and Heizer, R.F. 1968. Relationships among houses, settlement area and population in aboriginal California. In K.C. Chang (ed.), Settlement Archaeology. National Press, Palo Alto, California, pp.75-115. Coon, C. 1965. Caravan: The Story of the Middle East. Holt, Rinehart and Winston, New York. Cooper, M.C. 1928. Grass. G.P. Putnam and Sons, New York. Cranstone, B.A.L. 1971. The Tifalmin: a Neolithic people of New Guinea. World Archaeology 3:132-42. 1972. Environment and choice of dwelling and settlement: an ethnographic survey. In P.J. Ucko, R. Tringham and G.Dimbleby (eds.), Man, Settlement and Urbanism. Duckworth, London, pp.121-6. Crawford, H. 1978. The mechanics of the obsidian trade. Antiquity Lil: 129-32. Cribb, R.L. 1982. The Archaeological Dimensions of Near Eastern Nomadic Pastoralism: Towards a Spatial Model of Unstable Settlement Systems. Unpublished PhD dissertation, University of Southampton. 1983. On-site ethnoarchaeology. Archaeological Review from Cambridge 2(2):4—16.
References
233
1984a. Out of the trenches: the potential of spatial approaches in Australian archaeology. Paper delivered at the Australian Archaeological Association Conference, Tallebudgera ; Qld ; November 16-17. 1984b. Computer simulation of herding systems as an interpretive and heuristic device in the study of kill-off strategies. In J. Clutton-Brock and C. Grigson (eds.), Animals and Archaeology: 3. Early Herders and their Flocks. B.A.R. International Series 202, Oxford, pp.161-70. 1984c. Greener pastures: mobility, migration and the pastoral mode of subsistence. Production Pastorale et Societe 14:11—46. 1985. The analysis of ancient herding systems: computer simulation in faunal studies. In G. Barker and C.S. Gamble (eds.), Beyond Domestication. Academic Press, London, pp.75-106. 1987. The logic of the herd: a computer simulation of archaeological herd structure. Journal of Anthropological Archaeology 6:376-415. Dacey, M.F. 1973. Statistical tests of spatial association in the location of tool types. American Antiquity 38:320-8. Dahl, G. and Hjort, A. 1976. Having Herds: Pastoral Herd Growth and Household Economy. Stockholm Studies in Anthropology, No.2, Stockholm. Daker, N. 1984. Contribution a Petude de revolution de Phabitat bedouin en Syrie. In O. Aurenche (ed.), Nomades et Sedentaires: Perspectives ethnoarcheologiques. Editions Recherche sur les Civilizations, Memoir No.40, Paris. d'Allemagne, H.R. 1911. Du Khorassan au Pays de Bakhtiaris, Vol.1, [s.n.], Paris. Davis, S.J.M. 1984. The advent of milk and wool production in Western Iran: some speculations. In J. Clutton-Brock and C. Grigson (eds.), Animals and Archaeology: 3. Early Herders and their Flocks. B.A.R. International Series 202, Oxford, pp.265-78. de Bode, Baron C. 1845. Travels in Luristan and Arabistan. J. Madden, London. de Boer, W.R. and Lathrap, D.W. 1979. The making and breaking of Shipibo-Conibo ceramics. In C. Kramer (ed.), Ethnoarchaeology. Columbia University Press, New York, pp.102-38. Deetz, J. 1968. The archaeological visibility of food gatherers. In R.B. Lee and J. DeVore (eds.), Man the Hunter. Aldine, Chicago, pp.281—7. Demiruren, A.S. 1974. The Improvement of Nomadic and Transhumant Animal Production Systems. F.A.O., Rome. de Morgan, J. 1894-1901. Mission Scientiflque en Perse, Vol. II. Paris. de Planhol, X. 1959. Geography, politics and nomadism in Anatolia. International Social Science Journal Xl:525-31. 1979. Saturation et securite: sur l'organization des societes pasteurs nomades. In Pastoral Production and Society. Cambridge University Press, Cambridge, pp.29-42. de Schauensee 1968. Portable architecture. Expedition 10(3):32-9. Dickson, B. 1910. Journeys in Kurdistan. Geographical Journal xxxv:357-79. Digard, J.P. 1975. Campments Baxtyari: observations d ; un ethnologie sur des materiaux interessant l'archeologie. Studica Iranica IV: 117-29. 1979. De la necessite et des inconvenients, pour un Baxtyari, d'etre baxtiyari. Communaute, territoire et inegalite chez des pasteurs nomades d'Iran. In Pastoral Production and Society. Cambridge University Press, Cambridge, pp. 127-40. Dimbleby, G.W. 1976. Climate, soil and man. Philosophical Transactions of the Royal Society of Britain 275:197-208. Dogru, A.M. 1978. Koy adlar ve Anadolunun gegmi§ti. Turk Folklor Arastirmalari. Jan., pp.8197-9. Dole, G.E. 1968. Tribes as autonomous units. In J.Helm (ed.), Essays on the Problem of the Tribe. American Ethnological Society, Seattle, pp.83-100. Dominian, L. 1915. The people of north and central Asiatic Turkey. Bulletin of the American Geographic Society XLVII:832-71. Donmez, Y. 1964. A Yoriik (nomadic) settlement west of Karasu. Review of the Geographic Institute of the University of Istanbul 9-10:58-71. Duly, C. 1979. The Houses of Mankind. Thames and Hudson, London. Dupree, L. 1963a. Deh Morasi Ghundai: A Calcolithic Site in South Central Afghanistan. Anthropological Papers of the American Museum of Natural History, No.50, New York. 1963b. Prehistoric archaeological surveys and excavations in Afghanistan, 1959-60, 1961-63. Science 146:638-40.
References
234
Durrenberger, P. and Morrison, J.W. 1977. A theory of analogy. Journal of Anthropological Research 33:372-87. Dyson-Hudson, R. 1972. Pastoralism: self image and behavioural reality. In W. Irons and N. Dyson-Hudson (eds.), Perspectives on Nomadism. Institute of International Studies, Berkley, University of California Research Series 21, pp.30—47. Dzhavakhisvili, A.I. 1973. Stroitel'noe deb i Arkhitektura poselenii Iuzhnogo Kavkaza A-IH Tysiacheletii do N.E. Tiblisi (in Georgian). Eberhard, W. 1953a. Nomads and farmers in South East Turkey: problems of settlement. Oriens 6:32-49. 1953b. Types of settlement in South East Turkey. Sociologus 1:49-64. Edelberg, L. 1966-67. Seasonal dwellings of farmers in North West Luristan. Folk 9:374-401. Edelberg, L. and Jones, S. 1979. Nuristan. Akademische Druck, Graz. Edmonds, C.J. 1922. Luristan, Pish-i-Kuh and Qala Gariveh. Geographical Journal 59:335-56; 437-53. 1957. Kurds, Turks and Arabs: Politics, Travel and Research in North East Iraq, 1919-1925. Oxford University Press, London. Ehmann, D. 1975. Bahtiyaren-Persische Bergnomaden im Wandel der Zeit. Dr. Ludwig Reichert Verlag, Wiesbaden. Ekvall, R.B. 1961. The nomadic pattern of living among Tibetans as preparation for war. American Anthropologist 63:1250-63. 1968. Fields on the Hoof: Nexus of Tibetan Nomadic Pastoralism. Holt, Rinehart and Winston, New York. Erdmann, K. 1961. Das Anatolische Karavansaray des 13 Jahrhunderts. Verlag Gebr. Mann, Berlin. Eren, N. 1980. Yoriik Gogu 3. Turk Folklor Arastirmalari. Jan., pp.8863. Ertem, H. 1972. Han Ibrahim §ah excavations, 1970. Keban Project Activities, 1970. Middle East Technical University, Ankara, pp.69-76. Esin, U. 1974. Tepecik excavations, 1971. Keban Project Activities, 1971. Middle East Technical University, Ankara. Etherton, D. 1971. Algerian oases. In P.Oliver (ed.), Shelter in Africa Barrie and Jenkins, London, pp.172-189. Evans, C. 1983. On the Jube line: campsite studies in Kurdistan. Archaeological Review from Cambridge 2(2):67-77. Evans-Pritchard, E.F. 1940. The Nuer. Oxford University Press, Oxford. Fagre, T. 1979. Tents: Architecture of the Nomads. John Murray, London. Fairservis, W.A. 1961. Archaeological Studies in the Seistan Basin of Southwestern Afghanistan and Eastern Iran. Anthropological Papers of the American Museum of Natural History, No.48, Parti, New York. Fazel, G.R. 1973. The encapsulation of nomadic societies in Iran. In C. Nelson (ed.), The Desert and the Sown. Institute of International Studies, Research Series 21, University of California, Berkeley, pp.129-42. Feilberg, C.C. 1944. La Tente Noir. Nationalmuseets Skrifter, Etnografisk Roekke II, Copenhagen. Ferdinand, K. 1959-60. The Baluchistan barrel-vaulted tent and its affinities. Folk 1:27-50; 11:34-50. 1962. Nomadic expansion and commerce in Central Afghanistan. Folk IV: 123-59. Field, H. 1939. Contributions to the Anthropology of Iran. University of Chicago Press, Chicago. Flannery, K.V. 1972. The ecology of early food production in Mesopotamia. Science 147 (3663): 1247-55. 1976a. (ed.) The Early Mesoamerican Village. Academic Press, New York. 1976b. The early Mesoamerican house. In K.V. Flannery (ed.), The Early Mesoamerican Village. Academic Press, New York, pp. 16-24. Fletcher, R. 1977. Settlement studies (micro and semi-micro). In D.L. Clarke (ed.), Spatial Archaeology. Academic Press, New York, pp.47-162. 1981. People and space: a case study on material behaviour. In LA. Hodder, G. Isaac and N. Hammond (eds.), Pattern of the Past. Cambridge University Press, Cambridge, pp.97-128. Foley, R. 1981. A model of regional archaeological structure. Proceedings of the Prehistoric Society 47:1-7.
References
235
Forbes, R. 1831. Conflict: Angora to Afghanistan. Cassel and Co., London. Forge, A. 1972. Normative factors in the settlement size of Neolithic cultivators. In P.J. Ucko, R.Tringham and G.Dimbleby (eds.), Man, Settlement and Urbanism. Duckworth, London, pp.363-76. Frankel, D. 1974. Recent developments in the Cypriot Middle Bronze Age. In B. Hennessy (ed.), The Cypriote Bronze Age. Australian Studies in Archaeology 1, Sydney, pp.23-43. Fraser, D. 1969. Village Planning in the Primitive World. Studio Vista, London. Fried, M. 1968. On the concepts of tribe and tribal society. In J.Helm (ed.)7 Essays on the Problem of the Tribe. Proceedings of the 1967 Annual Meeting of the American Ethnological Society, Seattle, pp.3-20. Friedman, J. 1974. Marxism, structuralism and vulgar materialism. Man 9:444—69. Frodin, J. 1943-44. Les formes de la vie pastorale en Turquie. Geografiska Annaler XXVI:219-72. Garnett, L.M.S. (n.d.) The Women of Turkey and Their Folklore. David Nutt, London. Garrard, A.N. 1984. The selection of south-west Asian animal domesticates. In J. CluttonBrock and C. Grigson, Animals and Archaeology: 3. Early Herders and their Flocks. B.A.R. International Series 202, Oxford, pp.117-32. Garrod, O. 1946a. The Qashqa'i tribe of Fairs. Journal of the Royal Central Asian Society XXXHI:33-44. 1946b. The nomadic tribes of Persia today. Journal of the Royal Central Asian Society XXXiii:292-9. Garthwaite, G.R. 1978. Pastoral nomadism and tribal power. Iranian Studies 11:173—97. Geary, G. 1878. Through Asiatic Turkey, Vol.11. Low, Marston, Serle and Rivington, London. Gellner, E. 1982. What is structuralism? In A.C. Renfrew, M. Rowlands and B.A. Segraves (eds.), Theory and Explanation in Archaeology. Academic Press, New York, pp.97—108. Gifford, D.P. 1977. Observations of Modern Human Settlements as an Aid to Archaeological Interpretation. Unpublished PhD dissertation, University of California. 1978. Ethno-archaeological observations of natural processes affecting cultural materials. In R.A. Gould (ed.), Explorations in Ethnoarchaeology. University of New Mexico Press, Albuquerque, pp.77—101. Gilbert, A.S. 1983. On the origins of specialized nomadic pastoralism in Western Iran. World Archaeology 15:105-19. Girshman, P. 1964. Persia from the Origins to Alexander the Great. Thames and Hudson, London. Glatzer, B. 1982. Processes of nomadization in West Afghanistan. In P.C. Salzman (ed.), Contemporary Nomadic and Pastoral Peoples: Asia and the North. Department of Anthropology, College of William and Mary, Williamsburg, pp.61-86. Godelier, M. 1977. Perspectives in Marxist Anthropology. Cambridge Studies in Social Anthropology, Cambridge. Goff, C.L. 1966. New Evidence of Cultural Developments in Luristan in the Late 2nd and Early 1st Millennium. Unpublished PhD dissertation, Institute of Archaeology, University of London. 1981. An Archaeologist in the Making. Constable, London. Goldschmidt, W. 1979. A general model for pastoral social systems. In Pastoral Production and Society. Proceedings of international meeting on pastoralism, Paris, December 1976. Cambridge University Press, Cambridge, pp. 15-28. Goody, J. (ed.) 1958. The Developmental Cycle in Domestic Groups. Cambridge Papers in Social Anthropology, Cambridge University Press, London. 1973. (ed.) The Character of Kinship. Cambridge University Press, London. 1980. Living Archaeology. Cambridge University Press, London. Gulliver, P. 1958. Family Herds: A Study of Two Pastoral Tribes in East Africa, the Jie and the Turkana. Routledge and Kegan Paul, London. Giilokstiz, G. 1980. Dogu Anadolu'daki Sosyal Degisim Iginde Beritan A§ireti. Paper presented at the Sosyal Bilimler Ulusal Kongres, Ankara. Gummerman, CJ and Phillips, D.A. 1978. Archaeology beyond Anthropology. American Antiquity 43:184-91. Giirdal, M. 1976. Antalya Yuriikleri. Turk Ethnografya Dergisi xv:65-72. Haaland, G. 1969. Economic determinants in ethnic processes. In F. Barth (ed.)7 Ethnic Groups and Boundaries. Universitets Forlaget, Bergen, pp.58-73. Hakkarill Yilligi (Hakkan Yearbook), 1967. 1972. Giirsoy Matbaacilik Sanayi, Ankara.
References
236
Hambly, G. 1964. An introduction to the economic organization of early Qajar Iran. Iran 11:69-81. Hansen, H.H. 1961. The Kurdish Woman's Life. Nationalmuseet, Copenhagen. Hart, D.M. 1970. Clan, lineage, local community and the feud in a Rifian tribe. In L. Sweet (ed.), Peoples and Cultures of the Middle East II. Natural History Press, New York, pp.3-75. Hauptmann, N. 1974. Nors,un Tepe excavations, 1971. Keban Project 1971 Activities. Middle East Technical University, Ankara. Hay, W.R. 1921. Two Years in Kurdistan: Experiences ofa Political Officer, 1918-1920. Sidgewick and Jackson, London. Hedin, S. 1898. Through Asia (2 vols.). Methuen and Co., London. Heider, K.G. 1967. Archaeological assumptions and ethnographic facts: a cautionary tale from New Guinea. South West Journal of Anthropology 23:52-64. Helm, J. (ed.) 1968. Essays on the Problem of the Tribe. Proceedings of the 1967 Annual Meeting of the American Ethnological Society, Seattle. Herodotus, 1972. The Histories. Penguin, Harmondsworth. Herskovits, M.J. 1926. The cattle complex of East Africa. American Anthropologist 28:230-72. Herzfeld, E.E. 1935. Archaeological History of Iran. British Academy, London. Hesse, B. 1984. These are our goats: the origins of herding in west central Iran. In J. CluttonBrock and C. Grigson, Animals and Archaeology: 3. Early Herders and their Flocks. B.A.R. International Series 202, Oxford, pp.243-64. Hietala, HJ. and Stevens, D.E. 1977. Spatial analysis: multiple procedures in pattern recognition studies. American Antiquity 42:539-59. Hindness, B. and Hirst, P.Q. 1975. Pre-capitalist Modes of Production. Routledge and Kegan Paul, London. Hiscock, P. 1984. Siteless Archaeology. Paper delivered at the Australian Archaeological Association Conference, Tallebudgera, Qld. November 16-17. Hodder, LA. (ed.) 1978. Simulation Studies in Archaeology. Cambridge University Press, Cambridge. 1982. Symbols in Action: Ethnoarchaeological Studies of Material Culture. Cambridge University Press, Cambridge. Hodder, LA. and Orton, C.R. 1976. Spatial Analysis in Archaeology. Cambridge University Press, London. Hodges, H.W.M. 1972. Domestic building materials and ancient settlements. In P.J. Ucko, R. Tringham and G. Dimbleby (eds.), Man, Settlement and Urbanism. Duckworth, London, pp.523-30. Hodson, F.R. 1970. Cluster analysis and archaeology: some new developments and applications. World Archaeology 1:299-320. Hodson, F.R, Kendall, G. and Tautu, P. (eds.), 1971. Mathematics in the Archaeological and Historical Sciences. Edinburgh University Press, Edinburgh. Hole, F. 1962. Archaeological survey and excavations in Iran, 1961. Science 137:524-6. 1968. Evidence of social organization for western Iran, 8000 to 4000 BC. In S. Binford and L.R. Binford (eds.), New Perspectives in Archaeology. Aldine, New York, pp.245-66. 1974-75. Tepe Tula'i: an early campsite in Khuzistan, Iran. Paleorient 2:219-42. 1975. The sondage at Tepe Tula'i. Proceedings of the 3rd Annual Symposium on Archaeological Research, Muzeh-e Iran-e Badan, Teheran. 1977. Studies in the Archaeological History of the Deh Luran Plain: The Excavation of Chaga Sefld. Museum of Anthropology, University of Michigan, Ann Arbor, Memoir No.9. 1978. Pastoral nomadism in Western Iran. In R.A. Gould (ed.), Explorations in Ethnoarchaeology. University of New Mexico Press, Albuquerque, pp. 127-67. 1979. Rediscovering the past in the present: ethnparchaeology in Luristan, Iran. In C. Kramer (ed.), Ethnoarchaeology: Implications of Ethnography for Archaeology. Columbia University Press, New York, pp.192-218. n.d. Intrasite Variability among Pastoralists of Western Iran. Unpublished manuscript. Hole, F. and Flannery, K.V. 1962. Excavations at Ali Kosh, Iran, 1961. Iranica Antiqua 2:97-148. Hole, F., Flannery, K.V. and Neely, J.A. 1969. Prehistory and Human Ecology of the Deh Luran Plain: An Early Village Sequence from Khuzistan, Iran. Memoir No. 1 Museum of Anthropology, University of Michigan; Ann Arbor. Huntington, H.G. 1972. The rate of return from the Basseri's livestock investment. Man
7:476-9.
References
237
Hiitteroth, W.D. 1959. Bergnomaden und Yaylabauen im Mitteleren Kiirdischen Taurus. Marburger Geographische Schriften 11, Marburg. Hyames, E. 1972. Soil and Civilization. John Murray, London. ibn Khaldun 1967. The Muqaddimah: An Introduction to History (translated by F.Rosenthal). Princeton University Press, New York. Imhof, E., Leuthold, C.R., Winkler, E. and Vogt-Goknil, U. 1970. Unbekannte Turkei: Wo Ost und West sich Begegnen. Kummerley Frey, Bern. Ingersoll, D., Yellen, J. and McDonald, W. 1977. Experimental Archaeology. Columbia University Press, New York. Ingold, T. 1980. Hunters, Pastoralists and Ranchers. Cambridge Studies in Social Anthropology, Cambridge University Press, London. 1987. The Appropriation of Nature: Essays in Human Ecology and Social Relations. Manchester University Press, Manchester. Irons, W. 1968. The Turkmen nomads. Natural History Nov., pp. 44-51. 1971. Variation in political stratification among the Yomut Turkmen. Anthropological Quarterly 44(3): 143-56. 1975. The Yomut Turkmen: A Study of Social Organization among a Central Asian Turkic Speaking Population. Museum of Anthropology, University of Michigan, Ann Arbor, No.58. 1979. Political stratification among pastoral nomads. In Pastoral Production and Society. Proceedings of international meeting on pastoralism, Paris, December 1976. Cambridge University Press, London, pp.361-74. Irons, W. and Dyson-Hudson, N. (eds.), 1972. Perspectives on Nomadism. Brill, Leiden. Jacobs, L. 1979. Tell-i-Nun: archaeological implications of a village in transition. In C.Kramer (ed.), Ethnoarchaeology. Columbia University Press, New York, pp.175—91. Jacobson, T. and Adams, R.M. 1966. Salt and silt in Mesopotamian agriculture. In J.R. Caldwell (ed.), New Roads to Yesterday: Essays in Archaeology. Thames and Hudson, London, pp.466—79. Jarno, R. 1984. Le role de la tente dans la formation de Pespace villageois a Qdeir (Syrie): le jeu annuel de la sedentarisation. In O. Aurenche (ed.), Nomades et Sedentaires: Perspectives Ethnoarcheologiaues. Editions Recherche sur les Civilizations, Memoir No. 40, Paris. Jarring, G. 1939. On the Distribution of Turk Tribes in Afghanistan. Lunds Universitat Arsskrift, Lund. Jaubert, P.A. 1921. Voyage en Armenie et en Perse fait dans les Annies 1805-1806. [s.n.], Paris. Jentsch, C. 1973. Das Nomadentum in Afghanistan. Afghanische Studien 9, Anton Hain Verlag, Neisenheim am Glau. Johansen, U. 1965. Die Nomaden Zelte Siidost Anatolien. Bustan 1965(2):33-7. 1966. Giineydogu Anadolunun go^men gadirlar. Ankara University Dil ve Tarih Geografya Dergisi XXIV: 1-2. Johnson, D.L. 1969. The Nature of Nomadism: The Study of Pastoral Migration in South West Asia and North Africa. Dept. of Geography, University of Chicago Research Paper No. 118. Johnson, G.A. 1973. Local Exchange and Early State Development in Southwestern Iran. Museum of Anthropology, University of Michigan, Ann Arbor, No.51. 1978. Information sources and the development of decision-making organizations. In C.L. Redman et al. (eds.), Social Archaeology. Academic Press, New York, pp. 87-112. Kademoglu, O. 1974. Yoruklerde uq direkli karagadir. Mimarhk 74:27-9. Katouzian, M.A. 1978. Oil versus agriculture: a case study of dual resource depletion in Iran. Journal of Peasant Studies 5:347-69. Kemal, Y. 1976. The Legend of the Thousand Bulls. Collins and Harvill Press, London. Ker Porter, R. 1821. Travels in Georgia, Persia and Armenia, Vol.1, [s.n.], London. Khazanov, A.M. 1984. Nomads and the Outside World. Cambridge University Press, Cambridge. Khuri, F.I. 1970. Parallel cousin marriage reconsidered. Man 5:597-618. Kintigh, K.W and Ammerman, A.J. 1982. Heuristic approaches to spatial analysis in archaeology. American Antiquity 47:31-63. Klengel, H. 1977. Nomaden und Handel. Iraq XXXIX:163~9.
References
238
Kokten, K. 1976. Stone age explorations in the Keban Dam lake area, 1972. Keban Project, 1972- Activities. Middle East Technical University, Ankara, pp.5-10. Ko§ay, H.Z. 1976. Yenikoy mound excavations 1972. Keban Project, 1972 Activities. Middle East Technical University, Ankara, pp.185-204. Krader, L. 1959. The ecology of nomadic pastoralism. International Social Science Journal XI:499-510. 1979. The origin of the state among the nomads of Asia. In Pastoral Production and Society. Proceedings of international meeting on pastoralism, Paris, December 1976. Cambridge University Press, Cambridge, pp.221-34. Kramer, C. (ed.), 1979a. Ethnoarchaeology: Implications of Ethnography for Archaeology. Columbia University Press, New York. 1979b. An archaeological view of a contemporary Kurdish village: domestic architecture, household size and wealth. In C. Kramer (ed.), Ethnoarchaeology. Columbia University Press, New York, pp.139-63. Kudat, A. 1974. Normative and alternative systems of marriage among the Yoriik of South East Turkey. Anthropological Quarterly XLVI:27O—87. Lambton, A.K.S. 1953. Landlord and Peasant in Persia. Oxford University Press, Oxford. 1973. Aspects of Saljuq-Ghuzz settlement in Persia. In D. Richards (ed.), Islamic Civilization. Cassirer, Oxford, pp. 105-26. Land, C. 1966. The admirable tents of the Shahsevan. International Archives of Ethnography v.238-47. Lattimore, O. 1951. Inner Asian Frontiers of China (2nd edn.). Capital Publishing Co. and American Geographical Society, New York. Lee, R.B. 1972. !Kung spatial organization: an ecological and historical perspective. Human Ecology 1: 125-47. 1979. The IKung San: Men, Women and Work in a Foraging Society. Cambridge University Press, Cambridge. Lee, R.B. and DeVore, I. (eds.) 1968. Man the Hunter. Aldine, Chicago. 1976. (eds.) Kalahari Hunter-gatherers. Harvard University Press, Cambridge, Mass. Lees, S. and Bates, D.G. 1974. The origins of specialized nomadic populations: a systemic model. American Antiquity 39:187-93. Leroi-Gourhan, A. and Brezillon, M. 1966. L'habitation Magdelenienne No.l de Pincevant Pres Montereau. Gallia Prehistorie IX, Fascicule 2. Levine, L.D. and Cuyler-Young, T. (eds.), 1977. Mountains and Lowlands: Essays in the Archaeology of Greater Mesopotamia. Bibliotheca Mesopotamia, Vol. 7, Undina Publishers, Malibu. Libby, W. and Hoskins, F.E. 1905. The Jordan Valley and Petra, I. G.P. Putnam's Sons, New York and London. Ldffler, R. and Friedl, E. 1967. Eine Ethnographische Sammlung von den Boyr Ahmad, Siid Iran. Archive fur Volkekunde 21:95-207. Logashova, B.R. 1982. Transformation of the social organization of Iranian tribes. In P.C. Salzman (ed.), Contemporary Nomadic and Pastoral Peoples: Asia and the North. Department of Anthropology, College of William and Mary, Williamsburg, pp.53-60. Longacre, W.A. and Ayres, J.E. 1968. Archaeological lessons from an Apache wickiup. In L.R. and S.A. Binford (eds.), New Perspectives in Archaeology. Aldine, Chicago, pp. 152-9. Luke, J.T. 1965. Pastoralism and Politics in the Mari Period. Unpublished PhD dissertation, University of Michigan. Lynch, H.F. 1901. Armenia: Travels and Studies, Vol II. Longman and Co., London. Mackay, T.S. 1968. Olba in Rough Cilicia. Unpublished PhD dissertation, Bryn Mawr. Mallowan, M. 1977. Mallowan's Memoirs. Collins, London. Marsden, D.J. 1976. The Qashqa'i nomadic pastoralists of Fars Province. In The Qashqa'i of Iran. (World Festival of Islam, Whitworth Art Gallery), Manchester University Press, Manchester, pp.9-18. 1978. The Qashqd'i Tribe and Tribal Identity. Occasional Paper No.6 Centre for Development Studies, University of Swansea. Martin, M. 1980. Pastoral production: milk and fire wood in the ecology of Turan. Expedition 22:24-8.
References
239
Marx, E. 1977. The tribe as a unit of subsistence: nomadic pastoralism in the Near East. American Anthropologist 79:343-63. Masry, A.H. 1981. Traditions of Archaeological Research in the Near East. World Archaeology 13:222-39. Masson, V.M. 1972. Prehistoric settlement patterns in Soviet Central Asia. In P.J. Ucko ; R. Tringham and G. Dimbleby (eds.); Man, Settlement and Urbanism. Duckworth, London, pp.263-78. Masson-Smith, J. 1978. Turanian nomadism and Iranian Politics. Iranian Studies 11:57-81. Maunsell, F.R. 1901. Central Kurdistan. Geographical Journal XVIII:121-4O. Melgaard, J., Mortensen, P. and Thrane, H. 1963. Excavations at Tepe Guran, Luristan. Acta Archaeologica 34:7-137. Mellaart, J. 1960. Anatolia and the Balkans. Antiquity xxxiV:270-8. Mellink, M. (ed.) 1964. Dark Ages and Nomads: 1000 BC. Nederlands Historisch-Archaeologische Institut, Istanbul. Michaud, R. and Michaud, S. 1977. Caravans to Tartary. Thames and Hudson, London. Middle East Technical University, 1965. Yassihoyuk: A Village Study. Faculty of Architecture, METU, Ankara. 1974. Keban Project, 1971 Activities. Ankara. 1976. Keban Project, 197Z Activities. Ankara. Millingen, F. (Osmanbey) 1870. Wild Life Among the Koords. [s.n.], London. Minorsky, V.F. 1978. The Turks, Iran and the Caucasus in the Middle Ages. Variorum, London. Mitchell, S. 1974. The History and Archaeology of Galatia. Unpublished PhD dissertation, Oxford University. Monteil, V. 1959. The evolution and settling of the nomads of the Sahara. International Social Science Journal XL-.572—85. Mordtmann, A.D. 1925. Anatolien: Skizzen und Reisebriefe aus KleinAsien. [s.n.], Hannover. Morier, JJ. 1837. Some accounts of the Ilyats, a wandering tribe of Persia, obtained in the years 1814 and 1815. Journal of the Royal Geographic Society Vli:236. Mortensen, P. 1964. Early village farming occupation, Tepe Guran, Luristan. Acta Archaeologica 34:110. 1972. Seasonal camps and early villages in the Zagros. In P.J. Ucko, R. Tringham and G. Dimbleby (eds.), Man, Settlement and Urbanism. Duckworth, London, pp.293-8. 1974. A survey of prehistoric sites in Northern Luristan. Acta Archaeologica 45:1-47. Miiller, E. 1938. Die Herdwanderungen im Mittlemeergebiet. Petermanns Geographische Meitteilung LXXXIV:364-70. Murphy, R. and Kasdan, L. 1967. Agnation and endogamy: some further considerations. South West Journal of Anthropology 23:1-14. Murray, P. 1980. Discard location: the ethnographic data. American Antiquity 45:490-502. Murray, P. and Chang, C. 1980. The ethnoarchaeology of a contemporary herders' site in Greece. Paper presented at the Society of American Archaeology Conference, 1980. Narroll, R. 1962. Floor area and settlement population. American Antiquity 27:587-9. Naumann, E. 1893. Vom Golden Horn zu den Quellen des Euphrat. [s.n.], Munich and Leipzig. Nelson, C. (ed.) 1973. The Desert and the Sown: Nomads in the Wider Society. Institute of International Studies, University of California, Berkeley, Research Series No.21. Nissen, H.J. 1968. Survey of an abandoned modern village in Southern Iraq. Sumer 24:107-14. Nyerges, E. 1980. Traditional pastoralism in environmental perspective. Expedition 22:36-41. Oates D. 1976. The excavations at Tell Brak. Iraq xxxix:233-44. Oberling, P. 1974. The Qashqd'i Nomads of Pars. Mouton, The Hague. O'Connell, J.F. 1979. Room to move: contemporary Alyawara settlement patterns and their implications for Aboriginal housing policy. In M. Heppell (ed.), A Black Reality: Aboriginal Camps and Housing in Remote Australia. Australian Institute of Aboriginal Studies, Canberra, pp.97-120. Oppenheim, A.L. 1977. Ancient Mesopotamia: Portrait of a Dead Civilization. University of Chicago Press, Chicago. Orme, B. 1973. Archaeology and ethnography. In A.C. Renfrew (ed.), The Explanation of Culture Change. Duckworth, London, pp.481-92. 1981. Anthropology for Archaeologists: An Introduction. Duckworth, London.
References
240
Orton, C.R. 1980. Mathematics in Archaeology. Collins, London. 1982. Stochastic processes and archaeological methods of spatial analysis. Journal of Archaeological Science 9:1-23. Ozbek-Koroglu, M. 1980. Beritanli A§ireti ve Yerle§tirme Modeli. Unpublished MA dissertation, Middle East Technical University, Ankara. Ozdogan, M. 1977. Lower Euphrates Basin, 1977 Survey. Middle East Technical University, Lower Euphrates Project Publications, Ankara. Ozergin, M.K. 1965. Anadolu ; da Selguklu Kervansaraylan. Turk Tarih Dergisi 1965:141-70. Ozkan, S. 1978. Some notes on mobile or vernacular architecture. In L. Khan (ed.), Shelter II. Barrie and Jenkins, London, pp.30-2. Pahlavi, M.R. Shah 1961. Mission for My Country. Hutchinson & Co., London. Paine, R. 1971. Animals as capital: comparison among northern nomadic herders and hunters. Anthropological Quarterly 44:157—72. Palmieri, A. 1981. Excavations at Arslantepe (Malatya). Anatolian Studies 31:101-19. Parsons, H.L. (ed.) 1978. Marx and Engels on Ecology. Greenwood Press, Westport, Connecticut. Pastner, C M . 1978. Kinship terminology and feudal vs. tribal orientations in Baluch social organization: a comparative view. In W. Weissleder (ed.), The Nomadic Alternative. Mouton, The Hague, pp.261-74. Pastner, S. 1978. Conservatism and change in a desert feudalism: the case of South Baluchistan. In W. Weissleder (ed.), The Nomadic Alternative. Mouton, The Hague, pp.347-61. 1979. Pastoral Production and Society. (Proceedings of Institute Meeting on Nomadic Pastoralism). Cambridge University Press, London. Patai, R. 1951. Nomadism: Middle Eastern and Central Asian. South West Journal ofAnthropology VII:401-14. Payne, S. 1973. Kill off patterns in sheep and goats: the mandibles from A§van Kale. Anatolian Studies 23:281-303. Perkins, D. 1964. Prehistoric fauna from Shanidar Cave, Iraq. Science 144:1565-6. Perrot, J. 1962. Palestine-Syria-Cilicia. In R. Braidwood and G. Willey (eds.), Courses Towards Urban Life. Aldine, Chicago, pp. 147-64. Peters, E.L. 1960. The proliferation of segments in the lineage of the Bedouin in Cyrenaica. Journal of the Royal Anthropological Institute 90:29-53. 1966. The tied and the free: an account of a type of patron-client relationship among the Bedouin pastoralists of Cyrenaica. In J. Peristiany (ed.), Contributions to Mediterranean Sociology. Paris, pp. 167-88. Peters, E.W. 1976. Alisam: Ein Beitrag zur Anonymen Kerpic Architecture in Osten-Anatolien. (Mitteilung No.l, Institiit fur Bauern und Planen in Entwichlungslandern), Hannover. Peters, F.E. 1978. Romans and Bedouin in South Syria. Journal of Near Eastern Studies 37:315-26. Phillips, E.D. 1972. The Scythian domination of western Asia: its record in history, scripture and archeology. World Archaeology 4:130-49. Plog, FT. 1974. The Study of Prehistoric Change. Academic Press, New York. 1976. The Qashad'i of Iran. World Festival of Islam. Whitworth Art Gallery, Manchester. Rappoport, A. 1969. House Form and Culture. Prentice-Hall, New Jersey. Rawlinson, H.C. 1839. Notes on a march from Zohab through Luristan to Kirmanshah in the year 1836. Journal of the Royal Geographic Society 9:26-116. Redding, R.W. 1984. Theoretical determinants of a herder's decisions: modelling variation in the sheep-goat ratio. In J. Clutton-Brock and C. Grigson (eds.), Animals and Archaeology: 3. Early Herders and their Flocks. B.A.R. International Series 202, Oxford, pp.223-42. Reed, C.A. 1959. Animal domestication in the prehistoric Near East. Science 130:1629-34. Reid, JJ. 1978. The Qajar Uymag in the Safavid period 1500-1722. Iranian Studies 11:117-43. Renfrew, A.C., Dixon, J.E. and Cann, J.R. 1968. Further analysis of Near Eastern obsidians. Proceedings of the Prehistoric Society 34:319-31. Richards, D. (ed.) 1973. Islamic Civilization 950-1150. Cassirer, Oxford. Robbins, L.H. 1973. Turkana material culture viewed from an archaeological perspective. World Archaeology 5:207-14. Rosenfeld, H. 1965. The social composition of the military in the process of state formation in the Arabian Desert. Journal of the Royal Anthropological Institute 95:75-86, 174-94. Roux, G. 1966. Ancient Iraq. Penguin, Harmondsworth.
References
241
Roux, J.P. 1970. Les Traditions des Nomades de la Turquie Meridionale. Librarie AndrieuMaissonneuve. Rowlett, R.M. and Pollnac, R.B. 1971. Multivariate analysis of Marnian La Tene cultural groupings. In F.R. Hodson, G. Kendall and P. Tautu (eds.), Mathematics in the Archaeological and Historical Sciences. Edinburgh University Press, Edinburgh, pp.46-58. Rowton, M.B. 1973. Urban autonomy in a nomadic environment. Journal of Near Eastern Studies 32:201-15. 1974. Enclosed nomadism. Journal of the Economic and Social History of the Orient 17:1-30. 1977. Dimorphic structure and the para-social element. Journal of Near Eastern Studies 36:181-98. Rubel, P.G. 1969. Herd composition and social structure: on building models of nomadic pastoral society. Man 4:268-73. Sagona, A.G. 1984. The Caucasian Region in the Early Bronze Age. B.A.R. International Series 214(i), Oxford. Sahlins, M.P. 1961. The segmentary lineage: an organization of predatory expansion. American Anthropologist 63:322-45. Salzman, P.C. 1971. Movement and resource extraction among pastoral nomads: the case of the Shah Nawazi Baluch. Anthropological Quarterly 44:185-97. 1972. Multi-resource nomadism in Iranian Baluchistan. In W. Irons and N. DysonHudson (eds.), Perspectives on Nomadism. Brill, Leiden, pp.60-8. 1978. Ideology and change in Middle Eastern tribal societies. Man 13:618-37. 1980. Processes of sedentarization among the nomads of Baluchistan. In P.C. Salzman (ed.), When Nomads Settle. Praeger, New York, pp.95-110. Sarre, F. 1896. Reise in Kleinasien: Sommer 1895. [s.n.], Berlin. Sawyer, H.A. 1894. The Bakhtiari Mountains and upper Elam. GeographicalJournal IV:481-5O5. Schaffer, J.G. 1972. Prehistoric Baluchistan: a systematic approach. Unpublished PhD dissertation, University of Wisconsin. 1978. The later prehistoric periods. In B. Allchin and N. Hammond (eds.), The Archaeology ofAfghanistan from Earliest Times to the Timurid Period. Academic Press, London, pp.71-185. Schiffer, M.B. 1972. Archaeological context and systemic context. American Antiquity 37:156-65. — 1976. Behavioural Archaeology. Academic Press, New York. Schmidt, E.F. 1940. Flights Over Ancient Cites of Iran. University of Chicago Press, Chicago. Schurmann, H. 1962. The Mongols of Afghanistan. Mouton, The Hague. Schuyler, E. 1966. Turkistan. F.A. Praeger, New York. Service, E.R. 1966. The Hunters. Prentice-Hall, Englewood Cliffs, NJ. Shahrani, M.N.M. 1979. The Kirghiz and Wakhi of Afghanistan: Adaptation to Closed Frontiers. University of Washington Press, Seattle. Sheil, Lady M.L. 1856. Glimpses of Life and Manners in Persia. John Murray, London. Sherratt, A. 1981. Plough and pastoralism: aspects of the secondary products revolution. In I. Hodder, G. Isaac and N. Hammond (eds.), Pattern of the Past. Cambridge University Press, Cambridge, pp.261-305. Siirt II Yilligi: 1973 (Yearbook). Giirsoy Matbaacilik Sanayi, Ankara. Simek, J.F. 1984. Integrating pattern and context in spatial archaeology. Journal of Archaeological Science 11:405-20. Smith, A. 1983. Prehistoric pastoralism in the Southwestern Cape, South Africa. World Archaeology 15:79-89. Smith, H.H. et al. 1971. Area Handbook for Iran. US Army, Washington DC. Smith, P.E.L. and Mortensen, P. 1980. Three new 'early neolithic' sites in western Iran. Current Anthropology 21:511-13. Solecki, R.S. 1964. Cave art in Kutuin Ini, a Taurus Mountain site in Turkey. Man 64:87-8. 1979. Contemporary Kurdish winter time inhabitants of Shanidar Cave, Iraq. World Archaeology 10:318-30. 1981. An Early Village Site at Zawi Chemi Shanidar. Undena Publications, Malibu. Spooner, B. 1964. Kuch U Baluch and Ichthyophagi. Iran 11:53-68. 1969. Politics, kinship and ecology in S.E. Persia. Ethnology vni:139-52. 1971. Towards a generative model of nomadism. Anthropological Quarterly 44:198-210. 1972. The status of nomadism as a cultural phenomenon in the Middle East. In W. Irons and N. Dyson-Hudson (eds.), Perspectives on Nomadism. Brill, Leiden, pp. 122-32.
References
242
1973. The Cultural Ecology of Pastoral Nomads. Addison-Wesley Module in Anthropology, No.45, Mass. Stark, F. 1933. The Pusht-i-Kuh. Geographical Journal 82:247-59. 1934. The Valley of the Assassins. Penguin, Harmondsworth. Stauffer, T.R. 1965. The economics of nomadism in Iran. Middle East Journal XIX:284-3O2. Stenning, D.J. 1957. Transhumance, migratory drift and migration: patterns of pastoral Fulani nomadism. Journal of the Royal Anthropological Institute LXXXVII:57-73. 1958. Household viability among the pastoral Fulani. In J. Goody (ed.), The Developmental Cycle in Domestic Groups. Cambridge University Press, London, pp.92-119. Stober, G. 1978. Die Afshar: Nomadismus im Raum Kerman (Zentral Iran). Marburger Geografische Schriften, Marburg/Lahn, Heft 76. Stronach, D. 1963. Excavations at Pasargadae: first preliminary report. Iran 1:19-42. Siimer, F. 1980. Oghuzlar (Turkmenler): Tarihleri-Boy Teskilati-Destanlan. Ana Yayinlan, Istanbul. Sumner, W. 1979. Estimating population by analogy: an example. In C. Kramer (ed.), Ethnoarchaeology. Columbia University Press, New York, pp. 164-74. Sunderland, E. 1968. Pastoralism, nomadism and the social anthropology of Iran. Cambridge History of Iran I. Cambridge University Press, Cambridge. Sweet, L. 1974. Tell Toqaan: A Syrian Village. Museum of Anthropology, University of Michigan, Ann Arbor, Anthropological Paper No. 54. Swidler, N. 1972. The development of the Kalat Khanate. In W. Irons and N. Dyson-Hudson (eds.), Perspectives on Nomadism. Institute of International Studies, University of California, Research Series No. 21, pp.115-22. Swidler, W.W. 1972. Some demographic factors regulating the formation of flocks and camps among the Brahui of Baluchistan. In W. Irons and N. Dyson-Hudson (eds.), Perspectives on Nomadism. Brill, Leiden, pp.69-75. 1973. Adaptive processes regulating nomad-sedentary interaction in the Middle East. In C. Nelson (ed.), The Desert and the Sown. Institute of International Studies, University of California, Research Series No. 21, pp.23-42. Swift, J. 1977. Sahelian pastoralists: underdevelopment, desertification and famine. Annual Review of Anthropology 6:457-78. Sykes, E.C. 1898. Through Persia on a Side-Saddle. Innes and Co., London. Sykes, M. 1908. The Kurdish tribes of the Ottoman Empire. Journal of the Royal Anthropological Institute xxxvlii:451-86. Tapper, N. 1977. Pashtun nomad women in Afghanistan. Asian Affairs 1977:163-170. Tapper, R.L. 1974. Shahsevan in Safavid Persia. Bulletin of the School of Oriental and African Studies 2:321-54. 1979a. The organization of nomadic communities in pastoral societies of the Middle East. In Pastoral Production and Society. Proceedings of international meeting on pastoralism, Paris, December 1976. Cambridge University Press, Cambridge, pp.43-65. 1979b. Individuated grazing rights and social organization among the Shahsevan nomads of Azerbaijan. In Pastoral Production and Society. Proceedings of international meeting on pastoralism, Paris, December 1976. Cambridge University Press, Cambridge, pp.95114. 1979c. Pasture and Politics: Economics, Conflict and Ritual among Shahsevan Nomads of N.W. Iran. Academic Press, London. Tchalenko, G. 1953. Villages Antiques de la Syria du Nord. Libraire Orientaliste Paul Gaunthner, Paris. Thesiger, W. 1964. The Marsh Arabs. Longman, London. 1979. Desert, Marsh and Mountain. Collins, London. Thrane, H. 1964. Archaeological investigations in Western Luristan. Acta Archaeologica 35:153-69. Torrence, R. 1986. Production and Exchange of Stone Tools: Prehistoric Obsidian in the Aegean. Cambridge University Press, Cambridge. Tribe, D.E. 1950. Influence of pregnancy and social facilitation on the behaviour of grazing sheep. Nature 166:74. Trigger, B.G. 1968. The determinants of settlement patterns. In K.C. Chang (ed.), Settlement Archaeoloogy. National Press, Palo Alto, pp.53-75. Tuncdilek, N. 1962. Observations on rural settlement in two different regions of Turkey. Review of the Geographical Institute of the University of Istanbul 7 A7-56.
References
243
1963. Yayla settlements and related activities in Turkey. Review of the Geographical Institute of the University of Istanbul 9:58-71. Turan7 O. 1971. Selcuklular Zamaninda Turkiye: Siyasi Tarih, Alparslan'dan Osman Gazi'ye, 1071-1581. Nesriyat ve Matbaacilik Kollektif, Istanbul. Tiirkay, C. 1979. Osmanli Impiratorlugunda Oymak, Asiret ve Cemaatlar. Ba§bakanlik Ar§ivi, Tercuman Kaynak Eserler Dizisi. Ubicini, A. 1970. Letter on Turkey, 1855. Machette, Paris. Ucko, P.J. 1969. Ethnography and archaeological interpretation of funerary remains. World Archaeology 1:262-77. Vainshtein, S. 1980. Nomads of South Siberia: The Pastoral Economies of Tuva. Cambridge University Press, Cambridge. Vambery, A. 1868. Sketches of Central Asia: Russia Observed. Arno Press and New York Times, 1970. Vardiman, E.E. 1977. Nomaden: Schopfer einer neuen Kultur im Vorderen Orient. Economisch Verlag, Wien, Diisseldorf. Vincze, L. 1980. Peasant animal husbandry: a dialectical model of techno-environmental integration in agro-pastoral societies. Ethnology XIX:387-4O3. Voigt, M.M. 1977. The subsistence economy of a sixth millennium village in the UshnuSolduz Valley. In L.D. Levine and T. Cuyler-Young (eds.), Mountains and Lowlands. Undina, Malibu, pp.307-46. Wagner, M. 1852. Reise nach Persien und des Ldnden der Kiirden, Vol I. [s.n.], Leipzig. Watson, P.J. 1966. Clues to Iranian prehistory in modern village life. Expedition 8:13-23. 1978. Architectural differentiation in some Near Eastern communities, prehistoric and contemporary. In C. Redman et ai (eds.), Social Archaeology. Academic Press, London, pp.131-58. 1979. Archaeological Ethnography in Western Iran. Viking Fund Publications in Anthropology, University of Arizona Press, Tucson, No.57. Wenzel, H. 1937. Forschungen in Inner Anatolien II. Geographische Schriften der Universitd't Kiel, Schriften No. 7. Whallon, R. 1973. Spatial analysis of occupation floors I: the application of dimensional analysis of variance. American Antiquity 38:266—78. 1974. Spatial analysis of occupation floors II: the application of nearest neighbour analysis. American Antiquity 39:16- 34. 1984. Unconstrained clustering for the analysis of spatial distributions in archaeology. In HJ. Hietala (ed.), Intrasite Spatial Analysis in Archaeology. Cambridge University Press, Cambridge, pp. 242-77. Whallon, R. and Kantman, S. 1969. Early Bronze Age development in the Keban reservoir, East Central Turkey. Current Anthropology 10:128-83. Wheeler Pires-Ferreira, J. 1977. Tepe Tula'i: faunal remains from an early campsite in Khuzistan, Iran. Paleorient 1975-76-77:275-80. Whitelaw, T. 1983. People and space in hunter-gatherer camps: a generalizing approach in ethnoarchaeology. Archaeological Review from Cambridge 2(2):48-66. Wiessner, P. 1982. Beyond willow smoke and dogs' tails : a comment on Binford's analysis of hunter-gatherer settlement systems. American Antiquity 47:171-8. Williams, G. 1972. Eastern Turkey: A Guide and History. Faber and Faber, London. Wilson, Sir A T . 1926. The Bakhtiaris. Journal of the Royal Central Asian Society xlil:205-25. Winter, M.C. 1976. The archaeological household cluster in the valley of Oxaca. In K.V. Flannery (ed.), The Early Mesoamerican Village. Academic Press, New York, pp.25-30. Woodburn, J.C. 1968. Stability and flexibility in Hadza residential groupings. In R.B. Lee and I. DeVore (eds.), Man the Hunter. Aldine, Chicago, pp.25-30. 1972. Ecology, nomadic movement and the composition of local groups among hunter gatherers: an East African example and its implications. In P.J. Ucko, R. Tringham and G.W. Dimbleby (eds.), Man, Settlement and Urbanism. Duckworth, London, pp. 193-206. Woods, J.E. 1976. The Aqquyunlu: Clan, Confederation, Empire. A study in 15th/i6th century Turko-Iranian politics. Bibliotheca Islamica, Minneapolis and Chicago. Wright, G.A. 1969. Obsidian Analysis and Prehistoric Near Eastern Trade 7500 to 3500 BC. Museum of Anthropology, University of Michigan, Ann Arbor, No.37.
References
244
Wright, H.E. 1968. Natural environment of early food production north of Mesopotamia. Science 161:224-39. Wright, H.T. 1977. Recent research on the origin of the state. Annual Review of Anthropology 6:379-97. 1981. An Early Town on the Deh Luran Plain: Excavations at Tepe Earukhabad. Museum of Anthropology, University of Michigan, Ann Arbor, Memoir No.3. Yalgmlar, I. 1972-73. Volcan eteint de Nemrut et sa caldeira. Review of the Geographical Institute of the University of Istanbul 14:203-12. Yalman, A.R. 1977. Cenupta Turkmen Oymaklar. I and II. Kiiltiir Bakanligi Yaymlan (Ministry of Culture), Ankara. Yellen, J.E. 1976. Settlement patterns of the !Kung: an archaeological perspective. In R.B. Lee and I. DeVore (eds.), Man the Hunter. Aldine, Chicago, pp.47-72. 1977. Archaeological Approaches to the Present: Models for Reconstructing the Past. Academic Press, New York. Zagerell, A. 1975. Archaeological Survey in the North West Bakhtiari Mountains. In F. Bagherzadeh (ed.), Proceedings of the Second Annual Symposium of Archaeological Research, Iran. Muzeh-e Iran-e Bastan, Teheran, pp. 145-8.
GLOSSARY
Term
Language
Translation
Akkoyunlu
Turkish Turkish
Akraba Alacik Alpenwirtschaft
Turkish Turkish German
A§iret Beyleh Bonkuh Chadour Chul
Turkish/Kurdish Qashqa'i (Turkic) Qashqa'i (Turkic) Persian Persian
£adir Qardak Qatma ev
Turkish Turkish Turkish
Dag
Turkish Turkish Turkish Turkish Persian Persian Kurdish
Influential headman or landowner. Nomad confederacy and empire in Eastern Turkey, 15th century. Relatives or consanguinial kin. General term for trellis and other frame tents. An alpine pastoral economy based on seasonal vertical transhumance of livestock. Tribe. Minimal camp group. Large camp group. Tent. Storage platform along the inside wall of a tent (seeYukluk). Tent. Leaf-covered bower or summer shelter. Literally 'makeshift' tent; semi-cylindrical frame tent. Mountain. Pole (see nq direkli qadir). Old, ancient. House, dwelling. Winter quarters (see ki§lak). Rug or small carpet (see kilim) A nomad household. Archaeological mound. Lower order tribal section (Jaf Kurds). Tribe or tribal confederacy. Chief of a tribal confederacy. A tribal section corresponding to a named descent group. Capture, kidnapping. Black tent supported by poles. Headman or leader of Basseri oulad.
Aga
Direk Eski
Ev Garmsir Gelim Hane Hoyuk
Hoz D
Turkish
Ilkhan Kabile
Kurdish Persian Persian Kurdish
Kaqirma Kara qadir Katkhoda
Turkish Turkish Persian
245
246
Glossary
Kaymakam Khan Khel
Turkish Persian/Mongol Persian
Village headman. General term for chief. Tribal section (Baluch, Pashtun) or camp group
Kibitka
Turkish/Mongol
Kilim Ki§lak Kiz kaqirma Kom§u Ksour Kula Kurban Kurban bayram
Turkish Turkish Turkish Turkish Arabic Persian Turkish Turkish
Kurgan
Russian
Malik Mtistilman
Persian/Pashtu Turkish Turkish Turkish Persian
Covered wagon used on the Central Asian steppes, also applied to trellis tents. Small carpet or rug (see Gelim). Winter quarters. 'Girl capture', or elopement. Neighbour. Village compound or hamlet. Leaf-covered bower or summer shelter. Sacrifice. Feasting time at the end of the holy month of Ramazan. Mounded tombs in the Caucasus region, dating to the 2nd and 1st millennia BC. Tribal chief or headman (Pashtun). Muslim. A migratory unit (Yomut Turkmen). Migratory Turkish tribes of the 11th century. Agnatically defined migratory group among the Basseri. Tribe or large migratory unit. Headman (Jaf Kurds). An organized plundering expedition. Robbers, thieves, bandits. Metal shield for making yufka bread. Iron tripod to support §ac. Persian dynasty, 17th and 18th centuries. Art or artwork. Sand-dweller, or Bedouin. Summer quarters (see yayla). Persian dynasty, 2nd-4th centuries. A form of pastoral tenure in which livestock are looked after by relatives or clients. Turkic empire occupying much of Anatolia in the 12th and 13th centuries. Tribal or religious leader. Tray for serving meals. Black tent supported by poles. Loom. Tribal section (Qashqa'i). Copper bowl. Lower order descent group. Tribal section (Pashtuns, Shahsevan). Tribe (Bakhtiari).
Oba Oghuz(Oguz) Oulad Oymak Raiz Razzia Raubern Saq ayak Safavid Sanat Sandbewohner Sarhad Sassanian Saun
Turkish Kurdish French German Turkish Turkish Persian Turkish German Persian Persian Mongol
Seljuk/Selquk
Turkish
Sheikh Sini Siyah chadour
Arabic Kurdish Persian Kurdish Qashqa'i (Turkic) Kurdish Persian Persian/Turkic Persian/Turkic
Sa9
Spi Taifeh
Tas Tash Tayfa Tayfeh
(Jaf).
247
Glossary
Tell
Arabic
Tepe Te§i Tezek Tireh
Persian Turkish Turkish Persian
Toklu Topak ev Turkmen
Turkish Turkish Turkish
Uq direkli qadir Uluki§i Wandern Wilden Yapilan direk evler Yayla
Turkish Turkish/Kurdish German German Turkish Turkish
Yoruk/Yiiriik
Turkish
Yufka
Turkish
Yiikluk
Turkish
Yurt Zelt Zeltbewohnern Zemga
Turkish/Mongol German German Persian
Zoma
Kurdish
Archaeological mound (Turkish, hoyiik; Persian, tepe). Small hill or archaeological mound. Spindle. Dried dung burnt as fuel. Tribal section among western Iranian nomadic groups and Shahsevan. A herd consisting of male yearlings. Circular trellis tent. A major branch of the Turkish-speaking peoples. See also Yoruk. Three-poled (black) tent. Tribal or religious leader. Wanderers, nomads. Wild or barbarian. Houses made of poles or wattle-and-daub. High mountain pastures and summer camping ground. Literally 'wanderer'. Sub-ethnic group inhabiting southern and western Anatolia, not all nomadic. Also known as Turkmen. Thin 'pide' bread cooked on a domed metal shield or saq. Storage platform along the inside wall of a tent (see chul). Tent of Central Asian type. Tent. Tent-dwellers. A stone-walled dwelling partly covered by a black tent. A summer camp group of changing composition.
INDEX
abandoned tentsite 92, 166 abandonment 220 accretion 158 accumulation 34, 37, 40 of livestock 120 rate of 153 spiral of 40 activity areas 225 sets 124 space 4 affinal relations 39, 45, 48, 49, 54, 145,155,165,195 Afghan Pamirs 42 Afghanistan 220 aga 198, 200 age cohorts in stock 29 age-sex composition of herds 29, 32,67 agnatic camp groups 145, 165, 193, 195 core 48 descent 45, 46, 49, 51, 46 structure 146 agriculture 11, 12, 24, 27, 34, 61, 136, 144, 159,165, 195 agro-pastoral system 15, 25, 62, 196, 214 Akkoyunlu 57, 196 akraba 164 alaak 86, 88, 104, 166, 169 Algeria 158 Ali Kosh 216 alienation 40, 41 of land 202 Alikan Asjret 187, 196, 198, 202, 210 Alikan Mevki 219 Alikan tribe 197 Alikanli54, 101, 107, 110, 135-7, 140-7, 150-5, 189, 196-9, 200-8,211 alignment 140-1, 170 alluvial ash 185 deposits 114, 165,218,223 plain 13, 188 valleys 137 alpenwirtschaft 19 Amorites 10 Anatolia 16, 75-82, 86, 90-1, 95-6, 99, 101, 104, 107, 110-1,121,134,139, 158, 164, 187, 196, 205, 220 Anatolian nomads 115, 154 Anatolian plateau 8, 114, 162 angora goats 117
248
animal bone 131, 177, 188, 215, 218 pens 195 prices 117 products 69, 166 annexes 189 anomaly, ethnographic 5 anthropologists 53 apical ancestor 201 Aqqoyunlu Empire 154 Arabia 16 archaeological context 5, 68-9, 74, 76, 84, 97, 123 culture 65 data 192, 209 deposits 146 fieldwork 212, 227 material 161 mounds 223 objectives 171 profile 122 record 68, 75, 154 remains 22 sites 150 study 112 surveys 218 tell 139 visibility 69 architectural features 152, 213, 214, 221 form 6, 133, 153, 166 typology 107 Armenia 220 artifact density 172, 226 distribution 172, 178 texture 172 asiret 197, 199, 201, 203 ash dump 2, 176, 190,207,210, 214 aspect see tent orientation asymmetric reciprocity 82 Assyrian 11, 13 Atlas Mountains 25 baggage train 76, 198 Baharvand 38, 57,74, 95,100, 107, 140, 147 Bakhtiari 26, 57, 60, 100, 107, 141, 145, 157 Khans 155 campsites 146 Baluchi 10, 62 campsites 154, 221 Baluchistan 90-2, 96-8, 154, 169
barrel-vaulted bender 89 huts 89 tent 88 Barth, F. 39, 40, 51 Basseri 39, 40, 49, 53-6, 60, 134, 135-7, 144, 147-9, 200 bazaars 198 bedding platform 214 Bedouin 10, 11, 85, 92, 94, 150 campsites 75 migration 13 bender 88, 89 Beni Hasan 10 Berbers 54 Beritan tribe 110 Beritanli 98, 101, 104, 140, 144-7, 198, 205 Be§ikc.i, I. 35, 69, 155, 187-9, 196-7,200-11 beyleh 25, 39 bilateral family ties 49 Binford, L.R. 2, 68, 130, 226 bio-social factors 28 black burnished pottery 14, 220 black tents 84, 88, 90-1, 108, 142, 167,188, 205, 219 blood feud 46, 53-4, 200 boom and bust cycle 34 Boyr Ahmad 55, 141 bread oven 77, 189, 192-5 pit 80, 191 breeding stock 29, 33 brideprice 38, 39 butchery 124, 131, 182 Byzantine heritage 84, 114-5, 118, 123, 139,150,220 caldera 186, 188,211 camel bag 118, 123 camels 118 camp group 37, 45, 49, 121, 143-6, 199, 203 formation 20, 45, 48, 57, 110, 139, 143, 155 membership 110, 199 campsite 44, 49, 80, 137-8, 143, 157,164,184,187,211,214 contemporary, 76 formation 6, 137, 146, 147 organization 139, 144, 184, 214-5 temporary, 67 capital 34, 38-42, 119 agricultural, 34 investment 24
249
Index
losses 24
maintenance 37 pastoral, 23, 34, 37-40 requirements 40 resources 48 restricted, 34 caravanserai 150, 182, 196 carbon-14 dating 214 Central Asia 6, 42, 59, 86-9, 91, 95,149, 155, 220 nomadism in, 15 tents in, 96 ceramic 68, 75, 76, 78, 80, 191, 213, 214, 220, 221 cooking pots 92 forms 221 pots 66 sherds 188 styles 14, 127 chaleh 92 see also hearth
chiefdoms 48, 52, 104 Childe, V.G. 65, 66, 80 chronological arguments 11, 213 cW92, 214 Cimmerians 11 cities, formation of 155 climate 134, 164 climax vegetation 28 clustering 174 co-resident communities 20, 34, 39 groups 54 cognates 45, 49, 164, 195 collateral households 48 collective pasture rental 56, 200, 201 tenure 34, 56, 119 commerce 117, 119, 147 comparative archaeology 227 complex societies 10 composite dwelling 96-7, 154, 159, 160, 216, 223-5 herd 144 settlement 154-7, 216, 223 computer simulation of a herding system 29-34 confederacy 53-4, 199, 201-2 conflict, tribal 45, 55 consumption of food 21, 69 contract herding 34, 39, 40, 42, 119 cooking pots 75-7, 80, 124 cooperative groups 36, 143, 144, 227 corrals 66,67-9,96,105,115,118, 150-1,156, 173,180, 190, 192-4,195, 209-10, 220 courtyards 159 culling herds 29 cultivation 14, 25, 38, 61-2, 118, 135-6,139, 218 culture complex 76, 91 cycle of movement 28, 58 yirdak 90-1, 96 yuma ev 88, 89, 166, 185 Darius 1, 149 dark ages 11 defence 144 Deh Luran sequence 213 Deh Luran Plain 13 Deh Morasi Ghundai 220
demographic composition of flocks 29 of nomads 61, 189,191, 192, 208-9 density of settlement 117, 145, 155-8, 210, 228 descent groups 45, 49, 51, 120, 200 diachronic studies 15, 227 differential preservation 176 dimorphic chiefdom 26, 155 discard 172, 226 behaviour 227 organization of, 229 patterns 6, 224 zone 128-30, 132, 170, 173-6, 180,182, 214, 215 domain 100-1, 104, 105, 170 domestic activities 210 behaviour 124 complex 131-2, 138, 170-3, 180, 183, 184,191, 215, 218, 224, 226, 228-9 cycle 38 goods 69 units 193 domestication 9 dominant lineage 48 drought 31, 38 dry farming 13 durables 68, 111, 178 dwelling form 105-7, 160, 228 unit 161 dyeing yarn 210 dynamics, archaeological 5 dynastic cycle 58 dynasty 26 Early Bronze Age 185 Early Transcaucasian Culture 14, 220-1 earthenware 79 see also ceramic earthquake 160 eastern Anatolia 137, 144, 218, 221 eastern Turkey 139 ecological balance 135 controls 30 factors 48 imbalance 25 integration 15 succession 28 stratification 62 viability 40 egalitarianism 34, 40, 56, 147 Elburz Mountains 74, 91 Emirates 52 endogamy 49, 201 enforced sedentarization 60, 157 environmental capacity 30, 62, 64, 135 factors 4, 23, 139, 184 niche 22 stress 58 epidemic among livestock 31, 38, 40 essentialism 4 estate 55, 135 ethnic affiliation 10, 91, 105, 154 composition 62, 63, 228
migration 91 ethnoarchaeology 4, 5, 65-6, 122, 193, 196 ethnographic analogy 4, 5, 123, 214, 227 observations 62, 196, 199, 211 parallels 5 record 148, 227-8 studies 31, 45, 46,56,57, 112, 207 techniques 120, 187 ethnohistorical observation 62,69,74,108,112, 227 sources 15, 78 Eurasia 16, 27 evolutionary change 227 excavation 68, 170, 196, 212-4, 229 expendables 68, 69 exponential expansion 30 growth 33 extended family 40, 101, 193, 199,208 household 193, 208 external hearth 170, 192-3 famine 57 farming dry 13 mixed 27, 59, 63 Fars Province 26 faunal elements 178 evidence 215 population 215 remains 67 Fertile Crescent 10 fertility rates 28, 29 fixtures 66-9, 76-9, 84, 92, 97, 100, 105, 108, 110-1, 119, 151,195, 208, 223 flock leaders 28 flocks 37, 143, 204 floorplan 3, 91, 97-8, 105, 154, 191-4, 216, 221, 228 fodder store 192, 208 food production 9, 16, 68, 214 forces of, 22, 24, 27, 37 forecourt 184, 189, 191, 210, 216, 219, 221 fortified camps 154, 155 founding ancestor 201 functionalism 6, 172, 226 garmsir 134
genealogy 201 geographical scope 6 goat-hair tent 91, 195 grain 19, 82 grass 145 grazing effects of 188 fees 119 land 34, 135, 139, 143, 204 night, 209 rights 24, 42, 48-9, 202 selective, 28 shortage 31, 117-9, 198-9, 201 grinding stones 189, 193 group cohesion 146 guest accommodation 105, 148, 191,194-5, 205, 209, 210
Index
HajiFiruzTepe218 Hamidiye regiments 196 Han Ibrahim §ah 219 hane 199, 208 Haneans 10, 48,195, 200, 209, 214 hearth 2, 66, 84, 92,105, 110, 151, 170, 189, 191-3, 207, 214, 219, 221, 225 Hellenistic period 114 herding 38, 69, 124, 178 contract, 39 efficiency 40, 41 group 40 labour 24, 37-9, 48, 165 logistics 20, 44 system 32, 33 tactics 117,204 unit 34, 37, 39-41, 49, 144-7, 156, 194, 198, 200, 203 Herodotus 12, 59 heirlooms 75 high-risk pastoral strategy 34 historical evidence 10-2 Hittite raiding 11 hospitality 82, 88, 101, 104, 187 house forms 161 house plans 142 household 25-7, 30, 34-6, 44, 49, 113,143,149,183,205 autonomy 20, 145 composition 38, 193 consumption 42 equipment 69, 74, 75 flocks 36, 41 head 200, 205 husbandry 16 impedimenta 80, 81 inventory 74, 75 labour 35 organization 82, 98, 126, 131, 142, 214 size 35, 88, 101, 105, 195 space 85 unit 40, 44, 131 utensils 65 viability 36-9, 40, 42 Hulailan Valley 97 hunter-gatherer 20, 22, 48, 66-7, 80, 82, 155-6, 228 husbandman 37, 165 husbandly 16, 34 ibn Khaldun, 23, 53, 58 Ilkhan51 ingestion, rate of 28 inheritance 21, 34 integrated tribes 15, 26, 155 intensive agriculture 6, 120, 136 intra-site density 156 sampling 229 settlement 228 spatial analysis 161 invasion 58, 153 Iran 68, 74, 79, 82, 89, 91, 104, 111, 134, 145, 149, 153, 157-8, 183-5, 193, 196, 220 Iranian Revolution 60 Iraq 158 Iron Age 219 irrigation 135, 153 agriculture 13 canal 158, 212 Jamenites 10
250
kabile 144, 199-203, 209
Lurs51,96, 107, 108
leara $adir 86
karst7, 114
katkhoda 49 kaymakam 198 Khabur Valley 10, 11 Khamseh 26 Confederacy 51, 53, 149 khan 26, 42, 51, 55-6 Khuzistan 31, 68, 212, 216 kibitka 59, 86 kilim 82, 123, 165 kin support 36, 39, 41 kinetic factors 3, 4 studies 100 kinship 39, 44, 45, 53-4,133, 145, 146, 155, 157, 193-9, 209 Kirghiz 42, 63, 86 kislak 134, 165 leiz kaqirma 164, 165
kula 81, 90-1, 96-7, 108, 214, 221 Kurdish camps 139 ethnicity 197 nomad tribes 187 tribes 195-6, 198, 205 villagers 106, 154 Kurdistan 52, 61, 92 Kurds 69, 74, 104 kurgans 1 labour efficiency 40 force 24, 38, 39, 101 imbalances 41, 42 requirements 23, 24 resources 24, 42 shortages 39, 41 supply 35, 38, 39, 49, 56 labour-intensive system 12 labour-sharing 45 Lady Ella Sykes 78 Lake Van 185, 196-7,202 lambing season 29 land tenure see tenure land use 61 landlords 40, 200 Late Bronze Age 67 lattice-work 86 lava flow 137, 188, 210 laws 2 Levant 13 life-space 3, 131, 225 see also living space lineage 48, 62, 203, 209 maximal, 46 segment 45-8, 53 lithic technology 79 livestock 27, 30, 37-41, 55, 63, 97, 117,124,138,158,198,203, 209 holdings 32, 62 management 36 numbers 135 ownership 200 resources 55 wealth 35 living space 98, 99, 161 zone 132, 170, 173, 176, 215 Luri 38, 74, 79, 146, 151, 156, 213-4, 216 Luristan 15, 57-61, 91, 96-8, 135, 227
Malthusian controls 29, 40, 41 Mari archives 10, 11 marine fossils 7 market capitalism 34 forces 27, 40, 42,56, 119, 120, 166 niche 25 prices 28, 42, 117 trends 118 marriage 21, 38, 40, 62, 165, 201 isolates 49, 51, 53, 201 Marsh Arabs 54 Marxist theory 17, 52 material complexes 14 culture 65-83, 124, 211,221, 227 culture complex 76 remains 65, 83 maximal lineage 46, 200-1 mercenaries 54 Mesopotamia 12, 67, 91, 108, 155 metal containers 75 tripods 220 midden 2, 80, 161, 183, 190-3, 210, 214-5 middle range theory 2-6 migration 73, 118, 134, 144, 164, 166, 198-9, 227 argument for, 11, 58, 153 Bedouin, 13 community 18,19,22,120,197, 199 cycle 18, 54, 108 drift 58 episode 58 lifestyle 153 mode 21 orbit 107, 215 route 18, 28, 61, 154, 197, 212-4, 228 schedule 31, 49 track 20-2,59, 110-4, 120, 133, 146 trail 198 trains 82 trends 108, 120 military services 202 milk products 29, 138, 165 Millie's camp 122 minimal lineage segment 193, 195 mobile lifestyles 13, 14, 15 resource base 63 villages 44 mobility 97, 107,221 mode of production 3, 17-8 mode of subsistence 12, 17-8, 23-6, 30, 34, 61-2, 147, 153, 227-8 model building 3 modernization 60, 62 modular units 3, 144-5, 148, 156 Mongol incursions 12 Mongolia 86, 155 monument ruins 151 Morocco 92 mortality rates in flocks 28, 29 mounds 151 mountain regions 12 see also Taurus Mountains
251
Index
mud walls 155 multiple resource nomadism 39
overexploitation 27 oymak 49, 198
Near Eastern history 10 nomads 75 villages 69 Nemrut Dag 77-80, 110, 134, 185-8, 197,204,211 Neolithic sites 151, 185, 218 New Archaeology 4, 122 night grazing 209 nomadic activity 15 archaeology 66, 152, 211 architecture 80, 84, 224 camp 44-5, 69, 133, 153, 155-6, 161, 229 campsites 20, 68, 134, 216 culture 221 dynasty 196 entrepreneurs 12, 14 families 198 household 69 incursions 11 invasions 11 lifestyle 122 material culture 66, 68, 73-6, 80,83,211 migration 14, 97, 154 mobility 45 pastoralism 9, 18, 34, 81, 135, 228 pastoralists 66, 80, 82, 113 population 62, 115, 135 sections 15, 40, 42, 63 sedentarization 154, 156-61 settlement 67, 223-4 sites 68,75-6, 96,133,157,192, 212, 224, 225 spatial behaviour 224 tents 85, 100, 151 traits 97 tribes 199 nomadism compared with sedentism 122, 161, 166 definition of, 15, 19, 68 multiple resources of, 39, 119 tethered 18 untied 18 nomadization 5, 25-7, 59, 62-3, 111,120-1,153,223,228 North Africa 46 nuclear area 129, 131-2, 226 nuclear family 38, 40, 101, 193-5, 200, 203-8 nucleation 145, 153
pack animals 80-1, 101 Pahlevi Dynasty 60 paleobotanical evidence 28 Palestine 220 parallel cousin marriage 45, 165 Pashtuns 42, 91 pastoral accumulation 18, 63 activity 166 camp 68, 219 campsite 11,68, 183 capital 23, 34, 37-8, 40, 63, 117 economy 49 feudalism 42 herds 56 labour 35, 37 migration 16, 60, 118 mode of subsistence 24, 34, 147 nomadism 38 nomads 52 opportunism 58 peoples 2, 27, 30 production 11, 18, 21-2, 23, 27, 119-29 products 29 resources 13, 48, 52 sector 12, 41, 121 settlements 225 site 65, 183 society 55, 69 specialization 214 strategies 63 tenancy 42 tenure 40, 119 traditions 120 zone 137 pastoralism 24, 25-8, 68, 135, 143, 165 developed, 16 sedentary, 26 semi-sedentary, 16 pasture 24, 27-8, 45, 138 acquisition 56 allocation 56 land 27, 49 lots 42, 48, 56 rental 55, 56 rights 55, 202 patrilineal descent 45 patron-client relationships 36, 39, 42 patronage 39 peasantry 26 people of the hills 11 perishables 68, 69, 74, 76, 80, 97, 105, 176, 178 Persian Gulf 8 pits 219 political instability 25, 196 political office 105 population centres 224 composition 192 control 39 density 114, 210, 228 distribution 228 growth 135, 159 movements 121 pool 203 trends 153 portability 76, 107, 220 portable dwellings 105, 228
oba 49, 203 obsidian 185, 188 ocak 92 see also hearth occupation density 156 event 211, 228 levels 223 optimal grazing unit 144 living space 229 management 144 size 28 orientation see tent orientation Ottoman Empire 196-7, 202 oulad 49, 5\ over-grazing 40
hearths 220 items 69, 74, 84, 100 structures 66 superstructure 97, 106 post-depositional events 131 processes 3 posthole 191 potsherds 195, 220 pottery 66, 75 pre-depositional events 131 prehistoric sites 13 preservation 176 primary community 46, 48-9, 53-4, 57, 201 procurement strategy 22 productive animals 29 base 163 process 23 relations 21 strategies 20, 27, 59 unit 36 productivity 23-4, 28, 198 indices 29 property relations 44 public buildings 221 Qashqa'i 25-6, 39-41, 51-7, 60-1, 86-92, 100, 107, 135, 140-9 rank 82, 195, 205, 209 razzia, Assyrian 11 recessed hearth 92, 189 see also hearth reed screen 100, 210 reference group 49, 53 refuse disposal 124 organic, 218 primary, 3 secondary, 3 regional economy 154 relations of production 17-8, 37, 48 renewable products 17 renomadization 120, 121 rent collective 203 negotiations 118 pasture 55, 56, 110, 198, 200 reproductive life 29 potential 34 success 29 resettlement 160 residence associations 20, 24, 145, 227 group 20, 37, 44-8, 133, 147, 209, 228 hierarchy 27, 46, 132 residential groups 21, 229 patterns 44 unit 192, 203 resource-banking 25 restricted capital 34 Reza Shah 26, 60, 74, 96, 135 rib-strut tent 89 ridge-pole tents 90 ring model 226 rock-shelter 67 royal tents 88, 149, 155 ruined village 151 ruling class 18
Index
rustling 28 Sacjkara Yoriik 146 Safavid period 54, 196 Sahara 16 sale of livestock 166 Sanaydm 166, 167 Sanaydm Yayla 69, 76, 97, 111, 135,146, 162-5,172, 182-4 Sarmatians 12 Sassanian 68, 139, 150, 212 saturated milieu 59 saturation phase 156, 159 sautt system 42 Schmidt, E.F. 150, 157 Scythians 12 seasonal breeding 28-9 campsites 19, 110-1, 184, 215 cycle 6, 29, 56, 107, 108, 111 habitation 15 labour 39 migration 12, 18, 120, 227-8 movement 19 occupation 24, 67 pastures 19, 110, 136, 146 settlement 68, 88, 227, 228 secondary communities 48 security 44, 97, 144-5,159, 166 of tenure 110-2, 119, 146 sedentarization 5, 26, 39-42, 59, 62,74,108-11,120-1,147, 154-8,161, 228 enforced 60, 157 sedentarized nomads 114, 115, 133 sedentary agriculture 27, 135 architecture 224 community 23 pastoralism 26 sector 14, 25-6, 62, 63 settlement 155-7, 216-8 sites 76 village 195 villagers 13 sedentism 59, 122, 146, 161, 165-6 segmentary lineage 53, 54 system 53, 227 segmentation 26, 57, 121-2 selective grazing 28 pressures 27 self-sufficiency 36 see also subsistence Seljuk 18, 54, 57,139, 150, 154, 196 caravanserai 162 migrations 12 semi-aridity 28 semi-nomadism 10,15,16,19,59, 108 semi-sedentary 220 semi-urban 221 settled agriculture 59 population 153 village 36 settlement 15, 149 archaeology 155 cycle 158 density 117,145, 155-8 event 155 evidence 66, 67
252
hierarchy 48 history 152 pattern 108, 135-6, 141, 158, 223-4, 227-8 plan 6, 148, 158-61, 196, 209 size 3 structure 161 system 8, 20, 133, 214, 226 unit 133 settling nomads 158 Shah 153 Shahsevan 39, 42, 54-5, 60, 86, 95, 135 sheep corrals 198 sheep nomadism 63 sheikh 202 shelter 141, 151 shepherd households 41, 147 shepherding 39 labour 42, 204 shepherds 34, 37, 209 hired 40, 194, 204, 210 sherd scatter 124 sherds 76, 78, 189-91, 213-4, 219 shifting cultivators 20 Shirdashti 147 Sistan Region 13-4, 91, 220 site archaeological 150 density 153 fixtures 76 furniture 68, 79-80, 223 structure 4, 6, 172 siyah chadour 86 size threshold 46 skeletal remains 178-82 social structure 191 activity 145, 198 anthropology 19, 20, 227 distance 45, 155, 229 factors 4, 139 organization 100, 192-3, 198, 207-11 relations 9 status 148 stratification 41 structure 195 survey 187 socioeconomic factors 146-9 Southern Anatolia 117, 120 Soviet Georgia 220 Soviet Turkmenia 13 Soviet Union 11 spatial analysis 172, 191, 229 autocorrelation 130 behaviour 4 coherence 127, 130, 175 distributions 149 information 4 integrity 183 model 170 organizations, 100,123-8,147, 154,159, 209, 218, 226 patterning 4, 126 plots 128 relationships 155 structure 176 unit 183 specialization 9, 14, 26-7, 62 specialized areas 21, 129, 180, 182 domains 105 pastoral products 12 pastoralism 12
pastoralists 63 structures 107, 148 zones 129, 132 species diversity 28 spinning yarn 210 spring migration 198 state 48, 52 apparatus 54 ascendency 62, 135 authority 153 statics, archaeological 5 status 44, 82,101,105, 205, 210 Steppelands 6, 8, 12, 27 stock management 28 stone artifacts 79 footings 223 foundations 119 hearth 105 huts 111 platforms 213 tools 79, 185 village 187 storage jars 66, 76 platform 66, 100,105, 110, 189, 193, 216 structures 96 Strabo 11 stratigraphic layers 80, 191, 195, 213 record 80, 148 sampling 224 sequences 220 stratigraphy 68 structural organization 4, 131 subsistence 19, 80, 165 herd 117, 118 pastoralism 12 strategies 63, 215, 229 see also mode of subsistence Sultan Dag 76-7 Sumerian 89 summer camps 91, 96, 120, 134-5, 144, 165 campsites 74, 80, 139 grazing 135 migration 165 pasture 25, 55, 110-2, 134-6, 144, 198 quarters 107, 115, 198 resources 66 tents 100 superstructure 106 surface collection 122, 124 recording 172 survey 212, 218 design 115, 139,224 Suteans 10 Syria 14, 62, 75, 89, 94, 154, 157, 220, 224 tactical responses 119 taifeh 52 Taurus Mountains 7, 12, 84, 90, 111-4, 117-8, 134, 137, 162, 185, 198, 199 taxes 55 tectonic plates 7 telephone boxes 183 tell 151, 161 Tell Toqaan 157-8, 161 temporary campsites 67
Index
tent 44, 84, 205 camp 136-7, 139, 149, 150-1, 155-6 chief, 205, 207 complex 170 encampment 139 fabric 107 footings 80, 96 form 85, 91 foundations 151 orientation 141,170, 184, 205, 215-9, 224-8 sites 183, 188 size 101,104-5, 148 suburbs 154 typology 85-91 tent-dwelling tribesmen 155 tent-peg holes 166 tentdoth 84,106-8, 166, 184 tentpoles 99, 205 tentsite 68,107, 184, 213 platforms 119 tenure 34, 56, 61, 119 Tepe Guran 151 TepeTula'i 68,183,212-3 terraced tentsites 141 terrain 137-41, 170, 187, 219, 224-6 territorial aggrandisement 22, 58 conflict 53, 57 control 55 displacement 57 expansion 120 groups 49 instability 56, 58 stability 202 system 20-2, 52-4, 57, 62 unit 26, 203 territory 44, 54-5, 201, 229 access to, 20 test pits 183 tezek 138
Tibet 80 tied nomadism 59, 107, 120 tireh 51 toklulU, 118 tool kits 127 topak ev 86-8, 97 toss zone 130 toxic species 28 trade 14 tradesmen 74 transhumance 19,27,60,108,134, 215 migrations 20 pastoralism 13 pastoralists 165 transit camp 92-6, 105, 115, 134, 156, 163, 198
253
travellers 137, 150, 154, 186 trellis tent 86 tribal accretion 57 affiliation 65, 121, 133, 153, 157,196 chiefdoms 52 devolution 52 elites 42, 52 evolution 53 history 57 identity 55, 82, 198 institutions 63, 202 membership 48 migrations 163 militias 196 movements 163 names 120, 121 nomads 26, 199 organization 41, 55-7, 62, 199, 227 relationships 147 section 51, 55, 110, 121 segment 58, 117, 121, 197 segmentation 120 structures 52, 148, 199 territory 110 unit 26, 27, 51-5, 117,201 Tula'i 214, 215 Turkey 158, 224 Turkic emirates 114 incursions 12 invasions 196 migration 91 Turkmen 95, 151 162-3, 220 nomads 114 uq direkli qadir 87, 104, 166 uluki$i 202
under-cultivation 153 uniformitarian 227 unit of consumption 195, 200 management 42 production 34, 42, 200 unplanned settlements 100 unstable productive system 63 settlement system 119, 159 subsistence 153 untied migratory orbit 120 untied nomadism 59, 107 urban centres 15 elites 18 settlement 62 urbanization 14 use-life 76
usufruct rights 39, 42 Uzbek 86, 91 valuables 68, 69, 76 vegetation communities 28 viability, household 36 village 62, 67, 146, 153, 154 agriculture 63 house 106 pastoralism 16 pastoralists 214 plan 157 society 40 virilocal system 38 visibility, archaeological 69 Wakhan Corridor 63 walled structure 169 watch dogs 187 wealth 35, 40, 82, 101,105, 203, 209-10 Western Anatolia 200 White Revolution 60 wickerwork 96, 105 screen 91, 205 wild resources 82 winter campsites 137 encampment 151 grazing 120 pastures 134 quarters 107, 144 tents 107, 198 wood piles 210 wool 165 work zone 132 yayla 110, 115, 118-20, 134, 164, 198, 200-1 yaylag pastoralism 16 Yellen 2 Yoriiks 10, 32, 38, 42-3, 55, 61-2, 75, 87, 91, 104-7, 119-21, 134, 144-5, 150, 158,164,186, 200 yoghurt 75 Yomut43,86 Turkmen 38, 60 yukliik 92, 115
y«rr 86-8, 91, 141-2, 155 Zagros57, 134, 145,212,218 Mountains 12, 67, 216 Zawi-Chemi Shanidar 67 zemga97, 108, 111 zoma 199, 200-1, 203, 205, 209-10