The Social Organization of Early Copper Age Tribes on the Great Hungarian Plain 9781841717883, 9781407330334

Table of contents :
Front Cover
Title Page
Copyright
Dedication
Table of Contents
List of Tables
List of Figures
Preface
Acknowledgements
Chapter One: Introduction
Chapter Two: Tribal Social Theory
Chapter Three: Middle Range Theory
Chapter Four: The Archaeological and Geological Setting
Chapter Five: Research Design and Methodology
Chapter Six: The Sites and Assemblages
Chapter Seven: Integrative Units
Chapter Eight: Interaction Between Units
Chapter Nine: Conclusions
Bibliography

Citation preview

BAR  S1573  2006   PARKINSON  

The Social Organization of Early Copper Age Tribes on the Great Hungarian Plain

THE SOCIAL ORGANIZATION OF EARLY COPPER AGE TRIBES

William A. Parkinson

BAR International Series 1573 B A R

2006

The Social Organization of Early Copper Age Tribes on the Great Hungarian Plain William A. Parkinson

BAR International Series 1573 2006

Published in 2016 by BAR Publishing, Oxford BAR International Series 1573 The Social Organization of Early Copper Age Tribes on the Great Hungarian Plain © W A Parkinson and the Publisher 2006 The author's moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher.

ISBN 9781841717883 paperback ISBN 9781407330334 e-format DOI https://doi.org/10.30861/9781841717883 A catalogue record for this book is available from the British Library BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 1974 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by Archaeopress in conjunction with British Archaeological Reports (Oxford) Ltd / Hadrian Books Ltd, the Series principal publisher, in 2006. This present volume is published by BAR Publishing, 2016.

BAR

PUBLISHING BAR titles are available from: BAR Publishing 122 Banbury Rd, Oxford, OX2 7BP, UK E MAIL [email protected] P HONE +44 (0)1865 310431 F AX +44 (0)1865 316916 www.barpublishing.com

For my wife, Betsy – coach, teacher, saint, and, now, mother of Josephine Mei and Sarah Katherine.

Table of Contents

List of Tables ............................................................................................... vii List of Figures............................................................................................... ix Preface .......................................................................................................... xi Acknowledgements..................................................................................... xiii Chapter One: Introduction ............................................................................ 1 Chapter Two: Tribal Social Theory ............................................................... 9 Chapter Three: Middle Range Theory......................................................... 23 Chapter Four: The Archaeological and Geographic Setting........................ 39 Chapter Five: Research Design and Methodology...................................... 65 Chapter Six: The Sites and Assemblages .................................................... 81 Chapter Seven: Integrative Units............................................................... 123 Chapter Eight: Interaction Between Units................................................. 157 Chapter Nine: Conclusions........................................................................ 185 Bibliography .............................................................................................. 189

v

List of Tables

Table 4.1 Table 6.1 Table 6.2 Table 6.3 Table 6.4 Table 6.5 Table 6.6 Table 6.7 Table 6.8 Table 6.9 Table 6.10 Table 6.11 Table 6.12 Table 6.13 Table 6.14 Table 6.15 Table 6.16 Table 6.17 Table 6.18 Table 6.19 Table 6.20 Table 6.21 Table 6.22 Table 6.23 Table 6.24 Table 6.25 Table 7.1 Table 7.2 Table 7.3 Table 7.4 Table 7.5 Table 7.6 Table 7.7 Table 7.8 Table 7.9 Table 7.10 Table 7.11 Table 7.12 Table 8.1 Table 8.2 Table 8.3 Table 8.4 Table 8.5 Table 8.6 Table 8.7 Table 8.8 Table 8.9 Table 8.10 Table 8.11 Table 8.12

Late Neolithic and Copper Age radiocarbon dates ............................................................... 61 Magyarország Régészeti Topográfiaja sites ......................................................................... 82 Sites revisited during fieldwork ............................................................................................ 93 Sites revisited and collected during fieldwork...................................................................... 94 Dévaványa 115 – Site summary ........................................................................................... 97 Dévaványa 115 – Diagnostic ceramics ................................................................................. 97 Gyoma 28 – Site summary ................................................................................................... 99 Gyoma 28 – Diagnostic ceramics ......................................................................................... 99 Gyoma 163 – Site summary ............................................................................................... 101 Gyoma 163 – Diagnostic ceramics ..................................................................................... 101 Körösladány 14 – Site summary......................................................................................... 103 Körösladány 14 – Diagnostic ceramics .............................................................................. 103 Körösladány 15 – Site summary......................................................................................... 106 Körösladány 15 – Diagnostic ceramics .............................................................................. 106 Körösladány 16 – Site summary......................................................................................... 107 Körösladány 16 – Diagnostic ceramics .............................................................................. 107 Körösújfalu 4 – Site summary ............................................................................................ 109 Körösújfalu 4 – Diagnostic ceramics.................................................................................. 109 Okány 6 – Site summary......................................................................................................111 Okány 6 – Diagnostic ceramics ...........................................................................................111 Okány 16 – Site summary................................................................................................... 113 Okány 16 – Diagnostic ceramics ........................................................................................ 113 Vésztő 2 – Site summary .................................................................................................... 115 Vésztő 2 – Diagnostic ceramics.......................................................................................... 115 Vésztő 20 – Site summary .................................................................................................. 118 Vésztő 20 – Diagnostic ceramics........................................................................................ 118 Late Neolithic and Early Copper Age house areas ............................................................. 126 Revisited sites – Relative sizes ........................................................................................... 132 Periods represented at sites in the study area...................................................................... 136 Late Neolithic site clusters.................................................................................................. 140 Early Copper Age site clusters............................................................................................ 140 Middle Copper Age site clusters......................................................................................... 140 Comparison of Late Neolithic and Early Copper Age clusters........................................... 154 Comparison of Early and Middle Copper Age clusters ...................................................... 154 Pairwise correlation matrix of sizes.................................................................................... 154 Late Neolithic integrative units .......................................................................................... 155 Early Copper Age integrative units..................................................................................... 155 Middle Copper Age integrative units.................................................................................. 155 Possible interpretatios of stylistic attributes ....................................................................... 158 Diagnostic ceramic summaries by site ............................................................................... 158 Sites by cluster.................................................................................................................... 160 Visibility of ceramic attributes............................................................................................ 163 Assigned vessel types by site.............................................................................................. 163 Pedestal base indices........................................................................................................... 166 Lugs by site......................................................................................................................... 167 Lug shapes by site............................................................................................................... 169 Lug cross-sections .............................................................................................................. 169 Lug heights ......................................................................................................................... 169 Incised decoration by site ................................................................................................... 173 Dotted-linear incised decoration by site ............................................................................. 173

vii

Table 8.13 Table 9.1 Table 9.2 Table 9.3

Distribution of ceramic attributes ....................................................................................... 180 Comparative analysis of integrative units........................................................................... 186 Comparative analysis – Late Neolithic and Early Copper Age .......................................... 187 Comparative analysis – Early and Middle Copper Age...................................................... 188

viii

List of Figures

Figure 3.1 Figure 3.2 Figure 3.3 Figure 3.4 Figure 3.5 Figure 3.6 Figure 4.1 Figure 4.2 Figure 4.3 Figure 4.4 Figure 4.5 Figure 4.6 Figure 4.7 Figure 4.8 Figure 4.9 Figure 4.10 Figure 4.11 Figure 5.1 Figure 5.2 Figure 5.3 Figure 5.4 Figure 5.5 Figure 5.6 Figure 6.1 Figure 6.2 Figure 6.3 Figure 6.4 Figure 6.5 Figure 6.6 Figure 6.7 Figure 6.8 Figure 6.9 Figure 6.10 Figure 6.11 Figure 6.12 Figure 6.13 Figure 6.14 Figure 6.15 Figure 7.1 Figure 7.2 Figure 7.3 Figure 7.4

Pawnee cognized social structure................................................................ 34 Pawnee actualized social structure during the 18th Century ........................ 34 Pawnee actualized social structure during the early 19th Century ............... 35 Pawnee actualized social structure, 1859-1876........................................... 35 Trajectories of change in Pawnee social organization................................. 36 Idealized models for interpreting patterning ............................................... 36 Distribution of Late Neolithic Groups......................................................... 45 Distribution of Early Copper Age Tiszapolgár Culture............................... 45 Distribution of Middle Copper Age Bodrogkeresztúr Culture .................... 56 Relative chronology..................................................................................... 57 Radiocarbon dates – Tisza Group................................................................ 58 Radiocarbon dates – Herpály Group ........................................................... 58 Radiocarbon dates – Csőszhalom Group..................................................... 58 Radiocarbon dates – Lengyel Group ........................................................... 59 Radiocarbon dates – Early Copper Age....................................................... 59 Radiocarbon dates – Middle Copper Age.................................................... 59 Radiocarbon dates – Late Neolithic and Early Copper Age........................ 60 Political boundaries in the study area.......................................................... 72 Hydrology of the study area ........................................................................ 72 Surface sediments in the study area............................................................. 74 Location of MRT volumes in the study area ............................................... 74 Ceramic analysis form showing sample data .............................................. 76 Ceramic coding information........................................................................ 78 MRT sites - Southwest................................................................................. 91 MRT sites – North central ........................................................................... 91 MRT sites - Central...................................................................................... 92 MRT sites - East .......................................................................................... 92 Sites included in the analysis....................................................................... 93 Dévaványa 115 – Site collection map ......................................................... 96 Gyoma 28 – Site collection map ................................................................. 98 Gyoma 163 – Site collection map ............................................................. 100 Körösladány 14 – Site collection map....................................................... 102 Körösladány 15 and 16 – Site collection map ........................................... 104 Körösújfalu 4 – Site collection map .......................................................... 108 Okány 6 – Site collection map................................................................... 110 Okány 16 – Site collection map................................................................. 112 Vésztő 2 – Site collection map .................................................................. 114 Vésztő 20 – Site collection map ................................................................ 116 House areas by period ............................................................................... 124 Size of Late Neolithic and Early Copper Age house areas........................ 125 Number of sites per period ........................................................................ 131 Percentage of periods represented ............................................................. 131 ix

Figure 7.5 Figure 7.6 Figure 7.7 Figure 7.8 Figure 7.9 Figure 7.10 Figure 7.11 Figure 7.12 Figure 7.13 Figure 7.14 Figure 7.15 Figure 7.16 Figure 7.17 Figure 7.18 Figure 7.19 Figure 7.20 Figure 7.21 Figure 7.22 Figure 7.23 Figure 7.24 Figure 8.1 Figure 8.2 Figure 8.3 Figure 8.4 Figure 8.5 Figure 8.6 Figure 8.7 Figure 8.8 Figure 8.9 Figure 8.10 Figure 8.11 Figure 8.12 Figure 8.13 Figure 8.14 Figure 8.15 Figure 8.16 Figure 8.17 Figure 8.18 Figure 8.19 Figure 8.20 Figure 8.21 Figure 8.22 Figure 8.23 Figure 8.24

Comparison of site sizes............................................................................ 133 Linear regression of MRT size and core area ............................................ 134 Linear regression of MRT size and diffuse scatter size............................. 134 Site sizes by period.................................................................................... 135 Site sizes by period – Large outliers excluded ......................................... 135 Rank/size graphs of Late Neolithic and Early Copper Age sites............... 137 Early Copper Age single-period sites ........................................................ 137 Rank/size graph of Middle Copper Age sites ............................................ 138 Late Neolithic density contour map .......................................................... 139 Late Neolithic density contour map - East ................................................ 142 Late Neolithic clusters............................................................................... 142 Late Neolithic clusters by size................................................................... 143 Early Copper Age density contour map..................................................... 143 Early Copper Age density contour map - West ......................................... 146 Early Copper Age density contour map - East .......................................... 148 Early Copper Age clusters ......................................................................... 149 Early Copper Age clusters by size............................................................. 150 Middle Copper Age density contour map.................................................. 150 Middle Copper Age clusters ...................................................................... 152 Middle Copper Age clusters by period...................................................... 153 Diagnostic types by site............................................................................. 153 Sites included in the analysis..................................................................... 161 Vessel types by site.................................................................................... 162 Open vessel types by site........................................................................... 164 Index of pedestal base tops to flat bases by site ........................................ 165 Index of other pedestal base fragments to base tops ................................. 166 Lug decoration by site ............................................................................... 166 Lug heights by site..................................................................................... 168 Lug height by pierce type .......................................................................... 170 Lug location by site ................................................................................... 171 Incised decoration by site .......................................................................... 172 Relative percentage of incised decoration................................................. 173 Kisrétpart group sites in the study area ..................................................... 174 Dotted-linear decoration types by site....................................................... 175 Other incised decoration............................................................................ 176 Percentage of pierced pdestal base fragments by site ............................... 176 Pierce sizes of pedestal bases by site......................................................... 177 Diameter of everted rims by site ............................................................... 178 Diameter of flat bases ................................................................................ 178 Length of everted rims............................................................................... 179 Thickness of everted rims.......................................................................... 179 Lug shapes by site ..................................................................................... 181 Lug cross-sections by site.......................................................................... 181 Lug piercing by site ................................................................................... 182

x

Preface

This book is based on a doctoral dissertation entitled The Social Organization of Early Copper Age Tribes on the Great Hungarian Plain (Parkinson 1999). It describes the initial phases of the Körös Regional Archaeological Project in southeastern Hungary. Portions of the research described in this book have been published elsewhere (e.g., Parkinson 2002a, 2002b), and other aspects of the research have been developed further and published in other venues (e.g., Parkinson 2006; Parkinson and Gyucha In press). In the years since the dissertation was written, the Körös Regional Archaeological Project has actively been exploring several of the questions that arose out of it, especially those relating to the organization of specific settlements within microregions. The results of that recent research, which have been published in various places (e.g., Gyucha et al. In press; Gyucha, Parkinson, and Yerkes 2004; Parkinson, Gyucha, and Yerkes 2002; Parkinson, Yerkes, and Gyucha 2004; Parkinson et al. 2004; Sarris et al. 2004; Yerkes et al. In press), have significantly altered our understanding of Early Copper Age social organization. Nevertheless, the research described in this book never was published in its entirety and – even though we know a great deal more about the period now than we did then – the basic information about the ceramic assemblages and settlement patterns will be of general interest to those interested in European prehistory. I therefore have decided to publish the dissertation in its original form to make the primary data widely accessible.

xi

Acknowledgements

My interest in tribal societies was first piqued during a seminar taught by John OʼShea and Robert Whallon a few years ago. Several of the ideas central to this thesis would not have been formulated had I not the good fortune to kick around ideas with the other participants. In particular, I need to recognize a fellow student and very good friend – Severin Fowles – who wrote a paper for the seminar that has had a tremendous impact upon my thoughts about tribal societies (Fowles 1997).

whom this research would never have reached fruition. In particular, I would like to thank Csanád Bálint, the director of the Institute of Archaeology of the National Academy of Sciences. I would also like to thank Igor Grin, Imre Szátmari, Attila Gyucha, Gábriella Gábor, and the two Évas from the Békés County Museum, where I was so well received during the summer and fall of 1998. I look forward to several more years of working with them.

The seeds for a dissertation project on the Hungarian Copper Age were sown during a 007 gathering with John OʼShea and Ferenc Horváth at Dominicks in Ann Arbor in 1996. In the years that have passed, several people have contributed in various ways to see the project develop from an idea sketched out on a napkin to a big thick book. They all deserve my utmost thanks.

I am also happy to recognize financial support from the National Science Foundation (Dissertation Improvement Grant #9812677), Rackham School of Graduate Studies, the Department of Anthropology, and the Museum of Anthropology at the University of Michigan. Several friends helped with fieldwork and data collection. They include Jonathan “Huck” Hawes, Michael “Malakas” Galaty, Julian “Jay” Murchison, Severin “Spice” Fowles, my wife Betsy Parkinson, and my father-in-law, Dick Reichert. Without their donated time and effort I never would have finished. Farley Zuber washed the pottery and Kati Sebók drew several of the sherds.

I especially need to thank John OʼShea – my dissertation Chair and mentor – for keeping me honest over the years and never letting me slow down. It was due largely to his persistence, patience, and level-headed guidance that I was able to pull through and finish in a timely fashion. He and the rest of my committee – John Cherry, Carla Sinopoli, and Robert Whallon – never failed to read what I gave them, no matter how rough the draft. Their thoughtful comments and criticisms have made the final version of this manuscript a much better product.

Although the current project took only a few years to complete, I have benefited in innumerable ways from my constant interaction with the graduate students, faculty and staff in the University of Michigan Museum of Anthropology. I hope they all know that they have contributed to this dissertation in some way.

Ferenc Horváth was kind enough to guide me through several different dissertation topics over fish soup and beer in Szeged, where I first got my feet wet with prehistoric Hungarian ceramics. He eventually sent me off to Budapest to see his colleague Pál Raczky, who has been a constant help as I have struggled my way through the finer points of Hungarian Prehistory. It was on Dr. Raczkyʼs advice that I contacted János Makkay regarding the material at Vésztő and Örménykút. Dr. Makkay took the time to personally introduce me to people at the county museums and at the Institute of Archaeology. I cannot thank them enough for their assistance and tutoring over the last few years.

My family and in-laws have always supported my choices to do what I do, even when most of them still think I dig up dinosaur bones. Their unquestioning financial and emotional support over the years has been priceless to me, although I do not always let them know it. I do not know how to begin to thank my wife, Betsy, for her unconditional love, acceptance, and support. It was through her comfort that I continually found the strength to write, and in her eyes that I eventually saw the end of the tunnel. She inspires me each day by her drive to achieve, and sets an example I can only hope to live up to.

I was also fortunate to benefit from the good graces of several people in different institutions throughout Hungary, without xiii

Chapter One: Introduction

Do tribes exist? Or are they chimeras, imaginary compounds of various and, at times, incongruous parts, societal illusions fabricated for diverse reasons, but once created, endowed with such solid reality as to have profound effect on the lives of millions of people? The question is practical, because it does have consequences in daily life, and theoretical, because the notion of tribe has played a vital role in various social sciences, perhaps most conspicuously in anthropology.

as the term continues to be used in the discipline, tribes do exist. This being the case, the question then becomes one of agreeing on a precise definition for the term. This is one anthropological problem that may be better approached using archaeological – rather than ethnographic – data. The longterm perspective available to archaeologists allows us to track subtle changes in social organization that ethnographers are seldom at liberty to witness given the inherently short-term nature of the information at their disposal. By accessing the long-term perspective available to us from the information contained in the archaeological record, I believe we can eventually arrive at an answer to Friedʼs question.

This is how Morton Fried began his seminal work entitled, The Notion of Tribe (1975). In the twenty-four years since Fried posed this simple question – Do tribes exist? – anthropologists still cannot agree on its answer. Friedʼs own conclusion was that tribes are an aberrant form of social organization that occur only in very specific secondary social contexts (see also Fried 1968). Most cultural anthropologists have abandoned the concept entirely. As Elisabeth Colson (1986:5) began one article:

This dissertation explores the topic of tribal social organization using archaeological data from prehistoric Hungary. The argument presented in the following pages builds upon the concept of ʻlevels of social integration,ʼ proposed initially by Steward (1955), and developed further by Sahlins (1963, 1968) and Service (1971). In contrast to these now wellknown studies, which were concerned primarily with developing a method of analysis useful for explaining cultural evolution, the concern here lies with understanding the degree of variability exhibited by a particular form of social integration and its manifestations in different cross-cultural frameworks. By approaching tribes as societies that exhibit a particular form of social integration, it is possible to retain the concept – as a heuristic device for cross-cultural comparison – without having to accept all of the economic, ideological, and political trappings that earlier authors chose to associate with it – as an evolutionary stage. This results in a theoretical perspective that focuses upon the structural aspects of different societies as the critical variables that simultaneously lump them together, and split them apart into groups of varying degrees of homogeneity. Building upon the notion of the tribe as a dynamic, flexible form of socio-structural integration, I construct a methodology useful for modeling long-term changes within tribal social systems. This methodology is then applied to the Late Neolithic and Early Copper Age of the Great Hungarian Plain using spatial and ceramic data from several sites located in the Körös River Valley.

I do not know what is meant by ʻTribal Societies.ʼ ʻTribeʼ and ʻtribalʼ are slippery terms despite various attempts to pin them down so that they could be used analytically… ʻtribeʼ has been used with reference to the whole span of human groups, with perhaps the exception of the nuclear family. The Tribe On The Hill which Jack Weatherford published in 1981 is about the United States Congress with its associated staff and penumbra of lobbyists. Of course, the term also has a very specific legal definition in the halls of the United States government (see Sterritt et al. 1998; Beinart 1999). But the concept of tribe, and “The Problem of Tribe” (see Helm 1968) as an anthropological construct, still comes up quite frequently in archaeological literature, and a stray ethnologist can even be heard uttering the term every now and then. Despite various attempts to replace it with hyphenated terms that are more descriptive, such as “small-scale, semisedentary, egalitarian societies”, both the term and the concept persist.

I argue that the radical social changes that mark the transition

The question is therefore not whether tribes exist – as long 1

Chapter One

from the Neolithic to the Copper Age on the Great Hungarian Plain can best be understood as resulting from minor changes that occurred within a pre-existing social structure that was inherently flexible. This inherent structural flexibility manifested itself in a set of ʻstructural posesʼ (see Gearing 1958; see also Fowles 1997; Fowles and Parkinson 1999) which, when differentially expressed, produced over the long term what appear to be radically different social patterns. Fortunately, the long-term operation of these different social patterns left behind distinctive material correlates in the archaeological record. These material correlates – or ʻstructural signaturesʼ – can be used to model the changes in social organization that occurred about 4,500 BC.

operational definitions include culture, band, society, etc. The use of such discipline-specific terminology is a necessary evil within the social sciences, wherein the unit of analysis is seldom clearly defined. Regarding this problem, Marvin Harris (1979:15) noted that: A strong dose of operationalism is desperately needed to unburden the social and behavioral sciences of their overload of ill-defined concepts, such as status, role…group, institution, class, caste, tribe, state…and many others that are part of every social scientistsʼ working vocabulary. The continuing failure to agree on the meaning of these concepts is a reflection of their unoperational status and constitutes a great barrier to the development of scientific theories of social and cultural life (my emphasis).

Throughout the Late Neolithic, communities on the Plain were organized into complexly-integrated units that interacted intensively within well-defined, spatially discrete areas. This pattern gave way in the Early Copper Age to one dominated by more diffuse, less-complexly integrated units that interacted less intensively over larger areas. Despite the traditional tendency to envisage radical societal changes at this time (e.g. Bognár-Kutzián 1972), I propose that these changes occurred within the variability that was allowed by a single, albeit very flexible, social structure. This quality – of societies occasionally to rework their methods of social organization within certain structural bounds – is characteristic of tribal societies, and has been documented not only in other archaeological contexts (e.g., Braun and Plog 1982; Emerson 1999), but in ethnohistoric and ethnographic contexts as well (see also OʼShea 1989).

The ʻstrong doseʼ of operationalism suggested by Harris was never taken, and anthropologists currently find themselves inundated with a plethora of ill-defined terms which each seem to acquire their own definition depending upon the specific context within which they are employed. Nowhere is this problem more apparent than dealing with the term tribe. I have chosen to retain the use of the term ʻtribeʼ in this dissertation not because I wish to rekindle the polemic debate surrounding the supposedly inexorable process of sociocultural evolution (e.g., Band-TRIBE-Chiefdom-State [for example, Service 1971]), but rather because the term has a long history in cross-cultural anthropology, and because it denotes a form of social organization generally understood to refer to a wide range of social systems that regularly exhibit some degree of institutionalized social integration beyond that of the extended family unit, or band. Nevertheless, some may find my use of the term anachronistic, since it has come to be replaced by even more ambiguous phrases, such as ʻmiddle range societyʼ (e.g., Feinman and Neitzel 1984). This latter moniker attempts to place tribes somewhere Between Bands and States (Gregg 1991), as one book title puts it, and emphasizes the transitional and more ephemeral nature of tribal social systems.

This tendency of tribal social structures to produce radically different social patterns that fluctuate between 1) complexlyintegrated and intensively interacting systems, on the one hand, and; 2) loosely integrated and loosely interacting systems, on the other hand; is somewhat reminiscent of the ʻcyclingʼ often associated with chiefdoms (Anderson 1990), but in the case of tribal societies occurs in the absence of any institutionalized central authority and on a much smaller spatial scale. This phenomenon can be documented in other cultural contexts, and may itself be a criterion useful for distinguishing tribes from other decentralized, segmentary, forms of social integration, such as bands. Since these changes tend to occur gradually, over several generations, they often go undocumented in ethnographic descriptions of tribal societies, which are of necessity usually more limited in their temporal depth.

It is precisely this tendency – to view tribes as ephemeral ad hoc social constructions – that has resulted in the creation of a number of appellations, such as ʻtribeletʼ (e.g., Bocek 1991), ʻritualityʼ (e.g., Yoffee et al. 1999), and ʻtransegalitarian societiesʼ (e.g., Owens and Hayden 1997), which apply to only a few historically particular contexts and have limited, if any, utility in comparative cross-cultural analyses. Although cases occasionally arise when it is necessary to create new terms within the discipline, such neologisms have begun to run rampant within the field, and it is now necessary to begin reassessing their utility. To this end, the research presented here represents an attempt at stressing not the historically particular characteristics of tribal social systems, but their lasting – albeit somewhat elusive – structural nature.

ʻTribeʼ vs. ʻMiddle Range Societyʼ In a dissertation that claims to be about ʻtribesʼ it is necessary at some point to delineate precisely what this term means, and why it has been chosen over other possible alternatives that are much more in fashion these days. The word tribe is one of several arbitrary, operational definitions used by anthropologists to facilitate cross-cultural comparison (Bernard 1994; Kuznar 1997). Other examples of

By defining a tribe as a particular structural form – rather

2

Introduction

than as an evolutionary stage – it is possible to circumvent several of the problems that plagued earlier attempts at such classifications (e.g., Service 1971; Sahlins 1968; Fried 1968). In particular, a structural definition eliminates the necessity to associate tribes a priori with particular economic (e.g., horticultural, sedentary) or political (e.g., ʻegalitarianʼ) attributes. In fact, the majority of tribal societies, as defined above, could probably be characterized as small-scale, ʻegalitarian,ʼ sedentary, agricultural societies. Nevertheless, to define tribes in terms of these secondary attributes is to restrict the possibility of understanding their economic or political diversity. Rather, it is more prudent to allow the broadest structural definition of what is meant by the term tribe (see Chapter Two), and to allow empirical observation to determine the range of associated economic and political variability.

societies they study, they inadvertently neglect the range of structural variability exhibited by different societies over the longue durée. This leads to the unfortunate pigeon-holing of societies into different social classifications that simply do not bear themselves out over the long-term. While this argument also applies to more complex social classifications, such as chiefdoms and states, it is particularly with regards to bands and tribes that its effects are even more confusing. When viewed with the long-term perspective available through the archaeological record, societies that are traditionally called tribes tend to assume a variety of different configurations throughout their ontogeny. At times, they appear as discretely-defined, intensively-integrated entities with actively-maintained social boundaries. At other times, tribes appear as poorly-defined, loosely-integrated entities with few clear social boundaries. Ethnographically, the tendency has been to classify the former as ʻtribesʼ and the latter as ʻbands.ʼ When viewed over the long term (i.e., archaeologically), however, most tribal societies would most likely be found to oscillate between these two different patterns (see Fowles 1997, Fowles and Parkinson 1999).

Another benefit of defining tribes as exhibiting a particular structural form, rather than a bundled set of associated characteristics, is the successful ability of the definition to distinguish tribes from both more and less complex social forms. With reference to distinguishing tribes from bands, this is accomplished in a largely quantitative manner, for bands interact with, and are integrated into, larger units (for example, Serviceʼs [1971:72-98] notion of composite bands). Fried has made the argument against separating bands from tribes in an evolutionary context most elegantly:

To deny non-centralized social systems the flexibility to assume both a discretely-defined, complexly-integrated state, on the one hand, and a less clearly-defined, loosely-integrated state, on the other hand, is to ignore the archaeological fact that some societies regularly shift from one state to the other – these are tribes. In these cases, it can be assumed that the social structure contains within it a certain degree of organizational flexibility that allows the social segments to regularly interact and integrate at various levels beyond that of the band. Other egalitarian societies do not exhibit the inherent structures necessary to assume a clearly-defined, complexly-integrated state – these are bands.

Unfortunately… the arrangement of band and tribe in a general evolutionary sequence of forms leaves much to be desired. It seems possible, for example, to place both on the same temporal level and level of complexity, with the band representing the semi-permanent membership group sharing a common camp, and the tribe constituting the shifting congeries of camps (i.e., bands) that comprises an area of relative peace and which would include a demonstrably high frequency of intermarriage (Fried 1968:12).

In order to account for this empirical pattern, which emerges repeatedly in the archaeological record of tribal societies across the globe, it is necessary to assume a theoretical perspective of tribal social organization that allows for the dynamic nature of the societies in question. The definition proposed above allows for such flexibility by focusing upon the structural characteristics that characterize tribal societies.

Since Fried understood the major evolutionary development in human history to be associated with the transition from ʻegalitarianismʼ to ʻranked society,ʼ and since most tribes are essentially ʻegalitarian,ʼ he found it necessary to explain the reoccurrence of tribes in different geographic and historical contexts, “as a consequence of the impinging on simple cultures of much more complexly organized societies” (1975: 10). For Fried, pristine tribes did not exist, but occurred only in secondary contexts.

While it can be argued that the proposed definition distinguishes bands from tribes in a largely quantitative manner, the decentralized aspect of the definition sufficiently separates tribes from their traditionally more complex counterparts – chiefdoms and states. These latter ʻcomplex societiesʼ differ from bands and tribes primarily in their tendency to exhibit some form of institutionalized economic or political (or usually both) centralization. The segmentary, socially redundant, acephalous, structure of tribal societies makes it especially difficult for particular social segments to acquire and retain a centralized economic or political role. That is not to say that tribes are always decentralized, for this is certainly not the case. Rather, centralization in tribal societies tends not to

The perpetual argument over the distinction between bands and tribes, and whether either exists in anthropological reality, stems from an overreliance upon ethnographic and ethnohistoric information in the formulation of such social constructs. This tendency – to neglect the long-term perspective available through the archaeological record – results in a lopsided understanding of the range of variability associated with ethnographically documented tribal societies. Because ethnographers and ethnohistorians are necessarily limited in their ability to view long term processes at work in the

3

Chapter One

be politically institutionalized, and tends to occur in a manner reminiscent of Johnsonʼs (1978) sequential hierarchy – by consensus, rather than imposed from above.

While these two processes are intimately intertwined, I believe it is important to distinguish between them methodologically, since different types of archaeological data tend to speak more to one process than the other. By separating these traditionally ambiguous and overlapping processes into discrete analytical dimensions, it becomes possible to measure specific archaeological variables which relate predominantly to one dimension or the other. In keeping with the way these terms have been used traditionally (e.g., Steward 1955), I suggest that the term integration should be restricted to a group-level phenomenon that incorporates individuals into specific organizational units. In this sense, societies can be envisioned as integrating at various levels – the nuclear family level, the household level, the village level, etc. Interaction, on the other hand, refers to a more diffuse process that operates between individuals and groups. Of course, smaller units of integration presuppose increased interaction – members of families interact more with each other than with anyone else. The methodological challenges in dealing with tribal societies in archaeological contexts are, initially, to determine the size, scale and organization of the different integrative units, and second, to measure the extent to which the various units interact across the social landscape.

By assuming such a structural perspective, it becomes possible to identify particular ʻstructural signaturesʼ in the archaeological record. These signatures, which are the material correlates that are produced by the operation of particular social patterns over the longue durée, can then be used to model both the size, scale, and organization of the different social segments involved, and how these various segments interacted to varying degrees over a given period of time. By placing the analytical emphasis upon the structural processes that produced different structural signatures, this framework lends itself both to understanding diachronic change within particular prehistoric contexts, and to the cross-cultural analysis of different social forms. By modeling the different social patterns that generated different signatures, it becomes possible to incorporate dynamic concepts, such as Gearingʼs (1958) notion of structural pose, into our analyses of prehistoric social systems. This theoretical framework is developed further in the following chapter. Methodology – Measuring Interaction and Integration

Within this methodological framework, the units of integration are understood to be indicative of the size and scale of different social segments. For example, looking at the size and organization of houses in which people lived, it is possible to model how the basic segments of prehistoric societies were organized, or integrated – at the nuclear family level, or at the multi-family or lineage level. Similarly, by looking at the size and internal organization of villages or hamlets it is possible to model how these basic segments were integrated into corporate units of greater scale. Beyond the village or hamlet level, clusters of settlements frequently indicate integrative units of an even larger scale.

The theoretical perspective assumed throughout this dissertation has specific implications with regard to the archaeological methodology that needs to be employed for understanding the structural variability inherent in different tribal social systems. In particular, it is necessary to focus upon two analytical dimensions that are archaeologically accessible: 1. The units of integration; and, 2. The degree of interaction. By delineating the size, scale, and organization of integrated social segments within a society and the way in which those units interact to varying degrees, it is possible to generate a model that can be used to track changes in social patterns over the long-term. Such a model also facilitates the crosscultural analysis of trajectories of change in different cultural contexts.

The degree of interaction between these units can be assumed to co-vary with the degree to which they are integrated into segments of different scale. This being the case, by assessing the degree to which the members of different integrative units interacted with each other, it becomes possible to infer how these different social segments constituted themselves into structural units of even greater scale.

The acephalous, decentralized, segmentary nature of tribal societies has forced archaeologists to model tribal social organization by looking at changing patterns of integration and interaction over time (e.g., Braun and Plog 1982; Saitta 1983; Plog and Braun 1984; Creamer and Haas 1985; Haas 1990; Habicht-Mauche 1990). Unfortunately, we are seldom provided with precise definitions for these terms, which are often used interchangeably throughout the literature. There is some general consensus that integration refers to a group level phenomenon. Interaction, on the other hand, usually refers to a more general process that occurs at both the group and individual level.

Studies of prehistoric interaction range from general reconstructions of localized and long-distance trade networks, to intensive stylistic analyses of ceramic designs and lithic artifacts from particular sets of sites. These latter studies have a long history in archaeology (Leone 1968; Whallon 1968), and have undergone various theoretical revisions as the focus of the discipline itself has developed and changed (see Hegmon 1995; Stark 1998). Studies of prehistoric integration have a similarly long history in the discipline. While several of the same datasets – such as ceramic style – that are used to measure the degree of

4

Introduction

interaction (between individuals) also measure the degree of integration (between groups of individuals), there are specific sets of information that relate predominantly to the latter. For example, Hegmon (1989) discusses the use of architecture as evidence of integration in the American southwest. Other examples of archaeological datasets that attempt to measure social integration explicitly include rank/size analyses (Zipf 1949) and the spatial analysis of settlement patterns (see, for example, Flannery 1972). Recent developments in computer technology, and in particular the development of Geographic Information Systems (GIS) have added a strong component to the analysis of spatial variables, allowing more subtle patterns to be teased out of previously unmanageable datasets.

that the farmers and herders who inhabited the villages across the Plain were, to borrow Friedʼs phrase, ʻimpinged upon by more complex social forms,ʼ such as chiefdoms or states. Previous suggestions that these changes can be attributed to large-scale population replacement have recently fallen out of favor (see Bánffy 1994). Clearly, the social transformation that affected the villagers of the Plain in the mid fifth millennium must be understood in terms of the wide range of variability that occurs within tribal forms of social organization. The transition from the Late Neolithic to the Early Copper Age on the Great Hungarian Plain coincides with the inception of several technological developments that changed the trajectory of social evolution in Eastern Europe for several millennia. The copper ore sources in the Carpathian and Balkan mountains began to be intensively mined and smelted during this time (Jovanović 1982). Domestic animals, and especially cattle, began to be exploited not only for their meat, but also for ʻsecondary productsʼ such as milk and cheese (Chapman 1983; Sherratt 1981, 1983a). The plow was also introduced to the region (Milisauskas and Kruk 1991), permitting heavy alluvial soils to be brought under cultivation for the first time.

These different methodological tools for measuring prehistoric interaction and integration can be used to measure specific cultural variables within different archaeological contexts. By measuring different variables that relate to these two dimensions of analysis (i.e., integration and interaction), it is possible to reconstruct particular social patterns from their material correlates left behind in the archaeological record. The extent to which these structural signatures reflect actual social patterns will be determined largely by the chronological quality of the archaeological data itself. For example, those working in an area where local chronologies are more refined will be able to reconstruct structural signatures in much finer detail than those of us who must deal with archaeological periods that last for centuries. These different structural signatures, and the trajectories they assume in different societies are then broadly comparable in a cross-cultural framework, as are the inferred social patterns that produced them. The details of this methodology are discussed more explicitly in Chapter Three.

Roughly contemporary with these technological developments was a radical transformation in the way life itself was organized on the Plain. The dramatic changes documented in the archaeological record include: Changes in the spatial scale of ʻcultural groupsʼ. The three geographically-discrete ʻcultural groupsʼ that sub-divided the Great Hungarian Plain during the Late Neolithic were replaced by a single, relatively homogeneous ʻcultureʼ that extended across the entire Plain during the Early Copper Age – viz. the Tiszapolgár Culture (see Figures 4.1 and 4.2). The Late Neolithic ʻcultural groupsʼ, as they are known in Hungarian literature, each produced distinct ceramic assemblages, and exhibit differences in settlement type, settlement location and economic strategies (see Kalicz and Raczky 1987a). No such sub-divisions are readily identifiable during the Early Copper Age (see Sherratt 1982; contra BognárKutzián 1972).

This dissertation represents an attempt at operationalizing this theoretical and methodological framework by focusing upon the radical social changes that occurred on the Great Hungarian Plain ca. 4,500 BC. Tribal Social Change – The Great Hungarian Plain, ca. 4,500 BC The transition from the Late Neolithic (ca. 5,000-4,500 BC) to the Early Copper Age (ca. 4,500-4,000 BC) on the Great Hungarian Plain is marked by dramatic changes in the archaeological record – changes in the spatial scale of cultural groups, house form, settlement type, settlement location, trade networks, and mortuary practices (Bognár-Kutzián 1972; Raczky 1987a). These changes in material culture suggest that the population of the Great Hungarian Plain underwent a significant social transformation about 4,500 BC. This transformation affected not only inter-group relationships, as indicated by changes in trade-networks and settlement organization, but also intra-group relationships, as indicated by changes in house form and mortuary customs.

Changes in house form. The large, possibly multi-family, domestic structures of the Late Neolithic (10 - 20m long) were replaced by much smaller (ca. 5m long), less substantial dwellings in the Early Copper Age (BognárKutzián 1972; Goldman 1977; Kalicz and Raczky 1987a; Siklódi 1982, 1983). Changes in settlement type. The Late Neolithic settlement pattern, which combined the habitation of large tells (up to 6 ha) with large ʻflatʼ (i.e. horizontal) settlements (up to 11 ha), gave way in the Early Copper Age to the almost exclusive habitation of smaller, flat settlements (ca. 0.5-1 ha) (Bognár-Kutzián 1972; Chapman 1997a; Kalicz and Raczky 1987a; Sherratt 1984).

Throughout both of the periods in question there is no evidence of institutionalized social ranking. Nor is there any evidence 5

Chapter One

Changes in settlement location. In addition to being smaller than Late Neolithic settlements, Early Copper Age sites were less nucleated and more evenly dispersed across the landscape (Bognár-Kutzián 1972; Sherratt 1983b, 1984). For example, Sherratt (1997b:308) notes that in the Szeghalom region (ca. 68,000 ha), 106 Early Copper Age sites were recorded in an area where only 19 Late Neolithic sites were identified.

posed almost entirely of mortuary data. Thanks in large part to the extensive research that has been directed towards the Late Neolithic, it is relatively clear how the Great Hungarian Plain was organized during this period. The three geographically-discrete Late Neolithic ʻcultural groupsʼ were each clearly-defined, with markedly different ceramic assemblages, settlement types, and house types, within geographically-discrete exchange spheres (Biró 1998; Kalicz and Raczky 1987a). Each of these larger social groups was, in turn, organized into regional systems of integration in which several smaller sites were organized around a single large site (Makkay 1982; Raczky 1987; Sherratt 1984). Often this large ʻsupersiteʼ was a tell, and frequently the smaller sites were flat (i.e., horizontal) settlements. This complex form of regional integration – of smaller sites organized around a single supersite – disintegrated at the end of the Late Neolithic when the cultural groupings themselves disappeared, yielding to a single, relatively homogeneous cultural group – the Tiszapolgár Culture.

Changes in trade networks. The long-distance trade networks of the Neolithic, which brought goods from as far away as the Aegean, were re-structured and re-directed in the Early Copper Age to bring copper, gold, and chert from the Carpathians onto the Plain (Biró 1998; Sherratt 1987). Changes in mortuary practices. Large formal cemeteries were established in the Early Copper Age. These cemeteries usually occurred isolated in the landscape – entirely unassociated with settlement sites – and replaced the Neolithic pattern of burying the dead in and around settlement sites (Bognár-Kutzián 1963, 1972; Chapman 1997b).

In contrast to the Late Neolithic, the social organization of the Plain during the Early Copper Age is much less clear. Several years ago, Bognár-Kutzián (1963, 1972) suggested that the Early Copper Age Tiszapolgár culture could be divided into four spatially discrete ʻsub-groupsʼ. She distinguished each of these ʻsub-groupsʼ by minor variations in Early Copper Age ceramic assemblages. Since these sub-groups were based upon ceramic materials from cemeteries and stray finds, rather than from controlled excavations at stratified settlement sites, it is unclear whether they are indicative of regional variations in burial practices, or of small-scale temporal shifts in style. Subsequent research has demonstrated that Bognár-Kutziánʼs Early Copper Age sub-groups are not temporally and geographically continuous (Siklódi 1984), and they certainly do not represent the same kind of intensive regional integration that characterized the Plain during the Late Neolithic.

Changes in agricultural practices. Domestic cattle tend to dominate Early Copper Age assemblages, replacing the domestic/wild mix that dominate Neolithic assemblages (Bognár-Kutzián 1972; Bökönyi 1959, 1962; Sherratt 1983b; Skomal 1983). These changes document a radical abrupt transformation that affected almost every aspect of social organization — from the internal organization of settlements and houses, to the way settlements were integrated across the landscape. Such a radical shift in nearly every aspect of social life is unparalleled throughout the entire prehistory of the region. Nevertheless, our current ability to understand these changes is seriously hindered by the lack of attention that has been paid to Early Copper Age settlement sites.

The archaeological record of the Great Hungarian Plain documents social changes that cannot be simply explained in the context of ʻtraditionalʼ cultural evolutionary scenarios. No chiefdoms emerged suddenly out of the Late Neolithic tribal pattern, and no states colonized the region at the beginning of the Early Copper Age. To this end, it was necessary to formulate a research design that sought to answer specific questions regarding changing patterns of interaction and integration throughout the Late Neolithic and Early Copper Age.

Despite the obvious potential of this period for studying anthropological questions of socio-economic change within tribal social systems, the Early Copper Age remains one of the most poorly investigated periods in Hungarian prehistory. This is unfortunate for our archaeological understanding of the region, especially since the preceding Late Neolithic period has been the focus of much recent research and analysis (e.g., Drasovean 1996; Hegedűs and Makkay 1987; Horváth 1987; Raczky et al. 1994). Several Late Neolithic settlement sites have been excavated using modern techniques, as have many of the burials that frequently occur in and around the settlements. From this mortuary and settlement information, a picture of Late Neolithic socio-economic organization is beginning to crystallize (see Kalicz and Raczky 1987a). Systematic research into the Early Copper Age, on the other hand, has focused almost exclusively upon the analysis of a few large cemeteries, resulting in an unbalanced dataset com-

Research Design The research presented in this dissertation focuses upon a 2,000 km2 area in the Körös River Valley, in northern Békés County, eastern Hungary. This region was chosen as a study area because it has been the focus of intensive archaeological survey by Hungarian archaeologists since the late 1960ʼs (see Ecsedy et al. 1982; Jankovich et al. 1989) and because it contains one of the only excavated Early Copper Age settlement 6

Introduction

Structure of the Thesis

sites on the Plain – Vésztő-Mágor (see Hegedűs and Makkay 1987). Within this region, I analyze two separate lines of evidence that relate to the changing patterns of social interaction and integration during the Late Neolithic and Early Copper Age periods: 1) the spatial organization of settlement sites; and 2) regional ceramic patterns. These datasets are then used to interpret the structural signatures that produced them, and to model the social patterns that ultimately produced these signatures.

In this chapter, I have attempted to provide a brief outline of the argument that follows. The following chapter develops the theoretical framework I have outlined cursorily in this introduction. By tracing the development of the tribe concept as it has been used and abused by various anthropologists and archaeologists, I argue that the concept is most effective if it is approached not as a particular stage of cultural evolution, but as a structural form of social integration useful for crosscultural analyses.

The spatial analyses are based primarily upon data collected by various Hungarian research teams, and published in summary form in the Magyarország Régészeti Topográfiája (Archaeological Topography of Hungary) series (see Ecsedy et al. 1982; Jankovich et al. 1989). I revisted several of the sites recorded by the Hungarians during fieldwork in 1998 in an attempt to augment the information provided in the Magyarország Régészeti Topográfiája series, by determining the spatial extent of ceramic distributions and surface features at different Early Copper Age sites within the study area. During revisitation, controlled samples of ceramics were collected in order to assess the size of settlements in different periods at multi-period sites.

Chapter Three discusses the methodological correlates of this theoretical framework, and addresses the archaeological problem of inferring dynamic social systems from static material remains. The middle range theory and bridging arguments are presented and the problems of measuring social interaction and integration in prehistoric contexts are discussed. Chapter Four presents the archaeological background necessary for understanding the radical social changes that occurred on the Great Hungarian Plain, ca. 4,500 BC. It is necessary to provide this introduction to the archaeology of the region before presenting the specific research design in Chapter Five. This latter chapter discusses how the theoretical and methodological variables I discuss in Chapters Two and Three relate to the specific archaeological context – the Körös River Valley system.

The spatial analyses focus upon the size of settlements, and their spatial organization relative to other settlements, topographic features (water courses and soil types), and cemeteries. These variables are analyzed for both the Late Neolithic and the Early Copper Age within the study area and are used to model units of social integration within the study area. The spatial patterns that emerge from these analyses are then compared with the results of the stylistic analyses of Early Copper Age ceramic assemblages, which serve as a rough indicator of the degree of interaction between different sites across the study area.

Chapter Six provides an overview of the study area and presents the sites and assemblage included in the subsequent analyses . Chapter Seven presents the analysis of integration throughout the study area, based upon the spatial data and Chapter Eight presents the analyses of Early Copper Age interaction, based upon the stylistic data from the Early Copper Age ceramic assemblages.

The stylistic analyses are based upon ceramic assemblages collected during the course of surface investigations at several sites, and upon excavated assemblages from two sites in the study area – Vésztő-Mágor and Örménykút 13. The ceramic analyses employ a battery of different stylistic attributes to assess the degree of social interaction that occurred throughout the region. The ceramic patterns that emerge are then compared to the spatial patterns discussed above.

Chapter Nine integrates the analyses presented in Chapters Seven and Eight into a coherent model and attempts to place the study area into the wider temporal and geographic context of the Great Hungarian Plain, and into the wider context of anthropological archaeology.

While these two datasets deal with interrelated variables, and can be expected to covary to some extent, they speak to different aspects of social organization and when employed as independent lines of inference are more effective as analytical tools for investigating changing patterns of social organization. Having analyzed the data from within the study area of the Körös Region at a relatively high resolution, it is then possible to extrapolate the model from this specific area into the greater context of the Plain, and to delineate the basic integrative units and patterns of interaction that characterize each of the two periods. 7

Chapter Two: Tribal Social Theory

The Devleopment of the Tribe Concept

This chapter reviews the development of the notion of tribe in anthropological and archaeological thought, and extracts from this overview what I believe are the most useful elements of the concept for understanding long-term changes in social organization. It is argued that while the term tribe has been associated with a plethora of different types of societies – from family-based hunters and gatherers in Australia to the Congress of the United States (see Colson 1986) – the concept of tribe bears enough anthropological value to retain its use as a heuristic device for facilitating cross-cultural analysis. In large part, the problems that have plagued the concept in the past have stemmed from the fact that the tribe concept has been dealt with almost exclusively within the context of the short-term perspective dictated by the ethnographic record (see Wobst 1978). By approaching the concept of tribe from a primarily archaeological viewpoint, it is possible to circumvent the majority of the problems that have resulted in the unfortunate abandonment of the concept by both anthropologists and archaeologists in recent years (for a recent discussion of the abandonment of the tribe concept, see Beinart 1999).

Since the time of Morgan (1851), the concept of tribe has been plagued by the tendency of earlier generations of anthropologists to generate attribute lists that attempt to pigeonhole societies into different classificatory groupings. Early attempts at such classificatory schemes were based upon unilineal evolutionary paradigmatic approaches (see also Spencer 1896; Tyler 1871), wherein 19th century European civilization was envisioned as the ultimate pre-destined form of social organization to which all societies were inevitably progressing (see Trigger 1990). As I will demonstrate, several of the characteristics that were initially attributed to tribes within this teleological context continue to plague more recent formulations of the concept. Morganʼs (1851) initial social typology placed human societies into three developmental ʻstagesʼ through which he believed all societies necessarily passed – Savagery, Barbarism, and Civilization. Each of these stages was indicated by a particular technological repertoire, and was associated with a particular subsistence strategy and political form. This error – to group together societies based upon a plethora of characteristics which are understood to be intimately intertwined – was perpetuated throughout the following century in the works of various influential authors, such as White, Service, and Sahlins (see Feinman and Neitzel 1984 for an excellent discussion of the problems with ʻtypological approachesʼ). Nevertheless, Morganʼs initial discussion of tribal society set the terms for the way in which both the term and the concept would be employed during the next century.

I argue that the long-term perspective available through the archaeological record demands that the concept of tribe be approached in an essentially structural manner. By arriving at a structural definition of what it means to be tribal, it becomes possible both to delineate those socio-structural features that can be considered characteristic of tribes, and to construct an archaeological methodology that will allow us to identify specific material variables that can be measured to track long-term changes within (and between) the structure of different tribal systems. The critical variables assessed using this methodology will necessarily be structural in nature, and as such can be applied directly to the cross-cultural analysis of tribes in different geographic and temporal contexts (for an excellent disuccion of tribal social structure, see Kelly [1977, 1993]). This is by no means a novel idea, since it is essentially the stance that was taken by British functional-structuralist anthropologists during the early decades of this century (e.g., Radcliffe-Brown 1948, Evans-Pritchard 1940), and reiterates the point Steward (1951) attempted to make in his seminal article “Levels of Sociocultural Integration: An Operational Concept.”

Morgan used the term tribe to refer to linguistically homogeneous cultural units: Each tribe was individualized by a name, by a separate dialect, by a supreme government, and by the possession of a territory which it occupied and defended as its own. The tribes were as numerous as the dialects, for separation did not become complete until dialectical variation had commenced. Indian tribes, therefore, are natural growths through the separation of the same people in the area of their occupation, followed by divergence of speech, seg9

Chapter Two

mentation, and independence (Morgan 1995[1851]:93).

tive mechanism (see Radcliffe-Brown 1948:ix). The functional aspect of this perspective was based, in large part, upon Durkheimʼs concept of ʻsolidarityʼ (see Harris 1968:516).

Morgan envisioned tribes as forming due to a gradual outflow, or budding-off, of groups from a hypothesized geographic tribal centre. Over time, these emigrants would acquire distinct cultural traits and, eventually, linguistic differences, thus creating new tribes (see Morgan 1851:95).

Radcliffe-Brown delineated Andaman social structure as consisting of independent and autonomous small communities, each “leading its own life and regulating its own affairs.”

There are two main points to make here. First, Morgan cites as a causal factor in the formation of tribes “a constant tendency to disintegration.” As I will demonstrate, this notion persists in even some recent archaeological discussions of tribes, which are commonly understood as regionally-integrated systems that develop out of a quagmire of disaggregated bands (e.g., Braun and Plog 1982). Second, the principle of segmentation was already present in Morganʼs initial formulation of the concept as an anthropological classification of society.

These local groups were united into what are here called tribes. A tribe consisted of a number of local groups all speaking what the natives themselves regarded as one language, each tribe having its own language and its name. The tribe was of very little importance in regulating the social life, and was merely a loose aggregate of independent local groups… Within the local group the only division was that into [nuclear] families. These were the only social divisions existing among the Andamanese, who were without any of those divisions known as ʻclansʼ which are characteristic of many primitive societies (RadcliffeBrown 1948:23).

Durkheimʼs (1893) tangential contribution to the topic stressed the principle of segmentation, or mechanical solidarity, to distinguish less economically complex societies – what we now call bands and tribes – from those societies that exhibit organic solidarity, or economic specialization – chiefdoms and states. Although Durkheim was concerned explicitly with the development of the division of labor, his basic classificatory scheme carried with it the assumption that changing economic strategies occurred hand-in-hand with particular political forms. As Lewis Coster notes in his introduction to The Division of Labour:

Each of the tribal units occupied a particular territory, and spoke a different dialect. As was the case with Morgan, Radcliffe-Brown defined a tribe an essentially linguistically homogeneous region that was associated with a particular territory.

Durkheim…was, by and large, beholden to a structural explanation of moral phenomena…The essential differences between types of society were to be sought on the structural or morphological level. The causal arrow in the analysis of social phenomena went largely from productive relations and structural linkages between people to moral or legal systems of thought. (Coser 1984:xviii).

The largest political segment among the Nuer is the tribe. There is no larger group who, besides recognizing themselves as a distinct local community, affirm their obligation to combine in warfare against outsiders and acknowledge the rights of their members to compensation for injury (Evans-Pritchard 1940:5).

E. E. Evans-Pritchard, a student of Radcliffe-Brownʼs, assumed an explicitly structuralist perspective in his work The Nuer (1940):

Nuer tribes had no common organization or central administration, although they sometimes formed loose federations. In this formulation, a tribe was defined in terms of a group which was recognized by its members as constituting a coherent unit, particularly for the purposes of warfare and homicide retribution. Within the various tribal groupings of Nuer society, Evans-Pritchard noted several structural sub-divisions:

In Durkheimʼs work, the concept of segmentation – in the guise of mechanical solidarity – is combined with Marxist structural principles wherein different economic infrastructures produce different forms of superstructures. This basic structuralist concept of segmentation as being characteristic of less economically complex societies heavily influenced not only the pre-war British Structural anthropologists, but also the work of later writers, such as Steward, Sahlins, and Service.

A tribe is divided into a number of territorial segments and these are more than mere geographical divisions, for the members of each consider themselves to be distinct communities and sometimes act as such. We call the largest tribal segments ʻprimary sectionsʼ, the segments of a primary section ʻsecondary sectionsʼ, and the segments of a secondary section ʻtertiary sectionsʼ. A tertiary tribal section consists of a number of villages which are the smallest political units of Nuerland. A village is made up of domestic groups, occupying hamlets, homesteads, and huts (Evans-Pritchard 1940:5).

During the early decades of this century, several British anthropologists began working with tribal societies in different parts of the world, bringing a functional-structuralist perspective to the discipline. Influenced by French sociologists writing at the turn of the century, such as Henri Hubert and Emile Durkheim , members of the British school proposed an ethnographic method that combined a focus upon structure and function. This functionalist perspective lead RadcliffeBrown to a methodology that was cross-cultural in nature, and which focused upon each culture as an adaptive and integra10

Tribal Social Theory

Each of these various structural sections formed part of a segmentary system, “by reference to which it is defined, and, consequently the status of its members, when acting as such towards one another and to outsiders, is undifferentiated” (Evans-Pritchard 1940: 4). Like his mentor, Radcliffe-Brown, Evans-Pritchard envisioned these segments as integrating at various levels, each level determining the structural ʻdistanceʼ between the members of different segments.

tural integration to be used as a methodological device:

While the British structural-functionalist perspective proved extremely useful for describing social relations within static cultural contexts, it inevitably failed to formulate the significant sociocultural laws it had proposed to produce. Harris attributed this failure to the structural-functionalist tendency to allot social structure a central, primary, role to the expense of subordinating other techno-economic parameters (see Harris 1968:524).

He argued that the concept “provides a new frame of reference and a new meaning to pattern; and it facilitates cross-cultural comparison” (Steward 1955:52).

The cultural evolution of Morgan, Tylor, and others is a developmental taxonomy based on concrete characteristics of cultures. The concept of levels of sociocultural integration, on the other hand, is simply a methodological tool for dealing with cultures of different degrees of complexity. It is not a conclusion about evolution (Steward 1955:52).

Steward built upon Redfieldʼs (1941, 1947) distinction between folk societies and urban societies, noting that by establishing an empirically-based typology of integrational levels, it would be possible to examine the incorporation of smaller (what he called ʻsimplerʼ) societes into larger sociocultural systems, “... and to make generalizations about processes which go beyond what Redfield derived from the process of urbanization” (Steward 1955:53). To this end, Steward defined three basic integrational levels: the nuclear family, folk societies (or multifamily sociocultural systems), and states. He conceded that there are probably several levels of sociocultural integration between these three, but that “these are qualitatively distinctive organizational systems, which represent successive stages in any developmental continuum and constitute special kinds of cultural components within higher sociocultural systems” (1955:54). Steward suggested that the concept of sociocultural levels should be used as an analytic tool in the study of changes within particular sociocultural systems, which each consist of parts that developed at different times and which continue to integrate certain portions of the culture.

The structuralist concepts of segmentation and integration figured largely into Stewardʼs argument that societies should be approached in terms of varying levels of sociocultural integration (see Steward 1955). This idea carried over, in somewhat modified form, into the work of Sahlins and Service (1960). Initially, Steward intended the concept not as a component in cultural evolutionary theory, but as a tool for cross-cultural comparison. During this brief time, the tendency to lump together various political, economic, and social attributes became temporarily uncoupled. In Stewardʼs view, a particular structural characteristic – the level of integration – was used as the primary unit of societal analysis. It was only later, when the concept was coopted by Sahlins and Service (1960), that particular levels of integration became equated with particular stages of cultural evolution and were again associated with specific economic, ideological, and political criteria.

Service (1971) built upon Stewardʼs concept of levels of integration, but reincorporated an explicitly evolutionary component to its initial formulation. Despite the various critiques of his now (in)famous Band-Tribe-Chiefdom-State model (e.g. Fried 1968), the strength of Serviceʼs model lies in its focus upon the structural integration of societies:

Steward proposed the concept of levels of integration primarily as a tool for cross-cultural analysis as an alternative to what he called the traditional assumptions about tribal societies (Steward 1955:44). This traditional view was based upon three fundamental aspects of the behavior of members of tribal societies, which Steward rejected. He outlined these aspects in the following manner. First, tribal culture was a construct that represented the ideal, norm, average, or expectable behavior of all members of a fairly small, simple, independent self-contained, and homogeneous society. Second, tribal culture had a pattern or configuration, which expressed some over-all integration. Finally, the concept of tribal culture was understood to be essentially relativistic – meaning that the culture of any particular tradition was seen to be unique in contrast to cultures of other traditions. Steward (1955:46) suggested that while this conceptualization of tribal culture had been a tool useful for analysis and comparison, it was of little utility in dealing with culture change. In place of this normative perspective, Steward proposed the concept of levels of sociocultural integration.

If the general evolution of society consists, as some have said, of not only a multiplication of groups but also of an increase in specialization into economic and political parts, ritual units, and the like, then tribes have advanced over bands only in the sense of multiplication and integration of parts. This is why the present book chooses as the discriminating criterion of stages the form of integration. At each level the integration of parts is carried out differently (Service 1971: 132, original emphasis). Within this scenario, the defining characteristic of tribal social organization is the structured organization of segmentary units of a similar scale, usually lineages or groups of lineages (bands), via some integrative institution. According to Service, this institution usually takes the form of a pan-tribal sodality, which cross-cuts lineages and unites groups of bands into tribes. As Service (1971:100) notes:

Steward initially intended the concept of levels of sociocul-

11

Chapter Two

A tribe is of the order of a large collection of bands, but it is not simply a collection of bands. The ties that bind a tribe are more complicated than those of bands and, as we shall see, the residential segments themselves come to be rather different from bands (original emphasis).

autonomous segments that are held together by their likeness to each other (i.e., by mechanical solidarity) and by pantribal institutions, which cross-cut the primary segments. For Sahlins (1961), the segmentary lineage system is a substitute for the fixed political structure that tribal societies are incapable of sustaining.

This basic tenet – that tribes consist of integrated social segments – is not at all unreasonable. The problems arise when particular levels of sociocultural integration (e.g., tribe) are a priori associated with particular economic (e.g., horticultural) or mobility (e.g., sedentary) characteristics. However, the fact that previous authors inaccurately chose to associate different forms of social integration with these particular characteristics – thus contrsucting an idealized “Neolithic society” –does not undermine the validity of their use of this criterion for identifying units of different social systems that are amenable to cross-cultural analysis, as Steward initially intended.

Sahlins built upon Stewardʼs notion of levels of integration by linking varying levels of organization with sectors of social relations. Within this ʻsectoral modelʼ, “relations become increasingly broad and dilute as one moves out from the familial navel” (Sahlins 1968:16). Sahlins understood cooperation and social interaction to be most intense at the tribal “core” – the homestead and hamlet. Thus, the degree of integration decreases as the level of organization increases, and degrees of sociability diminish as fields of social relation broaden. In his own words:

Inherent in Serviceʼs concept of tribe is a certain degree of fragility, and a tendency towards disunity:

The model before us is set out in social terms. But more than a scheme of social relations, it is an organization of culture. The several levels of organization are, in the jargon of the trade, levels of sociocultural integration; the sectors, sectors of sociocultural relations… Functions are regulated by levels of organization, and transactions by sectors of relation (Sahlins 1968:16).

Considering the lack of institutional political means of unity and the absence of organic solidarity, and considering such grave sources of disunity as feuds, it seems remarkable that a tribe remains a tribe. It seems sensible to reaffirm that external strife and competition among tribes must be the factor that provides the necessity for internal unity (Service 1971:104; original emphasis).

Within Sahlinsʼ holistic approach, tribes can subsume an astonishing array of different societal arrangements, from segmentary tribes to chiefdoms (see Sahlins 1968:20). He envisioned many intermediate arrangements between these two ends of the tribal spectrum. These include: conical clans, segmentary lineage systems, territorial clans, dispersed clans, and local cognatic descent groups.

While the concept of levels of sociocultural integration, as Service uses it, provides a method useful for classifying different societal forms, it suffers from a static quality that does not account adequately for the degree of dynamic flexibility documented in the archaeological record. That is, even the roughly-hewn forms of social integration that Service employs suffer from the fact that they are themselves static idealizations of dynamic phenomena. Although Serviceʼs model allows for a certain range of variability within each of his forms of social integration (e.g., lineal and composite tribes), it does not account for the basic fact that the social structures, which themselves define the different evolutionary stages, inherently allow for a certain degree of integrative, or ʻorganizational flexibilityʼ (see Fowles and Parkinson 1999). This point will be addressed in greater detail below.

In addition to trying to blur the line between different social classifications, Sahlins also attempted to decouple the relationship between social forms and economic practices, “…while it is true that most tribesmen are farmers or herders, thus cultural descendants of the Neolithic, not all are… The Neolithic, then, did not necessarily spawn tribal culture. What it did was provide the technology of tribal dominance” (Sahlins 1968:3). Sahlinsʼ elegant, holistic, approach to cultural analysis continues in his later work (e.g., 1972), which is directed towards understanding primitive economic systems.

Marshall Sahlins also subscribed to a version of the basic Band-Tribe-Chiefdom-State evolutionary scheme and distinguished between bands and tribes in the following manner:

Friedʼs visceral reaction to the Band-Tribe-Chiefdom-State model, and to Service and Sahlins in particular, was based upon his paradigmatic assumption that social classification should be based upon the differential access to status positions available to individuals in different societies. This led to his tri-partite classificatory system of egalitarian, ranked, and stratified societies. Since Fried understood both bands and tribes to be essentially egalitarian in nature, he saw no need to sub-divide egalitarian societies into two discrete groups. In a series of articles (e.g., Fried 1968) and a book (Fried 1975), he launched a series of attacks upon the concept of tribe, arguing that tribes tend to occur only in secondary contexts, “as a con-

A band is a simple association of families, but a tribe is an association of kin groups which are themselves composed of families. A tribe is a larger, more segmented society. Without implying this as the specific course of development of tribes, we may nonetheless view a tribe as a coalescence of multi-family groups each of the order of a band (Sahlins 1961:324). In Sahlinsʼ view, tribes consist of economically and politically 12

Tribal Social Theory

sequence of the impinging on simple cultures of much more complexly organized societies” (Fried 1975:10).

Despite this fact, certain threads permeate each of the models, suggesting the existence of some ethnographic patterns that need to be considered while formulating an archaeologically useful notion of tribe.

Friedʼs critique certainly deserves consideration, not least because it constitutes the inception of the replacement of the term tribe by much more cumbersome phrases, such as ʻmiddle-range societies.ʼ This is unfortunate, in my view, for Friedʼs arguments seem to augment, rather than discredit the concept of tribe as a construct useful for cross-cultural analysis.

Attributes Associated with the Tribe Concept in Ethnology The various attributes frequently associated with the tribe concept include:

For example, Friedʼs contention, that tribes form only when less complex societies are affected by more complex ones, seems to beg the question: why do certain societies turn into tribes when they come into contact with states and empires, and others do not? Friedʼs inability to answer this simple question exposes the Achilles heel of his entire argument, which is based upon the untenable position that tribes exist only as discretely-defined cultural units, a notion wholly explicable by his dependence upon the ethnographic record. When viewed solely through the short-term perspective available through ethnography, the distribution of tribes across the globe would certainly seem to correlate with those regions which were heavily influenced by historical state-level societies: North America, New Guinea, South America, etc. Nevertheless, a closer look at the archaeology of these same regions would reveal that several tribes had emerged prior to contact, and indeed prior to the indigenous development or impact of state-level societies in these regions. Furthermore, even in the same areas where Fried argues that contact produced tribal systems, he fails to explain why certain societies, such as the Shoshone of California, or the Australian hunters and gatherers, never developed into tribal units, but remained un-integrated bands.

1. The concept of segmentation, or ʻmechanical solidarity;ʼ 2. A tendency towards entropy, or disunity; 3. The idea that tribes exist only as discrete entities, with well-defined social and geographic boundaries; 4. The idea that tribes are somehow ʻtransitionalʼ between less complex social forms, such as bands, and more complex forms, such as chiefdoms and states. While some of these attributes, most of which can be traced all the way back to Morgan, are extremely useful in formulating a concept of tribe that can be useful for understanding long-term changes in social organization; others are merely remnants of misconceived notions that should be summarily discarded. Segmentation In formulating an archaeological definition of tribe that will be useful in investigating long-term social change, it is most useful to focus upon those structural features that are likely to be identifiable archaeologically, and which are likely to indicate something about the nature of social organization. To this end, the principle of segmentation is particularly useful, and indeed seems to be a structural characteristic common to most egalitarian societies. Furthermore, it is possible through a variety of methods to measure the size and scale of different social segments, and to compare their organization both diachronically and cross-culturally (see, for example, Kelly 1985).

Friedʼs formulation of tribal society suffers from a static quality that precludes the possibility for tribes to assume a variety of different configurations throughout their ontogeny. The reason why tribes emerge in some instances of contact, and not in others, must have something to do with the structure of their social relations prior to contact. Some societies exhibit certain structural features – such as sodalities – that allow them to organize into more, and more complex, integrative units than other societies. These include tribes. Other societies lack the structural mechanisms necessary to integrate into these more complex units – these are bands. The structure of social relations prior to the time that societies are impinged upon by more complex ones necessarily determines the trajectories these societies assume after contact. Friedʼs inability, or unwillingness, to accept this basic fact stems from his overreliance upon the ethnographic record, which because of its short-term perspective is limited in its ability to track trajectories of change that occur on a much longer diachronic scale.

Tendency Towards Disunity In contrast to the useful idea that tribes are segmented, the notion that tribes tend towards disunity seems to be a vestigial characteristic that has been perpetuated by historical developments within the discipline. In Morganʼs initial formulation of the tribal concept, he argued that the reason tribes were segmented was because they were constantly fissioning. This basic notion carried through in the work of Sahlins and Service who saw entropy not as a causal feature in the evolution of tribes, but as the unfortunate result of a lack of centralization. In their view, tribes were plagued by external strife and it was only through constant competition with each other that they

This tendency – to construct classificatory systems based exclusively upon ethnographic and ethnohistoric examples – resonates throughout all of the models discussed above.

13

Chapter Two

managed to sustain any degree of cohesion. Warre was allotted a primary, central role.

Sahlins, on the other hand, used the term tribal to refer to the range of evolutionary forms that exists between bands and states, including chiefdoms. Within this scenario, tribes are distinct from civilizations primarily because the former are in a Hobbesian condition of war, “… Lacking specialized institutions of law and order, tribes must mobilize the generalized institutions they do have to meet the threat of war. Economics, kinship, ritual, and the rest are so enlisted” (Sahlins 1968: 12-13). Within the tribal form, Sahlins distinguishes between segmentary tribes and chiefdoms:

While there does seem to be a tendency for tribes to develop in groups, perhaps indicating some sort of inter-dependent relationship between them (see, for example, Braun and Plog 1982), the nature of these relationships, and in particular the nature of intra- and inter-tribal aggression, seems to vary widely (see Keeley 1996). At times, aggression in tribal societies consists essentially of intra-tribal feuds, occurring between family units (e.g., the Yanomamo; Chagnon 1983), at other times, it consists of all-out warfare between highly organized confederacies (e.g., the Iroquois; Snow 1994; see also Ferguson and Whitehead 1992). While there may, in fact, be some social logic behind these changing patterns of aggression, their existence should not lead us to presuppose a tendency towards disunity. Rather, it is more productive to envision different mechanisms that facilitate fission, at times, and fusion, at other times. This more accurately represents what happens within tribal systems, especially when they are viewed from the diachronic perspective of the archaeological record.

The segmentary tribe is a permutation of the general model in the direction of extreme decentralization, to the extent that the burden of culture is carried in small, local, autonomous groups while higher levels of organization develop little coherence, poor definition, and minimum function. The chiefdom is a development in the other direction, toward integration of the segmentary system at higher levels. A political superstructure is established, and on that basis a wider and more elaborate organization of economy, ceremony, ideology, and other aspects of culture (Sahlins 1968:20). As discussed earlier, Sahlins conceded that many intermediate arrangements stand between the most advanced chiefdom and the simplest segmentary tribe.

Tribes as Discrete Entities Another ethnographic fiction that has been perpetuated by the misrepresentation of tribal systems is the notion that tribes exist exclusively as discrete entities with very well-defined social and geographic boundaries. While some tribal societies certainly do exhibit clear boundaries, others appear as smears across the archaeological landscape, with few discernible internal or external boundaries. The segmented nature of tribal systems, combined with their tendency to fission and fuse given different social and environmental conditions, results in a social picture that assumes discrete boundaries at only isolated moments in time. The tendency of different segments within the system to constantly renegotiate their relationship with each other can preclude the formation of established social boundaries over the long term, usually resulting in a complicated archaeological picture with fuzzy lines approximating the borders between different prehistoric ʻgroups.ʼ

Unlike Service and Sahlins, who argued that tribes should be considered evolutionary stages between bands and states, Fried contended that tribes develop only in secondary contexts when band societies are impinged upon by much more complex societal forms. In this case, tribes are seen not as transitional entities on an evolutionary ladder, but as entities which develop in geographically transitional environments. While their views varied dramatically, all three authors conceded that their evolutionary models are based not upon long-term processes documented in the archaeological record, but on ethnographic examples. This focus upon the short-term perspective available through the ethnographic record results in the placement of tribes as transitional, ephemeral formations that occur between bands and states, evolutionarily and geographically (see Gregg 1991:1). An archaeological perspective of tribal social systems would suggest, rather, that tribes were the dominant social form on the planet for several thousand years following the end of the Pleistocene.

Tribes as Transitional Social Forms A final characteristic associated with tribes based upon ethnographic cases is the notion that they are transitional (read ephemeral) formations that exist evolutionarily or physically between bands and states. The idea that tribes are a stage on the evolutionary ladder dates back to Morganʼs (1851) unilineal stages of Savagery, which subsumes both bands and tribes, and Barbarism, which subsumes both tribes and chiefdoms. This basic idea was rephrased by Sahlins (1961) and Service (1971), both of whom were heavily influenced by Stewardʼs notion of multilinear evolution, and by the concept of sociocultural levels of integration. Service considered tribes to be transitional between bands, which are segmented and dis-integrated, and chiefdoms, which are centralized and ranked.

The last thirty years have witnessed the near abandonment of the tribe concept in ethnology in favor of, on the one hand, a tendency towards historical particularism with the analytical emphasis placed upon the cultural variables that distinguish one society from another. On the other hand, this trend has been accompanied by a tendency to employ classificatory schemata that refer ultimately to similar social forms, such as ʻmiddle-rangeʼ or ʻtransegalitarianʼ societies. As a result, the burden of exploring cross-cultural comparisons between tribal

14

Tribal Social Theory

societies has fallen largely upon the shoulders of archaeologists.

then it probably originated in order to perform those functions (Braun and Plog 1982:505).

This brief overview of how tribe – as both a term and a concept – developed within the context of anthropological social theory provides a basis for reviewing how it has been dealt with in archaeological contexts. In particular, the concepts of segmentation (mechanical solidarity), and integration tend to be used to characterize tribal systems in archaeological contexts.

Braun and Plog contended that these functionalist notions are rooted in the misconception that societies are relatively bounded discrete phenomena analogous to organisms. Rather, they suggested the adoption of an approach that emphasizes the adaptive role of integration as a response to socioenvironmental unpredictability. Citing case studies from the American Southwest and western Illinois, Braun and Plog argued that tribal social networks “will respond to increasing local risk with increasing, and increasingly formalized, regional connectedness” (1982:518).

Archaeological Concepts of Tribe Archaeological approaches to tribal systems have focused upon tribes as unranked regionally-integrated segmentary systems. Beginning with Strueverʼs (1968) discussion of the Hopewell interaction sphere, and continuing through Braun and Plogʼs (1982) seminal article, tribes have been dealt with primarily as nonhierarchical, regional social networks that develop as an adaptive organizational response to changes in the degree of local environmental unpredictability. While archaeological treatments of tribal systems have attempted to inject some temporal depth into discussions of their place in cultural evolutionary scenarios, even the most elegant archaeological models of tribes suffer from a static quality that does not sufficiently describe their flexible nature. Most archaeological approaches to tribal societies have focused upon the ʻevolution of tribal social organizationʼ – the process known as ʻtribalizationʼ (see Emerson 1999) – rather than on exploring the degree of variability exhibited in different tribal structures and trajectories. Nevertheless, certain organizational elements recur frequently in archaeological attempts at dealing with tribal societies.

Saitta (1983) attacked Braun and Plogʼs model on both methodological and theoretical grounds. In particular, Saitta expressed dissatisfaction with Braun and Plogʼs suggestion that tribal societies lack the “concrete decomposability” that would allow the assessment of the internal relationships between discrete institutional subsystems. This, he contended, leads Braun and Plog to consider tribal networks as a whole in relation to the outside: “environmental unpredictability is seen to affect the system, rather than sets of individuals conceivably differentially related within it” (Saitta 1983:822). He argued that Braun and Plog adopt a circumscribed notion of the spatial scale over which tribal social relations are transacted. Saitta suggested it would be more productive to model alternative tribal arrangements based upon the production and distribution of social surplus. In their reply to Saittaʼs comments, Plog and Braun (1984) attempted to clarify their position regarding the non-decomposability of tribal systems. Citing Simon (1973), they contended that with a given time span, and particular measurement techniques, the dynamics of only some hierarchical levels of a system may be observable. This notion is somewhat reminiscent of Stewardʼs (1951) concept of levels of integration, which in turn influenced Sahlinsʼ “sectoral model” of tribal organization (see Sahlins 1968: figure 2.1).

Following Middleton and Tait (1958), Gluckman (1965), Sahlins (1968) and Service (1971), Braun and Plog (1982: 504) defined tribal systems as: … social networks integrated by cross-cutting panresidential institutions, but lacking class structure or full-time segmental specialization. Such systems also may be termed “nonhierarchical,” in the sense that decision-making occurs primarily through consensus rather than through the full-time exercise of power by formally sanctified authorities.

Benderʼs (1985) work on the development of tribes during the Late Archaic of the American midcontinent stressed the role of social relations in the formation of tribal systems. Citing more elaborate burial rituals, increased production of socially valued artifacts, and far-flung material exchanges, she argued that these “flashes of ceremonialism” (cf. Tuck 1978) were most likely integral to alliance and exchange, and perhaps served as a locus for increasing social differentiation. Benderʼs explicitly Marxist viewpoint, which focuses upon adaptation in terms of social reproduction, was proposed in reaction to the prevalent notion that humans are “rational, cost-conscious, adaptive strategists,” a mirage which Bender argued “was conjured up by the Processualist insistence upon General Laws, which involves a quite unnecessary rejection of both specific history and of principles of social structure in favor of an assumed ecological common denominator” (Bender 1985:52; see also Hodder 1982). While Benderʼs

Braun and Plog were reacting to the traditional depiction of tribal regional social systems as developing in response to conditions of competition between, or conflict among, autonomous residential units. They argued that this traditional depiction relies upon two incorrect assumptions: 1. That local autonomy is a natural or universally sought condition in human social relations; and, 2. That if an institution generally performs certain integrative functions in its known ethnographic setting,

15

Chapter Two

approach is refreshing in its attempt to redress the environmental determinism that plagued much of archaeology during the early 1980ʼs, its failure to provide precise definitions of significant terms – such as tribe and tribal – limits its utility in the current discussion.

The second model of tribalization Haas outlined was the ʻcooperativeʼ model proposed by archaeologists, such as Braun and Plog (1982) and Bender (1985). This risk-based model, discussed above, emphasizes cooperation, integration, and communication as critical variables in the development of tribes. Haas noted that the process of tribalization can occur without the formation of concretely bounded tribes, and that “it might be possible to have a tribal form of organization, without having a discrete social unit readily identifiable as a bounded tribe” (Haas 1990:174). Haas argued that environmental stress and risk may have triggered an increase in social connectedness in prehistoric northeastern Arizona, hence lending some credence to the second model. But, “On a second level, there is evidence that the declining environmental conditions also stimulated conflict and raiding between groups on a broader, regional scale” (Haas 1990:188), as argued by the traditional theorists. A similar model has recently been suggested by Emerson (1999) for the later prehistory of the American Midwest.

Creamer and Haas (1985) contended that the archaeological correlates of a tribe will: …be manifested in a number of distinct economic and social patterns, including patterns in: settlement, architecture, centralized labor organization, surplus production, storage, specialization, rank, status goods, trade, boundaries and stress (Creamer and Haas 1985:740). Creamer and Haas were more concerned with discerning tribes from chiefdoms, rather than bands. Nevertheless, they provided an explicit definition of what they meant by tribe: The composite units within a tribe…are largely independent of one another economically, but politically and ceremonially interdependent. Production … is at a subsistence level, with limited surplus production. Existing surplus…is concentrated at the level of the household, community, or extended kin unit… With regard to status, while tribes are non-hierarchical, there may be some social “ranking” in tribal societies. Some kin units may have higher status than others, and some individuals may attain higher status…The tribal system is nevertheless decentralized, and the higher status individuals are neither placed structurally in leadership roles nor assigned responsibility for making decisions for the group as a whole (Creamer and Haas 1985:739).

Haasʼ approach to tribes led to his insightful conclusion that tribalization must be understood as a process, rather than as a “... one time historical event that suddently transforms a band society into a tribe” (Haas 1990: 189) – a position also reitterated by Emerson (1999:10). Although some recent archaeological treatments of the topic have stressed the more enduring, and stable characteristics of tribal societies (for examples, see Gregg 1991; see also Habicht-Mauche 1990, 1995), most discussions of this sort have mutated into discussions of “ethnicity” and “identity” (e.g., Jones 1997; see also Shennan 1989). In this regard, Emerson has recently noted:

This formulation of the tribal concept builds upon various previous models, including Service (1971), Sahlins (1961, 1968), Steward (1955), and Fried (1968). Here again, we find the basic notions of segmentation and regional integration combined with the idea that tribes are essentially decentralized and ʻegalitarian.ʼ

Ultimately it appears that, although more difficult than we might have imagined at first, it may be possible to identify ethnicity and tribalism, not, as is often attempted, as discretely bounded material clusters, but as processes reified in the material remains of past societies (Emerson 1999: 10; original emphasis).

In a later publication, Haas (1990:172) described tribes in the following manner:

“Tribalization,” Segmentation, Integration, and Decentralization

In simplest terms, a tribe is a bounded network of communities united by social and political ties and generally sharing the same language, ideology, and material culture. The communities in a tribe are economically autonomous and there is no centralized political hierarchy.

This brief overview of how tribes have been variously treated by archaeologists highlights several organizational elements that appear recurrently, and which must be considered in formulating an archaeologically operational definition of tribe. The majority of archaeological attempts at dealing with tribal systems have focused upon understanding their evolutionary position between bands and chiefdoms, or states. To some extent, this focus upon the evolution of tribal systems – the process of ʻtribalizationʼ – is to be expected, since archaeologists enjoy the luxury of the long-term perspective available through the archaeological record, and are therefore able to speak to questions of origin, unlike their ethnographic colleagues. The unfortunate aspect of this tendency to focus

Haas delineated two models that have been proposed for explaining the process of tribal evolution. The first he called the ʻtraditional model,ʼ which was offered by cultural anthropologists such as Service (1971), Adams (1975), and Sahlins (1961, 1968). The traditional model assigns a causal role to warfare in the evolution of the tribe, and envisions warfare as inducing consolidation through cooperation rather than conquest (see Haas 1990:172). 16

Tribal Social Theory

upon origin, and evolution, is the resultant lack of research that has been directed at understanding the cross-cultural variability that characterizes tribal societies. Nevertheless, the recurrent patterns that emerge from the way archaeologists have variously dealt with tribes deserve additional attention. While some of these characteristics carried over from ethnographically derived preconceptions of tribal societies, others only developed when tribes were approached archaeologically. The three most pervasive of these characteristics, which I believe are most useful in formulating an archaeologically useful definition of tribe, are segmentation, integration, and decentralization.

logical approaches to tribal societies is the idea that tribes are decentralized. This decentralization usually refers either to: 1) a lack of an institutionalized central political authority, which is usually associated with more complex forms of sociopolitical organization, such as chiefdoms or states; or to, 2) a lack of institutionalized economic specialization, which results in a social redundancy that prohibits the ability of one social segment to acquire and maintain a centrally powerful role. It seems that when most archaeologists write that tribes are decentralized, they are using the term in a rather loose manner, referring simultaneously to both the political and economic aspects of the term. There is some sense to this, since the economic is only vaguely distinguishable from the political in most tribal systems. This is precisely the point I believe Sahlins was trying to make with his ʻsectoral modelʼ of tribal relations, wherein he suggested that “… the social system … becomes weaker where it is greater: the degree of integration decreases as the level of organization increases” (Sahlins 1968: 16).

Segmentation Perhaps the most pervasive characteristic associated with tribal systems in both ethnographic and archaeological contexts is the idea that they are segmented. As noted above, the idea that tribes can be characterized by segmentary forms of organization can be traced back to Morgan (1851). Durkheim (1984) associated the term with mechanical solidarity, which later authors, such as Sahlins and Service, used to characterize bands and tribes, economically and politically (see also Kelly 1985). This notion carries over into archaeological approaches to tribal societies. Although different authors argue the degree to which mechanical solidarity – as it refers to the redundancy created by a lack of economic specialization between different social segments practicing the domestic mode of production (see Sahlins 1972) – can vary within tribal systems, there is some general consensus that social segments of roughly similar scale and composition replicate themselves at varying levels within tribal societies. The precise manner in which this integration occurs varies considerably within different tribal societies, but as a general rule it must involve at least some regular integration beyond the extended family unit, or band.

This brief overview of how the concept of tribe has been dealt with in ethnographic and archaeological approaches provides a framework upon which it is possible to construct an operational definition of the term that will be useful for exploring social changes over the long term. A Structural Definition of Tribe Given the multitude of definitions that have been proposed for the term tribe in the past – linguistically homogeneous territorial groups, regionally-interconnected social networks, groups that form in the face of warfare, the United States Congress, a Major League Baseball team – it is perhaps somewhat disturbing that I will now offer another. But the proposed definition is, in fact, not new. Rather, it is based upon those characteristics that have borne themselves out over time, and have been returned to repeatedly in anthropological and archaeological attempts at defining those societies that evoke the term tribal. The definition is based upon structural characteristics, rather than upon other (e.g., economic) attributes, in an attempt to cast the widest net possible. The particular structural characteristics that I propose define a tribe did not originate in support of any generalizing model of cultural evolution. They are elements that many scholars have attributed to societies they somehow believe to be different from other kinds of societies.

Integration Another characteristic that archaeologists frequently attribute to tribal societies is a particular form of integration that links together different social segments in a nonhierarchical manner. This concept can be traced directly to the British structuralists and to Stewardʼs (1955) work on levels of integration, but more often it appears in the archaeological literature in the sense intended by Service (1971) and Sahlins (1961, 1968) – as “…social networks integrated by cross-cutting panresidential institutions,” or sodalities. This structural characteristic is somewhat problematic, since all societies are – by definition of being a ʻsocietyʼ – integrated in some manner. Nevertheless, a certain degree of regularized, institutionalized, integration beyond that of the extended family unit (or band) seems to be what differentiates tribes from bands for most archaeologists.

I have argued above that three structural elements have traditionally been attributed to, or associated with, tribal societies in both ethnographic and archaeological contexts – viz. segmentation, integration, and decentralization. It seems appropriate then, that a structural definition of tribe should include these same characteristics. In the context of this dissertation, the term tribe refers to an essentially segmented and decentralized social structure within which social groups regularly interact with – and integrate at varying levels with – similar

Decentralization A final characteristic that repeatedly emerges in archaeo17

Chapter Two

to a position of central importance in structural analysis (Gearing 1958:1148).

groups beyond the level of the extended family unit, or band. These groups– or social segments – are similar to each other in demographic scale and political organization, creating a social redundancy that allows integration to occur at several levels simultaneously and resulting ultimately in an inherently flexible form of social organization.

Gearing initially applied this concept to the different structural poses assumed by the members of a Cherokee village within a single year. He convincingly demonstrated the way the village was structurally reorganized throughout a single year depending upon different social obligations: as an aggregate of households; as an aggregate of corporate clans; as a unit organized by clans and the body of elders; and as a unit organized by clans, elders and the war organization (Gearing 1958: 1151).

Organizational Flexibility The flexibility within this structure is created by individuals through the social rules they create and follow in their interactions with one other. These rules, which manifest themselves in systems of kinship, religion, and subsistence (among others), offer several options that can be selected given different social and environmental conditions. Defining the tribal form of organization in this structural manner allows the societies in question to be viewed not as static groups with rigidly defined social and geographic boundaries, but as dynamic groups of individuals who simultaneously participate in – and identify themselves as members of – a series of social units, each of which can in turn be viewed as a different level of structural integration (sensu Steward 1955).

Gearingʼs salient analysis brings to the forefront the fact “…that human communities typically rearrange themselves to accomplish their various tasks” (Gearing 1958: 1148). Although Gearing initially intended the term to refer to highfrequency (i.e., intra-annual) changes in the social organization of a single village, it is possible to adapt the notion of structural pose to refer to the processes by which groups of villages – or regions – interact and integrate differentially over the long term. Such an analysis allows tribal social systems to be understood not as static groupings, but as dynamic social systems created from a social structure that allows a variety of different social patterns to be expressed, given different social situations. As Fowles (1997:12) noted, “A tribal society has the potential at any given time to take on one of a multitude of organizational assortments. This flexibility is built into a tribeʼs form of integration.”

I understand the ʻorganizational flexibilityʼ in tribal societies (Fowles and Parkinson 1999), to be rooted in the nature of individual participation within the larger social system. Within any social system, individuals simultaneously identify themselves as belonging to a number of social groups: as members of a family unit, a lineage, a clan, a coresidential group, a trading network, a warrior society, etc. The particular association the individual chooses to express at any given time is dependent upon the social context of a particular situation. Although individuals living in ranked, or even stratified, societies can also shift their identity given different social settings, it is within segmented, and in particular within mechanically integrated and decentralized segmented social systems (i.e., tribes) that this flexibility assumes the widest range of possible configurations. While bands are limited in their ability to integrate regularly at any level beyond that of the extended family unit, or band, chiefdoms and states are similarly restricted in their ability to dis-integrate into autonomous segments, due to the existence of an institutionalized central authority which actively works against this.

The structural pose concept has profound implications with regards to the material correlates of tribal social systems over the long term. Since the structural organization of the segmentary units (that make up tribal societies) assumes a variety of poses throughout a single year, and since the corpus of structural poses exhibited in any given year can be expected to change over time in accordance with varying social and environmental conditions, then the scale and intensity of integration of tribal systems can be expected to fluctuate over time according to changing socio-environmental conditions. The trends that emerge from these long-term fluctuations can, in turn, be thought of as ʻsocial patterns,ʼ which apply not to the different ways a particular society is integrated throughout a given year, but rather to the generalized degree and scale of interaction and integration that a society assumes over a much longer period of time, perhaps hundreds of years. This last point has important ramifications with regards to our archaeological understanding of tribal social systems, and goes a long way towards explaining the problems encountered by ethnographers who have much less time depth at their disposal.

Structural Poses and Social Patterns As Fowles (1997) pointed out, this flexibility assumes several different temporal frequencies, and can be understood as functioning at a variety of levels. In its most ethnographically apparent form, it is similar to what Gearing called a ʻstructural poseʼ:

In order to modify Gearingʼs notion of structural pose to deal not only with the internal organization of a single village throughout a single year, but with the more generalized organization of social systems over the long-term, it is necessary to extend the concept to encompass the organization of social units within a region and over a much longer period of time.

The notion of structural pose…draws attention to the wellestablished fact that the social structure of a human community is not a single set of roles and organized groups, but is rather a series of several sets of roles and groups which appear and disappear according to the tasks at hand. The notion of structural pose elevates that known fact 18

Tribal Social Theory

In his application of this theoretical framework to the prehistoric American Southwest, Fowles (1997:63) argued that:

These material correlates – or ʻstructural signaturesʼ – are produced by the long-term operation of particular social patterns, which themselves are ultimately created within the context of an overall social structure. I understand this social structure to be inherent within the cosmology, kinship, and other social institutions to which the members of a particular social group subscribe. The modeling of this structure – the social rules that structure social behavior – is the ultimate goal of archaeological analysis.

The sociopolitical scale of Southwestern regions tended to expand and contract every two hundred or so years throughout prehistory. From this perspective it is clear that the tribal form of sociopolitical integration in both the Taos District and in the Greater Southwest after about AD 400 did not truly ʻevolve,ʼ but rather took on a series of different structural poses (original emphasis).

Latent and Active Flexibility

This modification significantly alters the original definition of the concept, but retains as a critical element the idea that “a human community does not have a single social structure; it has several. Put otherwise, the social structure of a society is the sum of the several structural poses it assumes around the year” (Gearing 1958:1149). Implicit within Fowles (1997) application of this concept to the Southwest is the idea that regions can be characterized by what I call ʻsocial patternsʼ, which relate to the relative degree of regional interaction and integration that occurs within a geographically-defined region over a given time.

The organizational flexibility inherent within a tribal social structure may, therefore, be latent, implicit and seldom expressed; or it may be active,explicit, and frequently expressed. Latent flexibility manifests itself in tribal contexts in the form of social rules about ancestry and structural relations that are recognized by the members of society but which figure only tangentially into the everyday workings of life. Active flexibility, on the other hand, usually takes the form of social rules that govern everyday interactions, and which offer the individual the option to participate in one of several social groups of which he or she understands himself or herself to be a member. Ethnographers are often at liberty to witness active flexibility at work within tribal societies – Gearingʼs (1958) article on structural poses is one such description. But the subtle workings of latent flexibility often emerge at a temporal scale too gross to be detected by ethnographic accounts.

These social patterns are the products of the repeated cycling of social organization through sets of structural poses whose formal characteristics evolve and change over time, themselves being contingent upon historical processes. In this regard, it is important to remember Giddensʼ notion that the consequences of oneʼs actions, whether intentional or not, impact future trajectories (Giddens 1984).

Within this framework, the archaeologically-identifiable ʻstructural signaturesʼ serve as a proxy through which it is possible to reconstruct particular social patterns, and ultimately the very social structures that produced them. By modeling how the structural signatures of a society change over time, archaeologists will find themselves in a position to model trajectories of structural change within different societies. The different trajectories that emerge from different cultural contexts can then be compared to each other in a broadly crosscultural manner.

It is useful to think of these generalized patterns as existing on a continuum of varying levels of integration and interaction. At one end of this continuum is the family level of integration, at the other end is the regional level of integration. The members of any tribal society regularly interact with – and are integrated into – all of these levels, depending upon what they are occupying themselves with at any given time: farming, trading, craft-production, feasting, etc. This being the case, the problem then becomes determining the frequency at which the individuals within the society interact with – and are integrated into – these varying levels.

Flexibility and Tribal Structure: An Ethnohistoric Example

Focal Levels of Integration

The theoretical perspective I have presented relies heavily upon several assumptions about how individual behavior is molded within a larger cultural context. Unlike the British structuralists (e.g., Radcliffe-Brown 1948), the structural model I propose is not predicated upon the idea that social structures evolve exclusively to fulfill particular functional roles. And unlike the later French structuralists (e.g., Lévi-Strauss 1976), I make no attempts to understand the psychological processes that bring these structures about. In an attempt to clarify this theoretical framework, and why it is critical for understanding the variability – and flexibility – that characterizes tribal societies, I offer the following ethnohistoric account of Pawnee social organization. In particular, the example demonstrates how latent social structures can influence social patterns,

Although integration occurs at these varying levels simultaneously, over the long-term particular integrative levels tend to assume more focal roles than others. For example, the members of some societies live in nuclear family units that are integrated into small dispersed hamlets. In other societies, people live in extended-family units that are organized into large nucleated villages. While there certainly are other intermediate levels at which integration occurs within both of these situations, in each case particular levels assume more critical – or focal – roles. Fortunately, these focal levels of integration frequently have material correlates, such as house size and settlement organization, which can be traced archaeologically. 19

Chapter Two

given changing socioenvironmental conditions.

South Bands, the principal chief of the Grand band was recognized as the primary chief of the entire sub-tribe, although it is unclear what, if any, special prerogatives he was able to exercise” (OʼShea 1989:65). OʼShea argued that the subtribes, “existed as independent, ethnic divisions within the larger tribe which tended to channel and structure interaction between bands” (OʼShea 1989:65).

In a recent overview of the archaeology and ethnohistory of the Plains Pawnee, OʼShea (1989) delineated how the social organization of these horticultural earthlodge villagers changed in response to their interaction with Europeans in the nineteenth century. OʼShea described Pawnee social structure in the following manner:

This brief description of Pawnee social organization suggests a social structure consisting of five levels of integration. This five-tiered social structure can be outlined as follows:

The Pawnee tribe should more precisely be considered a loosely bound confederacy of largely autonomous social and political units… The basic political unit of this organization was the village (whereas the household was the primary productive unit). The villages were, in turn, affiliated with one of the four Pawnee bands. The bands were grouped into sub-tribes, with the Grand, Tappage and Republican bands collectively termed the South Bands, while the Skidi remained distinct. Finally, all of the bands were linked together as the Pawnee tribe (OʼShea 1989: 61).

1. The earthlodge or household level; 2. The autonomous village level; 3. The band level; 4. The sub-tribe level; 5. The tribe level;

As early as the Eighteenth century the four Pawnee bands were each associated with particular territories:

Interestingly, this description of Pawnee social organization corresponds almost perfectly to Sahlinsʼ ʻsectoral modelʼ of tribal interaction, discussed earlier, wherein “relations become increasingly broad and dilute as one moves out from the familial navel” (Sahlins 1968:16). In other words, as Sahlinʼs (1968:16) noted ʻ[T]he system becomes weaker where it is greaterʼ – the degree of integration decreases as the level of organization increases, and degrees of sociability diminish as fields of social relation broaden.

The four principal Pawnee bands located their permanent villages in distinct geographical regions. The Grand were south of the Platte, and the Republican along the Republican river in Southern Nebraska and northern Kansas. The Skidi branch of the Pawnee maintained their settlements on the Loup river, north of the Platte… The locations of inferred Tappage settlements suggest the Big Blue River area might represent a similar home territory (OʼShea 1989:54).

Each member of the Pawnee tribe recognized each of the five levels of integration outlined above and identified their position within the society in terms of this social structure – which was encoded in systems of kinship, cosmology, and subsistence strategies. Nevertheless, the vast majority of daily events occurred within the realm of the autonomous village or household levels. These form the basic social segments upon which the remainder of the social structure was based. Within these basic segments, there existed a certain degree of active flexibility throughout the daily lives of its members – to garden or to hunt?; to raid or to feast? This is the kind of organizational flexibility Gearing (1958) models with his ʻstructural posesʼ. Beyond the village level, however, the social structure of Pawnee society served as a template for integrating and interacting at much less frequent intervals, encoding a latent scalar flexibility that could be actualized if necessary.

OʼShea suggested that the bands functioned primarily to facilitate mobilization of manpower for collective bison hunts, and for mutual defense; in addition to structuring exchange between villages during ʻbad yearsʼ. Beyond the band level, OʼShea delineated two higher integrative units: The Pawnee bands and villages were further integrated by two higher order structures that can be termed the subtribe and the tribe. The sub-tribe is the earliest division recorded for the Pawnee, distinguishing the South Bands (a.k.a. “Pani”) [i.e., Grand, Republican, Tappage] from the Skidi (a.k.a. “Panimaha”). The frequent representation of this distinction in Eighteenth century maps and reports attests to the significance of the division. That the terms themselves are Siouan in origin suggests a recognition of the division by other tribes as well as by the Europeans (OʼShea 1989:65).

In contrast to the household and village levels of integration, which were actualized on a daily basis, integration occurred much less frequently at the band level – probably only a few times a year. According to OʼShea, the bands each served to delineate hunting territories, and to structure exchange between villages during periods of strife. The only time that the basic segments – the households and villages – were integrated at the band level was during communal bison hunts, and for mutual defense. Nevertheless, the band also provided a struc-

The distinction between the sub-tribes was expressed by differences in ritual organization, dialect, and origin and migration myths. In addition, the sub-tribal groupings seem to have served some organizational functions: “[A]mong the 20

Tribal Social Theory

tural template when, during the Eighteenth and Nineteenth centuries, the 20-30 original villages coalesced into four band villages. Thus, while the Pawnee recognized themselves as belonging to a particular band, and throughout any given year would actualize this structural relationship only occasionally (e.g., during communal bison hunts), the bands ultimately provided the template for organizing life when forced into nucleation. As such, the band level of integration provided a latent structural feature that could be actualized if necessary. This pattern repeats itself at the sub-tribe level. According to OʼShea (1989:65), “When the bands began to coalesce, the villages of the South Bands would frequently combine, while the Skidi, although often located nearby, remained aloof.”

factors brought about the actualization of various levels of integration that previously were recognized – but not expressed – by the Pawnee, they did not bring about the levels themselves, as Fried (1968) would have us believe. The Pawnee recognized these higher structural levels, even if they seldom actualized them prior to contact. When it became necessary to coalesce into nucleated villages during the Eighteenth and Nineteenth centuries, the band level of integration, which had previously been actualized only sporadically throughout the year, provided the template for re-organization. When this coalescence into band villages occurred, the sub-tribal units – which previously served little purpose – influenced the location of the respective band villages. And it was not until the forced merger of the Pawnee into the combined village at Genoa that the tribal unit became actualized as a unified entity. Ultimately, these latent levels of integration provided the organizational flexibility necessary to deal with the changing social context the Pawnee found themselves forced into.

The tribal level was the least actualized integrative unit in Pawnee society. OʼShea contended that the tribal level only became significant during the Nineteenth century: South Band traditions depict the ultimate confederation of the Skidi and the South Bands as occurring only after a pitched battle in which the Skidi were forced to sue for peace. The increased importance of the Pawnee as a tribal whole resulted both from the effects of warfare and depopulation, and from the perception and treatment of the Pawnee as a single entity by the U.S. Government (OʼShea 1989:65-6).

It is important to note that throughout the dramatic social changes experienced by the Pawnee during the Eighteenth and Nineteenth centuries, their social structure did not change. Rather, different levels of integration within this social structure assumed an actualized, focal role at different times. Prior to the coalescence of the 20-30 villages into the four band villages, this focal level was assumed by each of the autonomous villages. During this coalescence, it was assumed by the band level. Finally, in the combined village at Genoa, the tribe (as a whole) assumed a focal role. As the focal role shifted between these various integrative levels, the changes were, in turn, reflected in the different settlement patterns.

Here again, the perceived social structure of Pawnee society ultimately provided the template upon which life became organized during the Nineteenth century. Despite the fact that this tribal level of integration was seldom actualized prior to this time, the social structure itself provided a sort of latent flexibility which could be accessed if necessary.

The Pawnee example is exceptional, for it provides an insightful glimpse into the long-term organizational changes of a tribal society that ethnographers are seldom at liberty to witness. This can be attributed to the excellent historical documentation of Pawnee society, as well as to the long history of archaeological research in the region, which combined allow the sort of in-depth analysis discussed above. But certainly while the example is exceptional, the processes it documents are not.

Despite the eventual forced merger of the entire Pawnee tribe into a single village around 1859, each of the structural units maintained a certain degree of autonomy: The layout of the Genoa village indicates the tenacity with which the bands continued to maintain their autonomy. Although all four bands lived within a single village, the lodges of each band were clustered in their own area, and on the hills surrounding the village each band maintained its own separate cemetery area (OʼShea 1989:66, and figure 5).

Similar patterns of changing integration can be documented in other ethnohistoric contexts (for example, among the Iroquois [Snow 1994]), and in other prehistoric contexts, such as in Neolithic Britain (e.g., Bradley 1996), in Neolithic Anatolia (see Yoffee et al. 1999:267), in the American Southwest (e.g., Neitzel 1999) and apparently in prehistoric Hungary (see Chapters Seven and Nine). Nevertheless, our understanding of these processes within tribal societies remains clouded by the absence of a theoretical and methodological framework for dealing with them archaeologically and ethnographically. I believe the theoretical framework I have outlined above, although not perfect, is a step in the right direction. By focusing upon the structural features which characterize tribal societies, it is possible to understand how their underlying social structures change over time. The various trajectories that emerge

Not only did the bands remain spatially discrete within the combined village at Genoa, but each of the sub-tribal units (the Skidi and the South Bands) exhibited their autonomy by occupying different ends of the village and constructing ramparts between them. This brief sketch of Pawnee social organization exemplifies how perceived social structures affect actualized social patterns, given different socioenvironmental conditions. In the Pawnee case, these socioenvironmental conditions can be attributed directly to contact with the US government, and with the decimation of previous modes of subsistence. While these 21

Chapter Two

from such an analysis can then be compared cross-culturally, in an attempt to discover even more general patterns. Summary This chapter has attempted to outline a theoretical framework for understanding variability and change in tribal societies. In this context, tribes are understood as segmentary, decentralized, societies whose members concurrently integrate into several different social units. This structural perspective has specific implications regarding the methodological approach that I believe needs to be adopted for modeling tribal social variability. In particular, it is necessary to focus upon two dimensions of analysis that will allow tribal structures to be measured and tracked over time – viz. interaction and integration. The following chapter outlines the archaeological implications of this approach.

22

Chapter Three: Middle Range Theory

This chapter builds upon the theoretical framework proposed in the previous chapter and presents the middle-range theoretical principles necessary for modeling long-term changes in tribal social organization. Although I am concerned primarily with constructing a methodology that can be applied to archaeological data in prehistoric contexts, the basic features of this methodology are equally applicable within ethnographic and ethnohistoric contexts – despite its materialist emphasis.

different tribal trajectories, it will be possible to gain an understanding of the variability – and flexibility – that characterizes tribal social organization. Integration and Interaction as Concepts Tribal social organization can be rationalized along two discrete but intimately interwoven dimensions:

I restrict my comments in this chapter to those aspects of the methodology that relate specifically to the theoretical framework I have presented in Chapter Two, rather than to the specific archaeological example that forms the central focus of this dissertation (see Chapter Five). As a result, the tone of the present chapter leans heavily towards the abstract, for it concentrates specifically upon translating general theoretical principles into specific measurable material variables.

1. The size, scale, and organization of the basic social segments (the households, hamlets and villages), and; 2. The way in which the members of these social segments interact with each other and integrate into actualized groups (bands, sub-tribes, etc.), either via kinship, cosmology, exchange relationships, or other panresidential institutions (i.e., sodalities, [see Service 1971, Fowles and Parkinson 1999]).

The theoretical framework I have presented – which approaches tribes in terms of structural characteristics – demands that tribal societies be analyzed along two dimensions that each relate to particular material variables that can be used to measure and track structural changes over time. The first dimension of analysis assesses the various manners in which the units of a society are integrated at varying levels. By measuring varying levels of integration using material variables, it is possible to model the size and scale of various social segments. The second dimension of analysis concentrates upon how the units of a society interact with each other differentially, and therefore can be used to model the structural relationships between different social segments.

These two dimensions delineate the lines along which social organization is constructed – and reconstructed – in tribal societies. The absence of an institutionalized central authority, combined with the fluid nature of this structural format, offers the members of tribal societies the flexibility to rework their social organization throughout a given year (active flexibility), and over much longer periods of time (latent flexibility). Any attempt to model such a flexible system must therefore account for such dynamism. Since it is the two-dimensional quality of tribal societies that make them so flexible and difficult to model, a methodology for understanding tribal social organization over the long-term must incorporate analyses directed along these same two dimensions: integration (of segments) and interaction (between segments). Both of these terms have a long history within the discipline, and they have gradually acquired nuances which have clouded their respective meanings. Although there is a general tendency to treat them as different social processes, we are seldom provided with precise definitions of the terms, which frequently are used interchangeably throughout the literature.

These two dimensions of analysis – interaction and integration – are by no means mutually exclusive, for one certainly relates to the other. Nevertheless, by treating each as an independent line of evidence, it is possible to generate models that more accurately represent the size, scale, and organization of particular structural elements. These models can then be used to track diachronic changes within different cultural contexts, and eventually to compare the various trajectories followed by different tribal societies over time. By gradually building up a comparative database that allows the cross-cultural analysis of

23

Chapter Three

There is some general consensus that integration refers to a group level phenomenon – individuals are integrated into groups. Interaction, on the other hand, usually refers to a more general process that occurs at both the group and individual level. So as not to muddy the waters further, it is necessary to clarify what I mean by the terms, and to outline why I think they are important as distinct methodological tools for modeling tribal social organization.

more importantly, the various levels that are detectable in a single society can change over time, providing an avenue for tracking the trajectories of changing levels of integration. For these reasons, the concept of integration is treated in the context of this dissertation not as a unified trait that characterizes tribal societies, but rather as an analytical dimension that can be modeled by measuring particular material variables (see below).

Traditional Approaches to Integration

Traditionally, two sets of data have been employed as evidence of integrative units in prehistoric societies: architectural patterning and mortuary analyses.

Most archaeological attempts at dealing with integration in tribal societies have employed the concept as Service used it – as a classificatory criterion which refers to the presence of pan-residential social institutions that distinguish tribes from bands – rather than as Steward intended it – as a methodological tool useful for facilitating cross-cultural analysis. Tribes are therefore commonly defined as ʻregionally-integrated social networksʼ (cf. Braun and Plog 1982). While this use of the concept – as a criterion for sorting out tribes from less complex social forms (i.e., bands) – is acceptable in the context of Braun and Plogʼs argument, to understand the term to refer only to an attribute that is exhibited by tribal systems is simply not sufficient.

Architecture as Integration The evidence most commonly cited for social integration in prehistoric contexts is architecture (e.g., Steward 1937; Wilk and McC. Netting 1984; Adler 1989, 1999; Lipe and Hegmon 1989; Bradley 1996; Coupland and Banning 1996; Labelle 1996). This is based upon the common sense notion that the size, shape, and organization of structures and settlements reveals a great deal about the relationships between the people who built and lived within them (see also Bordieu 1973, 1977). In the American Southwest, for example, the study of social integration through architecture can be traced back over a hundred years (see Lipe and Hegmon 1989). At the core of this methodology is the basic idea that:

Although tribal systems are integrated, they are not just integrated. Rather the members of such societies integrate at a variety of different levels, and into a plethora of different social units. Combined with the absence of an institutionalized central authority, this allows the focal unit of integration to shift from one unit to another, given different socio-environmental situations (see Chapter Two). This poses a particularly difficult problem for archaeologists who are concerned with modeling the various levels of integration over the long-term. Braun and Plog (1982) cumbersomely refer to this as the ʻnondecomposabilityʼ of tribal systems.

Many facets of architecture can be tied to social integration. Architectural structures reinforce social and cultural structures both spatially (by regulating movement and bounding social groups) and symbolically (by marking distinctions and by providing concrete associations with elements of history, myth or world view) (Hegmon 1989: 10).

The idea of non-decomposability suggests that the dynamics of only some hierarchical levels of a system may be observable over a given amount of time (Simon 1973; see also Braun and Plog 1982; Saitta 1983; Plog and Braun 1984). In the theoretical framework I have presented, this idea is reflected in the notion of latent flexibility – the idea that the members of societies may recognize particular structural principles that are seldom, if ever, actualized. The existence of these latent structural principles offers the members of the society a form of legitimized flexibility that can be accessed given the appropriate socioenvironmental conditions.

Although the primary focus in the American Southwest has been upon the role of integrative rituals as they relate to particular facilities, such as kivas and plazas (e.g., Hegmon 1989, Lipe 1989; Adler 1989); similar arguments have been made regarding the role of henges in Neolithic Europe (e.g., Bradley 1996), and earthworks in the Late Archaic of the American Southeast (Anderson 1999). Marcus and Flannery (1996) made a similar argument for Tierras Largas phase buildings at San José Mogote and for structures at Paso de la Amada, which they believe may have functioned as ʻmenʼs housesʼ or other such integrative facilities (see also Blake 1991).

Despite the archaeologistʼs inability to observe all of the hierarchical, or structural, levels within tribal societies, different levels nevertheless emerge as being relatively more important than others. While these archaeologically visible levels may not reflect the entire cognized structure recognized by different members of the society at hand, the very fact that particular levels are ʻactualizedʼ in some societies and are ʻlatentʼ in others is itself a very meaningful observation. Even

While research into architecture, as it relates to integrative rituals, has been an effective methodology for discussing higher-order structural integration, such a methodology can also be employd for understanding the organization of lowerorder social segments. As Ferguson (1989:173) noted: To fully investigate how architecture integrates society, we need to analyze both buildings and their open space struc24

Middle Range Theory

ture…At the same time, we need to pay attention to how the elemental groups of society are spatially segregated, as well as to the spatial and social mechanisms that integrated these groups.

While these two classes of information – architectural and mortuary patterns – are perhaps the most commonly followed archaeological paths to reconstructing integration in prehistoric tribal contexts, they do not exhaust all of the possible methods that could be employed. Nevertheless, they provide convincing evidence for the functioning of various integrative levels in the past.

Thus, while the presence of integrative facilities, such as kivas, plazas, and ʻmenʼs housesʼ can be employed as evidence for higher-order levels of social integration, so too can the size and organization of domestic structures (e.g., Wilk and McC. Netting 1984; Coupland and Banning 1996) and settlements themselves (e.g., Flannery 1972) be used as evidence for how the basic segments, or elemental groups, were themselves organized.

Traditional Approaches to Interaction The concept of interaction has a similarly long, but somewhat more jaded, history within the discipline. Unlike the concept of integration, which despite its various meanings has been applied relatively consistently to the organization of sociopolitical, decision-making, groups, the concept of interaction has been applied to nearly every unit of social analysis – from individuals to whole societies. With particular reference to tribal societies, interaction generally evokes a process that occurs between individuals; for example, in Sahlinsʼ (1968) ʻsectoral modelʼ of tribal society, wherein various sectors of social interaction reflect different levels of integration. In archaeological contexts, two methodological tools have dominated reconstructins of prehistoric social interaction – the analysis of stylistic elements, and the analysis of exchange spheres.

Mortuary Integration The second class of data commonly cited by archaeologists as evidence of social integration comes from mortuary studies (see, for example, Goldstein 1980). The location and place people choose to bury their dead is frequently viewed as reflecting a unit into which they were perceived having been integrated during life. As OʼShea (1984:21) noted: The theoretical basis for relating mortuary patterns to aspects of societal organization is by now well established in archaeology, and rests on the assertion that there is a systematic relationship between an individualʼs station in life and the treatment received in death (cf. Binford 1971; Saxe 1970).

Stylistic Analyses Early attempts at assessing interaction in prehistoric contexts using archaeological data emphasized the role of stylistic variables, and began a tumultuous debate about the various functions, and applications, of style in archaeology. In a recent discussion of the topic, Michelle Hegmon (1995:8) noted:

Thus, communal burial grounds, such as the Late Woodland Juntunen site in northern Michigan, are often considered as indicative of some sort of structural mechanism that integrates dispersed groups (see OʼShea 1984). Bradley (1996) has discussed a similar phenomenon with regards to barrows in Neolithic Britain, and Goldstein (1980) has demonstrated a similar relationship between formal disposal areas and corporate groups in the Lower Illinois Valley during the Mississippian period. But such arguments are not restricted to the location of burial grounds with reference to settlement sites, they also apply to spatial differentiation within cemeteries and ossuaries. As such, mortuary studies can be used to identify patterns of integration at a variety of different social levels, from the lower-order integration of individual social segments within a single settlement to the higher-order integration of several clusters of communities within a region (e.g., Goldstein 1980; OʼShea 1996).

Many archaeologists and other anthropologists have recently put forth various definitions of style, have identified several kinds of style, and have developed at least as many techniques of stylistic analysis… It sometimes seems as though we have almost as many approaches to style as we have works on the topic. Indeed, the analysis of the relationship between style and its various cultural roles has generated a daunting amount of literature over the years, and for good reason. Style is an extremely powerful construct that can help to bridge the gap between archaeological material and the people who created it (see Hegmon 1995:7). As such, the analysis and interpretation of style in material culture forms one of the most essential tools available to the archaeologist for reconstructing patterns of social organization and interaction in prehistoric contexts.

In addition to the spatial placement of particular burials, burial clusters, and cemeteries or ossuaries on the tribal landscape, frequently the grave goods associated with particular individuals can provide a great deal of information regarding the various social institutions in which they participated in life. For example, OʼShea (1996) argues for the existence of sodalities based upon differential grave goods included in female burials in the Early Bronze Age cemetery at Mokrin, in northern Yugoslavia.

Beginning with Deetzʼs (1965) groundbreaking research on the relationship of stylistic variables in Arikara ceramics to rules of residence and descent, a flurry of articles appeared which built upon, modified, and reapplied this initial model within different cultural contexts (e.g., Longacre 1964; Leone 1968; Hill 1968; Whallon 1968). In his analysis of stylistic 25

Chapter Three

variables to reconstruct prehistoric Iroquoian social organization, Whallon outlined the basic theoretical assumptions that characterize much of this early research:

Voss (1980) argued that these two seemingly exclusive theoretical approaches should be merged into a more comprehensive theoretical approach that understands stylistic behavior as a function of personal identity. Based largely upon Friedrichʼs (1970) ethnographic work with Tarascan pottery painters, wherein she argued that only certain design features were indicative of interaction and others were not, Voss suggested that:

The hypothesis that measurable relationships exist between style and social organization is founded upon two basic assumptions. The first of these is that style has many aspects and levels of behavior which may be analytically distinguished and measured… The second assumption basic to this hypothesis is that the nature of the diffusion of stylistic ideas and practices, both within and between communities, will be determined by the nature of interaction among artisans. The aspect of style concerned, the rate of diffusion, and the directions and limits of diffusion will be conditioned by the kind, frequency, and channeling of interaction among the producers of the stylistic frequency (Whallon 1968: 223).

Personal identity is a dynamic function, formed through social interaction and involved in the projection of beliefs concerning the self toward significant individuals in the social environment. As an identity function, therefore, stylistic behavior is molded through social interaction and should therefore be reflective of such interaction, but is also behavior directed toward others in the social realm. Such behavior transmits information concerning, among other things, social affiliation and the expectations of those within the social realm (Voss 1980: 274).

The basic theoretical premise that underlied these early attempts at reconstructing social interaction from stylistic elements began to fall under attack during the 1970ʼs and 80ʼs (e.g., Plog 1976; 1978; 1980).

This led Voss to argue that stylistic variability should incorporate at least one dimension related to group affiliation and one related to the specifics of social group interaction. In this framework, certain design features are expected to be more indicative of social groups and group affiliations, and should therefore be distributed homogeneously within regions and sites. Other attributes should be more reflective of patterns of interaction, and shoud therefore be distributed clinally throughout regions (see Voss 1980:275).

In their discussion of the use of style for reconstructing prehistoric social interaction, Braun and Plog (1982; see also F. Plog 1977) delinated two theoretical approaches that they suggest dominate such inquiries. The first of these they called the cognitive, or social interaction theory of stylistic variation, which rests on the basic idea that stylistic attributes tend to be shared among groups that interact more intensively (Braun and Plog 1982:509). Within this theoretical framework, which characterizes most of the early works cited above, it is assumed that more intensive interaction between groups will increase the similarity between them, and will result in a decreased homogeneity within any one group. Braun and Plog argued that this approach assumes that style is assumed to be imprinted and socially passive, and that it does not hold up to ethnographic testing.

The debate over the role of style in archaeology crystallized with the exchange between Sackett and Weissner. In that exchange, Sackett (1977, 1982, 1985) proposed that style should be viewed as isochrestic in nature. Within Sackettʼs scheme, stylistic variability was understood as resulting from passive enculturation: An artifact can be regarded from two contrasting, but fully complementary points of view. In the first it is perceived in action, as a thing that was manufactured and in turn used in a succession of activities that made up daily life in a given cultural setitng. Our interest in this case concerns the ends it served… When considering an artifact from this point of view we are thinkng of its function…Just as any artifact has an active voice which connotes function, so it has a passive voice which connotes style. In this latter case we are viewing it not as an actor in a variety of roles but instead as a signpost or banner advertising the arena in which the roles are being performed (Sackett 1977: 370; original emphasis).

Braun and Plog called the second theoretical approach to stylistic analysis information exchange theory. This approach accepts as its basic assumption that: within nonindustrial societies, differences in stylistic behavior result more from social constraints on the choosing of alternative decorative options during the act of decoration, than on the social context in which a person learned his or her decorative repertoire (Braun and Plog 1982: 510). Examples of this school of stylistic analyses include the works of Wobst (1977), Braun (1977) and Conkey (1980). Braun and Plog suggested that information exchange theory is more appropriate for the analysis of interaction in tribal social systems, since this perspective tends to gain more credence from ethnographic case studies (e.g., Hodder 1977, 1979; Wobst 1977).

According to Sackett, the passivity of style is rooted in the notion that the choices made by artisans are socially transmitted, therefore “the degree of similarity among the choices that are made in two historically related loci depends upon the intensity of social interaction shared by their occupants” (Sackett 1977:371).

26

Middle Range Theory

Weissner (1984, 1985), on the other hand, suggested that style should be viewed as an active phenomonon that communicates social and personal indentities in order to define social relations. In her view, style is viewed as a kind of iconography that is purposely used for signalling ʻemblemicʼ or ʻassertiveʼ messages of group membership or of individual expression, respectively. In a more recent discussion, Wiessner (1990) observed that the different approaches to style are based upon different sets of criteria:

trariness in technological traits that generates variability in material culture patterning (Stark 1998:6). The dichotomy that is traditionally drawn between active and passive conceptualizations of style is rephrased in terms of a more general distinction between structure and agency (cf. Dietler and Herbich 1998:245 for discussion), and Bordieuʼs (1977) theory of practice is frequently cited as the appropriate theoretical perspective for uniting these two seemingly opposed viewpoints.

Sackettʼs classification is based primarily on whether style plays an active or passive role in communication…My categories of emblemic and assertive style are based on quite different criteria, with by far the most important of these being specificity of referent…Styles with distinct referents are those that I call emblemic and those with more vague associations I call assertive (Wiessner 1990: 108).

For example, in their ethnoarchaeological analysis of the production, distribution, and consumption of ceramics in western Kenya, Dietler and Herbich (1998) argued that: Material style can serve as a useful concept for archaeologists attempting to investigate the social role and meaning of material culture only if it is seen as the objectified result of techniques (rather than as straightforward objectified information); and more specifically it must be seen as the result of characteristic ranges of responses to interlinked technical, formal, and decorative choices made at all stages of a chaîne opératoire of production (Dietler and Herbich 1998:246).

This lead Wiessner to suggest that it “…may be best to move away from classification of style and see style from a unified perspective as a means of non-verbal communication to negotiate identity” (Wiessner 1990:108). More recent discussions of stylistic analysis (e.g., Carr and Neitzel 1995; see also Conkey and Hastorf 1990) have attempted to move towards a unified theory of stylistic analysis. Unfortunately, as Wiessner noted:

They contended that Bordieuʼs notion of habitus –the idea that people are disposed to act in certain ways due to the influence of structures of material conditions – “…offers a means of situating both material culture and the chaînes opératoires and social actors responsible for its production and transformation within a framework that mediates structure and agency” (Dietler and Herbich 1998:246). Since Bordieuʼs habitus “can generate patterned actions that appear regulated as if resulting from rules,” Dietler and Herbich suggested that:

Even in an ethnographic study with full information on history and context available, it can be very difficult to discover the symbolic meanings underlying style… Dreary as it is, the only solution seems to be careful descriptions and analysis … of as many different classes of artifacts as possible, the linking of these with information on economy, exchange, politics, settlement patterns, etc., and in the end, an attempt at interpretation of meaning underlying certain symbols and metaphor (Wiessner 1990:111).

…it allows the perception of how practice both reproduces and transforms structure as it adjusts to demands. Rather than seeing practice as predetermined by a static set of cultural concepts or structures (e.g., some sort of rigid mental templates), the habitus is a dynamic relational phenomenon which is both an historical product and agent (Dietler and Herbich 1998:247).

Indeed, the most effective applications of stylistic analyses in particular archaeological contexts are those that assume a more synthetic perspective, incorporating different approaches to understanding the meaning of different variables (e.g., Voss 1980; Plog 1980, 1990; Braun 1990; Hegmon 1995).

While Bordieuʼs (1977) theory of practice and Giddenʼs (1984) structuration theory both provide interesting theoretical insights into the relationship between structure and agent (see also Ortner 1984), and are therefore extremely useful concepts for bridging the theoretical gap between the actions of individual humans and the societies in which they live, the application of these more generalizing sociological principles to specific cultural examples must be conducted carefully, and with an eye towards averting the tendency to defer all knowledge to dogmatic assumption. As Hegmon (1998:269) noted:

The long-standing debate of stylistic analysis in archaeology has recently been restructured to incorporate technological studies from the French tradition (e.g., Stark 1998). Finding its roots in Leroi-Gourhanʼs (1993) concept of chaînes opératoires –embedded “operational sequences” – the French tradition of technologie has recently been turned to as a possible way out of the style vs. function dilemma in North American archaeological circles. This approach, demonstrated most elequently by authors such as Lemonnier (1986, 1989, 1992) and Lechtman (1977), is based upon the notion that:

…we must guard against simply adopting these new (stylish) words and applying them to old concepts. That is, habitus is not the same thing as culture, and agency cannot be equated with the behavior of an individual in an opti-

Behaviors and techniques are guided by human choices, and most steps in any technological process can be carried out using several alternative approaches. It is this arbi27

Chapter Three

mality model.

to discern patterns of interaction that occur within particular societies and at a more localized (geographic and social) level, the analysis of long-distance exchange systems has been used primarily to reconstruct patterns of interaction that occur between societies and over larger (geographic and social) areas (e.g., Flannery 1968; Renfrew 1975). As a result, these two lines of evidence refer to different scales of social interaction that operate at different levels within tribal societies.

Rather, Hegmon suggested, we may benefit more by assuming a more theoretically eclectic perspective that draws upon the most useful aspects of different approaches. Whereas earlier stylistic analyses of the “information exchange” and “social interaction” variety (see above) focused primarily upon the social context of distribution and consumption of material items, authors writing in the French tradition of technologie tend to stress the social context of production with as much, if not more, vigor (e.g., Dietler and Herbich 1998; Maceachern 1998; Gosselain 1998). To some extent, this shift in analytical focus is reminiscent of the distinction Sackett (e.g., 1982, 1990) drew between “isochrestic choices”, which he understood to be particular choices made by individuals based upon learning, habit, and enculturation, and “style;” which referred to a subset of those choices that indicated group membership (see also Hegmon 1998:267). The main difference between the technologie and isochrestic approaches is rooted in the perceived role of the individual within the social context. Whereas Sackett envisions the individual as passively acting within a particular social environment, the technologie proponents, based largely upon the theoretical works of Bordieu (1977, 1980) and Giddens (1984), argue that individuals and the operational sequences they conduct to produce technological products are an active element within that social environment. That is, they are part of a societyʼs habitus.

Since the analysis of long-distance exchange by definition entails the identification of ʻforeignʼ materials within particular archaeological contexts (cf. Earle and Ericson 1977), and since these materials (e.g., obsidian, amber, tin) can frequently be traced through various chemical and macroscopic methods to their points of origin, such analyses are particularly wellsuited to the materialist-based methods of archaeology. As such, they provide a way of tracing general patterns of interaction between distant regions, allowing the archaeologist to reconstruct, on the one hand, “interaction spheres” which generally refer to large geographic regions that incorporate networks of related autonomous societies (see Caldwell 1964; Struever and Houart 1972). On the other hand, the analysis of long-distance trade often allows for the delineation of more infrequent patterns of face-to-face and down-the-line exchange that occur over even larger areas and between largely unrelated societies (e.g., Flannery 1968b). These may be called “exchange spheres”. In a recent overview of interaction spheres in Mesopotamia, Yoffee (1993) traced the concept to Caldwellʼs (1964) description of the Hopewell:

Although there is certainly much to be gained from this recent focus upon the social context of technological choices, the most fruitful research will most likely incorporate aspects of style and technology into its methodology. As Hegmon (Hegmon 1998:278) noted:

He invented the term interaction sphere to denote that there were social, ideological, and trade connections among populations that shared thereby a restricted corpus of material culture – in the Hopewell case of pipes, figurines, copper axes – namely, those objects associated with the interment of the honored dead (Yoffee 1993:257).

We must guard against assuming that technology (whether technological style or technological choices) is necessarily a better avenue of exploration than style or other aspects of material culture. Researchers focusing on style have (mostly) stopped arguing about which perspective on style is best and have usefully combined different perspectives. I hope new arguments that pit style against technology do not take their place. Whether style or technology is the best subject will depend on the research question and on the particular study. A combined approach will probably be preferable in many cases.

The nature of interaction within such a sphere is such that, “certain material features found over a large area reflect a set of cultural relations that transcend localized nests of institutions and distinct peoples embedded within it” (Yoffee 1993: 258). The existence of interaction spheres is normally recognized by the presence of diverse classes of material culture and generally assumes a relatively high degree of interaction, whereas the existence of exchange spheres is based upon the presence of only one class of material and generally assumes less intensive interaction. Consider the Hopewellian example, wherein a series of different artifact types (e.g., pipes, figurines, ceramics) are cited as evidence of an interaction sphere (Caldwell 1964), that incorporates a number of autonomous societies that it is assumed share some ideological principles, thereby subsuming a group of societies that interacted rather intensively, despite remaining autonomous. At the same time,

Despite the various problems associated with particular theoretical and methodological approaches, stylistic and technological analyses have provided effective means of assessing interaction, primarily through the analysis of ceramic variables (e.g., Deetz 1965; Whallon 1968; Braun 1977; Voss 1980; Plog 1990; Hegmon 1995). Long-Distance Trade and Exchange Whereas the analysis of stylistic variables tends to be used 28

Middle Range Theory

obsidian from Yellowstone and shells from the Gulf Coast are regularly found in Hopewell sites, indicating a larger exchange sphere within which a different sort of interaction occurred. This latter interaction is assumed to be much less frequent and most likely occurs via down-the-line exchange between largely unrelated societies.

principles. Integrative Units By defining the size, scale, and organization of integrative units in this empirical manner, it is possible to construct a structural model that essentially builds itself from the bottomup. Each unit is understood to represent a particular cognized social units (e.g., families, clans, sub-tribes), which may or may not be inferentially identifiable through their material correlates. In general, five units constitute the basic levels that can be used to characterize integration within tribal societies:

Despite the related nature of these two concepts, interaction spheres and exchange spheres provide a way to measure regional and inter-regional interaction, respectively. When combined with stylistic and technological analyses, which can further clarify the nature of interaction within a region, it is possible to generate a model that approximates how the various integrative units (as reflected in architecture and mortuary treatment) actualized into specific structural configurations over time.

1. The immediate coresidential unit. This unit frequently, but by no means exclusively, corresponds to the nuclear or extended family group or lineage. The archaeological correlates of this unit include the size, shape, and organization of discrete domestic architectural features (e.g., houses, pithouses, longhouses). This unit normally subsumes the basic unit of production and pooling of resources.

Reconstructing Levels of Integration The first step necessary for modeling social organization in prehistoric contexts is to determine the size, scale, and organization of the basic social segments – the constituent elements of the society in question. In archaeological contexts this is possible initially by assessing the size of coresidential units that are indicative of different levels of integration. The size and scale of even the most basic segmental units (i.e., households and villages) vary considerably in tribal societies, and can be related to different cross-cultural patterns of postmarital residence (see Ember and Ember 1995) and economic organization (Sahlins 1972; Flannery 1972). Hence, by measuring the size and organization of domestic structures and how they are organized into settlements it becomes possible to reconstruct these basic integrative levels. These basic integrative units, which may themselves vary from one structural pose to another over the course of any given year, can be used to model the active flexibility that characterizes the structure of the society.

2. The intermediate coresidential unit. This unit can correspond to any number of social groups; such as the extended family, the lineage, or the band. The archaeological correlates of this unit include the size, shape, and organization of discrete groups of domestic architectural features within settlements (e.g., room-blocks, house-clusters, longhouse clusters). Some pooling of resources may occur within this unit. 3. The settlement. Depending upon its size, this unit can correspond to any number of social groups; including the extended family, lineage, band, sub-tribe, or tribe. The archaeological correlates of this unit include discrete settlements (e.g., villages, hamlets, or isolated houses). Although some pooling of resources may occur throughout an entire settlement, to a large extent the frequency to which this occurs will depend upon the organization of the settlement itself (cf. Flannery 1968; Wiessner 1982).

Beyond the settlement, it is frequently possible to identify settlement clusters throughout the landscape that correspond to even larger integrative units. Not only can the spatial relationship between settlements be used as a method of discerning integrative units that operated over the long-term, but frequently the relationship between settlement sites and other special purpose sites, such as cemeteries or shrines, can also effectively be used to define the location of different boundaries in the social landscape (see Stark 1998a). The patterns that emerge from any such spatial analyses can then be compared to patterns of interaction to determine whether they represent the presence (or absence) of specific social boundaries over the long-term. By assessing the degree to which such boundaries were actively maintained or passively transcended, as indicated by patterns of interaction across them, it is possible to generate some conclusions regarding the nature of the relationship between the groups at hand. Such combined analyses may therefore reveal the operation of more latent structural

4. The settlement cluster. The settlement cluster can correspond to any number of social groups, depending upon the number, and organization of, lower-order units integrated within it. The archaeological correlates of this unit include settlement clusters, or groups of settlements organized around, or separated by, special purpose sites, such as cemeteries or shrines. Some pooling of resources may rarely occur at this level, but is likely to take the form of communal feasts, or dances. Clusters may be either diffuse or discrete. 5. The supercluster. The supercluster can correspond to any number of social groups, depending upon the number of lower-order units integrated within it. The archaeological correlates of this unit include groups of settlement clusters. 29

Chapter Three

Not all tribal societies are expected to exhibit all of these levels of integration. But by delineating which levels are exhibited in different societies over varying periods of time, and by determining the size, scale and organization of each of those levels, it is possible to track how the various levels change over time.

the Classic Horizon to much smaller houses (average = 62 m2) during the Historic Horizon. This pattern leads Hollinger to infer a corresponding shift in preferred post-marital residential patterns, which he attributes to the various effects of European contact. I argue for a similar change in house form from the Late Neolithic to the Early Copper Age in eastern Hungary (see Chapter Seven).

By focusing initially upon the organization of coresidential integrative units, it is possible then to use the organization of these units as a baseline for assessing the possible occurrance of other integrative institutions – such as sodalities –that, by definition, are not based upon coresidentiality.

Changes in the Intermediate Coresidential Unit The intermediate coresidential unit refers to any structural sub-division that occurs between the immediate coresidential unit (the house or longhouse) and the settlement (the village or hamlet). Depending upon the cultural context, such units may or may not be exhibited. Or, in contrast, several such units may exist at different levels, reflecting various levels of internal structure within a particular settlement (village). Intermediate units may reflect a variety of different social levels (e.g., lineages, bands, sub-tribes), and they frequently form the basis of some unit of pooling or decision-making within a larger social framework. As such, changes that occur within the size and organization of intermediate coresidential units are likely to reflect changes in the structure of intra-village pooling and decision-making. In addition, such intermediate units are likely to serve as a basis for the mobilization of manpower for conducting communal activities within the settlement, such as building earthlodges or longhouses. These intermediate units frequently form the basic units of fission and fusion. They are generally the units that leave to form independent settlements, and also the units that coalesce to form nucleated settlements (see, for example, Chagnonʼs [1983] discussion of Yanomamo fission and fusion).

Such an analysis requires the comparison of identical integrative units over time. By tracking how each of these units change with respect to each other, it is possible to infer a variety of different social mechanisms that can be held accountable. Changes in the Immediate Coresidential Unit As noted above, the immediate coresidential unit in tribal societies frequently, but not always, corresponds to some sort of nuclear or extended familial group or lineage. This unit also subsumes the vast majority of production and pooling of resources that occur on a daily basis within a society. As such, changes that occur within the scale and organization of the immediate coresidential unit may reflect changes within the structure and nature of the familial group; and the immediate pooling and production of resources within the ʻdomestic mode of productionʼ (Sahlins 1972). In addition to these, cross cultural analyses also suggest that the size of domestic structures – the material correlate of the immediate coresidential unit – may also be indicative of postmarital residence patterns. For example, Divale (1977) and Ember and Ember (1995) suggested that houses greater than ca. 60 m2 are statistically associated with matrilineal social patterns. Similarly, other research suggests that the shape of domestic structures (e.g., round versus square; cf. Flannery 1968) may reflect something about the structural relationship between immediate coresidential units, since square dwellings lend themselves more readily to agglomeration and accretion than do round dwellings – implying that the tendency to build square or rectangular houses may be indicative of a more fluid relationship between households.

In the Pawnee example, intermediate coresidential units do not emerge until the villages begin to coalesce into the four ʻband villagesʼ during the early 19th Century. Then, the extended lineages that each previously formed autonomous villages aggregated into spatially discrete earthlodge clusters in the ʻband villagesʼ. During this time, OʼShea notes a cyclical pattern of fission and fusion along the lines of such intermediate units. When the Pawnee moved into a single settlement around 1856, even more intermediate units emerge – at both the sub-tribal and band levels. In stark contrast to the stability reflected in the immediate coresidential unit, the Pawnee organized into a variety of different intermediate units, which can each be traced both backwards and forwards in time to different structural phenomena in the preceding and succeeding periods. In the absence of such excellent historic information, it will frequently be impossible in prehistoric contexts to associate these intermediate units with particular social levels (e.g., bands). Nevertheless, the very presence or absence of such levels, and the way they are organized over time, elucidates the nature of intra-village organization.

Interestingly, throughout the various trajectories of change undergone by the Pawnee during the 18th and 19th Centuries (see Chapter Two), the structure of the immediate coresidential unit remained essentially intact. The size of the earthlodges and their internal organization during the 18th Century persist throughout the 19th century – reflecting the resiliency of this basic social segment in Pawnee society during a period of otherwise unmitigated change. But such stability is not always the case. For example, in the Oneota ʻcultureʼ of the Upper Midwest, Hollinger (1995) has argued for a significant shift in house floor area – from big houses (average = 171m2) during

Changes in the Settlement The settlement – the farmstead, camp hamlet, or village – can

30

Middle Range Theory

vary considerably in size, scale and organization, and as such corresponds to a wide variety of social levels – from individual nuclear families (e.g., farmsteads) to entire societies (e.g., the Pawnee combined village). As such, changes in the size, scale, and internal organization of the settlement often correspond to the actualization of more latent structural principles. For example, the initial organization of the settlement among the Pawnee during the 18th century consisted of relatively small, essentially autonomous villages that were inhabited by extended family groups, or lineages. During the early 19th century, the 20-30 previously autonomous villages coalesced into four large villages organized at the band level. This pattern of nucleation becomes even more pronounced during the latter half of the 19th century, when the entire tribe was forcibly merged into a single village near Genoa. In this particular case, the reasons for merger were clear, and can be traced to the increased impact of the United States government on the entire region. In other cases, wherein changes in the settlement cannot be attributed directly to the impact of more complex societies, the causes frequently are external, and relate to factors such as the organization of warfare or subsistence practices. But just as frequently, changes in the settlement can relate to problems of scalar stress upon internal organizational structures (e.g., McGuire and Saitta 1996).

a single tribal cluster. This change marked the actualization of latent social relationships that the Pawnee had previously recognized only as it related to the demarcation of different hunting territories. Focal Units of Integration In addition to providing a convenient method for parsing apart the actualized social structures of prehistoric societies, the analysis of social organization along different integrative units allows the researcher to identify particular focal levels of integration within each actualized configuration. These focal levels, as I have chosen to call them, refer to the unit at which particular societal processes – such as aggression, competition, and mobilization – assume primary roles within the various structural configurations assumed by different societies. For example, if the members of a hypothetical society are organized into nuclear family groups (immediate coresidential unit) that live in small autonomous hamlets (settlement), that are dispersed evenly across the landscape (settlement cluster), the nature of aggression and competition within that society is likely to operate on a much different scale than if the societal members lived in extended family groups (immediate coresidential unit) organized into large, nucleated villages (settlement) that formed discrete, territorially-bounded clusters (settlement clusters). The same is true for the focus of social competition and the mobilization of resources. Different structural configurations result in differential roles being assumed by each of these processes.

In much the same way that the immediate and intermediate coresidential units correspond to particular pooling and productive capacities, such activities also occur at the level of the settlement, albeit in a different fashion. Whereas immediate coresidential units (or households) tend to pool their resources on a daily basis, sharing often occurs between members of settlements on a less frequent scale. Flannery (1972) has argued that different types of interactive behavior are likely to influence architectural styles – “nucleated villages with rectangular houses” vs. “circular-hut compounds” – that offer different benefits with regards to the organization and intensification of resource production.

This implies that, to some extent, the structural configuration of the society at hand lends itself to accommodating different sorts of social processes. For example, as the actualized integrative structure of a society becomes more complex – incorporating more individuals into more elaborately organized coresidential units – processes that occur between the various units, such as social competition, tend also to operate at increasingly higher levels. As such, the degree of complexity exhibited by the actualized integrative structure of a society has particular implications with regards to the degree of social, political, and economic inequality possible within that society. In general, societies that exhibit relatively simple actualized integrative structures are also less likely to exhibit significant degrees of inequality, since inequality is itself a relative concept that refers to differential access to resources, materials, or social power between the members of a particular group. If the settlement also constitutes the focal unit of pooling within a society, then it is nearly impossible for an individual to achieve some degree of economic inequality within that unit. Similarly, given more complex structural configurations, the nature of warfare is likely to operate at a higher level – including more individuals and focusing more at the inter-village level, than at the inter-family level.

Changes in the Settlement Cluster and Supercluster The settlement cluster – the village or hamlet cluster – generally forms the largest spatially-discrete integrative unit in prehistoric tribal societies. In extreme cases, settlement clusters can themselves form even larger groupings – or superclusters – across the landscape. Settlement clusters may vary dramatically in size – from clusters of a few isolated houses to clusters of large, nucleated villages – and as such can correspond to a plethora of different social levels, from extended families to entire societies, or even groups of societies. Changes within the settlement cluster generally correspond to the actualization of latent structural principles. As the Pawnee example demonstrates, during the 18th century the coresidential settlement cluster consisted of several autonomous villages that were located in geographically-discrete ʻband territoriesʼ. This pattern changed during the early 19th century, when the four band villages constituted

31

Chapter Three

Reconstructing Interaction

that occurred within a particular society or within a small group of closely related societies that inhabited a particular region.

The second dimension of analysis concentrates upon how the members of a society interact with each other differentially, and therefore can be used to model the structural relationships between different social segments. Having delineated a proposed methodology for measuring the size, scale, and organization of various coresidential integrative units, it is now necessary to demonstrate how analyses of interaction can be used to infer higher-order social groupings that occur beyond the settlement, or which themselves cross-cut residence.

2. The analysis of trade networks. In contrast to stylistic analysis, the analysis of trade networks generally is employed to model the degree of interaction that occurred between unrelated societies. Since each of these analytical techniques generally refers to a different level of societal interaction, each must be treated independently.

Defining the Temporal and Spatial Scales of Analysis

Intra-Regional Interaction

The first task necessary in any attempt to reconstruct the nature of interaction in prehistoric societies is to define the temporal and spatial scales of analysis. As noted above, the temporal scale will be determined largely by the nature of the data at hand. The chronological precision of various material sequences will determine at what temporal scale such analysis can occur. For example, in the American Southwest, where dendrochronology provides very precise dating for ceramic sequences, it is frequently possible to discuss prehistoric interaction at a similar scale of temporal precision – at the decadal or generational level. In other parts of the world, where ceramic sequences can be related only to much longer periods of time, perhaps hundreds of years, it is possible to discuss interaction only as a generalized process that occurred over a much longer temporal scale. In these cases, it is important to be aware that any patterns of interaction will refer not to specific interactive spheres, but to the operation of various interactive spheres over time. Nevertheless, such long-term patterns will be broadly comparable to patterns that were produced over a similar length of time.

Within a geographically-defined region, interaction typically occurs at a variety of levels within prehistoric tribal societies. Nevertheless, certain structural levels typically assume focal, or primary, roles between different integrative units at different points in time. This being the case, the archaeological challenge then becomes distinguishing a) which levels assumed this focal interactive role, and b) how these higherorder levels related to the various integrative units between which interaction occurred. As noted above, the analysis of stylistic variables has proven to be particularly effective for modeling intra-regional interaction in prehistoric tribal contexts. By articulating such stylistic analyses with the analysis of integrative units (see above), it becomes possible to generate models that more accurately describe the degree of interaction that occurred at different levels within a society over a given amount of time. Having defined the coresidential integrative units in the manner discussed above, it is possible to generate a set of expectations for the manner in which stylistic patterns relate to particular interactive patterns. Since it can be assumed that a high degree of interaction occurred within the settlement, it is necessary for any such stylistic analysis to examine the manner in which various settlement clusters, if any existed, interacted. To this end, such analyses must focus upon the stylistic (and technological) variability exhibited both within and between different settlement clusters (see Plog 1990; Hegmon 1995).

Since the nature of tribal society is such that social boundaries are fluidly restructured over time (Barth 1969), and since the configurations of actualized structural patterns can vary dramatically over the landscape, it is necessary to define the spatial scale of analysis in geographic, rather than social terms. In other words, the spatial scale of analysis should be determined by particular geographic and topographical features, and not by social features, such as the geographic extent of a particular material ʻcultureʼ. By defining the spatial scale in this independent manner, it is possible to generate models of interaction that can then be compared to each other to delineate trajectories of social change within a region.

Following Voss and Young (1995), and Carr (1995) it is useful to assess two different stylistic attribute characteristics – distribution and visibility – that can be used to interpret various regional patterns. These stylistic patterns can then be compared to the spatial patterns of settlement clusters to further clarify their meaning within the specific social context. In general, it can be assumed that more visible stylistic attributes are likely to reflect emblematic, or group-marking behavior. On the other hand, attributes with low visibility are more likely to reflect isochrestic variation. As such, the respective degree of visibility of the attribute in question will refer essentially to discrete social realms. Thus, in order to accurately assess the degree of interaction between various social segments

Reconstructing Interaction Between Segments I have outlined above the various manners in which archaeologists traditionally attempt to model social interaction in prehistoric contexts. Two types of analyses have dominated such attempts: 1. The analysis of stylistic variables, usually within ceramic or lithic assemblages. Stylistic analyses are generally employed to model the degree of interaction 32

Middle Range Theory

that occurred within a region it is necessary to assess stylistic variability along both types of stylistic variables.

it is possible to track the various actualized structural changes that occurred as the Pawnee were increasingly impinged upon by nearby tribes and by the US government.

By analyzing high- and low-visibility stylistic variables with reference to their spatial distribution, and by assuming that the relative visibility of an attribute relates to its stylistic role, as a means of information exchange or as an isochrestic variable respectively, it then becomes possible to generate a model of regional interaction that approximates the manner in which the various social segments interacted within a geographically-defined region (see Figure 3.6).

During the Eighteenth century, and probably for a considerable length of time prior to that, each winter the Pawnee lived in large earthlodges (Immediate Coresidential Unit) in small autonomous villages (Settlement). These lodges were situated within particular territories that corresponded to particular geographic areas that, in turn, were associated with particular bands (Settlement Cluster). Throughout the remainder of the year, when the Pawnee were mostly involved with hunting bison, they dispersed into hunting camps which were structured in an entirely different manner. Within these camps, which were much smaller than the permanent winter villages, they lived in hide tipis (Immediate Coresidential Unit). These camps (Settlement) were generally located at good hunting locations within the band territories and were moved frequently following the herds (see Roper 1990, for discussion). The combination of these two structural poses into the annual round of the Pawnee produces active flexibility within the system, and, if viewed as a cyclical process – as a social pattern that operates over roughly a century – leaves behind a particular ʻstructural signatureʼ – viz. the material correlates of each of the integrative units.

Inter-Regional Interaction In much the same way that the analysis of stylistic variability is an effective means of assessing interaction between related social segments within a region, the analysis of trade networks provides a means of assessing interaction between unrelated societies. On the one hand, ʻinteraction spheresʼ, which are traditionally based upon the coexistence of multiple types of material data within a broadly-defined region, are frequently used to infer the existence of closely-related autonomous societies. On the other hand, ʻlong-distance exchange spheresʼ, which normally incorporate only one or two material types, are frequently used to model interaction between unrelated societies in discrete geographic regions.

Although the Pawnee recognized higher-order structural relationships (such as the sub-tribe and tribe) beyond the band level, these structures were seldom – if ever –actualized during this early period. The vast majority of a Pawneeʼs life was spent working with members of his immediate family and the other members of the hunting camp, or village. During some parts of the year there was occasion to work together with the members of other nearby villages to communally hunt, or for defense. This situation dramatically changed as the Pawnee coalesced from 20-30 villages into (nucleated into) four band villages during the early Nineteenth century.

By investigating interaction along these two axes, it becomes possible to delineate the manner in which interaction occurred between various social segments. The Pawnee Revisited In order to demonstrate the utility of this method, I return to the Pawnee example I introduced at the end of the last chapter. To this end, I present three structural models that correspond to the three organizational modes assumed by Pawnee society. The first model applies to the structure of Pawnee society during the Eighteenth century, when they were organized into autonomous villages. The second to the early Nineteenth century, when they coalesced into ʻbandʼ villages. The third to the period after about 1859, when the entire tribe settled at the combined village near Genoa.

Although the immediate coresidential unit (the earthlodge) remained the same during the early Nineteenth century, the size, scale and organization of the other integrative levels of the permanent village changed dramatically. The size of the band villages (Settlement) was much larger than it had been during the early period, regularly exceeding 80 earthlodges. This pattern towards nucleation reached its apex in the latter half of the Nineteenth century, when the entire tribe was forcibly merged into a single village around 1856.

Figure 3.1 represents the cognized social structure of Pawnee society during the historic period. Figures 3.2, 3.3, and 3.4 represent the various actualized structures assumed by the Pawnee throughout the Eighteenth and Nineteenth centuries. Each figure divides the actualized social structure into two poses – the ʻpermanent village poseʼ and the ʻhunting camp poseʼ – which each represent the different structures assumed by the Pawnee during their annual round. Within each of these poses, I have attempted to represent the various social, and material correlates that relate to each of the integrative units discussed above. The main point of this exercise is to demonstrate that by consistently measuring the size, scale, and organization of particular integrative levels within Pawnee society,

In the combined Pawnee village near Genoa, two intermediate coresidential units emerged. Not only did the members of various bands live in earthlodges that were clustered into discrete groups, the members of each band continued to work their own fields and place their dead into separate cemeteries. In addition to this band-level of organization, there was also a distinction at the sub-tribe level. According to the map provided by Hyde (1974), and reproduced in OʼShea (1989: figure 5), the Skidi and the South Bands (sub-tribes) each in-

33

Chapter Three

���������������������������������������� ���������������������������

������

�����

���������

����������������

����

������������������

�����

�������

�����

����������

�������

���������� ��������

������ ������ ���

�������� ������ ������ ���������

���������������������������������������������������������� ������������������� �����������������

����������������

������������������ ����

����������

������������ ������������������

��������� ������������������

������ ���

���

��������������� �����

����������������� �������������������

������������������

�����������������

�������������� ��������������������

������������������

������������������ �����������������

����������

���������� ��������������

��������� ������������������

������ ���������� ������ ��������� ������ ��� ���� ��������

34

����

������������������

��������������

������������

������������ ������������������

���

����������������

���

���������� ���������������

���

���������

Middle Range Theory

������������������������������������������ ���������������������� ������������������� �����������������

������������������

����������

����������������

������������������

�����

���������������

����

������������ ������������������

��������� ������������������

������ ���

����������������� �������������������

������������ �����

������������������ ����������������������

�����������������

����������������

������������������

����

������������������

����������������������

���������� �����������������

�������������������

������������ ������������������

���������� ������������������

��������� ������������������

������������

���

���

���������� ���������������

���������

������ ���������� ������ ��������� ������ ��� ����� ���� ��������

����������������������������������� ��������� ��������������������� �������� �����������������

������������������

����������

����������������

���

�����

�������� ������������ �������������������

����

��������� ������������������

������ ���

������������ �����

����������������� �������������������

������������������

�����������������

������������������

���

��������������

�������� ���������������� �������������������� ������������������� ���������� ������������������

35

�����������

����������

���

������������ ������������������

���

��������� ������������������

������ ���������� ������ ���������� ����������

����������������

���������� ���������������

������������������

��������������

���

���

���������

Chapter Three

Trajectories of Change in Pawnee Social Organization –�Permanent Village Pose Integrative Level

18th Century

Settlement Cluster

Early 19th Century

�������

������������

�������

������������

�������

������������

Band Territory

Tribal Territory

����������

������������������

����������

������������������

����������

������������������

Settlement

Late 19th Century

N/A

��������������������� ���������������������

Combined Tribal Village

Band Village

Autonomous Village

������������������

����������

Intermediate Coresidential Unit

�����

����������

N/A

Sub-Tribal Settlement

Earthlodge Cluster ?

Immediate Coresidential Unit

������ ���

������������������ ������������������

����������

�������� ������

�������� ������

�������� ������

Earthlodge

Earthlodge

Earthlodge

������������ �� ������ ������ �������������

Possible Patterns in the Data:

Spatial Data:

Artifactual Data: Between Settlement Clusters

Settlement Spacing Discrete Clusters 1 Discrete Clusters Discrete Clusters Diffuse Clusters 2 Diffuse Clusters Diffuse Clusters No Clusters No Clusters

Homogeneous Heterogeneous Heterogeneous Homogeneous Heterogeneous Heterogeneous Homogeneous (Between Sites) Heterogeneous (Between Sites)

3

DISCRETE CLUSTERS

RI

VE

Within Settlement Clusters Homogeneous Homogeneous Heterogeneous Homogeneous Homogeneous Heterogeneous N/A N/A

DIFFUSE CLUSTERS

RI

R

VE

RI

R

VE

R

SETTLEMENTS 1

DISCRETE CLUSTERS OF SETTLEMENTS

�

�

Figure 3.6��

2

DIFFUSE CLUSTERS OF SETTLEMENTS

3

NO CLUSTERS OF SETTLEMENTS

Idealized models for interpreting the patterning in the spatial and artifactual data. By combining these two lines of evidence, it will be possible to assess the degree of autonomy and interaction between the settlement sites within the study area.

36

Middle Range Theory

habited spatially discrete settlement areas that were separated by sod ramparts within the combined village. Nevertheless, despite these dramatic changes in social organization, the basic scale and organization of the immediate coresidential unit – the earthlodge – remained the same. This ethnohistoric example elucidates how the cognized social structure influences the actualized material patterns that can be used to characterize the integrative structure of a society over a given period of time. In addition, it demonstrates that by measuring the size, scale, and organization of different integrative units within a particular social system, it is possible to delineate trajectories of change over time. Since this example is based primarily upon ethnohistoric and ethnographic information, it is possible within each pattern to associate a particular cognized socio-structural level (i.e., band) with a particular actualized coresidential unit. In prehistoric archaeological contexts, this is frequently impossible. Nevertheless, while it may not be possible in prehistoric contexts to associate particular integrative units with particular cognized levels, by measuring the scale and organization of each level, it becomes possible to model how actualized patterns vary with respect to each other over time. That is to suggest that if the cognized models are causal, then they can be inferred by the specific patterns they generate. Summary: Modeling Tribal Social Structure This chapter has attempted to outline the middle-range theoretical principles that are necessary for modeling tribal social structure. Although I am primarily concerned with constructing a model that can be employed in prehistoric archaeological contexts, several of the basic tenets I have proposed are equally applicable within ethnographic and ethnohistoric contexts. A methodology for understanding the high degree of flexibility that characterizes tribal social organization over the long-term must incorporate analyses directed along two dimensions: integration (of segments) and interaction (between those segments). I have attempted to contextualize the various ways these concepts have been treated in anthropology and archaeology, and have suggested that there is much benefit to be gained by approaching them as distinct, yet intimately-related, methodological constructs. In particular, I have argued that it is possible to model the size, scale and organization of units of integration (i.e., immediate coresidential unit, intermediate coresidential unit, extended coresidential unit, and settlement clusters) through architectural and spatial information. On the other hand, the distribution of stylistic variables and the operation of exchange networks provide a method of assessing how these various integrative units interacted across the prehistoric landscape. By measuring the social structure of tribal societies via these basic material variables, it is possible to track trajectories of change along both axes – integration and interaction. 37

Chapter Four: The Archaeological and Geological Setting

This chapter provides the relevant background to the prehistory of the Great Hungarian Plain by placing the region into its geographic and archaeological context. The chapter begins with a brief discussion of the geologic history of the Carpathian Basin and the complex geomorphological processes that sculpted the Great Hungarian Plain during the later Pleistocene and Holocene. I then provide a period-by-period overview of prehistoric archaeological research on the Plain, leading up to a discussion of Late Neolithic and Early Copper Age social organization. The chapter ends with a discussion of the complicated chronological relationships between various Late Neolithic and Early Copper Age archaeological cultures throughout Central and Eastern Europe and the Balkans.

also laid down several hundred meters of sediment on top of the Pannonian deposits. Within this context, the rivers that eventually became known as the Danube and the Tisza gradually began incising and filling the basin with large amounts of fluviatile sediment brought down from the surrounding uplands (Pécsi 1964).

The Geomorphological Setting

The Geomorphology of the Great Hungarian Plain

This section briefly outlines the geologic and geographic setting of the Great Hungarian Plain within the Carpathian Basin. A more detailed discussion of the study area can be found in Chapter Six.

The eastern half of Hungary, extending from the Danube floodplain to the foothills of the Transylvanian mountains in Romania, is a flat, poorly-drained, plain that has been gradually filled by alluvial and aeolian deposits since the Pleistocene, creating a complex geomorphological landscape with only subtle changes in relief. This is the area known as the Great Hungarian Plain, or Nagy Alföld.

The topography of modern Hungary is traditionally divided into several discrete landscapes. The areas of Quaternary deposition are known as the Great and Little Plains (Nagy and Kis Alföld, respectively). The hilly and mountainous region to the west of the Danube is called Transdanubia (Dunatúl), and the Northern Mountains define the northernmost section of the country, extending into Slovakia (see Figure 4.1).

Geological History of the Carpathian Basin As the Danube winds out of central Europe towards the Black Sea, it passes through a large basin that it played a role in sculpting – the Carpathian Basin. The basin formed during the later Tertiary as the Alpine, Carpathian, and Dinaric mountain chains folded up around it. Gradually it became inundated by the Tethys, a large marine trough that Pécsi (1964:3) suggests was, “the Mediterranean of those days.” Throughout the Mesozoic, large quantities of limestone and dolomite were deposited at the floor of this prehistoric sea – the materials that today form the basement material throughout most of modern Hungary.

The westernmost geomorphological feature of the Great Hungarian Plain is the area between the Danube and Tisza rivers (Duna-Tisza Köze). This sandy interfluve was created during the later Pleistocene as the alluvial fan of the Danube was blown to the southeast, forming an extensive series of dunes that extend to the floodplain of the Tisza. This area remained unoccupied throughout much of the prehistory of the region, creating a large geographic and social boundary between the rest of the Plain to the east, and Transdanubia to the west. To the north and east of this sandy interfluve are a series of alluvial fans that separate the floodplain of the Tisza from the northern mountains (see Sherratt 1997a:275, figure 11.3). This string of alluvial ʻconesʼ was formed by smaller streams as they flowed out of the northern mountains into the Tisza during the Pleistocene and Holocene. Another Pleistocene alluvial fan – the Nyírség – rises in the northeastern most corner of the Plain. Initially deposited by a series of rivers

During the Pliocene, the massif in the center of the basin gradually began to subside, yielding itself to another inland body of water – the Pannonian Sea. Having deposited a large amount (up to 3,000 m in some places) of sandy, clayey sediment, the Pannonian Sea slowly transformed into a series of isolated freshwater lakes. Subsequent uplifting and subsidance gradually turned these freshwater lakes into rivers, which 39

Chapter Four

(Tisza, Szamos, Kraszna, and Bodrog) during the Pleistocene, the Nyírség was uplifted during the early Holocene, creating a substantial boundary to later streambeds. The westernmost margin of the Nyírség is abutted upon by the Hajdúság, a slightly more elevated pocket of aeolian loess. The southeastern corner of the Plain is occupied by the Maros alluvial fan, which was formed by a complex series of braided rivers that deposited thick layers of sand and sandy gravel during the Pleistocene and Holocene.

to be grazed on the more sandy areas that are not as suitable for intensive agriculture (see Pécsi and Sárfalvi 1964). The Plain is technically the westernmost extension of the Russian steppes, and throughout its long prehistory had an environmental character much different than today. Classified as a forest-steppe environment, the Great Hungarian Plain harbored a wide variety of different plant associations. According to Kosse (1979:48), the various plant associations of the forest-steppe are heavily dependent upon water supply and sodium content of the soils on which they occur. In areas of low alkalinity, pastures tend to form. In higher alkaline areas, sagebrush and salt-meadow grasses dominate. On alkaline soils in elevations above the floodplain, tartar maple-oak forests border upon elm-ash-oak groves in the floodplains.

Flowing between these various Pleistocene formations today are a series of rivers that originate in the east. The largest of these, the Tisza river, flows southwest into the Plain from the Ukraine. The present course of the Tisza is somewhat deceptive, for it is currently meticulously regulated by a complex system of dams and levees. Prior to the construction of these water management features, the Tisza wandered freely across the western half of the Plain, depositing thick fluviatile sediments as it meandered across the soft loess surface. During the Pleistocene, the Tisza flowed much further to the east than it does today, entering the Plain from south of the Nyírség and carving out the Berettyó-ér valley. During the Holocene, its course shifted to the north of the Nyírség, where it currently enters the Plain at the Tokaj Gate (Pécsi and Sárfalvi 1964).

Gyulaiʼs (1993) recent synthesis of pollen cores allows a general reconstruction of the environment within the Carpathian Basin throughout the Holocene. Building upon the work of Szujkó-Lacza (1991), Gyulai suggests that the Atlantic Phase, which corresponds roughly with the early Neolithic, was characterized by mixed oak forests on loess soils (i.e., oak, elm, lime, and ash with hazelnut). During the beginning of the Subboreal Phase, towards the end of the Neolithic (i.e., after ca. 5,000 BC), there was an invasion of beech associated with a general cooling trend, which resulted eventually in the creation of large deciduous forests in combination with parkland steppe areas during the later Copper Age. During the subsequent cool period, the parkland steppe gradually was replaced by a mixture of beech and oak forests. By the end of the Late Bronze Age, during the late Subboreal, the general continental character that had previously characterized the region gave way to species of Subatlantic type, dominated by beech and hornbeam (see Gyulai 1993:13-18).

The floodplain of the Tisza is traditionally divided into three discrete regions – the Upper, Middle, and Lower Tisza regions. The region to the east of the Tisza, consisting mostly of Holocene alluvial deposits, is called the Tiszántúl. Created by the continuous meandering of the Tisza, Berettyó, and Körös rivers throughout the Holocene, this region was annually inundated by early spring and early summer floods. Prior to its drainage and regulation during the nineteenth century, this portion of the country was a complex series of swamps and oxbows, interspersed by alluvial silts and clays and pockets of wind-blown and redeposited loess (Pécsi 1970).

The Archaeological Setting

Soils and Environment

The complex geomorphological history of the Great Hungarian Plain has left behind a region that is, in many ways, ideal for archaeological research. The Carpathian Basin forms a topographically discrete unit within the European landscape and the Great Hungarian Plain forms an even more discrete geomorphological region within that basin. Separated from the rolling hills of Transdanubia by the Danube and Tisza rivers, the rich alluvial soils of the Great Hungarian Plain are surrounded by imposing geographic boundaries, creating an almost insular environment that to this day is considered a distinct cultural area with more affinities to the east and south, in Romanian Transylvania and the Yugoslavian Banat, than with the west and north.

The frequency of flooding in the Great Hungarian Plain has largely inhibited the process of soil formation. As a result, the vast majority of the region is characterized by skeletal (i.e., alluvial, aeolian) soils with little or no humic formation in those areas where flooding occurred, and steppic soils (chernozem) in those areas that were high enough to avoid inundation. These latter soils were described by Pésci (1964:84) as “the most fertile type of soil in Hungary”. Areas that were perennially under water are characterized by meadow and swamp soils, which tend to be rich in clay and peat (see also Sherratt 1997a:276). The present vegetation of the Great Hungarian Plain is currently given over to industrialized agriculture. Sunflower, rice, wheat, barley, and corn are harvested on the remnants of large collective farms, which have been replaced more recently by small individual farmsteads. A long tradition of stockbreeding persists in the region, and cattle, sheep, and horses all continue

The extensive deposition of wind-blown loess during the end of the Pleistocene and the beginning of the Holocene, combined with the successive movement of several active streambeds across the Plainʼs soft loess surface, left behind a surface landscape largely devoid of the raw materials that prehistoric people were so fond of using, thus transforming 40

Archaeological and Geological Setting

the Plain into a virtual laboratory for prehistoric archaeologists. Since the prehistoric inhabitants were forced to look elsewhere for metal and stone, which occur in abundance in the surrounding mountainous areas, the presence of materials such as obsidian, chert, copper, and gold in different archaeological contexts provides the prehistorian with the ability to identify patterns of exchange in a manner unparalleled in most other parts of the world.

and alluvial fans in the northern Alföld to the sandy plain between the Danube and the Tisza (Kertész 1996:16-17). Interestingly, despite the relatively high density of Mesolithic sites, Kertész did not discover any Early Neolithic (Körös) sites in the Jászság. The closest Early Neolithic sites are located in the Middle Tisza valley. This led Kertész to suggest that differences in site location are likely the result of, on the one hand, geological, hydrological, and climatic considerations; and, on the other, the differing ecological demands of hunters and gatherers (i.e., Mesolithic populations) as opposed to farmers (i.e., Early Neolithic populations):

While these various geological events have created a region that in many other respects offers benefits that can be found in only a few other parts of the world, these same geomorphological processes have obscured much of the early prehistory of the Plain, leaving a picture of the past that fades to black sometime during the Mesolithic.

The Mesolithic sites in the Alföld lie on the well-watered alluvial plain (the Jászság area) and in sand regions on the dune ridges associated with marginal floodplains and water-courses (the sandy plain between the Danube and the Tisza as well as the Nyírség area). In other words, the Mesolithic sites lie in ecological niches which were rejected by the Körös culture. The Körös culture preferred levees suitable for agriculture. This is the main reason why Mesolithic and Early Neolithic Körös groups lived side by side for a long period of time in the Alföld (Kertész 1996:25).

Upper Paleolithic, Epipaleolithic, and Mesolithic The earliest prehistory of Hungary can be traced to the Lower Paleolithic at Vértesszőlős (see Kretzoi and Dobosi 1990), and several Middle Paleolithic sites have been documented throughout Transdanubia and in the Bükk Mountains, but the earliest prehistory of the Great Hungarian Plain begins sometime during the Upper Paleolithic. As Kertész (1996:5) has recently noted, “…the Great Hungarian Plain (the Alföld) remained a terra incognita until recently with regard to the Palaeolithic and Mesolithic”.

Kertészʼs recent research is beginning to shed light upon a much-disputed topic in Hungarian archaeology – the origins of the Neolithic.

Although some Upper Paleolithic and Mesolithic sites had been identified on the Plain during the early decades of this century (e.g., Szeged-Öthalom, Hugyaj/Érpatak, TószegÁldozó, see Banner 1936; Hillebrand 1925; Sümeghy 1944), the cultural makeup of the region prior to the Neolithic remained largely unknown. As a result, reconstructions of these earlier periods were therefore subject to much speculation (for a recent discussion, see Makkay 1996). In recent years, however, new evidence regarding the later Pleistocene and early Holocene habitation of the Plain is beginning to emerge, thus clarifying the relationship between the early Neolithic agriculturalists and their hunting and gathering predecessors.

Early and Middle Neolithic The Neolithic of the Great Hungarian Plain is characterized by a trend towards increasing regional differentiation in ceramic styles, settlement patterning, and resource exploitation. This trend began with the intrusion of the earliest farmers onto the Plain during the Early Neolithic (Körös Culture). It gained pace gradually throughout the Middle Neolithic (Alföldi Vonaldiszes Kerámia, or Alföld Linear-Decorated Pottery AVK), and finally reached its most heightened and regionally differentiated level during the Late Neolithic (Tisza-HerpályCsőszhalom Complex).

Of particular interest in this regard is the work of Kertész (1996), who has identified several Upper Paleolithic, Epipaleolithic, and Mesolithic sites in the northwestern corner of the Plain, in the Jászság area. Kertész notes that Upper Paleolithic and Epipaleolithic sites in the Jászság tend to be located on Pleistocene alluvial fans. In contrast, Upper Paleolithic sites in the center of the Plain tend to be covered with thick Quaternary deposits. For example, Kertész (1996: 12) notes that the cultural layer at the Szeged-Öthalom campsite was found at a depth of 4.5 m, while at Madaras-Téglavető it was between 6.5 and 7 m deep. The depth of deposits on top of these early sites precludes their systematic investigation at the center of the Plain. Mesolithic sites, on the other hand, are located in a wider variety of geological contexts in the Jászság, and can therefore be expected in different contexts throughout the Plain. These range from marginal subsidences

The Early Neolithic Körös Culture The earliest agriculturalists on the Great Hungarian Plain, known as the Körös Culture, appeared sometime during towards the end of the seventh millennium BC (cal.), shortly after the Neolithic became well-established in the southern Balkans. Unlike their southern contemporaries in the Thessalian Plain of northern Greece (see Demoule and Perlès 1993), wherein domesticates regularly make up 90-95% of faunal assemblages, there are generally fewer domesticates in the Early Neolithic Körös sites (ca. 75%; see Bökönyi 1988; see also Tringham 1971:92). The Körös culture is traditionally discussed as part of a much larger cultural complex that encompasses the Great Hungarian Plain and eastern Romania known as the Körös-Criş complex (see Bognár-Kutzián 1944; Tringham 1971:91-96). Sometimes the northern Yugoslavian 41

Chapter Four

Early Neolithic variant (Starčevo) is also included, thus creating the Körös-Starčevo-Criş complex (see Makkay 1996 for a recent discussion). Kalicz and Makkay (1977) assigned a few sites in the northeastern Carpathian Basin to a sub-group of Körös – the Szatmár group – since these sites were spatially distinct and appeared to be contemporary with the later phases of Körös. However, more recent research (see Makkay 1996) has demonstrated that these sites should be considered transitional between Körös and AVK, or even early AVK (e.g., Kalicz 1998).

he located almost one hundred Middle and Late Mesolithic sites in an area of approximately 100 km2 (see Kertész 1994; Kertész et al. 1994; Kertész 1996). This evidence for a rather intensive occupation of the northeastern Plain during the later Mesolithic, combined with the contemporary lacuna of sites between the Jászság area and the distribution of Körös sites further to the south, have led Makkay to suggest that: … the Körös culture did not penetrate further north partly because it was repelled by the heavy, water-logged clay of the low wet backswamps and partly because the area was inhabited by a Late Mesolithic indigenous population. Which for some reason or other, resisted Neolithization in spite of the fact that for some time it lived beside the northern border of the Körös culture. And when this Mesolithic population finally accepted the Neolithic way of life, it did so on its own terms. By this time, however, the decisive element was not the geographic boundary and environment, but rather the different ethnic background. (Makkay 1996:41).

As is the case with the Early Neolithic throughout all of Europe, the process of “Neolithization” within the region has received much attention. Two questions have dominated research into the Hungarian Early Neolithic. The first concerns how the region came to be occupied by ceramic- and foodproducing groups. The second addresses how those ostensibly invasive groups related to indigenous Mesolithic hunting and gathering populations. Until the recent work of Kertész (see above) in the Jászság area the relationship of the Early Neolithic agriculturalists to their Mesolithic predecessors, and contemporaries, has remained the subject of speculation.

The relationship between Late Mesolithic hunters and gatherers and Early Neolithic farming populations has been the topic of much discussion in European prehistory (e.g., Keeley 1992; Price and Gebauer 1992; Whittle 1996:43-71, 1997). It is likely that future research in the Great Hungarian Plain will further elucidate the particular means by which “Neolithization” occurred via different processes of diffusion, acculturation, and migration throughout different parts of the continent.

In a recent discussion of the topic, Makkay (1996:35) noted: Traditionally, research of the Körös-Starčevo-Criş (KS or KSC if the Transylvanian distribution of the culture is also involved) has been mainly concerned with regional chronologies, typological nuances, and the refinement of the internal chronology of the culture complex. Considerably less effort has been devoted to a better understanding of the Körös distribution, or the research of its settlement patterns and even less to the explanation of large-scale patterns of historical development.

The Middle Neolithic Alföld Linear Pottery Culture (AVK) The Middle Neolithic of the Great Hungarian Plain is known by a variety of different names and acronyms in the literature – Alföld Linear Pottery Culture (ALP, or ALPC) in English, Alföldi Vonaldíszes Kerámia (AVK) in Hungarian. All of these terms attempt to differentiate this Linear Pottery Culture from its Central and Northern European successors, the Linienbandkeramik (LBK) and Stichbandkeramik (SBK).

Makkay has for a long time commented on the interesting fact that the distribution of the Early Neolithic Körös Culture did not extend to the alluvial fan of the Great Eastern River, nor into the Upper Tisza valley (cf. Makkay 1982 for discussion). This has led him to suggest: … that the northern expansion of the Körös culture in the Tisza valley was blocked not by geographic factors, but rather by an ancestral (and, possibly, linguistic) border which in the Neolithic did not coincide with any major natural geographic boundary, but which might easily have coincided with an earlier, Mesolithic, or even Pleistocene geographic boundary or area and, also, with a corresponding cultural boundary which must already have existed in the Mesolithic. (Makkay 1996:41)

The AVK apparently developed in situ within the Great Hungarian Plain, directly out of its Early Neolithic (Körös and Szatmár) predecessors (Kalicz and Makkay 1977; Makkay 1982), sometime during the middle of the sixth millennium BC. Roughly contemporary with the transition from Starčevo to Vinča further south in Yugoslavia, the transition from Körös to AVK in the Plain is also characterized by ceramic changes. In the Great Hungarian Plain, this period is marked by the appearance of incised, linear decoration on open-shaped vessels. While AVK pottery is typically decorated with white paint, the early Vinca pottery is typically dark and burnished. Roughly simultaneous with this development in the eastern Carpathian Basin, the LBK appears in Transdanubia (for a recent synthesis of the Neolithic in Transdanubia, see Kalicz 1998). Makkay (1982:42-43) suggests that the AVK developed out of the Szatmár II group in eastern Hungary, and the LBK from

The northern part of the Alföld also follows the boundary between the continental and Mediterranean climatic regimes, and it may therefore have been less suitable to growing vernalized crops, such as winter wheat, perhaps providing another explanation for the distribution. Nevertheless, the overall pattern of site distribution has recently been bolstered by Kertészʼs research in the Jászság area (see above), where 42

Archaeological and Geological Setting

the Bicske Group.

The Late Neolithic Tisza-Herpály-Csőszhalom Complex

Kosse (1979) noted that AVK sites occupy a much wider range of environments than sites of the preceding Körös Culture, penetrating into forest zones hitherto unoccupied during the Neolithic. Correlated with this shift in settlement expansion, Kosse (1979:150-151) recorded a shift towards the more intensive exploitation of locally available domesticated animals and cultivars:

The trend towards regional differentiation that began initially with the sub-division of the Carpathian Basin into two discrete cultural complexes – AVK (on the Plain) and DVK (Dunatúl Vonaldiszes Kerámia, or Transdanubian Linear Pottery) – continued throughout the Middle Neolithic in the Great Hungarian Plain (e.g., Szakálhát, Esztár; Tiszadob). This resulted in the division of the region into three discrete ʻculturalʼ groups during the Late Neolithic, known as the Tisza-HérpályCsőszhalom complex (see Figure 4.1). Within the wider context of central, eastern, and southeastern Europe, this complex is roughly contemporary with Lengyel (I-II) in Transdanubia, and the Petreşti culture in Transylvania (for discussions see Kalicz and Raczky 1987a:25-27; Raczky 1988).

In the Linear Pottery culture, the wheat-sheep-aquatic/wild resource economic pattern of the Körös culture underwent several changes, the best documented of which is the shift to locally available domesticated animals. Cattle replaced sheep as the most important species, and pigs, which had been almost negligible at the Körös sites, became equally or more important than sheep. Aurochs and wild swine are both indigenous to much of temperate Europe, and the increase in the number of their domesticated counterparts at the Linear Pottery sites was without doubt the result of local domestication (Kosse 1979:150-51).

In their recent synthesis of Late Neolithic research on the Plain, Kalicz and Raczky (1987a) traced the concept of a younger, final, phase of the Neolithic to Tompaʼs (1929) influential article: [Tompa] assigned the Lengyel culture of Transdanubia to the younger, second phase of the Tisza culture, and traced its development to the Transdanubian expansion of the Tisza population. In fact, Tompa had correctly identified the culture complexes that are even today regarded as the main representatives of the Late Neolithic in Hungary (Kalicz and Raczky 1987a:11).

This basic trend has been explored further by Vörös (1994), who documented a shift in the relative number of domestic to wild animal bones in AVK sites (94-98% domestic vs. 2-6% wild), and in later AVK/early Tisza sites (63-78% domestic vs. 22-37% wild). A wider range of domesticated plants, cereals, and legumes, began to be utilized during this period. In addition to the Early Neolithic complex of einkorn and emmer wheat, two-row hulled barley and millet, breadwheat and lentils also began to be cultivated.

Today, the Late Neolithic of the Great Hungarian Plain is understood to be represented by three regionally-discrete ʻcultures,ʼ as they are known in Hungarian literature. These sub-groups – Tisza, Herpály and Csőszhalom – were initially identified by differences in ceramic styles, and have more recently received support by the identification of regional differences in settlement patterns and subsistence (Kalicz and Raczky 1987a:13-14).

Towards the end of the Middle Neolithic, about 5,100 BC, the previously homogeneous AVK cultural assemblage gave way to more spatially-discrete ceramic groupings within the Great Hungarian Plain (Bognár-Kutzián 1966; Kalicz and Makkay 1977). As Sherratt (1997b:280) noted: On the lower and middle Tisza and along the Körös, the Szakálhát group is distinguished by pottery ornamented with paint outlined by incised decoration. To the north of this the Esztár group occupied a wide territory on the Nyírség, Szamos and upper Berettyó, in which pottery with dark painted ornament was used. Various smaller groups emerged on the fringes of the northern mountains: Szilmeg in the Bükk foothills, Tiszadob in the Sajó valley, and Bükk in the mountains themselves, with exceptionally fine incised and occasionally painted pottery. All of these groups are distinguished by their characteristic finewares, which occur in smaller quantities as imports in each othersʼ areas.

General Trends During the Late Neolithic The Late Neolithic of the Great Hungarian Plain is most frequently associated with the northernmost penetration of tells, and tell-based lifeways, into the European continent (Makkay 1982; Raczky 1985; Kalicz and Raczky 1987a; Chapman 1997a). For the most part, the large tell settlements that are typically associated with the Late Neolithic cultures of the Plain were founded during the end of the Middle Neolithic, Szakálhát phase: The lowest layers of most tell settlements and contemporaneous single-layer settlements are now thought to represent the period that was until recently termed late Szakálhát or Szakálhát-Tisza transition on the basis of pottery styles. (Kalicz and Raczky 1987a: 14)

Sherratt attributed this small-scale cultural distinctiveness to economic differentiation, and to increased trade between the obsidian-bearing regions in the mountainous Tokaj region and the Plain.

Thus, although tell settlements are a sort of fossil directeur of the Late Neolithic on the Plain, they frequently were founded earlier, and therefore were subject to longer durations of 43

Chapter Four

habitation than their ʻflatterʼ counterparts, which occured even more frequently in the Late Neolithic Landscape. To highlight this point, Kalicz and Raczky (1987a:15) differentiate between three basic Late Neolithic settlement types, based upon layout, thickness, and areal extent:

In addition to these various shifts in settlement form and distribution, the manufacture of ceramics also underwent much change during the Late Neolithic: One conspicuous feature of this period is a basic change in pottery technology as compared to preceding periods. Chaff was no longer used for tempering, and the qualitative differences between coarser and finer wares practically disappeared…The new pottery forms making their appearance in this period include various amphora-shaped vessels and high pedestalled bowls. (Kalicz and Raczky 1987a:19).

1. Genuine tell settlements. These are very denselysettled sites that were inhabited for long durations, resulting in the vertical accumulation of habitation layers regularly reaching heights of 3-4 meters. 2. Tell-like settlements. Similar to tells, but generally with less long durations of habitation, these sites generally extend over larger areas than tells, and rise to a modest height of 1-2.5 meters.

Despite this ubiquitous shift in pottery manufacturing technology, “[R]egional differences can be observed in pottery ornamentation that are culturally diagnostic for the various cultural units of the Late Neolithic in the Great Hungarian Plain” (Kalicz and Raczky 1987a:20). These regional differences will be discussed below.

3. Single-layer horizontal settlements. By far the most frequent type of settlement in the Tisza region, horizontal settlements are characterized by thin cultural layers (25-50 cm).

Late Neolithic subsistence patterns continued many of the trends established during the Middle Neolithic. Despite a heavy reliance upon domestic animals, regularly one-quarter to one-half of Late Neolithic faunal assemblages are of wild species – aurochs, roe deer, red deer, and wild pig (Kalicz and Raczky 1987a:24). Of the domestic animals, cattle and pig increase in frequency, and sheep/goat generally decrease. These domesticates supplemented a diet otherwise based upon the cultivation of domestic plant species (e.g., einkorn wheat, millet, barley, and peas ), the collection of wild nuts and berries, and fishing (see Hartyáni et al. 1968; Hartyáni and Máthé 1979; Kalicz and Raczky 1987a:24).

Kalicz and Raczky (1987a:16) note that horizontal settlements tend to occur in greater frequency in the northern Tisza territory, and that tells, and tell-like mounds occur in more frequency in the southern Alföld, particularly around Hódmezővásárhely and Szeged. These various settlement types are frequently found side-by-side. In some cases, they are combined into single, complex settlements wherein a tell-like settlement is incorporated into a larger, more dispersed horizontal settlement (cf. Korek 1961:9-26; Sherratt 1984:33-36; Kalicz and Raczky 1987a:17). Several authors have noted a dramatic decrease in the number of sites dating to the Late Neolithic (e.g., Makkay 1982; Kalicz and Raczky 1987a:16). A good deal of this is undoubtedly due to the fact that Late Neolthic horizontal settlements are quite difficult to identify during the course of surface survey, and they therefore frequently go undetected. Nevertheless, a general trend towards nucleation is typical of the period (see, for example, Makkay 1982).

Geographic Distribution of the Groups The subdivision of the Plain into three discrete cultural groups occurred gradually throughout the Late Neolithic: The beginning of the Late Neolithic over the entire Great Hungarian Plain is presently equated with the appearance of the formative, early phase of the Tisza culture. It has also become clear that the Herpály and Csőszhalom complexes did not evolve directly from the local Middle Neolithic groups, but from this remarkably uniform early Tisza culture (Kalicz and Raczky 1984: 131-2).

The recent excavation and publication of several important Late Neolithic sites (e.g., Raczky 1987a) is further clarifying the extent of intra-site organizational variability that occurred within each of the different sub-groups. Late Neolithic sites on the Great Hungarian Plain are sometimes fortified with ditches (e.g., Gorzsa [see Horváth 1987], Csőszhalom [Raczky et al. 1994]); and some have both ditches and ramparts (e.g., Öcsöd-Kováshalom [Raczky 1987b]). In general, the houses that occur on Late Neolithic settlements are large, rectangular structures, with internal sub-divisions and plastered floors.

During the early phase of the Late Neolithic, the Tisza culture extended from the floodplain of the Tisza River, in the west, to the foothills of Transylvania, in the east: By the classical phase, this uniform early Tisza complex had disintegrated into smaller cultural units: the Herpály culture and Csőszhalom group in the east and northeast [respectively]. In the Körös region and in the Middle and Southern Tisza region the development of the Tisza culture was uninterrupted. (Kalicz and Raczky 1987a:14).

Burials frequently occur in and around Late Neolithic settlements, either in unoccupied parts of the settlements, in refuse pits, or in small clusters near the settlements (Kalicz and Raczky 1987a:23). Generally lacking in grave goods, Late Neolithic burials were often covered in red ochre and placed in a contracted position inside wooden coffins. 44

Archaeological and Geological Setting

Csõszhalom Culture

N

Danube River

Budapest

Tisza River

Tisza Culture

Transylvania

Körös River

Herpály Culture

Lake Balaton

Maros River Carpathian Mountains

Danube River

Sava River

Belgrade

= 100 Km

Figure 4.1� Approximate distribution of Late Neolithic 'Cultural Groups' on the Great Hungarian Plain, ca. 5,000-4,500 BC. � Map adapted from basemap created by László Zentai (Department of Topography, Eötvös Loránd University).

Tisza River

N

"Lucska"

Danube River

Budapest

"Basatanya"

Transylvania

Tiszapolgár Culture "Kisrétpart"

Lake Balaton

Körös River

Maros River Carpathian Mountains

Danube River

"Deszk"

Sava River

= 100 Km

Belgrade

Figure 4.2 � Approximate distribution of the Early Copper Age Tiszapolgár Culture, ca. 4,500-4,000 BC. Bognár-Kutzián's � (1972) "sub-groups" - Lucska, Basatanya, Kisrétpart, and Desk - are also included. Basemap created by László � Zentai (Department of Topography, Eötvös Loránd University).

45

Chapter Four

Kalicz and Raczky subdivide the territory occupied by the classical Tisza culture into two distinct zones, “one extending from the Aranka stream to the Körös region, the other from the Körös rivers to the upland region and to the Csőszhalom territory. The Körös valleys form a kind of frontier zone between these two main areas” (Kalicz and Raczky 1987a: 14; see Figure 4.1).

and tell-like settlements near major rivers indicates their prominence as foci of trade. The tell sites of the Tisza culture range in size from ca. one to four ha, and the tell-like mounds generally cover larger areas (e.g., 6 ha at Gorzsa; and 11 ha at Szegvár; see Kalicz and Raczky 1987a:16). Horizontal settlements are even more variable, ranging in area from 1-12 ha: It has been suggested that the coexistence of these three settlement forms perhaps reflects some sort of economic or social hierarchy. However, these suggestions have been mostly based upon the presence of single-layer settlements concentrated around major Tisza mounds (Makkay 1982: 111-164; Horvath 1985:93), and only the extrinsic criteria of this assumed hierarchy, focusing on settlement size, have been outlined; causal relationships yet remain to be proven. (Kalicz and Raczky 1987a:16)

The Herpály culture occurs over a much smaller geographic area within the Berettyó and Sebes-Körös river valleys in the eastern Alföld. During its later phases, the Herpály culture extended further east into Transylvania, in modern-day Romania (Kalicz and Raczky 1987a:14). The Csőszhalom group, in the Upper Tisza and Bodrog valleys, has received the least attention of the three groups, and it therefore remains the most poorly understood. The type-site, Polgár-Csőszhalom, has recently been the focus of intensive research, and is beginning to clarify the relationship of this sub-group to its southern neighbors (see Raczky et al. 1994; Raczky et al. 1997).

Kalicz and Raczky prefer to attribute such variability in settlement form to adaptations to localized environments. Noting the tendency of Tisza sites to incorporate tells, or tell-like mounds, into a larger horizontal site, Kalicz and Raczky (1987a:16) suggest that, “…the entire area of these singlelayer settlements had not been occupied simultaneously and that their large areal extent can probably be attributed to the periodic shift of smaller settlement nuclei.”

As noted above, these three discrete groups were initially identified by differences in ceramic production and decoration techniques. More recent research has demonstrated that the various differences between the three groups extended to other aspects of social organization as well. In addition to obvious differences in ceramic assemblages, more subtle patterns in subsistence patterns, and settlement organization are beginning to further distinguish one from the other.

Tisza settlement sites are frequently fortified with ditches, or are surrounded by fences, and the houses on them are usually very closely-spaced. Discrete house clusters, separated by small open areas, have been identified at sites such as ÖcsödKováshalom (Raczky 1985, 1987) and Kisköre (Korek 1986, 1989). Horvathʼs (1987) excavations at HódmezővásárhelyGorzsa identified houses organized around an unoccupied central area.

The Tisza Culture The Tisza Culture has received the most attention of the three groups that occupied the Great Hungarian Plain during the Late Neolithic. Its impressive tell sites, distinctive ceramics, and decorative figurines have drawn the attention of professional and avocational archaeologists since Móraʼs (1930) excavation at Csóka in 1907. But it was not until the 1972 publication of Korekʼs monograph that the Tisza Culture received comprehensive treatment (see also Korek 1989). Subsequent excavation of various Tisza sites has further clarified the internal organizational features that characterize the settlements of this culture, and which differentiate them from its northern (Csőszhalom) and eastern (Herpály) neighbors (see Kalicz and Raczky 1987a).

Tisza houses, such as those at Hódmezővásárhely-Gorzsa, Öcsöd-Kováshalom, and Vésztő-Mágor, tend to be rather large, and frequently have internal sub-divisions. For example, Horváth (1987:38) described House 2 in Phase C at Gorzsa as follows: The groundplan of house 2 resembles an inverted angular U. A rectangular room (3.60 x 7 m) adjoined and connected the two rooms of the northeastern wing (6.20 x 5.80 m and 5.80 x 6.20 m, resp.) from the northwest with room 4 (5 x 5.80-6 m) of the southwestern wing, that adjoined room 5 (5-4.2 x 8.4 m) whose orientation corresponded to the longitudinal axis of the southeastern wing, but was not quite parallelly aligned. Room 6 (3.6 x 3.6 m) lying at the northwestern end of the house showed a somewhat differing mode of construction, but had nonetheless clearly belonged to this building. A slightly narrowing passage lay between the two wings (1 to 1.50 m wide). The overall length of the house was 20.2 - 20 m, its width 12.8-13 m. Remains of a doorway could be observed to the first room on the southern side and between rooms 1 and 2; the doorways between the other rooms had been destroyed by

Tisza ceramics are characterized by deeply-incised meandric patterns on unpolished, usually open-mouthed, vessels. The incised meanders normally occur in decorative panels, and the vessels are normally fired to a pale orange or bright-red color (Kalicz and Raczky 1987a:20; Korek 1989). In contrast to ceramics of the Herpály and Csőszhalom groups, which are typically painted with various colors, Tisza ceramics are painted only occassionally with wide black bands (see Kalicz and Raczky 1984; Raczky 1985; Kalicz and Raczky 1987a). Makkay (1982) has suggested that the location of Tisza tells 46

Archaeological and Geological Setting

later pits.

function and had perhaps been the house of a special personage (a shaman or priest), or that it had been a village shrine, or perhaps both (Hegedűs 1983:7).

The exceptional preservation of the house at Gorzsa, with walls preserved up to a height of 70 cm, can be attributed to the fact that it was destroyed by fire. Kalicz and Raczky (1987a:18) observed that the basic pattern of the houses at Gorzsa – large, with internal divisions and plastered floors – also occur at the Herpály type-site, and at Öcsöd-Kováshalom, another Tisza site (see also Kalicz and Raczky 1984:99-107; Raczky 1985:105). In addition, “recent excavations suggest that similar houses can be assumed also for other sites of the Tisza and Herpály culture, as well as for the Csőszhalom group, since parts of similar houses have been uncovered on the Csőszhalom tell” (Kalicz and Raczky 1987a:18).

It is tempting to overemphasize the ritual importance of this peculiar building and its associated assemblage. Its excellent preservation – due to its destruction by fire and the subsequent abandonment of the entire settlement – has very likely produced an assemblage that appears to be much more unique than it probably is, or was. Other domestic shrines, such as Room 3 of the U-shaped building(House 2) at Gorzsa, have produced similar, although by no means as extensive and elaborate, assemblages. Similar ritual areas have been identified within domestic structures at the site of Herpály (Kalicz and Raczky 1987b), suggesting that domestic ritual activities were not restricted to the Tisza culture.

Domestic shrines are quite common throughout the later Neolithic of southeastern Europe. Within the Tisza culture on the Great Hungarian Plain, clay altars and ʻsacrificial pitsʼ occur frequently in otherwise typical domestic structures see Kalicz and Raczky 1987a:22). But in addition to ritual activity areas within houses, there is growing evidence on some Tisza settlements for areas that were set aside specifically for ritual use. For example, Hegedűs and Makkay (1987) have tentatively argued for the presence of a village shrine at the site of Vésztő-Mágor. At that site, the excavators recorded a rectangular building, ca. 13 m long, with an internal sub-division, that was rebuilt several times. During its final construction, the floor of the building was replastered three or four times. In addition to typical domestic artifacts (storage jars, lithics, loom weights, bone tools), the building also contained an impressive array of non-utilitarian ʻculticʼ objects, including a large female figurine (ca. 80 cm high), a ʻlibation tableʼ, and various other non-utilitarian items (for a complete discussion, see Hegedűs and Makkay 1987:92-103).

Clusters of burials frequently occur in Tisza settlements. Kalicz and Raczky (1987a:23) have noted that: In the Tisza and Csőszhalom cultures the burials form small clusters within the settlement, whereas in the Herpály culture the burials found within the settlement are mostly of special or unusual type, with the ʻrealʼ graves lying outside the settlement boundary in small groups. This latter form in fact marks the separation of settlement and cemetery; practices indicative of a similar process have been noted on Tisza sites since the deceased were usually interred in the temporarily unoccupied parts of the settlement, suggesting that the deceased were thought to remain members of the community. (Kalicz and Raczky 1987a:23) Tisza burials tend to be in a contracted position, interred in wooden coffins, and frequently wrapped in rushwork mats interments, although a few secondary cremation burials are known from Öcsöd-Kováshalom (Kalicz and Raczky 1987a: 24).

These finds and their find circumstances, and the presence of fish scales and the debitage from stone working led K. Hegedűs to conclude that, ʻin spite of the finds of cultic character, this structure can be regarded as a residential buildingʼ (Hegedűs 1977:32). However, the possible function of this building must be reconsidered in view of the lavish assemblage of cult finds… (Hegedűs and Makkay 1987:97)

Burial goods are generally scarce in Tisza contexts, although they tend to increase in quantity throughout the period (Kalicz and Raczky 1987a:23). Tisza grave goods include limestone and marble beads, bone pins, red ocher, and occasional copper rings. Spondylus pendants and armrings, which tend to be associated with Tisza females, are unknown in Herpály contexts.

In their more recent interpretation of the building, Hegedűs and Makkay (1987:103) have tentatively suggested that:

The Herpály Culture

… the number and nature of these cult artefacts undoubtedly transcends the bounds of an individual ritual and the room appears to have been a place where sacrifices had been offered on more than one occasion. Consequently, this inner area or room can probably be interpreted as a domestic shrine that, however, also played a specific role within the entire community (that does not necessarily imply that some of the other contemporary houses of the Vésztő tell did not contain a similar cult room). One possible interpretation is that this building had some special

Situated in the eastern part of the Great Hungarian Plain, the Herpály culture was initially distinguished from the Tisza culture based upon its distinctive painted pottery. Generally highly smoothed or polished, pottery from Herpály sites is frequently decorated with a variety of painted patterns, which were applied after firing: The early Herpály phase is characterised by patterns painted in black with fine, delicate strokes, and the following phase by a combination of red painted wide bands 47

Chapter Four

framing finely painted white patterns, while the late phase by the predominance of white painting (Kalicz and Raczky 1987b:20).

that site. Domestic shrines, similar to their Tisza counterparts, are frequently found in Herpály houses: The excavations at Herpály did not uncover one single building that could unambiguously be interpreted as serving a sacred or ritual purpose, or that may have been a sanctuary. On the other hand, various built-in large altars of various form and basins were observed in almost each and every house. The charred grains, animal bones, and grinding stone fragments, often found plastered into or under these altars and basins can most probably be associated with some ritual or ceremony centered around these features…(Kalicz and Raczky 1987b:120-21).

Vessel forms include: cups, bowls, pedestalled bowls, ʻfruit standsʼ, jugs and amphorae (see Kalicz and Raczky 1987b: 116). In addition to painting, these vessels frequently have pointed knobs, and incised punctate decoration. The pottery of the Herpály culture is in many respects very similar to that of the Csőszhalom group (see below). Whereas settlement patterns of the Tisza culture generally combine all three settlement types, Kalicz and Raczky noted that:

Treatment of the dead also differed between the Herpály and Tisza cultures. Kalicz and Raczky (1987a:23) contend that burials found within Herpály settlements are somehow special or unusual, and that the normative procedure was to bury the dead outside the settlements in small groups. In addition, Herpály burials occur less frequently in wooden coffins. Boar tusk pendants and boar mandibles occur in Herpály and Csőszhalom burials.

The most characteristic settlement form of the Herpály culture appears to have been the tell and tell-like sites. The tells or tell-like mounds generally lie along well-definable points of the local drainage network, on islets or small peninsulas almost complete surrounded by water…These tells are surrounded by smaller and larger single-layer settlements, occupied for briefer periods of time, that are separated from the central settlement by a watercourse (Berettyószentmárton and Zsáka). (Kalicz and Raczky 1987a:17).

In addition to these general differences between the Tisza and Herpály cultures, there are also several, more subtle differences, that differentiate the material culture of the two groups. These include differences in oven form, storage bins, and figurine forms (cf. Kalicz and Raczky 1987b:19). Subtle differences also occur between the subsistence practices of the two groups. Evidence for fishing is largely lacking in the Herpály culture (cf. Bökönyi 1985:71; Kalicz and Raczky 1987b:122), and there may have been a more heavy emphasis on hunting and the domestication of the aurochs than in Tisza contexts (Kalicz and Raczky 1987a:24).

Several authors have observed that Herpály settlements are regularly spaced a few kilometers from each other (cf. Makkay 1957; Kalicz and Raczky 1984; Kalicz and Raczky 1987a:17). Although the sizes of the tells vary, they are generally understood to be smaller than Tisza settlements (