Mediterranean Archaeological Landscapes: Current Issues 9781934536285

Citation preview

Mediterranean Archaeological Landscapes: Current Issues

This page intentionally left blank

Mediterranean Archaeological Landscapes: Current Issues

Edited by

Effie F. Athanassopoulos and LuAnn Wandsnider

UNIVERSITY OF PENNSYLVANIA MUSEUM OF ARCHAEOLOGY AND ANTHROPOLOGY PHILADELPHIA

Copyright © 2004 University of Pennsylvania Museum of Archaeology and Anthropology 3260 South Street Philadelphia, PA 19104-6324 First Edition All Rights Reserved

Library of Congress Cataloging-in-Publication Data Mediterranean archaeological landscapes : current issues / edited by Effie F. Athanassopoulos and LuAnn Wandsnider. — 1st ed. p. cm. Includes bibliographical references and index. ISBN 1-931707-73-1 (hardcover : alk. paper) 1. Mediterranean Region—Antiquities. 2. Excavations (Archaeology— Mediterranean Region. I. Athanassopoulos, Effie-Fotini. II. Wandsnider, LuAnn. DE60.M417 2004 930'.09822--dc22 2004008734

Printed in the United States of America on acid-free paper.

Contents Figures and Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Preface and Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi 1

Mediterranean Landscape Archaeology Past and Present, Effie F. Athanassopoulos and LuAnn Wandsnider . . . . . . . . . . . . . . . . . . . . . 1

2

Less is Better: The Quality of Ceramic Evidence from Archaeological Survey and Practical Proposals for Low-Impact Survey in a Mediterranean Context, Timothy E. Gregory . . . . . . . . . . . . . . . . . . . . . 15

3

Sampling Sinop: Putting Together the Pieces of a Fragmented Landscape, Owen Doonan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

4

The Disjunction between Mediterranean and Near Eastern Survey: Is it Real?, T. J. Wilkinson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

5

Artifact, Landscape, and Temporality in Eastern Mediterranean Archaeological Landscape Studies, LuAnn Wandsnider . . . . . . . . . . . . . 69

6

Historical Archaeology of Medieval Mediterranean Landscapes, Effie F. Athanassopoulos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

7

Historical Contingency, Nonlinearity, and the Neolithization of the Western Mediterranean, C. Michael Barton, Joan Bernabeu, J. Emili Aura, Lluis Molina, and Steven Schmich . . . . . . . . . . . . . . . . . . . . 99

8

Time, Scale, and Interpretation: 10,000 Years of Land Use on the Transjordan Plateau amid Multiple Contexts of Change, J. Brett Hill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

9

World-Systems Theory and Regional Survey: The Malloura Valley Survey on Cyprus, P. Nick Kardulias and Richard E. Yerkes . . . . . . . . . . 143

10 From Density Counts to Ideational Landscapes: Intensive Survey, Phenomenology, and the Sydney Cyprus Survey Project, Michael Given . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 11 The Archaeology of Modern Greece, Lita Diacopoulos . . . . . . . . . . . . . 183 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235

v

This page intentionally left blank

Figures and Tables FIGURES

Figure 2.1 Figure 2.2 Figure 2.3

Sydney Cyprus Survey Project: study area . . . . . . . . . . . . . . . . . 16 Eastern Korinthia Archaeological Survey: study area . . . . . . . . 17 Australian Paliochora-Kythera Archaeological Survey: study area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Figure 2.4 Example of ChronoType hierarchy . . . . . . . . . . . . . . . . . . . . . . 22 Figure 2.5 Chronological scheme for the Eastern Korinthia Archaeological Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Figure 2.6 Processing team photograph (in field), EKAS . . . . . . . . . . . . . . 31 Figure 2.7 Processing team photograph (in field), EKAS . . . . . . . . . . . . . . 31 Figure 2.8 Processing team photograph (in field), APKA . . . . . . . . . . . . . 31 Figure 2.9 Processing team drawing (in field), EKAS . . . . . . . . . . . . . . . . . 32 Figure 2.10 Processing team drawing (in field), APKAS . . . . . . . . . . . . . . . 32 Figure 2.11 Processing team drawing (in field), EKAS . . . . . . . . . . . . . . . . . 32 Figure 3.1 The Sinop promontory, with topographic zones indicated . . . . 38 Figure 3.2 CORONA photograph of the Sinop promontory . . . . . . . . . . . . 39 Figure 3.3 Vernacular architecture in the Sinop hinterland . . . . . . . . . . . . 40 Figure 3.4 The major geological regions of Sinop promontory . . . . . . . . . . 44 Figure 3.5 Plan of the settlement and cemetery of Karapınar (Nohutluk quadrat, Karasu valley) . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 3.6 Survey tracts in the Demirci valley . . . . . . . . . . . . . . . . . . . . . . 50 Figure 3.7 Survey tracts in the central Karasu valley . . . . . . . . . . . . . . . . . 51 Figure 4.1 Site recovery from recent Near Eastern surveys compared with those from Greece . . . . . . . . . . . . . . . . . . . . . . . 57 Figure 4.2 Tells and small sites plotted according to archaeological periods from the area of Tell Beydar, eastern Syria . . . . . . . . . . 61 Figure 4.3 The number of archaeological sites in the Jabul survey area of north-central Syria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Figure 4.4 The number of sites according to archaeological periods excavated and surveyed in the upper Tabqa area, Syria . . . . . . . 62 Figure 4.5 The extension of Roman, Byzantine, and Islamic rural settlement into the upland behind Antioch . . . . . . . . . . . . . . . 63 Figure 6.1 Map of the NVAP survey area with identified Byzantine/ Medieval sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Figure 6.2 NVAP sites with Byzantine material in the southern part of the survey area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 vii

Figure 6.3 Figure 6.4 Figure 7.1 Figure 7.2 Figure 7.3 Figure 7.4 Figure 7.5 Figure 8.1 Figure 8.2 Figure 8.3 Figure 8.4 Figure 8.5 Figure 8.6 Figure 8.7 Figure 9.1 Figure 9.2 Figure 9.3 Figure 9.4 Figure 9.5 Figure 10.1 Figure 10.2 Figure 10.3 Figure 10.4 Figure 10.5 Figure 11.1

NVAP sites with Byzantine material in the northern part of the survey area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 View of the northern side of Mt. Polyphengi and the town of Nea Nemea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Research area in eastern Spain . . . . . . . . . . . . . . . . . . . . . . . . . 100 Ubiquity of land use for each chronological phase in each valley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Areal extent of different levels of land-use intensity . . . . . . . . 110 Lithic accumulation rates, in pieces per kilometer squared per millennium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Local density analysis for grid cells . . . . . . . . . . . . . . . . . . . . . . 114 Map showing the Wadi al-Hasa study area in west-central Jordan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Degree of settlement movement throughout the Holocene for upper quartile sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 Mean use intensity by FAO categories of soil loss . . . . . . . . . . 137 Mean PSL by use-intensity interval in upper quartile sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 Upper quartile Roman sites . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Upper quartile Byzantine sites . . . . . . . . . . . . . . . . . . . . . . . . . 140 Total PSL in northern and southern areas by period . . . . . . . . 141 Map of Cyprus showing location of Athienou Malloura and other ancient sites on the island . . . . . . . . . . . . . . . . . . . . 144 Map of the project area showing sites identified during the survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Map of Site 1 (Malloura) and adjacent area, including Site 2 (Magara Tepesi) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Viewshed analysis showing portion of project area visible from four hilltop sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Distribution of threshing sledge flints in the project area . . . . 161 Map of Cyprus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 Map of Sydney Cyprus Survey Project survey area . . . . . . . . . 171 Distribution of Roman pottery in the northeastern part of the SCSP survey area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 The Tamassos plain, looking north from the area of the Roman cemetery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .174 Distribution of Medieval–Modern pottery in the northeastern part of the SCSP survey area . . . . . . . . . . . . . . . 176 One of the endangered threshing floors (alonia) in Aroniadika, Kythera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 viii

Figure 11.2 Map showing location of Lakka Skoutara, Eastern Korinthia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 Figure 11.3 Lakka Skoutara showing distribution of houses, alonia, and Modern period artifact densities . . . . . . . . . . . . . . . . . . . . . . . 196 Figure 11.4 Typical house in the basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 TABLES

Table 3.1 Table 3.2 Table 6.1 Table 7.1 Table 7.2 Table 7.3 Table 7.4

Table 8.1

Overview of nested Black Sea Trade Project . . . . . . . . . . . . . . . 48 Relative ceramic densities for eight sample zones in the Karasu valley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Styles of Byzantine decorated pottery from NVAP sites . . . . . . 94 Regional chronology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Calculation of Temporal Index . . . . . . . . . . . . . . . . . . . . . . . . 105 Neolithic vs. Paleolithic: percent of area surveyed for Temporal Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Neolithic vs. Paleolithic: percent of area surveyed for Settlement Intensity Index in square kilometers per millennium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Chronological periods for the study area . . . . . . . . . . . . . . . . . 131

ix

This page intentionally left blank

Preface and Acknowledgments

T

he Mediterranean landscape record is recognized for its length and richness and the opportunity it offers to study the interaction between humans and their landscape over the short, medium, and long term. With this volume our intent is to explore a variety of current archaeological issues in the context of specific Mediterranean landscapes from southern Spain through Greece and Cyprus to Jordan and from Antiquity to Recent times. We are interested in anthropologically informed approaches, and our goal is to explore themes that have wide applicability and significant methodological and interpretive implications. The introductory chapter offers a brief history of Mediterranean landscape archaeology, identifies current issues in Mediterranean landscape studies and connects them to broader changes in the discipline of archaeology. The remaining chapters discuss fundamental issues and interpretations of Mediterranean archaeological landscapes. We consider methodological issues—efficiently applying chronometric information to surface contexts, sampling the landscape at multiple temporal and spatial scales, landscape taphonomy and its relation to settlement pattern reconstruction, assumptions about time based in artifact and monument survey, developing approaches to chronologically impoverished time periods, such as the Medieval and Post-Medieval—of general importance to Mediterranean landscape studies. Subsequent chapters offer interpretations of specific Mediterranean archaeological landscapes that transcend standard settlement approaches. Authors variously consider sophisticated ecological treatments of the human-landscape interaction through time as well as assume a world systems perspective. Finally, we explore the social and cultural meaning of survey data, the ideational landscape, and we examine the rich material culture of modern Greece, the present cultural landscape, for its information content. In this fashion, we map the wide interpretative potentials of the rich Mediterranean landscape. This volume was born in presentations at the Electronic Symposium “Crossroads in Mediterranean Landscape Archaeology” at the 66th Annual Meeting of the Society for American Archaeology, April 2001, in New Orleans. Prior to the symposium, authors and other participants had already read papers previously posted on a website; thus we were able to launch into a stimulating discussion on issues attending landscape archaeology in the Mediterranean. This volume reflects and extends the dynamic exchange that occurred at the symposium. We would like to thank the Research Council of the University of NebraskaLincoln for a publication subvention grant. We are indebted to the authors for their participation in the original symposium and their contributions. We would

also like to thank the anonymous reviewers for the improvements they suggested. Finally, we are grateful to Walda Metcalf of the University of Pennsylvania Museum Publications for her commitment to this project and her assistance and guidance through the publication process.

xii

Mediterranean Archaeological Landscapes: Current Issues

xiii

This page intentionally left blank

1 Mediterranean Landscape Archaeology Past and Present Effie F. Athanassopoulos and LuAnn Wandsnider

R

ecent studies of Mediterranean landscapes have emphasized their diversity, their fragmentation, and the high degree of contact between their diverse areas, that is, their connectivity (Horden and Purcell 2000). Moreover, the Mediterranean landscape record is recognized for its length and richness and the opportunity it offers to study long-term interaction between humans and their landscape, however landscape is defined. At the same time, the particular histories of archaeological perspectives that have dominated fieldwork in the region make it difficult to compare with other areas, for example, the New World. Thus, with this volume, our intent is to address issues of relevance not only to Mediterranean archaeology but to landscape archaeology in general. There has been a dramatic expansion in the theoretical approaches—both anthropological and classical—assumed by researchers here over the last 25 years. As well, over the same time span, a huge volume of field survey projects have been carried out in the Mediterranean arena (summarized in Cherry 2003:138–40). For these two reasons, it is appropriate to take stock of what we have learned, identify lacunae, and consider new approaches to our understanding of the rich surface landscape record of the Mediterranean. Where the Archaeology of Mediterranean Landscapes volumes (Barker and Mattingly 1999) emphasize technique and method geared toward understanding population processes, our goal with this volume is to explore theoretically diverse interpretative themes and the methods that make those approachable. The Side by Side volume (Alcock and Cherry 2004) strives to make comparative sense of the many intensive Mediterranean surveys that have been conducted over the last twenty years. Complementarily, this volume deliberately explores paradigms—from anthropology, history, and other disciplines—within which Mediterranean landscape studies are currently being conducted.

Mediterranean Archaeology: A Brief History Until recently, the paradigm within which most archaeological research has been conducted in the Mediterranean was strongly tied to the classical tradition and stressed the primacy of texts over material remains. Archaeology’s concern was to provide the material evidence of the truth revealed in texts

2

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

(Kardulias 1994b). Furthermore, since the eighteenth century, Mediterranean archaeology developed in close association with art history. The interplay of classicism and romanticism (e.g., Winckelmann [Marchland 1996:7–16]) came to define archaeological practice and interest in monumental architecture and objects of art. In addition, archaeology in the Mediterranean has had strong ties to nationalism. Most excavation projects tended to concentrate on periods of the greatest past glory as modern nations sought to recreate their past in the nineteenth century (e.g., Abu El-Haj 2001; Diaz-Andreu and Champion 1996; Kohl and Fawcett 1995; Meskell 1998; Morris 1994; Shanks 1996; Silberman 1990; 1995). Thus, lines were drawn, dividing the continuous flow of past time into periods of greater and lesser importance and, therefore, of greater and lesser scholarly interest. Given these trends, it is important to examine the growth of archaeology in the Mediterranean in its historical and social context. Archaeology’s subservient relationship to philology significantly influenced the direction of fieldwork. The prevailing model since the end of the nineteenth century has been large-scale excavations at urban or religious sites. The interest concentrated on the elites, while the archaeologically invisible (e.g., rural population) were excluded from the picture. Furthermore, archaeology’s task was seen primarily as an organizational one, to generate and order a vast body of data. Substantial effort has been invested in the compilation and classification of excavated data in multi-volume site reports and catalogues (Morris 1994). The production of such reports created the need for large numbers of specialists who dedicated their energies to the analysis of particular categories of material culture. Dyson argues that the “big-dig” was built on the organizational model of corporations and encouraged specialization and fragmentation. Preoccupation with artifact typologies and chronology building were on the top of the research agenda. Thus, “students were sent to study the fragments of red figure vase painting in some college museum or to produce a corpus of loomweights from some major site…” (Dyson 1982:89). In the post–World War II period, patterns of archaeological fieldwork changed. The prohibitive cost of the “big-dig” along with new perspectives of scientific humanism (Fotiadis 1995) led to the rise of surface survey as a cheaper and faster alternative, albeit still second-class, form of field archaeology. To a large extent this new kind of fieldwork capitalized on the strong topographic tradition that had developed in the region since the late eighteenth and early nineteenth centuries. Regional surveys in the New World (Willey 1953) and in Mesopotamia (Adams 1965; Adams and Nissen 1972) preceded the development of similar approaches in the Mediterranean. However, the early Mediterranean regional surveys were primarily an outgrowth of local research traditions and concerns, often with textual impetus (Cherry 2003:141–42) and framed by a pervasive scientific humanism (see below).

MEDITERRANEAN LANDSCAPE ARCHAEOLOGY PAST AND PRESENT

In Italy the first systematic regional project was the survey of South Etruria (Potter 1979; Ward-Perkins 1962; 1972; Ward-Perkins et al. 1968). In the 1950s and 1960s it was amongst the first in the Mediterranean to use surface survey as a tool for writing landscape history. It was undertaken by the British School at Rome and built on a long tradition of archaeological research in the region (e.g., Ashby 1927). It took place at a time when suburban growth and agricultural activity were bringing sites to the surface at an unprecedented rate (Potter 1979; Ward-Perkins 1955). It started as a small-scale field-walking program and gradually increased in intensity. Over a period of twenty years it served as a laboratory for the development of field survey methodology, integrating field-walking with the study of extant rural remains, excavations at selected sites, paleoenvironmental studies, and documentary research. It produced a record of settlement and land use from late Prehistoric to Medieval times (Barker 1991:1–2). The South Etruria survey represents one of the most significant achievements of Mediterranean landscape archaeology in the postwar period. It provided the model and inspiration for the next generation of survey projects organized from the 1970s on (e.g., Attema et al. 2002; Barker 1995a; 1995b; Barker and Lloyd 1991; Coccia and Mattingly 1992; 1996; Dyson 1978; 1981a; 1981b; Hayes and Martini 1994; Malone and Stoddart 1994). Currently, the South Etruria evidence is being re-analyzed. Advances in ceramic analysis and the availability of computer technology (e.g., Geographic Information Systems or GIS) have led to the re-evaluation and enhancement of this important data set (Belcher et al. 1999; Patterson and Millett 1998; Patterson et al. 2000). A comparably signal project is the Minnesota Messenia Expedition (MME) in Greece. This regional survey started in the 1950s and continued until the mid1970s (McDonald and Rupp 1972), serving as a model for the Aegean in subsequent decades. The innovations introduced piecemeal by MME included the adoption of a regional approach, attention to the environmental setting and its economic potential, interest in sites and settlements of all periods, ethnographic studies, and concern with later periods that were traditionally the domain of history. Guiding all of these studies (and evident in archaeological regional studies into the 1980s) was an over-riding concern for demographic and economic processes, consistent with the scientific humanism (Fotiadis 1995) that pervaded social sciences. The apparent success of MME led to the proliferation of regional surveys in Greece. In 1981, a conference (see Keller and Rupp 1983) on survey archaeology in the Mediterranean was held in Athens, Greece, underlining the increasing role of survey in Mediterranean archaeological fieldwork. The papers covered a wide range of topics including research strategies, field methods, and interpretation. They demonstrated that archaeological survey was widely employed, although in highly variable ways for highly varying purposes (Cherry 2003:138–41; see also Dyson 1982). In a thoughtful essay concluding that volume, John Cherry (1983) reflected on the unique contributions of survey in the Mediterranean and offered

3

4

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

a series of prescriptions for improved survey—diachronic perspective, regional scope, multi-disciplinary team effort. In the main, these recommendations, along with other innovations, have been implemented by a variety of Mediterranean survey projects over the last twenty years, all of them primarily oriented toward understanding past demography, economy, and political organization. Intensive surface exploration has now become as common as excavation in the Aegean. A large number of the first generation of regional surveys that built upon the foundations laid by MME has been completed with results published (Boeotia [Bintliff and Snodgrass 1985], Berbati-Limnes [Wells and Runnels 1996] Northern Keos [Cherry, Davis, and Mantzourani 1991], Laconia [Cavanagh et al. 1996; 2002], Methana [Mee and Forbes 1997], Nemea [Wright et al. 1990], Pylos [Davis et al. 1997; Davis 1998], Southern Argolid [Jameson et al. 1994; Van Andel and Runnels 1987]).

Mediterranean Archaeological Landscapes: Current Issues Following two decades of intensive archaeological survey in the Mediterranean, it is possible to recognize several issues that require attention. The early extensive surveys in the Mediterranean comprised isolated studies with interests in particular chronological periods. In contrast, the “New Wave” (Cherry 1994) of intensive surveys of the 1980s and 1990s have produced a rich diachronic record of the prehistoric as well as the historic countryside, meaning that issues of change and stability over the medium and long term can be addressed. How to render and interpret that change and stability is an important matter. Athanassopoulos (this volume) suggests that the Annales framework may be usefully employed in this capacity, given its emphasis on the articulation of the general and the particular and the linking of different time scales into an integrated narrative. A second significant contribution is the incorporation of geomorphological studies, which have become a standard component of regional field surveys around the world, albeit to meet a variety of goals. The diachronic perspective in combination with geomorphological study has focused attention on the links between changing landscape and changing human occupation. Questions regarding human-environmental interaction have led to the development of new perspectives. Thus, currently, changes in the environment are more likely to be attributed to anthropogenic factors rather than simply to climatic fluctuations (e.g., Van Andel and Runnels 1987, but see Rackham and Moody 1996). McGlade’s (1995; 1999a) eco-dynamics looks specifically at the convolution of anthropogenic and climatic forces as they play out in Mediterranean landscape contexts. Similarly, multi-disciplinary co-operation and increasingly sophisticated techniques (Gillings et al. 1999) have led to the creation of complex spatial databases, facilitated by innovations in GIS hardware and software, that can answer a broad range of demographic and environmental questions (Barker and Mattingly 1999).

MEDITERRANEAN LANDSCAPE ARCHAEOLOGY PAST AND PRESENT

Nevertheless, in several other areas field surveys have been less effective. In general, a significant portion of the work has been dedicated to the development of methods and research design, as well as to issues of site definition and collection strategies. Less explicit effort has been directed to interpretation. Social and economic inferences based on survey data tend to be limited and post hoc. Interpretive models remain general with little concern for specific chronological or regional contexts. The prevalent model identifies alternating phases of dispersed and nucleated settlement patterns (e.g., Bintliff 1999b). Invariably, dispersed patterns of settlement have been associated with small holdings, intensive forms of agriculture, and periods of economic expansion (e.g., Cherry et al. 1991; Halstead 1987; Runnels and Van Andel 1987). In contrast, nucleated patterns represent control of the land by an elite or the imposition of political control by an external power (e.g., Alcock 1993; Renfrew and Wagstaff 1982). This binary opposition of dispersed and nucleated patterns pays little attention to different land-use histories and imposes a uniformity in interpretation at the expense of regional diversity. Several chapters in this volume (Barton and colleagues, Kardulias and Yerkes, and Hill) consider similar patterns but with respect to other interpretative frameworks that are explicitly grounded in bodies of robust theory. To some extent these basic patterns may be an artifact of the methodology as well as the topography of the surveyed areas. For one, different surveys, with different conventions and varying survey intensities, report dramatically different numbers and qualities of archaeological sites (Cherry 1994; 2003). Secondly, the selection of regions in part determines the nature of the settlement record and even the date of the recorded remains. For example, survey in upland areas can produce meaningful results only if attention is paid to landscape taphonomy. Downslope movement, erosion, or terracing are all factors that complicate the picture, making the distinction between site and off-site material tenuous. On the other hand, survey in the lowlands where long-term settlement and intensive land use have been pervasive might produce an incomplete picture of earlier remains (Bintliff et al. 1999; Wilkinson, this volume), underscoring again the importance of geomorphological considerations seen in current Mediterranean work (Barton and colleagues, this volume; Doonan, this volume; Given, this volume; Gregory, this volume). Furthermore, implicit assumptions about “pottery rich” and “pottery poor” periods remain to be clarified. The supply and consumption of pottery and other artifacts is known to fluctuate significantly in different chronological periods (Millett 1991). Still, lack of readily identifiable material from particular periods is often taken at face value, leading to inferences about depopulation or abandonment of a region. Factors related to high or low ceramic visibility need to be addressed in order to interpret social and economic patterns that are period specific. After all, the “one size fits all” generalizing approach has resulted in an over-simplified interpretive scheme where very different social and economic structures may be homogenized. The model of alternating cycles of demographic-

5

6

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

economic growth and contraction pays little attention to social and historical context. Moreover, such analysis is not sensitive enough to issues of scale. Whether we are dealing with a prehistoric, small-scale society or a multi-regional empire ought to be important in interpretation since the social structures of production, control, and distribution are radically different in each case (Crumley and Marquardt 1987). Kardulias and Yerkes (this volume) incorporate scale in their study of the Malloura Valley and link local processes and supra-regional structures in dynamic relationships. A final important issue is the integration of the results. Field surveys started as low-budget operations of a small team of researchers and by now have grown to expensive multi-disciplinary efforts. The teams consist of a large number of specialists who operate in a setting similar to the “big-dig” model discussed earlier. The danger here is that the “big dig” approach will also prevail in the publication of the results, with emphasis placed on the organization and presentation of massive amounts of data generated by multi-year field efforts. Thus, integration of the varied databases and data sources should be on the top of the agenda of the next generation of surveys. Doonan (this volume) and Athanassopoulos (this volume) explicitly consider the integration of survey results in the spatial and temporal domains, respectively.

The Broader Picture Other issues become apparent when Mediterranean archaeological landscape studies are considered within the context of the grand sweep of archaeological history. Pervasive changes in Anglo-American archaeology over the last four decades have affected all aspects of the paradigm within which archaeological landscapes have been approached, from metaphysical principles, to theoretically based interpretative frameworks, to methods and techniques. As noted above, current archaeological landscape studies in the Mediterranean are in part a continuation of earlier local practices (Cherry 2003). More recent studies also reflect the evolution of landscape studies within archaeology, themselves an outgrowth of two related trends. The first is the early 1900s emphasis on geographical method and interpretation in Britain seen in O. G. S. Crawford’s (1953) work, later evolving into a British emphasis on environmental and economic archaeology seen in the work of Vita-Finzi and Higgs (1970). Second is the scientific humanism (Fotiadis 1995) of the post-WWII era, expressed as environmental functionalism in Anglo-Americanist archaeology (Trigger 1989) and made operational in the 1950s New World settlement archaeology (e.g., Virú valley project in Peru [Willey 1953]; Lower Mississippi valley survey [Phillips et al. 1951]). The stimulus for the latter was provided by Steward’s cultural ecological studies in the American Great Basin (Steward 1937; 1938), although it rapidly assumed its own dynamic.

MEDITERRANEAN LANDSCAPE ARCHAEOLOGY PAST AND PRESENT

Importantly, in settlement archaeology, individual sites no longer are viewed as representative of a culture. Rather, sites are considered part of a larger integrated network and play diverse and complementary roles. Moreover, settlement patterns represent a powerful abstraction of reality that allows for the integration of complex detailed information (Fish 1999), thus allowing the application of demographic, cultural ecological, political, and economic geographic principles to archaeological data (Trigger 1989). Changes in the settlement patterns were viewed in terms of internal transformations, sometimes driven by population growth. Settlement patterns, thus, became fundamental in the study of social, economic, and political transformations over the long term (Trigger 1989:284). Regional survey is the primary vehicle for obtaining a portrait of regional settlement patterns (Fish 1999). After some delay, survey projects based on the Virú valley model multiplied rapidly in Latin America, Mesopotamia, and the Mediterranean, as described above. Emphasis was placed on systematic extensive surface survey of regions ranging from several hundred to several thousand square miles. The goal was to define the extent of “the settlement system,” delineate broad problems, formulate hypotheses regarding site function, demography, land use, and polity that could be tested and refined through subsequent intensified investigation (Parsons 1972:133–34). Under the influence of the New Archaeology, problemoriented, processual questions encouraged the adoption of a systemic framework and the study of past human societies within ecological and economic paradigms. Scientific positivism placed emphasis on research design, analytical models, and explanation (Watson et al. 1971). Regional research goals promoted the use of locational models and quantitative forms of spatial analysis. Techniques for the reconstruction of demographic trends and population dynamics were incorporated into the research design. Many methodological issues, especially sampling strategies, became the subject of lengthy polemical debates whose spirit was captured in Flannery’s famous parody (Flannery 1976). Overall, Americanist New Archaeology contributed significantly to the refinement of field survey methodology. It offered the tools to investigate the material record of all social groups including the archaeologically invisible, the rural population, that earlier elite-centered archaeological research had ignored. However, in general its explanatory framework placed emphasis on change as adaptive response where humans fulfill system needs and react to environmental stimuli. It did not leave room for human agency and, although it “sought to locate the peasant farmstead and to produce accounts of daily life, the epistemological framework of New Archaeology in effect also denied history to people in the past” (Moreland 1991:14). The New Archaeology was soon challenged by other charges of environmental determinism and universalism as well as for its de facto emphasis on the economic and ecological (Trigger 1989). Since the early 1980s, numerous critiques of the New Archaeology have resulted in studies that emphasize other aspects of the archaeological record and that explore other theoretical approaches as evidenced by the expansion of “settlement pattern studies” to “landscape studies.”

7

8

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Current Mediterranean landscape studies also differentially reflect changes in archaeological epistemology. The avowed scientific positivism explicitly embraced by the early New Archaeology during the 1960s and 1970s has been supplanted by a mitigated objectivism (Wylie 1989a; 1989b; 1992; 1995) in which knowledge about the past is constituted in terms of independently and provisionally substantiated inferential tools. Understanding emerges as these tools are applied and the hermeneutic spiral is traversed (Hodder 1999). In classic settlement archaeology applications, two sorts of inferences are needed: one to establish settlement role and the other to establish settlement occupation time (Binford 1992). Settlement role is often approached in terms of settlement size as measured by site size, made operational in terms of area of distribution, mound size, or relative frequencies of ceramic sherds. For example, Sanders and colleagues (Sanders 1965; Sanders et al. 1979) explicitly ground these middle-range tools in ancillary studies that relate settlement size to these proxy measures. To establish time of occupation, ceramic sherd presence (and, recently, absence; Dewar 1991) is used. All sites with sherds from a particular time period are then assumed to be contemporaneous and the settlement pattern for that particular time period can thus be reconstructed. These settlement inferential tools, however, do not consider the many formation processes besides settlement that influence area of distribution, mound size and ceramic sherd frequency. Nor do they substantially engage the assumption of contemporaneity (Chapman 1999; Wandsnider 2004). In part because of the poverty of available inferential tools, settlement archaeology is no longer the only approach to the archaeological landscape. Finally, just as we have seen change in the frameworks and inferential tools available in archaeology, the conceptualization of the archaeological record has changed radically. In the late 1800s, it was construed as a powerful testimony to human achievement, demonstrating progress and sometimes regression (Daniel 1975). At least in the Americas in the early 1900s, it is referred to as a degraded residual of the rich, textured past (Dixon 1913). Because of the multiple roles that material culture plays in social, political, and ideational domains (as articulated by the New Archaeologists) and because of its reflexive nature (as elaborated upon by workers in postprocessual archaeology), the archaeological record is now perceived of as a rich record of past conditions and of human agency. Beyond monuments, the information potential of mundane cultural phenomena such as domestic architecture, terraces, and ceramic distributions as well as of geological deposits has been recognized. Moreover, formation process research (summarized in Schiffer 1987) has gone far to articulate the parameters of material history, which records events and processes in time very differently than do oral or textual histories. In spite of this more expansive understanding of the archaeological record, it remains for scholarship to address the integration of diverse material records, textual records, and ethnographic records, each with their very different strengths and weaknesses.

MEDITERRANEAN LANDSCAPE ARCHAEOLOGY PAST AND PRESENT

Approaches to Mediterranean Archaeological Landscapes These interpretative elaborations and epistemological and ontological shifts are reflected in the evolving conduct of landscape archaeological studies in the Mediterranean and elsewhere. In the new millennium, the windfall of postmodernism, the critique of the New Archaeology, and the ongoing critique of that critique (!) have resulted in a healthy diversity of archaeological approaches, as recently surveyed in Americanist archaeology (Hegmon 2003). Reflecting this diversity as well as shifts in epistemology and the understanding of the archaeological record, we see at least four rather different approaches to the archaeological landscape, each depending to varying degrees on surface survey. The first is the modified continuation of environmental functionalist settlement pattern study already discussed above. Again, such studies, which rely heavily upon regional surface survey, further the goal of demographic, economic, or political reconstruction (Billman and Feinman 1999; Bintliff and Sbonias 1999; Fish and Kowalewski 1990). Current settlement pattern approaches are technically very sophisticated and pay attention to sampling issues and the impact of geomorphology on site preservation (e.g., Bintliff et al. 1999; Wells 2001) and landscape formation (Barton et al. 1999). They incorporate remote sensing and GPS technology to locate and describe settlements, and geographic information systems technology is used to represent and analyze settlement distributions. In their study of the Burgundian landscape of France, Crumley and Marquardt (1987) also considered issues of scale, region, and boundary and how these changed through time. In the Mediterranean, settlement pattern studies abound as evidenced in the recent Mediterranean Landscapes series (Barker and Mattingly 1999). Second is the approach that emphasizes human agency and the reflexivity of landscape monuments, which so far has been little explored in Mediterranean landscape studies. In this approach, landscape is not merely the external world we see but a construction, a way of seeing the world. It is closely linked with broader historical structures and processes and represents a historically specific way of experiencing the world developed by and meaningful to certain social groups (Anshuetz et al. 2001; Cosgrove 1984:13–15; Knapp and Ashmore 1999). While ecological approaches treat the landscape as an object external to perception, social and symbolic approaches view it as a social construct, the expression of ideas that the analyst must attempt to understand. Several studies (Ashmore and Knapp 1999; Bender 1993; Bradley 1998; Tilley 1994) exemplify this shift in approach. These studies offer a contextual analysis of monuments, natural places, and historical landscapes (Leone 1988; Kealhofer 1999; Yamin and Metheny 1996). In the Mediterranean, Alcock’s analysis of the sacred landscape (Alcock and Osborne 1994) and Chapman’s (1994) analysis of recent Balkan monuments are two such examples. In both of these cases, the focus is on monuments and little use is made of survey data per se. Given (this volume) attempts to correct this situation.

9

10

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

A third approach focuses on the archaeological landscape, its taphonomy and evolution (McGlade 1995; 1999a; 1999b; Rossignol and Wandsnider 1992; Wandsnider 1998) over the medium and long term. Settlement archaeologists emphasize the geomorphological aspect of landscape taphonomy and how this impacts settlement pattern reconstruction and thus functional interpretations of demographic, economic, and political processes. The aspect of landscape taphonomy we aim to highlight here, however, incorporates a more broad definition of taphonomy (à la Dibble et al. 1997; see Barton and colleagues, this volume, and Wilkinson, this volume), one that focuses on the evolution of the landscape as it participates in necessarily interlinked cultural and natural processes. This approach sees the past knowable in terms of a constellation of conjoined processes operating with different temporalities; the material histories of artifacts, monuments, and surfaces offer a means to learn about these conjoined processes (Wandsnider 2004). Such an approach relies on survey as well as excavation data. Other than McGlade’s work in Spain (1995; 1999a) and recent work in southern Turkey by the Rough Cilicia team, this approach is little evidenced in the Mediterranean. Wandsnider (this volume) deals with the metaphysics behind this approach and Hill (this volume) offers a case study. Finally, attempts to relate landscape monument construction to evolutionary principles, e.g., in terms of costly signaling (Neiman 1997) or cultural elaboration (Dunnell 1999), or to institutional theory (Earle et al. 1998) have been little explored in the Mediterranean. Arguing from a group selectionist stance, Sterling (1999) sees Early Dynastic pyramid construction as a “wasteful” activity that depressed population levels to a point below which they were significantly affected by unpredictable Nile flood events, resulting in overall increased reproductive fitness. Such an approach could be tied into topographic survey of monuments in the Mediterranean, but to date has not.

Mediterranean Archaeological Landscapes: The Volume From the 1960s onward, Mediterranean archaeological practice increasingly reflects an awareness and incorporation of perspectives emanating from anthropological archaeology as practiced in Britain and the Americas. Current practice in the Mediterranean, while certainly not uniform, tends to emphasize intensive survey for the location of settlements and the reconstruction of settlement patterns. Interpretation of past social, economic, and political processes follows, sometimes rather intuitively as noted above, from the portrait of the settlement system yielded by survey. Because of shifts in metaphysical, epistemological, and ontological domains, we see modification and enhancement to the regional studies or settlement approach to Mediterranean archaeological landscapes, but also the appearance of other approaches. With this volume, we explore issues fundamental to the study of the Mediterranean landscape as revealed through systematic survey. As well, interpretative

MEDITERRANEAN LANDSCAPE ARCHAEOLOGY PAST AND PRESENT

studies of the archaeological landscape that reflect enhanced settlement studies and other approaches are presented. Cross-cutting and recurring methodological themes center on matters of temporal and spatial scale, both of data collection and interpretation, and on unit design. In their discussions, authors consider the relationship among and between the information provided by archaeological landscape documentation, excavation, historical texts and the varying temporalities of these different sources of data. They also reflect on the units used for describing and interpreting past landscapes, which in the past have been derived from anthropology, geography, and history. They consider the importance of the diachronic perspective, the role of geomorphological studies, the meaning of sherd presence/absence, and the nature of integrating multiple records. In that the larger questions addressed by Mediterranean researchers entail comparative analysis, another recurring theme is that of data quality and comparability. Finally, with this volume, we do not attempt to cover the full geographical breadth of the Mediterranean. Rather, we emphasize particular methods and interpretations in the context of specific archaeological Mediterranean landscapes. The case studies involve data sets from Prehistoric as well as Historic periods. Regarding the latter, our goal is to focus on “non-traditional,” “later” eras (Medieval, Ottoman, Modern) and address the complexities of archaeological fieldwork and interpretation in textually rich contexts (Athanassopoulos, this volume; Diacopoulos, this volume; Given, this volume).

Fundamental Issues In the first part of the volume, methodological issues fundamental to Mediterranean landscape studies are considered. The control of time in surface contexts is essential for the diachronic approach to be successful. Gregory addresses the issue of the temporal quality of ceramic data, the backbone of survey analysis and interpretation in the Mediterranean. He introduces ChronoTypes as methodological units that necessarily affect interpretation. Gregory also proposes a low-impact scheme that preserves archaeological artifacts in their context. Landscape surfaces and subsurfaces cannot all be documented at the same level of intensity. Hence, integration of results across different spatial scales is critical. Here, Doonan discusses issues of sampling design and nested survey in the diverse environment of the Black Sea region of the port of Sinop. Regional survey here has employed a combination of general and intensive strategies in conjunction with excavation and systematic underwater survey to study the hinterland of the most important port on the south shore of the Black Sea from the mid-first millennium BC to the first millennium AD. The next chapter by Wilkinson highlights another fundamental issue, that of landscape taphonomy. Through an examination of the imbalance and lack of comparability between Near Eastern and Mediterranean surveys, Wilkinson considers landscape taphonomy and its relation to settlement pattern reconstruc-

11

12

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

tion. To a large extent this disparity seems to be a result of different regional research traditions but also reflects different aspects of landscape taphonomy, discussed further by Barton and colleagues and Hill in their chapters. Building on the issue of ceramic presence/absence, Wandsnider considers the nature of surface remains and archaeological interpretation in light of the palimpsest nature of archaeological landscapes. This chapter considers the formation of site and non-site deposits, arguing that their histories are equally problematic, but also temporally variable and rich. Finally, Athanassopoulos discusses the temporal integration of survey and other data sources. She considers the relation between textual sources and regional survey data and examines the utility of integrative schemes like the Annales that recognize the varying temporalities of these diverse databases. This chapter delineates patterns of socioeconomic change in the Medieval period.

Interpreting Mediterranean Archaeological Landscapes The second portion of the volume offers interpretations of specific Mediterranean archaeological landscapes that transcend standard settlement approaches. Barton, Hill, and Kardulias and Yerkes all present enhanced settlement approaches that are theoretically well situated; sophisticated interpretation is enabled because the theoretical framework is explicit and well tied to field procedures. Given explores the meaning of the ceramic landscape while Diacopoulos looks at modern landscape phenomena. In all of these case studies, matters discussed at length in the Mediterranean Landscapes series (Barker and Mattingly 1999)—geomorphology, data cataloging, GIS applications—are of necessity incorporated, reflecting the state of the art in Mediterranean studies. Expanding upon other work (Barton et al. 1999) to survey the neolithization process in several valleys, Barton and colleagues emphasize the dynamic relationship between humans and their landscapes and examine the spatio/temporal dynamics of human ecology in five valleys of Mediterranean Spain during the first half of the Holocene. Integrating results from multiple seasons of study, they find that there is considerable variation across small distances and suggest that neolithization, even at this small regional scale, was not a uniform process. Using similar ecological concepts, but with an explicit concern for temporal scales of analysis, Hill uses settlement data spanning Holocene agro-pastoralism in the Wadi-al-Hasa, west-central Jordan. This study compares and contrasts settlement histories of adjacent ecological zones and evaluates different models of settlement change in conjunction with ecological and culture history. Hill argues that different theoretical approaches are suited to interpret data with different temporal scales of analysis. Kardulias and Yerkes explore the utility of world-systems theory as an interpretive framework that cross-cuts economic, political, and social dimensions. Within

MEDITERRANEAN LANDSCAPE ARCHAEOLOGY PAST AND PRESENT

this framework, the authors examine the settlement dynamics of the Malloura Valley in central Cyprus as a case study. Given identifies a disjunction between methods and interpretation in Mediterranean regional surveys. Rich survey data sets are rarely utilized to answer questions as to how people perceived the landscapes in which they lived and worked. He argues that a phenomenological approach based on artifact distribution patterns is possible and long overdue. Finally, Diacopoulos adds another dimension to Mediterranean landscape studies, that of the present cultural landscape. She argues for the expansion of current survey techniques and the need to incorporate the Modern periods in regional archaeological studies. Her study emphasizes the necessity of considering the human aspect of the present cultural landscape, including indigenous perceptions of heritage, history, and national identity. Cherry (2003) opines that this generation may be the last to effectively document the rich Mediterranean surface record, which is in jeopardy as development continues apace. As the Mediterranean landscape evolves, reflecting the values and impact of the modern world, our efforts to elicit meaning and make sense of this rich heritage must intensify. Our labors today, both in the field and in interpretation, greatly affect future research in the kinds of questions that can be posed and the degree to which they can be answered. We hope that the issues addressed in this volume will further the development of effective research strategies in Mediterranean landscape archaeology.

13

This page intentionally left blank

2 Less is Better: The Quality of Ceramic Evidence from Archaeological Survey and Practical Proposals for Low-Impact Survey in a Mediterranean Context Timothy E. Gregory

S

ampling, recording, and collecting strategies stand at the heart of any archaeological survey project. All analysis is based on the collected data, and those, in turn, depend on the strategies set out at the beginning of the project. This much is obvious, and most of the recent criticism of archaeological survey has been directed at perceived weakness in these techniques. Space does not allow a full consideration of scholarly critiques of archaeological survey, but it is fair to say that current opinion tends in the direction of recommending greater, rather than lesser, collection of data—and that often means a 100% surface coverage of an area and 100% counting and/or collection of artifacts, based on the argument that to do otherwise will bias the evidence and lead, in particular, to the under-representation of certain periods in the past. Serious question of sampling in survey method was first raised by Ammerman (1981; 1995) and then by Fish and Kowalewski (1990) and Cowgill (1990), and, although Plog (1990) expressed his misgivings about the abandonment of sampling strategies, the direction in archaeology over the past ten years has definitely been in the direction of “fullcoverage survey” (e.g., Bintliff et al. 1999, discussed below). Further, within such an intensive regime, 100% collection of artifacts has often been advocated (i.e., not only 100% counting of artifacts, but also the complete collection and archiving of all archaeological objects!).

A Conceptual Framework for Low-Impact Survey: The ChronoType System This chapter does not directly address the broader issue of sampling, although the author believes that sampling is inherent in all archaeological work and that we cannot hide that fact by a false belief that “100% coverage” avoids it. Rather, this issue—in survey at least—always is a matter of intensity of coverage, which can be defined and measured in a wide variety of ways. While ignoring the broader question of sampling, this chapter will argue that low-impact methods of recording and collection can be devised that will yield sound results, at a fraction of the cost

16

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 2.1. Sydney Cyprus Survey Project: study area (from Given and Knapp 2003: Fig. 1.1).

and with greater attention to the integrity of the archaeological material, for the purpose of resurvey or re-analysis. The consideration of such low-impact techniques is derived from several sources. First, on the basis of ethical concerns, we maintain that survey archaeology ought to impact the study area as little as possible, thus allowing only a minimum disturbance to the surface and the artifacts lying on it; secondly, we see such an approach as the best way to maximize time and resources (including financial), especially in situations of rapid development, where survey must be carried out quickly; additionally, we argue that low-impact survey is most appropriate in the first phase of survey, where later, more detailed analyses of individual places are planned and where the material should be left as undisturbed as possible; finally, especially in the Mediterranean area but probably also in many other parts of the world, governmental and grass-roots organizations are certain to put greater and greater limitations on the degree of intervention archaeologists may make on the landscape. Instead of bemoaning this trend, we argue that archaeologists ought to embrace this opportunity and develop the techniques that will allow them to work constructively with local officials in the preservation of archaeological heritage. It should be pointed out that the suggestions made in this chapter are presented as an interrelated whole, since they have emerged in specific fieldwork environments. They are, however, separable and flexible, and one may divide what

THE QUALITY OF CERAMIC EVIDENCE FROM ARCHAEOLOGICAL SURVEY

Figure 2.2. Eastern Korinthia Archaeological Survey: study area.

is essentially a taxonomic system, designed to impose systematic nomenclature and maintain chronological control (as much as one may on the basis of artifact dating), from the implementation of that system in specific fieldwork situations. The chapter certainly does not advocate a “one-size-fits-all” approach, but realizes that different projects will need to select approaches that are appropriate for their own situation. Further, the suggestions made here were, for the most part, derived from surveys in “artifact-rich” environments, where a regime of 100%-retrieval would have been completely beyond our ability to implement, and the suggestions, especially for fieldwork, may be most appropriate to such situations. Furthermore, the techniques presented here grew out of specific situations where limitations were put on the collection of artifacts, but our position is that many of these techniques should be applicable in other areas where such constraints do not exist. Decisions on this, of course, are matters for individual projects to consider, but we would argue that procedures modeled in this direction should be considered in various places. Overall, these suggestions resulted from experience with the Sydney Cyprus Survey Project (SCSP, Fig. 2.1), the Eastern Korinthia Archaeological Survey (EKAS, Fig. 2.2), and the Australian Paliochora-Kythera Archaeological Survey

17

18

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 2.3. Australian Paliochora-Kythera Archaeological Survey: study area.

(APKAS, Fig. 2.3), as well as years of dealing with pottery in many surveys, where I have been able to see first-hand what actually happens to the finds that are brought in for analysis by the ceramic experts. Much of what follows was developed within the framework of SCSP, between 1995 and 1997, in collaboration with Nathan Meyer. A. Bernard Knapp, Director of SCSP, provided the setting and the encouragement for the development of these ideas and R. Scott Moore provided valuable reaction and insight. The conceptualization described here was a fully joint product of Meyer and myself; its first presentation in print has recently appeared in the final publication of SCSP (Given and Knapp 2003:14–16). In subsequent years the systems have been modified and field-tested in two newer projects, EKAS and APKAS, and changes made to fit the situations there; the differences between what is presented here and the final publication of SCSP are therefore largely the result of different situations rather than any disagreement with final SCSP procedures. In particular, the methods discussed here have been influenced by a permit situation in which one project (EKAS) has essentially been prevented from collecting artifacts in “off-site” areas, so the methods were in part a response to necessity. But I would argue that these methods, while perhaps not

THE QUALITY OF CERAMIC EVIDENCE FROM ARCHAEOLOGICAL SURVEY

appropriate for all survey projects, are effective in their own right and they have elements that can be considered by any project. The preconditions for our method are, first, good spatial control (based on GIS and printed aerial photographs and topographic maps), so that field teams know precisely where they are, and, second, good geomorphological control, so as to avoid (as much as possible) the mixing of artifacts that come from different geomorphological contexts. Among the results of this method is the creation of relatively small survey units: the mean size of units from EKAS 2000 was 3,242 m2, a plot approximately 57 m on a side. Third, and most importantly, we made use of a tightly organized taxonomy for artifacts—the so-called ChronoType (CT) system, pioneered from 1995 onward by SCSP (Given et al. 1999). The name ChronoType is derived from the concept of the “fossil-type” used originally in geology and subsequently in archaeology. The suffix “chrono-” was added to create a neologism and to emphasize the concept that individual identifications had to carry with them, first and foremost, some chronological definition (however broad), since chronology is the first and most important deliminator in survey (or any) archaeology. Looked at from one point of view, the ChronoType system is nothing more than a systematization of naming systems used by ceramic (and other) specialists today. It profits from suggestions made by many ceramicists dealing with large quantities of excavation material (Byrd and Owens 1997; Slane 1987; Tomber 1990), although in most instances they suggest methods appropriate for one project rather than something more broadly applicable. The taxonometric aspect of the CT system, by contrast, is conceived explicitly to facilitate comparison of data across projects as well as to encourage standardization and data rigor within a given project. In addition, through a system of aliases or alternate names, it allows specialists to make use of and look up alternate names for objects that may be more familiar to them. The CT naming system is, as much as possible, based on standard terms and definitions of artifacts used by specialists in different periods. This means that the system is flexible, open-ended, and (except at the lowest level of distinction) not based on a specific definition of “wares” that may be appropriate for one period but not for others. In addition, the system allows multiple identifications, so that a scholar should be able to search the artifact database using different definitions. Thus, it should not matter for analysis if an amphora is identified by Dressel type, Bernice type, Williams and Peacock number, or by any other system: the database is set up to recognize these identities. The CT naming system can therefore afford to be eclectic and flexible, accommodating terms used by different specialists for what is essentially the same kind of artifact (although, naturally, there is one “official” CT name for each kind of artifact). Although, for convenience, the term “ware” is used in many CT names, the concept is not basic to CT definitions since “ware” is used differently by scholars working in different periods: in some cases an individual CT may be

19

20

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

what most scholars call a “ware,” but in other cases it may be defined by decoration (e.g., types of transfer prints on the same “ware,” types of combing on LR2 amphorae) or some other characteristic that has chronological significance. The CT system is based on what one may call an interlocking web of hierarchically based dimensions:1 1. Dimension 1: Class (monitors general artifact morphology) 2. Dimension 2: Subclass (monitors function, mostly) 3. Dimension 3: Subdivision (monitors chronology, the time the artifact was manufactured/used, either by period or by specific dates) 4. Dimension 4: Stylistic elements (e.g., slip, form, design elements) The goal of the system is to allow any artifact to be assigned to specific categories within this structure. The system is rigidly hierarchical and any object must fall in one or another of these categories, based on its characteristics and the knowledge of the scholar making the designation. Another way of looking at this is to examine these categories of characteristics, from the most general to the most specific, arranged as follows: Class: e.g., Pottery Subclass: e.g., Fineware Subdivision: e.g., Roman Fineware Type: e.g., Roman Red Slips Subtype: e.g., African Red Slip Form: e.g., Form 50 Subform: e.g., Form 50B The Classes of materials (i.e., the grossest subdivision within the system) are, at present, the following: 1. Pottery (including tile) 2. Stone/lithics 3. Metal 4. Glass 5. Terra-cotta (non-“pottery”) 6. Bone/shell 7. Other The Subclasses of pottery are: 1. Coarse 2. Medium coarse 3. Amphora/jar 4. Fine 5. Semi-fine 6. Kitchen or cooking 7. Lamp 8. Pithos 9. Tile

THE QUALITY OF CERAMIC EVIDENCE FROM ARCHAEOLOGICAL SURVEY

The Subclasses are meant to be readily identifiable to the trained observer, but they are not based exclusively or even primarily on the physical characteristics of the clay (i.e., how coarse or fine it is), but rather they carry imbedded in them an element derived from what we may call an “assumed function.” In practical terms, for pottery, this means that elements such as surface treatment, wall thickness, and module play a role, along with the size of inclusions, in determining assignment to the Subclass category. Thus, fine pottery is defined as relatively expensive tableware and it would include Medieval glazed plates and bowls, which were manufactured in a fairly coarse fabric, as well as Classical Black Glaze and Modern transfer-print plates. Kitchen/cooking pottery is defined as made with a very stony, coarse (usually red), and very friable fabric, used in Antiquity and the Middle Ages for making cooking pots and similar vessels.2 Semi-fine pottery is used to designate what in Greece is often called “Plain Ware,” i.e., unslipped/unglazed pottery that would have been appropriate for table use—it may be restricted to the ancient period (and is very different from what one means by “Plain Ware” in Cyprus and elsewhere). Pithos is distinguished by its large size, the thickness of fabric, and other characteristics, while tile embraces both ceramic rooftiles and bricks, and ceramic lamps are self-explanatory. Parenthetically, the careful reader will note that tiles and ceramic lamps might more precisely be grouped under “terra-cotta” than “pottery,” and indeed in EKAS this was done. The ChronoType nomenclature system allows such changes, and comparisons of data among projects can easily be accomplished with the appropriate database massaging. The hierarchical character of the CT system is designed, in part, to respond to the reality of the imperfection of our knowledge, but to maximize the information that can be derived from an analysis of an object. Thus, an artifact specialist may, in fact, be unable to determine the precise identification and chronology of an object, either because it lies outside his/her area of specialization or because the object itself is in a fragmentary condition that makes precise identification impossible. For example, a given sherd might be identified as either an ARS Form 104 or 106, and this can be written on a sheet of paper by the specialist, but this is difficult for analysis, especially with electronic means, and a less experienced expert might simply identify the sherd as ARS or even as a Roman redware. What we need is a system that will extract as much information as the specialist is able to provide, while realizing that there will be—in our imperfect world—lapses in knowledge. The CT system allows the specialist to identify this hypothetical sherd as either ARS 104 or 106, or as “ARS 104/106,” or as “ARS,” or as “Red-slip, Roman.” Each of these ChronoTypes is distinct and each carries with it the appropriate identifiers of class, subclass, subdivision, etc., and a chronological range. Naturally, the more precisely the object can be identified, the narrower the range of date that can be supplied. For broad ChronoTypes, which characterize the great bulk of material derived from survey, the chronological range is necessarily wide. Thus, we have a ChronoType for each of the subclasses of pottery that is, in fact, chronologically unknown (i.e., defined by us as “Ceramic Age,” from Middle

21

22

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 2.4. Example of ChronoType hierarchy.

Neolithic onward). Nonetheless, even the rude bits of coarseware that one often finds can often be roughly ChronoTyped, e.g., “Coarse Ware, Post-Prehistoric,” or “Coarse Ware, Medieval-Modern.” Certainly, this information will not be very specific, but it can be nonetheless important in our attempt to understand the chronology, as well as the functional attributes, we wish to assign to the areas where survey is carried out. In addition, this strengthens our ability to make negative statements, such as “no pottery identified as prehistoric has been found” in a certain area. Figure 2.4 provides an example of hierarchy for an actual ChronoType; it is based on the realization that a given object may well “actually” have been ARS Form 50B, but that, for one reason or another, it can only be identified as something further “up” the hierarchy. All of the categories are, of course, expandable—theoretically infinitely— although there is a real benefit to keeping the number of broader categories relatively small, while the finer categories (more precise ChronoTypes) will expand considerably. One significant advantage of the CT system is its flexibility in the face of continued growth of knowledge. Thus, if a specialist, at a later time, is able to provide greater precision about individual objects (i.e., to assign objects, or a whole

THE QUALITY OF CERAMIC EVIDENCE FROM ARCHAEOLOGICAL SURVEY

group, to a more precise ChronoType, either one that already existed or a newly developed CT), that is easily done in the appropriate database, and all dates and analyses based on those objects are automatically updated and changed. Likewise, if scholarly consensus calls for the re-dating of certain classes of objects, it is simple to change the chronology assigned to a certain ChronoType, and all the objects are then automatically re-dated, along with analyses based upon those dates. Ultimately, our goal is to move away from our present practice of recording sherds according to standard periods and the assignment of these periods to individual ChronoTypes. Our plan it to replace the periods by simple date ranges (presumably using BP dates), thus allowing considerably greater flexibility in our ability to create different chronologies for analysis (looking at 100-year increments, for example, rather than being limited by the standard periods currently in use). This may also allow us to reduce the number of ChronoTypes in the system, although the ramifications of this have yet to be fully worked out. The difficulty in implementing this change, however, is a serious disagreement among ceramic specialists (in particular) concerning absolute chronology, especially for the Prehistoric period. Thus, we are, at present, still working with a relatively complex and cumbersome periodization scheme, and one that has to be modified somewhat from region to region, because of differences in chronological schemes. Figure 2.5 shows the periods currently being used in EKAS.3 Note that the periods are, themselves, hierarchically arranged (although this isn’t shown here), commonly with two or more levels of subordination below the main categories, and considerable (necessary) overlap. As mentioned, the number of individual CTs is theoretically unlimited. At present the number of CTs being used for analysis in EKAS is 966, a considerably higher number than the 270 used in SCPS (even though many of the Cypriotspecific CTs have been essentially eliminated for work on the Greek surveys). This is, simply, a result of the longer time the CT system has now been used in the Greek surveys and the greater range of ceramic expertise that has been applied in the definition of the system. One expects that continued study of EKAS/APKAS material will result in the multiplication of CTs defined and utilized in these projects. The result of this system is that the artifact specialist need only record a minimum of information about a specific object, or rather a group or “batch” of objects that are the same CT and that are otherwise indistinguishable for the purpose of analysis. Thus, our practice has been to lump together objects that receive the same description into a category called a “batch,” so as to avoid recording repetitive data. Using pottery as an example, the specialist really needs to record only the name of the piece of the vessel represented (rim, handle, body sherd, etc.), the colour of the fabric, any surface treatment, and comments and observation, along with the ChronoType. The class, subclass, etc., the period, and the function—the most important variables for the analysis of artifacts—are all embedded in the CT and can be derived from them for purposes of analysis. In practice, however, the recording of redundant information, especially for characteristics such as color or the kind of

23

24

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

PERIOD

DATES

Palaeolithic Mesolithic Preceramic Age Ceramic Age Ancient Prehistoric Neolithic Early Neolithic Middle Neolithic Late Neolithic Final Neolithic Final Neolithic–Early Helladic I Bronze Age Minoan Early Bronze Age Early Helladic Early Minoan Minoan Pre-Palatial Early Helladic-Middle Helladic Early Helladic I Early Helladic I–II Early Helladic II Early Helladic III Middle Bronze Age Middle Helladic Middle Minoan Minoan Palatial Late Bronze Age Late Helladic Late Minoan Late Bronze Age–Classical Late Helladic–Modern Present Middle Minoan III–Late Minoan I Minoan Post-Palatial Middle Helladic–Late Helladic Middle Helladic–Late Helladic I Middle Helladic–Late Helladic III Late Helladic I–IIA Late Helladic II Late Helladic III Late Helladic IIIA Late Helladic IIIA-B Late Helladic IIIB Late Helladic IIIC

+8300 BC 8300–6800 BC +6300 6300 BC–AD 2000 6300 BC–AD 700 6300–1050 BC 6800–3200 BC 6800–5800 BC 5800–5300 BC 5300–4500 BC 4500–3100 BC 4500–2650 BC 3100–1050 BC 3100–1050 BC 3100–2050 BC 3100–2000 BC 3100–2050 BC 6800–2000 BC 3100–1680 BC 3100–2650 BC 3100–2150 BC 2650–2150 BC 2150–2000 BC 2000–1680 BC 2000–1680 BC 2050–1675 BC 2000–1470 BC 1680–1050 BC 1680–1050 BC 1680–1075 BC 1680–323 BC 1680 BC–AD 200 1750–1490 BC 1490–1000 BC 2000–1050 BC 2000–1580 BC 2000–1065 BC 1680–1500 BC 1600–1420 BC 1420–1065 BC 1420–1310 BC 1420–1190 BC 1310–1190 BC 1190–1065 BC

Figure 2.5. Chronological scheme for the Eastern Korinthia Archaeological Survey.

THE QUALITY OF CERAMIC EVIDENCE FROM ARCHAEOLOGICAL SURVEY

PERIOD

DATES

Late Helladic IIIC–Submycenaean Post-Prehistoric Ancient Historic Protogeometric–Hellenistic Protogeometric–Archaic Protogeometric Geometric Geometric–Archaic Archaic Archaic–Classical Archaic–Hellenistic Classical Classical–Hellenistic Hellenistic Hellenistic–Roman, Early Roman Roman, Early Roman, Late Roman–Medieval Roman–Medieval, Early Ancient–Medieval Ancient Historic–Medieval Roman–Modern Bronze Age–Roman Medieval Medieval, Early Medieval, Late Medieval, Ottoman/Venetian Medieval–Modern Modern Modern, Early Modern, Present Unknown

1190–1050 Bc 1050 BC–AD 2000 1050 BC–AD 700 1050–31 BC 1050–500 BC 1050–700 BC 800–700 BC 800–480 BC 700–480 BC 700–323 BC 700–31 BC 480–323 BC 480–31 BC 323–31 BC 323 BC–AD 250 31 BC–AD 700 31 BC–AD 250 AD 250–700 31 BC–AD 1800 31 BC–AD 1200 6800 BC–AD 1800 1050 BC–AD 1800 31 BC–AD 2000 3100 BC–AD 700 AD 700–1800 AD 700–1200 AD 1200–1537 AD 1537–1800 AD 700–2000 AD 1800–2000 AD 1800–1960 AD 1960–2000

Figure 2.5 cont’d.

inclusions, may be useful, especially in the definition of new CTs and the retrospective separation of material into new CTs. Projects using the CT system seek to publish one example of as many individual CTs as possible, with profile and photograph where appropriate. A plan is now underway to link the main CT database with representative photographs and profile drawings, in part to facilitate identifications and also to indicate clearly what is meant by individual CTs. A CT-based study collection of pottery and some other artifact classes has been set up at the Ohio State

25

26

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

University Excavations at Isthmia, and an online version of this is available at http://isthmia.osu.edu/study. Obviously, these photographs will not allow secure identification of all the CTs, and it is hoped that we will be able, in the reasonably near future, to implement a plan to prepare macroscopic photographs of the CTs, to begin to provide some petrographic information about them. It should be stressed, however, that (except in special cases) the CT system is designed only to separate out “wares” that are distinct as a result of visual examination (i.e., without the aid of a microscope).

The ChronoType System in Fieldwork, Collection, and Analysis Beyond its use as a descriptive and analytic tool, the ChronoType system has significant ramifications for fieldwork and the collection and examination of artifacts in the field. Naturally, as mentioned above, the CT system, as a taxonomic device, may be separated from any consideration of collection and field description. These latter have, however, been developed in SCSP, EKAS, and APKAS as part of the same integrated processing and analytic scheme, so they will be discussed here, although the present chapter does not assume their full integration will always be desirable or useful. As a first principle of any field technique, we may note that all methods of artifact recording in a survey context must strive to represent the totality of what is, in fact, on the surface of the ground in the area being investigated. All survey techniques strive to accomplish this, with greater or lesser success (Fentress 2000; Mattingly 2000; Given and Knapp 2003:9-13). In virtually all contexts this involves sampling. This sampling can take place at a variety of levels: (1) sampling at the level of the region, by deciding what blocks of land will be examined, (2) sampling at the level of the survey unit, determined by the intensity of observation (based in large part on the spacing of field-walkers) but also by the intensity of recording (determined largely by the kind of observations made and the degree of intensity with which artifacts are recorded, both of these usually, although not always, connected with the amount of time spent on a specific survey unit), and (3) sampling in terms of the number and type of artifacts that are described in the field and/or collected for later analysis and study. We would distinguish between the description of artifacts and their collection and removal from the field, since these are not at all the same thing. Most projects are aware of this, in the sense that many (probably most) artifacts on the surface in a survey area are, in fact, not seen at all by the team members (for various reasons including intensity of observation and visibility); others are seen and described in some way in the field, ranging from mere counting to complete infield description; and, all or a portion of those that are noted are removed from the field and taken for storage and (presumably) later study. Each of these levels of recording represents different kinds of sampling.

THE QUALITY OF CERAMIC EVIDENCE FROM ARCHAEOLOGICAL SURVEY

Thus, something approaching a 100% collection strategy is impossible on practical grounds in any but the most restricted areas, but, as said above, we think it is also inappropriate on ethical grounds, since this involves striping whole areas of their archaeological material, even if research suggests that the “archaeological surface” will tend to renew itself over time. The question, therefore, should not be whether we will sample, but what sort of strategy is most appropriate and effective for gathering information that will allow observations about the material on the ground, in both qualitative and quantitative terms. Despite the important comments of, for example, Bintliff et al. (1999:153–56), the strategy adopted by most surveys has been the identification and collection of “diagnostic” sherds, usually meaning “feature sherds” (rims, handles, bases) as well as those with specific kinds of surface decoration (especially ancient painted and slipped vessels).4 The difficulties with such a strategy are (1) that certain body sherds and pieces of coarse or cooking ware vessels will normally not be noted or collected, even though they often are, in fact, highly diagnostic, and (2) that such a method provides absolutely no indication of the quantity of artifacts of various types originally on the ground. Thus, the identification of two or three Late Helladic IIIC (LHIIIC) pottery sherds will probably suggest a LHIIIC “presence” in a specific location, but this will not provide any indication of the “strength” of the LHIIIC component nor any comparison between that and other periods that might also be present there. In fact, such an identification is little more than a “dot on the map,” a presence-absence piece of evidence that does not help as much as it could in any kind of more sophisticated analysis. The comments of Bintliff and colleagues (1999) are apropos here and reinforce the difficulty of the still-common distinction between “on-site” and “off-site” data. Thus, they point to the practice whereby Bintliff and Dickinson “revisited historical sites to enlarge the collection of a very sparse prehistoric component” by lying down on the ground and looking along the plane of the soil surface, “seeking the tell-tale ‘wet biscuit’ shapes and textures” of prehistoric pottery (Bintliff et al. 1999:158). Later they suggest that greater attention to coarsewares might result in “considerable revision…of present views on the location and density of population foci” (Bintliff et al. 1999:159) for the Sub-Mycenaean and Early Byzantine “Dark Ages.” While the techniques they suggest are potentially useful and certainly entertaining, we would suggest that the application of the ChronoType system may also contribute to our understanding of these difficult periods, in part because of the attention it pays to coarse wares and material that is not normally regarded as “diagnostic.” The ChronoType system, then, as a means to conceptualize and divide the universe of potential artifacts on the surface, can be used to devise an in-field as well as laboratory identification scheme, and a method for collecting artifacts for later analysis and refinement of identification. As mentioned, these procedures are based on the desire to record both the different kinds of artifacts in a survey unit and their approximate number originally on the surface without resorting to a

27

28

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

100% collection strategy—something which, to repeat, would be absolutely impossible in an artifact-rich environment. Briefly, our method of survey is typical for most projects of this kind. The crew is arrayed at a spacing of 10 m between walkers. Each walker is equipped with two tally counters (clickers), a compass, and a number of plastic zip-lock bags. Each walker observes the ground surface in a 2 m swath only. That is, the range of surface observation is 1 m to the walker’s right and 1 m to the left (meaning that we are examining a 20% sample within the survey units and that scatters less than 8 m in diameter could potentially be missed).5 All artifacts larger than a thumbnail that appear in the 2 m swath are counted, divided into categories that vary slightly between EKAS and APKAS, with APKAS providing more categories for this numerical record. In addition to counting, certain objects are retained in the plastic bags for analysis. The system for retaining these objects is conceptually based on the ChronoType system discussed above. Naturally, field-walkers cannot be expected to know all the different CTs and they do not have the time to distinguish them as they move across the landscape. Field-walkers were therefore instructed to “pick up” (we intentionally do not use the term “collect,” in part to reflect the distinctions between the Greek - and place in their bag every potential lithic fragment, term perisyllégo- and syllégo) coin, or other “rare” artifact, such as figurine or sculpture—unless these were extraordinarily numerous, in which case the team leader would determine a proper recording/collection strategy. For common objects—normally pottery, tile, and slag—team members were told to “pick up” one piece of each artifact type that appeared different from all the others they had observed. In practice this meant that team members picked up the first piece of pottery they observed and each piece thereafter that, in their judgment, was different. The axes of difference in this process are the following: fabric (color, texture, inclusions), surface treatment (slips, glazes, decorative elements), ceramic body part (rim, neck, neck/shoulder, handle, body, foot), and thickness (casually estimated). In theory then, the “bagging” of materials—by a team of, say, five field-walkers— from a unit could include up to five pieces that were for all practical purposes identical. In addition, since team members were told to “bag” pottery that was identical but preserved different parts of a vessel, a 5-person team could pick up 25 examples that represent the same ChronoType. One should remember that the team members were not expected to know the ChronoTypes, but since ChronoTypes are defined by visible differences, there will be a reasonable relationship between identification of ChronoTypes by artifact specialists and what was actually picked up. In addition, we told team members that, if they were uncertain, to pick the items up. Thus, team members almost certainly “erred” in the direction of picking up artifacts as “different” that the specialists would not put into separate ChronoTypes. Thus, the specialists commonly had to group together some artifacts that the team members viewed as different, because of our knowledge of the ceramics (or lack thereof). Our experience, however, was that team members rarely missed artifacts

THE QUALITY OF CERAMIC EVIDENCE FROM ARCHAEOLOGICAL SURVEY

that would, in fact, have been useful in the identification of periods or functions at individual locations (discussed below in the section on experiments and testing). This, of course, is the ultimate goal of the procedure—on the one hand to represent the kind of artifacts on the ground, with some indication of their original number, while on the other to avoid overwhelming the process with an attempt at 100% collection. The theoretical basis behind this is easy to comprehend. If, for example, a team of 5 people picked up one example of a certain CT in one survey unit, but 25 examples of the same CT in another unit, it would be fair to say that the CT was much more highly represented in the second unit (or at least that it was more highly represented in the sample of the unit seen by the members of the team). This can be expressed in more concrete terms by saying that the higher the number of artifacts picked up and assigned to a given ChronoType, the greater the number of sherds of that ChronoType that will likely have been, originally, on the surface—not accounting, of course, for the effect that soil color, visibility, and background disturbance may have had on the collection of sherds.6 Naturally, there is a limit to the number of pieces collected in this system (25 for 5 fieldwalkers) and therefore the procedure is “closed-ended” and will tend to underrepresent CTs with very large numbers—tiles of a certain type, for example. Nonetheless, experience suggests that for pottery (as opposed to tile) the maximum number of collected pieces was almost never reached and the phenomenon can probably be modelled using various statistical manipulations (such as the Pottery Index discussed below). Other projects naturally grapple with similar problems, as can be seen in the contribution by Barton and colleagues (this volume). Based on experimental analysis in which we modeled actual ChronoType sherd frequency vs. that measured using our data collection procedures in SCSP, Nathan Meyer derived the following relationships (Given and Knapp 2003:50–51, with changes in letter designations here [by the author] in the interest of clarity): A=c+v+b where: A = adjusted count of pottery in a unit c = count of pottery in a unit v = correction for visibility b = correction for background confusion Precisely how visibility and background confusion should be factored here was a matter of debate. Though SCSP certainly did not claim to solve this problem for all situations, the experimental data allowed an approach in the context of that project (reflected in the equation above), while this is something that has to be worked out for EKAS/APKAS (Meyer and Schon in Given and Knapp 2003:52–57). To return to the relationship between the artifacts picked up and those on the ground, consider a situation where, in a certain unit, there are 100 sherds actually on the surface divided into ChronoType X and ChronoType Y. If we discover 50

29

30

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

of these sherds in the ratio of 1:4 (10:40), the probability that the 51th sherd will be of ChronoType Y is higher than that it will be of ChronoType X. The actual probability would be complex to calculate and would need to take into consideration the relative obtrusiveness of the sherds, the degree of clustering, etc. Our goal, however, was not to increase the overall amount of pottery counted but to affect that number as little as possible while rearranging how individual ChronoType batches are weighted within that total. After doing various exploratory analyses of the data, we settled on a weighting of the counted ChronoTypes as follows (again, taken from Meyer in Given and Knapp 2003:50–52, with changes in letters by the author): W = c + ([c – ac]/10) where: W = weighted count of sherds in a ChronoType batch c = count of sherds in a ChronoType batch ac = average count of sherds per ChronoType batch The next step was to create an index that could be used across units of different sizes and displayed on maps in a way where differences could be discerned. In order to display the quantity of a given ChronoType in a unit we devised the following formula, where W is from the equation above: P = s(W/a) where: P = projected quantity of a ChronoType in a unit, adjusted for area a = area of the unit (from GIS determinations) s = a scaling number (10,000 was used) In SCSP the goal of this manipulation was the creation of a Pottery Index, which would allow calculation of a weighted number that represented the “strength” of various chronological components (and ultimately individual functional categories—although this was not done in SCSP). In EKAS/APKAS we have not arrived at a point where we can do this, but we think we have the data that will ultimately allow such reasonably sophisticated manipulation of the information partly on the basis of experiments carried out in those projects.

Data Processing and the Field Processing Team In part because the permit granted by the Greek government to EKAS has not allowed objects observed in DUs (Discovery Units or survey units) to be removed from the field for processing and analysis elsewhere, the laboratory has had to come to the field. The Field Processing Team, made up of at least one artifact specialist and two or three assistants, has the task of recording, photographing, and drawing objects picked up by the survey teams. The Processing Team analyzes the material in much the same way it would in a laboratory. The finds are divided up into “batches,” groups that are distinct along the axes of difference mentioned above (including different batches for handles, rims, etc.). Next, the artifacts are

THE QUALITY OF CERAMIC EVIDENCE FROM ARCHAEOLOGICAL SURVEY

identified by ChronoType and this information entered onto a recording sheet that will later be keyed into the finds database. In addition, selected artifacts are drawn and/or photographed using a digital camera. The selection of objects to be drawn and/or photographed depends on a variety of circumstances. On the one hand, this may include the “documentary” purposes of demonstrating through publication and other means the range and types of artifacts encountered at specific places and how we define the various artifact types. In this regard, the state of preservation of an object will commonly play an important role, since we want to select for illustration objects that are relatively well preserved so as to demonstrate effectively the characteristics of the given ChronoType. Artifacts illustrated for this reason will normally include those that members of the Processing Team feel quite confident in identifying. On the other hand, objects whose identification is questionable, for one reason or another, are more frequently documented with drawings and photographs, since it is expected that they will more likely be identified through consultation with other members of the project staff or, later, with other specialists. Since the cost of making drawings (which are later digitized) and digital photographs is relatively low, the tendency is to draw and photograph as many artifacts as time will allow, so as to

Figure 2.6. Processing team photograph (in field), EKAS.

Figure 2.7. Processing team photograph (in field), EKAS.

Figure 2.8. Processing team photograph (in field), APKA.

31

32

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 2.9. Processing team drawing (in field), EKAS.

Figure 2.10. Processing team drawing (in field), APKAS.

Figure 2.11. Processing team drawing (in field), EKAS.

make as complete as possible a collection of descriptive information for later study and analysis. Figures 2.6–2.11 provide examples of some of these field illustrations. The following discussion provides an example of the scale of material encountered and recorded by a survey project, in this case EKAS, during the 1999 and 2000 seasons. During those years some 120,000 artifacts were observed and recorded by the survey teams, and of those approximately 18,000 (or 15.3%) were picked up by the survey team members and described by the Processing Team. For these same seasons, approximately 1,200 artifacts were drawn and 600 photographed. All of these are available for further analysis and study. In addition, we were allowed to collect artifacts in certain controlled situations, especially in what we call LOCAs (Localized Cultural Anomaly, essentially the same as what is meant by POSI [Place of Special Interest] in PRAP [Pylos Regional Archaeological Project], SCSP, etc.). In the LOCAs, during 1999 and 2000, we collected 640 artifacts, and these form the beginning of an ongoing study collection, naturally closely linked to the CT nomenclature and available for inspection by interested scholars. This collection was significantly increased in subsequent seasons.

THE QUALITY OF CERAMIC EVIDENCE FROM ARCHAEOLOGICAL SURVEY

Purposefully experimental procedures were carried out during the 2000 season of EKAS, involving an area (Rachi Boska) where intensive survey units had been investigated in 1999. In LOCA 9002, a grid of 40 (10 m x 10 m) squares was laid out over the area. Within most of the squares (30 out of 40) a ChronoType sample was taken from a 5-m2 circle within the grid square (a 5% areal sample of the whole). Ten of the squares were “experimental” in that they were sampled using four different techniques: (1) “Dogleash CT”: ChronoType collection of a 5-m2 circle within the grid square; (2) “Dogleash All”: complete collection of that same circle (achieved by separately bagging what was left after ChronoTyping); (3) “Grid Cell CT”: ChronoType collection of the rest of the square; and (4) “Grid Cell All”: complete collection of the rest of the square. With 4 separate bags, we essentially had 4 types of samples: (1) ChronoType of the circle; (2) complete collection of the circle, by combining pottery from bags 1 and 2; (3) ChronoType collection of the whole square, by combining bags 1 and 3; and (4) complete collection of the square, by combining all 4 bags. The purpose of this exercise was to determine the optimum sampling strategy for all squares. By comparing samples from circles with samples from entire squares we would see if the small sample provided by a 5-m2 circle would be sufficient to provide satisfactory qualitative and quantitative information from the square. By comparing complete collection with ChronoType collection we could determine if collecting ChronoTypes would provide a representative sample of the artifacts in the LOCA. Upon examination of the artifacts in each sample we found that only one pottery type appeared in Complete Collection of all 10 squares that did not appear in ChronoType collections of the same squares. These results are reassuring, especially since ChronoTyping generally reduced the quantity of material picked up by around one third. The situation for the circular samples was similar. ChronoType collection of circles often was, in fact, the same as complete collection: generally, 0–3 sherds were added by complete collection of a circle and complete collection added a significant amount of material (>10 sherds) in only a few instances. These figures inspire confidence in the most minimal of our sampling options. The data recovered from circles was adequate to produce qualitative and quantitative information, yet took around one fourth of the time of the squares. Further experiments were conducted in the 2001 EKAS field season and these will be reported once the material has been fully analyzed. Some scholars will undoubtedly observe that the absence of a large storehouse of artifacts from these projects seriously inhibits our ability to reconsider the initial identifications and the preliminary conclusions drawn from them. The artifacts themselves cannot be re-studied by specialists (either members of our teams or others). On the other hand, we should point out, first, that these techniques are used mainly in the first phase of survey, what we call the “discovery phase,” and in the more intensive LOCA phase we do indeed collect a significant quantity of material that can be used for later analysis. The approach in all of these projects, however, has essentially been that of “siteless survey” and we do rely

33

34

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

heavily on what in other projects might be called “off-site” data, and we have been, so far, pleased with our ability to use the drawings and photographs from the Processing Teams to refine and revise identifications made in the field. Finally, as mentioned above, one of the purposes of this “low-impact” approach is to leave as much as possible of the archaeological material more or less in situ in the survey areas. As a result, resurvey and various kinds of re-investigation are especially appropriate and, in a certain sense, the whole of the survey universe remains an enormous “warehouse” where we “store” the artifacts from the survey. Experiments carried out in 2003 with such re-investigations seem especially encouraging.

Conclusions These procedures are still new and we are constantly making changes in definition and increasing the number of ChronoTypes. Thus, considerably more experimental data will be needed before their usefulness can be fully demonstrated. The nomenclature system, the heart of the ChronoType hierarchy, however, is at worst benignly consistent: few scholars would disagree with an increased rigor in taxonomic usage. In this regard it is a practical, perhaps even simplistic “cousin” to the more ambitious schemes of Sebastian Heath, Pedar Foss, and others. We are, in fact, working along very similar lines, and one imagines greater congruence of design in the years to come, without ever imagining that there will be complete uniformity of method. Certainly, as these systems develop, it will be possible to collaborate, and scholars working in one system will be able to make easy use of data generated in others, something that everyone working in survey archaeology should applaud. The ChronoType system facilitates recording and, perhaps most importantly, increases our ability to use the artifact data in (especially electronic) analysis. The perceived danger here, of course, is that the CT system forces specialists to place artifacts into categories where they may not properly fit. Nonetheless, this objection is based on a misunderstanding of the system, since the nomenclature can be expanded endlessly to cover various shades of uncertainty, overlap, ignorance, and disagreement among specialists. Key to the development of this system, as mentioned above, is the generation of further experimental data, especially from full pick-ups (and collections) of whole survey units. Ultimately we will bolster this with resurvey studies to test, not only our ability to see the objects on the ground, but more importantly to know and be able to record what the artifacts are with an accuracy that will inspire confidence from others interested in survey method. Although the techniques discussed in this chapter grew, to a certain degree, from the necessities of permit restrictions, we would emphasize, first, that the broader framework of the ChronoType system was developed in SCSP, in Cyprus, where there were no such restrictions. Second, we believe that the techniques described here represent a set of workable and valid methods in survey archae-

THE QUALITY OF CERAMIC EVIDENCE FROM ARCHAEOLOGICAL SURVEY

ology. They will certainly not be applicable to every situation, even in the Mediterranean, and they will need further development and modification as research goes forward. Such a “low-impact approach,” however, should not be seen as a “second-best,” but as a means that can produce high-quality results at less cost and less disturbance of the archaeological record and that may be used with positive results in many contexts.

Notes 1. Ceramics will be used as the primary example of CT classes since they are, arguably, more complex—or at least more fully studied—and our work has focused more fully on the definition of ceramic ChronoTypes. Other CT classes (metal, stone, etc.) can be divided in similar ways. 2. It is true that much larger vessels were sometimes made in this “cooking ware” fabric, especially in the Roman period, and this could potentially obscure the connection between subclass and function; nonetheless, amphorae with cooking ware fabric were normally made with much thicker walls, and in the CT system sherds of this type would properly be classed as medium coarse. 3. Dates for Prehistory are notoriously difficult to establish and are subject to modification. These have been provided, in the present instance, by Daniel Pullen and Thomas Tartaron. For further discussion of the problems of prehistoric chronology, see Warren and Hankey (1989) and Manning (1995). 4. The following is a summary of what a few important recent surveys in Greece have said about their collection strategy. For the Nemea Valley Survey, see Wright and colleagues (1990). On “tracts”: “Collections are made of all potentially diagnostic pottery (i.e., all but plain body sherds), all chipped stone, and any other types of materials” (Wright et al. 1990:604). On “sites”: “Standard procedure involves the collection of all artifacts found within circles 5 sq. m. in area…after which ‘grab’ samples of potentially diagnostic artifacts are collected from each of the four quadrants defined by these transects” (Wright et al. 1990:606). Davis and colleagues (1997) detail procedures for PRAP, including that transects were spaced 15 m apart and “All ‘feature’ artifacts (in the case of sherds, all but coarse, undecorated body sherds) were collected and brought to our workrooms, where they were identified and, at the end of each season, permanently stored by tract as a reference collection in the Museum of Chora” (Davis et al. 1997:401). See also Davis et al. (1997:401, and notes 28 and 29) for discussion of “transect-and-grab techniques” of “site” investigation. The Methana survey (Mee and Forbes 1997:35) ruled out the total collection of artifacts from sites, reckoning that a medium-sized site would yield approximately 44,500 artifacts—and they recorded 103 sites. Instead, they opted for a “representative sample, preferably of diagnostic artifacts.” The process was a “grab sample,” which in practice tended to consist of feature sherds and lithics. As the selection procedure was subjective, the sample was bound to be biased, but an attempt to compensate for this by processing the artifacts from the density counts did not yield much additional information and was discontinued.

35

36

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

The project (Mee and Forbes 1997:39–40) counted off-site pottery after the first season, but did not collect any: “Total off-site artifact collection would have reduced our progress across survey transects to a snail’s pace and generated large quantities of mostly undiagnostic material. Quite apart from the time-element involved in processing it, there was the potential problem of transporting the material back to base…The alternative, to collect only diagnostic pieces, was equally unacceptable.” 5. The significance of the “2 m swath” as opposed to “full” coverage of the survey unit is important from a theoretical and analytic point of view and it represents a significant difference between practice at SCSP (where coverage was supposedly total within the survey unit) and that at EKAS/APKAS. This may have ramifications for the interpretation of the data that have not yet been worked out, although my own working assumption (not necessarily shared by Nathan Meyer) is that the difference between these two methods is not significant and that, in fact, field-walkers cannot see anything more than a 2 m swath, even if they wander over their line of walk, especially if it is 15 m wide (as it often is in comparable surveys). This, however, is something that requires further discussion and elaboration. 6. This observation relates to the number of sherds on the ground and does not attempt (at this stage of analysis) to account for the fact that vessels will have been of different sizes and that larger vessels will naturally have “produced” more sherds; this can, however, be addressed at the analytic stage and the CT system will assist in this by allowing us to factor the number of sherds by a “size index” that may be attached to each CT (or to each CT subclass). For discussion of size of vessels as a factor in quantitative analysis of pottery, see Fentress and Perkins (1989).

3 Sampling Sinop: Putting Together the Pieces of a Fragmented Archaeological Landscape Owen Doonan

T

he Sinop Regional Archaeological Project (SRAP) is an interdisciplinary archaeological investigation of the engagement of the Sinop1 (Turkey) hinterland with a succession of Black Sea cultures and economies over the past 5,000 years. The Sinop promontory extends about 25 km into the Black Sea at the midpoint of its southern (Turkish) coast (Fig. 3.1). Sinop is shut off from the Anatolian plateau by the Pontic mountains and set in an ideal location to control trans-Pontic and east-west coastal communications (Hiebert et al. 1997). Sinop port has the best deep-water harbor along the south coast and has functioned as a fulcrum of Black Sea communications since the establishment of the Greek colony of Sinope in the late seventh century. The Sinop promontory offered a rich and diverse hinterland to this strategic port. It provides an excellent case study for clarifying the engagement of local communities in the broader cultural, political, and economic systems around the Black Sea. The Black Sea is an unfamiliar region to archaeologists working in the Mediterranean in part due to the lingering effects of the Iron Curtain and in part due to the lack of research that has been done on the Anatolian coast (notable exceptions: Alkım, Alkım, and Bilgi 1988; Burstein 1976; Bryer and Winfield 1985). The Black Sea economic and cultural community started to emerge as early as the fourth millennium BC, when ceramics found by SRAP show close stylistic and technological parallels to traditions in the western and northern regions of the Black Sea (Bauer 2002; Doonan 2004). Evidence for the emergence of a Black Sea cultural koine is stronger in the early first millennium BC on the eve of Greek colonization (Doonan 2004). Colonists from Miletus established Sinope in the late seventh century as an emporion through which contacts with the metal-rich eastern coast and the agriculturally rich northern shores of the Black Sea were established and strengthened (Akurgal and Budde 1956; Boysal 1959; Doonan 2002; 2003; 2004). Sinope’s relationship with its hinterland changed profoundly from the fourth century BC. Persian influence interrupted Sinope’s connections with its eastern colonies, prompting a new focus on local agricultural production as the foundation of the city’s economy. From this time through Early Byzantine times (the seventh

38

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 3.1. The Sinop promontory, with the various topographic zones indicated. Each zone was sampled in the extensive survey of the promontory. Zones were defined based on their ecological and physical properties, their relationship to the port, coasts, potential communications routes, and cultural-historical and environmental factors. (1) Boztepe, (2) Sinop near hinterland, (3) Inceburun, (4) West coast valleys (outer promontory), (5) West coast (rough to Ayancik), (6) Karasu valley, (7) East coast valleys (outer promontory), (8) East coast (rough near Gerze), (9) Middle highlands near Erfelek, (10) Highlands, (11) Kızılırmak-Gökırmak valley.The position of Sinop in the Black Sea is indicated in the inset.

century AD) the economy of the promontory is characterized by expanding agricultural and industrial production. The growth of the hinterland economy was reversed during the Arab and Turkish invasions (ninth through eleventh centuries AD) and the bitter struggles between the Christian kingdoms of Byzantium and Trebizond (a former Sinopean colony in the eastern Black Sea) and the Seljuk Turks. The port of Sinop became a critical node in the political and economic systems of the Black Sea, but its relationship to its former hinterland on the

SAMPLING SINOP

Figure 3.2. CORONA photograph of the Sinop promontory highlighting the Boztepe, Demirci, and Karasu valley sample quadrats. Dark areas are heavily forested, offering almost no potential for systematic archaeological survey.

promontory weakened. This pattern continued through the Ottoman Turkish domination of the Black Sea (AD 1461–1923), although in the latter years the coastal hinterland showed some expansion as cash crops like tobacco were introduced (Doonan 2004). During the Ottoman period Sinop was a major shipbuilding center supplying warships for the Ottoman empire, and the inland areas show evidence of widely dispersed subsistence-level settlement.

Conditions on the Sinop Promontory This coast is notable for its high annual rainfall (up to 2400 mm in the eastern Black Sea) and rugged terrain. The dense forests of the central and eastern parts of the coast have been renowned for shipbuilding timber since antiquity (Doonan 2002). The dense vegetation of this region makes archaeological visibility very low. Forested and fallow areas offer close to zero visibility. Over 50% of Sinop province is classified as forest by the Turkish government, making it the most heavily forested province in the country. Heavily forested areas appear as black on Figure 3.2. Survey results in these areas are particularly difficult to obtain because stone and masonry architecture were seldom used outside of Sinop port and its suburbs before this century (Fig. 3.3).

39

40

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 3.3. Vernacular architecture in the Sinop hinterland. Houses like this one have traditionally been constructed of wood that decays rapidly following abandonment. The terracotta roof tiles commonly used in the lowlands and slate in the highlands are quickly removed after the abandonment of a house, leaving few traces to be detected by archaeologists.

The surface conditions on the Sinop promontory make it necessary to sample rather than to conduct anything approximating a full-coverage survey (Fish and Kowalewski 1990). More importantly, SRAP seeks to understand social, cultural, and economic links between the port and distant parts of the promontory. A sampling strategy analogous to those employed in multi-sited ethnographies (Marcus 1995) can help to coordinate research goals and sample distribution. A clearly defined set of research goals within a multi-site research framework can be useful in assessing whether a sufficient sample has been examined to support viable conclusions and if not, what further investigations may be necessary. Of course, any archaeological landscape study is bound to be “multi-sited,” so how might such an approach be distinctive? Research topics like globalization, diaspora formation, and trade are situated in a network of locales and environments suited to analysis in multiple research arenas (Appadurai 1990; Bauer and Doonan 2002; Shami 2000). Rather than focusing on multiple places, multi-sited research follows the processes that connect distant physical and

SAMPLING SINOP

conceptual locations (Marcus 1995:105). These processes are the primary focus of multi-sited research rather than the patchwork of places that are connected by them. This approach has been especially fruitful in analyzing cultural production in multi-sited communities, linked by modern communications technologies ranging from telephones and TVs to internet chat rooms (Appadurai 1990; Clifford 1994). Ethnographers can “follow the people” (Marcus 1995:106), goods, or processes (Marcus 1995:105–10) to trace the locales in which a multisited program can be implemented. Processes that have connected the communities of the Sinop promontory include communication, interaction, production, consumption, and identity formation. Each of these processes maps differently across the land- and seascapes around Sinop. The following brief discussions sketch out how topographic and historical factors have impinged on each of these processes in the Sinop hinterland based on historical examples. The multi-sited nature of these processes is emphasized and in particular the close integration of terrestrial and maritime arenas. Each discussion concludes with a few thoughts on how focusing on the process (as opposed to focusing on the places) should affect the process of sampling in the Sinop region.

Communication Communications on the Sinop promontory are heavily dependent on the sea. Coastal communities (and in particular the port of Sinop) have facilitated communications between the hinterland and the broader Black Sea world in many cultural horizons from as early as the mid-third millennium BC. Factors conditioning overland communications include physical topography, environment, and social geography (infrastructural, political, and demographic factors). In different seasons and topographic contexts ridges, terraces overlooking river valleys, and high mountain passes have provided effective transportation routes through the Sinop promontory. Plentiful winter and spring rainfall and melting snows cause violent spring flooding that washes away bridges and lowlying roads in the river valleys. Highland routes are reliable for long-distance travel after the snows melt in the spring, but crossing between these highlands and the outer promontory requires passage through the dramatic folded canyons most vulnerable to spring flooding. Like terrestrial landscapes maritime seascapes structure interaction by a combination of physical, economic, and cultural factors. Physical features such as prevailing winds, surface currents, and coastal topography are strong determinants controlling the basic potential of a maritime space for interaction. In the Black Sea long-distance deep-water routes follow the prevailing surface currents between Sinop and the Crimea (Ballard et al. 2001; Hiebert et al. 1997). The eastern side of the promontory is protected against the prevalent winter storm patterns coming from the northwest. Coastal routes connected local and primary

41

42

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

ports and access along these routes is more reliable along the coast to the east and south than along the north and west. Maritime and terrestrial interaction are normally considered independently of one another. However recent research in Denmark has highlighted the importance of the coastal interface in connecting maritime and terrestrial communications systems (Westerdahl 2001). Coastal topography and infrastructure are important to creating and maintaining an effective transition between communications systems on land and at sea. Our investigation of communications needs to sample in each of these environments to yield patterns of settlement, production, and infrastructure (roads, bridges, hans, churches/mosques, and so on). We need to examine whether the coastal infrastructure to the west of Sinop is more or less developed than that to the east. Where and when were investments in the overland transportation networks more intensive? And finally, the underwater research program can provide essential information on the links of hinterland and coastal communications routes with overseas communities.

Interaction Interaction takes many forms (Schortman and Urban 1994), each of which maps differently over a particular geography. Warfare, migration, trade, and other forms of interaction flourish in different spatial and social contexts. In Sinop the mountain paths with minimal potential for high-volume trade were on occasion effective conduits for military raiding or seasonal exchange (e.g., transhumance or traveling bazaars). In most periods the rugged Pontic mountains were a deterrent to interaction between Sinop and the inland Anatolian plateau, although they did provide the opportunity for sudden military ventures (Kashka invasions of the Hittite kingdom, Mithridatic wars, Seljuk capture of Sinop). Trade across the mountains has been difficult in some periods because of the low population in the highlands (Seljuk and Ottoman times). In some cases large-scale migrations over land have taken place (e.g., the Turkmen and others that came into the area following the Turkish conquests of the thirteenth century). Maritime interactions also map differently onto different seascapes. A coast studded with many ports might facilitate trade or periodic raiding but discourage large-scale migration. A dense network of ports could encourage or discourage trade and warfare based on historical factors. Interaction also depends on social structures that are formed and maintained through repeated contacts. Maritime interaction was a feature of Black Sea culture from before the initial Greek colonial expansion (ca. 650 BC) onward, laying a foundation for later trade systems (Doonan n.d.). “Pre-trade” contacts may have helped lay the groundwork for later intensification of long-distance relationship building, while the spread of technological and economic information could have stimulated interest in developing mutually beneficial relationships. In several horizons a

SAMPLING SINOP

colonial community or diaspora facilitated interactions through the port of Sinop (e.g., Greek colonial). The communications geography does not stop at the water’s edge (Hiebert et al. 1997). In an underwater context the types and contents of submerged vessels could illuminate the nature of maritime interactions. However, different kinds of vessels and different circumstances of deposition should complicate interpretation of sunken warships, merchants, or fishing vessels. It may be difficult to compare the intensity of deep water vs. coastal interaction because of the different probabilities that a given vessel would fail in these very different contexts. Research designed to trace the processes of interaction needs to take into account the different intensities and geographic configurations of these various kinds of interaction.

Economic Production Economic production in terrestrial environments is dependent on topography, distribution of resources, demography, and infrastructure. The ecological concept of patchiness may provide a useful analogy for evaluating productive potential in inhabited landscapes (Winterhalder 1983). Patches are local discontinuities in a landscape with distinctive properties that affect the behavior of particular organisms and processes. Patchiness is the degree to which patches are widely or unevenly dispersed through a landscape (Winterhalder 1994:33). Patches in a human productive landscape might be identified with certain soil types, topographic conditions, raw materials, and social factors, such as access to communications or potential for specialization in a complex regional economy. Taking patchiness into account could provide guidelines for the appropriate scales at which various kinds of production are visible (e.g., kitchen gardening, industrial olive production, old growth forest clearing, and so on). By taking the scale of these processes into account it should be possible to estimate the nature and intensity level of landscape studies needed to investigate them (Wandsnider 1998). Inland patches might be defined based on the availability of resources, suitability for agricultural or pastoral production, and/or industrial facilities. Coastal production can take advantage of a complex maritime-terrestrial catchment area. Production in maritime environments depends on the distribution of resources, particularly fish, and on various political and demographic factors structuring access. Strabo, Aelian, and other ancient sources provide information on the distribution of different fish species around the Black Sea (Doonan 2002). The rhythms of seasonal variability and spatial dispersal of resources can encourage greater or lesser degrees of specialization within and between inland and coastal communities depending on the organizational context of production. Our study of production on land investigates patches in a productive landscape in which landforms are divided into four zones on the basis of their underlying geology (Fig. 3.4): (1) rugged highland zones (deformed flysch bedrock);

43

44

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 3.4. The major geological regions of Sinop promontory, including the outer volcanic, the central mixed marine, and highland flysch zones. Major faults are marked with thin straight lines. The primary faults occur where the Cretaceous volcanic bedrock has thrust up above the later (Neogene) mixed marine deposits and along the west coast where the promontory has thrust up with respect to the sea bed. Severe folding in the highlands has produced the dramatic relief apparent in the topography. Rainfall is indicated at Ayancik, Sinop, and Gerze in mm/annum.

(2) the promontory itself consisted largely of plains dissected by riverine valleys (bedrock of lime-rich sandstones with isolated limestone deposits); (3) coastal valleys and colluvial marshy deltas; and (4) isolated volcanic masses along the north coast. The forests of the highlands have been important sources of raw materials since at least the fifth century BC (Doonan 2002). Most of zones two and three have plentiful water and lime-rich sandy soils. Isolated deposits of clays are found in the bottoms of several coastal valleys, but otherwise there are few deposits of economically significant minerals. In the highland flysch zones patchiness is dependent partly on the distribution of plant types: coniferous vs. broadleaf forests offer very different raw materials (especially timber) and food products (arboreal nuts and fruit); treeless meadows provide agricultural land and extensive grazing for animals. Topography has also been critical to the engagement of some places in regional production: timber resources that were difficult to remove were left untouched while accessible timber along the coasts and in valleys offering easy access to coastal ports was used for

SAMPLING SINOP

shipbuilding and other uses. Patchiness in the highlands could also be dependent more on social and historical factors (transportation, deforestation, siting of settlement nodes) as much as on ecology. Valley systems from the ridges to the lower slopes make appropriate units for study. We have studied the Karasu and Demirci valley systems to date using sample quadrats (see below). In periods when the hinterland economy was intensive (Roman, Ottoman), coastal production sites and ports made use of the raw materials and agricultural products from the hinterland and maritime zones to produce goods for export. Maritime components of production in Sinop include fishing and trade. At some times raw materials may have been brought in from outside the promontory (e.g., shipbuilding timber in Ottoman times). In these cases underwater research and close analysis of raw materials may be critical to understanding the complex distribution of sources. Fishing has long been a valuable industry in Sinop—the pelamydes (bonito) of Sinope were prized in antiquity, earning very high prices in the Mediterranean (Doonan 2002). Coastal facilities for processing the fish must have been present, although clear evidence for them is still lacking. Our sampling program seeks to establish units for study based on the complementary activities carried out in these various zones. In periods with integrated industrial economies the scale of agriculture and production are sufficiently great that relatively limited samples can suggest that a study area was likely to have been involved (or not) in the regional economy. However sampling for evidence of dispersed subsistence economies is more difficult because the scale and density of production are so much less. Over time SRAP is looking to achieve a balanced distribution between geographic zones covering potential sources of raw materials, communication routes, and loci of production. A stratified program of survey, remote sensing, excavation, and materials analysis allows us to identify the contexts and nature of production.

Consumption The products of local, regional, and distant producers are consumed in another multi-sited process. Consumption can be traced in and around habitation sites, refuse disposal areas, cemeteries, and special purpose sites (sanctuaries, temporary markets, and others). Understanding consumption as a multi-sited process requires that we achieve good control of the origins of goods that are then used and deposited elsewhere. It is also important to identify different kinds of use contexts for consumed goods based on the nature and surface distribution of material assemblages. Onsite gridded pick-up, geochemical and geophysical studies can be used to refine the interpretations of sites, and excavations can retrieve in situ material for dating and analysis. Settlements and industrial sites are the largest and may be most effectively sampled using a rather coarse-grained sampling strategy. Other

45

46

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

kinds of sites like cemeteries and sanctuaries are generally located in more marginal spaces and more difficult to detect except in a relatively extensive sampling program.

Identity Formation Cultural production can be traced through a number of historically contingent processes spread through a landscape. In the Sinop region one of the most critical cultural processes has been identity formation, as the region has been colonized and contested by distinct groups in repeated episodes from the Neolithic onward. Identities are inscribed at many scales in the landscape. Identity can be traced through what Knapp and Ashmore (1999:11–13) call constructed and ideational landscapes. Constructed landscapes are created by an intensive investment of energy. Within and between communities people inscribe, read, and reproduce identities through constructed landscapes of villages, gardens, fields, roads, cemeteries, and temples. These transformations build up and decay over time, creating a palimpsest of superimposed features (Wandsnider 1998). The types and styles of material culture used in different contexts would be prioritized in a study of identity. The special purpose sites like cemeteries and sanctuaries would be particularly critical, so a sampling program would need to be sufficiently extensive to detect these. Reliable evidence about ceramic assemblages can help to establish basic frameworks for exploring models of identity in particular locales (e.g., colonial vs. indigenous, socially stratified vs. non-stratified, locally distinctive vs. regionally integrated social groupings). Combining studies of spatial variation in the formal, decorative, and technological aspects of material assemblages we can start to explore the processes of rupture and integration between coastal and inland populations produced by repeated episodes of colonization in the Sinop region (Shennan 1989). Identity is also established in ideational landscapes, making it necessary to sample in topographic locales not necessarily suited to studies of production. Sacred locales are often sited in marginal spaces such as mountaintops, caves, and the middle of the sea (Bradley 2000). To trace the path of identity we must also investigate these marginal spaces, examining geological formations likely to produce caves, high exposed places, unusual natural features, and areas that look onto such features. The sampling program needs to be sensitive to contexts likely to have been focal points in local Bronze Age, Hittite, Paphlagonian, Greek, Roman, and Ottoman ideational landscapes.

Applying the Multi-Sited Approach: SRAP The multi-sited approach engages fieldwork and research design in an active dynamic partnership. To date the SRAP has carried out one season of extensive archaeological survey, three seasons of systematic archaeological field survey,

SAMPLING SINOP

geomorphological and geophysical research, and one season of sondage excavations at the site of Sinop Kale N.W. SRAP has adopted a nested program of archaeological investigations that records archaeological evidence at several spatial scales and levels of resolution: 1. Site-level studies (geophysics, excavation, high-resolution pick-up) 2. Intensive surveys of high visibility on- and off-site sample quadrats (ca. 25% sample of 1 km2 sample areas) 3. Systematic surveys of larger areas (ca. 2–5 km2) that examine a range of visibility conditions

Site-Level Studies Most archaeological projects expend the majority of their time and energy studying sites, the relatively dense clusters of archaeological evidence that collect in places where people have lived, thrown out trash, worshiped, made things, or buried the dead. SRAP refers to such places as “loci” to avoid the confusing implications that have built up around the word “site.” Each locus is a cluster of archaeological material (pottery, tiles, daub, lithic debris, and so on) that is distinct enough to suggest that some intensive activity took place there. Naming a locus is a judgement call—clusters of archaeological material can form through a host of different human and natural formation processes (Schiffer 1987). Loci are different from the usual use of the term “site” in that a “site” might actually be composed of multiple loci. For example, a series of houses with spaces between them might be evidence of a village settlement. Such a settlement would often be called a site in the archaeological literature. However, if the spaces between clusters are clear enough to suggest distinct units we would refer to each cluster as a locus. Once a locus or series of related loci is identified we decide whether further close analysis is warranted. There are several options of high-resolution surface study and excavation available (listed roughly in order of increasing expense): high-resolution surface pick-up, gridded geochemical sampling, geophysical studies, and excavation. High-resolution pick-up records the density and nature of finds in well-defined collection units (gridded 2 m2 sample boxes, linear transects), allowing more precise definitions of specific activity areas (Fig. 3.5). All finds are photographed with a digital camera, and all ware types are characterized according to gross physical properties (hardness, color, inclusions, and so on). Detailed photographs are keyed to the descriptions of ware types to assist in later identification of these types on digital photos. The high expense involved in these studies means that they are applied in selected on-site contexts. Because the area of intensive human activity always extends beyond the structures and refuse piles that form the most visible loci, onsite studies attempt to interpret sets of loci and the spaces around them in relatively coherent units.

47

48

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Table 3.1. Overview of nested Black Sea Trade Project. INVESTIGATIVE

DATA ANALYSIS

OBJECTIVES

THEMES EXPLORED

METHOD (LAND)

Small-scale excavation

Collection of in-site evidence for temporal, seasonal, and functional analysis

Defines economic and chronological data from single period sites

Production Consumption

Field survey (continuous full coverage)

Spatial analysis of site distribution and non-site features

Settlement pattern, Density, diversity and intensity of usage

Communications Interaction Production Identity

Field survey (non-contiguous sampling)

Spatial analysis of assemblage variation in on-site and off-site contexts

Distribution of ceramic types onand off-site; functional definition

Production Consumption Identity

Geomorphology and palynology

Evaluate landscape development, impact of human occupation through time

Defines local ecological and landscape changes, anthropomorphic context

Production

Geophysics and geochemistry

Site definition, function, and specialization

Defines site extent Production and structure, speCommunications cialized use of space

Ceramic and floral analyses

Site function, distribution and spread of local and imported products

Identify economic patterns

Production Consumption Identity

Historical and ethnographic

Land-use evaluation and specialization, customs registers

Defines land use, infra-structure, and resource procurement

Communications Interaction Production Consumption Identity

Archaeological finds from loci are subjected to close analysis. Pottery, building materials and industrial debris are subjected to a program of physical and chemical analyses that helps to characterize the date and origins of finds. Laboratory analyses are tied in with the ware types recorded in the field. These studies

SAMPLING SINOP

Figure 3.5. Plan of the settlement and cemetery of Karapınar (Nohutluk quadrat, Karasu valley). This locus was recorded using 100% pick-up of 1 m transects at 10 m intervals. Lighter shading indicates higher concentration of ceramics (up to a maximum of 117,529/ ha). Arrows indicate the direction of transects and the numbers indicate the number of transects walked in each transect.

are the foundation of the multi-sited analysis of production, consumption, and discard (Table 3.1). Stylistic analysis of finds and interpretations of special finds help to clarify the processes of identity formation. Thus site-level studies not only permit us to interpret particular loci, they also form the basis for analyzing multisited processes.

Quadrat Surveys The next level of scale and resolution in our data collection is the sampling of high-visibility fields in sample quadrats. In three seasons of intensive on- and offsite survey we have investigated over 350 tracts in an intensive survey methodology that documents the spatial distribution of ceramic densities and types from plowed and planted fields that offer high visibility. In each sample zone non-

49

50

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 3.6. Survey tracts in the Demirci valley where six survey quadrats were sampled (west–east): (1) Uzungurgen S, (2) Uzungurgen N, (3) Eldevüz, (4) Kümes, (5) Demirci, (6) Keçioglu.

contiguous tracts were walked by field-walkers spaced at 10 m intervals. Our ceramic pick-up strategy collected quantified data on ceramic assemblages from on- and off-site tracts. The team collected 100% of the archaeological material from 1 m transects spaced at 10 m, counting, weighing, photographing, and characterizing these materials in the field. Only highly diagnostic forms and examples of ware types from loci were removed for further study, and we left 95% or more of the collected ceramics in the field. By photographing all finds we were able to collect rich data not only on eyecatching types, but also on the general composition of assemblages. By characterizing ware types in each tract according to gross characteristics, we can match digital photographs to field descriptions, reducing the errors introduced by highly differential lighting and other factors in the field. Depending on visibility and other considerations, the total area investigated in each zone ranged between 5 and 30 ha representing 10–40% of each sample zone’s area. This strategy allows systematic study of the circulation of local and imported material in the hinterland. Seven quadrats were sampled in the Demirci valley survey (Doonan and Gantos n.d.) and eight in the Karasu valley. Quadrat locations were chosen to represent a broad range of topographical and ecological potentials (Fig. 3.6). In the Karasu valley, quadrats were placed to test the proposition that this valley may have been the primary communications corridor for the interior of the promon-

51

SAMPLING SINOP

Figure 3.7. Survey tracts in the central Karasu valley including two sample quadrats: (1) Kılıçlı, (2) Nohutluk.

Table 3.2. Relative ceramic densities for eight sample zones in the Karasu valley; BR=Bronze Age, H=Hellenistic, R=Roman, O=Ottoman. DISTRICT NAME

Nohutluk/ Haciog˘lu

AREA SURVEYED

AVG. DENSITY SHERDS/ HA

PROXIMITY RANKING

Sinop

East

North

PREDOMINANT TYPES (IN ORDER OF FREQUENCY)

142,580 m2

1205

5

2

7

R, O, Br

Sarsi

58,920 m2

398

7

8

4

R, Br

Uzungurgen

52,250 m2

320

8

1

8

R, O, Br

188,330 m2

107

1

3

2

O, H/R

42,690 m2

93

4

4

8

O, Br

284,560 m2

50

6

6

3

O, Br

Osmaniye

77,700 m2

21

2

5

5

modern

Akliman*

103,150 m2

10

3

7

1

H/R, O, Br

Bostancili Kılıçlı Dibekli

* The surveyed Akliman zone does not include the site of Harmene, which has very low visibility because of Ottoman and more recent improvement of the harbor area and dense forest covering the rest of the site.

52

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

tory. Two coastal patches (Akliman, Bostancili) investigated whether the coast was the focus of industry, settlement and the primary communications outlet for the valley. Bostancili was located just beyond the ancient suburban spread of Sinope. Akliman and Sarsi were in a position to take advantage of delta marshes. Osmaniye and Dibekli held commanding positions on the flanks of the delta. Kılıçlı, Nohutluk, and Uzungurgen overlook the river floodplain along the possible communication route (Doonan et al. 2001). In the Nohutluk and Kilicli quadrats we surveyed over 50 tracts (approximately 20 ha total), representing elevations of 40–175 m (Fig. 3.7). Significant differences were observed in on- and off-site ceramic density in comparison with coastal sites (Table 3.2). The average density of sherds/ha in Nohutluk was more than 10 times greater than in Kilicli and 100 times that near the historically documented coastal port of Harmene, reflecting a significant inland occupation in Roman and Ottoman periods.2 The low density of ceramics from quadrats in the outer parts of the valley contradicts the model that this valley was a major communications corridor. In fact, the highest densities of ceramics are in those quadrats closest to the east coast (Table 3.2). This pattern may suggest that the east coast was a more significant focus of settlement, production, and communications than the river valley. The quadrat sampling system has permitted us to gather comparable data from dispersed locales in the Sinop promontory. This system takes advantage of the relatively sparse areas with good visibility and permits sampling based on the multi-sited research questions discussed above. On the other hand this sampling method skews our data collection toward the agricultural fields cleared in Ottoman and later times. Certain types of loci like the typical mounded Prehistoric settlements and Ottoman villages are under-represented. For these reasons we are complementing the quadrat sampling system with more continuous coverage of broader transects.

Transect Sampling The most extensive form of systematic survey employed by SPRS investigates more contiguous fields representing a variety of field conditions in 2–3 km 2 sample transects. Transects are defined to cross-cut sets of patches in a landscape. Field-walking is still organized in tracts with field-walkers collecting quantified on- and off-site data in evenly spaced (10 or 25 m intervals) parallel field transects (not to be confused with the larger transect sample units). In comparison to the quadrat samples, the intensity of counting, weighing, and ware characterization is reduced in order to permit more extensive coverage. This sampling method is more effective than quadrat sampling for estimating settlement density and patterning and recording locus types that tend not to be preferred for agricultural fields. Scarps, springs, riverbanks, and similar landforms that often provide valuable information about subsurface conditions are recorded more effectively.

SAMPLING SINOP

Special features that are notorious for bad visibility but frequently associated with sacred activities (e.g., springs, caves, and mountaintops) are incorporated into this sampling strategy.

Discussion and Prospects Multi-sited research design is permitting the SRAP to investigate the threads and conjunctures that integrated the disparate parts of this diverse region. By focusing on connecting processes rather than just the places of interest we can prioritize our sampling choices to address problems situated in multiple places and at multiple scales more effectively. In the decade following the publication of Fish and Kowalewski’s influential 1990 volume, Mediterranean archaeologists have been moving away from sampling in survey and more toward full coverage. “Full coverage” does not necessarily mean recording everything, but the strategy is to cover a large percentage of the study area. This is a reasonable approach if the research is oriented toward understanding the catchment around a particular settlement or the settlement history of a specific valley. Despite the soundness of full coverage for some research agendas, projects that aim to understand processes that play out over a larger spatial scale may be better suited to other kinds of research structures. Many processes are not situated in a continuous landscape or at a scale that can be practically addressed by fullcoverage survey. Such processes may play out in fundamentally diverse environments (maritime/terrestrial, cultivated coastal / forested highland) and thus have little potential for full coverage in any meaningful sense. In these cases a sampling program tailored to clearly defined research agendas provides a more feasible approach. A sampling program does not have to be standardized or random to be effective, but can make use of information gathered to drive further research (Orton 2000). By discussing the relationship between the sampling program and research questions transparently, it is easier to develop critical target areas for further sampling. Some of the themes that will be more fully addressed by the SRAP field sampling program in future seasons include (1) pre-colonial communications, production and identity; (2) Ottoman production, consumption, and the relationships between diverse ethnic communities; and (3) highland production, consumption, and communication with the coasts. East-west oriented transect samples incorporating the Sarsi and Hacio¤lu quadrats will clarify the distribution of pre-colonial settlements. By extending transects into non-forested patches on the west side of the river we can clarify the degree to which settlement in this part of the promontory was tied in to the east coast. On-site geophysical and gridded pick-up of materials on pre-colonial settlements will clarify production and consumption patterns. Ottoman period production and consumption will be investigated by intensive documentary research and ethnohistorical investigations in the field (Table 3.1). The highlands will be

53

54

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

investigated by a combination of sampling strategies since to date SRAP has conducted only very limited investigations in these environments. Highland quadrat and transect surveys will contribute substantially to our understanding of the ideational landscape of Sinop. Survey of highland transects in the southeastern part of Sinop province that might cross-cut potential communications routes could clarify the communications routes linking the coast to inland regions in some periods and perhaps also help us to understand the margins of Sinopean territory in Hellenistic times.

Concluding Thoughts These are still relatively preliminary thoughts on multi-sited research and archaeological survey. The data gathered to date will permit us to develop models of interaction for the Sinop promontory as physical analyses of materials from the survey progress (Bauer 2002). It is hoped that they will prompt further reflection on the importance of the conjunctures and threads that connect the patchworks of places documented in regional archaeological projects. Our sampling designs must be tailored not only to local topographic conditions and standard practice, but also should reflect significant cultural-historical processes within our research areas. One positive trend in recent decades has been toward integrated regional multi-year landscape-oriented archaeological projects in the Mediterranean. The multi-year structure permits a degree of flexibility and reflexivity in research design as research questions evolve. This is not a call to free-form research, with each survey designing its own idiosyncratic format. If surveys are not comparable their results are hardly useful. On the other hand, by considering the scales and contexts in which specific processes are visible in the landscape we can address problems that lie in the conjunctures between archaeological places.

Notes 1. Sinop is the name of the modern town and the province and will be employed in most situations. “Sinope” is the name of the ancient Greco-Roman port city located in the same place as the modern town and will be employed only when referring to the port itself during these periods. 2. In general, ceramic density figures in Sinop represent Roman occupational patterns, since Roman settlements are characterized by much more prolific ceramic production than earlier and later period settlements.

4 The Disjunction between Mediterranean and Near Eastern Survey: Is It Real? T. J. Wilkinson

A

brief perusal of the survey literature suggests that there is an ever-widening gap between archaeological surveys in the Near East and those in Mediterranean regions, with the former area appearing to be falling further behind in terms of the accurate recovery of settlements. In many parts of the Near East the tell has remained the primary objective of survey; not only are many smaller sites missed by regional surveys but even some mounded sites are under-represented. In contrast, in the Mediterranean region site recovery has improved to the point that even ephemeral scatters that show up only seasonally are being successfully logged (Ammerman 1995 and reply by Davis). Although there appears to be a stark contrast between survey standards in these two areas, does this mean that we should throw our hands in the air and say that no comparison is possible between the worlds of Near Eastern and Mediterranean surveys? Although there can be little doubt that many Near Eastern surveys are conducted at a level of intensity that seems hopelessly unrealistic, in terms of site recovery, a more measured view would suggest that settlement data can be recovered more successfully for some archaeological phases than others. In other words conventional full-coverage methodologies (Sumner 1990) may be satisfactory for recovering the basic settlement patterns of certain periods, but may underrepresent other periods such as those of the Iron Age and later. It should also be noted that certain types of off-site features occur only during specific archaeological periods or only in certain areas so that their record is influenced by the survey technique employed as well as landscape taphonomy. Finally, I note how the most recent surveys in northern Syria are successfully demonstrating the degree and scale of Hellenistic, Roman, and Early Byzantine settlement. Here I concentrate on survey results recovered from the northern Levant and upper Mesopotamia. I will also mention recent surveys conducted in the southern Levant, but for reasons of space I will not deal with the complications of the settlement record of southern Mesopotamia (but see Wilkinson 2000 for a summary of recent Mesopotamian surveys). Before moving on to the main thrust of this chapter, it is necessary to summarize my own perception of survey techniques in the Near East. Overall, large-scale full-coverage type surveys have been conducted over many parts of lowland Mesopotamia and upper Mesopotamia between the Tigris-Euphrates rivers (the

56

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Jazira), as well as further west in the Jabbul plain, the Biqa, the Orontes valley, and the Amuq. In contrast, in the southern Levant earlier tell-based surveys have been superceded by much more intensive site surveys (e.g., Finkelstein 1998), which in many cases include an explicit landscape component (Dar 1986; Gibson and Edelstein 1985). In addition some workers (e.g., Banning 1996) have introduced a geoarchaeological component into surveys so that it has become possible to recognize buried sites in the landscape. However, despite the landscape approach of many Levantine surveys, off-site recovery techniques are often lacking. Overall, survey techniques in the southern Levant are therefore starting to become more closely aligned with the intensive survey methodologies practiced further to the west in the Mediterranean. This shift in the southern Levant toward intensive survey is partly conditioned by topographic considerations because many Levantine surveys have been conducted in upland or hilly terrain. This necessitates the employment of intensive survey techniques that are capable of recognizing the smaller or more elusive sites that are so characteristic of such broken terrain. In contrast, within the extensive plains of Syria and Iraq (and to a lesser extent Anatolia), tell-based settlement patterns continue to be readily recovered using “traditional surveys.” Or are they?

Near Eastern and Mediterranean Surveys: How They Compare In his article “Frogs Round the Pond,” John Cherry noted how increases in the intensity of survey over recent decades have resulted in a commensurate increase in the recovery of settlements (Cherry 1983: Fig. 1). By comparison, surveys conducted in the Jazira of northern Syria, northern Iraq, and southern Turkey between 1980 and 1990 have not yet attained the level of data recovery of the post-1970s Greek surveys (Fig. 4.1). Rather, these Upper Mesopotamian surveys provide an average site density in the region between 1 site per 10 km2 and 1 site per km2, which, according to the Cherry Chart, falls between the level of recovery of “earlier Greek surveys” and those surveys of the post-1970 period. Significantly, these site recovery figures in the fertile semiarid steppe are much lower than those recorded in the arid and marginal regions of the Negev and northern Jordan. In contrast to moist areas to the north, where much settlement is nucleated in the form of tells, these sub-desertic areas are characterized by numerous small trace occupations dispersed across the desert. Although it is clear that we are not comparing like with like, it is significant that one of the reasons behind this high rate of site recognition is that processes of site formation and preservation are different in the two areas, as will be discussed below. In terms of the three categories of survey defined by Alcock for Greece (1993: Table 3), most Near Eastern surveys fall within her type C (i.e., extensive, nonsystematic surveys conducted over large areas and which recover only the higher

57

THE DISJUNCTION BETWEEN MEDITERRANEAN AND NEAR EASTERN SURVEY

10000 Greek surveys since c. 1970 Earlier surveys in Greece Large-scale Italian surveys 2

Near Eastern surveys

10

10

si

te

s

pe rk

m

2

rk m

2

NUMBER OF SITES

1000

pe

1

2

1

si

te

3

6

5

1

100

si

te

pe

r1 0

km

4

si te

pe

r1 00

km

2

7

1

7

9 10 10

100

1000

10000

SURVEY AREA [km 2 ] 1. 2. 3. 4. 5.

Uvda (Negev) Umm Jimal Samaria N Jazira Project Tell Beydar Survey

6. 7. 8. 9. 10.

Amuq Kurban Hoyuk Hamoukar Survey Dhamar Survey Wadi al-Hasa (Jordan)

Figure 4.1. Site recovery from a number of recent Near Eastern surveys compared with those from Greece.

end of the site hierarchy). Some fall within Class B (i.e., relatively systematic small-scale surveys often conducted by single individuals, or somewhat larger, more extensive coverages). Only in the southern Levant do we see the most intensive surveys of Class A (intensive surveys with good levels of control, usually conducted by means of transects to recover loci of human activity, rather than simply “sites”). Nevertheless, in Upper Mesopotamia “off-site” control and more intensive, systematic coverage have been employed over significant areas and have therefore provided a greater degree of control than would otherwise be possible (e.g., Wilkinson and Tucker 1995).

Landscape Taphonomy and Site Preservation The above statistics concerning site density demonstrate clearly that counts of the total number of sites will vary according to how one defines an archaeological site. Thus in the desert fringe the often subtle traces of individual structures or activity areas can hardly be accorded the same weight (in terms of population or archaeological significance) as the much larger frequently multi-period sites of more humid areas.

58

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Even during the earlier stages of landscape archaeology there was an awareness that the cultural landscape had suffered progressive attrition of features through time, with some elements being added while others were lost. Schiffer has formalized landscape transformation processes into those resulting from cultural and those caused by environmental processes (C- and N-transformations, respectively). As early as 1972 Christopher Taylor provided an important waypoint in the debate and his conceptual framework of “landscapes of destruction” and “landscape of survival” enabled archaeologists to take account of the likelihood of feature survival when assessing the landscape record on Britain (Taylor 1972). For example, above the limit of cultivation in the uplands there was an increased likelihood of landscape features or entire landscapes surviving, whereas in the lowlands (where processes of settlement and cultivation had endured over millennia) there was a much greater chance of earlier landscapes being lost. This simple conceptual framework received less attention than it deserved, but for the British Isles it has now been revised by Tom Williamson into “a particularly complex kaleidoscope of patterned creation and structure destruction” (Williamson 1998:6 and Fig. 3). From the Taylor-Williamson model it is clear that landscape transformations have played a fundamental role in the preservation or loss of landscape and settlement data. Strictly speaking the term “taphonomy” applies to processes that act upon biological features after death and before they become fossils. Here I have adapted the term to landscape studies because similar principles apply. For example, the progressive removal of stones from a wall to build a new one, or the wholesale removal of settlements to expand cultivated areas leads to the progressive loss of archaeological data from landscapes of destruction. In a similar manner, as has been discussed in a number of geoarchaeological texts, geomorphological processes can either erode away parts of the landscape or entomb landscape features beneath an obscuring blanket of sediment. Suitably amended, the Taylor model can be effectively applied to the ancient Near Eastern landscape. Landscapes with the greatest probability of feature survival occur in deserts and high mountains, whereas progressive loss of features is at its maximum in areas of long-term cultivation and rather less so in marginal zones of settlement. Finally the coastal zone experiences a patterned loss and survival depending upon coastal sedimentation and currents.

Zone 1: Zone of Preservation in Deserts Because most deserts lack the potential for irrigation they are not normally occupied by long-term sedentary settlement. Nevertheless, if settlement did take place it is unlikely that their remains would be erased by subsequent settlement. Consequently, archaeological remains can be remarkably well preserved in deserts.

THE DISJUNCTION BETWEEN MEDITERRANEAN AND NEAR EASTERN SURVEY

Zone 2: Zone of Preservation in Mountainous Areas At the other extreme, mountains provide unsuitable conditions for sustained human activity so that permanent settlement rarely occurs above a certain altitudinal limit. When it does occur, for example if the area suddenly becomes important for mining or highland pastures are adopted for upland camps (equivalent to the Turkish yayla), then archaeological remains may be preserved with some degree of clarity. Site and feature preservation can occur as long as the characteristic geomorphological processes of mountain areas (mass wasting, talus development, and so on) have not destroyed them.

Zone 3: Intermediate Zone (marginal rainfed steppe, with fluctuation settlement levels) The Jazira region of Syria, Turkey, and Iraq is dominated by tells that attained their maximum size often during the middle part of the third millennium BC. Optimum landscape preservation may occur in the intermediate zone because in wetter areas sustained human settlement has increased the loss of earlier features by attrition, whereas in drier areas even though landscape transformation processes are less marked, long-term settlement has not been sufficiently heavy in its imprint to cause features to be embedded into the landscape. Through the millennia this area has experienced several cycles of settlement rise and decline, the cause of which continues to be debated, but which appears to result from a range of climatic, social, economic, and political changes (see reports in Dalfes et al. 1997). In those areas where there has been a retreat in the margin of settlement, this has allowed the early phases of settlement and landscape to be visible.

Zone 4: Zone of Attrition Whereas Zone 3 exhibits a partial landscape record in which occasional earlier features and sites remain, in the “zone of attrition” there is decreased chance of feature survival because long-term settlement has been virtually continuous over some 9,000 years. This vast zone includes the Levantine coastal plain and much of Israel, Lebanon, western Syria, as well as parts of Turkey, Jordan, and Yemen. Although it is impossible to do justice to such a complex zone, it is important to bear in mind that in such areas a significant but unknown part of the archaeological landscape and settlements may have been lost by attrition. This is particularly well illustrated for parts of Yemen and the southern Levant where valley sides and floors are carpeted by staircases of terraced fields, the construction of which has resulted in earlier structures being dismantled or destroyed. In such areas it is often possible to see only occasional fragmentary remains of archaeological sites in the form of dispersed scatters of sherds and fragments of pre-existing

59

60

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

walls. For example in the southern Levant at the site of Sataf, a small ChalcolithicEarly Bronze I site was completely buried by later terracing (Gibson et al. 1991), and in the Samarian uplands many hilltop sites have been heavily disaggregated by later settlement activity (see Finkelstein and Lederman 1997).

Cycles of Settlement Nucleation and Dispersal and the Recovery of Settlement Patterns It has become apparent in recent years that some parts of the Fertile Crescent have experienced a cyclical fluctuation between phases of nucleation exemplified by tell-based settlement and phases of dispersal. This compares with Greece and the Aegean where recent intensive surveys have also shown an alternation between a nucleated and dispersed pattern of settlement (Sbonias 1999a:7). For example, in the western Khabur Valley of northern Syria archaeological sites can be classified into the following morphological categories: (a) small, low sites of 1–5 ha and < 5 m height; (b) tells that are > 5 m high and can be from 0.5 ha area and larger; and, (c) extensive low sites, usually adjacent to tells, referred to here as lower towns. When these different site categories are plotted according to their ceramic phase of occupation (Fig. 4.2), we see that settlement has fluctuated from a phase of mixed small and tell sites during the Neolithic and Chalcolithic periods, followed by a marked phase of nucleation during the third millennium BC at which time settlement was concentrated almost exclusively on tells. There then followed a complex transition during the second millennium BC that culminated during the Middle Bronze Age or Late Bronze Age in the decay of the pre-existing centers, the appearance of small lower towns, and the initial dispersal of smaller settlements into the countryside. This process of dispersal was well underway by the Iron Age, especially during the period of the Neo-Assyrian Empire, at which time many parts of the Jazira and west-central Syria were settled by a rash of rural settlements. Although there is a significant degree of regional variation, this substantial growth in small rural settlements can be recognized for many parts of northern Syria and northern Iraq. It appears that this process of settlement dispersal was initiated in the early first millennium BC, or slightly earlier, but various manifestations of it continued during the later first millennium BC, the first millennium AD, and later. It is this Iron Age and later pattern of small dispersed rural settlement that has been severely under-represented by traditional surveys. This is particularly problematic because Neo-Assyrian texts provide a detailed record of both rural settlement (Fales and Postgate 1995) and deliberate schemes for resettlement of displaced populations on often-marginal land (Oded 1979). Such records suggest that Iron Age settlement consisted, in part, of a widespread dispersal of small farmsteads and villages. Whereas traditional mound surveys either missed most of these sites, or grossly under-represented them (except Oates 1968), more recent surveys have demonstrated that most early first millennium BC settlements were small and were widely dispersed across the landscape

THE DISJUNCTION BETWEEN MEDITERRANEAN AND NEAR EASTERN SURVEY

Figure 4.2. Tells and small sites plotted according to archaeological periods from the area of Tell Beydar, eastern Syria.

(Morandi 2000; Wilkinson and Barbanes 2000). Recent survey records therefore bring a measure of harmony to the archaeological and textual data. Similar processes of dispersal can be recognized in western Syria, although in addition Iron Age dispersal is frequently accompanied by continued settlement on tells (Sader 2000:68). This mixed pattern of settlement contrasts with that found in eastern Syria and northwest Iraq where during the Iron Age settlement on tells was eschewed in favor of lower towns distributed around the base of tells. Although this settlement of lower towns may be partly explained by the convenience of siting houses at the foot of a tell, especially when political conditions were more stable during the phase of Neo-Assyrian administration, it was also apparently deliberately encouraged by the Assyrian authorities themselves (Fales 1990). This may be because tells as defensible strongholds formed centers of power that may have represented a threat to the authority of the Assyrian regime. It is therefore possible to speculate that in western Syria, in areas which were probably more loosely controlled by the Assyrian authorities, it seems that there was greater autonomy and the inhabitants were able to live on their traditional settlements located on tells. Consequently, for the Iron Age in western Syria tell-based survey will probably under-represent somewhat the settlement pattern by only indicating those settlements that were on tells. In contrast, in the Jazira, tell-based survey gives a totally misleading view of Iron Age settlement. A noteworthy feature of the Iron Age is that parallel patterns of settlement dispersal appear to have occurred in very different regions and apparently for different reasons. Thus the so-called Israelite settlement of the southern Levant

61

62

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 4.3. The number of archaeological sites in the Jabul survey area of north-central Syria.

Figure 4.4. The number of sites according to archaeological periods excavated and surveyed in the upper Tabqa area, Syria.

appears to resemble the pattern of dispersed settlement of the Syrian/Iraqi Jazira, albeit for quite different reasons. In this region, at the end of the second and the beginning of the first millennium BC there was a vigorous settlement of, for example, the hill country west of the Jordan (Finkelstein and Lederman 1997) and the Acco plain (Lehmann 2001) that resulted in the establishment of numerous small settlements on formerly sparsely settled uplands. Again surveys of moderate intensity have resulted in the retrieval of settlement patterns that would otherwise have been missed by conventional surveys focused on lowlands. In western Syria, parts of southern Turkey, Lebanon, as well as much of the southern Levant, the pattern of dispersed settlement that commenced in the Iron Age (e.g., Finkelstein 1998) continued into the Hellenistic, Roman, and Early Byzantine periods (Anderson 1998:451). As noted by Alcock for the southern Levant

THE DISJUNCTION BETWEEN MEDITERRANEAN AND NEAR EASTERN SURVEY

(1994:182–83), there was, however, a considerable amount of local and regional variation during the Hellenistic period. Nevertheless, in northwest Syria and parts of southern Turkey, there had been a massive increase in small settlements by Roman times. This is evident in the Amuq plain (Yener et al. 2000), the Jabbul (Fig. 4.3 based on Schwartz et al. 2000), and the Biqa valley (Marfoe 1979: Fig. 9), as well as along the Euphrates Valley (Fig. 4.4; Algaze et al. 1991: Fig. 26; Wilkinson 1990: Fig. 6.2). In addition to their low survey intensity another biasing factor in traditional Near Eastern surveys is that they concentrate upon extensive plains (e.g., Braidwood 1937; Marfoe 1978) at the expense of the surrounding uplands. Although it is arguably true that the bulk of Bronze Age settlement may have been concentrated on the broad expanses of cultivable plains, this is not true for later periods. For example, the pattern of upland Late Roman/Byzantine settlement, well known from the magnificent studies of Tchalenko (1953) and Tate (1992), are still biased because they were primarily focused upon the periods that were well represented architecturally, namely the Late Roman–Early Byzantine periods. Consequently remains earlier than the Hellenistic were hardly recorded (if at all). Although it is clear that there was a major phase of upland colonization during the Roman/Byzantine period, the scale of earlier occupation remains uncertain. Further clarification of this process of dispersal is supplied by Marfoe’s surveys and analyses of the Biqa Valley, Lebanon, which suggest that there was a substantial upsurge in settlement on piedmont slopes in the Hellenistic and Roman periods (Marfoe 1978:625–59). However, because Marfoe’s surveys did not penetrate into the uplands themselves, he was not able to fully describe the extent of this settlement. This issue is now being clarified by recent surveys of the Amuq Valley Regional Project, which demonstrate not only the degree of increase in settlement during the Hellenistic-Roman and Byzantine periods, but also that it represents a

Figure 4.5. The extension of Roman, Byzantine, and Islamic rural settlement into the upland behind Antioch.

63

64

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

significant expansion away from previous Bronze Age and Iron Age settlements, which largely were rooted in the valley floors (Fig. 4.5; Casana 2003). The foregoing sketch of a large body of available data suggests that if one is interested in the Early Bronze Age of Upper Mesopotamia, a well-structured and careful full-coverage survey may provide an acceptable picture of the prevailing settlement pattern. As a result, when earlier extensive tell-based surveys have been resurveyed using more intensive techniques, there has been little resultant change in the settlement pattern of the Early and Middle Bronze Ages. For example, the important survey of the Balikh by Peter Akkermans (1993) provided a good record of the Bronze Age and earlier periods but was less successful for the Iron Age and later periods when smaller, dispersed settlements became the norm. Consequently, resurvey resulted in the recovery of many new sites (Wilkinson 1998: Fig. 7). With the addition of the off-site component the recognition of archaeological activity becomes a more complex process. This is because off-site sherd scatters are produced by a range of non-settlement factors, in many cases being attributable to the application of manure or composts. In Upper Mesopotamia, where according to the contained ceramics such field scatters are dominantly of Early Bronze Age date, they form concentrated scatters that extend up to several kilometers away from major tells (Wilkinson 1994). On the other hand, in western Syria and southern Turkey, as well as in southern Mesopotamia, field scatters are significantly later in date. For example, in the Amuq plain, along the Turkish Euphrates, and in the lower Balikh Valley in Syria, field scatters seem to be mainly of Hellenistic, Roman, or Early Byzantine date (Wilkinson 1990; 1998), and in southern Mesopotamia they contain Seleucid, Parthian, and Sasanian sherds. When small sites occur within such field scatter zones they are often barely recognizable within the dense “background noise” of off-site material; therefore, unless sampling schemes are designed to disentangle such scatters, they tend to either be missed or confused with low-density occupations. In the Amuq plain of southern Turkey surveys have been designed to deal with this more elusive component of the offsite record (Verstraete and Wilkinson 2000), but it is clear that the recognition of the lowest level of the site hierarchy for these periods requires a much more intensive level of survey and sampling than has hitherto been practiced. Overall, even though the traditional Bronze Age tell survey may produce reliable results for sedentary settlement, these techniques entirely miss the off-site component of the Bronze Age landscape. A similar problem relates to the mobile element of settlement that occurs in the form of pastoral camps, and the ephemeral traces of these sites continue to be under-represented in much of the northern Levant and Upper Mesopotamia (but see Danti 2000 for promising new results).

Discussion Owing to the significant differences in the intensity and application of field survey it could be argued that a fault line does indeed exist between Near

THE DISJUNCTION BETWEEN MEDITERRANEAN AND NEAR EASTERN SURVEY

Eastern and Mediterranean surveys. This is an oversimplification, however, because in reality survey techniques in the Middle East can be differentiated into those that focus on extensive plains that are sprinkled with tells versus areas of broken terrain that tend to be characterized by more dispersed settlements. Therefore, throughout the Levant (both north and south) surveys in the uplands fringing the eastern Mediterranean coast have necessarily started to adopt the methodologies of Mediterranean surveys, whereas those covering the tell-dominated plains further to the east in northern Syria and Iraq have usually maintained the techniques of traditional extensive full-coverage surveys. Therefore, broadly speaking, the problems posed by the terrain as well as the sites themselves have partly determined the types of techniques employed. This does not mean that Levantine surveys can claim to be as intensive or provide as accurate a representation of the settlement record as those further west. Nevertheless, the results of recent surveys in the Levant are following a trajectory not unlike those developed further to the west. Survey results continue to be difficult to interpret because frequently when surveys in different terrain types are compared, these result in very different trends in settlement. For example in southern Turkey, surveys focused on the uplands result in a strong bias toward later (i.e., Post-Hellenistic) settlement (e.g., Blanton 2000), whereas those that concentrate on lowlands exhibit a stronger bias toward Bronze and Iron Age settlement. Consequently when reviewing settlement trends in such areas it is crucial to recognize that not only were different processes of settlement taking place in different periods (viz. dispersed settlement into the uplands in the Post Iron Age), but also the plains and uplands represent two very different taphonomic zones. It is a challenge for future surveys to attempt to disentangle these two interrelated processes. How do we therefore compare the results of other surveys with those being undertaken around and within Mesopotamia? First, many of the tell-based surveys do indeed seem to provide a reasonable picture of settlement for the Early Bronze Age and locally for the Middle Bronze Age too. However, this is at the expense of the off-site record as well as settlement on the uplands and pastoral nomadic settlement in the semiarid margins (but see Bernbeck 1996). Overall, in the Near East traditional surveys have proved to be woefully inadequate for accurately representing settlement of the Iron Age, Hellenistic, and later periods. For these periods settlement dispersal started during the second half of the second millennium BC, so that by the Iron Age many areas seem to have experienced a very pronounced extension of settlement into the steppe or uplands. Although this process did not take place throughout the Near East it has been recorded in much of the Jazira (Morandi 2000; Wilkinson and Barbanes 2000), the Jabbul plain (Schwartz et al. 2000), and the Levant (Finkelstein 1997). Furthermore, during the Hellenistic, Roman, and Late Antique periods the development of a rural landscape dominated by a dispersed pattern of rural settlement became even more pronounced in the southern Levant (Broshi 1980), the northern Levant

65

66

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

(Yener et al. 2000; Schwartz et al. 2000), as well as the Euphrates region of southern Turkey and parts of northern Syria (Wilkinson 1990; Algaze et al. 1994). The recent surveys support and locally amplify Alcock’s statement that for the Hellenistic period there was an increase in settlement numbers and associated population (Alcock 1994). However, such an increase is more evident in the northwest Levant, because further east in Upper Mesopotamia any such increase appears to have taken place during the Iron Age, in part as a result of the dramatic series of deportations initiated by the Neo-Assyrian kings (Wilkinson and Barbanes 2000). Recent surveys undertaken at a greater level of intensity and with more regard for smaller sites now present a picture of a densely settled rural countryside during the late first millennium BC and early first millennium AD for much of the Levant and northwestern Syria and southern Turkey (Casana 2003). In many areas this now resembles the “busy countryside” described by Lloyd for Italy (1991; see also Barker 1996). Although it may be that in certain areas such a growth of rural settlement and rural agricultural production may indeed have been associated with the implantation of new cities by the Seleucid rulers (Grainger 1990:110–19; also Alcock 1993:181), in other places this pattern of dispersed settlement appears to have been built upon foundations that were laid during the Iron Age (e.g., northern Iraq: Wilkinson and Tucker 1995). The striking change in settlement structure that took place between or during the mid–late second millennium BC and the first millennium BC appears to represent a major shift from some form of “city state” type settlement system during the Bronze Age to the territorial empires of the first millennium BC and later. Such changes, which were probably associated with the collapse of the socio-economic pattern of the Late Bronze Age, may be linked to wide-ranging changes in land tenure, as well as shifts in the exercising of power and population redistributions. However, before such links between settlement structure and the political economy can be established it is crucial for settlement surveys to be conducted in both lowlands and nearby uplands to a rigorous and standardized methodology. It is also necessary to take into account differences in landscape taphonomy (both physical and cultural transformations), as well as the dynamics of the different types of landscape. In the case of landscape dynamics, landscapes dominated by tells often appear to be remarkably stable so that not only do the tells themselves remain as fixed points in the landscape—sometimes for millennia—but also landscape features can become concentrated or “embedded.” As a result, features such as “linear hollows” or off-site sherd scatters become recognizable despite their subtlety. On the other hand, settlement landscapes that comprise smaller dispersed settlements often exhibit a very different dynamic with sites shifting from place to place every few generations or centuries. Consequently, one might expect any associated offsite features to be less recognizable, although this depends on the age of the phase of settlement as well as on a number of cultural factors.

THE DISJUNCTION BETWEEN MEDITERRANEAN AND NEAR EASTERN SURVEY

To conclude, it is interesting to reflect why there has been such a neglect of the rural settlement pattern in many parts of the Near East. In part this bias appears to be a product of the period of interest of the researcher. In Upper Mesopotamia archaeologists whose primary interest was the settlement of the Early Bronze Age have conducted numerous surveys, and in many areas these rather coarse-grained surveys appear to have done an adequate job. Sites of the Classical, RomanParthian, and Sasanian-Byzantine periods have usually been under-represented by mound-based surveys because they fall outside the period of interest of the Bronze Age specialists. Also, rather conveniently, the survey methodologies chosen have allowed such late “annoyances” to slip through the survey mesh. In contrast, in Greece and much of the Mediterranean the Classical and Hellenistic periods have occupied a more central position in the academic research agenda and consequently there has been more of an incentive to provide an accurate measure of these rural landscapes. Nevertheless, the tide has now turned and the recent appearance of a spate of new volumes on the Iron Age and later periods appears (we hope) to be ushering in a new phase of regional interest in the Near East (Bartl and Hauser 1996; Bunnens 2000; Hausleiter and Reiche 1999; Liverani 1995). Clearly, despite the tendency for traditional surveys to adequately represent the pattern of Bronze Age settlement, Near Eastern surveys need to become more intensive in order to provide a better representation of long-term trends in settlement.

Acknowledgments This chapter was originally presented as a paper at the meeting “Crossroads in Mediterranean Landscape Archaeology,” organized by Effie F. Athanassopoulos and LuAnn Wandsnider, at the Society for American Archaeology (April 18–22, 2001), in New Orleans. A related paper on a more developed theme was then presented at a meeting hosted by John Cherry and Susan Alcock in Ann Arbor, Michigan, in April 2002. I am very grateful to the organizers of both meetings for working to make these meetings a success and for bringing them to publication. I also wish to thank Jesse Casana and Jason Ur of the Oriental Institute CAMEL Lab, who contributed ideas and data to versions of this chapter.

67

This page intentionally left blank

5 Artifact, Landscape, and Temporality in Eastern Mediterranean Archaeological Landscape Studies LuAnn Wandsnider

I

ntensive survey over the last several decades has detailed an archaeological surface record in the Mediterranean that Cherry (1983:395, emphasis in original) describes as “likely to consist of a virtually continuous spatial distribution of material over the landscape, but a distribution extremely variable in density.” In addition, geoarchaeological work, often coupled with survey, has demonstrated just how dynamic Mediterranean surfaces have been. Both of these field practices, intensive survey and geoarchaeology, were carried out in part to enable regional settlement pattern studies, to collect accurate, reliable, and precise data about past settlements and their location with respect to each other and with respect to aspects of the landscape (Cherry and Shennan 1978). As a result of intensive surface survey and geoarchaeological work, a paradox has become apparent. This paradox is that the surface archaeological record is of such a quality that the settlement mode of interpretation, the impetus for the high quality work undertaken in the Mediterranean, may be inappropriate. That is, it demonstrates that rather than a record of settlements awaiting discovery and definition by the archaeologist, the archaeological record is better considered a record of places with different material histories, both cultural and natural. Viewed in this way, the surface record, currently interpreted in terms of settlements and using empirical, historical conventions, instead becomes a source of information on the human condition at a variety of spatial and temporal scales. Below, I make the case for this paradoxical situation (see also Holdaway and Wandsnider n.d.; Wandsnider 2004; Wandsnider and Holdaway n.d.). Archaeologists working elsewhere and not explicitly concerned with the archaeological landscape have come to a similar realization, concluding that a metaphysical shift in the approach to archaeological deposits appears warranted. Given this shift, how should interpretation of Mediterranean archaeological landscapes proceed? Two additions to the archaeological tool kit are suggested: a library of potential multi-temporal processes and a series of three interpretative tools.

70

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

A Paradox in Mediterranean Archaeological Landscape Studies Over the last several decades, the Mediterranean surface record primarily has been studied using the interpretative vehicle of regional settlement patterns (Barker et al. 1999; Kardulias 1994a; Keller and Rupp 1983). Systematic surface survey is the main tool used to collect data on the location and type of past settlement. Settlement type is determined using common sense and correlational dating based on ceramics establishes the time period or periods of occupation (Binford 1992). The questions driving such studies have been mostly implicit rather than explicit (Jacobsen 2000; Morris 1994), embedded in a tradition of scientific humanism, wherein the “universal man,” as he interacts with his environment, is the object of scrutiny (Fotiadis 1995). Observed settlement patterns have been interpreted empirically and rely on an historical concept of agent, time, and causation wherein settlement patterns are taken to reflect culture history, rendered with increasing sophistication (Jacobsen 2000:5). As discussed in the introductory chapter, regional settlement pattern studies have been avidly pursued throughout the world, no less so than in the Mediterranean. But, the surface archaeological record and the tool of systematic surface survey, used to find and document sites (usually inferred to represent settlements of some sort), were, in the Mediterranean, explicitly recognized as problematic. In a seminal paper on archaeological survey in the Mediterranean, Cherry (1983:379) summarized the recognized intractabilities of survey data employed in settlement pattern studies: a. The necessarily coarse chronological framework, seldom more precise than a century and often much vaguer than that, which results from the use of aggregate collections of poorly preserved surface material, so that b. maps of site distributions based on survey data to some degree (often unknown) must be taken to represent “palimpsests” of sites, not all of which were necessarily in use simultaneously… c. The smearing and blending of surface finds, whether by natural or human agencies, means that small sites may often go unrecognized and sites of all sizes and types may be difficult to define accurately in spatial terms. d. Information about the internal organization and function of sites is usually very difficult to obtain.

These limitations of the surface archaeological record and the survey product were cited by Hope Simpson (1983) as an indictment of stand-alone survey as a method for discovering and reporting on past settlements and of the conduct of intensive regional settlement studies. Cherry (1983), in response, offered a number of prescriptive measures to ensure and understand data consistency and quality, including an emphasis on intensive survey and methodological refinements. Many

ARTIFACT, LANDSCAPE, AND TEMPORALITY

of these measures were adopted by Mediterranean practitioners in the “New Wave” (Cherry 1994) of regional studies. For example, Gallant (1986) offers a method for reliably defining sites relative to regional sherd density (see also Cherry et al. 1991; Given et al. 1999). Numerous surveys (e.g., Barker 1995b; Bintliff et al. 1999; Given et al. 1999; Van Andel and Runnels 1987) are concerned with the degree to which evidence for sediments are buried or eroded away, leading to an incomplete picture of settlement. It is in part this latter concern, along with that of paleoenvironmental reconstruction, that led to the incorporation of a geomorphologic specialist on the field team. Reflecting on the state of the discipline in 1994, Cherry noted encouraging signs in that previously ignored cultural remains were now receiving the attention that Prehistoric and Classic rural domestic landscapes had earlier received. Thus, ritual activities, urban centers, and the Late Antique were now also being approached using intensive survey. But, he (Cherry 1994:103–105) also identified several continuing problems. That is, progress had been made in finding sites, but classification of sites according to functional type using surface materials was proving difficult. Targeted excavation and geophysical work were suggested as possible remedies. Also, archaeologists had documented numerous “off-site” artifact scatters, but the treatment, reporting, and interpretation of those off-site scatters he found to be widely divergent. Cherry (1994:104–105) noted that “surprisingly little progress has been made toward a consensus about the taphonomic processes that produce such distributions, or toward an understanding of why in some regions artifacts are to be found everywhere, while in others they are relatively rare outside of ‘sites’.” Mediterranean archaeologists continue to soldier on using the regional settlement vehicle of interpretation, but acknowledging the difficulties of their database in doing so. For example, Cherry and colleagues (1991:34) and Given and colleagues (1999:23) explicitly acknowledge that in designating “sites” from deposits with complex formational histories and in moving from “sites” to “settlements,” they are engaging in interpretation, in the conscious construction of analytic units using explicitly detailed criteria. Other regional studies have taken one further step. Having identified sites through survey and interpreted them as settlements, Bintliff (1999b) and Sbonias (1999a; 1999b) go on to identify another series of problems that attend the use of archaeological settlement data to address issues of regional demography. They offer a series of conventions to be used to surmount these “methodological hindrances” (Binfliff 1999b:21) and recognize that their demographic estimates must be considered provisional. The paradox is that through the application of technically sophisticated intensive survey and attention to geological surface processes, the problems they were intended to solve—the spatial definition of sites, of chronological assignment, and of functional interpretation—have not disappeared. Indeed, sophisti-

71

72

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

cated survey work sometimes coupled with targeted excavation in the general eastern Mediterranean (e.g., Barker 1995a; 1995b; Cherry et al. 1991; Given et al. 1999; Van Andel and Runnels 1987) throws into sharp relief the stubborn nature of these difficulties. John Chapman explicitly discusses these difficulties. For example, when relatively higher sherd densities are observed at a site, does this indicate more intensive deposition, a higher site population or nucleation, or a longer occupation span (Chapman 1999:69)? And multiple sites, dated with coarsely resolved ceramics, are often treated as contemporaneous enough to permit settlement reconstruction, when in fact they may not be. Indeed, in at least two cases, high-resolution radiocarbon dating of associated buried sediments revealed the complex formation histories of surface palimpsest deposits (Chapman 1999; Whitelaw 2001), emphasizing that “contemporaneity” is very much dependent on the resolution of the measuring device, in this case, ceramics; the degree of contemporaneity amongst and between archaeological deposits may not allow for the valid reconstruction of entities like settlement systems. These observations as well as the fact that surface archaeological deposits are still considered problematic even after the application of sophisticated technologies and methods for finding, parsing, and interpreting sites and settlements suggests that perhaps the problem lies not with the surface archaeological deposits but with the vehicle used to interpret those deposits. Below, I suggest abandoning the regional settlement studies vehicle and embracing the complex formational nature of archaeological landscape deposits.

Time in Archaeological Landscape Studies The abiding technical issues alone are enough to challenge the regional settlement pattern studies vehicle, but another challenge comes from the ontological domain. Here, the empirical, historical, reconstructive mode of interpretation commonly utilized to interpret the formationally complex landscape is threatened. The issue again ties back to the palimpsest (Cherry 1983; above) nature of the archaeological landscape. As Binford (1981b) emphasized in his discussion of the Pompeii premise, the archaeological record is qualitatively different from the materials and behaviors observed during quick-time ethnographic study (see also Foley 1981a; 1981b). It is denser than that documented during short-span ethnoarchaeological observation. This being the case, Binford and Foley argue that such apparently temporally thick deposits are owed to deposition over generations and, thus, are ideal for studying longer-term organization and structure, rather than short-term behavioral events (see also Clark 1994). Over the last decade, these powerful observations have been rendered more nuanced. Rather than being taken as only referable to longer-term phenomena, archaeologists are exploring the potential of archaeological deposits for informing on a variety of processes operating at a variety of tempos, from short (behavioral),

ARTIFACT, LANDSCAPE, AND TEMPORALITY

to medium term (ecological, demographic), to longer term (geologic, tradition, or mentalité). “Time perspectivism” is the label applied to this pursuit (Bailey 1987; Knapp 1992a; Murray 1999) and two expressions of it are evident. Bailey (1983; 1987), Butzer (1982), and Dewar and McBride (1992) independently offer various hierarchical temporal schemes derived from biological or ecological domains. Drawing from history, Barker (1995b), Bintliff (1991a; 1999a), and contributors to Knapp (1992b) recruit Braudel’s schema of événementsconjoncture-longue durée (see Athanassopoulos, this volume) as another way to organize temporal interpretations. Attempts at multi-temporal interpretation to date have produced tantalizing results. For example, the account for Boeotia from 600 BC to AD 700 (Bintliff 1991a) presents different data classes (texts, regional site distributions, erosional sequences) and recognizes that they are sensitive to processes with different temporalities. The intersection of processes with different temporalities is not considered, however. Similarly, multi-scalar temporal processes are discussed generally for the Biferno valley, with important geographic differences emphasized (Barker 1995b). While the spatial scale of analysis changes through time, hierarchical explanation, i.e., a consideration of the interplay between processes operating at different tempos, is not attempted. In part, the difficult task is how to present history at multiple temporal and spatial scales. While anti-historicist in orientation, but not explicitly embracing time perspectivism, Olivier’s (1999) analysis of the Iron Age “Princely” grave at Hochdorf provides a useful model. Here Olivier focuses on three different classes of archaeological material (artifacts, the interred person, and the funerary monument) and recognizes general processes operating at different tempos. The first concerns the life cycle of objects deposited in the grave, their manufacture, use, and burial, and operates over a few years to a few decades. The second concerns the recognition of the status of the deceased, from death to burial, and operates over the span of a few days to a few weeks. Finally, the relationship of the living with the dead person, extending from the erection of the funerary monument to later re-occupation or rearrangements, lasts over generations and perhaps for centuries. The important point is that Olivier creates and uses interpretative units that draw on the strength of the archaeological materials but also are sensitive to the processes under study. Other multi-temporal interpretations rely too fundamentally, I suggest, on interpretative units that are quasi-ethnographic entities, i.e., occupations, sites, and settlements. In light of our formational understanding of archaeological materials, Smith’s (1992:29–30) observation is germane: we study not Mayan “house floors” but rather Mayan “house floor series.” Similarly, our sites are not settlements; rather, they may be settlement series. As he also notes, exactly how to interpret a house floor series and, by extension, a settlement series, remains underexplored. Consideration of a multi-temporal model of agent, time, and causation is nonetheless especially important for Mediterranean archaeological landscape

73

74

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

studies because it forces archaeologists to abandon the ready empirical, historical interpretive crutch (Clark 1994; Olivier 1999). That is, it opens the doors to supra- (e.g., the Annales mentalités) and sub-historical (e.g., mortuary practices) processes, which will require theoretical and methodological attention in order to be studied archaeologically (see Hill, this volume). It also emphasizes the likelihood of complex processual linkages (McGlade 1995), for example, between deforestation and political cycles, as well as hierarchies of causation (Bailey 1983).

A Metaphysical Shift In the preceding, I have described the paradox that exists between our current formational understanding of the archaeological landscape and how it is documented, in terms of units that are functional, synchronic, and quasi-ethnographic entities. Equally frustrating, by approaching the archaeological landscape using the settlement pattern vehicle, archaeologists are locked into interpretive schemes that are historic or ethnographic in nature. Other interpretative temporalities are difficult to consider. All of this argues for a metaphysical shift in how archaeological landscapes are conceptualized and approached along at least three related dimensions. The first is from a focus on people and individual activities to that of places or landscapes (Binford 1982; Smith 1992). Most archaeological settlement pattern survey is predicated on the notion of recovering or reconstructing common sense, ethnographic entities like settlements. Such entities make sense in the lived lives of humans. But, the material artifactual and landscape constructions of humans have lives very different from their creators. It is these materials, especially the landscape, that become the focus. As a result of place use with different intensities and deliberations, a cultural landscape of places was (and is constantly) constructed (Tilley 1994). To see place in space, to map the relief of the cultural landscape, we cannot only focus on sites and settlements: a landscape perspective is necessary. Related to the above is Dunnell’s (1992) case for a shift from a site-based approach to an artifact or feature-based approach to archaeological landscapes. Dunnell argues that each land parcel has its own formational or taphonomic history. Moreover, sites, dense concentrations of archaeological materials, are too coarse a unit with which to decompose that history. Rather, by focusing on artifacts, artifact assemblages, features (and surfaces, Sullivan 1978), the possibility of tracing out formation histories for individual (arbitrarily defined) landscape parcels exists. A third dimension is related to the nature of material, be it artifact or landscape. We understand that materials participate in human lives in both tangible and intangible ways, that they are manipulated by people and that, in turn, people are manipulated by them. They have mass, are traversed, are differentially

ARTIFACT, LANDSCAPE, AND TEMPORALITY

mutable, and have varying, culturally specific capacities for performing in people lives. Thus, it is too simplistic to argue that people exploit or are constrained by their environment and it is equally too simplistic to say that people create the landscape. Rather, people, artifacts, and landscapes are in a constant dialogue that some (Green 1997; McGlade 1995; 1999a) argue is unparseable. But, of course, as analysts, we can approach the artifacts and the cultural landscape however we wish, if it buys us insight. In sum, the overall metaphysical shift is from that of a focus on quasi-ethnographic kinds (sites and settlements) to that of formational (and therefore, historical) material variation, or from so-called essentialism (Dunnell 1980; 1986; Lyman et al. 1997; Ramenofsky and Steffen 1998) to ontological materialism (but see Binford 1992 for comment on these terms). This metaphysical shift registers in the consideration of Mediterranean archaeological landscapes: where formerly functional, ethnographic, and historic models of agent, time, and causation were embraced, today we see increasing use of multi-temporal, multi-processual modes of interpretation.

Studying Mediterranean Archaeological Landscapes It is all very well to conceptualize surface archaeological deposits as palimpsest deposits owed to the operation of multiple processes (cultural and natural) with different temporalities. It is an altogether different task to interpret the contingent human-nature trajectories from those palimpsest deposits. To do so requires several tools, which are only partially developed. Here, I focus on two, both of them middle range (Binford 1981a) in nature. The first is concerned with developing a library of human and natural processes, the second, with taking advantage of the material histories of artifacts, features, and surfaces to interrogate landscape formational histories. Smith (1992:25) notes that Braudel’s temporal schema with its hierarchy of different sociocultural processes and constraints was empirically derived from his historical studies of the Mediterranean. McGlade (1999b; see also Knapp 1992a) suggests that such processes (along with biological and ecological processes) have their own inherent tempo. We currently have only a shadow of an understanding of human and natural processes with known ranges of temporalities, spatial ranges, and collateral processes. For the Mediterranean, such a library might contain information on the rebound rates of forests under range of different conditions, something currently under construction by landscape ecologists working there (Goldammer and Jenkins 1990; Naveh and Lieberman 1994). Palynologists working in concert with other paleoenvironmental scientists may provide the longer view of this process (Bottema 1991; Bottema and Woldring 1984). Other processes include the dynamic and modes of colonizer-colonized relationships. For example, Elton (1996) and Wells (1999) summarizing Roman occupation

75

76

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

in the provinces describe a similar pattern. Roman legions were garrisoned in newly acquired territory until rebellion was unlikely, usually a generation in length. At this time, legions were relocated to other areas. With historical data on colonizer-colonized dynamics from a series of different contexts, along with attending studies on the associated effects of different colonizer-colonized relationships, some of the dynamic we likely see archaeologically may become interpretable. Here historical documents are essential for building this body of reference knowledge (Sabloff 1986; Smith 1992) and Murray (1997) suggests other research tactics. The process library is important not purely because we desire an archaeological signature of particular processes. Such a library is necessary before we can begin to determine the temporal or spatial scales at which archaeological observation can occur and at which the interaction between and among other processes becomes possible (Bailey 1983). For example, Rackham and Moody (1996; see also Moody 1997) argue that it is not “climate” that is responsible for Early Antiquity erosional deposits documented in eastern Mediterranean contexts but rather “weather,” i.e., spatially and temporally unpredictable local rainfall events. This is a debate with many facets: on the nature of the relationship between “climate” and “weather,” between precipitation events and slope failure events, and between the archaeological proxies and the systemic realm phenomenon they measure. This is a relationship that we can understand in the here and now, when populations of slopes and potential slope failures are considered (which is what regional documentation is so good at providing). The second important addition to the interpretative tool kit includes inferential tools based in material history, the strength of archaeological deposits. Such tools depend on our understanding of the elements of archaeological deposits— artifacts, assemblages, features, and surfaces—as differentially performing and differentially enduring time travelers. Artifacts may have been designed to have relatively short use-lives (e.g., ceramic tea cups used by travelers on the Indian Rail System before the advent of paper and plastic cups) or very long use-lives (e.g., burial vessels). They may be designed to be transported over great distances (amphorae) or as furniture essentially remaining in one place (pithos). They may be designed to signal social status (as in blackwares employed in lieu of iron wares). And, their archaeological signature, i.e., sherds, is related in systematic ways with the original vessel; that is, we can identify an attending Mickey Mouse Law (Flannery 1973) with important implications: small vessels break into fewer sherds per vessel than do massive vessels, other things being equal (Fentress and Perkins 1989). The design and then the actual use to which vessels are put determine the rate at which individual artifacts enter archaeological deposits (Ammerman and Feldman 1974; David 1972; DeBoer 1974). Thus, comparing relative frequencies of sherds from vessels with different sizes and also with different uselives may permit a closer reading of residential occupation dynamic (all other things, e.g., scavenging, visibility, being equal). That is, we expect a basal

ARTIFACT, LANDSCAPE, AND TEMPORALITY

frequency of storage vessels and roof tile sherds. However, the frequency of sherds from short-lived vessels, with greater opportunities to be introduced into residential deposits, should be sensitive to the occupation dynamic (e.g., Varien and Mills 1997). In contrast to that discussed below for features, the iconographic capacity of one sherd is likely low. The iconographic capacity of sherd aggregates remains to be fixed. Given (this volume) presents information that such aggregates have a monument-like iconographic capacity for contemporary Cypriot farmers. Artifacts differentially accumulate on surfaces and in deposits because of the dynamic associated with their manufacture, use, and emplacement, loss, or discard. Features such as walls, houses, temples, and baths, on the other hand, reflect a very different dynamic. They are introduced to the landscape where, depending on construction material, they have the potential to be almost eternal. This enduring capacity means that they have the potential to serve as temporal buoys against which subsequent activities may be monitored. That is, once constructed, they offer the potential to see how such features were maintained, modified, or destroyed. For example, in Hellenistic and Roman Western Rough Cilicia (southern Turkey), we see various building trajectories, temples converted to baths, tombs converted to temples, and so forth. Presumably the directionality of the conversion is not random, but such remains to be explored. Features may be deliberately or not imbued with great iconographic capacity. This capacity, in turn, may make them lightning rods for how subsequent occupants react to extant monuments. In a more recent context, Chapman (1994) reports on the deliberate destruction of monuments with local historical significance (see also contributors to Bradley and Williams 1998). Because of their enduring material qualities, features offer a unique means to perceive events that likely occurred outside the modal range of human behaviors (Wobst 1978), that is, very rarely and under great stress. And, feature history, documented through its construction, modification, and destruction using simple but powerful archaeological observation on sequencing (Sullivan 1992) and perhaps rendered using Harris-matrix-like (Harris 1979) sequence models (e.g., Bleed 2001), provides insight not otherwise obtainable. Moreover, their initial construction indicates that a certain threshold has been breached. That is, the construction of a fortification wall variously signals the presence of a threat, the nature of that threat in light of extant siege craft (McNicoll and Milner 1997), that someone or something is to be protected, that labor and resources existed to construct that wall, and/or that a political infrastructure to organize construction exists. As Fletcher (1995) documents for the growth of cities, once constructed, the very materiality of features constrains subsequent modifications, removals, and so forth. People may well endure the inconveniences and worse of extant features because of the high cost of remodeling them to conform to current needs. At some point, however, another threshold may be breached, wherein those high costs become tolerable or necessary.

77

78

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Surfaces are similar to features in that they potentially host and remember different place histories (Sullivan 1978). They are places where artifacts and features accumulate or not. Their use-lives are dictated by geomorphological activity and perhaps also by human activities that differentially affect the quality of surfaces as work and living spaces. For example, surfaces are created and maintained as agricultural terraces are constructed. On these surfaces may accumulate the incidental artifacts associated with agriculture as well as those owed to deliberate manuring. Should the terraces no longer be maintained, the potential exists for some surfaces to become eroded and others buried as terrace walls fail. Like deliberately installed features, surfaces offer a slate on which the hand of man and nature may write. Given these elements with differential sensitivities to process, how shall archaeological documentation and analysis proceed? Current practice emphasizes reliably tallying quantities of things (sherds and so forth) because these somehow (and these linkages are not often specified) tell us about numbers of people (relatively or absolutely). But, the synchronic premise that undergirds such practice conflicts directly with the multi-temporal understanding of palimpsest deposits articulated above. In light of this understanding of the unique information capacity of archaeological elements, three kinds of inferential tools have been used by archaeologists. The first and most common, point indicators, rely on the presence/absence of particular material remains to inform on particular possible conditions. The presence of an African Red Slip sherd in interior Rough Cilicia indicates some kind of relationship with the greater Mediterranean market. An assembly house at Asar Tepe (Rough Cilicia), for example, signifies that at least at one point in time, the multifarious conditions supporting the creation of an assembly hall existed. Similarly, the presence of baths and fortification walls signal that yet other conditions are in place at some restricted point in time. Point indicators are typically used comparatively—on which hilltops are fortified cities found—or, populations of point indicators—numbers of African Red Slip sherds—are also informative. Span indicators, on the other hand, refer to assemblages of artifacts or features that accumulate over a particular span of time. An assemblage of pristine features would suggest a very short occupation span; an assemblage of features showing repeated modification would suggest a rather longer or more intense occupation span. An assemblage composed only of domestic ware sherds, again, reflects a brief domestic occupation; an assemblage also containing storage sherds suggests longer domestic occupation. The relative ratios of material culture with different life spans offer another way to understand the prevailing conditions over the long term. A number of Prehistoric archaeologists (Bamforth and Becker 2000; Binford 1977; Holdaway et al. 2004) have explored the utility of span indicators derived from time-averaged (Stern 1994) chipped stone assemblages. The same needs to be attempted for assemblages from more complex societies.

ARTIFACT, LANDSCAPE, AND TEMPORALITY

The material histories of artifacts, features, and surfaces are one last source of inferential power. Small, battered artifacts suggest they have spent some time in the plow zone (Wilkinson 1982); pristine sherds may reflect recent looting activity. Features with little modification vs. those enhanced through repeated building events vs. those showing deliberate destruction suggest different longerterm roles of those features in the cultural landscape (Chapman 1994; 1999). Finally, the work of Van Andel and colleagues (Van Andel and Runnels 1987; Van Andel et al. 1997; Van Andel et al. 1995; Van Andel and Zangger 1990; Van Andel et al. 1990) has demonstrated how surface and landscape history can inform on the configuration of the human-natural landscape. Point indicators, span indicators, and material histories, of course, are not unproblematic inferential tools. The middle-range research required for them to perform meaningfully is not to be denied and remains on the horizon.

Conclusion Archaeological landscape studies in the Mediterranean are at a critical crossroads. Because of the intensity and great technical sophistication with which archaeological survey has been done here, we increasingly appreciate that we deal with a formationally complex landscape that potentially informs on more than the historicist time dimension emphasized by most settlement pattern surveys. How are these other interesting temporalities to be approached? Here I have identified two methodological needs: one is a more complete library of potential short-, medium-, and long-term processes at work in creating the archaeological landscape; the second consists of inferential tools based in the material histories of artifacts, assemblages, features, and surfaces. With these various conceptual and methodological tools in place, the potential of the rich archaeological landscapes of the Mediterranean to better complement the equally rich textual record is boundless.

Acknowledgments I thank my Rough Cilician compatriots—Nicholas Rauh, Matt Dillon, Michael Hoff, and Rhys Townsend—for their insights. Thanks also to Effie Athanassopoulos and other symposium participants for stimulating discussions on Mediterranean archaeological landscapes and to Simon Holdaway for introducing me to the term “time perspectivism.” Errors in thought remain my own. This chapter is one in a series of reports that explore aspects of time perspectivism in archaeological deposits from the Mediterranean and elsewhere. A followon paper to this chapter was presented in the Side-by-Side Conference, hosted by Susan Alcock and John Cherry at the University of Michigan (Ann Arbor) in spring 2002 (see Wandsnider 2004).

79

This page intentionally left blank

6 Historical Archaeology of Medieval Mediterranean Landscapes Effie F. Athanassopoulos

I

n the last twenty-five years Mediterranean landscape archaeology has produced a rich surface record of Prehistoric as well as Historic times. Still, interest in the archaeology of the Medieval and Post-Medieval past remains limited. Mediterranean archaeology in general has been content to focus on the more glorious and remote time-periods, the Bronze Age, the Classical, the Roman. Many practitioners would suggest that the Medieval, the Post-Medieval, and the Early Modern pasts lie beyond the domain of archaeology. The existence of documents has relegated this segment of the past to the domain of history. Thus, understandably, landscape archaeology projects have been ambivalent about the treatment of the material record of the recent past. They have been meticulous enough to record its material remains but have been less successful in integrating these “later” periods in their grand narrative. This situation calls for a renewal of methodology and interpretative approaches, issues that the current generation of regional surveys should address. The extant rich material record of the recent past creates new opportunities but also presents new challenges. In order to utilize it properly we need to develop new tools and create a “dialogue” between the material remains and the existing textual record. These methods currently exist only in rudimentary form. Thus, this chapter will address questions that are fundamental to the study of historical materials. Some of these questions are: What is the linkage between texts and artifacts in contemporary discussions of historical periods in Mediterranean landscape archaeology? Is archaeological research supposed to fill in details and add to history? Should the archaeological record be viewed as independent, capable of yielding information that is equal to the historical? How can the two be combined? I intend to examine these questions in the context of the Medieval period. This chapter addresses issues linked to the relation between historical and scientific approaches, historical and anthropological archaeology. The rationale for selecting the Medieval period as the focus for this discussion reflects the growing need to develop context specific narratives informed by historical approaches. In the eastern Mediterranean the Medieval past is one of the historical periods that have been neglected by conventional archaeological research. However, the new

82

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

rich database produced by landscape archaeology projects has opened new opportunities and created new challenges. Before I address these questions, however, I would like to discuss the historical context and current state of archaeological research of the Medieval period in the eastern Mediterranean.

The Medieval East Byzantium, as the Eastern Roman empire is called, is one of the least known societies from an archaeological viewpoint. As defined conventionally, the Byzantine Empire starts with the foundation of the city of Constantinople in AD 324 and ends when the city was captured by the Ottoman Turks in AD 1453. Studies of material aspects of Byzantium revolve around art, religious architecture, and iconography. Little is known about everyday life from an archaeological viewpoint. The archaeological study of Byzantium lags considerably behind that of Medieval Western Europe, which in the last few decades has developed into a major branch of Historical archaeology (e.g., Gerrard 2003). Part of the reason for such a state of affairs is the lack of written sources dealing with aspects of everyday life. Records of the central government, provincial administration, land owners, merchants, etc., have not survived. Such texts would be most useful to archaeologists since they “speak” the language of material remains. In addition, the historical development of academic disciplines can explain the lack of interest in the material aspects of the Medieval period in Greece and the eastern Mediterranean in general. Byzantine studies are a by-product of Greek and Roman studies and have not been held in high esteem by philologists or classically trained archaeologists. Actually, Byzantium was dismissed by several intellectuals of the Enlightenment as “a tragic epilogue to the glory of Rome,” “a worthless collection of oracles and miracles,” “a tissue of rebellions, insurrections and treachery” and was characterized as “a thousand year history of decline” (Mango 1984; Runciman 1977). The view prevalent at the time was that Byzantium had no development on its own, that it remained fossilized and unchanging in a world of growth and development, which culminated in the Renaissance and finally in the Age of Reason. Such attitudes are rejected today. Nevertheless, they are of historical importance because they affected the growth of the field of Byzantine studies, including archaeology. In contrast, interest was centered on the remains of Classical antiquity. This preference resulted in the loss of a significant part of the Medieval layers in excavations of sites carried out in the late nineteenth and the earlier part of the twentieth century. The “uninteresting” Post-Classical layers were routinely removed, in the majority of cases without even being recorded, a common practice that only changed in the late 1970s and 1980s. Such practices are also related to the role that archaeology came to play in the process of Greek nation-building in the nineteenth century. The young discipline of archaeology was concerned with questions of cultural continuity and its task was to establish the link between the “glorious”

HISTORICAL ARCHAEOLOGY OF MEDIEVAL MEDITERRANEAN LANDSCAPES

Classical past and the reborn Hellenic state. This was a vision shared by native and foreign scholars alike who accepted ancient Greece as the birthplace of the European spirit and Western civilization. Byzantium was not part of that vision (Athanassopoulos 2002; Herzfeld 1982; Kardulias 1994b). This bleak picture, however, has been rapidly changing. In the last twenty years the study of Byzantine material culture, and ceramics in particular, has attracted significant interest (e.g., Dark 2001; Deroche and Spieser 1989; Lock and Sanders 1996; Papanikola-Bakirtzis 1996; 1999; Sanders 2000; 2001; 2003; Vroom 2003). An additional factor contributing to an increasing interest in the remnants of the Medieval past is the tremendous growth of regional surveys in Greece, Italy, and other regions of the Mediterranean. Regional archaeological surveys have provided data on rural settlement, land use, agrarian systems, subjects that the available written sources do not cover or cover poorly. Diachronic coverage is one of the great strengths of surveys, providing data for Prehistoric and Historic periods, including the Medieval as well as the Early Modern and Modern periods. Because of their diachronic focus, regional surveys mark a significant break with the tradition of selective study of the past, prevalent in countries like Greece.

Text-Aided Archaeology The recent rise of regional field surveys underscores the importance of the relation between texts and archaeology. This is an issue that several branches of archaeology have confronted, including Historical archaeology and Medieval archaeology. How to handle documents and material objects has concerned the discipline of Historical archaeology since its inception (Schuyler 1978; 1988). In the early phase of development of “text-aided” archaeologies, the textual record served as the frame and the principal source against which archaeology played a secondary role. This approach assumed that the historical record was somehow a more reliable source than the archaeological. It promoted the view of “archaeology as the handmaiden of history.” Today, there is skepticism about viewing archaeology as subservient to written sources. In contemporary Medieval archaeology a pervasive theme is the relation between artifact and text. There is a revolt against the sole dependence on texts. Champion (1990) has suggested a term that has become quite popular by now, “the tyranny of the historical record.” It refers to the fact that the archaeology of the Historic period in Europe has been set by history and the historic vision of the past. Not only have materially based studies of archaeology been subordinated to those of the literary record, but the entire conceptual framework of questions and evidence has been limited by historical concerns ultimately rooted in a vision of European uniqueness and implied superiority (Champion 1990:91). Texts and artifacts tend to report on different kinds of past processes, the general nature of archaeological evidence being that it seldom speaks the language

83

84

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

of historical events. Texts are most often concerned with short-term events (battles, treaties, expeditions). In general they provide information about the elite and tend to focus on political and social aspects of their behavior. Archaeology, on the other hand, tends to inform us about the “textually invisible” and disenfranchised members of a society. The archaeological evidence records longterm processes and mundane behavior related to daily aspects of life such as farming practices and habitation (Kosso 1995:182–88; Small 1995). Archaeologists also emphasize the biased nature of texts. They often view texts as elite products and distorted ideology and they point to the restricted nature of effective literacy in most periods of the historical past. Some scholars argue that most textual data are subjective reflections of the aspirations and worldview of the elite. Texts are material culture and, like artifacts, they must be interpreted in relation to the social context of the time when they were produced. Actually, texts are artifacts used in the differentiation and constitution of self, class, and society and are implicated in the negotiation and manipulation of social relations (Moreland 1991:23; 2001). Furthermore, texts and artifacts are different forms of expression. They may be approached as interdependent and complementary or as independent and contradictory. Oddly enough both of these views are viable (Paynter 2000:14). Several Historical archaeologists have explored these issues in innovative ways (Beaudry 1988; Deagan 1988; DeCunzo 1995; Galloway 1991; Leone 1988; Little 1992; 1994; Schuyler 1978; 1988). In Historical archaeology these are matters that have been debated intensely. Ultimately, the answer infringes on questions of both epistemology and the study of meaning, the limits of logical positivism, and the possibility (or impossibility) of an emic perspective (Paynter 2000:14–16).

Landscape Archaeology in Context Landscape is a concept that cuts across many disciplinary boundaries. It has been used widely in several fields, including geography, anthropology, and archaeology (Ashmore and Knapp 1999; Bender 1993; Cosgrove 1984; Cosgrove and Daniels 1988; Hirsch and O’Hanlon 1995; Tilley 1994; Tuan 1977; Ucko and Layton 1999; Yamin and Metheny 1996). In archaeology the approaches to landscape have changed drastically over time, from a predominance of economic and ecological perspectives in the 1960s to more recent “post-processual” (postModern) ones that focus on social and symbolic aspects (e.g., Ashmore and Knapp 1999). In general, landscape archaeology as practiced in the Mediterranean has placed emphasis on economic and social aspects, resource exploitation, land use, and rural settlements. Issues of ideology, meaning, and social power have not had a significant impact so far (but see Given, this volume). The methodology of Mediterranean surveys is based on British landscape archaeology and American settlement archaeology developed in the 1950s and earlier (Dyson 1982; Fotiadis 1995; Parsons 1972). Refinement of field method-

HISTORICAL ARCHAEOLOGY OF MEDIEVAL MEDITERRANEAN LANDSCAPES

ology continued from the 1960s on under the influence of the “New” or processual archaeology. The principle aim of processual archaeology was to transform archaeology into a science, through the use of cross-cultural comparisons, generalizing models, and the study of past human societies within an ecological framework. One of the shortcomings of this approach and its application to societies with an extensive written record is its inadequate treatment of the relation between documentary and archaeological data. Frequently, in the polemic literature produced in the early years of processual archaeology, history is presented as the chronicling of unique and particularistic events. Processual archaeologists “used an anachronistic conception of history as a straw man against which to pit their new science” (Moreland 1992:113). Because of the epistemological framework they have adopted, processual archaeologists have not been able to incorporate history in their reconstructions of daily life and attempts to reach the ordinary people of the past. The emphases on system and process and the creation of general laws were ahistorical constructs that minimized the impact of purposeful human thought and action and neglected the variety of past human experience. Change was seen as adaptive response rather than as the product of human creativity (Moreland 1991; 2001:24). Several scholars have taken these issues further. Trigger (1984), Patterson (1990), and Kohl (1998) embed archaeological theory within the context of Western culture. Trigger distinguishes between nationalist, colonialist, and imperialist archaeologies. American archaeology began as a colonialist endeavor but with the advent of processual archaeology took on the characteristics of an imperialist archaeology. Of importance here are the views of processual archaeology toward different regional and national traditions, which were typically approached as trivial knowledge. This move to trivialize national/regional traditions, the domain of culture history, which was under attack at the time, seems to be “the ideological device of elevating the interest of a segment of world society to the status of a universal as a means to hide the particularity of that segment’s point of view” (Paynter 2000:21). Thus, from their inception Mediterranean regional surveys embraced a framework that undermined the importance of historical context and the contribution of the textual record in favor of broad generalizing themes.

The Annales Perspective and Archaeology So far, I have discussed the reasons why archaeological survey data is difficult to correlate and integrate with textual data. This is more so the case in the Mediterranean where Classical archaeology as a discipline grew as a subordinate to philology and history (Dyson 1993; Morris 1994). Because of the difficulties in reconciling textual and archaeological evidence there has been increasing attention to the approach of the French Annales School, which recognizes that processes responsible for the present occur at many different tempos. The Annales

85

86

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

School, or French Structural History, has developed a multi-disciplinary approach, which combines history with geography, sociology, anthropology, and psychology, along with an emphasis on quantitative studies. One of the main elements that make this paradigm attractive to archaeology is Braudel’s theme of the interrelation of three scales of time: the long term (longue durée), the medium term (conjoncture), and the short term (l’histoire événementielle) (Braudel 1972). These three different scales or rhythms of time operate contemporaneously. Change results from the interaction, the conflicts and contradictions between the structures associated with these three time scales (Iannone 2002:72). In archaeology the Annales approach has been applied to different cultural sequences (e.g., Bintliff 1991b; Knapp 1992b). More recent studies explicitly based on the Annales have explored the long- and medium-term cycles exhibited by the Maya state (Iannone 2002) and the archaeology of the event in the context of Colonial Australia (Staniforth 2003). Barker has employed the Annales approach in the Biferno valley project, in Italy (1995b). Similarly, Bintliff (1991a) has utilized survey data from Boiotia, Greece, to explore the efficacy of this perspective in reconstructing the long-term history of the region. However, his example has not led to the application of this approach to other regional archaeological sequences in Greece. For archaeology, the most important contribution of the Annales school is its de-emphasis of the historical event as the focus of research and the emphasis on longer-term historical processes (Dyson 1993:201). The Annales paradigm has the potential to provide useful means of linking regional survey data to texts. Survey information would fall on the long-term scale, since it consists of economic data in their environmental context; it corresponds to the slow process of change usually seen in the countryside. The medium term corresponds to cycles of expansion and contraction operating in the landscape at different periodicities (400–5,000 years have been suggested). Textual material, on the other hand, would correspond to the short term (événementielle), and supply a political narrative linked to short-term events (Small 1995). Fundamental ideas associated with the Annales paradigm are a stress on generalization and a search for wider insights beyond the unique and the particular. These are goals that have been emphasized also by processual archaeology and are compatible with the goals of archaeological landscape projects. Furthermore, because the Annales approach incorporates a combination of viewpoints, it seems to promise success where processual perspectives have failed. It does put emphasis on social and economic history but also leaves room for the role of ideology and perceptions of the world in past societies (mentalités). In general this framework has inexhaustible potential but several problems remain to be solved. For example, it is not clear how to articulate the long-term aspects of history with cycles of rapid change or how important is the role of random factors. The Braudelian model is innovative because it helps to conceptualize the relationship of structure and event. However, critics of this approach

HISTORICAL ARCHAEOLOGY OF MEDIEVAL MEDITERRANEAN LANDSCAPES

argue that it tends to emphasize the structural scaffolding of history and to downplay the role of contingent events in the emergence of social-spatial structure. The Annaliste narrative is constructed on the primacy of geo-historical processes over the role of human agency. It concentrates “on the anonymous masses instead of conspicuous individuals and on continuities and regularities instead of rapid changes” (Horden and Purcell 2000:39). Moreover, it is not explicit just how the different temporalities intersect in a dynamic sense and the primacy accorded to the longue durée presents us with an essentially linear view of history (McGlade 1999b:146–47). In spite of these criticisms, it is the emphasis on the long-term processes that makes the Annaliste approach so attractive to archaeology. Regional archaeological projects have concentrated their efforts on refining the methodology and maximizing the extraction of information from surface artifacts and assemblages. However, they are still struggling with issues of interpretation. Frequently, in the publication of the results of regional projects the objective is to organize and present the massive amounts of data that these multi-year field efforts generate. Little effort goes into integrating the varied databases and data sources. This is particularly the case with the “later” material (Post-Roman). At present, the lack of integration of the different data sources reminds me of a similar situation in the field of Historical archaeology. Robert Schuyler suggested in 1988 that Historical archaeology is stuck in “Phase I” and has a hard time moving into “Phase II” (borrowing terminology common in the CRM environment). The symptoms of this situation are easy to recognize. Most archaeological reports assume a final format that divides into an archaeology section, a history section (not always present), and a section on methodology. To quote Schuyler (1988:40): It is as though a social historian, such as Borchert, divided his final monograph into: Part 1, historical conclusions based on documents written on white paper, Part 2, historical conclusions based on documents written on yellow paper; general conclusions, none.

Indeed, several examples of final reports of survey projects follow a similar structure. They continue proudly the well-established tradition in Classical archaeology of detailed description and copious presentation of data. They emulate the large excavation projects in their operation, and like those “they encourage specialization and fragmentation…cataloguing and ordering of decontextualized material” (Dyson 1993:195). Most energy is expended on these aspects and little room is left for integration and interpretation. Davis has expressed it very well: “Mediterranean survey, like Classical archaeology in general, runs the risk of becoming an essentially atheoretical enterprise of parochial interest” (Davis 1996). Although this situation affects the presentation of most historical periods discussed in the final reports of regional survey projects, it is particularly acute when it comes to the Medieval and Post-Medieval periods.

87

88

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Case Studies In a chapter of this length it is possible to discuss only a few case studies. In the remaining section I will review studies that explore different ways of integrating survey data and texts in the context of the Medieval and Post-Medieval periods. Finally, I will attempt to apply the Annales perspective to the region of Nemea, in the northeastern Peloponnesos, Greece, which has been the focus of an intensive regional survey project, the Nemea Valley Archaeological Project (hereafter NVAP, see Wright et al. 1990). Moreland (1991; 1992) analyzes Italian hilltop settlements of the ninth century that preceded incastellamento (the growth of fortified sites on hilltops). These hilltop sites should be viewed in the context of agrarian re-organization and the tightening of social relations. To understand the significance of these sites it is important to look at the estates of large monasteries like Farfa and San Vincenzo. Documents in these archives record the donation of lands and provide details of the boundaries of such properties. Hilltop sites seem to have been involved in the close administration of the monastic property. They should be seen as focal points for the increased control of local populations. They had a prominent role in the collection, collation, and recording of information. Documents had a similar role: they point to an increased control over social relations; along with the hilltop sites, they were the means by which that control was exercised. Some of the texts are explicit about the relationship between the creation of a castello and the control of people; they make it clear that the purpose of the foundation was the consolidation of lands and the aggregation of people. Thus, at the level of the peasant producer, the relocation in the hilltop settlement was a physical manifestation of the relations of power exercised by the elite. Moreland attempts to show that the landscape is full of meaning, which had an important role in the creation and reproduction of a network of relations of power. He emphasizes ideological aspects, worldview, best expressed by the Annaliste term mentalités. Mentalités derive from social and economic contexts and are tied to notions of equilibrium and maintenance of the structure. Furthermore, mentalités could be viewed as an active force in promoting or stifling change (Iannone 2002:73). In Greece, a large number of intensive regional surveys have produced evidence of settlement and land use in the Medieval and Post-Medieval periods (e.g., Boiotia [Bintliff and Snodgrass 1985; Vroom 2003], Southern Argolid [Jameson et al. 1994], Keos [Cherry et al. 1991], Nemea [Wright et al. 1990], Berbati-Limnes area [Wells et al. 1990; Wells and Runnels 1996], Methana [Mee and Forbes 1997], Laconia [Cavanagh et al. 1996; 2002], Pylos [Davis et al. 1997; Davis 1998]). These surveys have identified periods when the density and pattern of settlement changed considerably. These shifts are usually interpreted as representing cycles of economic expansion and contraction and population fluctuations. Patterns of dispersed settlement have been often associated with dense

HISTORICAL ARCHAEOLOGY OF MEDIEVAL MEDITERRANEAN LANDSCAPES

population, access to commercial markets, and intensive agricultural activity (e.g., Jameson, Runnels, and Van Andel 1994). On the other hand, economic contraction is associated with nucleated settlements located close to the arable land to allow the population to minimize movement to and from the fields, while practicing subsistence agriculture (Runnels and Van Andel 1987:327). In the case of the island of Melos nucleated phases of settlement seem to correspond to periods when the island was closely linked to external polities that exerted direct economic and political control over the island. During these times the population was concentrated in the nucleated settlements. To bring the population together in one or a few settlements serves to articulate the economy of the island to the controlling outside polity and may be described as a “colonial model” (Wagstaff and Cherry 1982a; 1982b). Furthermore, economic processes have been associated with geomorphological evidence, specifically localized soil erosion and alluviation episodes (Van Andel and Runnels 1987; Van Andel et al. 1990). The number and size of rural sites has been interpreted as evidence of different agricultural strategies and patterns of land tenure. A dispersed pattern has been associated with intensive subsistence agriculture and small holdings while a nucleated pattern (also called traditional) points to an extensive mode of agriculture and possibly to periods when land has been in the hands of an elite minority (Alcock 1993; Cherry et al. 1991; Davis 1991; Halstead 1987). Overall, the models of interpretation advanced by regional surveys in Greece have been predominantly economic. The research design is a uniform one that encompasses both Prehistoric and Historical periods. In most cases, the inclusion of written records in the interpretation of the material is an afterthought. Fortunately, in recent years more effort is being devoted to the integration of archaeological and textual data and the development of context-specific interpretations. Projects that have been successful in opening new paths for the study of “later” periods, notably the Post-Medieval to Early Modern times, are the Boiotia regional survey (Bintliff 1996) and the Pylos Regional Archaeological Project (Davis 1998; Bennet et al. 2000). These regional surveys have invested in the analysis of unpublished documents from the Ottoman archives. The relevant documents, primarily Ottoman tax registers, are extremely detailed. Kiel (1999) has been able to extract a great deal of information of interest to survey archaeology, including a complete breakdown of villages and towns in the region of Boiotia, crop production by variety and yields, industrial production, and detailed demographic data. On the basis of this new evidence, the Ottoman conquest of Boiotia does not seem to have disrupted settlement, in spite of the fact that many historians view it as a traumatic divide (Bintliff 1996:6–7). The growth of the region in the late fifteenth and throughout the sixteenth century recorded in the Ottoman registers can be tied in with the flourishing settlement picture documented by the Boiotia archaeological survey. This “dialogue” between the two data sources has helped archaeologists to identify the pottery of this period (Vroom 1998; 2003). Similar work has been undertaken for the

89

90

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

region of Pylos. It combines the study of unpublished historical documents of the Ottoman period (early eighteenth century) with the results of archaeological fieldwork (Zarinebaf et al. 2004).

Medieval Patterns in Nemea In my own analysis of the Medieval material remains in the Nemea region (Athanassopoulos 1993; 1997) I encountered several difficulties, one of those being the lack of written records at the local level. This is a problem for most of Greece, which in the Middle Ages belonged to the extremely centralized Byzantine state centered on Constantinople, whose detailed official and tax records have vanished (Mango 1980). Thus, the archaeological evidence provides the primary means by which we can develop an understanding of the Medieval countryside in the regions that formed the “Eastern Roman empire.” An additional problem is a coarse ceramic chronology in the region due to a clear disinterest among previous generations of archaeologists to record and date Medieval ceramics, which were found in excavations of Classical sites and were routinely thrown away.1 Such practices prevailed in the past but, fortunately, have changed in the last two decades.2 Ceramics are the most common find of surveys. Since they come from the surface and not from a stratified context, our present ability to date surface pottery is dependent on the existence of well-dated comparative collections from excavations. Although our knowledge of local pottery is rapidly improving, still, in most cases we have to work with ceramic sequences spanning one to two centuries or more, which clearly do not provide sufficient chronological control. Similar problems are encountered by surveys in many other regions of the Mediterranean, notably in Italy (Patterson 2000). Furthermore, fluctuation in the circulation of glazed pottery is an important factor that has been addressed only recently. Current work on material from Corinth indicates that there is a dramatic change in the appearance of Medieval ceramics toward the end of the eleventh century when the quantity of glazed pottery increases and its appearance undergoes a complete metamorphosis (Sanders 2000). Sanders attributes this change to the growth of urban centers during that time and believes it represents an increase in the disposable income of the city’s lower classes, who could now afford the more costly glazed tablewares. This is precisely the pottery that is highly diagnostic and has been used by archaeological surveys for chronological identification. These patterns may have important implications for the interpretation of survey data. Sanders ponders whether the increase in Middle Byzantine sites identified by regional surveys is related to the greater quantity and visibility of glazed pottery rather than to a growth in rural population, as is often assumed (Sanders 2000:172–73). In order to develop a Braudelian framework in the context of landscape archaeology of the Medieval period in Greece, it is important to improve our understanding of ceramic production and circulation of diagnostic wares in

HISTORICAL ARCHAEOLOGY OF MEDIEVAL MEDITERRANEAN LANDSCAPES

Figure 6.1. Map of the NVAP survey area with identified Byzantine/ Medieval sites (Rosemary J. Robertson).

different regional contexts (e.g., Vroom 2003). At this point, we can only suggest programmatic goals. At the level of Braudel’s conjonctures, medium-term trends, there are a number of social and economic cycles recorded in the archaeological and the historical records. In the Nemea region the archaeological evidence points to a proliferation of sites dating to the twelfth to thirteenth centuries AD (Figs. 6.1–6.3). The majority of these sites are located near the arable land, on the lower slopes of the hills surrounding the Nemea valley, and in smaller valleys in the southern part of the survey area. There are two large sites, sites 600 and 704, which consist of dense scatters of archaeological material spread over several fields. Site 600 (Fig. 6.2), located near the Sanctuary of Zeus and the Classical Stadium, covers an area of approxi-

91

92

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 6.2. NVAP sites with Byzantine material in the southern part of the survey area (Rosemary J. Robertson).

mately 34 ha. Site 704, to the south, has an estimated size of 47 ha. The Classical city of Phlius (Fig. 6.3) has a substantial Medieval component concentrated in the SW corner of the site (Alcock 1991: Fig. 19). Smaller sites vary in size; they usually cover less than a hectare (e.g., site 602, 0.7 ha [Fig. 6.3]; site 510, 0.6 ha [Fig. 6.2]). These remnants of Medieval settlements must reflect the intense level of agricultural activity in the Nemea area at the time. This interpretation is also supported by the documentary evidence that exists for other regions, in particular the cadaster of Thebes that deals with central Greece. It reveals a complex pattern of land ownership with scattered, small land holdings and dense settlement (Harvey 1989:63–64; Hendy 1989:12; Lemerle 1979:198–99). In general, the period of the eleventh and twelfth centuries is characterized by the growth of landed estates in most Byzantine territories, a trend that transformed most peasants into dependent farmers (Hendy 1989:18, 71). In some regions of Byzantium entire villages and estates became the property of powerful land owners. But this does not seem to be a pattern that encompassed all territories. For example, in the region covered by the Theban tax-register, in central Greece, the accumulation of property occurred in a more piecemeal way, the result being a less consolidated pattern of land ownership (Harvey 1989:74, 78). Furthermore, textual evidence reporting on social conditions of late twelfth century Athens presents a picture of hardship and poverty for the lower classes. Michael Choniates, the Metropolitan of Athens, acted to lighten the tax burden of the Athenians who had been adversely affected by the imperial policies that favored

HISTORICAL ARCHAEOLOGY OF MEDIEVAL MEDITERRANEAN LANDSCAPES

Figure 6.3. NVAP sites with Byzantine material in the northern part of the survey area (Rosemary J. Robertson).

the growth of large estates at the expense of peasant holdings (Kolovou 1999; Angold 1995:205). The evidence from Nemea points clearly to a dispersed pattern of settlement usually associated with small holdings and intensive agricultural practices. However, in the absence of textual records, it is not possible to speculate about the status of the peasants or the scale of land ownership. Still, the significant amount of glazed pottery recovered from fields in the Nemea area provides direct evidence of the participation of the Medieval inhabitants in extensive trade networks. It indicates growth in the disposable income of the peasant households, who were able to purchase the more expensive glazed tablewares. Moreover, it implies economic growth in the countryside and increase in trade. Economic expansion in Nemea might be associated with commercial activities in the major urban center of the region, Corinth. Commercial activity in Corinth from the late eleventh century on was greatly stimulated by the Venetians (Harvey 1989:216). Venetian presence in Corinth is attested by documents and increasingly so by the archaeological evidence from the on-going excavations. In turn, increased availability of trade opportunities may have contributed to the opening up of new fields and an increase in the area of land under cultivation in rural areas like Nemea. The Byzantine pottery present in the NVAP survey area includes a wide variety of well-known styles of decoration (sgraffito, slip painted, glaze painted and/or a combination of these main decorative types; Table 6.1). Most of these

93

Table 6.1. Styles of Byzantine decorated pottery from NVAP sites. The numbers represent catalogued sherds in each category.

NVAP Site

Fine Sgraffito

Incised Sgraffito

1 3 7 203 300 303 501 509 510 513 514 515 600 601 602 603 605 704 801 804 901 902 910 925 927 Phlius

2 2 5 2 8 2 1 -

3 1 3 2 2 1 2 6 3 4 2 8 4 1 1 2

Painted Incised Sgraffito 1 1 1 1 1 1 1 2 1 -

Painted Sgraffito Sgraffito 1 1

1 1 4 1 2 2 5 7 3

Measles Green & Brown Painted 1 2 1 1 1 3 14 1 1 1 1 5 2 1 3 -

Slip Painted 1 6 3 5 4 1 1 1 -

Glaze Zeuxippus Matt Painted Painted 1 4 2 1 1 -

1 1 1

1 1 1 1 1 8 3 1 1 3 1 3 2 1 5

Plain Glaze

Italian Imports

NVAP Site

4 1 2 5 2 1 2 3 1 3 1 3 4 8 1 1 1 13

1 5 5 1 4 3

1 3 7 203 300 303 501 509 510 513 514 515 600 601 602 603 605 704 801 804 901 902 910 925 927 Phlius

HISTORICAL ARCHAEOLOGY OF MEDIEVAL MEDITERRANEAN LANDSCAPES

styles were common in the twelfth and the thirteenth centuries, based on stratified evidence from the Corinth excavations. Sanders has revised the chronology of glazed wares and documented the fluctuations of the different types present in the Corinthian deposits from the late eleventh to early fourteenth centuries (Sanders 2000:160–61; 2003). As mentioned earlier, toward the end of the eleventh century the volume of glazed pottery increases dramatically in Corinth. In the twelfth century there is a wide variety of new styles produced by the Corinthian pottery industry which have short life spans, about a generation or so. The Frankish conquest of the city in the early thirteenth century does not appear to have affected local pottery production. However, there is a radical transition toward the end of the thirteenth century, when the Corinthian market is initially contested and eventually dominated by Italian imports. The NVAP evidence indicates that there is a dramatic increase in the circulation of glazed pottery in the countryside in the twelfth to thirteenth century. However, unlike the Corinthian trend of the late thirteenth century, Italian imports have a limited distribution in the NVAP survey area. The few finds primarily occur at sites associated with Frankish presence/control. Here, it is important to note that Byzantine/Medieval pottery, besides being a dating tool, can also provide a wealth of information about changing norms regarding food consumption and social display (Blake 1978; Vroom 1998; 2003). Furthermore, the symbolism present in the popular decorative themes of Byzantine pottery is a rich unexplored area that offers access to the mentalités of the population (e.g., for birds and hunting scenes see Von Wartburg 2001). In addition, appearance/disappearance of specific forms of pottery document changing eating habits and table manners. One characteristic form, the chafing dish, disappears from the Corinthian deposits in the late eleventh century when new forms (bowls, plates, cups) replace it. The chafing dish is usually associated with communal meals, whereas bowls and plates indicate a shift to individualized place settings (Sanders 2000:171–72; for changing dining habits in the Byzantine period see Vroom 2003:313–33). In the Nemea area, the next cycle, roughly from the late thirteenth to fifteenth centuries, provides a contrast to the patterns documented for the twelfth to late thirteenth centuries. After the Latin conquest of AD 1204 the political and social structure of Byzantine Greece became extremely fragmented and decentralized; prolonged conflict, civil wars, invasions, and the plague seem to have taken a heavy toll on the rural population. There are numerous narrative and documentary sources that point to demographic decline and impoverishment in the countryside (Longnon and Topping 1969; Panagiotopoulos 1985). Archaeological evidence of settlement and agricultural activity in Nemea becomes scarce. A complex of sites, consisting of a fort/tower along with remains of an extensive village on the top of a precipitous hill, dominates the region (Fig. 6.2, sites 901, 902, 910; Fig. 6.4). The steep hill, identified in historical sources as Polyphengi (Kordosis 1981), controls the passes to the south and the west, routes of signifi-

95

96

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 6.4. View of the northern side of Mt. Polyphengi and the town of Nea Nemea. Photograph by the author.

cance in the Medieval period. Diagnostic pottery from these sites includes Italian imports (protomajolica, Italian painted sgraffito, majolica) ranging in date from the late thirteenth to fifteenth centuries. Thus, at present, the temporal scale of the Braudelian conjoncture lies at the edge of current archaeological capabilities. The existing chronology is sufficient to document contrasting patterns and the presence of change but it is not refined so that we can analyze the actual processes or events of change in great detail. When it comes to shorter time scales, those that span a generation or so, we have to rely on information that comes strictly from textual sources (Bon 1969; Kordosis 1981; Longnon and Topping 1969; Panagiotopoulos 1985). For example, there is a wealth of documentary material from the end of the seventeenth to early eighteenth century (1685–1715) when the Venetians gained control of the whole Peloponnesos (Davies 1994; Frangakis and Wagstaff 1987; Panagiotopoulos 1985; Sauerwein 1969; Topping 1972; 1974; 1976; 2000; Wagstaff 1978). These sources provide data on demography, settlement, agricultural practices, and general conditions in the countryside. Future chronological refinement of archaeological material may add another dimension here but currently it is not possible to combine or contrast patterns based on the documentary and archaeological evidence of the Post-Medieval period in Nemea.

HISTORICAL ARCHAEOLOGY OF MEDIEVAL MEDITERRANEAN LANDSCAPES

Epilogue I would like to close by voicing concerns similar to those expressed by archaeologists working in other cultural areas (e.g., Smith 1992). The Annales perspective is not a panacea that will provide instant illumination of the historical past, but it ties in well with current concerns in Medieval and Historical archaeology in general. More than anything else, it forces us to confront issues related to the effective utilization and correlation of the archaeological and historical records. Furthermore, the significance of varying temporal scales should not only enter the picture at the level of interpretation but also at the level of research design and methodology (Smith 1992:31). So far, the research design of regional surveys has not been sensitive to the particular historical contexts under investigation. The next generation of surveys should include careful consideration of chronological issues in both their theoretical (Braudelian temporal rhythms) and practical dimensions. These are matters that landscape archaeology projects need to address in order to justify more effort invested in fieldwork and the accumulation of regional archaeological data. They represent the new challenges that have to be met so that regional surveys can fully contribute to our understanding of the evolution of the Mediterranean landscape.

Acknowledgments I would like to thank the American Council of Learned Societies for a Fellowship during the academic year 2000–2001 that enabled me to undertake research for this chapter. I would also like to thank the co-directors of the Nemea Valley Archaeological Project, John Cherry and Jack Davis, for use of data from the survey. In particular, I would like to thank Jack Davis for his support and for making the NVAP databases available in electronic format. Thanks also to LuAnn Wandsnider and to many other symposium participants for their ideas, comments, and suggestions. Figures 6.1, 6.2, and 6.3 should not be reproduced without permission of Rosemary J. Robertson.

Notes 1. Charles Morgan, one of the pioneers, describes the situation eloquently: “Plates and bowls, fragmentary and whole, were consigned to the dump, their bright green or yellow glazes signaling them for oblivion as Medieval rubbish. But despite this almost universal contempt for things Byzantine, a few pieces found inconspicuous shelter in the darker corners of the storehouses on the Acropolis, and some even attracted, on very rare occasions, casual notice in publications” (Morgan 1942:1).

97

98

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

2. The excavations of Corinth have produced a rich record of material culture of the Byzantine and Frankish periods (Williams and Zervos 1988; 1990; 1991; 1992; 1993; 1994; 1995; 1996; Williams, Barnes, and Snyder 1997; 1998; also several studies in Williams and Bookidis 2003).

7 Historical Contingency, Nonlinearity, and the Neolithization of the Western Mediterranean C. Michael Barton, Joan Bernabeu, J. Emili Aura, Oreto Garcia, Lluis Molina, and Steven Schmich

O

ver the past decade, a new synergistic perspective on the relationships between humans and the natural world has emerged in a variety of natural and social science disciplines (Balée 1998; Butzer 1996; Crumley 1994; Pyne 1997; Sheridan 1995; Zimmerer 1994). An important characteristic of this perspective is the recognition that humans and cultural systems have played an integral role in the development and maintenance of ecosystems world-wide. At the same time, humans are still subject to diverse ecological constraints, even in the context of complex society. This means that human society is constantly reshaping the intertwined cultural and natural components of the socioecological landscape on which its members and their descendents must operate. This broader socioecological dynamic—one that includes culture as well as nature (see Barton et al. 2001; Winterhalder and Smith 1981)—operates over varying, but often long time periods as was recognized decades ago by Braudel (1980) in the humanities and Butzer (1982) in the natural sciences. Very often the environmental consequences of human action only are apparent over the course of many generations. Historical disciplines, and especially archaeology with its overt focus on long-term cultural change and human-environmental interaction, are essential for understanding the temporal dynamics of socioecosystems (McGlade 1995; McGlade 1999a). Socioecological processes also operate at varying spatial scales. As human interaction and social organization vary, both across space and at different spatial scales (e.g., household, community, state), so too do the ways in which humans interact with the non-human aspects of the environment. Hence, the state of a socioecosystem at any particular place is equally a product of spatialdependent as well as time-dependent processes. While human societies operate and change within the framework of general evolutionary processes, we cannot expect a cultural system to follow any particular linear developmental trajectory. Rather, the outcomes social evolution will be strongly contingent on socioecological history.

100

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Figure 7.1. Research area in eastern Spain and valleys discussed in text.

Over the past decade, we have been working to better understand the longterm temporal and spatial dynamics of socioecosystems in eastern Spain. Our research actively integrates archaeological survey and excavation to build regional data sets for modeling Prehistoric landscapes and their changes. Because we are interested in the dynamics of human-environmental interaction, this has involved developing a suite of methodological protocols that are relevant to many of the issues under consideration in this volume, including appropriate units for data collection and interpretation, replicability and comparability of data recording and analysis, and theoretical approaches to the archaeological record at regional scales. In this chapter, we review the protocols we have developed for the survey portions of this research program and discuss some the results that are coming out of our study.

Research Setting The Mediterranean Basin possesses characteristics that make it uniquely suited for studies of long-term socioecosystem dynamics. It has a coherent ecological definition, being comprised of a series of interrelated environmental settings around the Mediterranean seaboard and its hinterlands, yet shows a great deal of ecological diversity, often over small distances. The Mediterranean Basin also

THE NEOLITHIZATION OF THE WESTERN MEDITERRANEAN

has a history of human occupation that extends back a million years or more, encompassing the human transformation from simple consumer and ecosystem participant to ecosystem modifier and eventually to ecosystem manager (Butzer 1996). This region witnessed some of the earliest well documented modern humans, the first of our genus associated with large-scale impacts on plant and animal communities; the earliest known of agro-ecosystems and their attendant effects on the biological and physical landscape; and the world’s earliest development of social complexity and urbanism. Our ongoing research project is located in the western end of the Mediterranean Basin, in Alicante Province of the País Valenciano, Spain. Within a region of over 1,800 sq km, we have conducted intensive survey in seven valleys of the rugged uplands that border the Mediterranean littoral. In this chapter, we report on our work in four of them: the Polop Alto, Penaguila, middle Serpis, and Alcalá valleys (Fig. 7.1). Elevations within this region range from 200 m to > 1500 m, and rainfall measures 600–900 mm annually, permitting dry farming throughout. Human occupation extends back into the Middle Pleistocene, and has been more or less continuous at least since the early Upper Pleistocene. The regional prehistory recently has been reviewed in detail elsewhere (Aura and Pérez-Ripoll 1995; Barton et al. 1999; Bernabeu and Juan-Cabanilles 1994; Villaverde et al. 1998), and an overview of relevant chronological frameworks is presented in Table 7.1. We focus here on the transition from Pleistocene foraging to Holocene food producing economies—marking a fundamental change in the way in which humans interacted with the natural environment.

Table 7.1. Regional chronology. PERIOD

Middle Paleolithic

DATES

>34,000 BP

Upper Paleolithic

34,000–14,000 BP

Final Paleolithic–Mesolithic

14,000–8,000 BP

Geometric Mesolithic

8,000 BP–5,600 BC

Neolithic I

5,600–4,500 BC

Neolithic IIA

4,500–3,700 BC

Neolithic IIB

3,700–2,900 BC

Neolithic IIC

2,900–2,300 BC

101

102

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Protocols for Systematic Survey Conceptual Frameworks While our ultimate interest is in the ecology of interacting human and natural systems, these systems are no longer extant and their information, energy, and material flows cannot be observed directly. However, landscapes are the geographical context in which socioecosystems operate and are altered by socioecological processes. In recent years, “landscape” has become a catchword that embodies a wide variety of concepts (see McGlade 1995; Rossignol and Wandsnider 1992; Ucko and Layton 1999; Waters and Kuehn 1996). Here, we use the term to mean the earth’s surface and surface sediments, along with its physical, biotic, and social constituents. Social and natural processes that structure and change socioecosystems very often have physical outcomes on landscapes, depositing, rearranging, or removing these constituent materials. Landscape studies can serve as a way to get at many of these no-longer-observable processes we seek to understand. In order to do this, we must record cultural and natural characteristics of modern landscapes that are only the most recent manifestation of the long-term, cumulative products of a very long history of social and natural processes and their interactions. In this sense, the things that we study—whether architecture or artifacts, soils or landforms—are profitably viewed as accumulations that have been deposited at varying rates and differentially preserved. Because of the cumulative nature of landscapes—including behavioral residues—and because of the complex intertwining of cultural and natural processes, we take a taphonomic approach to the archaeological record. That is, we focus on understanding the suite of processes that are responsible for the accumulations of artifactual and other materials that make up modern landscapes. These processes include not just initial deposition, but subsequent alteration, transport, and loss. Rather than treating all but the first of these sets of processes (i.e., discard or deposition) as noise that distorts the archaeological record, we see all taphonomic processes as potentially informative. In practice, this means that we record land use and geomorphology along with artifact and feature counts in our survey units, and include evidence for transport and other postdepositional alteration in our artifact analysis along with technological, functional, and stylistic features. It also means that we treat the archaeological record as a continuous, spatially variable, long-term accumulation rather than a set of imperfect snapshots of the past (Gillings and Sbonias 1999; Zvelebil et al. 1992).

Field Methods Our research focus on changes in subsistence economy, the spatial configuration of human land use, changes in settlement permanence, and associated changes to biotic communities and surface sediments has structured our field

THE NEOLITHIZATION OF THE WESTERN MEDITERRANEAN

and analytical protocols. Hence, our work has taken place at something of a meso-scale, intermediate between extensive, site-focused surveys and longterm excavation of particular human settlements. We want to cover large enough areas to encompass most of the normal, day-to-day activities and natural processes that characterize community-level socioecosystems. For this reason, we have chosen to invest considerable research effort in pedestrian survey and extensive subsurface testing rather than long-term excavation of a single or few sites, and to intensively survey in a series of compact valleys rather than extensively across an entire region. At this meso-scale, we endeavor to collect information about landscapes rather than sites, employing a patchbased methodology. Because we are interested in the interactions between humans and other components of their ecosystems, and seek to identify and explain spatial variation in artifact accumulations across landscapes, we need to systematically collect data where material culture is rare or absent as well as in locales where it is abundant. This means that we cannot limit our survey units to those locales commonly termed ‘sites.’ The considerable literature on “non-site” or “off-site” survey encompasses a wide variety of techniques (Dunnell 1992; Ebert, Camilli, and Berman 1996; Foley 1981b; Zvelebil et al. 1992). Our survey units are geographically defined ‘patches’ from which we systematically collect data—including but not limited to information about cultural materials—whether or not artifacts or other behavioral residues are found in the unit. Such geography-based units are called patches in fieldwork deriving conceptually from the ‘new ecology’ and seeking to assess spatial variation in ecological parameters and processes (e.g., Collins et al. 2000); we follow this usage here. We generally (though not exclusively) use the small terraced fields that are pervasive throughout the Mediterranean to delineate our study patches. In the first surveys conducted in this region (the Penaguila and parts of the Serpis valleys) contiguous areas were completely covered. Subsequently, we have employed a multi-stage strategy of stratified, random sampling followed by selected sampling to cover larger areas while maintaining the potential for statistical modeling. Within each patch we record a suite of standardized data about the modern landscape and collect all Prehistoric artifacts. Some of the data we record, such as landform and surface soil characteristics, provide information about landscape development over time. Others, such as current land use and surface visibility, help us better assess artifact accumulation processes. We make extensive use of GIS tools to organize and rescale the data we collect. By using a GIS to map patches, we can overlay aerial and space-borne imagery that provides further information about soils, sediments and geomorphology, and vegetation and land use. GIS also facilitates quantitative analysis and ecological modeling from patch-based data (see Barton et al. 1999; Barton et al. 2002; Bernabeu et al. 1999 for more detailed description of field methods).

103

104

MEDITERRANEAN ARCHAEOLOGICAL LANDSCAPES : CURRENT ISSUES

Analytical Approaches Our perspective that the archaeological record is part of a more inclusive socioecological landscape that has been differentially accumulating across space and time has shaped our approach to analysis of the data we collect. We assume a priori that artifact accumulations are palimpsests rather than cultural snapshots (see Binford 1981b). Furthermore, we recognize at the outset that these accumulations are not simply the result of cultural site formation (sensu Schiffer 1987), but are also shaped by and record the long-term dynamics of landscape formation that are ongoing today. Chronology building is a fundamental aspect of archaeological research, and distinguishing the changing nature of human use of each patch is essential to modeling the temporal dynamics of human ecosystems across space. However, this is a more complex task for surface assemblages where episodes of greater human activity (and higher artifact accumulation rates) are not separated stratigraphically by sediment that accumulated more rapidly than artifacts during intervals of minimal activity. Archaeologists commonly use chronologically sensitive artifacts to estimate the age of surface artifact. Our analytical protocols systematize this generally subjective procedure while also recognizing the often palimpsest nature of material culture accumulations. We use the combined presence/absence of a variety of artifact forms to estimate systematically the likelihood that artifacts accumulated in a study patch during any of several temporal intervals. In the case of the work presented here, the intervals are Middle Paleolithic, Upper Paleolithic, final PaleolithicMesolithic, Neolithic I, and Neolithic II (Table 7.1 and 7.2) (see Barton et al. 1999; Barton et al. 2002; Bernabeu et al. 1999 for detailed descriptions of these procedures). These probability estimates, which we call Temporal Index (TI, ranging from 0–1), allow us to model where human activities took place and the extent of the landscape used at different times in the past. We further make the reasonable assumption—based on the positive relationship between diversity and sample size (Jones et al. 1983)—that the time period(s) with the strongest chronological signal (i.e., the highest TI values) will be the one(s) during which the greatest portion of an artifact assemblage accumulated. Hence, we also weight a measure of artifact abundance by Temporal Index to estimate geographic variation in intensity of land use within each time interval (see Barton et al. 1999; Barton et al. 2002; Bernabeu et al. 1999 for detailed descriptions of these procedures). We have designated this latter value Settlement Intensity Index (SII). In essence, we have simply systematized the more subjective chronological assessments that archaeologists commonly make in survey projects. In so doing, we endeavor to make these assessments more consistently replicable while reducing inter-investigator variation in the interpretation of archaeological materials. Because they are quantitative and linked with spatially defined data collection units (i.e., calculated for each landscape patch rather than only for sites), measures

105

THE NEOLITHIZATION OF THE WESTERN MEDITERRANEAN

Table 7.2. Calculation of Temporal Index (TI) (after Barton et al. 1999). MP: Middle Paleolithic, UP: Upper Paleolithic, N1: Neolilthic I, N2: Neolithic II. PERIOD

Late Neolithic

Early Neolithic

TEMPORAL INDEX

0.9

0.7

present: N2 tools or N2 ceramics

0.5

0.1

0

present: present: present: ceramics or Neol tools, blade tech. ground stone ceramics or ground stone

present: artifacts

present: N/A

absent: N/A

absent: absent: backed tools N/A and N1 tools

absent: N/A

absent: N/A

absent: artifacts

present: (backed tools and [ceramics or N1 tools]) or N1 ceramics

present: present: Neol tools, backed tools ceramics or or ceramics ground stone

present: blade tech.

present: artifacts

present: N/A

absent: N/A

absent: N/A

absent: artifacts

absent: absent: absent: N2 tools and N2 tools and N/A N2 ceramics N2 ceramics present: Late Upper backed tools Paleolithic– Mesolithic absent: Neol tools and ceramics Upper Paleolithic

Middle Paleolithic

present: backed tools

0.3

present: present: present: backed tools backed tools lithics or blade tech. or blade tech.

absent: absent: absent: ceramic densi- ceramic densi- N/A ty