Social Complexity and Complex Systems in Archaeology [1 ed.] 0367478587, 9780367478582

Citation preview

Social Complexity and Complex Systems in Archaeology

Social Complexity and Complex Systems in Archaeology turns to complex systems thinking in search of a suitable framework to explore social complexity in Archaeology. Social complexity in archaeology is commonly related to properties of complex societies such as states, as opposed to so-called simple societies such as tribes or chiefdoms. These conceptualisations of complexity are ultimately rooted in Eurocentric perspectives with problematic implications for the field of archaeology. This book provides an in-depth conceptualisation of social complexity as the core concept in archaeological and interdisciplinary studies of the past, integrating approaches from complex systems thinking, archaeological theory, social practice theory, and sustainability and resilience science. The book covers a long-term perspective of social change and stability, tracing the full cycle of complexity trajectories, from emergence and development to collapse, regeneration and transformation of communities and societies. It offers a broad vision on social complexity as a core concept for the present and future development of archaeology. This book is intended to be a valuable resource for students and scholars in the field of archaeology and related disciplines such as history, anthropology, sociology, as well as the natural sciences studying human-environment interactions in the past. Dries Daems is Assistant Professor in Settlement Archaeology and Digital Archaeology at Middle East Technical University. He is also affiliated with the Sagalassos Project at University of Leuven. His research interests include social complexity, agent-based modelling, material studies, and human– environment interactions.

Social Complexity and Complex Systems in Archaeology

Dries Daems

First published 2021 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 52 Vanderbilt Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2021 Dries Daems The right of Dries Daems to be identified as author of this work has been asserted by him in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloguing-in-Publication Data Names: Daems, Dries, author. Title: Social complexity and complex systems in archaeology / Dries Daems. Description: London; New York, NY: Routledge/Taylor & Francis Group,   [2020] | Includes bibliographical references and index. Identifiers: LCCN 2020040820 (print) | LCCN 2020040821 (ebook) |   ISBN 9780367478582 (hardback) | ISBN 9781003036968 (ebook) Subjects: LCSH: Social archaeology. Classification: LCC CC72.4.D34 2020 (print) | LCC CC72.4 (ebook) |   DDC 930.1–dc23 LC record available at https://lccn.loc.gov/2020040820 LC ebook record available at https://lccn.loc.gov/2020040821 ISBN: 978-0-367-47858-2 (hbk) ISBN: 978-1-003-03696-8 (ebk) Typeset in Times New Roman by KnowledgeWorks Global Ltd.

To my parents who always believed in me

Contents

List of figures List of tables Acknowledgements 1 Introduction Social complexity  1 Complex systems thinking  4 Complex systems thinking in archaeology  6 Structure of the book  13

ix xi xii 1

2 The history of complexity in archaeology Introduction  15 Social evolution  16 Urbanisation and state formation  21 Systems thinking and complexity  27 Complex systems thinking  33

15

3 Conceptualising social complexity Introduction  63 Building blocks of social complexity  65 Complexity formation  81 Outcomes of complexity  93 Multi-scalar dynamics of change and stability  107

63

4 Social complexity trajectories in Anatolia Introduction  117 Case study: Social complexity in Anatolia  123

117

viii  Contents 5 Conclusions Reflections on the case study  197 Reflections on the conceptual model  202 Reflections on the future  210 Bibliography Index

197

212 247

Figures

1.1 Different levels of theory and empirical data within an overarching conceptual framework 11 1.2 Gartner’s hype cycle 12 2.1 Structure of a complex system 35 2.2 A small world network 42 2.3 Preferential attachment structure 43 2.4 Different types of rank-size distributions 48 2.5 Theoretical scaling graphs with super-linear, linear, and sub-linear growth 49 3.1 Flows of energy, resources, and information 64 3.2 Non-linear flow of temporal change 74 3.3 Canonical loop of collective decision making 83 3.4 Positive feedback loops of information processing and problem solving 85 3.5 Basic model of complexity formation 92 3.6 Energised crowding as a generative driver of community formation94 3.7 Flows of energy and resources between society and nature in a social metabolism model 105 3.8 Adaptive cycle 108 4.1 Gini coefficients for New World and Old World polities 119 4.2 Composite complexity metric based on PCA from Seshat project121 4.3 Complexity trajectory of the Konya Plain 124 4.4 Map of case study of social complexity trajectories in southwest Anatolia 125 4.5 Main (late) chalcolithic sites discussed in the text 126 4.6 Main EBA sites discussed in the text 133 4.7 Main MBA and LBA sites discussed in the text 149 4.8 LBA polities in Anatolia 153 4.9 Main EIA, MIA, and LIA sites mentioned in the text 164 4.10 MIA and LIA sites in the area of Lake Burdur, with indication of the study area of the Sagalassos project 173

x  Figures 4.11 MIA and LIA poleis identified by the Copenhagen polis Centre in the Aegean and Anatolia 4.12 Hellenistic city foundations in Anatolia 4.13 The territory of Sagalassos 4.14 Main Hellenistic sites in southwest Anatolia mentioned in the text 5.1 Diachronic overview of flows of energy, information and resources, push-pull dynamics and complexity mechanisms of diversification, integration, and intensification 5.2 Basic model of complexity formation and its application on the level of communities and polities

182 184 188 190 201 204

Tables

3.1 4.1 4.2 4.3 4.4 4.5

Push-pull dynamics on multiple scales Social complexity trajectories in the chalcolithic Social complexity trajectories in EBA Social complexity trajectories in MBA and LBA Social complexity trajectories in IA Social complexity trajectories in the Hellenistic period

89 131 145 161 179 193

Acknowledgements

It does not take a village to write a book. But it does take a village to give the writer the opportunity to do so. Many parts of this book go back to my PhD research, which I started back in 2014. A long journey and this means many people to thank. I cannot mention everyone by name so first a general “Thank You” to everyone who helped and supported me over the years. Special thanks to all the people who helped with making this work see the light of day. Thanks to Danae and Bas for their help with making some of the figures and Ralf for his feedback on parts of the text. An important word of thanks goes to my PhD supervisor and mentor Jeroen Poblome. Thank you for your support, intellectual guidance and stimuli which set me on my way to explore this fascinating topic. Gratitude to all of my collaborators and colleagues with whom I had the pleasure of working with over the last few years, within and beyond the Sagalassos Project. Special thanks to Peter for the whiskeys we shared at Sagalassos, to Sam for being an amazing office-mate and friend, and to Bas for the many great times we’ve shared over the years. I could always turn to them if I needed advice, discussions or share a drink. Thanks as well to the best circle of friends one could wish for: Maarten, Hans, Kristof, Stijn, Hanske, Frauwe, Griet, Eef, Leen, and Lise. Immense gratitude to my family who have always supported me, especially my parents who have always been there for me. Their unconditional support and love has made me into the person I am today and in everything I do, I aim to make them proud. A final word of thanks is reserved to a special person in my life. Thank you Danae, for your love and support over the years. It is a blessing to share my life with you. Thanks again to all of you for your help, support and love to make this book possible. You are a wonderful village and I’m truly grateful to be part of it.

1

Introduction

The aim of this book is to champion complex systems thinking as a major conceptual approach to study social complexity in archaeology. At the beginning of the new millennium, Stephen Hawking famously claimed that the 21st century would be the ‘century of complexity’. In the first two decades of that century, complexity science has started to gain ground as a core theoretical and methodological approach. Studying complexity entails trying to understand the underlying patterns and common mechanisms in physical, biological, social, and technological systems. It is about uncovering how and why the adaptive interactions of agents within these systems produce evolution and change expressed in – oftentimes surprising – emergent behaviour. Over the last few decades, archaeology has also increasingly come to realise the enormous potential of this approach. Summarising some of the trends and evolutions of the use of complex systems in archaeology is one of the goals of this book. Yet, it is not a retrospection. There is still much left to be done to develop complex systems thinking as a major research framework in archaeology. It is not my intention here to advance complex systems thinking as a new paradigm for archaeological theory and practice. The days of the ‘paradigm wars’ in archaeology are long gone and good riddance to them. Paradigmatic modes of thinking result in overly simplified characterisations of research traditions that “reduce ‘the other’ to a fixed point which provides leverage against the current dynamic” (McGlade and van der Leeuw, 1997, p. 1). Archaeologists today prefer to stress the plurality of theoretical approaches to construct interdisciplinary, integrative, and synthesising research frameworks (Altschul et al., 2018). Instead, I aim to integrate complex systems thinking as part of a set of conceptual and methodological tools derived from archaeological theory, social practice theory, and resilience science to build a multi-faceted conceptual framework for the study of social complexity.

Social complexity Over the last 12,000 years, human societies have changed in dramatic ways. We have gone from small hunter-gatherer groups to highly urbanised communities and industrialised nation-states in a globally interconnected world.

2  Introduction These changes are typically considered indicative of a massive increase in social complexity. Yet, what exactly constitutes social complexity and how it changes over time is not altogether clear and not always made explicit. Social complexity is often used as a catch-all concept for a variety of different processes and characteristics. It has been noted that “Inequality, large-scale networks of cooperation, institutionalized leadership, and hierarchical forms of governance all are central elements of the human career, often lumped under the rubric of social complexity” (Feinman, 2017, p. 459). Archaeological definitions of social complexity commonly involve aspects of scale, functional differentiation, and hierarchical power structures. To illustrate the point, I refer to a definition by Gary Feinman, who stated that social complexity is: “The extent of functional differentiation among social units, [which] may be vertical or horizontal; vertical complexity is hierarchical governance with a degree of concentration in decision making and power, [whereas] horizontal complexity is the differentiation of a population into various roles or subgroups.” (Feinman, 2012, p. 36) The emergence of social complexity has been ascribed to a variety of drivers, including environmental pressures, population growth, innovation and technology transfers, warfare, peer-polity interaction, surplus production, and redistribution and individual agency. Trends of increasing complexity have been mostly noted in: (1) Agriculture (Boserup, 1965; Clark and Haswell, 1964; Minnis, 1996; Nelson, 1996; Wilkinson, 1973); (2) Technology (Arthur, 2015; Nelson, 1996; Wilkinson, 1973); (3) Competition and warfare (Carneiro, 1970); and (4) Socio-political control and specialisation (Carballo et al., 2014; Feinman, 2011; Spencer, 2014; Tainter, 1988). Traditionally, archaeologists consider social complexity as a defining property of ‘complex societies’, as opposed to ‘simple societies’, with the former evolving out of the latter. This is the premise of social evolutionary approaches which will be discussed in the next chapter. While such perspectives have been rightfully criticised and largely abandoned in contemporary archaeology, the moniker of ‘complex societies’ is still fully in vogue. Complex societies can be generally defined as those human societies consisting of large numbers of people, many social and economic roles and large permanent settlements (Barton, 2014). In this book, I will explore social complexity from the perspective of complex systems thinking, focusing particularly on the properties and effects of social interaction and information transmission in human societies. A related definition states that “complex social systems are those in which individuals frequently interact in many different contexts with many different individuals, and often repeatedly interact with many of the same individuals over time” (Freeberg et al., 2012, p. 1787). This approach transcends qualitative dichotomies between simple and complex societies, and

Introduction 3 emphasises instead the (quantitative) differences in scale of flows of energy, resources, and information in societies as the defining aspects of social complexity trajectories. Traditional complexity characteristics such as monumental architecture, political institutions, and economic specialisation are seen as emergent phenomena produced by these underlying flows. Social complexity is a topical subject in archaeology. An interesting debate that has been raging in the last few years involves the so-called ‘moralising gods’ hypothesis or ‘Big Gods’ theory (Norenzayan et al., 2016; Whitehouse et al., 2019b). This theory states that religious beliefs emerged as evolutionary by-products of human cognitive development and were repurposed as moralising tendencies upheld by supernatural surveillance and punishment. It is also claimed that supernatural beliefs reinforce social coordination as a prerequisite for the development of social complexity. The main point of contention is whether the belief in supernatural beings imposing a moral order preceded the development of ‘complex’ societies or not. In case of the former, it is suggested that the moral order imposed by supernatural beings provided a prosocial mechanism to overcome the classic free-rider problem and facilitate social cooperation among non-king members of society, thus allowing group sizes to grow beyond the limits of direct, face-to-face relations, and social complexity to increase (Whitehouse et al., 2019b). Others have disputed such a direct, causal relationship (Norenzayan et al., 2016) and argue instead that a more general standardisation of ritual practices among a large population precedes the emergence of moralising gods and is a far more important factor in the emergence of large social groups and social complexity (Whitehouse et al., 2015). In an attempt to settle this debate – and advance the study of social complexity in general – the Seshat: Global History Databank project is building perhaps the most ambitious comparative research project in archaeology to date, compiling an online database currently containing data from 414 societies covering 30 regions across the world from the past 10,000 years.1 By conducting time series analysis on this dataset, the Seshat project argued that moralising gods only emerged once the rise of social complexity had crossed a certain size threshold (Whitehouse et al., 2019b). In a scathing response, a group of scholars produced heavy criticism regarding the methods, data quality, and biases unaccounted for in the study (Beheim et al., 2019). Other research groups such as the Database of Religious History project2 have also questioned the coding practices and reliability of data collection by the Seshat project (Slingerland et al., 2019). This criticism prompted a string of responses by the Seshat team, in which they defend and uphold their original analysis and results (Savage et al., 2019; Turchin et al., 2019; Whitehouse et al., 2019b). While the importance of this debate – including the original claims as well as the concerns raised against them – must be acknowledged, I will not dwell on the detail of the argument. Instead, I want to use this debate to illustrate the importance of proper data collection, coding practices, and

4  Introduction methods, as well as stress how working in a transparent framework geared towards openness and reproducibility can stimulate the development of ideas and the future advancement of our discipline. The creation of openly accessible databases such as the Seshat Databank is an essential part of the development of archaeology as a scientific discipline. Even though the process is still showing some growth pains, I believe that efforts such as these constitute an important way forward for our discipline by facilitating quantitative analysis and synthetic research on topics such as social complexity. Recently, it has increasingly been argued that we are on the brink of a new era of ‘big data’ in archaeology which will reshape our discipline (Gattiglia, 2015; Graham et al., 2016; Huggett, 2020; White, 2016). More data allows us to distil patterns we could not possibly gather from small datasets. The potential of big data analysis to lift archaeological debates to a higher level is exciting indeed. However, to unlock this potential it is essential to provide both bottom-up and top-down embedding of databased approaches: “… initiatives seeking to marry quantitative and qualitative historical research must work from the ground up … and thereby amassing more granular, accurate and meaningful data on each.” (Slingerland et al., 2019, p. 14) This quote captures the idea that big data analysis in archaeology can only be pursued when supported by a foundation of in-depth expert knowledge and fine-grained data collection, that is, by adding a dimension of contextualisation and embedding. At the same time, it is essential to complement big data analysis with proper theory, conceptual knowledge, and hypotheses to guide our analysis and move beyond ‘blind’ interpretations driven by radical empiricism (Coveney et al., 2016; Huggett, 2020). This book aims to contribute to such an enriched approach by building a general conceptual framework to describe and explain social complexity trajectories. Such a framework needs to be able to deal with the messiness and limitations of archaeological data, while also allowing to draw out wider patterns of interpretation. The theoretical starting point for this approach is situated in complex systems thinking.

Complex systems thinking This book centres on elucidating the potential of complex systems thinking for the study of social complexity in archaeology. I already outlined some archaeological definitions of social complexity earlier. The first question to consider is whether this conceptualisation matches that of complexity as a property of complex systems. The answer is generally no. Only few archaeologists have been considering complexity from the perspective of complex systems thinking. This field originated during the 1970s and 1980s out of

Introduction 5 developments in various disciplines – including physics, biology, chemistry, mathematics, general systems theory, and cybernetics – and has since been applied to a wide range of disciplines. ‘Complexity’ is a general term that has been defined in many different ways. To quote eminent sociologist and network scientist Duncan Watts: “Nobody really agrees on what makes a complex system ‘complex’ but it’s generally accepted that complexity arises out of many interdependent components interacting in nonlinear ways” (Watts, 2011, p. 141). It is, however, a misconception that complexity is inevitably difficult to understand. It may be, well…complex, but that does not mean it remains irreversibly elusive from our understanding. So, what exactly do we mean when we say a system is complex? It is of course tautological to state that complexity is that what defines a complex system. A complex system can be generally defined as “a system whose properties are not fully explained by an understanding of its component parts” (Lewin, 1992, p. x). Frequently cited examples include cities, economies, societies, the brain, the internet, and ecosystems. Obviously, complexity must be something very general if it can be considered a property of such a wide variety of systems. It is indeed telling that textbooks and general introductions to complexity science rarely offer a direct definition of what exactly complexity is. Instead, they tend to focus on a number of key properties of complexity, such as emergence, non-linearity, self-organisation, adaptation, and information processing. The behaviour of a system that displays properties which cannot be explained by the dynamics and properties of its components is called ‘emergent’. Emergence derives from the non-linear aggregation of behaviour, properties and processes on a higher level that transcend mere summation of lower level effects. Human societies, for example, are emergent phenomena, arising from the multitude of interactions between people. Without people or their interactions, societies cannot exist. Yet, it is also impossible to assess how every individual interaction contributes to society and what its effects are. This is what we mean when we say that societies emerge in a non-linear fashion from its constituent interactions. Complex systems are also self-organised. This means that they have no central system of control governing behaviour. Individual components within a complex systems interact and make decisions based on locally-defined information that give rise to global patterns. Complex systems are dynamic because their components continuously adapt through an autocatalytic (i.e. self-­ sustaining) loop of information transmission, information processing, and decision-making. Throughout this book, I will use two terms when discussing the scientific framing of complexity. These are ‘complex systems thinking’ and ‘complex systems science’. With the former, I address the epistemic tenets of complexity, whereas the latter refers to its theoretical and methodological operationalisation. I prefer to talk in terms of complex systems thinking as, to me, it provides a more comprehensive approach to complexity. I am, however,

6  Introduction aware that others might consider complex systems science to be a more comprehensive term than I have defined it here and would consider it to cover both sides of the spectrum that I paint here. One might be tempted to take this explanation of complex systems and conclude that it is all about trying to look at the bigger picture of macrolevel systems. This conclusion, however, does not fully capture what complex systems thinking is. Merely focusing on the big picture inevitably leads to loss of information at higher resolution. You could compare it with someone standing on top of a mountain. From this vantage point, you would have a great view over the surrounding landscape. However, without binoculars you will not be able to make out animals on the plains below. Complex systems thinking would be like bringing binoculars to the summit. It provides a way to direct attention and focus across scales within the system, and manipulate between what is currently foreground and background (Kepler, 2019). One more thing to address is the convergence (or lack thereof) of complex systems and complex societies as defined earlier. If anything, only a weak parallel can be drawn. Complex societies often consist of differentiated structures such as social groups, classes, specialized labour, etc. Evaluations of the emergence of new political actors, levels of organisation, and hierarchical social relations are all aspects of interest from a complexity perspective, yet they do not match the full scope of a complex system. No inherent equivalence exists between complex societies in an archaeological sense and the more general phenomenon of complex systems (Auban et al., 2013, p. 53). All human societies, be they classified as socially complex or simple, are intrinsically complex systems in the sense of open systems requiring energy input, regardless of their size or organisational structure. The crucial question remains how to interpret the unmistakable changes in social complexity in human societies from the Pleistocene until today. I argue that, from a complex systems perspective, the traditional characteristics of complex societies should not be considered as direct expressions of social complexity as such, but rather as emergent phenomena arising from social complexity trajectories. The latter are driven by transformations in energy capture, information processing, and the transmission of information and resources.

Complex systems thinking in archaeology This book is not ground-breaking (that might seem like a self-deprecating statement but that does not mean it is not valuable) and it is definitely not the first application of complex systems thinking in archaeology. Over the last few decades, several important works have already paved the way. So far, however, we have not yet seen a comprehensive overview of complex systems thinking in archaeology or an attempt to propose a synthesising

Introduction 7 conceptual framework to study social complexity in a long-term historical perspective. This book aims to contribute towards addressing these gaps. One might raise the question why we even need complex systems thinking as a framework to study the past. I am well aware that for many archaeologists, the introduction of a ‘foreign’ theoretical framework, especially when coming from the so-called ‘hard’ sciences, can come across as intimidating. In the natural sciences, there is also the cliché of the old physicist moving to other disciplines (such as biology) to ‘solve’ their longstanding problems. Introducing complex systems approaches in archaeology might feel to some like opening the door to the physicists and inviting them in, only to find ourselves kicked out from our own party. Anyone not familiar with complex systems and their theoretical underpinnings, might indeed wonder why complex systems thinking would offer a suitable framework for archaeological research. If we put this question into general terms, we discern the same duality touched upon earlier regarding the difference between complex systems thinking and complex systems science. That is, whether the epistemic tenets of complex systems thinking are valid for understanding the past. And if so, whether complex systems science offers the right tools to study the past. The philosopher of science Adrian Currie offers a suitable approach to frame this question by considering a discipline’s epistemic situations and its epistemic resources. An epistemic situation can be interpreted as “the challenges scientists face when generating epistemic goods in a particular context” (Currie, 2018, p. 15). We can understand this as the inherent possibility of finding/generating suitable data to study a given phenomenon – in this case, social complexity – and the epistemic suitability of a given framework to capture these phenomena. Epistemic resources on the other hand, are the knowledge, capacities, evidence sources and methodological tools available to scientists to study any given epistemic situation. First, let us address the epistemic situation. The phenomenon we are interested in here is social complexity in the past. The question is therefore whether we can study social complexity in the past in general, and through complex systems thinking in particular. The answer to the first question can be quite short: Yes! Social complexity has long been one of the core topics of archaeology, and is still touted as one of the grand challenges for archaeology in the 21st century (Kintigh et al., 2014). Throughout this book, I will present many examples of successful studies of social complexity in the past. Regarding the second question, it should be clear to the reader that the answer – for me at least – is an equally short and strong ‘yes’. Why else would you be reading this book? Still, it is necessary to elaborate. The past is inherently knowable because it leaves ‘traces’, that is, an epistemic causal descendant of that event in the present. In other words, it carries information about that event (Currie, 2018). As archaeologists, we use the material traces from people in the past to learn about them and the world they lived in. To do so, we are generally

8  Introduction limited to those activities leaving material remnants to be observed in the archaeological record. With the passing of time, more traces disappear. The view of the archaeologist is therefore always limited and incomplete. This does not mean that archaeology is epistemically impossible or that we have nothing to work with. Archaeologists uncover traces of the past and generate new data every day. One way to improve the epistemic situations in archaeology is to improve our epistemic resources. As stated by Lewis Binford (1968, p. 22), the limitations of doing archaeology lie not in the archaeological evidence itself, but rather in the resources that we bring to its investigation. These improvements can entail methodological advances to capture data, such as with the introduction of 14C dating. The track I will explore here in this book, however, is improvement in the conceptual and theoretical frameworks to understand data. Philosopher of science Derek Turner (2007) discusses the dual role of background theory as a catalyst for knowledge. On the one hand, background theory can limit our knowledge. This is what happens when scientific ‘ways of doing’ become so entrenched they limit the questions that are to be asked as well as the various ways to answer them. On the other hand, background theory can also amplify our knowledge by allowing new connections to be built that could lead to new ways of thinking and new questions to be asked. Being solidly grounded in the background theory of one’s discipline makes it easier to see links with other fields that complement the own background. Creative combinations of relevant background theories to match a given epistemic situation can in these cases stimulate innovative perspectives. The creative combination of converging strands of data, methods, theories and insights from various disciplines to create a common groundwork of explanation subscribes to a tradition of research subsumed under the general moniker of ‘consilience’ (Wilson, 1999). The principle of consilience entails collecting evidence from independent sources to build stronger support for one’s conclusions. This is exactly the kind of approach that has driven complex systems scientists to explore the common mechanisms of complexity in a wide range of systems to great avail. In archaeology as well, we have a long tradition of looking beyond our own disciplinary boundaries for suitable theories and methods to explore the past, to the extent that historical scientists have been described as ‘methodological omnivores’ (Currie, 2015). I strongly believe that there is a natural match between archaeological ways of thinking and doing on the one hand, and the tenets of complex systems thinking and complex systems science on the other. While I will extensively make use of the methodological tools of complex systems science in this book, the main goal will be to explore complex systems thinking as a suitable epistemic framework to study social complexity in the past. This is done first and foremost by building a conceptual model of social complexity, integrating the tenets and concepts from complex systems thinking with elements from archaeological theory, social practice

Introduction 9 theory, and resilience science. To make such an extensive interdisciplinary cross-fertilisation work, we need to be explicit about transposing ideas and theories between disciplines. Andreas Wimmer (2006) identified four modes of cross-disciplinary concept migration: tool transfer, methodological analogy, model migration, and metaphor move. For this discussion, I will focus particularly on the latter two. Model migration entails the transfer of the theoretical propositions and empirical terms of a model into a new situation. Metaphor moves entails the usage of illustrative devices to explain complex processes or models that are difficult to understand intuitively in a novel way. Both of these heuristics have their own scope and explanatory potential. A metaphor can be potentially applied to a wide range of cases but provides less detailed insight into the workings of these systems. A mathematical model can only be applied to specific cases for which it is valid, but allows us to exactly trace the dynamics of those systems. Using consilience to create new conceptual frameworks requires creative combination of background theories by using the proper mode of cross-disciplinary concept migration. It is absolutely essential when transposing concepts or ideas from one discipline to the next, that we are aware of its original role as a heuristic device, and thus of its scope and explanatory potential. Metaphors tend to be dismissed as mere rhetorical window dressing; however, they play an important role in innovative thinking and can reveal potentially fruitful connections that lead to new insight. A metaphor entails an intrusion from one domain into another. As a result, metaphors never fully fit onto the new context. Yet, it is from this tension between the familiar and unfamiliar that the innovative potential of the metaphor arises. Metaphors can therefore act as lynchpins for scientists from different disciplines to engage in dialogue and increase understanding of each other’s work. However, we have to be aware of their limitations. One of the main differences between metaphors and models is the limited susceptibility for empirical testing of the former. This means that as scientists, the metaphor cannot exist by itself. This is why a useful metaphor has been described as “an invitation to hard work” and that “every science must start with metaphor and end with algebra” (Gray and Macready, 2019, p. 131). What is meant is that different degrees of formalisation are required to build scientific frameworks. The metaphor can only ever be a starting point for further methodological and mathematical operationalisation. Yet, the full trip from metaphor to algebra is the journey of a discipline and beyond the scope of a single book. What this book aims to accomplish is to take some steps on this journey. Some headway has been made in applying complex systems thinking in archaeology and a significant portion of this early work has been aimed at addressing social complexity. As of yet, however, there has not been an attempt to build a comprehensive conceptual framework using the core tenets of complex systems thinking to study social complexity in archaeology. Complex systems science is built around terms such as emergence,

10  Introduction adaptation, networks, evolution, phase transitions, self-organisation, criticality, as well as even more metaphorical concepts such as the edge of chaos and the butterfly effect. Some of these terms have been used to describe patterns and processes in the past, more often than not in a metaphorical sense. From the preceding discussion one can surmise that I do not oppose such metaphorical application. However, as stated before, the metaphor is an invitation to hard work. The first steps may have been taken, but there is still a long journey ahead of us. To ensure that in this endeavour we retain the match between the epistemic situation and epistemic resources discussed earlier, we have to keep in mind that all theories and concepts brought in through this interdisciplinary cross-fertilisation need to be suited for archaeological applications. To do so, we need a multi-layer model consisting of different levels of theory and evidence. This way, we can build a bottom-up approach, starting from theories covering micro-level processes to high-level theories covering general patterns of stability and change. For this approach, it is essential to work through creative combinations of middle range theories (MRT) to bridge the gap. MRT were first proposed by the sociologist Robert Merton (1968) as sets of theories that lie between the minor, but necessary, working hypotheses that evolve during day-to-day research, and the all-inclusive systematic efforts to develop a coherent theory that explains uniformities of social behaviour, social organisation and social change (Merton, 1968, p. 39). MRT mediate between general theories that are too distant from the observed behaviour and idiosyncratic descriptions of such behaviour without further degree of generalisation or wider application. MRT were first used in archaeology by Lewis Binford (1977) who adapted the term as it was defined by Merton to denote a way of bridging the ever-changing behavioural dynamics of human societies in the past and the static phenomena or traces of the archaeological record we are left with today. In other words, MRT give evidential relevance to material remains (Kosso, 2001). Here, I combine Binford’s approach to MRT as a way of bridging theory and data with its original meaning of middle-range theorizing, in order to connect, on the one hand, different sets of theories on an epistemological level, while on the other hand, confronting, investigating, and integrating the role of the actual archaeological data in these sets of theories. In this sense, I follow the interpretation of MRT by Norman Yoffee (2005, pp. 186–187), who stated that: “Levels of archaeological theory exist, if indeed they exist at all as discrete levels, as a hierarchy of propositions that afford linkage between matters of data collection (and the primary analysis of data) and the process of inference within which patterns of data are held to represent social phenomena. The levels of theory I demarcate are hierarchical only in degrees of abstraction, not in chronology of employment (or even in importance, which can be debated).”

Introduction 11

Figure 1.1  Different levels of theory and empirical data within an overarching conceptual framework

In this book, I will develop and apply a conceptual framework consisting of multiple levels of theory, each with varying scope and applicability (Figure 1.1). It is through the creative connections of the constituent components that this overarching framework will be able to provide an encompassing model of social complexity. One of the main challenges for building a suitable theoretical framework to describe and understand social complexity based on complex systems thinking, is that it necessarily has to cover multiple scales of analysis. From individual actions, interactions and behaviour to collective societal dynamics. Such a framework will have to deal with what is considered fundamental and what is emergent. Fundamental objects or processes are those elements that cannot be decomposed further into constituent elements. In the framework proposed here, the fundamental elements are social interactions and information transmission. This need not mean that there is no theory available that distinguishes constituent components in these elements. Neuroscience, for example, has extensively studied the components of human decision-making underlying much of our interactions (Friston, 2010). However, these are not essential for our purposes here, much in the same way as, for example, molecular chemistry not having to explicitly deal with quantum mechanics for most of its applications and studies.3 Moving from the fundamental to the emergent, this model aims to understand how social interactions and information transmission give rise to human societies embedded in a framework of stability and change in social-ecological systems. For such a model, we need to elucidate the connections across scales. In physics terms, the process of relating theory for smaller scales to larger ones is called ‘coarse-graining’. This entails that processes and objects on lower resolution levels can be described by medium

12  Introduction and high-resolution levels, but not the other way around. Simply put, it entails a way to understand behaviour at a fundamental level by referring to a coarse-grained (simplified) representation on a higher (or emergent) level. This process inherently causes information loss. It is virtually impossible to incorporate every interaction between every person to explain the structures of societies. Even if we could, it would not result in any meaningful knowledge about the system due to the non-linear nature of dynamics in complex social systems. Yet, we also gain something in return. Coarse-graining provides perspective to distil patterns and connections that create structure out of seeming disorder. It provides a pathway to move from the messiness and multitudes of archaeological data to build structured models with explanatory power regarding the underlying drivers of system dynamics. It is my hope that the model proposed in this book will prove useful for fellow archaeologists and other scientists interested in long-term patterns of social complexity trajectories. I fully believe that we have only scratched the surface of what complex systems thinking could bring to our discipline. This book aims to uncover more of this surface, even if only to show how deep we can still go. At the same time, it is advisable to keep in mind the warning posited by Jacob Freeman and colleagues (2017, p. 1), who stated: “Yet, the history of theory in archaeology is littered with concepts and models that practitioners initially find useful, but then abandon because, although initially stimulating, the concepts fail to generate testable hypotheses that advance research”. This quote reminds of the famous Gartner’s hype cycle for emerging technologies (Figure 1.2) which outlines how, following a moment of technological innovation, expectations for the innovation rise rapidly and lead to a peak in visibility, which is followed by a

Figure 1.2  Gartner’s hype cycle Source:  Creative commons, made by jeremykemp and reused under CC 3.0 license

Introduction 13 negative hype period where the technology does not live up to expectations. The final stage of the cycle is the phase of technological maturation where both the potential and limits of the technology are understood. The application of complex systems thinking is still gaining ground in archaeology but I do not know where we currently are. Maybe we are still rising towards the peak of inflated expectations, or perhaps some have already started to slope down towards disillusionment. What I do know, however, is that there is still a wide and unexplored plateau of productivity out there, waiting to be discovered. If anything, I hope this book can at least point towards the general direction of this plateau and its potential yields. Our discipline only has to benefit from it.

Structure of the book This book is built on three main chapters following this first introductory chapter. The second chapter will present an overview of the study of social complexity in archaeology, starting with its usage in social evolutionary perspectives. I then continue with a discussion of urbanisation and state formation, as two major macro-scale processes commonly associated with the emergence of social complexity in archaeological discourses. Next, I will discuss the role of complexity in systems theory and social systems thinking in archaeological research. Finally, I will outline the general properties of complex systems and their applications in archaeology in recent years. I will focus in particular on the results of four fields, correspond largely to four methodological approaches: network science, settlement scaling, cultural evolution and agent-based modelling.4 By discussing the general theoretical underpinnings for complex systems approaches, as well as highlighting the achievements and limitations of these works, this part will pave the way for the in-depth conceptualisation of social complexity in the subsequent chapter. The third chapter forms the core of the book and presents the conceptual model of social complexity formation. Following the tenets of complex systems thinking, I identify the fundamental building blocks generating social complexity as an emergent property in human societies. These building blocks are social interaction and information transmission. ‘Bundles’ of interactions and associated information are gathered as social practices, which in turn are grouped in social structures. I then discuss how these building blocks result in the emergence of social complexity by outlining a model of complexity formation. This model starts from selection pressures acting as input information for (collective) decision-making strategies, using mechanisms of problem-solving to generate collective action strategies that can act as pushing or pulling forces for societal development. This iterative problem-solving loop of information transmission, information processing and decision-making continuously produces a series of outcomes that affect the flows of energy, resources and information, and hence, social complexity trajectories. I then make this process more specific by discussing some

14  Introduction outcomes of complexity formation, most notably community formation and polity formation. These can be considered as generalised models of change on different societal scales, which act as the counterparts for urbanisation and state formation discussed in the second chapter to illustrate the potential of the model. After discussing communities and polities as particular loci for social complexity formation, I extend the focus on social dynamics to include its environmental ramifications. I first outline the energetic costs of complexity that can result in collapse and transformation of societies, before moving on to elucidate the theory of social metabolism as a suitable framework to approximate flows of energy and resources between societies and their natural environment. Finally, I will present the concept of adaptive cycles and panarchy as the high-level theoretical lynchpin to combine all of the previous elements into a coherent framework of multi-scalar dynamics of stability and change in social-ecological systems driven by social complexity trajectories. It has been lamented that studies of social complexity too often fail to include interdisciplinary perspectives or apply overly narrow lenses of inquiry, both in spatial/temporal and theoretical scope (Chacon and Mendoza, 2017, pp. 2–3). While it is not the main goal of this book, it does aim to address both issues. The wide scope and theoretical richness of the third chapter provides a multi-dimensional model of complexity trajectories. In the fourth chapter, this model will be applied to an extensive case study focusing on southwest Anatolia (modern-day Turkey) from Chalcolithic to Hellenistic times to illustrate the potential and explanatory power of the conceptual model. In the fifth and final chapter, I will provide a concluding synthesis, summarizing the main points and conclusions of the book. This chapter will emphasise the added value of social complexity as a core concept for studying past human societies and complex systems thinking as a core conceptual approach. I will also provide some general outlines towards further development of the model, particularly ways of operationalising it in future work, as well as posit some recommendations for future research avenues.

Notes 1. Publications and data of the Seshat Project are openly available from: http:// seshatdatabank.info/. Results of studies based on the Seshat database will be explored in more detail in subsequent chapters. 2. An online encyclopaedia for quantitative and qualitative data on religious cultural history: http://religiondatabase.org. 3. That being said, there is a field called quantum chemistry which deals exactly with the problems of applying quantum mechanics in chemical systems. What is considered fundamental and what is emergent is very much a matter of the appropriate scale of analysis. 4. It should be noted, however, that these approaches are not completely separate endeavours and are not necessarily mutually exclusive.

2

The history of complexity in archaeology

Introduction This chapter will present an overview of the topic of social complexity in the study of the human past. This is not meant to be an exhaustive overview, but rather a guide to trace the general lines of development and provide some background to the approach advocated in this book. Social complexity has a long pedigree in the study of human history and can be traced back to social evolutionary approaches from the second half of the 19th century. Ever since, social complexity is commonly considered a property of ‘complex societies’, as opposed to other, so-called ‘simple’ societies (Lull and Micó, 2011). In this sense, social complexity is commonly considered to pertain to the rise and development of (social, political, and economic) hierarchies and inequality associated with societies consisting of large, dense populations, socio-political stratification, formal information systems, and socio-economic specialisation. The concept is generally used to describe trajectories of societal evolution and change on two scales: settlements/communities and polities/societies. For the former, the transition from village to urban communities is particularly seen as an important transition phase of social complexity. In the latter, societies are ranked in a number of discrete categories characterised by a certain degree of complexity. This categorisation of complexity often implies an inescapable, teleological evolution from egalitarian and smallscale societies, such as ‘Bands’ or ‘Tribes’ to increasingly socially stratified and large-scale societies such as ‘Chiefdoms’ and ‘States’. Cities and states are traditionally considered the end-point of trajectories of increasing complexity on both levels and will be considered in detail later on in this chapter. Even when dissociated from any explicit teleological implications, implicit biases can be highly problematic if assessed without a proper framework and should be approached with caution. This chapter will start with a discussion of the development of social evolutionary approaches in archaeology. This will be followed by an overview of some features of urbanisation and state formation. Next, I will discuss applications of systems theory in archaeology as precursors to complex

16  The history of complexity in archaeology systems thinking. Finally, an alternative approach to social complexity will be presented based on complex systems science, which will form the basis for the model of complexity presented in the next chapter. I will finish the chapter with four main areas of complex systems approaches in archaeology: network science, settlement scaling, evolutionary approaches, and agent-based modelling.

Social evolution Social evolutionary approaches became increasingly popular in the second half of the 19th century, through the works of scholars such as Herbert Spencer, Lewis Morgan, Edward Tylor, Karl Marx, and Friedrich Engels (Sanderson, 2001). Social evolution uses the principles of biological evolution to explain (historical) processes of social change. Biological evolution is built on three central pillars: natural selection, variation and inheritance through genetic transmission. These pillars already go back to the works of the English naturalist Charles Darwin in his work On the Origin of Species (1859). The natural environment offers certain selection pressures that allow certain individuals and species to either thrive or decline, depending on its characteristics, during the competitive struggle for resources. These characteristics are not set in stone but can change through genetic mutation and the interaction of individuals with their environment, resulting in variability in the natural fitness of an individual, typically measured through the amount of viable offspring it can produce. ‘Fitter’ individuals transmit their genes to a higher amount of offspring, thus having better chances to sustain their lineage and transmit their characteristics through the population. In biological systems, species increase complexity as a response to selection pressures to improve their fitness in an ever-evolving bid to adapt to environmental changes. Competitive interaction occurs between species and environment, as well as among different species. Processes of bilateral evolution have also been described as the ‘Red Queen effect’, where two connected species, predator and prey, must keep on adapting and evolving, not only to gain a reproductive advantage, but to maintain their respective places within an ever-changing evolutional ‘rat-race’ (Van Valen, 1973). The phrase ‘survival of the fittest’ is often attributed to Darwin, however it was originally coined by a contemporary scholar, the English sociologist Herbert Spencer in his work Principles of Biology (1864). Spencer is considered one of the key figures of ‘Social Darwinism’ in the last part of the 19th and early 20th centuries. As part of this movement, the principles of biological evolution were used as an explanatory framework for human behaviour and the historical development of human societies, particularly, so-called complex societies such as the state (Lull and Micó, 2011). Social evolution holds at its core that human groups undergo directed processes of

The history of complexity in archaeology 17 social change driven by adaptation to external circumstances resulting in an inherent tendency to increasing complexity over time. The latter part of that definition in particular is problematic. The idea is that continued adaptation drove societal development along various stages in a progressive trajectory towards increasingly complex modes of sociopolitical configurations. Crucially, however, in contrast to Darwin’s model of biological evolution, Spencer did not propose equivalent causal mechanisms of change in social evolution (Manning, 2020). Without proper causal mechanisms, social evolution can only be defined in general terms, lacks explanatory value and does not stand up to scrutiny as an analytical framework. Even more problematically, explicit links between social complexity and (evolutionary) fitness were not only based on Darwinian biology, but also rooted in the works of 18th century scholars of the Enlightenment, such as Montesquieu, Miller, and Adam Smith (Chapman, 2007, p. 13). Based on these works, notions of successful complex societies came to be appropriated by Western societies and embedded in a wider framework of ‘Eurocentrism’ and Western cultural superiority, among others in a bid to justify Western colonialism (Morris, 2013, p. 2). Within the zeitgeist of the 19th century, scholars interpreted the nation-state – as the most prevalent form of social organisation at the time – as the culmination of human evolution and therefore as the best possible form of human society. Increasing social complexity growing out of adaptive processes was seen as part of a teleological and ‘inevitable’ trajectory. Through the biological analogy, highly complex societies within evolutionary trajectories of societal development were associated with notions of ‘successful’ societies, outperforming their ‘simpler’ counterparts. Let us, for example, take the work of John Fiske published in 1873 with the telling title “The Progress from Brute to Man”. If we just briefly look at the opening sentence of his work: “The chief difficulty which most persons find in accepting the Doctrine of Evolution, as applied to the origin of the human race, is the difficulty of realising in imagination the kinship between the higher and - the lower forms of intelligence and emotion.” Fiske (1873, p. 251) It is clear that a certain ranking of ‘higher and lower forms of intelligence and emotion’ is presupposed here. A few sentences later, Fiske asks how “a race endowed with such a capacity for progress be genetically akin to those lower races of which even the highest show no advance from one generation to another” (Fiske, 1873, p. 252). While he primarily emphasises the difference between humankind and its primate ancestors rather than explicitly discussing differences between human races or societies, implicit biases of success and failure are clear for all to see. Many evolutionary approaches assume an inherent tendency towards greater complexity (Tainter 1996). Complexity was thus long assumed to be a desirable thing, and the logical

18  The history of complexity in archaeology result of surplus food, leisure time, and human creativity in past societies (Bronowski 1973; Steward 1972). Driven by the core tenets of empiricism as developed during the Enlightenment, scholars such as Lewis Henry Morgan, Edward Tylor, and John Lubbock increasingly strove to document the proposed trajectory with ethnographic data (Robb and Pauketat 2012, p. 6). It was supposed – out of the idea of a unitary conception of human nature – that all societies developed in similar and comparable ways, and therefore passed through the same series of successive steps in a linear and teleological fashion, culminating in the State (Lull and Micó, 2011, p. 136). This way, these proposed trajectories were thought to be applicable to all human societies throughout time and space. The American anthropologist Lewis Henry Morgan used anthropological observations of ‘primitive’ contemporary societies to reconstruct the past stages of more advanced societies (Morgan, 1877). Morgan identified three main phases: ‘Savagery’, ‘Barbarism’, and ‘Civilisation’, of which the first two were again divisible in three levels, ‘Lower’, ‘Middle’, and ‘Upper’. Definitions of these phases and the transitions between them were based on technological advancements. For example, the manufacture of pottery vessels as the crucial technological innovation which separates Upper Savagery from Lower Barbarism or the development of a phonetic alphabet as the technological innovation marking the transition from Upper Barbarism to Civilisation. Although contributing greatly to the development of anthropology as a discipline, evolutionary discourses came under scrutiny in the early years of the 20th century. In particular within the framework of German idealist philosophies, most famously propagated by Georg Hegel, unilinear evolutionary trajectories of societal development were increasingly criticised. The image of a unitary human nature was substituted with myriad of juxtaposed cultures. The movement of ‘Historical Particularism’, most notably advocated by Franz Boas, criticised the inappropriate methodology of the evolutionary comparative method as it relies upon ordering of synchronic data to make diachronic inferences (Sanderson, 1999, pp. 15–16). Instead Boas argued for the need for contextual study of human cultures in and by themselves and in their own worth (Lull and Micó, 2011, p. 148). This led to the abandonment of attempts at uncovering general regularities and evolutionary trajectories of universal validity in human societies. During the second half of the 20th century, scholars became increasingly dissatisfied with overt particularism in anthropological and historical research and were again looking to include comparative perspectives. A countermovement started where ideas from 19th century evolutionism were adapted and appropriated within a neo-evolutionary, framework. In anthropology, the works of Julian Steward and Leslie White were fundamental for this resurgence (Steward, 1949; White, 1949). Much like earlier evolutionism, different stages of human societies were postulated.

The history of complexity in archaeology 19 However, the importance of technological innovations was combined with a strong emphasis on different forms of political organisation. More specifically, the implementation of political hierarchies and the institutionalisation of leadership in society, ranging from simple and hardly formalised types of leadership to permanent, centralised and highly regulated forms (Lull and Micó, 2011, p. 161). While neo-evolutionism retained a basic succession of societal configurations, this trajectory was dissociated from explicit teleological notions of unilinear development. White (1949, 1943) considered the key concept to trace different stages of societal complexity to be consumption of energy. Theoretical foundations for this key understanding were found in the first two laws of thermodynamics. It was argued that human cultures provided a notable exception to other systems bound to this fundamental law of physics as humans could postpone inevitable energy loss by harnessing technology and technological innovation to gather external energy, resulting in increasingly elaborate culture systems.1 A similar stance would later be propagated by Bruce Trigger, who argued that evolution in human history was driven by a directionality involving an overall tendency towards larger, more differentially articulated structures requiring greater per  capita expenditure of energy for their operation (Trigger, 1998, p. 10). Conceding to previous concerns regarding the particularity of individual societies and culture, Steward (1972) argued that not all intermediate steps of societal development needed necessarily be fulfilled within a similar, unchangeable trajectory, allowing more varied, multilinear trajectories of development. Important contributions to the neo-evolutionary approach were made in particular by the anthropologists Elman Service and Morton Fried. Service proposed different steps of societal development consisting of ‘bands’, ‘tribes’, ‘chiefdoms’, and ‘states’, based on formative elements such as group size, group cohesion, form of leadership and nature of the means of subsistence (Service, 1962). Fried on the other hand divided human societies in ‘egalitarian’, ‘ranked’, ‘stratified’, and ‘state societies’ based on Marxist ideas of growth of social inequality (Fried, 1967). In both classifications, each of the categories functioned as a scale of reference in order to classify different societies based on modes of socio-political institutionalisation. As they have been described up until now, these models were largely formulated on the basis of anthropological observations and defined through generalising features of societal development. These descriptive modes of social organisation decontextualised societal development by lifting processes out of their spatial and temporal context. Criticism from archaeologists led to a call for ‘materialisation’ of these concepts in the 1960s and 1970s. However, the archaeologist – diligent as (s)he may be – can never directly dig up institutions or socio-political organisations. A way had to be found to interpret the archaeological material as a reflection

20  The history of complexity in archaeology of institutional organisation in society. It was therefore essential to identify those formative elements of material culture which could be linked to the prevalent political configuration of society and could thus be assigned explanatory value. A prominent example of archaeological research trying to tackle this problem is found in Kent Flannery’s analysis of socio-political complexity (Flannery, 1972). His approach was based on the analysis of exchange of information and different levels of decision-making institutions within processes of administrative centralisation and mechanisms of control. A connection can be made with the cybernetic approach of Henry Wright (1978, p.  56) who defined the state as “a society with specialised decision-making organisations that are receiving messages from many different sources, recoding these messages, supplementing them with previously stored data, making the actual decision, storing both the message and the decision, and conveying decisions back to other organisations”. Degrees of socio-­political complexity can therefore be studied through the number of levels within the settlement hierarchy of a given society (Wright, 1969). Chiefdoms possessed one hierarchical level of decision-making to control the basic level of communities organised in villages, whereas state societies possessed two or more hierarchical levels that functioned as mechanisms of control. In conclusion, (neo-)evolutionary approaches to societal change have long focused on transformations between different classes of society, such as from chiefdom to state. As Barton (2014) remarks, this implies that a certain number of fundamental properties co-occur throughout a number of societies across time and space, which allows them to be characterised as belonging to one stage or another. Additionally, it implies that such societies stay in a sufficiently stable form of equilibrium until a sudden ‘jump’ to another stage takes place and a society rapidly reorganises to meet the properties and characteristics of the new stage. When approaching human societies from a complex systems perspective, however, it becomes clear that these various phenomena and properties supposedly characterising a specific stage of development, “do not necessarily co-occur or coevolve, although their trajectories can converge to varying degrees in some cases… If there are universals in the rise of complex societies, it is more likely that they will be found in the underlying processes or algorithms that drive the evolution of complexity” (Auban et al., 2013, p. 34). What exactly constitutes these underlying processes will be elucidated in the following parts of this chapter and the next. The suggested way forward will build on flows of energy and information processing as essential buildings blocks of social organisation and complexity. I will argue that information processing constitutes an important driver of societal change, whereas energy processing constitutes its overall constraints. First, however, I will focus on two processes that have traditionally been strongly intertwined with social complexity: urbanisation and state formation.

The history of complexity in archaeology 21

Urbanisation and state formation Urbanisation and state formation constitute two major modes of social organisation. Both are commonly considered “end-points” in trajectories of complexity (Pluciennik, 2005). I already mentioned how the state was considered the pinnacle of organisational development in human societies from a social evolutionary perspective. Likewise, the emergence of cities and associated urban processes are commonly connected to an upsweep in social complexity on a local and regional scale (Bintliff et al., 2007). In this part, I will discuss some of the general properties of cities and states and how they come to be. It should be noted that this focus on cities and states does not mean that I consider these to be inherently synonymous with complex societies. In this book I apply a complexity theory perspective, in which all human societies can be considered complex systems. Yet, for most of the history of our discipline this is not how complex societies have been conceptualised. The discussion outlined here will be contextualised in a broader approach of complex social systems in the next chapter. To give a full account of the scholarship on both topics would require volumes of their own and cannot be accomplished here. Instead, I will trace the main outlines of their emergence and traits, before discussing their place in social complexity trajectories. Urbanisation Cities have been a common research topic in a number of disciplines including social geography, sociology, anthropology, archaeology, history, economics, demography, psychology, urban planning, and ecology. At the most general level, cities can be seen as a particular form of settlement consisting of a significant spatial concentration of people (Lagopoulos, 2009, p. 1).2 However, exactly how to define and distinguish the city from other forms of settlement is a debated issue. According to George Cowgill, “it is notoriously difficult to agree on a cross-culturally applicable definition of ‘the’ city, but we cannot do without definitions altogether…No single criterion, such as sheer size or use of writing, is adequate.” (Cowgill, 2004, p. 526). Urban sociologists have in the past frequently used static, quantitative properties such as population size or density to define the city. Gideon Sjoberg (1960, p. 83) for example speaks of a limit of “10,000 perhaps only 5,000 persons”. Such a static delineation between cities and other forms of settlement is rather arbitrary as it hardly takes into account the context in which these settlements operated. Static definitions therefore have a hard time accounting for the dynamic nature of urban communities. In the 1920s and 1930s sociologists at the University of Chicago, most notably Ernest Burgess and Louis Wirth, drawing from biological and evolutionary concepts, considered the city as consisting of a specific population creating a

22  The history of complexity in archaeology new environment to live in. Wirth’s (1938) basic definition of the city as a “large, dense, and permanent settlement of socially heterogeneous individuals” already goes a step further by incorporating differences in social roles of its inhabitants. This social heterogeneity is integrated within an overall diversification of settlement roles. Cities can therefore also be considered to occupy a distinct functional role within a larger settlement pattern (Osborne and Cunliffe, 1996). Owens (1991, p. 3) for example stressed the paramount function of the city as a defensive bastion for most of antiquity. But is there any canonised set of functions performed both in and by cities? Several definitions have been put forward, centred on a number of traits deemed indispensable for a city. Oft-quoted is the definition of the German sociologist Max Weber in his work on ‘The City’3: “To constitute a full urban community a settlement must display a relative predominance of trade-commercial relations with the settlement as a whole displaying the following features: 1) a fortification; 2) a market; 3) a court of its own and at least partially autonomous; 4) a related form of association; and 5) at least partial autonomy and autocephaly.” (Weber, 1958, pp. 80–81) We can highlight the emphasis on the defensive function of a city. Fortification walls are indeed commonly considered essential historical identifiers for an urban settlement. The supposed predominance of a commercial nature, however, suggests that this definition cannot be used in archaeological contexts for pre-industrial and pre-capitalistic cities. When looking at other lists of traits for identifying urban settlements, one cannot overlook Gordon Childe’s concept of the Urban Revolution (1950) and his list of associated traits which offered a comparative perspective on urbanism. The list entails: (1) increasing settlement size; (2) centralised accumulation of capital resulting from the imposition of tribute or taxation; (3) monumental public works; (4) invention of writing; (5) development of science; (6) long-distance trade in luxuries; (7) emergence of a class-stratified society; (8) craft specialisation; (9) political organisation based on territorial principles rather than kinship; and (10) development of art. This ‘Weberian’ approach of classification by identifying traits present in known instances of a given phenomenon but not in others, and using these to identify hitherto unknown instances of the phenomenon, has the advantage of being explicit and allowing comparison with other entities across time and space. However, it also carries the inherent risk of creating circular reasoning. One can always wonder on what basis certain traits are selected or whether this selection happens with an implicit view of what the concept is supposed to define. Depending on the context it can be argued which exact traits on the checklist are essential to trace this heterogeneity and whether they truly represent a shift towards urban communities or not.

The history of complexity in archaeology 23 Should writing for example always be considered an indispensable element of urbanisation? Certain settlements such as Uruk in Mesopotamia can easily be considered cities, yet existed hundreds of years before the invention of cuneiform (Van De Mieroop, 2004, pp. 19–23). Scholars have indeed frequently stressed the insufficiency of such ‘checklist’ approaches (Osborne, 1996). Using lists of traits will generally work better to describe a given phenomenon but will have problems explaining why it arose in the first place. What we need to consider therefore are the processes behind the formation of urban communities. Both with Weber and Childe, what appears to be a concise list actually consists of highly different, yet to a certain extent convoluted traits. One way to deal with this convolution is to clearly distinguish between three related concepts, which are often used interchangeably: ‘urban’, ‘urbanisation’, and ‘urbanism’ (Smith, 2003, p. 12). This tripartite division is in part relatable to Harvey’s (1973, 1969) tripartite ontology of space, consisting of absolute, relative, and relational space. ‘Urban’ refers to the material characteristics of a city in its geographic and territorial locality. ‘Urbanisation’ entails a geographical process of agglomeration, covering population movement from rural to urban areas. Finally, ‘urbanism’ refers to the general phenomenon of construction of the city-community as related to social, political and economic aspects (Purcell, 2007, p. 252; Smith, 2003, pp. 12–16). Definitions of the city generally employ traits from all three dimensions. When using these concepts interchangeably without sufficiently clear distinction, one convolutes aspects of materiality, demography, and community formation without distinguishing between drivers or feedback mechanisms. Developments in processes of urbanisation and urbanism are, for example, the main driving mechanisms for the development of urban communities in a physical sense. It must be noted that developments in any one of these dimensions can be linked to a wider process of state formation, but need not necessarily be interpreted within such a framework. In the absence of a single centralised power, the development of cities is for example sometimes explained through the mechanism of peer-polity interaction, with a high degree of homology between local autonomous centres (Renfrew and Cherry, 1986). Greek poleis during the Archaic and Classical periods are considered prime examples of such peer polities (Snodgrass, 1986). Peer-polity development is driven by processes such as competitive emulation, warfare, symbolic entrainment and increasing flows in exchange of goods. In the next chapter I will discuss in more detail how these and other drivers can intensify processes of community formation and settlement networks as key properties of complexity trajectories. First, however, I will shift focus from cities to states as a form of social organisation traditionally seen as the end-point of macro level complexity trajectories.

24  The history of complexity in archaeology State formation In the previous part, I focused on the scale of communities and cities, with particular attention for the transformative processes of urbanisation and urbanism. Throughout much of history, individual communities were often integrated within socio-political units on a higher scale, such as states and empires. In this part, I will focus on state formation, both as a converging and parallel trajectory to urbanisation. The development of cities and states are indeed often considered together as expressions of social complexity trajectories, and it has even been suggested that both emerge simultaneously (Adams, 1966). Some have even gone as far as positing a causal relationship, with claims such as that “cities are only found in societies that are organised as states” (Fox, 1977, p. 24). Going back as far as the 13th century, Arab historian Ibn Khaldun (1332– 1406) argued for the primacy of the establishment of authority in dynastical state structures and considered urban development an inevitable secondary effect of this development. The immense degree of co-operation between people required to build great cities was considered to be only achievably through coercion by a strict ruler within the framework of a state society.4 The same convolution of cities with state formation can be found in Childe’s Urban Revolution. This concept is essentially built around the notion of centralised concentration and management of production surpluses, allowing an increased differentiation in social roles in society through the division of labour. The urban environment itself is then considered the spatial and material expression of this core societal development (Renfrew and Bahn, 2012, p. 167). Another way cities and states are linked is when urban communities emerge in the wake of imperial expansion, with the imperial apparatus stimulating urban development following military conquest as part of a larger strategy of imposing measures of centralised control (Renfrew, 2008, pp. 37–39). More recent studies have dissociated urbanisation and state formation, suggesting that many urban settlements preceded initial processes of state formation (Cowgill, 2004; Jennings, 2016). Others have stated that “the formation of urban communities neither requires nor produces the state” (Osborne, 1996, p. 2). Hansen (1998) for example discusses the Yako in Nigeria, with urban settlements of 2,000 to 11,000 inhabitants, yet with no form of central government. Communities were instead controlled by a crisscrossing system of patri- and matri-clans. In Mesopotamia, urban activity was centred on temples functioning as redistributive centres, preceding palace-centred socio-political units by nearly half a millennium (Robertson, 1995). Urban development should therefore not exclusively be linked to particular forms of socio-political organisation and did not necessarily require a state level organisation. These studies re-emphasise the need to transcend unidirectional evolutionary approaches and instead urge us to address urbanisation and state formation as part of

The history of complexity in archaeology 25 social complexity trajectories in non-linear and multi-directional patterns (Gyucha, 2019, p. 4). State societies have been identified in many parts of the world. Near Eastern states, most notably Egypt and Mesopotamia, are considered to be some of the oldest examples (Bang and Scheidel, 2013). The concept of state has also been used to describe the development of complex socio-political configurations in the America’s (including among others the Tiwanaku, Olmec, Maya, Inca, and Wari civilisations) and Ancient China (with the Han dynasty as most cited example).5 In its broadest sense, the term ‘state’ denotes a form of highly organised socio-political community (Hansen, 1998; Lull and Micó, 2011). The question that can immediately be raised is what exactly is meant with the qualification of ‘high organisation’. Following the discussion earlier in this chapter, it should not come as a surprise that such terminology is derived directly from these social evolutionary approaches. Comparison of ‘high’ forms of social organisation, especially with other, supposedly ‘lower’ forms, is inherently fraught with biases impeding proper analysis. Let us therefore first consider some characteristic elements of the state as a specific form of political organisation outside of any evolutionary trajectories. A recent volume highlighted five main characteristics of the emergence of premodern states: (1) population growth and resource availability; (2) formation of territorial boundaries and monopoly on the use of force within those boundaries; (3) expansion of specialised production and exchange; (4) investment in infrastructure and monuments; and (5) ideological sanctification of political authority (Sabloff, 2018). Trying to define what constitutes a state is inseparably connected with asking how it emerged. Various theories have been posited as explanatory mechanisms of state formation: 1 Trade and economic exchange are considered essential drivers of complex political administration to manage political and economic structures (Wright and Johnson, 1975). 2 Conflict theories concentrate on the various ways in which the state is developed as a mechanism for the benefit of elite groups, most famously advocated in Marxist thought (Fried, 1967). 3 Population pressure inducing the expansion of socio-political structure to manage and control growing population numbers (Johnson and Earle, 2000). 4 The irrigation hypothesis as a specific instance of the development of an elaborate bureaucracy, especially in Mesopotamia, to manage largescale irrigation infrastructure (Steward, 1949). 5 Circumscription theory positing political development and state formation emerging out of competitive warfare over limited environmental resources (Carneiro, 1970).

26  The history of complexity in archaeology Modern scholarship on the state often refers to Max Weber (1972) who offers a definition of the state as a specific type of political organisation “if and insofar as its administrative staff successfully claims the monopoly of legitimate physical coercion in the implementation of its order” (Translation from Scheidel 2013, p. 5). More recent definitions mirror Weber’s emphasis on regulation and coercion, as in the definition by Sanderson (1999, p. 56), who regarded the state as ‘a form of socio-political organisation that has achieved a monopoly over the means of violence within a specified territory’. Tilly (1992, p. 1) stressed that Weber’s view on the monopoly of physical coercion should be nuanced and should rather be thought of in terms of priority rights of coercion, by defining states as “coercion-wielding organisations that are distinct from household and kinship groups and exercise clear priority in some respects over all other organisations within substantial territories”. Carneiro (1970, p. 733) defines the state as “an autonomous political unit, encompassing many communities within its territory and having a centralised government with the power to draft men for war or work, levy and collect taxes, and decree and enforce laws”. Different kinds of coercive power can be observed. Maisels (2010, p. 3) defines the state as ‘control over people and territory exercised from a centre through specialised apparatuses of power that are: (1) military; (2) administrative; (3) legal; and (4) ideological. Power in this sense should be seen as the ability to direct and benefit from the actions of others. Turner defined power as “the capacity to dictate, to varying degrees, the actions of others, whether individuals or collective units” (Turner, 2003, p. 9). The emergence of state polities is inseparably connected with power structures. Power strategies are the means by which ruling segments combine sources of power to pursue their political goals (Demarrais et al., 1996; Mann, 1986). Economic power can generally be most easily controlled because of its material nature. Natural resources and human labour can be concentrated, controlled and defended against others. Social relationships, military power, ideology, and information are in general less easy to control. The capacity to control behaviour in state societies should be seen across all aforementioned sources of power. Development of power structures inherently entails the creation of social inequality. Price and Feinman (2010, p. 2) define social inequality as “unequal access to goods, information, decision making, and power”, adding that the structure of unequal social relations is essential to higher orders of social organisation and is basic to the operation of more complex societies. This is a common approach, relating the development of social inequality to hierarchical modes of organisation. Human society operates through the tension between dominance and equality, between hierarchical and egalitarian, between modes of behaviour that feature or privilege the group to these that accentuate individuals. Permanent social inequality is almost universally seen as a major attribute of social complexity among humans. This need not mean, however, that egalitarian societies must necessarily be defined as ‘simple’ and hierarchical societies as so-called ‘complex’.

The history of complexity in archaeology 27 One way to resolve this tension field can be found in the concept of heterarchy, introduced in archaeology by Carole Crumley (1995), to provide an alternative heuristic device for exploring the multiple sources of power inherent in many societies. The concept leaves more room for alternative forms of power (not only socio-political or economic), such as religious, labour or client power. A society never fully adheres to either hierarchical or heterarchical modes of organisation. Instead, structures move along a continuum between heterarchy and hierarchy as context and values change. Additionally, both modes are structural basins in a wide continuum of complementary mechanisms across different scales. For example, hierarchically organised communities might be interconnected in a heterarchical regional network (so-called peer polities), which can in turn again be integrated in hierarchical imperial structures. The work on state formation is extensive and doing justice to the full scholarship on the topic cannot be accomplished within the scope of this part. This book is not specifically about states but rather about how these social entities relate to the emergence and development of social complexity. In the next chapter, I will readdress state formation as part of a wider discussion on polity formation from a social complexity perspective. Before addressing complex systems themselves, however, it is essential to sketch one of the main precursors to complexity science. I will focus particularly on systems thinking and its conceptualisation of complexity, both as a general discipline and as applied in archaeology.

Systems thinking and complexity Over the course of almost a century, the principles and aims of systems thinking have permeated through a wide range of disciplines including: planning and evaluation, education, business management, sociology, psychology, cognitive science, sustainability, environmental sciences, ecology, biology, physics, earth sciences, and historical sciences (Cabrera et al., 2008). Its fundamental concepts were developed in the early parts of the 20th century in disciplines such as biology, ecology, psychology, and cybernetics. Systems thinking covers a wide variety of approaches, methods, and theories used for conceptual thinking about all aspects of the genesis, dynamics, and operationality of various kinds of systems. As systems thinking was applied in an increasingly broad variety of scientific disciplines, definitions of what exactly constitutes a ‘system’ proliferated as well. The American philosopher of science Merrilee Salmon even argued that: “there is not one sense of ‘system’ that can be captured by a single definition, or because any definition broad enough to cover all the legitimate uses of ‘system’ would be so vague that anything at all would count as a system” (Salmon, 1978, p. 176)

28  The history of complexity in archaeology At its broadest, a system can be defined as “any set of things and the relationships between them” (Marchal, 1975). While technically correct, this definition provides little practical use. Systems, in this sense, include anything “from card catalogues to airplanes to economies” (Klir, 2001). Extending this minimal definition, a system can better be considered as a set of things, interconnected in such a way that they produce their own pattern of behaviour over time (Meadows 2008). The use of systems thinking as a structuring framework in scientific thought on the ontological level, rather than a fixed set of directives, has allowed for some degree of ambiguity resulting in myriad of systems-based models (Cabrera et al., 2008). Many of these models highlight different approaches towards problem solving in systems thinking and often appear more or less complementary or even interchangeable to the external observer. Nevertheless, amongst the multitude of different approaches, some general similarities characteristic for the ‘systems approach’ can be discerned. According to Mingers and White (2010) this entails: (1) viewing the situation holistically, as opposed to a reductionist view, as a set of diverse interacting elements within an environment; (2) recognising that the relationships or interactions between elements are equally or more important than the elements themselves in determining the behaviour of the system; (3) recognising a hierarchy of system levels and the consequent ideas of properties emerging at different levels, and mutual causality both within and between levels; and (4) accepting, especially in social systems, that people will act in accordance with differing purposes or rationalities. These four aspects are essential to systems thinking as they involve the essence in the way we conceptualise and try to understand the world around us. Social systems thinking Social systems refer to a particular type of system, more specifically “those comprised of humans, their various aggregate creations (groups, formal organisations, economies, social institutions, etc.), and the relationships amongst them” (Castellani and Hafferty, 2009, p. 58). Another definition describes a social system as “any group of people who interact long enough to create a shared set of understandings, norms, or routines to integrate action, and established patterns of dominance and resource allocation” (Westley et al., 2002, p. 107). The aspect of mutual interaction, as we will see, will prove crucial in the conceptualisation of social systems as complex systems later on. One of the first to rigorously apply systems thinking to human societies was the sociologist Talcott Parsons (1977), who defined a system as a stable set of interdependent phenomena delineated by analytically established boundaries, which relates to an ever-changing external environment. A social system is defined accordingly as a system of social

The history of complexity in archaeology 29 interactions between reciprocally oriented actors. It consists of roles, collectives, norms and values. According to Parsons, the social system operated as an integrative pattern within a larger level of the general action system. Parsons’ model of social systems can be considered an equilibrium model as tendencies towards deviance are met by the system through feedback of control mechanisms aimed at maintaining its current structures. As argued by Spencer-Brown (1972), all systems are the result of an act of ‘distinction’ consisting of establishing a border between the internal system dynamics and its external environment. The act of drawing a boundary, distinguishing between system and environment, is central to the very definition of a system. This is also the core element of Maturana and Varela’s (1980) concept of autopoiesis. Originally developed in biology, the concept of autopoiesis can foremost be considered an ontological property of systems analysis and has its intellectual roots in cybernetics (Padgett and Powell, 2012, p. 55). It was applied to social systems by the German sociologist Niklas Luhmann, who stated that an autopoietic approach to social systems requires a single operation that possesses connectivity as a starting point for self-reproduction of the system (Luhmann, 1995). In this view, social systems are essentially systems of communication, consisting of coupled components with operational, but not energetic closure (Haynes, 2017, pp. 10–12). The crucial elements in Luhmann’s social systems are not people, but rather the connective processes of communication that are used by the system to create its own structures through a self-referential organisational distinction from its environment. Systems draw on a sense of operational closure – that is wholly containing the own operations within these boundaries – to define themselves. The role of social systems is then to provide order by managing the complexity of the environment. Parsons’ action systems and Luhmann’s systems theory have often been criticised for losing the individual in their rigid and deterministic system frameworks (see for example Mills 1959, pp. 25–49). Social systems can generally be considered as (1) indecomposable totalities (holism); (2) aggregates of autonomous individuals (individualism); or (3) systems of interrelated individuals (systemism) (Bunge, 1999, p. 4). Most systems thinkers combine to some extent a number of elements of these three different extremes, however, many critiques on social systems thinking stems from the fallacy of mistaking systemism for holism. Rather than rejecting the system concept altogether for its supposed determinism, we should look to properly reintegrate the individual in a systems-based perspective. One of the first to react against the static conceptualisation of systems by Parsons was the sociologist Walter Buckley, who argued for a more dynamic type of systems, drawing from Bertalanffy’s general systems theory. Buckley (1967) argued for a ‘process model’, as derived from the predominant view of American sociology in the early 20th century, led by the Chicago school with key

30  The history of complexity in archaeology authors such as Albion Small, George Herbert Mead, Robert Park, and Ernest Burgess. The process model centred on a series of events inducing processes of system maintenance and change. As a result, social systems carry an inherent tendency to undergo continuous structural elaboration, conceptualised through the process of ‘morphogenesis’. Social systems thus came to be considered as a complex, multifaceted, fluid interplay of widely varying degrees and intensities of association and dissociation (Buckley, 1967, p. 18). Structural elaboration occurs as participants learn more effective ways of relating to one another and adapting to their environment (Johnson 2008, p. 472). Due to this tendency towards structural elaboration, social systems inherently evolve towards a more complex state. I have already discussed some of the problems associated with one-way conceptualisations of increasing social complexity in the first part of this chapter. In the next part, I will show how complex systems thinking can be used as an alternative framework for system development. Social systems in archaeology Systems thinking was most famously applied in archaeology by the British archaeologist David Clarke. His seminal book Analytical Archaeology (1968) was one of the first and most extensive attempts to integrate archaeological practice and analysis into a consistent system-based theoretical framework. Clarke defined a system as “any intercommunicating network of attributes or entities forming a complex whole” (Clarke, 1968, p. 43). This means that a system is basically considered an ensemble of attributes. The emphasis on attributes, most notably in material objects, allows the identification of essential elements by differentiating ‘meaning’ from ‘noise’. Three different types of attributes are distinguished: (1) inessential attributes which are not relevant to the study at hand and which consequently do not figure in the system as defined; (2) essential attributes, which are those essential variables that are part of the system and whose values may change as part of system changes; and (3) key attributes, those essential attributes in the system whose successive transformation values are co-varying in some specific relationship with successive values of other similar attributes. The identification of relevant attributes depends on the system in question. Focusing on material culture, Clarke distinguished different hierarchical systemic levels building on the level of the attribute. These are: artefact, artefact-type, assemblage, culture, and culture group (Clarke, 1968, p. 21). Depending on the scale of analysis, archaeological systems can be materialised as a system of attributes within a population of artefacts, a system of artefacts within a changing cultural assemblage, or a system of social attributes within a changing society, and so on. None of these categories are

The history of complexity in archaeology 31 ever fixed as changes can be observed in the archaeological record for each one of them. Clarke identified a number of primary mechanisms of change, including random variation, multi-linear development, invention, diffusion, and cultural selection. This raises the question of how the archaeologist can know whether observed differences, for example changes in a specific type of pottery, are meaningful or merely the result of unintended and partially random variation? The answer lies in the identification of repetition in material attributes, suggesting the endowment of meaning. Repeated similarities or regularities are considered systematically correlated attributes that give recognizable identity to objects. It must be noted however, that similarities, for example typological regularities, may not necessarily be simple ‘one-to-one’ regularities. Instead, a polythetic attribute system is required to define membership of a given type. These regularities are then considered to follow from certain limiting conditions or constraints imposed by physical or social action. The importance of these constraints lies in the predictive information we can get from them as they generate a greater degree of regularity. Clarke’s definition of a system emphasises the transfer of information among system components, which occurs whenever constraints restrict the variety of outcomes in the system. As in many other approaches in systems thinking, also outside of archaeology, this definition builds on the tenets of cybernetics and information theory, where ‘messages’ constitute an ordered selection from an agreed set of selected variety, which must be differentiated from disturbances that do not represent any part of the essential message and are therefore termed ‘noise’ (Ashby 1956, 121–160; Cherry 1957, pp. 303–307). The role of positive feedback mechanisms, amplifying small deviations into large differences, operating on such information transfers has been especially highlighted as a factor of system development in the past (Flannery, 1968). From this perspective, the sociocultural system at large can be considered an elaborate behavioural information system. Within the overall system, a number of subsystems can be discerned: social, religious, psychological, economic, and material. The social subsystem, for example, is generated by a set of acquired ideas or information, which continuously reproduces itself by conscious and subconscious imprint (Clarke, 1968, p. 105). The close similarities to Bourdieu’s concept of habitus are apparent, but interestingly, the self-replicating aspect of the social subsystem is rather reminiscent of the previously discussed concept of autopoiesis. Material culture is considered an information subsystem in its own right, consisting of patterned constellations of artefacts which outline the behavioural patterns of a sociocultural system and embody that system’s technology (Clarke, 1968, p. 129). However, Clarke also explicitly notes that every subdivision of an overall sociocultural system into component subsystems is merely an arbitrary conceptualisation of different aspects of the same network. It can therefore be surmised that

32  The history of complexity in archaeology the same set of general postulates may be relevant in each of these arbitrary subsystems within the same system and therefore display the same set of inherent qualities. Archaeology as a discipline can therefore be expected to trace the same inherent ‘behaviour’ determined by the overall sociocultural system. Clarke drew extensively from Cherry’s (1957) theory of signs in considering material culture as flows of information, using: (1) designata, a set of roles or activities that a particular artefact was intended for; (2) percepta, information conveyed to an observer in the act of perceiving an object or artefact; and (3) concepta, information contained in the abstract idea of an object or artefact conceived in the brain of a person or potential artificer, recalled from memory. Clarke’s model of cultural systems remains a prime example of a rigorously defined and internally consistent model in archaeology. His contribution to the theoretical development of the archaeological discipline can hardly be overstated. In the words of Plog (1975, p. 210) “[Clarke’s Analytical Archaeology is] by far the most complex and thorough effort to apply general systems theory to archaeology”. Yet, his work has also received criticism over the years. As these critiques also take direct aim at the general application of systems theory in archaeology, it is worthwhile to repeat some of it here. Most notably, Salmon (1978) discarded the applications of systems theory in archaeology. While not dismissing the potential of a systems approach altogether, she found the definition of archaeological systems to be insufficiently clear and the applications by archaeologists such as Flannery and Clarke to have limited utility. At most, she claimed, the archaeologist could adopt some useful concepts such as ‘feedback’ and ‘equilibrium’, and even these, according to Salmon, were originally derived from the fields of physics and engineering, rather than systems theory. Others have also criticised the systems approach for being deterministic in nature. It was argued that human beings did not inhabit a specific sphere within an objective and external environment, but instead lived in an embedded world that they are able to understand and act in culturally (Ingold, 2000). These reproaches explicitly rejected division of the world in separate spheres such as economic, social, ritual, etc. (Robb and Pauketat 2012, p. 10). It must be emphasised however, that Clarke himself already noted that every subdivision of the overall sociocultural system into component subsystems is merely an arbitrary conceptualisation of different aspects of the same network. The use of systems theory in archaeology tended to diffuse emphasis upon single causes and develop explanations that encompassed continuity, gradual change, and sudden transformation (Robb and Pauketat 2012, pp. 9–10). Still, a main point of critique addressed the emphasis on studying the system through its subsystems. The criticism on the limitations of such reductionist approaches was most succinctly formulated by American physicist

The history of complexity in archaeology 33 Philip Anderson, Nobel Prize laureate in 1977, in his seminal paper More is Different, he writes: “The ability to reduce everything to simple fundamental laws does not imply the ability to start from those laws and reconstruct the universe… The behaviour of large and complex aggregates of elementary particles, it turns out, is not to be understood in terms of a simple extrapolation of the properties of a few particles. Instead, at each level of complexity entirely new properties appear, and the understanding of new behaviours requires research which I think is as fundamental in its nature as any other.” (Anderson, 1972, p. 393) The challenge ushered by Anderson, did not go unnoticed and it was increasingly realised how complex interconnections and interactions between system components could give rise to wholly new and unexpected behaviour, so-called emergent phenomena. It is exactly these non-linear interactions between subsystem components and the emergent phenomena they produce that are the subject of the field of complex systems thinking. This rich and interdisciplinary research field is increasingly finding its way into a wide range of disciplines, including archaeology. In the next part, I will first briefly discuss the history of complex systems science and its general properties, before finishing this chapter with a series of recent applications of complex systems approaches in archaeology.

Complex systems thinking Complexity science originated during the 1970s and 1980s out of developments in various disciplines – including biology, chemistry, physics, mathematics, general systems theory and cybernetics. Traditionally, these fields work by assuming stability, equilibrium, linear change, cyclicality, and robustness in systems generating simple behaviour. Yet, for many real-life systems such as organisms, cities, the environment, and societies, these core axioms do not hold. A new paradigm was needed to adequately address the complex properties and behaviours of such systems. As we saw previously, a system can be generally defined as a ‘complex whole of related parts’ (Cabrera et al., 2008). The idea behind describing a system as complex (as in exhibiting ‘complex’ behaviour) is that of the system as more than the mere sum of its constituent parts. One of the most crucial elements of a complex system is its display of emergent properties. These properties arise when aggregates of identical elements obtain new, emergent characteristics which are not directly derived from the summation of existing characteristics of individual constituent elements (Holland, 2014, p. 4). A classic example is that of H2O molecules obtaining the emergent characteristic of ‘wetness’ not found in the individual molecules themselves, whereas properties such as weight are always

34  The history of complexity in archaeology direct aggregates coming from the summation of constituent elements (Ball, 2004). Complex systems thinking has gained enormous momentum ever since the foundation of the Santa Fe Institute (SFI) in 1984.6 The central theoretical nexus of SFI is that of complexity as a property of ‘complex adaptive systems’ (Gell-Mann, 1994; Holland, 1995; Kauffman, 1993), with a strong methodological focus on mathematics and computational tools in formal modelling approaches. SFI prides itself on combining critical and creative thinking, using rigorous mathematical frameworks to evaluate and test hypotheses, while at the same time encouraging scholars to leave the well-trodden paths of conventional science and venture into the unknown or unexpected (Gumerman and Gell-mann, 1994, p. 5). Through this combination, many scholars affiliated with SFI have used complexity science to markedly impact the course of science over the last few decades. Properties of complex systems A complex system can be generally conceptualised as “a system whose properties are not fully explained by an understanding of its component parts” (Lewin, 1992, p. x). It is “a system in which large networks of components with no central control and simple rules of operation give rise to complex collective behaviour, sophisticated information processing, and adaptation via learning or evolution” (Mitchell, 2009, p. 13). A representation of the overall structure of a complex system and its main dynamics can be found in Figure 2.1. The basis of any complex system consists of a multitude of constituent elements that interact, often in relatively simple ways (Holland, 2014; Mitchell, 2009). These systems do not only consist of many elements. In order for them to be complex, these elements also need to be mutually interacting. This does not necessarily mean that every element communicates directly with all others in a synchronous fashion. Rather, the interactions follow network patterns with highly similar properties (Haynes, 2017, p. 14). Out of these interactive dynamics, emergent properties arise that are not directly derived from the summation of existing characteristics of the constituent elements (Holland, 1998). In this sense, complex systems thinking offers precisely the kind of solutions needed to go beyond the limits of reductionism impinged on traditional systems thinking (Lewin, 1992, pp. ix–x). Examples of complex systems include ecological systems such as forests, biological systems such as ant colonies or the human brain, artificial systems such as the World Wide Web, or social systems such as cities, economies, communities and societies. Despite all research efforts (or perhaps precisely because of it), the very concept of complexity has proven to be difficult to pin down. Complexity is often used as a descriptive term, whose origins and development can remain something of a black box. It has been stated that “one of the hurdles

The history of complexity in archaeology 35

Figure 2.1  Structure of a complex system Source:  Wikimedia commons, image used under CC licence; Credit to creator Acadac.

in defining a theory of complexity, and with it, developing a fundamental, helpful approach is that there is no uniformity in the meaning of complexity” (Sitte, 2009, p. 25). The term can, for example, refer to an aspect or subpart of a system, as well as the magnitude and variety of the overall system. We have to be aware that different dimensions of complexity can exist, sometimes simultaneously within the same system. These can be considered different aspects or manifestations of complexity, but none of them is complexity per se. Correspondingly, it has been stated that “there is no single quantitative measure of complexity, but various aspects of it can be defined and measured” (Mayfield, 2016, p. 58). A number of key properties are commonly stressed: non-deterministic and non-linear behaviour, emergence, self-organisation, adaptation, non-equilibrium, information processing, and interdependencies among system components (Holland, 2014; Mitchell, 2009; Thurner et al., 2018). Approaching human societies through a complex systems lens entails a clear focus on information flows, decisionmaking, interactions at multiple scales of organisation, feedback mechanisms, and non-linear dynamics in which individual agency generates system-level emergent phenomena (Byrne, 1998; Miller and Page, 2007). I already noted how information processing is considered a central element in systems thinking. Likewise, a key emergent property of complex systems is their capacity for computation and transmission of information among its components and its environment. The transmission of information, both intra-system (between system components) and inter-system

36  The history of complexity in archaeology (between the system and its environment) makes complex systems adaptable. Two-way feedback mechanisms operate between simple constituent behaviour and emergent collective behaviour. On the one hand, positive feedback mechanisms occur when change in one direction makes the system even more prone to keep changing in that same direction throughout successive system states, whereas negative feedback consists of counterbalanced change continually guiding the system towards the current equilibrium (Bentley and Maschner, 2007, p. 245). The ways a complex system develops depend very much on the connections between system components, the information that is exchanged and the combination of positive and negative feedback dynamics. Due to their adaptive nature, I consider complex systems and complex adaptive systems (CAS) here to be largely synonymous. CAS can be defined as large networks of interacting components with no central control and simple rules of operation, exhibiting dynamic emergent behaviour that cannot be reduced to the sum of its individual parts, sophisticated information processing, and is responsive to its environment via learning or evolution (Holland, 1995). System components receive information regarding the environment and alter their behaviour in response to that information, again transmitting information on their current state to other components. The main difference with general systems thinking is that, in terms of structure and agency, CAS components are not merely structuring parts of the overall system, but are also capable of exerting their own agency (Barton, 2014). System development then results out of both internal and external challenges and stimuli. Referring back to the structure of Figure 2.1, human societies and its organisational structures emerge from a multitude of social interactions between individual people that are structured through social practices performed at a given time and place, giving rise to collective patterns of behaviour in the form of institutions and social organisation. This overall social organisation in turn exerts feedback mechanisms back onto the behaviour of constituent agents. In CAS, agents’ actions and choices are fundamental as the outcomes of individual decisions are not merely averaged away within the overall workings of the system but may be magnified as a result of other decisions and therefore decide the direction the system takes (Bogucki, 2003, p. 98). For example, it has been argued by Arthur (1989) that in knowledge-based economic systems, decisions for initial investment in knowledge acquisition are rewarded with rapid accumulation of experience, stimulating functionality and efficiency of economic processes. Agents adapt their behaviour to the information received from higher levels of the system and their environment, including other agents and their behaviour. As a result of these feedback mechanisms acting upon social interactions, system behaviour evolves over time. It are these changing dynamics over long-time perspectives that are the subject of complex systems applications in archaeology.

The history of complexity in archaeology 37 Complexity is often conflated with the related but distinct concept of chaos. A chaotic system is characterised by infinitesimal changes in initial conditions leading to wholly different system dynamics and outcomes. This has been metaphorically described by Edward Lorenz (1963) as ‘the butterfly effect’, where at one point the flaps of the wings of a butterfly on one side of the world may change air currents until it ‘causes’ a hurricane across the other side of the world, whereas with the next flap nothing happens. This uncertainty has profound consequences for the ways such systems are conceptualised and studied as the unpredictability of non-linearity magnifies any initial uncertainties in the system state. Additionally, the character of the interactions among subsystem components is generally far more decisive for the overall system behaviour rather than their inherent characteristics (Barton, 2014). Chaotic systems have no fixed equilibrium state but are rather characterised by multiple equilibria between which it can oscillate. Somewhat unintuitively, chaotic systems are entirely deterministic – as in governed by a fixed set of rules and undisturbed by external noise – even though they are inherently unpredictable because of the impossibility of defining initial conditions (Bentley, 2003, p. 4). For a chaotic system, knowledge of past states is meaningless given that it has no bearing on the subsequent system states, even if we know the laws regulating changes in the system. Complex systems, by contrast, are neither static nor chaotic but have dynamic structures embodied in the patterns of interactions between components (Cilliers, 2001, p. 140). Complexity lies in the intermediate zone between static order and unstable chaos. This zone has been famously called the ‘edge of chaos’ (Kauffman, 1993; Packard, 1988). These intermediate configurations have been described as critical states or ‘self-organised criticalities’ (Bak et al., 1987; Bak 1996). A self-organised criticality can be defined as a quasi-steady system state in which small inputs accumulate until a sudden collapse occurs at random intervals and intensities (Bak, 1996). As systems move periodically into critical states, change becomes episodic, and sometimes dramatic (Boulton et al., 2015). Dynamical selforganising systems are said to constantly reorganise, releasing stress/tension to evolve towards this critical steady-state. This general system property also suggests that rapid and large-scale system changes need not necessarily be connected to major causal factors but can also result from an accumulation of bottom-up, small changes and stimuli passing a threshold value and resulting in a system-wide ‘tipping point’ (Gladwell, 2000). These dynamics are inherently non-linear, meaning that the scale of system-level changes is not necessarily proportional to the scale of the phenomena that trigger it. Complex systems can sometimes absorb a large amount of perturbations without any meaningful or notable changes in system configurations, whereas at other times, comparatively small perturbations could result in a cascade of changes with huge consequences for the overall system state and further system dynamics.

38  The history of complexity in archaeology The unpredictability of emergent phenomena developing out of constituent interactions has strong implications for the ways complex systems develop, especially on large temporal scales. It has been noted that in applying complexity discourses over deep-time frames, the role of disjunction, that is bifurcations into multiple potential pathways for a given system, is essential (Bintliff 2003, p. 81). These bifurcation points tend to converge into a limited number of recurrent system states with similar properties, so-called attractor states. Interaction between system components at the moment of bifurcation can generate transformative behaviour, inducing self-reinforcing mechanisms and dynamics steering the system into a given direction (McGlade and van der Leeuw 1997, p. 12). This raises the question to what extent contingency lies at the basis of change in complex systems, especially when applied to human societies (Bintliff, 1999, 1997). This is not to say that complex systems are unknowable and system change happens randomly. Whereas the trajectory of a complex system is difficult to predict, its past phase transitions can be understood. Complex systems approaches therefore need to combine an understanding of system changes and phase transitions, as well as the emergence of order and structure which keep the system in a given basin of attraction. It should be noted that “what is predictable, and what contingent, often depends upon the scale under consideration” (Gould, 1999, p. xvi). Contingency in this sense relates to the theory of punctuated equilibrium, which particularly deals with the pace of change in a given system, balancing long-stretched periods of stability with rapid transformation phases (Gould, 2007). Such phases of discontinuity can be considered “thresholds of change where the role of human agency and/or idiosyncratic behaviours assumes paramount significance in the production and reproduction of societal structures” (McGlade and van der Leeuw 1997, p. 11). To understand such episodes of change induced by tipping points and other threshold effects, multi-scalar perspectives are needed. One of the major aims of complexity systems science is to discover how dynamics at a small-scale translate into emergent phenomena at a larger scale, or what emergent properties can be expected. As multitudes of components in complex systems interact in heterogeneous ways, with variable frequency and intensity, the scale and direction of change in the higher-level structure of organisation within the system need not necessarily be proportional to the scale and direction of the constituent phenomena that trigger it (Barton, 2014). In other words, connections between constituent interactions and emergent system behaviour are non-linear. Moreover, dynamics in complex systems are driven by feedback loops, resulting in the effects of system dynamics to become part of the causes. This stands in stark contrast with linear explanations of system dynamics, which need cause and effect to be clearly separated. As a result, no linear correlation can be drawn between the size or intensity of a system input and the corresponding system outputs, meaning that even small perturbations could result in major

The history of complexity in archaeology 39 system-wide consequences (Bak, 1996; Ball, 2004, p. 227). The property of small changes in system inputs radically altering system output has been described as sensitivity to initial conditions. The components of CAS are organised into nested groups that can be represented as structured networks or organisational hierarchies. As the system grows more complex, a deeper nested structure emerges (Barton, 2014). Because of the bottom-up genesis of nested, multi-scalar structures in complex systems from the foundational interactions of its constituent components, interactions among components within one of the nested groups tend to be more frequent and stronger than interactions between groups at any level in the nested hierarchy (Barton, 2014). In a biological system, for example, cells of the heart interact more directly with other heart cells than they do with lung cells, even though both are part of the cardiovascular system. Likewise, households within a specific community share stronger social ties and interact more frequently with other households within the community than with those of another community, even if both are part of the same polity. Units that make up the nested hierarchies of CAS, can often continue to function even when links to other units are broken. For example, if a larger polity, say the Roman empire, were to disintegrate and lose its ties of connectivity holding communities together, social life within these communities may continue unaffected on several levels. This property of stronger connectivity between units of similar types and the ability of units to operate semi-independently of other units in a nested hierarchy is a property of CAS called ‘near decomposability’ (Auban et al., 2013; Simon, 1962; Wu, 2013). One of the key properties of information transmission in complex, multiscale systems is uncertainty reduction, observed particularly in biological systems driven by deep-time, evolutionary dynamics (Flack et al., 2013). Lower-level units in hierarchical systems reduce uncertainty by coarse-graining environmental information, meaning they compress complex behaviour into simplified representations to distil regularities. This coarse-grained information is then used as a predictor for future system states and necessary course of action. Over time, coarse-grained variables used as predictors for system behaviour converge as they build up accuracy and new levels of organisation consolidate in nested dynamical processes (Flack, 2017a). The result is a recurrent micro-macro transmission of information, which could effectively be considered a form of (locally optimised) computation (Flack, 2019). Through recurrent iteration, this micro-macro transition becomes structured, giving rise to a regulatory structure for higher-level units to direct lower-level constructive processes and reduce variance on the component level (Smith and Morowitz 2019, p. 126). The reciprocity between upward construction and downward regulation can be considered a “feed-down” relation or downward causation. These topdown relationships allow higher hierarchical levels to create order on the

40  The history of complexity in archaeology lower scales without directed, centralised intervention. The result is a so-called ‘information hierarchy’ (Flack, 2017b, 2017a; Flack et al., 2013). These information hierarchies entail the emergence of new functionalities on higher levels that can influence behaviour on lower levels using coarsegrained information generated by uncertainty reduction in information transmission. In the next chapters, I will develop this approach of uncertainty reduction in information transmission for human societies as a key property of social complexity trajectories. Complex systems thinking in archaeology So far, I have traced the major outlines of the historical usage of social complexity and social systems in archaeology. In the previous part, I also outlined the main properties of complexity theory. Proponents of complexity science often stress its value in providing general tools to describe and understand complex systems. Its power lies in providing scaffolding for interdisciplinary research by combining theories, methods, and tools from various fields and integrating them in coherent scientific frameworks, which are not necessarily discipline-specific, but rather question-specific. Complex systems thinking seeks to transcend traditional disciplinary boundaries and instead focuses on topics and questions of interest. In recent years, archaeologists have increasingly turned towards complex systems thinking as a conceptual and methodological framework to study long-term trajectories of social complexity. In the remainder of this chapter, I will discuss some of the major areas of application for complex systems thinking in archaeological research, moving from more general conceptualisations to specific fields of applications. I will not attempt to provide an exhaustive overview, but will highlight a number of examples that cover large parts of the spectrum of approaches that have been applied to archaeology. Specifically, I will focus on four main fields: (1) network science; (2) settlement scaling; (3) cultural evolution; and (4) agent-based modelling. Network science Complex systems can be defined as large networks of interacting components producing emergent collective behaviour. Networks are a crucial aspect of complex systems studies, and network science is one of its major tools (Thurner et al., 2018). Examples of networks include the internet, power grids, friendship networks, citation networks, ecological food webs, trade networks, neural networks, etc. The term ‘network’ is used ubiquitously, often serving as nothing more than a shorthand for connectivity. This colloquial usage of the term, however, is not necessarily linked to network science. This field studies entities (nodes, vertices or actors) and the connections (links, edges or ties) between them in a formal and mathematical

The history of complexity in archaeology 41 way (Newman, 2010). It describes and explains how things are connected and interact through the properties of these links. It is only by understanding these properties that the emergent behaviour of the network as a whole can be understood. Properties of links include whether they are undirected (symmetric) or directed (asymmetric), the strength of the ties, and whether they consist of positive or negative relationships. Over the last decades, scholars from a wide variety of fields have developed an extensive array of mathematical, computational, and statistical tools for analysing, modelling, and understanding networks. To provide an encompassing overview of these tools would be beyond the scope of this book.7 Instead, I will provide some general outlines of basic network measures, followed by some more advanced applications in archaeological networks. One important network property is that of centrality, indicating how important any given node is within the network. Various measures can be used to study centrality. For example, for every node, the number of connections (degree) with other nodes can be counted. Nodes with much larger degrees than others are called hubs and often hold a central position within the network. One example could be trade links between settlements, with outgoing links for export and incoming ones for import. Hubs in the trade network would be those settlements with a high number of trade links, for example an important port site. The full distribution of node degrees in a network is called a degree distribution and contains crucial information regarding the overall structure of the network. A sequence of edges leading from one node to any other in the network is called a path. Average path lengths in the network can indicate the speed at which information, diseases, cultural ideas, etc. can travel through the network. Some networks contain groups of nodes that are strongly interconnected with only few inter-group links, for example friendship networks consisting of smaller cliques of friends inter-connected through links between one or few nodes. In network terms, these groups are called clusters or communities. Several techniques exist to identify and evaluate clusters in a network. One important approach focuses on modularity as a measure of cluster quality to identify communities by progressively removing edges from the original network (Girvan and Newman, 2002). The outcome of the analysis is a dendrogram that allows (visual) inspection of where and when different communities emerge from within the network. Network science does not only encompass these analytical tools, but also includes network modelling, using complex network models such as gravity models to identify, simulate, explore and explain properties of real-world networks. Two network models that have been particularly influential for archaeologists will be discussed here: the small-world model and the scale-free model. Many real-world networks display the famous small-world effect, in popular culture referred to as “six degrees

42  The history of complexity in archaeology

Figure 2.2  A small world network (made by the author)

of separation” (Watts and Strogatz, 1998). It was noted that real-world networks often are neither completely ordered nor completely random, but rather exhibit important properties of both. Small-world networks form an intermediate type of network in-between regular and random networks where networks are highly clustered while the average path length is typically relatively low and increases only as the logarithm of the number of vertices in the network (Figure 2.2) (Newman, 2010, p. 10). The implication is that every node is only a few steps removed from any other node in the network, meaning that information can quickly pervade the network. The second model considers so-called scale-free networks (Barabási and Albert, 1999). It was discovered that when one plots the degree distribution of large, real-world networks, these are often not normally distributed, but are rather highly skewed and follow a long-tailed or power law distribution.8 This means that most nodes in the network typically have less than the average amount of links, whereas some (the hubs) are much more strongly connected. One of the underlying mechanisms of power law structures is preferential attachment in network growth (Figure 2.3). When a new node enters the network, it gets attached to existing nodes with a probability based on its number of links, that is, a well-connected node is more likely to gain new

The history of complexity in archaeology 43

Figure 2.3  Preferential attachment structure (made by author)

connections. The idea of preferential attachment is derived from the “richget-richer effect” observed in personal wealth distributions. The economist Herbert Simon (1955) was the first to mathematically show that this effect gives rise to power-law distributions. His ideas were then adopted by British physicist Derek Price to explain cumulative advantages in citation networks of scientific papers (Price, 1976). It was only with Barabási and Albert’s (1999) seminal paper that the term preferential attachment was coined and gained widespread recognition as an important generative mechanism for scale-free networks. Let us now turn to some examples of network science applications in archaeology. In recent years, a number of excellent reviews of network science in archaeology have been published.9 I will here only provide some main outlines and refer the reader to these other publications for more in-depth overviews and discussions. Ever since the 1960s and 1970s, precursors to network science such as graph visualisation and analysis were used in archaeology, for example for exploring spatial relationships (Terrell, 1976), seriation (Kendall, 1969), trade networks (Pitts, 1965), and settlement interaction systems (Irwin, 1978). One of the earliest examples of using network science specifically for archaeological research questions entailed the use of Proximal Point Analysis to model geographic networks and inter-island movement in the

44  The history of complexity in archaeology Solomon Islands (Terrell, 1977). Despite these early applications, network science never really got an extensive foothold in archaeology over the following decades (Brughmans and Peeples, 2017, p. 4). It is only in recent years that archaeological network analysis, particularly social network analysis (SNA), started to take off. SNA applications focus on the interdependence of actors and their actions, links between actors as channels for flows of resources, networks as structural environment providing opportunities or constraints on individual action, and network structures as lasting patterns of relations among actors (Wasserman and Faust, 1994). SNA is particularly suited to study diffusion of (material and immaterial) resources, mostly in but not limited to, socio-economic systems such as networks of pottery production and distribution (Brughmans and Poblome, 2012) or settlement networks and itineraries (Graham, 2006a). Fiona Coward combined SNA with small-world networks – albeit more on a descriptive than analytical level – to discuss the linkages and interactions between social groups in the Epipalaeolithic and Early Neolithic Near East (Coward, 2010). She used these descriptions to argue for the existence of tightly knit kin-based and proximity-based groups that were interconnected over a larger geographical area through a series of weak ties. A collection of papers in an edited volume by Alex Bentley and Herbert Maschner (2003) focuses specifically on the usage of scale-free networks in archaeology. They argued in particular for the utility of scale-free network growth to study social inequality in prehistoric societies. In one of the papers, it was suggested that house size variability in the North Pacific can be explained by group size and status of the chieftains, and therefore reflect the emergence of social inequality through the rich-get-richer effect (Maschner and Bentley, 2003, p. 57). A combination of small-world and scale-free network properties was used by Shawn Graham to study social networks of individuals associated with the Roman brick industry in the Tiber valley (Graham, 2008, 2006b). He developed a multi-scalar network model of individuals involved in patronage and manufacturing networks, switching between local and global scales. Graham concluded that these networks exhibited both egalitarian (nodes having similar degrees) and hierarchic (power-law degree distribution or scale-free network) small-world structures at different points in time. Another case study combining small-world and scale-free networks assessed artefact-type distributions in South Scandinavia during the Early Viking Age based on attested type co-presences in settlements as proxy for interaction networks (Sindbæk, 2007). It was suggested that site hierarchies could explain the differences in relative volume of imported goods and raw materials. This research indicated the existence of a small number of hubs displaying a markedly higher degree, as well as a high level of geographical clustering, suggesting a small-world structure. However, Sindbæk (2007, p.  70) himself already raised an important issue with these results,

The history of complexity in archaeology 45 pertaining to the problem of assigning behaviour from dynamic models to network structures identified in the archaeological record. A next set of examples to highlight the use of network modelling are spatial interaction models. The central focus of these models is to explain flows between different sets of located entities. The so-called ‘gravity model’, has been particularly influential. This model was originally developed in the field of geography to give a model-based account of the evolution of shopping centres in Leeds (Wilson et al., 1977). It was then transposed from geography to archaeology by Rihll and Wilson (1987) to determine how the relationships between sites are conditioned by geographical space and to what extent their interactions transcended geographical constraints. The case study of Rihll and Wilson focused on the origin of polis in mainland Greece during the ninth and eighth centuries BCE. To simplify things, all settlements were a priori considered to have held the same importance, disregarding all power structures and hierarchies. Simply put, the model simulates a measure of influence of every settlement over its surrounding area. On the one hand, the approach allows us to bridge current gaps in our knowledge of a large amount of settlements. The authors themselves indicated that the main goal of this paper was “to fill in gaps in data or knowledge through a capacity for inference based on hypothesised spatial relationships which are thought to have existed” (Rihll and Wilson, 1987, pp. 11–12). Many of the sites in this particular dataset, for example, have never been excavated. Our knowledge of their emergence and development is therefore rather limited. On the other hand, the model does not take into account any previously existing information as priors to inform further parameterisation. This information was only used for subsequent calibration and validation of the model. Using this model, the authors tried to ‘predict’ which poleis would have been most important and held the strongest expected pulling power for flows in the network based on site location. Comparison of their results with available data showed a reasonable fit, correctly identifying important sites such as Athens, Corinth, Argos, and Thebes to be in the top 10 ranked sites across different settings for degree of centralisation in network flows. At the same time, sites such as Kromna, Merenda, and Kalyvia in the Mesogeia plain of Attica were flagged as prominent settlements as well. However, very little research has been done at these sites and our information on their role and importance in Antiquity is limited. This type of modelling could potentially be used to guide future research efforts at these sites. A more recent example of the use of spatial interaction models is a study of maritime interaction during the Aegean Bronze Age by a collaboration between the archaeologist Carl Knappett and theoretical physicists Tim Evans and Ray Rivers (Knappett et al., 2008). This study took the gravity model as a basis for analysis, but extended it in order to better account for prior information on site importance. They identified 39 sites, represented

46  The history of complexity in archaeology as vertices that were assigned a fixed carrying capacity reflecting local resource availability and a variable to indicate a site’s relative importance, thus acting as a “centre of mass” for the surrounding areas. Links between sites were given a weight depending on physical distance and effort spent towards interaction with another site. The idea of gravitational attraction was implemented in the model by weighing costs and benefits of maintaining connections. This way, rather than looking for the most likely networks, the authors looked at the most efficient networks in the sense of optimising the costs and benefits that the network demands and provides (Evans et al., 2012, p. 11). The resulting network model is used to study the impact of the volcanic eruption of Santorini on the island of Thera on Late Minoan civilisation. Comparing pre- and post-eruption times, the model revealed that the removal of one node (the site of Akrotiri located on Thera) initially had little effect but later on could have resulted in an unsustainable situation and the disintegration of the network by increasing exchange costs. It should be noted, however, that their decision to increase exchange costs in order to obtain the observed network breakdown is not grounded in empirical data (Brughmans, 2013, p. 652). The results of the study should therefore not be considered actual predictions of past processes or their underlying causal dynamics, but rather as a tool to explore the parameter space of various hypotheses to guide validation with new data in future research (Evans et al., 2012). Network science has by now become firmly entrenched in archaeology as an analytical tool to describe, model and explain links between entities of interest. Perhaps it cannot yet be considered part of the standard toolset, but certainly many archaeologists are by now more or less aware of its potential and implications for the field. That does not mean that no hurdles are left to overcome. The ever-growing availability of computing power and ready-made statistical packages for the application of standard network tools and approaches such as SNA has in recent years certainly contributed to increasing applications in archaeology. However, herein lies also one of its greatest challenges. These techniques are directly adopted from available software packages but rarely adjusted to specific case studies or the peculiarities of archaeological data in general. As a result, a lack of critical engagement results in limited awareness of the underlying assumptions of these tools as they are coded in statistical packages. Yet, such understanding is particularly essential when moving from identification of network patterns in our data towards explaining them through underlying social processes (Brughmans, 2013, p. 641). One way forward for this problem might be found in the current movement towards open science. In contrast with proprietary software, underlying assumptions and models can be consulted both in the source code and accompanying documentation of open-source software, allowing the full research process to be transparent and reproducible.

The history of complexity in archaeology 47 A second issue pertains to a more general archaeological caveat which states that the archaeological record never fully reflects the complexity of past social interactions. As a result, no analytical tool can ever be expected to fully capture this complexity. This is especially the case for network tools such as SNA (Brughmans, 2011). It is therefore essential to keep in mind that the concept of network as a social phenomenon should not be confused with that of the analytical tool we use to try and make sense of it (Knox et al., 2006). Finally, I already mentioned the issue related to adding dynamics to static network structures derived from the archaeological record. Such an addition of a dynamic component to our models is essential in order to move from description of network structures to explanation of its dynamics and emergent behaviour. At the same time, this is also one of the major challenges for archaeological applications of network science. Indeed, it has been stated that “identification of emergent self-organising properties does not explain how this behaviour came about and what it meant for the individuals creating it” (Brughmans, 2013, p. 648) and that “characterising a property as emergent is at best a general description and never an explanation” (Kohler, 2012). Settlement scaling One approach of complex systems thinking which has increasingly found its way into archaeology is that of settlement scaling. Scaling theory builds on models from biology, urban economics, economic geography, complex systems and regional science to look at empirical regularities in properties of settlement systems across scales of magnitude. One particularly wellknown precursor to the application of scaling laws in social systems is ‘Zipf’s Law’, named after the American linguist George Kingsley Zipf, which states that city size decreases in inverse proportion to its rank within the same urban system (Zipf, 1949). That last element is an important point as scaling laws are indeed only valid between cities within the same urban system and do not predict how cities scale between different systems (West 2017). Studies of settlement sizes focus on the identification of hierarchies and rank-size distributions (Drennan and Peterson, 2004; Johnson, 1980). When plotting settlement populations against the rank of the city on logarithmically scaled axes, a straight line indicating a log-normal distribution is expected to appear (Figure 2.4). The degree of divergence from this straight line indicates the degree of integration of cities within the wider system, being either strongly integrated hierarchical systems (concave or primate line) or decentralised, autonomous polities (convex line). Some have criticised rank-size approaches as being speculative and poorly grounded in theory (Smith, 2017). Archaeologists have increasingly turned towards another approach to scaling studies, one grounded in a quantitative

48  The history of complexity in archaeology

Figure 2.4  Different types of rank-size distributions (made by author)

framework consisting of hypotheses and mathematical relationships generating predictions regarding regularities in the attributes of settlements related to their population size. This approach can be generally subsumed under the moniker of settlement scaling. Settlement scaling developed out of studies of regularities in body sizes and metabolic rates (i.e. the amount of energy expended per unit of time) in biological organisms. Already in the 1930s, the Swiss biologist Max Kleiber discovered that metabolic rates of organisms scale to the ¾ power of the animal’s mass.10 This ‘economy of scale’ indicates that as organisms get larger, their energy expenditure becomes more efficient (West, 2017). This is no trivial matter, as without these efficiency gains, that is when assuming a linear scaling relation, surface area increases by the power of two, whereas volume grows by the power of three, meaning that heat generated per unit of mass increases far more than the total area allowing for the dissipation of heat. If metabolic scaling was linear, organisms the size of elephants would have the same metabolic rate as a mouse and would literally cook to death For a long time, it remained unclear as to why metabolic rates followed a quarter power scaling law according to body size. It was only with the works of physicist Geoffrey West and biologists James Brown and Brian Enquist that a general model of metabolic scaling was developed. They discovered

The history of complexity in archaeology 49 that the observed energy efficiencies are the result of properties of the distri­ bution structures of organisms, that is, their cardiovascular system (West et al., 1997). The general idea is that metabolic scaling is not limited by surface area (as would be expected by a linear scaling law) but rather by the rates by which energy and materials can be transported throughout the body. The cardiovascular system approximates a fractal structure, where each part of the structure looks the same across different scales. As a result of this fractal structure, space-filling properties of the network result in efficiency gains inducing economies of scale.11 In recent years, scholars have started to transpose the underlying tenets of metabolic scaling to develop a unifying framework of scaling laws in settlement patterns and socio-economic production and growth. These works particularly focus on super-linear effects (increasing returns to scale) of social interaction and sub-linear effects (economies of scale) of urban infrastructure (Figure 2.5) (Bettencourt et al., 2007). The foundational assumptions of settlement scaling are (1) human interactions consist of exchanges of information and goods that take place in physical space; (2) the intensity, productivity and quality of individual-level efforts are mediated and enhanced through interaction with others; (3) any human activity can be thought of as generating benefits and incurring costs; (4) human effort is bounded; and (5) the size of human agglomeration is both

Figure 2.5  Theoretical scaling graphs showing super-linear, linear, and sub-linear growth (made by author)

50  The history of complexity in archaeology a consequent and a determinant of the agglomeration’s productivity (Lobo et al., 2019, p. 10). In a seminal paper, Luis Bettencourt and colleagues (2007) demonstrated that urban infrastructure scales with population size in a sublinear pattern, with an approximate exponent β in the range 2/3 ≤ β ≤ 5/6 (maximum value of β ≈ 0.83). This means that for every unit of growth in population, a less than proportional growth in infrastructure and services can be presupposed. This makes sense, as cities doubling in population numbers need not necessarily double its infrastructural services such as railways or sewage systems but can instead partially intensify usage of existing infrastructure (i.e. economies of scale). Conversely, this study showed that the relationship between number of granted patents and population sizes of US urban areas scales superlinearly (maximum value of β ≈ 1.29), which means that as cities grow in population size, their innovative potential grows more rapidly. The phenomenon of higher than proportional (socio-economic) output compared to (infrastructural) input is commonly recognised in economics as increasing returns to scale (Krugman, 1991). In urban economics, these results are described as productivity gains that result from economies of scale, the mobility of labour, knowledge spill-overs, and other effects of agglomeration economies. The combination of infrastructural economies of scale with increasing returns on interaction can be explained by assuming that cities tend to grow such that costs of moving within the settlement are balanced with the benefits of the resulting social interactions (Bettencourt et al., 2007). The apparent similarities between organisms and cities in scalar advantages induced by energy efficiencies as they increase in size, has led to the usage of the term ‘urban metabolism’ in analogy with biological metabolism (West 2017). The central notion is that spatial proximity induces improvements in flows of information, thus creating spill-overs, both on a social and economic plane, such as when trading goods, sharing information or developing cooperative strategies. One particular type of information flow is that of learning and transfer of knowledge. As a higher concentration of people gathers, chances of having more knowledgeable agents nearby increase accordingly. In highly clustered networks, interactions can take place with higher frequency and are affected by more rapid feedback loops. By decreasing the distance between nodes within a clustered network ‘lag time’ in the transfer of information is reduced, as well as transportation costs for moving people and goods. The mathematical model of urban scaling is based on the number of people in a settlement, average output per person, travelled distances, and probabilities of encountering other people (Bettencourt, 2013).12 The model revolves around a set of basic assumptions: (1) human settlements can be considered spatial concentrations of human interaction; (2) people arrange themselves in space so as to balance the costs of moving around with the

The history of complexity in archaeology 51 benefits of the resulting interactions; and (3) socio-economic outputs are proportional to the total number of social interactions within the population. In this model, cities are posited as “social reactors”, where increased numbers of social interactions due to population increase or aggregation result in magnification of socio-economic outcomes, both positive (community formation and economic growth) and negative (scalar stress) (Ortman et al., 2014, 2015; Smith 2019). The conceptualisation of cities as social reactors will be discussed in more detail as part of the theoretical framework in the next chapter. One of the seminal applications of settlement scaling is a study by Scott Ortman and colleagues (2014), who worked in the Pre-Hispanic Basin of Mexico, using data from 1,500 settlements covering over two millennia and four major cultural periods. This study showed that, in accordance with the general tenets of settlement scaling, total settlement area increased with population size, on average, according to a scale invariant relation with an exponent in the range 2/3 ≤ β ≤ 5/6. This means that as population numbers increase, settlement density increases first, before extending in size. It has been suggested that the fundamental processes behind settlement scaling in antiquity operated very much in the same way as those of today’s cities. Likewise, this research also showed that scaling laws operated in similar ways in villages and cities, suggesting that the common urban/non-urban dichotomy should be nuanced, and perhaps discarded altogether (Ortman and Cofey 2015). Another study looked at 173 cities from Medieval Europe from the 14th century AD to determine the relationship between population and settled area as predicted by settlement scaling theory (Cesaretti et al., 2016). Here as well, clear patterns of densification of cities with increasing population size were found, consistent with patterns observed in contemporary cities. Again, the observed empirical similarities in the correlation between settlement area and population size supports the hypothesis that settlements share common organisational principles across time and space, resulting in clear patterns in the relation between socio-economic networks and structured urban spaces. It was rightly noted that the successes of settlement scaling in archaeological applications indicate that theories of contemporary urban processes may indeed be applicable to the past, not necessarily because of matching macro-historical structures and institutions, but rather because of fundamental similarities in micro-level behaviours and their emergent outcomes (Cesaretti et al., 2016, p. 17). Settlement scaling constitutes a promising field of research, which is proving an added value to the toolset of archaeologists studying settlement systems in the past and will no doubt produce many interesting studies in the years to come. Still, the framework is not fully established yet and several challenges still lie ahead if it is to gain a household position in the field.

52  The history of complexity in archaeology A first major issue relates to the framework of settlement scaling in general. It has been pointed out that the exponents of settlement scaling have not found the same sound theoretical foundations as those of metabolic scaling. For the latter, it was mathematically shown that space-filling properties of cardiovascular system in organisms give rise to the observed quarter law scaling (West et al., 1997). Such space-filling processes can only be loosely connected with the densification of settlements under growing population sizes. Both processes do not operate according to the same mathematical properties, and as a result, the latter does not attain the fractal-like structures resulting in the efficiency gains observed in metabolic rates. As a consequence, the observed divergences from the expected slope exponent of the power law in settlement scaling are still larger than those observed in metabolic scaling. Geoffrey West explained this higher degree of variation as reflecting the much shorter time cities have had to organically evolve toward the idealised optimal configuration represented by the scaling curves (West 2017, p. 287). Further studies are needed to substantiate this hypothesis. A second issue pertains more specifically to the application of settlement scaling in archaeology. So far, it has proven exceedingly hard to confidently observe super-linear trends indicative of increasing returns to scale of socio-economic activities in archaeological data. Whereas many studies have provided clear indications of economies of scale in ancient cities, it is distinctly harder to find reliable proxies for super-linear returns to scale. Ortman and Cofey (2015) argued in two case studies of the Central Mesa Verde and Middle Missouri populations, that households in larger settlement are on average more productive and that household productivity should be reflected in the total roofed space that was used for daily living, as well as production, storage and consumption of goods. In another paper, Ortman and colleagues (2015) use public monument construction rates and total domestic mound area from the pre-Hispanic Basin of Mexico to argue for super-linear scaling of socioeconomic rates with population. All of these proxies are to a large extent also related to infrastructure and are therefore not fully reflective of the socio-economic productivity of a settlement. This is quite clearly illustrated in the average exponent values of scaling graphs identified from archaeological data. It has been demonstrated that a true measure of aggregate interaction or socio-economic output approximates an exponent value of 7/6 (Bettencourt, 2013), whereas most exponents in archaeological case studies fall between 2/3 and 5/6, the latter being the typical exponent for measures of infrastructure and economies of scale (Ortman et al., 2016). Other potential economic markers could be quantities of imports or craft production output (Smith, 2017). One recent paper takes data from the Peruvian Central Andes during the 15th century to use material outputs (such as the increased frequency of metal objects and production related goods, reorientation of agricultural practices, indications for increased meat consumption and improved health) across settlements and households

The history of complexity in archaeology 53 as proxy measures for economic expansion and improved living standards, stating that these are consistent with the expectations generated by the settlement scaling framework (Ortman et al., 2016). While the link between material correlates and improved living standards is not wholly unproblematic, in this case it seems a valid course of reasoning. However, the authors did not actually take these data into account when trying to calculate the socio-economic output of these settlements, but instead focused again on infrastructural proxies such as the distribution of farming and administrative settlement areas and patio group areas. It can be concluded that, so far, no clear proxies for socio-economic output have been found that can attest super-linear growth and increasing returns to scale through the archaeological record. Doing so will prove to be perhaps the biggest challenge of settlement scaling in archaeology. Cultural evolution Some might wonder why I return here to evolutionary approaches, having discussed and discarded social evolution earlier as a viable framework to study social complexity. The confusion is warranted. Scholars have often used the terms social and cultural evolution interchangeably. I have discussed extensively how social evolution is not based on Darwinian principles, but rather hails from the works of Herbert Spencer assuming a progressive, ladder-like, unilinear process of increasing complexity lacking clear causal mechanisms of change. Social evolution was developed, first in anthropology and later adopted in archaeology as a way to classify past and present human societies. The field of cultural evolution, by contrast, is a far more interdisciplinary endeavour. It combines ideas from archaeology and anthropology with evolutionary biology, ecology, psychology, genetics, sociology, economics, and computer science. It has been noted that “cultural evolution is the theory that cultural change in humans and other species can be described as a Darwinian evolutionary process, and consequently that many of the concepts, tools and methods used by biologists to study biological evolution can be equally profitably applied to study cultural change” (Mesoudi, 2016, p. 481). It is crucial to note that, in contrast to social evolution, cultural evolution explicitly states the mechanisms of change to posit a proper analytical framework with explanatory value. More specifically, it entails a system of variation, differential fitness and inheritance as essential drivers of change in culture (Creanza et al., 2017). ‘Culture’ in this sense entails any socially (rather than genetically) transmitted information, such as beliefs, knowledge, skills or practices (Mesoudi, 2016, p. 481). Yet, the field should not be considered as “neo-Darwinian” in that it does not necessarily use the mechanisms of genetic inheritance – such as random mutation – that biologists following Darwin discovered and that were integrated into evolutionary theory during the evolutionary synthesis.

54  The history of complexity in archaeology The goal of cultural evolution is to use evolutionary drivers and dynamics to explain social behaviour, which is both rooted in deep-time history as well as continuously expressed in social practices and interaction. It effectively places all forms of human societies throughout history on the same footing regarding drivers of social change. This does not mean that all social behaviour or evolutionary outcomes are considered to be equivalent, but merely that the same or similar underlying drivers generate change both in the past and present. The field of cultural evolution has grown immensely over the last few decades. Its first milestones consisted of a set of mathematical models based on population genetics and modes of cultural transmission, developed by pioneers such as Cavalli-Sforza and Feldman (1981) and Boyd and Richerson (1988). The former explored the transmission of cultural traits through three main mechanisms: from biological parents (vertical social learning), peers (horizontal social learning) and unrelated members of the parental generation (oblique social learning). They also added models of cultural mutation using random mutation, cultural selection and cultural drift, all modelled after processes of genetic mutation. The latter added psychological realism to this perspective by introducing various mechanisms of preferential trait copying, including copying from successful or prestigious individuals (indirect or prestige bias), popularity (frequency-dependent bias or conformity, as well as negative frequency-dependence or anti-conformity), and intrinsic trait characteristics (content bias). They also studied the interaction between genetic and cultural evolution, more specifically the conditions for genetic evolution of social learning, which led to the development of the field of gene-culture coevolution (Feldman and Laland, 1996; Gintis, 2011). A recent bibliometric analysis of the field of cultural evolution based on co-authorship patterns identified seven main clusters: (1) mathematical modelling and dual-inheritance theory; (2) computational biology and cultural niche construction; (3) cognitive linguistics and experimental cultural evolution; (4) behavioural ecology; (5) biological anthropology and archaeology; (6) evolutionary psychology; and (7) cross-cultural and phylogenetic studies (Youngblood and Lahti, 2018). Throughout these main clusters, three cultural evolutionary approaches have seen a fair amount of application in archaeology: (1) human behavioural ecology; (2) phylogenetics; and (3) cultural transmission theory (Garvey, 2018, p. 2). The first one has been particularly popular in subsistence studies of prehistoric hunter-gatherers and their material remains (Bettinger et al., 2015). Second, phylogenetics entails the construction of phylogenetic trees as a tool for explaining evolutionary patterns, and has been incorporated in archaeology through the use of cladistics to map relationships among artefacts. The goal is to trace changes and stability in artefact traits as reflecting inheritance versus adaptation in the material domain (O’Brien et al., 2016). Finally, cultural transmission theory is commonly approached through the

The history of complexity in archaeology 55 quantification of artefact variability to identify learning biases and identify meaningful differences in human behaviour (Bettinger and Eerkens, 1997). For now, I will continue to discuss evolutionary approaches as a fruitful field of application for complex systems thinking in the study of past societies. The goal here is not to give an exhaustive overview of the field, but rather to focus on the general level of cultural evolutionary approaches regarding the development of socio-cultural complexity, including the evolution of large-scale human cooperation, the nexus between population size and complexity, cognitive drivers of cultural attractors, and the causal role of culture as a driver of change in human societies. One of the most fundamental debates in cultural evolution with great importance for archaeology pertains to the evolutionary basis of cooperation and the development of complex organisational structures. The central question in the emergence of cooperation is why individuals perform behaviour that is beneficial to another without direct benefit to themselves (Hamilton, 1964). One explanation focuses on cultural group selection which states that people may favour altruistic cultural norms when biases in social learning (e.g. conformity) generate intra-group cultural homogeneity and inter-group diversity, which results in different groups having variable altruistic cultural traits that can be selected upon if they offer benefits to the group but are costly to individuals (Boyd and Richerson, 2009). The emergence of large-scale cooperation in human groups was a crucial process for the emergence of complex societies (Turchin, 2016). This type of cooperation can be defined as a systematic form of cooperation and exchange among ephemeral interactants (e.g. strangers without kin relation), often involving many contributors or co-operators (Chudek and Henrich, 2011, p. 219). The development of this form of ‘ultrasociality’ is considered a major evolutionary transition in the history of humankind, paving the way for the type of societies we encounter in our history books and still see today (Turchin, 2013). Large-scale cooperation is a complex process that emerges from several interacting dynamics operating on different scales. Cooperation typically evolves first in small-scale units such as kin groups, favouring selective processes that select cooperative behaviour that lead to the development and integration of larger collectives. A second phase in the evolution of social cooperation led to the development of large-scale sociality and associated increases in social complexity. It has been noted that ultrasociality is so closely related to social complexity that “these two concepts may in fact be thought of as simply different approaches to be the same general phenomenon by different scientific disciplines: evolutionary science (ultrasociality) and anthropology/archaeology, as well as complexity science (social complexity)” (Turchin, 2013, p. 66). One important line of cultural evolutionary research is the correlation between demographic factors such as group sizes and population movement with the evolution of socio-cultural complexity. This factor was already discussed in the seminal work of Cavalli-Sforza and Feldman (1981), and has

56  The history of complexity in archaeology in recent years been explored further (Aoki, 2018). It is commonly observed that larger-scale societies have larger populations. The question is, however, whether these trends are correlated, and if, conversely, falling population sizes also results in reduced socio-cultural complexity. A well-known example pertains to the original settlers of Tasmania. Upon settlement around 10,000 years ago, who were cut off from the Australian mainland and lost many complex tools and skills including winter clothing, fishing spears and boomerangs (Henrich, 2004). It was argued that this loss of cultural complexity was due to a strongly reduced population size, resulting in a reduction of skilled individuals to learn from and generate innovations. A similar approach was applied to archaeological data from the Palaeolithic, where it was argued that changes in complex technological and social traits such as abstract art, the bow and musical instruments coincide with changing population densities (Powell et al., 2009). Recent work in cultural evolution has started to experimentally explore the validity of the link between population size and complexity. It has been noted that archaeological studies can only test the outcome of the population-complexity model, but not its underlying mechanisms and assumptions (Mesoudi, 2016, p. 487). Recent experiments confirmed the predictions of the model regarding the higher presence of knowledgeable agents in larger groups, providing more people to learn from and support higher levels of cultural complexity in various tasks (Derex et al., 2013; Muthukrishna et al., 2014). Archaeological analyses, however, maintain that little evidence for this hypothesis can be found and that toolkit complexity in hunter-gathering societies rather correlated with environmental risk (Collard et al., 2013, 2005). The link between population size and social complexity remains debated and some case studies have yielded conflicting results. I will return to this debate as it developed in archaeological circles in more detail in the next chapter. Cultural evolutionary models can be used to interpret patterns identified in the archaeological record. Conversely, archaeological data can be used to independently test the outcomes of these models. It has been stated that “to fulfil archaeology’s potential, we should continue to develop models specifically tailored to archaeological circumstances, and explore ways to incorporate the rich contextual data produced by archaeological research” (Garvey, 2018, p. 1). Archaeologists have since then become increasingly aware of these issues and have started to work to mitigate them. O’Brien and Lyman noted that studying processes of change – whether in biological evolution or in the archaeological record – needs its own set of requisites regarding acceptable units of analysis (O’Brien and Lyman, 2000, p. 187). They also pointed out that even biologists are still wrestling with the concept of a species as a unit of analysis and that some struggle is still to be expected for archaeological applications of evolutionary theories. Another important issue pertains to a common statement that ‘culture is too complex for simple models’. This criticism is not only levelled at cultural

The history of complexity in archaeology 57 evolution specifically, but at archaeological modelling, and by extension any type of social systems modelling as well. The idea that the dynamics underlying the actions of human beings in producing culture – with which we differentiate ourselves from all other species on the planet – could be captured in models consisting of a series of equations or formal statements appears ludicrous to many scholars. I will discuss why that argument does not make much sense in final part of this chapter on agent-based modelling. For now, it suffices to point out that this type of criticism bypasses the actual purpose of using formal models in scientific research. They are not meant to fully replicate a given phenomenon – be it culture, complexity, cities, ecological systems, or any other kind of entity – but are rather meant as tools to formalise statements as logical arguments fitting into a coherent whole (Mesoudi, 2016, p. 489). Agent-based modelling One point of criticism levelled at general systems approaches is that they are often not concerned with individual actions, interactions and decision-­ making processes, focusing rather on general properties and structures of the system. As discussed earlier, complex systems thinking goes beyond strictly top-down conceptualisations by focusing on the bottomup emergent properties of individual behaviour. Agent-based modelling (ABM) is a tool where such a bottom-up starting point is elevated to the very core of the approach (Epstein and Axtell, 1996). ABM simulates agents as autonomous entities, acting on imperfect local knowledge of other agents and their environment, following a set of stochastic rules with probabilistic outcomes that cannot a priori be predicted. The aggregate outcome of these rules could be potentially undesirable or even directly opposite to the goals of individual actors. In other words, ABM captures emergent behaviour. ABM also allows for heterogeneity among agents by representing each agent as an autonomous individual, following its own set of rules, behaviours and goals. The ability to study emergent dynamics through heterogenic agents in a bottom-up approach is one of the key differences between agent-based models and equation-based models that assume homogeneous populations exhibiting similar behaviour. One of the most commonly heard questions when considering modelling in archaeology or the social sciences in general, is “why would you model this?” (Van Der Leeuw, 2004; Chattoe-Brown, 2013). In the previous part, I mentioned the common criticism that culture is too complex to be meaningfully captured by simple models. Yet, most people seem to be fine with the idea of modelling physical, chemical, or biological processes. When it comes to the social sphere, however, things become more precarious. Surely – they say – human cultures and the psychological/cognitive processes underlying our decision-making are too complex to be cast in a simple model or a set of equations? Yet, modelling is used without qualms

58  The history of complexity in archaeology to study dynamics in modern human societies for a wide range of things such as traffic flows, the spread of diseases, technological innovation, and many more. In a seminal paper, Joshua Epstein listed 16 reasons for modelling social systems (Epstein, 2008). These can be reduced to seven main purposes: prediction, explanation, description, theoretical exploration, illustration, analogy, and social interaction. Three main scientific purposes for computational modelling can be identified: hypothesis testing, theory building and methodological development (Mithen, 1994). These extend four main benefits: (1) enforce conceptual clarity; (2) understand how things change; (3) infer past behaviour from a static archaeological record; and (4) test other quantitative methods (Wurzer et al., 2015, p. 7). ABM allows us to systematically explore ideas about dynamics and behaviour in the past, and compare them against available data to test our theories (Van Der Leeuw, 2004). In other words, “models are interactive tools that provide opportunities to follow the implications of one’s ideas to find where they are inconsistent and, from this, to build better-informed explanations of empirical data.” (Premo, 2006, p. 92). ABM is indeed an ideal tool for generating and testing hypotheses, and has been described as constructing ‘behavioural laboratories’ (Premo, 2006). They allow the archaeologist to ask ‘what if’ questions and generate a set of plausible answers (Graham, 2006a, p. 60). At the same time, ABM is not merely a ‘computerised’ form of counterfactual fiction. The main difference is that in a computer model, all inherent assumptions and decisions need to be explicitly and formally stated. It is this formal nature that allows the model to be used as a testing ground to conduct controlled, repeatable experiments during the scientific process. An essential difference with other individual- and action-centred approaches in archaeology is that all steps of the process are formalised in mathematical relationships and implemented in a computational environment.13 Following Stephen Gould’s famous thought experiment, if we were to re-wire the tape of life, the chances of evolutionary processes to end up at the same place, or even anywhere near the life we know today, depending on the contingencies of history, are slim to none (Gould, 1989). ABM allows us, not only to rewind the tape over and over again, but to determine exactly the parameters of its initial conditions and see what happens. ABM is particularly suited to study dynamical systems changing over time. Earlier, I already mentioned Shawn Graham’s work on the social networks of the Roman brick making industry. He discussed how archaeological and historical data sources are often static in nature and do not focus on the actual patterning of interconnections (Graham, 2006a, p. 48). He implemented his network model into an ABM to test the effect of different patterns of connectivity on the diffusion of information, thus creating a dynamic model allowing for evolution over time. These dynamics of information diffusion were modelled for the Antonine Itineraries (a collection of

The history of complexity in archaeology 59 225 lists of stopping places along various Roman roads across the Roman Empire) for the provinces Iberia, Britain, Italy, and Gaul and compared based on the relative speed of transmission. It was concluded that Britain and Gaul were less cohesive than Iberia and Italy in terms of network fragmentation, and that similar internal information diffusion patterns are shared by each set of provinces. It was moreover indicated that Iberia had the fastest internal diffusion rate, followed by Gaul, whereas Britain and Italy had a slower but more internally consistent diffusion rate. These results suggest that innovations would be adopted slower, but more thoroughly, by the latter two. This finding was then compared with rates and densities of inscriptions per region, suggesting that the predictions generated by the model appear quite close to the findings in this dataset. The idea could potentially be tested further, for example, against a dataset on diffusion of innovations in technologies and ideas, in order to strengthen the explanatory power of the hypothesis. This provides a good example of how ABM can guide future research efforts by generating testable hypotheses. ABM can be a very potent tool of historical analysis and exploration. Simulation models in general have been around in archaeology for some time now, long before the term agent-based model emerged during the 1990s. One of the oldest examples is Wobst’s simulation model of Palaeolithic social groups (Wobst, 1974). Wobst used evolutionary theories to test the validity of identifying these groups as band societies as well as applied Monte Carlo simulations to predict group sizes. Since the turn of the millennium, ABM applications in archaeology have been clearly on the rise. A fundamental early contribution was Kohler and Gumerman’s (2000) edited volume of ABM on dynamics in human and primate societies. The volume brought together a number of influential scholars and clearly showed the potential breadth of archaeological ABM. The volume also contained a key publication of one of the seminal ABM in archaeology, the Long House Valley model, also known as the Anasazi model (Axtell et al., 2002; Dean et al., 2000). This model was built to explore the relationship between climatic circumstances, resource availability, settlement location, and population growth in Long House Valley (Arizona) in the period 800–1350 CE. The goal of the model was to compare simulated results with the known settlement and population growth history of the Anasazi. The spatial dimension of the model is based on actual paleoenvironmental, climatological, and archaeological data from the Long House Valley in Arizona and thus represents a realistic GIS environment. Agents in the model were conceptualised as households of, on average, five persons. As the model runs, it simulates household responses to food availability, including relocation in the case of shortage or removal from the simulation in the case of sustained lack of suitable resources. Once all households have completed the entire decision-making process, yields are calculated and the cycle starts again.

60  The history of complexity in archaeology The simulation results plot the total Anasazi population over time against the archaeological data. Early simulation runs tended to considerably overshoot the estimated population sizes (Dean et al., 2000, pp. 190–191). Recent implementations, however, seem to provide a relatively good fit.14 It has been questioned, however, to what extent these optimised runs reflect improvements in measurement accuracy and our knowledge of the parameters involved, or whether they are the result of better curve fitting efforts (Wurzer et al., 2015, p. 42). Following the development of the Anasazi model as a household name in archaeological ABM, a number of research groups started to emulate this model in their own efforts. The most notable project building on this seminal work is the Village model developed by the Village Ecodynamics Project (VEP) (Kohler and Varien, 2012; Kohler et al., 2012a, 2012b). The goal of this project was to understand the co-development of society and environment in the emergence of leadership in small-scale societies through dynamics driving village aggregation, growth, and depopulation in southwestern Colorado between 600 and 1300 CE. As with the Anasazi model, the VEP model simulates maize production depending on climate, soil quality, and farming intensity, as well as social learning and usage of other resources such as wood, water, and game to study the impact of resource procurement strategies on demographic patterns (Kohler et al., 2007). The model simulates the degree to which populations exploited various resources and carrying capacities of the landscape. Results showed that initial village formation is often associated with increased group conflicts. As groups increase in size, hierarchical structures needed to be developed to allow sustained growth. Models such as Anasazi and Village are good examples of computational approaches to study social-ecological systems in the past. Recent archaeological scholarship has increasingly focused on human-environment interactions, as part of a wider shift in aspirations of archaeologists aiming to offer their insights to contemporary society, not only for the sake of knowledge of the past, but of the present as well by contextualising present and future changes in socio-ecological dynamics in long-term dynamics of societal change and human-environment interactions (Wurzer et al., 2015, p. 8). ABM plays an essential part in this shift. A final example that I want to highlight here constitutes an extension of the Village model, built to simulate the coevolution of hierarchy and warfare during the Pueblo II period (890–1145 CE) in the Pueblo Southwest (Crabtree et al., 2017). Agents in the model represent households who farm maize, hunt deer and leporids, raise turkeys, fetch water and fuel, trade resources, and react to local variability in environmental productivity through settlement relocation to more productive areas or intensifying resource exploitation. The model then simulates the emergence of cooperation and leadership in hierarchical structures over three scales. First, collaboration of households within groups, then the growth of leadership

The history of complexity in archaeology 61 within groups, and finally the formation of groups-of-groups (polities) during competition over arable land. The results of the simulations suggest that an autonomous village model could not sufficiently explain the observed structures of social organisation during the Pueblo II period (890–1145 CE). Instead, it is suggested that overarching polities existed during the Pueblo II and probably into the Pueblo III period (1145–1285 CE). For these periods, it is suggested that Pueblo culture consisted of a complex hierarchical society. Later in the Pueblo III period, more autonomous organisations operated until the final depopulation of the area in the late 13th century. Some of the underlying assumptions and tenets of this model of hierarchy formation will feature in the model of social complexity developed in the next chapter as well, where they will be discussed in more detail. In the last few years, ABM has gradually, but steadily, moved from being a conceptual curiosum to a tried and tested tool, or in other words, started to gain methodological maturity (Lake, 2014). One of the signs that, after being around for decades, archaeological simulations in general and ABM in particular, have finally started to come of age is that “simulations increasingly provide results that are useful to researchers who were not involved in the modelling process” (Wurzer et al., 2015, p. 6). ABM have also started to become part of educational curricula and more educational resources are becoming available. A recent trilogy of papers provides open access tutorials and learning materials for ABM in archaeological courses, signalling that the approach has indeed started to gain widespread acceptance as a valid conceptual and methodological approach (Crabtree et al., 2019; Davies et al., 2019; Romanowska et al., 2019). It can be expected that ABM will hold an increasingly important place in archaeological research in the near future. Efforts such as these are essential to familiarise the next generation of archaeologists with these tools and their potential for archaeological research. By extension, the same could be said for all other approaches outlined earlier, and for the concept of social complexity and complex systems thinking in general as well. I hope that the current book and the model proposed in the next chapter can contribute to this trend.

Notes 1. The argument is expressed through the formula: C = E × T (degree of cultural development equals the amount of energy times technological efficiency). 2. But see Fletcher 1995 for a discussion on low-density urban areas. 3. Published posthumously in 1924 as part of his book on economy and society: Wirtschaft und Gesellschaft. Grundriß der verstehenden Soziologie, translated in English in 1958. 4. Ibn Khaldun, The Muqaddimah, Translated from the Arabic by Franz Rosenthal (London: Routledge and Kegan Paul, 1986) 5. See different contributions in Faulseit 2016; Maisels 2010; For the Han empire see contributions in Morris and Scheidel 2009; Kidder et al. 2016 and the contribution of Anna Razeto in Creekmore and Fisher 2014.

62  The history of complexity in archaeology 6. My own thinking on complex systems has benefited greatly from participation to the annual Complex Systems Summer School at SFI in 2019. I would highly recommend anyone interested in social complexity and/or complex systems to apply for this program. 7. For good introductions to network science see (Newman, 2010; Newman et al., 2006). 8. For a more detailed overview of the mathematical properties of power laws and how to use them in network science see (Clauset et al., 2009; Newman, 2010, pp. 247–261). 9. See among others: Brughmans 2010, 2013; Brughmans and Peeples 2018; Brughmans, Collar, and Coward 2016; Collar et al. 2015; Peeples 2019. 10. Scaling laws build on the structures and properties of power law distributions, where a quantity of interest Y is plotted against some measure N of the size of the system, expressed in the formula Y = cNb, where c is a constant and b is a fixed exponent. In contrast with Gaussian bell curves, power law distributions typically have no meaningful average value. Instead they consist of many small events and few bigger events, albeit more than expected in a normal distribution. In these cases, events or entities are correlated and not random. When b = 1, a linear relationship occurs, whereas b < 1 is considered sublinear and b > 1 a super-linear relationship. 11. For a more detailed introduction to the properties of fractal structures and its ramifications for metabolic scaling, the MOOC “Fractals and Scalings” by the Complexity Explorer website offers a comprehensive account. 12. For a comprehensive overview of the mathematical framework of settlement scaling see Lobo et al. 2019. 13. The modelling software NetLogo (https://ccl.northwestern.edu/netlogo/) has become the de facto standard in archaeological ABM due to its unique coupling of intuitive design and computational power (Wilensky 1999). 14. See for example the Anasazi model implemented in NetLogo (Stonedahl and Wilensky, 2010).

3

Conceptualising social complexity

Introduction In the previous chapter, I looked at the history of social complexity in archaeology. This overview included social evolutionary approaches, cities and states as the perceived pinnacles of social complexity trajectories, and complexity in social systems thinking. I concluded the chapter with recent applications of complex systems thinking in archaeology. I particularly highlighted four fruitful avenues of past and current research in social complexity and complex systems in archaeology: network science, settlement scaling, evolutionary approaches, and agent-based modelling. In this chapter, I will present a roadmap for current and future research on social complexity and complexity trajectories in the form of a conceptual model of social complexity trajectories. This model will be applied to an extensive case study in the next chapter. In a recent paper, Elizabeth Hobson et al. (2019) distinguish two approaches to social complexity: a descriptive approach which allows systems to be ranked in terms of their relative complexity, and a causal approach which identifies the mechanics that cause one system to be more or less complex than another. Here, I pursue the second approach. The conceptual framework presented here aims to uncover underlying dynamics and causal processes of social complexity trajectories. I conceptualise human societies as complex systems capturing, transmitting and processing flows of energy, resources, and information to sustain their population and maintain or expand their organisational structures. While societies cannot exist without being physically sustained by the processing of energy and resources, its social structures can only emerge through the exchange and processing of information. Moreover, shared information also creates the channels that allow more energy and resources to be appropriated and circulated, such as social relations and exchange networks (van der Leeuw, 2007, 2019). This feedback loop is represented visually in Figure 3.1. Multiple definitions of information have been proposed, starting with the seminal works of Claude Shannon, William Ashby, Colin Cherry, and

64  Conceptualising social complexity

Figure 3.1  Flows of energy, resources, and information (made by author)

others, giving rise to the fields of cybernetics and information theory in the 1950s and 1960s. These works focused strongly on information as part of a communicative process, albeit one where the transfer of information was dissociated from the transfer of meaning as a way to separate structural information from irrelevant noise generated by communication disturbances (Shannon and Weaver, 1949). Information in this sense was inherently connected to uncertainty in the transmission of messages. What we are interested in here, however, is not the average level of uncertainty in communication calculated through Shannon’s entropy measure, but rather how meaningful information is transmitted and processed in the context of human societies. In a general definition, we may think of meaningful information as “any signal that relates an observation to an existing set of meanings” and the information processing capacities of a social collective as “all the available means [it] has at its disposal to register and interpret information” (van der Leeuw, 2007, p. 217). The latter includes all of its expertise, skills, means for social learning, problem-solving structures, and social cohesion mechanisms. Throughout the chapter, I build a case to consider information processing and uncertainty reduction strategies as drivers of social complexity trajectories. I outline how social interactions and information transmission – bundled as social practices and entrenched in social structures – can be considered the foundational building blocks of social complexity. I then build a conceptualisation of complexity as a problem-solving tool in collective decision making, driven by sets of selection pressures and giving rise to formative push-pull dynamics. Next, I trace some of the social outcomes

Conceptualising social complexity 65 of such complexity trajectories through the lens of communities as social reactors, dynamics of polity formation, and the energetic costs of social complexity leading to societal collapse and transformation. I conclude the chapter by integrating the concepts of the adaptive cycle and panarchy as part of a theoretical framework to describe and understand multi-scalar dynamics of change and stability in human-environment interactions and social complexity trajectories.

Building blocks of social complexity When developing a model using complex systems thinking, we must start from the basic constituent units that give rise to the emergent phenomenon of interest. In this case, the model must start with social interactions as the core element of information transmission in human societies, and then detail how organisational structures and social complexity emerge. It can be stated axiomatically that social systems can only originate when people interact with each other. No interaction means no communication, transfer of information or exchange of ideas that form the core of human social life. Moreover, interactions need to be grounded in some type of social groups for any social structure to take form. This means that, at its most fundamental level, any form of social system is constituted by social actions and interactions. At the same time, interaction cannot be random. Social systems consist of social roles, rules, norms, culture, etc. which act as constraints to human behaviour. These systems offer a platform for interactions to take place. This means that, if they are to be meaningful, social interactions must be geared towards this collective sociality. I will here combine practice-­ based approaches with information theory to conceptualise the emergence of social systems from the recursive relationship between individual (inter) actions and organisational structures. In this framework, social practices play a mediating role in the formation of communities and form the basis for the development of social complexity. Information transmission forms the social scaffolding that allows these processes to take place. In this part, I will first discuss the role of social interactions and information exchange in the emergence of social complexity. I then explain how social interactions are structured and turned into social practices. Next, I show how social practices act as loci for information selection and transmission that form the basis for the development of social organisation. Social interaction and information transmission Human beings are a social species. Social interactions and the exchange of information is engrained in what it means to be human. Interaction between group members is a core element in group formation, the development of societies, and the emergence of social complexity. In this book, I consider social complexity as an emergent property generated by the transmission of

66  Conceptualising social complexity information through social interactions. Let us therefore start with taking a closer look at the role of interactions and information processing in social groups. It should not come as a surprise that there is a limit to the amount of interactions anyone can maintain. It is simply impossible to talk to everybody in a large group on a regular basis. Today, we could potentially interact with a host of people across the planet barring some practical limitations such as available time (as well as technology and language barriers). However, even if you were to try to interact with everyone you came across, the question is how meaningful the interaction would be, and what you would remember from it afterwards. These two things go to the very core of every interaction: it serves to transmit some form of information and this information needs to be processed by our brains to be retained. One of the crucial constraints on the potential amount of social interactions are our cognitive limits to information processing. Studies have shown that these cognitive constraints play a big part in determining group sizes and social life (Dunbar et al., 2014; Freeberg et al., 2012). Integrating findings from archaeology, anthropology, cognitive psychology, evolutionary biology, cultural evolution, and others, these studies not only pertain to homo sapiens, but also our ancestors and closest living relatives. The ‘social brain hypothesis’ posits an intricate link between brain size – and more specifically neocortex volume – and the size of basic social units. The hypothesis suggests that complexity of primate and hominid societies was/is created by mutual interactions, defined by social group sizes (Dunbar et al., 2014). But what does that mean? At the core of the argument lies the assertion that constraints must exist on the size of a social group due to limits in information processing. For human groups, studies have indicated that group sizes are limited to about 150 people on average. This is also known as ‘Dunbar’s number’, after the English anthropologist Robin Dunbar who conducted some of the seminal work (Dunbar, 1993). This limit corresponds roughly to the number of people we can interact with on a regular basis. A classic example is Anthony Forge’s study of normative factors explaining settlement size of Neolithic cultivators in New Guinea (Forge, 1972). Forge argued that stabilisation of villages occurred at an average population size of 150 people and explained this process as an attempt to maintain face-to-face relationships within a social group. Yet, you just have to take a walk through your village or city to see that some groups can greatly exceed these numbers. At the university, some classes easily go beyond this number, seemingly without any effects on the efficiency of information transmission.1 In these cases, however, you are not actively or intensively interacting with all members of the group. The threshold of 150 people is also not a hard boundary and there are several ways for groups to transcend it. In Forge’s study, villages could fission and divide into two distinct communities, or they could undergo an internal (horizontal) sub-division into distinct social groups. By internally subdividing,

Conceptualising social complexity 67 each unit still adhered to the proposed threshold, thus maintaining smallscale community relationships within larger groups. Mechanisms of internal subdivision to allow increasing group sizes include the development of social roles, status differentiation and division of labour. Once social groups transcend this first population limit, another threshold appears at around 400–500 people. This threshold was also recognised by Martin Wobst (1976, 1974), who linked it to the minimum required population size for the practice of endogamy within one’s own community. Dunbar’s number therefore refers in the first place to face-to-face groups, rather than social groups at large. The former refers to social networks on the level of the individual, whereas the latter entails those on the level of a community or society. This suggests that there is not one single constraint, but rather that the size of social units can be deconstructed across a hierarchically nested sequence following a regular pattern. Moving up the hierarchy, group sizes increase, whereas the strength of the bonding between people within the groups systematically decreases. Initiating and maintaining relationships and social groups requires social capital, that is, a ‘fixed quantity of social effort that we can invest in each of our friends and acquaintances’ (Dunbar et al., 2014, p. 45). The main idea is that the sizes of social units emerge out of the limits imposed by the ability to handle relationships of different intensities. The importance of every relationship is thus determined by the amount of effort or time invested in it. Dunbar himself has suggested a series of layers – his so-called ‘circles of friendship’ – characterised by thresholds of ca. 5–15–50–150 people, corresponding, respectively, to an intimate group, support group, bands, and communities. Note that each subsequent limit is three times the size of the previous one. The regularity of the pattern is taken as an additional indication that these thresholds could indeed be linked to the cognitive structure of the brain. It has been noted that the exact value of each threshold is subjected to variation across groups and societies (West, 2017, pp. 304–309). This has led some to question the validity of inferring cognitive selection pressures for the size of social networks (de Ruiter et al., 2011; Powell et al., 2017). The one-sided focus on brain physiology as a determining factor ignores our social capacities in developing cultural mechanisms, practices, and social structures that humans develop to counter such deficiencies. These studies argue that human information processing cannot be understood as a simple product of biological factors and that, if we are to explain increasing group sizes and the development of communities, we must look beyond biological constraints (Read and Van der Leeuw, 2008). In a seminal paper, Gregory Johnson (1982) argued that internal subdivisions – in the sense of hierarchical organisation – emerge even in small groups from six individuals onwards. He relates this threshold with ‘scalar stress’ associated with decreasing consensus in decision-making processes and decrease of effectiveness in group performance. Johnson then scales up his argument by stating that social groups generally surpass such

68  Conceptualising social complexity small-group limits and must therefore have mechanisms allowing them to overcome communications stress. The key lies in the development of social organisational structures as information transmission mechanisms. Johnson distinguishes between simultaneous and sequential hierarchies, defining the former as organisational structures in which system integration is achieved by exercising control and regulatory functions by a relatively small proportion of the population, whereas the latter is defined as a structure where consensus is achieved sequentially along different levels of social groups integrated within a given society (Johnson, 1982, pp. 396–403). For example, a sequential decision-making process could encompass consensus reached first within nuclear families, then in extended families, which eventually is passed on to consensuses reached on a larger group level, for example a tribe, as happened with the Hopi in the US Southwest. By relating intra-group sizes (for example, nuclear families of up to six people) to the number of basal groups on any given level (for example, up to six households within an extended family or six extended families within a tribe), Johnson could account for observed increasing group sizes to the integration of increasingly more levels of organisation, without having to discard the initial organisational limit observed in his small-group studies. Societies and groups in this sense are considered sets of nested social structures (Carter and New, 2004). The key to the issue of the link between group size and social complexity is therefore not merely the quantitative increase in social interactions, but the development of information transmission mechanisms across different scales. As social groups grow, they do not simply get larger but rather self-organize to better process socially transmitted information and more effectively make decisions (Auban et al., 2013, p. 56). The question is then, how do these information transmission mechanisms develop? As mentioned earlier, biological factors alone do not offer an explanation why things such as social organisation, cultures, cities, and states emerged. Yet, they provide a good starting point. I will argue here for the crucial role of information processing and uncertainty reduction strategies as fundamental drivers of change in social groups, resulting in a range of emergent outcomes including social complexity. To understand how humans process information, we must start from individual cognition and build our way up to collective sociality. People think, dream, and interact to make sense of themselves and the world around them. This process comes with uncertainty. Not every piece of information is necessarily useful or meaningful. How can we separate useful information from ‘noise’ in the barrage of impressions and sensory inputs that we receive on a continuous basis? Several mechanisms of information selection and processing have been proposed, but one in particular can be highlighted: prediction (Clark, 2016). Human brains have evolved as powerful predictive devices. By constantly trying to predict incoming sensory signals, we can engage with the world through thought and action.

Conceptualising social complexity 69 One powerful strategy is trying to guess the structure and shape of incoming sensory stimuli based on prior information. Correct guesses allow the individual to anticipate and act accordingly. Failed guesses generate a prediction error based on a degree of uncertainty compared with one’s priors that can be used as renewed input for future guesses. This process of recalibration trough ‘predictive processing’ allows learning to update one’s experience and perception of the world. The result is a dynamic, self-organising (social and biological) system using flows of information to constantly reconfigure the perception of oneself and the outside world. The effects of prediction are seen all throughout our experience of daily life. Think of the strange feeling you have when taking a sip from a cup of what you thought was tea, only to find out it is coffee. The mismatch between your prediction and sensory input becomes almost tangible as your brain adjusts its prediction to match reality. The predictive brain is “not an insulated inference engine so much as an action-oriented engagement machine” (Clark, 2016, p. 1). This means that prediction happens based on external input, and is therefore inherently connected to the external world. It also means that it needs to conciliate externally generated input with internally driven actions. People act and interact within social environments, as part of social groups. The external world therefore consists for a large part of other people trying to act appropriately within a community, society, or any other collective social unit. To do so requires inferencing other people’s intentions from their behaviour. Doing so, however, requires background knowledge and social embeddedness. External stimuli and internal drivers must be integrated in such ways as to allow individuals to align their actions and interactions as part of mutually reinforcing social structures. This also means that the individual as a biological prediction machine does not give us the full picture. People are not able to form these collective units merely by predicting their (social) surroundings on an individual basis. This means that biological mechanisms of uncertainty reduction alone cannot explain why and how social collectives emerge. Instead, we must look at social, collective mechanisms of information transmission and uncertainty reduction that allow people to interact and live together in structured and predictable ways. In the next part, I will focus on two social mechanisms of information transmission. First, the formation of social practices as ‘bundles of information’ operating as selection mechanisms for the transmission and processing of information. Second, the embedding of information in material surroundings as a way of transmitting information in a social group. Social practices as bundles of information In this part, I will elucidate how information becomes structured in society through four dimensions of structuration: time, space, social, and material. The main mechanism behind this structuration is the development of

70  Conceptualising social complexity social practices. Social practices have been studied extensively in the field of sociology, particularly through practice-based approaches most famously advocated by Pierre Bourdieu and Anthony Giddens. The idea of practice-based approaches is to find a way to mediate between agency-centred and structure-centred approaches by integrating individual actions and social structures into a single conceptualisation of social practice. Five key components of social practices can be distinguished: (1) interaction, (2) social agents, (3) communication, (4) social knowing, and (5) coupling (Castellani and Hafferty, 2009, p. 38). The first two components, interaction, and social agents refer to the aspect of (social) agency, whereas numbers three and four are subsumed under ‘symbolic interaction’, which is related to structuring properties in social systems. Social interaction occurs when two or more people ‘encounter’ each other – that is, create an episode of mutual awareness – supplemented by communication (Turner, 2003, p. 4). These encounters can be both focused, that is, actively engaging its participants in mutual information exchange, or unfocused, where people maintain mutual awareness and implicit communication but do not engage with each other directly (Goffman, 1972). Communication is considered here in a broad sense as any exchange of information, regardless of the medium, rather than verbal exchanges per se. Social knowing involves aligning social practices with the worlds in which humans live, that is, familiarising oneself with the prevalent ‘ways of doing’ in a social system. Coupling refers to the intersection between agency and structure that allows social practices to be connected, attached, and merged with elements of other practices or their constituent components. Ease of coupling is determined by the inherent plasticity of social practices because of the constant renewal of its structures through recursive performance. Constituent components of one kind of practice can be used within and between other practices, constituting a connective structure that holds complex social arrangements, for example a community, in place (Shove et al., 2012). The consistent repetition of a coupled set social practices can be considered the most essential formative process of a society. An interconnected set of social practices constitutes a social system, to the extent that it has been argued that “social complexity theory begins with the assumption that a social system is a type of social practice” (Castellani and Hafferty, 2009, p. 44). Giddens stated that social life consists neither of a collection of individuals, nor of a social structural order by itself, but should rather be seen as a process built around the actualisation of social practices (Giddens, 1984). Rather than conceptualising the individual and social structure as two distinct elements, he proposed to view both aspects as being two sides of the same coin in a recursive loop of ongoing social dynamics. Social practices are then structured by the ‘binding’ of time and space to social systems. The philosopher Ludwig Wittgenstein argued that meanings are enacted through practices embedded in space and time. Giddens used Wittgensteinian premises, to state that structures are produced and reproduced in specific

Conceptualising social complexity 71 contexts. Structuration therefore entails a process involving human beings’ ongoing actions as they occur through the flow of time. Structures consist specifically of rules and resources that people use in daily life to engage in interactive practices, both constraining and enabling the total possibility space of social action. Social structures are not merely instantiated through social interactions, but also, in turn, shape the social content of practices across time and space as to make them internal to social relations. Giddens proposed the so-called ‘duality of structure’ to conceptualise the role of social structures as medium of social action as well as being reproduced by those very actions. As a result of this aspect of reproduction, social relations between actors can be organised in durable patterns embedded in time and space, thus allowing the constitution of social systems. In this view, structures are not pre-existing external elements limiting the potency of individual actions but refer to intersecting sets of rules (guidelines for actions) and resources (media for transformative capacity) acting as structuring properties of social practices, much in the same way as language is used to structure speech. Social structures must be interpreted as some sort of ‘virtual order’ rather than a distinct social reality and are only actualised during processes of social practice. As a result, structures can be both enabling and constraining, rather than merely imposing external barriers to social action. Bourdieu’s conceptualisation of social practice centred on ‘habitus’, referring to a general and unconscious set of interpretative and motivational guidelines or ways of acting and thinking (Bourdieu, 1977, 1986). Habitus can be interpreted as the structuring apparatus of potential actions that is obtained naturally during upbringing. Bourdieu provided a more elaborate account of the context-specific reproduction of social structures encountered with Giddens. He conceptualised society as a range of social spheres composed of interrelated fields (political, economic, cultural, scientific, etc.). All are fields of struggles, with desired resources at stake. These societal fields are defined by a combination of horizontal functional differentiation and vertical stratification. The range of practices acquired through one’s habitus is always field-specific and determined by a specific kind of capital (economic, cultural, symbolic, and social). These specific types of capital are the socially effective forms of power that structure the social space (Malaina, 2014, p. 480). To recapitulate, social practices in a basic definition can be considered ‘a routinised type of behaviour’ (Reckwitz, 2002, p. 249). Practices emerge out of the recurrent iteration of ways of doing that attain meaning beyond the here and now. Performances actualise patterns of behaviour, but it is only through successive repetition that these become ‘encoded’ over time. This recurrence can be referred to as an ‘attractor state’ for cognition, interaction and behaviour (Boyer, 2018, p. 263). Social systems often face recurring needs, requiring recurring goals to be met with recurring actions (Bailey, 1994). It is this threefold recurrence which produces patterned replication over time resulting in ordered systems. That does not mean that

72  Conceptualising social complexity social systems are deterministic or follow from linear relations between actions and structures. Recurrence merely reduces maximal randomness. Structured practices therefore allow social groups to reduce uncertainty by providing internalised ways of doing as templates for action and interaction within the system. The idea of uncertainty reduction in the transition from individual-­ level actions to coordinated system-level interactions has been formalised through the concept of social entropy (Bailey, 1990). The concept of entropy was originally coined in physics as a property of thermodynamic systems. In this context, social entropy can be considered as a measure of the probability of events and order in the face of unpredictability and uncertainty in organisational structures of a society (Netto et al., 2017). It is important to stress that the reduction of entropy is not a deterministic process induced on a system level. Instead, it emerges out of the bottom-up coordination of actions between agents pursuing a wide range of goals. These agents operate within an operational space defined by spatial, temporal and organisational parameters (Pickett et al., 2005). Through the concurrence of these three spheres, a coherent and interconnected set of practices, and the people involved with them, can be defined as belonging to a social system. To operationalise this framework for archaeology, I will add a fourth dimension, that of the material. Social entropy provides a way to explain how social systems deal with uncertainty occurring in the transition from individual actions to systems of interactions. Until proven otherwise, we can assume people possess some degree of free choice in their decision making and can therefore initiate a multitude of potential actions. Yet, out of this enormous possibility space, routinised patterns of behaviour still arise to allow social groups to proliferate and reproduce. This is possible because out of the huge amount of potential actions, only a limited set is selected and actualised. Social practices constitute an important mechanism of selection and uncertainty reduction, providing guidelines for the actor on how to behave or communicate within a given social group, and therefore, effectively how to actualise actions. The key lies in the recursive nature and dual role of practices as active and generative forces in social structuration. New actions are grafted upon those preceding it. Practices are therefore always inherently contextualised. They must necessarily take place in a social and material setting, be they located in space and occur at some point in time. Essentially, this context constitutes an ‘information environment’ that reduces the potential interaction space. Let us break this down. To explain the emergence of social organisation, community formation and social complexity out of constituent interactions and practices, a bottom-up framework with mechanisms of micro-macro transitions is needed. I will discuss four structuring dimensions (social, temporal, spatial, and material) of social practices that provide such mechanisms.

Conceptualising social complexity 73 Let us start with the social dimension. To properly distil the emergence of social systems, we need to explain these formative dynamics within a given social organisational unit. Some inspiration can be found in the micro-­ sociological approach of Peter Blau. Blau (1964) conceptualised social behaviour and the emergence of macro-structures from bottom-up face-to-face interaction between human agents, more specifically, through imbalanced exchanges. He argued that extended patterns of social exchange would give rise to enduring organisational forms with qualities beyond those of the individual people in the organisation. Even though large-scale associations develop out of elementary exchange, he notes that emergent properties outweigh the dynamics of small-scale exchange transactions as not all aspects of society can be reduced to day-to-day interactions. This of course sounds familiar in terms of complex systems thinking, which conceptualises human societies as emergent structures from simple interactions between multitudes of agents in a non-linear way, which can therefore not readily be reduced to the dynamics of these foundational interactions. Explaining the transition from micro-level interactions to macro-level social structure was also the primary outset of the phenomenological approach, originally conceived by Alfred Schutz, but mainly advocated by Peter Berger and Thomas Luckmann. Their approach centred on elements of ‘habitualisation’ and ‘institutionalisation’ (Berger and Luckmann, 1966). It is the concept of habitualisation that is of interest at this point. Habitualisation constitutes the construction of a fixed pattern of actions by actors dealing with frequently re-occurring situations. Habitualised actions provide a template for these situations. For example, in theory, there may be many ways to make a ceramic vessel. However, by becoming habitualised, one set of actions is selected and actualised. Through iteration, actualised patterns of behaviour are granted purposive direction and specialisation. In short, it is transformed into a social practice. The vessel is made in a certain way for a certain purpose, for example a jug to pour wine. From then on, any person undertaking these actions is freed from the pressure of making conscious choices repeatedly at every step of the process, as would be the case in undirected or unstructured trajectories. It provides a stable background for human activity to develop with a minimum of decision making, thus freeing up energy for conscious decision making whenever this may be necessary. In other words, the background of habitualised activity acts as a stage for deliberation and innovation in conscious actions. Besides social structuration, space and time can be seen as “a primary means of structuring social encounters and so producing and reproducing social relationships” (Laurence and Wallace-Hadrill, 1997, p. 219). The temporal dimension seems a straightforward one. One action precedes another and could potentially influence the subsequent one, but not the other way around. However, time should not merely be reduced to a linear progression of events.

74  Conceptualising social complexity A more sophisticated conceptualisation of time and temporal changes in archaeology was developed by the Annales school, which formed the leading current of historical analysis in France from the 1930s to the 1970s. The core tenets of the Annales are centred on a tripartite division of temporal change and a conceptualisation of the co-evolution of time and structure (Bintliff, 1991). History is said to develop on three parallel levels operating at different speeds. First, at the level of individual time, momentary actions make up the bulk of conventional political histories. These interact with processes unfolding on the second level, that of social time where social structures and institutions developed over multiple decades or even centuries. Third, mid-term processes are integrated in the level of geographical time (also colloquially referred to as the longue durée), a deep time perspective where changes are slow and often invisible. Key factors operating on the last scale include geography, climate, and demography. The Annales historians argued that history is always the result of multiple processes operating on different temporal scales. To understand one event on the level of individual time, it must be contextualised in a wider framework covering temporal scales up until the longue durée. For now, it suffices to state that all action is embedded and structured by time operating on different scales. Gavin Lucas (2005) proposed a multi-­ scalar perspective of time through the ‘echo’ or retention length of an event.2 Older events with longer retention lengths could in this sense have greater impact than something more recent, thus transcending conventional linear progressions in time and temporality. Let us take a simple, abstract example (Figure 3.2). Events A–B–C–D occur in chronological order. When looking at the last element, event D, from a linear time perspective, we would assume the immediately preceding event C to have had a more direct influence on D compared to B and A. However, in a multi-scalar perspective of time, event A could have the most influence on D, whereas C only has limited influence, and B none at all due to differing retention lengths. Opening the conceptualisation of time beyond a single progression of interlinked events in a chain of causality, prima facie expands the scope of the possibility space of social practices and their constituent components. As we have seen earlier, social practices are constructed out of the routinised iteration of behaviour, actions, and interactions. If all actions can be potentially influenced by every action in the past, a theoretically infinite

Figure 3.2  Non-linear flow of temporal change (made by author based on Lucas, 2005, p. 36)

Conceptualising social complexity 75 potential combinations exist for new social practices to be constructed. Yet, we discussed how structured patterns of behaviour emerge that become encoded in a set of social practices over time. This is possible because humans act as ‘uncertainty reduction devices’ by selectively bundling and transmitting information. Even though the potential possibility space is infinite, the actualised possibility space follows distinct, socialised patterns. Every act of selection preceding the genesis of social practices is defined by its historicity. Once a selection is made, it becomes entrenched in a subsequent ‘pathway of development’. It is important to note that this is not a return to a linear conceptualisation of time. The multi-scalar conceptualisation of temporal events with differing retention length still holds here. In fact, it can be suggested that events with longer retention length have a higher probability of being selected as components for generating social practices. The simultaneous reductive and productive role of historicity has been called ‘generative entrenchment’ (Wimsatt, 2007). The idea is that historically defined processes – that is, those processes emerging over time – are built cumulatively by accumulating dependencies in patch-dependent processes. However, once such a dependency is created, it limits the potential set of subsequent connections. In our earlier example of a sequence of events from A to D, once a system arrives at D, the historicity of the earlier progression may eliminate E and F as viable future options, steering it towards G or H instead. The latter two may be equally viable and could therefore both be potentially selected to continue the pathway. Which one is actualised could depend on a wide range of factors. This uncertainty is captured by the concept of ‘contingency’. Contingency can be considered a conditional probability, in the sense of the likelihood of an event given past states (Currie, 2018, p. 207). Contingency precludes determinism by probabilistically selecting viable pathways of development. While societies are constantly going through changes, formative processes are not driven by random mutations across the state space. Dynamics in complex systems typically tend towards a limited number of recurrent system states, so-called attractor states. Crucially, a system evolving towards an attractor state can also reach a point where it may face several potential outcomes with distinct qualitative properties, called a bifurcation point. It is at these bifurcation points that contingency works as a conditional probability in shaping future pathways. Path-dependency therefore holds the middle between randomness and determinism and is an essential element of uncertainty reduction. Let us now return to the question of what guides the selection process in actualising patterns of behaviour through social practices. We already indicated that the key to uncertainty reduction lies in iteration and structuration. But how do actors decide which patterns of behaviour to iterate? In a completely ‘blank-slate’ world, no clues exist for which actions and interactions are viable, effectively reducing all interactions to a random process,

76  Conceptualising social complexity prohibiting the structured transmission of information. However, as we have seen, historicity ensures that certain pathways of development become actualised. Moreover, the world around us is not a blank slate. It is already filled with people and things that offer ‘hooks’ for renewed action and interaction. I will now discuss in some more detail how people reduce uncertainty by using and reshaping their surroundings for building information environments. This discussion focuses particularly on the remaining two dimensions of social practices that have not been explicitly discussed yet: space and materiality. While one may be tempted to consider the spatial dimension more easily definable and see it as a mere passive backdrop for human action, it is, like temporality, neither straightforward nor absolute. The British social geographer David Harvey (1973, 1969) has argued for a tripartite ontology of space, consisting of absolute, relative, and relational space. Absolute space can be considered a ‘practical’ interpretation of space as a preexisting container providing the background for human action and interaction, existing independent of things appearing inside it. It corresponds to the spatial dimensions in Euclidean geometry or Newtonian physics. Next, relative space interprets spatial dimensions as constituted from of the relations between different objects. It is therefore a space created by things and people, rather than an independent background for life. Third, relational space is an inherent characteristic within individual elements themselves, in contrast with relative space, which is explicitly determined by the external relationalities and juxtaposition between different objects and people. The latter two dimensions are not quantifiable or geometrical in the sense of absolute space. They rather constitute a qualitative space created out of the embedding of people and things within the functional background of social practices. They are thus also inherently tied to temporality as discussed earlier. We can for example look at the works of Anthony Giddens, who explicitly drew temporal and spatial dimensions into the analysis of social practices through the concept of ‘locale’, defined as the temporal and spatial context in which social practices are manifested (Giddens, 1984). Locales can essentially be situated within any spatial setting, a room, a house, a street corner, a town, a city, etc. Spatial properties and performance of social practices can also be integrated in the concept of ‘place’, defined as ‘lived space’, ascribing meanings, identities, and memories that actively shape people’s daily practices and experiences (Feld and Basso, 1996; Low and Lawrence-Zúñiga, 2003; Meskell and Preucel, 2007; Rodman, 1992). Places offer spatial contexts for people to orient themselves and act within culturally constituted landscapes based on heterogeneous social knowledge and experience (Robb, 2007, p. 9). In short, practices and places are continuously and recursively redefined. To adequately cover this mutual influence, we must consider space as an effect of practice, as well as the effects of space on practices (Netto, 2017, p. 127). The question then arises, how does this work? Places act as

Conceptualising social complexity 77 structuring properties of social (inter)actions because of their role as information environments. This is made possible through the duality of cognition and space, that is the cognitive properties of space and spatial properties of cognition. Vinicius Netto et al. (2018, 2017) raise five main points in their work on the effects of space on entropy reduction: 1 Cognition is situated and extended: Cognitive activity occurs within an environment and is shaped by the coordination with external elements. 2 The built environment is loaded with information: Cognitive processes trigger associations with environmental elements through incorporation of socially acquired information. 3 Spatial information relieves cognitive work: Short-term memory is constrained by limits to information processing. Symbolic off-loading of information transforms the structural environment into an informational environment through mnemonics and cues for action. 4 Cognition is temporally situated: Response capacity can be extended through interactions with the spatial environment in a continuous updating of plans in response to changing conditions. 5 Spatial information serves action: Information encoded in the environment enables others to build inferences for their own actions that may or may not correspond to the original meaning. To create an information environment, space alone does not suffice. This entails not merely the physical properties of space in a Euclidean sense but also how space is used for human activities to provide context for social life, as well as its material context (Bryant and Jary, 1997). The information environment then becomes part of the practice at hand through its spatial association with meaningful performances. It has been noted that “understanding specific practices always involves apprehending material configurations” (Schatzki et al., 2001, p. 3). Conversely, in becoming entangled with social practices, locations themselves can also allow inference of meanings regarding the activities that took place there. After all, this is what archaeologists do all the time when we dig up buildings and look for clues as to their function. Information is contained in the physical world around us, and therefore inherently deposited in the objects and structures providing the setting for human life (Hidalgo, 2015). These objects allow people to communicate messages, coordinate our social and professional activities, and transmit knowledge and knowhow as the necessary ‘software’ that allow information processing to take place. The material environment is thus transformed into a societal frame of reference. Embodying information in matter requires and enables people to push their cognitive capacities beyond biological limits, often beyond what a single individual or group could ever achieve. Think, for example of such projects of enormous complexity such as the LHC collider in Switzerland,

78  Conceptualising social complexity which requires huge amounts of knowledge and knowhow to be brought to completion. One need not only think of such megalomaniac projects to find projects exceeding the individual capacities of a single individual. Think for example of the many specialised and highly skilled workers needed to complete major architectural projects, such as the Great Pyramids of Egypt or the Parthenon in Athens. Also, more mundane undertakings such as the construction of irrigation systems, building of ships or specialised artisanal production generally exceeded individual capacities of knowledge and knowhow. To transcend these individual limitations, people need to collaborate and form social and economic networks that allow us to collectively embody more knowledge and knowhow. One crucial element to form such networks is the manufacture and distribution of artefacts, helping us to increase our capacity to collectively process information. Obvious examples are written records such as books, inscriptions, law texts, etc. However, carrying information is an inherent property of every object, including everyday objects such as utensils, weapons, and ceramic vessels. These objects can then be considered the physical embodiment of information and context-specific properties that this information helps carry. The same holds true for the built environment at large. Without the assistance of the material as a regulatory factor for information transmission, our sensory system would have much more difficulties coping with the increased demands resulting from increased social interactions. The material frame of a community has a slower replication rate and possesses more inertia than social activities, communication, and movement. It has therefore been suggested that major changes in nature and size of a community can only be sustained if a new assemblage of material aids to interaction and communication is developed (Fletcher, 1995, p. 7). The material thus provides an integrative framework for daily life and a frame of reference for active behaviour as it carries non-verbal signals about the patterning of space and time. At the same time, material entities can also obstruct active behaviour due to the inertia acting as barriers to signal transmission. Information is stored and accumulated in networks of people and objects. The creation of these networks has been called – somewhat poetically – the ‘crystallisation of imagination’ (Hidalgo, 2015). The underlying idea is that the complexity of the information embedded in the material framework created out of these networks is a reflection of the complexity of the network itself, or in other words, the manifestation of the limits of its imagination. It has been noted, however, that we should be critical of the convenient assumption that social actions and verbal meanings have direct causal connection to the material environment of a community (Fletcher, 2004, p. 115). Instead, the relationship between material behaviour and active behaviour should be carefully considered on multiple scales of analysis. First, the relationship with the small-scale spatial and temporal patterning of social life should be elucidated. Second, this should be related to the middle level

Conceptualising social complexity 79 of behavioural parameters of human interaction. Third, the behavioural aspects of social life should be integrated in a view on large scale constraints of energy and resource supply, which affect the capacity of a community to replicate itself and its material context. To transpose this to the creation of information environments as contexts for social interaction, we can turn to Amos Rapoport’s (1982, 1988, 1990, 2006) model of material environment-behaviour interaction. Rapoport defines his approach with three questions: What characteristics of human beings influence characteristics of built environments? What effects do built environments have on people? What mechanisms link humans and the built environment? Rapoport distinguishes between three levels of meaning: (1) Low level meaning focusing on mnemonic cues for identifying the uses for which certain material settings are intended, enabling users of a building, city, or space to behave and act appropriately and predictably; (2) middle level meaning where deliberate statements about identity, status, wealth, power, and other traits are communicated through buildings and cities; and (3) high level meaning consisting of a type of symbolic representation that only exists within the context of a specific cultural and religious system. Together, these three levels constitute the overall negotiation of meanings expressed by the built environment. This approach clearly emphasizes the recursive relationship between human action and built environment as mutually constituent components. It is through the combination of social, spatial, temporal and material structures that information is ‘bundled’ in social practices that form the scaffolding for the development of social organisation, community formation and social complexity. Before moving on to the model of complexity formation, however, there is one more building block that I want to address, the emergence of social structures out of the structuration of social practices. Emergent social structures In the previous part, I discussed how social practices can be considered crucial scaffolds of social interaction through their role as bundles of information, guiding human behaviour in social contexts. In this part, I will outline how externalisation of information – that is the creation of an information environment – allows social structures to emerge. I already alluded to two concepts advocated by Peter Berger and Thomas Luckman – habitualisation and institutionalisation – but only discussed the former. Let us now turn to the second one to start off the discussion on the emergence of social structures and organisation. Remember that habitualisation entails the construction of a fixed pattern of actions in re-occurring situations. In principle, this iteration could perfectly well occur on the level of the individual. By contrast, institutionalisation occurs with reciprocal typification of habitualisation among multiple actors. Institutions can be defined as systems of interrelated rules and norms which prescribe

80  Conceptualising social complexity particular roles and shape social relations, thus acting as regulators of social interaction (Currie et al., 2016, p. 200). Institutions are rarely created instantaneously but imply a sense of historicity as they are built up from a history of actions performed among a certain social group (Berger and Luckmann, 1966, p. 52). Institutions are created and maintained by iterative interactions among human agents through structured practices. One typical example of institutionalisation is the development of markets. This type of institution also entails a form of scheduling, linking people to specific times and spaces to perform certain activities, that is, develop habitualised social practices performed in specialised places. Effective markets successfully solve cooperating problems by linking micro-level behaviour (selling, buying, bargaining) between heterogeneous actors to workable and cost-effective organisational structures (Blanton and Fargher, 2016, p. 70). To be able to function in a market requires the development of institutionalised patterns of interaction, allowing people – even if they are total strangers – to reliable interact in a way that is familiar for all parties involved. In short, uncertainty needs to be reduced to a minimum. Uncertainty reduction can happen through knowledge of the relevant social practices, and associated behaviour. For example, when partaking in a bidding auction, this requires knowledge of how the auctioneer will act and interact with the bidders, as well as being able to gauge the audience and wait for the right time to launch a bid to maximise chances of success. Operating within such a context is highly contingent on having the right background knowledge in a personal, practical, and institutional level. For an outsider, the process borders on the edge of total chaos. For the insider, however, the intricacies of the game are all too clear and maximally employed to his or her advantage. As certain practices are favoured within a given community, the actions associated with them are cast in specific procedures, guidelines, and ­structures  – in short, institutions – to build social organisations around the standardised outcomes of such practices. The agents participating in a social system are thus principally responsible for the very construction of this system in the first place. As such, they could, in principle, change any prevalent set of institutionalised practices at all times. Yet, as time goes by, existing trajectories become increasingly entrenched in historical pathways of development. Children born into an existing social system are from the onset ‘socialised’ in accordance with its prevalent norms and values. For these new generations, the social system already appears a given objective reality. In summary, the social system is first externalised out of the interactions between numerous agents, next it is objectified as a fixed set of practices is casted in an institutional framework, and finally, as it becomes increasingly perceived as an objective reality, the social system produced by these agents themselves, will in turn start to act back upon these producers (Berger and Luckmann, 1966, p. 57). Over time, an institutional environment can be created that greatly constrains the continued ability to change the internal set-up of society. Existing structures become entrenched as they

Conceptualising social complexity 81 prevent the creation of new avenues, even when faced with changing external stimuli. Once established, a system of institutional norms creates an interlocking of interests that keep it in place, even if individual devotion to the underlying values starts to wane (Parsons, 1990, p. 326). It has been observed that whereas organisations facing many different challenges within the same field initially show considerable variability, over time, they increasingly show a marked homogenisation around a smaller range of organisational responses. DiMaggio and Powell (1983) have tried to explain the apparent homogenisation in institutional structures and responses through the concept of isomorphism. The original definition of institutional homogenisation by Hawley (1968) states that this process acts as a constraining factor that forces one unit in a given population to resemble units that face the same set of environmental conditions. DiMaggio and Powell identify three different mechanisms of what they call ‘institutional isomorphic change’: coercive, mimetic, and normative isomorphism, respectively, stemming from political influence and problem of legitimacy, standard responses to uncertainty, and professionalisation. These processes of interlocking interests despite changing circumstances and progressive homogenisation of institutional structures, effectively creates pathways of development. Once a given society embarks upon one such pathway, sunk costs associated with this pathway make it increasingly difficult to break out of this set pattern. Institutions are therefore prime providers of societal stability and structure. They work by constructing structured avenues for iterating practices in social contexts. Institutions effectively act as uncertainty reduction mechanisms on a macro-societal level. Institution building thus forms a bridge between individuals and society in the formation of social complexity.

Complexity formation In the first part of this chapter, I discussed the building blocks of social complexity – interactions, practices, and organisational structures – through their role as mechanisms for information transmission and uncertainty reduction. Having laid the foundations, it is now time to build the superstructure. In this part, I will discuss the drivers and mechanisms of social complexity formation, followed by its outcomes on the level of communities and societies in the next part. Whereas the previous part detailed what the core elements of social complexity are, this part will address the question of how it emerges, followed by why in the next part. Selection pressures and collective decision making The principal causal factors in social systems are often related to the material conditions of human existence, that is, the demographic, ecological, technological, and economic forces at work in social life. Their causal significance

82  Conceptualising social complexity as selection pressures for system development is derived from basic human needs concerning subsistence and reproduction. These selection pressures operate within communities and societies as socio-politically and geographically bounded systems. Development of social systems occurs through both endogenous and exogenous drivers. In addition to these material drivers, other factors such as warfare, information management, long-distance trade, and elite competition have all been highlighted as potential drivers of complexity formation (Johnson and Earle, 2000; Sanderson, 1999; Turchin et al., 2015). The American sociologist Jonathan Turner (2003) identifies three levels of selection pressures for social organisation. Micro-level forces such as emotions, needs, and status are generated out of direct encounters and face-to-face interaction. These give rise to meso-level dynamics in social groups, such as social segmentation, differentiation and integration. Meso-level social groups organise themselves through the creation of larger socio-cultural organisations driven by macro-level forces such as demography, production, distribution, regulation, and reproduction. These act as key selection pressures on ‘the institutional core’ or basic elements of human social institutions, including economy, kinship, religion, polity, law, and education. Turner’s framework provides a starting point to discuss the formation of organisational structures and social complexity. It is important to note that dynamics in complex systems can never be reduced to mono-causal processes. Social complexity trajectories are generally driven by a combination of multiple causal processes. These causal factors exert effects in a wide range of domains and act as selection pressures for societal development. Selection pressures act as a set of key ‘forces’ operating on different scales and underlying emergence and development of social structures (Turner, 2003, p. 6). This should not be considered as a return to functionalist approaches in sociology, as these forces are not static functional requisites, but rather causal factors that, depending upon their valances, probabilistically exert varying degrees of selection pressures on social organisation (Sanderson, 1999, pp. 8–9). I will focus here on eight main types of selection pressures: subsistence, cooperation, competition, interaction, distribution, production, governance, and demography. I will return to these selection pressures in more detail in one of the next parts when discussing multiscalar push-pull dynamics. A probabilistic approach to multi-level selection pressures for social organisation can help untangle the underlying factors generating social complexity and transcend reductionist mono-causal approaches. To formalise this probabilistic approach in a complex systems framework, I will integrate it in a model of decision making and information transmission in collective action. To this end, I turn to an algorithmically formalised model3 of decision-making processes proposed by Cioffi-Revilla (2005) who postulated that various driving forces, or stimuli, operate on the emergence and subsequent development of communities through a dual

Conceptualising social complexity 83 canonical – as in undergoing variations on a recurring theme – loop of signal detection, information processing, and problem solving resulting in either successful adaptation or failure of social organisation. This dual loop consists of a ‘fast process’ of crisis and opportunistic decision making through collective action, which feeds a ‘slow’ process of socio-political development or decay. The model is designed to start from a ‘blank’ initial state of complete egalitarianism, to take into account the development of social complexity from pristine state development until present and even future dynamics of social organisation. The basic form of the model can be represented as: A ← 〈( K ) ∧ (C | K ) ∧ ( N | C ) ∧ (U | N ) ∧ ( P | U ) ∧ ( S | P )〉 Each symbol denotes a distinct step in a fast process event and the angular brackets indicate that the conjunction of events is ordered from left to right. The algorithm is represented graphically, along with potential outcomes, in Figure 3.3. The model starts when a given social group (K) without any a priori defined system of social organisation is faced with situational events (C).

Figure 3.3  Canonical loop of collective decision making (adapted by author from Cioffi-Revilla, 2005)

84  Conceptualising social complexity These events can be organised along a variety of lines, including both stresses and opportunities, endogenous and exogenous processes, social or physical in nature, and human or environmental induced. The nature of these situational events as stimuli of social development is purposefully left unspecified to facilitate wide applicability of the model. Once the situational event occurs, it must be correctly perceived by the social group and induce an understanding of the need to undertake a collective action (N). To do so, the situational event must be correctly perceived and separated from alternative perceptions so that the right measures to deal with it come to be understood by the community. If and only if both conditions are met, collective action (U) can be undertaken. For example, a community may identify the situational event – say a drought – and perceive the need for collective action to be undertaken, but fail to identify the right measures – building irrigation channels – to succeed. Situational changes may or may not persist (P), resulting in either weak socio-political developments (W), or in collective action success (S). It is important to note that these decision-making processes based on input information do not entail fully rational or fully informed decision makers such as would have been presupposed for the Homo economicus of classical economics. We should rather think of biases and bounded rationality to be inherently part of every decision. Undertaking collective action and completing the cycle does not necessarily lead to the development of social complexity. However, when a social group or community is repeatedly successful in managing or overcoming serious situational changes, probabilistic selection pressures may occur under a specified set of conditions, yielding a long-term (slow) probabilistic accrual (or loss) of emergent socio-political complexity and development (A). It has been argued that human-environment relations, used here in the sense of the input signals for the decision-making process and its eventual outcomes, are characterised by positive feedback loops of problem solving, information processing capacity, and new situational challenges (Figure 3.4) (van der Leeuw, 2007). These positive feedback loops reinforce the process of complexity trajectories initiated by the initial set of selection pressures. As a result, variable pathways of development – also depending on specific situational events – will develop and result in different outcomes that are inherently context specific. One element of the loop in Figure 3.4 is left to discuss, that is the lynchpin at the top of the cycle: mechanisms of problem solving and decision making. These will be discussed in more detail in the next part. The algorithmic approach outlined here shows how complexity can accrue through the iterated superposition of decision-making strategies, rules, and other collective action measures. It should be stressed, however, that a progressive increase of complexity, should by no means be considered unavoidable. At every step of the process, the possibility of failure exists. Sustaining collective action over time becomes ever more difficult as

Conceptualising social complexity 85

Figure 3.4  Positive feedback loops of information processing and problem solving (adapted by author from van der Leeuw, 2007)

it increasingly requires a highly interconnected system of flows of people, materials, energy, and information (Blanton and Fargher, 2016, p. 38). Two points can be raised to conclude this part. If complexity is indeed difficult to maintain, why does history then seem to point towards a trend of increasing complexity over time (Morris, 2013; Turchin et al., 2018a). Second, if complexity can indeed be considered a problem-solving mechanism emerging from basic information processing and decision-making strategies, why does not every society grow consistently in scale and complexity (Ortman, 2019, p. 185)? Both observations are partially related. The first thing to note is the role of preservation bias in the historical sample. One consequence of the view of complexity as a problem-solving tool is that we would expect societies to grow more complex over time, and more complex societies to stay in existence longer. From a statistical point of view, societies that exist longer have a bigger chance to leave behind material traces for us to find and study, thus skewing the distribution of material evidence towards more complex societies. Additionally, more complex societies tend to express their organisational structures in more durable and elaborate ways. One obvious example is the construction of monumental architecture. In the previous chapter, I already briefly touched upon how scaling theory has shown that economies of scale can provide strong benefits to societies.

86  Conceptualising social complexity Yet, despite these intrinsic benefits, not all societies inevitably initiated trajectories of increasing complexity. First, it should be stressed that even though complexity formation can be considered a problem-solving process, it does not come without (energetic) costs. I will return to this point later on in this chapter. For now, it suffices to say that as societies continue to implement changes to situational events over the long term, these measures will trigger future challenges which, in turn, need to be dealt with. Over time, a society builds an “ever denser scaffolding structure of related and interacting institutions” (van der Leeuw, 2016, p. 168). This process of problem solving inherently carries the risk of focussing on frequently occurring challenges over rare ones, increasing the dangers of unknown challenges over time. This could result in a situation where a society does not have the required expertise to deal with such challenges or its consequences, provoking drastic changes which can be generally be subsumed under the moniker of a “tipping point” or societal transformation. Complexity mechanisms In the previous part, I discussed selection pressures as drivers of complexity trajectories and collective decision making as its main modus operandi. The latter works through iterated loops of information processing and problem solving. The missing piece to operationalise these loops in the current framework is the aspect of complexity mechanisms. As a starting point, I want to highlight an approach to measure ‘subjective’ or relative complexity. In this approach, the degree of complexity development always depends on the available frames of reference, starting from a ‘reference simplicity’. A given society can only be considered ‘complex’ compared to another society, which may in turn be considered ‘simple’, of course without endowing any moral connotations to any such labels. To measure the development of complexity, we can use a simple equation by Efatmaneshnik and Ryan (2016): K ( S ) = F (µ( S ) ⋅ D( SR )) In this equation, the complexity K of system S is a function of input μ, the size of the minimal description of a given context and D, a distance function. Complexity formation can then be considered some form of ‘distance’ or amount of system change compared to an initial input value. However, not every form of societal change necessarily constitutes the development of social complexity. To approximate changes in complexity trajectories, we therefore need to have some way of defining an ‘initial’ state, as well as approximating the amount of change along relevant parameters of development. Here, I consider social complexity as the degree to which societies can capture, process and transform flows of energy, resources, and information

Conceptualising social complexity 87 in response to stimuli, challenges and opportunities. Any approximation of complexity trajectories therefore needs to incorporate how a given system impacts these flows and associated changes over time. In this part, I will focus particularly on mechanisms of complexity formation as the connecting element between input processes and outcomes. I highlight three key mechanisms: diversification, intensification and integration. Diversification induces differentiation within and between systems. Integration connects system components and produces interactions to generate emergent behaviour. Finally, intensification refers to the (relative) investment of energy and resources in a given process. Diversification as a property of human societies has been considered extensively from an evolutionary perspective. Early approaches to construct diversity measures were developed mainly for comparison of evolutionary perspectives on cultural change, and therefore strongly borrowed input from earlier advances made in ecology to quantify biodiversity (Bobrowsky and Ball, 1989). Kenneth Bausch (2001) distinguished three fundamental tenets of ‘differentiation theory’: (1) differentiation is the master trend of societal evolution; (2) it is directed by societal needs; and (3) it increases adaptation, generality and inclusivity of a society. Diversity is thus an essential element of conceptualising system changes over time. Diversification by itself does not necessarily equal increasing complexity. The behaviour of components must be integrated in order to function as a complex system. Structural organisation limits and channels behaviour, granting a dimension of behavioural predictability and internal system coherence. Complexity thus occurs when different components become connected, start to interact and generate novel information which determines further system dynamics. What makes complex systems truly ‘tick’ are the connections between the constituent components of social systems, such as individuals, groups, and institutions. As more and more system components become interconnected, an increasingly smaller number of key productive strategies start to become strongly interdependent and draw in energy or labour investment. Processes of intensification and specialisation induce increased efficiency, and process optimisation characteristic for transitionary complexity phases. Formalising this general structure gives us a model of: input information (I), selection pressures (X), complexity mechanisms (M), and output (Y). This output can then feature as (part of) new input I, creating a recursive loop of system dynamics as was elucidated in the previous part: Y ← (X) ∧(X | I) ∧ (M | X) The angular brackets indicate that the conjunction of events is ordered from left to right. X can be considered an element of a given system state, originating out of a combination of I from prior system outcomes and external

88  Conceptualising social complexity stimuli. Information is then evaluated according to decision-making processes based on internalised knowledge in the system in accordance with selection pressure X and transformed into a new system response Y through a mechanism M (Daems, In Press). In this standard model, complexity formation occurs through decisionmaking processes in response to selection pressures, using complexity mechanisms to obtain an outcome (which may or may not be the outcome desired in the decision-making process). This structure is of course a simplification given that complex systems are driven by multiple causal factors producing system outcome in non-linear ways. In practice, we would have to consider the synergy between multiple selection pressures acting upon different mechanisms and producing a range of outcomes. Still, the simplified representation helps us to make sense of the different components of system dynamics and the nature of their interrelations. Following the conceptualisation of complexity as a function of a given input value and a distance measure, it can be stated that the complexity distance is generated by the interplay between differentiation in structural organisation, growing connectivity between differentiated components in a structured whole and the intensity of energy investment in each component. Push-pull forces In the previous parts, I outlined how social complexity trajectories can emerge from selection pressures inducing a canonical loop of information processing and decision making, using mechanisms of complexity formation that result in collective action measures. The model I have outlined up until now, however, is still incomplete. What is lacking is how these processes influence flows of energy, resources, and information. In this part, I will introduce multi-scalar pushing and pulling forces that, respectively, concentrate and dissipate energy, resources and information at a given scale in a given social unit. Push-pull dynamics have been mainly used as explanatory factors for population nucleation by migration, population aggregation and other demographic processes (Kohler and Sebastian, 1996; Leonard and Reed, 1993; Zimmermann, 1996). Adler et al. (1996) consider push and pull dynamics as, respectively, exogenous and endogenous drivers of population aggregation. Here, I will apply a broader definition. Generally speaking, push-pull dynamics are those factors influencing organisational structures (Turner, 2003, p. 5). More specifically, I define push-pull dynamics as forces operating on various levels and domains, in and between social units, that provide stimuli for the creation, development and disbandment of organisational structures through the aggregation/dissipation of flows of information, capital, people, and resources. Pull dynamics are those processes influencing the aggregation of information, capital, people, and resources,

Conceptualising social complexity 89 thus contributing to community formation and complexity development. Push dynamics are all forces resulting in the disbandment of any such structures and processes. Even though specific forces have a primary role as either ‘pushing’ or ‘pulling’, depending on the perspective or scale of analysis, the same parameter can be said to both ‘push’ and ‘pull’ in a given process. For example, when focusing on locational benefits on the level of the landscape – such as the pres­ ence of resources – locations with high potential may be said to “pull” people in, whereas if the focus would be on optimal foraging decision making, individual agents may be “pushed” towards different locations depending on the parameters of their decision making. The effects of push-pull dynamics are scale-specific and need to be analysed on the relevant level. Michael Smith (2014) compiled a set of driving forces commonly identified in archaeological literature to explain the nucleation of people in settlements. These include: (1) defence; (2) political administration; (3)  institutional forces; (4) economies of scale; (5) labour pooling; and (6)  central place functions. Additionally, he points out other drivers such as population growth, environmental change, and colonisation. These drivers can all be considered pulling forces for settlement aggregation. We can identify three levels of interaction of interest here: micro-scale dynamics on an intra-community level, the meso-scale of inter-­c ommunity interactions, and the macro-scale of polities. Earlier, I already highlighted eight main types of selection pressures for complexity formation (subsistence, cooperation, competition, interaction, distribution, production, governance, and demography). These can also be considered the general categories for push-pull dynamics affecting social systems. It is not my intention here to provide a comprehensive overview of all potential push-pull dynamics or discuss particular drivers in detail. I will merely offer a handful of examples to illustrate the process on multiple scales (Table 3.1). Table 3.1  Push-pull dynamics on multiple scales (made by author) Selection pressures

Process

Scale

Primary role

Subsistence Competition Production Production Subsistence Interaction Competition Cooperation Governance Competition Governance

Resource availability Security needs Labour opportunities Production intensification Land availability Scalar stress Group fission Group fusion Central place formation Peer-polity interaction Polity power structures

Micro Micro Micro Micro Micro Micro Micro Meso Meso Macro Macro

Pull Pull Pull Pull Push Push Push Pull Pull Push Pull

90  Conceptualising social complexity Let us start with the level of people and communities on a local and (micro-)regional scale. This level constitutes the locational patterns of settlements in the landscape and the movement of people, resources, and information between them. Some examples of micro-level forces are environmental drivers such as resource availability, or social drivers such as the need for defence and labour opportunities. Multiple selection pressures are usually combined to give rise to locational patterns. Once a settlement is established, it can act as a node in the landscape, pulling in people and resources from the outside. One of the most prominent examples is the city acting as an attraction pole for rural-to-urban migration. Communities cannot keep on drawing in people and grow indefinitely. Earlier, I already discussed how cognitive limits to information processing constrain group sizes. Groups can overcome some of these limits by developing organisational structures and create information environments. Still, push factors always remain constraining factors on community development. Cognitive limits are important push factors in their own right. As socials groups grow larger, communication loads cause decreasing consensus in decision-making processes. This phenomenon has been described as ‘scalar stress’, a term first coined by archaeologist Gregory Johnson (1982) to denote intra-group conflicts as a result of increasing group sizes. Others have used the terms social stress (Düring, 2013), communication stress (Fletcher, 1995; Meier, 1972), or density-dependent conflict (Birch, 2013a) to essentially denote the same process. As more people aggregate in higher densities in nucleated settlements, activities and communications become more densely interconnected and interaction interferences can occur. Earlier, I already discussed how scalar stress can be overcome through the development of information transmission and uncertainty reduction mechanisms as part of the establishment of an information environment. The development of such an information environment provides pathways for signal consistency and higher fidelity of information during transmission (van der Leeuw, 1981). A second mechanism to deal with scalar stress and intra-group conflict is fission. Group fission can occur in social groups growing in size without developing proper higher-level integrative institutions capable of mitigating scale-related social stresses. The process generates two or more smaller units of manageable size (Bandy, 2004). Usually, one of the resultant groups retains its position, whereas the other(s) resettle elsewhere. As fission processes are repeated and the (regional) landscape starts to fill up – often related to patterns of regional population growth – the costs of fission and relocation rise accordingly given that it will either involve conflicts with existing groups on nearby lands or relocation to progressively more distant lands (Bandy, 2004). At some point, the costs of fission and relocation will outweigh the constraints exercised by the desire to retain face-to-face societal structures in favour of additional integrative mechanisms that allow social groups to expand in size. This might shift the group towards other attractor states requiring larger group sizes such as the desire

Conceptualising social complexity 91 for endogamy within the own community (Wobst, 1976, 1974). The integrative mechanisms needed to enact such changes could consist of both hierarchical (vertical) and heterarchical (horizontal) organisational structures such as, respectively, the establishment of a chief or ruling group, or the subdivision of a settlement into neighbourhoods. In addition to the development of intra-group integrative mechanisms, inter-group conflicts between adjacent polities would have increased as well as the (regional) landscape started to fill up and land availability decreased. Overcoming such conflicts need not necessarily imply (violent) conflict as tensions might be solved through several strategies, including not only direct conquest, but also intimidation, alliances, religious legitimisation, and marriage. As certain groups succeed in developing integrative mechanisms, their increased size could allow them to establish themselves as powerful pulling forces in local and regional landscapes. This adds a new dimension to inter-group conflicts, where some groups will establish structures of control over others. Different trajectories are possible. A powerful group can establish dominance over a series of dependent groups, or it can merge smaller units into its own social structures. The latter case is denoted as group fusion, giving rise to intricate patterns of subsequent fission-fusion dynamics (Griffin, 2011). Out of these fission-fusion dynamics emerges a differential pattern of community structures. Environmental circumstances, selection pressures and successful collective decision making can result in certain communities obtaining positions of comparatively greater power and importance in local and regional settlement networks. These can be considered central places that act as pulling forces for people, resources, and information from a wider environment by exerting specific functions within the wider settlement pattern. Five main central place functions have been identified: (1) administration; (2) security; (3) craft and industry; (4) trade; and (5) cult (Gringmuth-Dallmer, 1996, p. 8). Going beyond a qualitative function-based definition of centrality, settlement nodes exert centrality in flows of energy, resources and information through each of the functions it performs within the wider settlement network (Meijers, 2007). Centrality can then be defined in terms of the scope of capturing quantitative degrees of interaction, rather than qualitative (presence/absence) functionalities. The framework of push-pull dynamics outlined so far has been limited to communities and networks of communities. This is of course only part of the picture as these dynamics also play out on macro-level social units such as chiefdoms, states, and empires. I will discuss these in more detail in the next part of this chapter. Here, I will briefly outline one parameter of push-pull dynamics on the macro level, that of peer-polity interaction (PPI). I use the term ‘polities’ here as a catch-all concept for the aforementioned macro-level socio-political configurations. PPI was first applied by Colin Renfrew to explain “the growth of sociopolitical systems and the emergence of cultural complexity” (Renfrew, 1986, p. 1).

92  Conceptualising social complexity It essentially entails the full range of interactions – including competition, imitation, and exchange of goods and information – between autonomous socio-political units located in relative spatial proximity that drive the growth and development of socio-political units (Daems, Accepted(a)). The last element listed in Table 3.1 entails polity power structures. This refers to the role of macro-level polities, including not only states and empires but also other types such as chiefdoms or tribes, in redefining socio-political configurations throughout their controlled territory. This impact can be established through political and economic incentives, mutual diplomacy, or the unilateral exertion of political authority. I will discuss the role of power structures in polity formation in more detail in one of the next parts and will therefore not elaborate on it here. To summarise, the model of social complexity trajectories presented here consists of selection pressures inducing a canonical loop of information processing and decision making, using mechanisms of complexity formation that produce outcomes with a pulling or pushing effect, respectively concentrating or dissipating the flows of energy, resources and information that produce social complexity. A basic scheme of this model is represented in Figure 3.5. In order to operationalise this basic scheme, we need to make the outcomes more specific and apply the model to particular processes. In the next part, I will discuss outcomes of complexity trajectories related

Figure 3.5  Basic model of complexity formation (made by author)

Conceptualising social complexity 93 to community formation and polity formation. I will then add the missing dimension of the energetic costs of complexity formation, and incorporate this in an overarching model of human-environment interactions.

Outcomes of complexity In this part, I will describe some of the outcomes of complexity trajectories emerging from the building blocks of complexity and developed by the complexity formation model built on selection pressures, decision-making processes, and push-pull dynamics outlined in the previous parts. So far, I have discussed the origin and development of social complexity, using general terms such as social groups, units, societies, and systems. The reason is that the model of complexity trajectories presented here is intended as a general model that can be applied in a wide range of cases. The specific context and structure of the social system inherently influences its complexity trajectory and can therefore not be excluded from analysis. I will therefore briefly detail – in general terms – the emergence and development of complexity in two specific types of social systems: communities and polities. For the former, I will particularly focus on communities as spatial and social aggregations of people, and their role as social reactors for interaction and information transmission. For the latter, I focus on the projecting of power structures and information environments across a wider territory as the main properties of polity formation. Finally, while complexity is generated by interaction and information transmission within social units, it cannot increase ad infinitum. I will therefore discuss the costs associated with social complexity to explain how the potential of energy capture sets the limitations to complexity development. Typically, the loss or decrease of social complexity is discussed in terms of societal collapse or transformation. I will reframe this discussion in terms of the capacities for information processing and energy capture of a complex social system as part of human-environment interactions. Communities as social reactors At the beginning of this chapter, I discussed information transmission in social groups and cognitive limits on information processing as important constraints on group sizes. To overcome these constraints and allow for group sizes to grow and social complexity to increase, social units develop practices and organisational structures and construct an information environment, allowing group members to offload cognitive stress into the environment. In this part, I will discuss one of the major loci for social interaction, information transmission and complexity formation, that of the community. Communities entail the social plane upon which shared, day-to-day activities and interactions take place to form the main unit of socialisation

94  Conceptualising social complexity for a group of people (Smejda and Baumanova, 2015, p. 53). A seminal definition of community defines it as a group of people who reside together, engage in face-to-face interaction on a daily basis and derive a cohesive social identity through shared membership of a local social network (Murdock and Wilson, 1972). A community can be considered a general term for social collectives, covering a range of more specific configurations such as villages, towns, and cities. When referring to community dynamics, I denote general social processes that transcend the boundaries of those specific categories. Following the building blocks of social complexity outlined in the first part of the chapter, we can consider social interactions as the basic drivers of community formation. We can take as our basic axiom that, as more people come together, a higher amount of (potential) interactions can occur. As a result, the effects of those interactions increase according to the amount of people participating in them. Simply put, larger societies tend to be considered more complex than smaller ones. The core of the argument holds that larger group sizes lead to exponential increases in amounts of social interactions, in turn driving increased social complexity. The relationship scales super-linearly given that the number of possible pair-wise links between people increases faster than the absolute number of people.4 Bringing more people together can happen through population growth, nucleation (increasing density) or the influx of people (aggregation). These demographic drivers can be subsumed under the single moniker of ‘energised crowding’. The term was first coined by the architectural historian Spiro Kostof (1999) and has come to denote a process of increased faceto-face interaction among members of a community. Energised crowding acts as the nexus between demographic drivers on the one hand, and social outcomes on the other (Smith, 2019). These outcomes can be both negative (scalar stress) and positive (community formation and economic growth) (Figure 3.6). Let us discuss some of these outcomes in more detail. Looking at community formation through the lens of social complexity dynamics, the starting point is the intensity of interaction within a social group. Communities have been described as hubs of social interaction with multiplicative effects

Figure 3.6   Energised crowding as a generative driver of community formation (adapted by author from Smith, 2019)

Conceptualising social complexity 95 for social and economic processes, or, in other words, ‘social reactors’ (Bettencourt, 2013). As the amount of interactions grows, the effects of energised crowding increase in non-linear ways. The field of settlement scaling explores the regularities in the multiplicative effects of social interaction across settlements of different sizes (Bettencourt, 2013; Bettencourt et al., 2007; Bettencourt and West, 2010). I have already discussed some of the theoretical and mathematical tenets of scaling theory in the previous chapter. In this part, I will focus on the link between social interaction as a driver of energised crowding, information transmission and the role of communities as social reactors. Face-to-face interactions have been the main form of communication and information transmission in communities throughout most of human history, until the advent of modern electronic-based communication (Storper and Venables, 2004, p.  352). Information transmission is also the main driver behind the social and economic effects of social reactors. One particularly potent mode of information flow is social learning and the transfer of knowledge. As a higher concentration of people gathers, chances of having more knowledgeable agents nearby to stimulate knowledge ‘spill-overs’ increase accordingly. The multiplicative effects of the spatial concentration of economic activities is called agglomeration economies (Krugman, 1991; Scott and Storper, 2015; Storper, 2010). Social reactors emerge out of the diffusion of information and knowledge through social networks of people and institutions (Bettencourt et al., 2007). Positive feedback loops of interaction generate more than proportionate increases in the effects of social reactors. The resulting outcomes include most notably an intensification of interaction and higher socio-economic productivity, as well as the diversification of social and economic activities leading to increased interdependencies within an (urban) community (Bettencourt and West, 2010). The conceptualisation of communities as social reactors is based on their role as information intensifiers. The positive and negative effects of intensification can be defined as ‘energised crowding’. This process finds its clearest expression in the formation of urban places. Cities have been described as “the hubs of innovation, engines of wealth creation and centres of power, the magnets that attract creative individuals, and the stimulant for ideas, growth and innovation” (West, 2017, p. 8). Urban communities provide the stage for increased socio-economic interactions, mechanisms and processes that drive economic growth and other processes of exponential expansion. At the same time, the negative effects of urban life on the social and mental well-being of urban citizens have long been a dominant theme in the social sciences (Milgram, 1970; Simmel, 2012). Other examples that have been linked to differing degrees of urbanisation are the rise of poverty (Teitz and Chapple, 1998), crime (Glaeser and Sacerdote, 1999), and health problems (Winsborough, 1965). While the effects of energised crowding described here are mostly associated with urban communities, settlement scaling studies have shown that the

96  Conceptualising social complexity same quantitative relationships between population size and socio-­economic outcomes can be found in villages or other non-urban communities (Ortman et al. 2014; Ortman and Coffey, 2017). This suggests that village aggregation and urbanisation are actually expressions of similar underlying dynamics. The differences between urban and non-urban communities from the perspective of energised crowding must therefore be discussed on a quantitative rather than a qualitative basis (Ortman and Cofey, 2015). Villages often display similar features such as scalar stress and group fission (Alberti, 2014; Bandy, 2004) and develop mechanisms of community formation such as of social hierarchies (Birch, 2013b), or ritual activities to stimulate group integration (Froese et al., 2014). Conversely, social reactors intensify the dynamics of social interaction in communities and its effects increase as settlements grow in size. Group fusion can therefore be considered a major driver of communities as social reactors (Bandy, 2004; Birch, 2013b). While the continuum between villages and cities is theoretically valid and can be empirically approximated through settlement scaling approaches, in many cases not enough data is available or its resolution is insufficient to actually reconstruct continuous settlement systems. In these cases, applying discrete classes of communities can still be useful, even if only as a heuristic tool to capture discrete snapshots of change in continuous processes. Additionally, while I agree with the basic idea of a continuum between urban and non-urban5 communities, I still want to highlight that at least the nature of their information environments is qualitatively different. Many of the characteristics of an ‘urban transformation’ can be generally subsumed under the umbrella of social entropy reduction. The regular lay-out of street grids, the construction of public buildings – often given prominence in central positions within the community and being expressed in monumental architecture – and the emergence of public administration are all part of the process of building a structured community and settlement organisation that provides cues for people participating in social life. This first and foremost covers the own community, but, crucially, also an increasingly larger portion of outsiders. One crucial aspect of urban communities is their participation in wider networks of exchange in goods and information with other communities on a local, regional, and inter-regional scale. This is an essential part of central place formation discussed earlier. As communities develop into central places through the participation in such networks, people from outside the own community will increasingly visit the settlement. Earlier, I already mentioned the example of market development. Markets require fixed times and spaces and a set of institutional elements (such as weights, exchange currencies, sale procedures, etc.) to reduce uncertainty and streamline patterns of action and interaction. The general process of urban transformation entails an interrelated set of such uncertainty reduction processes in a wide range of domains – including social, political, economic, and religious spheres – collectively creating an information environment that

Conceptualising social complexity 97 is intelligible for people both internal and external to the own community. Rather than speaking of urbanisation per se, I would propose to call this transformation the shift from internalising to projecting information environments. In the next part, I will discuss a second outcome of social complexity trajectories, moving the higher scale of polity formation. I will argue that developing power structures in tandem with projecting information environments is a key conditio sine qua non for polities to emerge. Polity formation So far, I have focused on what can be considered the micro and meso level of complexity formation. I started off with covering the foundational building blocks – interactions, practices, and social structures – of social complexity. I then outlined a model of complexity formation based on selection pressures, collective decision-making mechanisms, and push-pull dynamics. The model shows how organisational structures emerge out of the superposition of the outcomes of a recursive loop of information transmission, processing, and decision making. I then outlined some of the outcomes of complexity formation on the level of communities as social reactors. In this part, I want to discuss the outcomes of complexity trajectories on the macro level through the lens of polity formation. In the previous chapter, I already discussed some of the literature on state formation as the pinnacle of social complexity trajectories from the perspective of evolutionary and Eurocentric discourses. An unequivocal focus on the state blinds out the wide variety of polity structures attested in the archaeological and historical record. Various terms are commonly used to denote specific forms of polities, including simple foraging societies, complex foraging societies, horticultural societies, extensive agricultural societies or intensive agricultural societies, tribes, chiefdoms, and states. These terms cover either the main form of subsistence strategy employed by a society or social group or its form of political structure. To avoid the biases of social evolutionary approaches, I will talk here only in general terms about groups or societies with particular social, political, and economic properties as polities. Their genesis and subsequent development are subsumed under the moniker of polity formation. Polity formation inherently entails the formation of internal social stratification. The model of decision making and complexity formation outlined earlier provides a formalised approach to link information input and processing to the development of social organisation, hierarchical structures, and collective action measures. Undertaking collective action involves the development of some form of (horizontal) cooperation or (vertical) leadership within social groups. For the former, suitable mechanisms need to be devised to ensure that the benefits of cooperative behaviour outweigh the costs. For the latter, potential leaders need to be available and willing to

98  Conceptualising social complexity take on their new role, whereas the group as a whole need to be convinced (either voluntarily or through coercion) to become followers by adhering their objectives and strategies to those of the potential leader. The various ways in which leadership and cooperation emerge in social groups, as well as the outcomes of these processes related to the development of social and political organisational structures, is the subject of collective action studies in fields such as archaeology, political philosophy, cultural anthropology, and sociology. I consider here mainly the ‘proximate’ drivers of cooperation, rather than the ‘ultimate’ sources of prosocial behaviour pursued by evolutionary psychologists (Blanton and Fargher, 2016, p. 4). The latter argue that the emergence of prosocial behaviour, rewarding cooperative or altruistic behaviour and punishing non-cooperators, dates back at least to smallscale hunter-gatherer groups during the Pleistocene. Martin Nowak and Roger Highfield, for example, describe humans as “SuperCooperators”, arguing that social cooperation was a decisive factor in the evolutionary development of our species (Nowak and Highfield, 2011). The literature on the emergence of cooperation in (cultural) evolutionary studies is extensive and will not be discussed in detail here. Two major discourses in collective action approaches can be discerned: non-coercive cooperation aimed at mutual benefit and coercive cooperation developing out of the competitiveness of human nature.6 Cooperation can be defined as “actions that require individuals to incur some cost or risk associated with other individuals receiving a benefit”, whereas collective action problems can be seen as events where “the optimal strategy from the perspective of an individual differs from the optimal strategy viewed from the perspective of a group” (Carballo et al., 2014, p. 99). Additionally, two different kinds of collective action strategies can be distinguished. On the one hand, common-pool resource management aims at maintaining sustainable strategies of resource exploitation (in contrast to the classic notion of ‘tragedy of the commons’ driven by self-interested actors) such as for example herding strategies used on common grazing grounds (Ostrom, 1990). On the other hand, public goods systems are systems of limited membership developed to gain mutual benefit from collectively produced resources for members of the group only (Blanton and Fargher, 2016, pp. 30–31). Collective action forms the foundations of polity formation, but does not necessarily explain how it develops. A suitable starting point can be found in a theory on the genesis of political order and inequality proposed by Carles Boix. In his theory, Boix (2015) starts with the assumption of a blank slate initial state, with no formal political institutions, equal distribution of resources and simple technologies of production and war. Agents follow two strategies to survive: a productive strategy, allocating resources to production and exchange of goods and services, or a predatory strategy directing resources at appropriating the assets or returns of other individuals. Starting from this initial state, he considers technological innovation as

Conceptualising social complexity 99 the primary engine of change. Interestingly, technological change is defined as taking place when “individuals solve problems and attempt to master their environment, interacted with climate and geography to modify the economic (and military) endowment of individuals and to generate growing economic differences across individuals” (Boix, 2015, p. 7). Boix then traced under which conditions spontaneous cooperation emerges or breaks down to give rise to exploitative strategies. His model of evolutionary game theory states that social cooperation based on informal rules and personal interactions is only sustained if relative equality between agents is maintained. In all other scenario’s, those individuals with a competitive advantage on a military level or who are at an economic disadvantage, will have incentives to develop exploitative strategies. This suggests that social systems built on social inequality are far more likely than the continuation of a pre-polity state of sustained cooperation. This is corroborated by a recent study on the origin of wealth inequality in 1990 Western Eurasian sites, which suggested that inequalities emerge spontaneously following the transformation of agricultural societies from a labour-limited to a land-limited form (Bogaard et al., 2019). Building on the idea of clumped resources as the basis of dominance hierarchies – derived from the field of behavioural ecology – they argue that the transmission of existing wealth disparities from one generation to the next exceeds the levelling power of productive capabilities. Over time, this will result in accumulated increases in wealth disparity and inequality. Another research group led by Tim Kohler (2017) likewise argued that small initial disparities in wealth and resource availability resulted in extensive social inequalities, even though they actually stressed the multiplier effect of human labour over time. Both studies essentially highlight that small differences in initial conditions are exacerbated by positive feedback loops that result in pathways of development characterised by entrenched structures of social inequality. Once inequality becomes entrenched in a social group, systematic conflict is unavoidable unless an overarching structure is created to maintain peace and internal cohesion. Earlier in this chapter, I discussed how hierarchies and other social structures emerge in growing social groups. Boix (2015) highlights technological innovation, particularly related to production and warfare, as the main driver of polity formation out of such groups. A recent study showed a strong covariation among a number of generative factors – termed coactive causal processes – such as collective action, commercialisation, increased material standard of living, degree of urbanisation, population growth and production intensification (Blanton and Fargher, 2016, p. 254). Yet, finding covariance between these traits does not necessarily provide an adequate answer. On the contrary, the argument begs the question as these supposed causal drivers are also used as the defining traits for the outcomes of the process. Instead, I argue here that the key properties of polity formation are the projection of power structures across a wider territory beyond the own

100  Conceptualising social complexity community and the integration of this territory into the existing information environment of that community. Archaeology has a long tradition of settlement and landscape studies that has focused extensively on the relationship between a community and its hinterland. The concept of Siedlungskammers or ‘settlement chambers’ originated in the 19th century German tradition of historical geography to denote micro-environments defined by natural boundaries (e.g. rivers or mountains) covering a small geographical area with sufficient resources to sustain a nucleated community (Bintliff 2009). The extension of power structures across a natural environment allowed a polity to gain control over necessary resources and energetic potential. The extent of this natural environment is not static. Population size may grow or decrease, resulting in a need for more or less resources. At the same time, a range of internal dynamics such as elite competition may result in a polity wanting to expand its territorial claims. Expanding polities will, at some point in their trajectory of development, need to deal with adjacent polities laying claims to lands of their own. Once these polities are pushing for expansion towards mutual boundaries, interpolity conflicts may arise. The American sociologist Randall Collins developed a model of geopolitics to explain inter-polity conflict and territorial change (Collins, 1995). His model was originally developed to explain the collapse of the Soviet Union, but has been applied to historical polity formation as well. The model includes territory size and associated geopolitical resources, logistical loads (the costs related to projecting military power from the home base) and marchland position (the amount of conflict boundary zones a polity faces). In a mathematical implementation of this model, Peter Turchin (2003) demonstrated that positive feedback between territory and geopolitical resources, combined with negative feedback by logistical loads results in first-order dynamics and metastability. He then extended this theory by incorporating a form of collective action ‘fitness’ and intra-group solidarity, ethnic assimilation and the presence of meta-ethnic frontiers to develop a mathematical model of state formation displaying second-order dynamics and sustained oscillations in polity cycles that match the empirical data. Polity expansion first and foremost entails the integration of a larger spatial area into existing networks of flows of energy, resources and information. For the whole system to continue functioning, flows need to be commensurate according to the position in the network (van der Leeuw, 2007, p. 219). This idea links back to central place theory as discussed earlier and centre-periphery approaches such as Wallerstein’s world system, in which higher densities of individuals ‘aligned’ with the polity – that is who share its perspectives, beliefs and ways of doing – can be found in the central areas, resulting in more rapid and effective information transmission. While the intensity of flows generally tapers off towards the peripheral areas, they still need to be sufficiently strong to maintain overall cohesiveness and willingness to cooperate.

Conceptualising social complexity 101 One recent study of polity growth redefined evolutionary trajectories of social complexity in informational terms (Shin et al., 2020). Based on PCA analysis of the dataset compiled by the Seshat Project7 – covering 285 polities across the world – they found that polity growth typically entails the development of additional layers of administration and information processing technologies. The analysis also indicated that the development of socio-political structures is first preceded by increases in the scale of the polity (in the sense of the aggregate variation in social complexity captured by the first principal component), followed by an improved information processing apparatus and economic system (for example an administrative system of records and standardised weights to facilitate exchange). Once these new socio-political structures are established, a renewed phase of scale increase can occur (Shin et al., 2020). This suggests that polity growth is constrained by thresholds in scale and information processing. Once a polity grows beyond a certain scale, additional information processing structures are needed to maintain internal cohesion among all areas of the polity and avoid social breakdown. In turn, as innovations in efficient information processing and transmission develop, an information threshold is crossed which allows for additional growth. The interplay between these thresholds results in a discontinuous trajectory of growth, consisting of slow accumulation of innovations in information processing and transmission, punctuated by episodes of transformative change. Polity growth is not infinite. It is inherently capped by constraints in the capture and consumption of energy and resources to sustain its dynamics. Even today, we have not yet reached the energetic boundaries of our planet (even though we seem to be trying our hardest to get there). Most polities in history were not able to transcend the practical constraints on the transmission of energy and resources necessary to sustain their growth. It can therefore be stated that as limits to information transmission defines the boundaries of potential complexity formation, energetic costs of complexity sets the boundaries for actualising complex formation. In the next part, I will relate the limits on energy capture to the collapse and transformation of complexity in societies and human-environment interactions. Costs of complexity: Collapse and transformation Earlier in this chapter, I conceptualised social complexity as driven by the outcomes of iterated loops of information input, processing and decision making. In this sense, complexity can be defined as a problem-solving tool (Tainter, 2006, 1996). ‘Problems’ can be considered in the general sense of ‘situational changes’ covering all internal and external stimuli and challenges that a social group encounters throughout its existence. With every iteration of the decision-making loop, subsequent strategies, and solutions become superimposed, eventually generating a costly apparatus consisting

102  Conceptualising social complexity of multiple, partially overlapping structures of administration, laws, and measures of socio-political organisation, as well as an intricate set of social norms, values, and various avenues of communication between people, social groups, and central administration, all of which are costly to maintain. Each iteration, even when successful, requires more effort and energy to be invested to maintain the existing measures of socio-political development along with developing new ones. This energetic pressure will often induce additional situational events, entailing qualitatively new risks, some of which involve a larger spatial scale and/or longer time frame which pushes the system along its trajectory of development, often at an increasing speed (van der Leeuw, 2007, pp. 214–215). This again requires even more measures to be undertaken and continues driving complexity development. Social complexity can be considered a costly development, whose maintenance requires constantly renewed energy input in the form of expenditure of labour, money and/or time. As complexity formation is a recursive process, there is a characteristic flow among decision-making loops where societies generally tend to first use simple and cost-effective efforts with high returns. As iterations continue, solutions to maintain societal structures become more complex and costly, with diminishing proportionate marginal returns upon investment (Tainter 1996). This general phenomenon of diminishing returns means that complexity always has an increasing cost and recursive loops of development cannot go on forever without additional energy being brought in. In a sense, every iteration of socio-political development increases the possibility of failure or stasis. Episodes of developmental failure have generally been interpreted as ‘societal collapse’, which can be defined as a process of rapid simplification and loss of an established level of social, political, or economic complexity. Whereas collapse is sometimes ascribed to a single dominant factor, usually a set of causal factors is highlighted as inducing the widespread decline of social, political, and economic complexity. These factors can be both internal (e.g. civil strife or unrest, moral decline) and external (e.g. warfare, environmental depredation). More recently, concepts underscoring the endurance of social configurations, such as resilience, sustainability, reorganisation, and transformation, have come to the fore. In a recent volume edited by Ronald Faulseit (2016), contributors sought to reassess traditional conceptualisations of collapse through the lens of resilience and societal transformation to capture the full extent of possible outcomes of transitionary phases. Collapse was defined in this volume as a rapid (over a few generations) decline in socio-political complexity or the demise of a particular political system (Faulseit 2016, p. 5). Resilience on the other hand, was considered in a broader sense as the vulnerability of a particular unit and its ability to adapt to, cope with, or transform when facing both acute and chronic stresses. It was also noted that resilience is not about “stability around a single state, but rather the possibility of multiple socioecological states that maintain the primary

Conceptualising social complexity 103 functional relationships of the socioecological system” (Redman et al., 2007, p. 118). Societal collapse is often associated with dramatic and irreversible effects on cultural continuity, socio-political complexity, and economic patterns of production and distribution. This approach has recently been reassessed through studies of long-term socio-political resilience and transformation, which integrated phases of collapse and reorganisation into models of cyclical or episodic trajectories of complex social systems. A model of spatial-temporal dynamics in early complex polities links the cyclical nature of polity building and collapse with a the concept of self-organised criticality (SOC) in fission-fusion dynamics driven by population growth and spatial expansion (Griffin, 2011). The model output replicates the cyclical development of hierarchical settlement patterns and is validated by empirical rank-size data of settlement patterns from Mesoamerica. Griffin outlines four main micro-level drivers of development: (1) population growth and migration; (2) allocation of resources between general population and ruling elites; (3) internal competition between rival factions; and (4) external competition between polities. The cyclical recurrence of system dynamics suggests continuity in the underlying drivers, even after apparent episodes of collapse. While we might intuitively expect that the socio-political and economic configurations of polities in subsequent cycles are markedly different, we also often see remarkable similarities and continuity. In these cases, the system revolved around an attractor state within the same basin of attraction. In the next part, I will conclude the reframing of the complexity formation debate in terms of energetic constraints induced by humanenvironment interactions through the concept of social metabolism. Complexity, energy, and human-environment interactions So far, I have discussed the development of social complexity out of selection pressures inducing iterated loops of signal detection, information processing, and problem solving through complexity mechanisms, as well as the pushing and pulling powers influencing the accumulation and dissipation of flows of energy, resources and information. The key thread throughout all of these elements is the role of information transmission and uncertainty reduction in social complexity formation. At this point, I will extend the conceptualisation of societies as information processing units to incorporate the flows of energy and resources between societies and their natural environment. Like all complex systems, human societies need to derive energy from their environment to maintain and develop their organisational structures. To sustain social life, flows of energy and resources need to be exploited from nature and redirected towards society. Axiomatically, every society relies to some extent on its environment to derive the necessary energy and resources. To look at how societies generate the necessary energy to develop and maintain its external structures, how they apply various response

104  Conceptualising social complexity strategies to environmental and social stress factors, and how some strategies result in potential system transformations and regime shifts, we must look beyond the limits of a social unit as bounded or isolated. Societies never exist in a vacuum but operate within, and are affected by, the dynamics, cycles, and pulses of the ecological context through positive and negative feedback loops (Ostrom, 2009). Studies of human-environment interactions have traditionally been based on conceptualisations of ecosystems in equilibrium states, where change is exceptional and considered as ‘noise’ that must be analytically suppressed (Holling et al., 2002a). It has been increasingly realised that moving beyond the limitations of equilibrium-based approaches requires shifting towards more fluid, dynamic, and non-equilibrium (or at least multiple equilibria) perspectives (Schoon and van der Leeuw, 2015). Already in the 1970s, resilience thinking emerged as a counter narrative against prevalent equilibrium-based models of ecosystem dynamics (Cote and Nightingale, 2012, pp. 476–478). In a seminal paper, Holling (1973, p. 14) advanced the concept of ecological resilience as the capacity of systems to absorb disturbances while retaining its main population or state variables. In other words, the ability of a system to remain organised around a set of processes, structures, and functions without losing its identity or continuity. This is a different view compared to the more traditional engineering approach, which assumes a single steady-state and defines resilience as the time needed for the system to return to equilibrium after it experienced a disturbance. Resilience has also proven to be a popular concept in archaeology as archaeologists increasingly seek to participate in debates with wider contemporary relevance regarding sustainable development, long-term dynamics in coupled human-environment systems, and response options to environmental challenges (Barton et al., 2012; Redman and Kinzig, 2003; Redman, 2005; Schoon and van der Leeuw, 2015; van der Leeuw and Redman, 2002). More recently, human-environment interactions have been commonly studied through the framework of social-ecological systems (SES) as the complex coupling of social and ecological spheres (Berkes et al., 1998; Ostrom, 2009; Schoon and van der Leeuw, 2015). A SES can be considered a coherent bio- and geo-physical unit linked with social actors and institutions though structures of resource appropriation and governance. This approach particularly focuses on identifying key variables that could affect long-term sustainability of these social-ecological systems and allow for cross-case comparisons (Schoon and van der Leeuw, 2015, p. 170). Since the 1990s, the concept of social metabolism has gained considerable attention in human-environment studies as a framework to trace, analyse and interpret the exchange of energy and resources between nature and society. Social metabolism can be generally defined as “the particular form in which societies establish and maintain their material input from, and output to, nature and as the way they organize the exchange of matter and energy

Conceptualising social complexity 105 with their natural environment” (Fischer-Kowalski and Haberl, 1997). From a metabolic perspective, human societies must be considered inherently embedded in nature, as they affect, and are affected by, the dynamics, cycles, and pulses of their ecological environment through relationships of exchange of energy, materials, and information (de Molina and Toledo, 2014, p. 22). The concept has been used both as a theoretical framework for explaining socio-environmental change and as a set of methodological tools to analyse specific flows of biophysical behaviour (Weisz, 2007). While metabolic approaches were first developed in biology and ecology, they were soon applied to social collectives such as cities and production systems (Wolman, 1965; Ayres and Kneese, 1969; Meadows et al., 1972). The collective metabolism of a social group minimally equals the sum of the biological metabolisms of its individual members, in addition to the requirements of maintaining the social organisation of the group and all other extra-somatic energetic needs. Broadly defined, social metabolism entails the entirety of biophysical exchanges in matter and energy between society and nature (de Molina and Toledo, 2014). Three types of material and energy flows can be distinguished – input flows, intra-system flows, and output flows – subdivided in five functions: appropriation, circulation, transformation, consumption and excretion (Figure 3.7). These metabolic functions can operate on two levels: individual or biological and collective or social. For example, appropriation at the individual level entails human beings extracting oxygen, water, and biomass from nature to survive. At the social level, a collective unit of individuals connected through social

Figure 3.7  Flows of energy and resources between society and nature in a social metabolism model (made by author)

106  Conceptualising social complexity relations (for example a family, company, or community) also extracts matter and energy from nature to ensure maintenance and reproduction. The distinction between the individual and collective level of energy appropriation can be extended towards all phases of the metabolic process. The division between individual and collective metabolism corresponds to a distinction made by the American biophysicist Alfred Lotka (1956), between endosomatic use of energy in nutrition (bio-metabolism) and the exosomatic use of energy by tools (techno-metabolism). It has been argued that the flow of endosomatic metabolism remains fairly constant in time and is directly related to population size, whereas exosomatic metabolism is far more variable (Giampietro et al., 2011, p. 187). To approximate exosomatic metabolism, a contextualised analysis of socially determined practices is needed, as the ways human beings are organised in society will determine the ways in which they affect, transform, and appropriate nature, which in turn conditions the ways in which societies are configured (de Molina and Toledo, 2014, p. 60). Given that endosomatic energy needs per capita generally remain stable, the development of social organisation can only take place through the expansion of socio-metabolism beyond the addition of the bio-metabolisms of all its members, or in other words, through an expansion of exosomatic energy dissipation. In this way, the ratio between endosomatic and exosomatic energy has been used as an indicator of the level of material complexity of societies (Giampietro, 2003). For the past such calculations can only happen on a very coarse-grained level due to the limitations in data availability and resolution. One of the main functions of social systems is to store energy, among others in the form of food reserves, money, social and institutional structures, and other capital and assets. Social systems ‘work’ by converting external energy derived from the environment into internal exergy. Complex societal systems can generally apply four different strategies of exergy use: (1) storage for future use; (2) system maintenance; (3) buffering; or (4) luxury consumption, that is exergy consumption not leading to one of the other outcomes (Muys, 2013, p. 43). In an evolutionary perspective, the latter is generally expected to be eliminated by selection pressures in times of crisis or system disturbance. Larger and more complex societies would then require more efficient ways of buffering external and internal disturbances and thus more prosperity and higher standards of living. As a society grows more complex, it tends to increasingly struggle with keeping up the pace due to the diminishing returns on investment and higher energetic requirements to maintain its socio-political structure. The increased dissipation of stored energy by social systems to increase complexity in response to disturbance events is of crucial importance for understanding the properties of resilience in these systems. It has also been noted that surplus energy within a complex system is generally quickly dissipated and that human societies rarely had surplus energy (or exergy) readily

Conceptualising social complexity 107 available for exploitation (Tainter, 2016, p. 35, 2011, p. 90). Increasing complexity therefore does not always lead to greater energy flow within the system, as greater complexity also results in more costly economic and administrative structures (Tainter, 2006). Due to the process of diminishing returns on investment, societies are increasingly put under strain unless new sources of energy are found. Those exceptional events in human history where innovations occurred in the exploitation of new energy sources with enormous potential were of such primordial importance that they are colloquially known as revolutionary transitions such as the ‘Agricultural Revolution’ or ‘Industrial Revolution’. Following the view of complexity as a problem-solving tool presented earlier, any available energy would be quickly tied up with and exploited by new measures of collective action. Inexpensive energy exploitation measures generate increasing complexity at first, but through a combination of generated problems associated with these exploitative measures (for example waste production) and diminishing returns on investment setting in, solutions need to be developed to deal with these problems, intensifying energy capture but again increasing complexity (Tainter, 2016, p.  35). Complexity formation therefore generates higher degrees of energy exploitation but can only be sustained on the condition that energy is effectively transmitted and consumed to sustain societal dynamics. Constraints on social complexity development are therefore set, not by energy surpluses, but by available flows of energy fuelling the different stages of the social metabolism.

Multi-scalar dynamics of change and stability So far in this chapter, I have outlined the main components of a conceptual model of social complexity trajectories. Starting with the foundational building blocks of complexity, the model covers the main drivers of complexity formation such as selection pressures, decision-making mechanisms and push-pull dynamics. I then outlined a series of outcomes of complexity formation, focusing on communities as social reactors and polity formation through the projection of power structures and information environments. Finally, I extended the informational approach with a conceptualisation of flows of energy and resources as constraining factors of complexity formation. Combined, the different parts of this chapter have provided a series of theoretical and methodological supports to conceptualise social complexity as the outcome of flows of energy, resources and information. In the final part of this chapter, I will extend these strands of thought into a conceptual framework using the concepts of the adaptive cycle and panarchy. This framework is one of the key conceptual and methodological cornerstones of resilience theory and ecology, developed as a high-level theory of stability and change in socio-environmental systems. I will use

108  Conceptualising social complexity the adaptive cycle as a heuristic framework to describe and understand multi-scalar dynamics in social-ecological systems. Modelling change and stability: Adaptive cycles Drawing from ecology, economics, institutional theory, and complex systems theory, the adaptive cycle provides an integrative framework to trace change and stability in the dynamic behaviour of coupled socio-ecological systems (Gunderson and Holling, 2002). It is commonly used as the lynchpin for integrated analyses of socio-environmental systems across different temporal and spatial scales, bridging the gaps between archaeology, the geosciences and cultural anthropology (Widlok et al., 2012). The adaptive cycle has also been put forward as a catalyst for the implementation of resilience theory in archaeology, which would allow archaeologists to be more active participants in debates on contemporary issues such as climate change and sustainability (Redman and Kinzig, 2003). As a result, the potential of adaptive cycles as a conceptual framework in archaeology has gradually gained recognition.8 Applications cover a range of topics such as the formation and disbandment of human groups, economic systems, settlement patterns, material production, agricultural subsistence, trade networks, population movements, and more. The adaptive cycle (Figure 3.8) describes system dynamics in terms of potential for change, defining the range of possibilities for system development though accumulated potential, and connectedness between system variables to measure system resilience (Holling and Gunderson, 2002, pp. 32–33). System trajectories move along these three axes through four phases: exploitation (r), conservation (K), release (Ω), and reorganisation (α). The first two phases of the cycle – r and K – respectively, refer to rapid initial growth and the sustained conservation of the system. They are derived

Figure 3.8  Adaptive cycle (adapted by author from Gunderson and Holling, 2002)

Conceptualising social complexity 109 from r/K selection theory, originally developed in ecology (Pianka, 1970). With r-strategists, we denote explorative species quickly occupying ecological niches through extensive dispersal, rapid growth, and high reproduction rates. By contrast, K-strategists tend to have slower growth rates and lower reproduction rates. The Ω and α phases are derived from economics and refer to the release of an increasingly integrated system and the associated loss of accumulated potential, followed by the reorganisation of the system as it enters a new cycle. Dynamics in the cycle describe processes of episodic change in non-linear system dynamics. Flows move in a slow ‘front’ loop from r to K, inducing incremental changes and accumulation of potential within a basin of attraction, punctuated by a ‘back’ loop towards Ω, α, and back to r, consisting of punctuated episodes of rapid transformation that create emergent outcomes. The r-phase is one of rapid initial growth characterised by low connectivity between system components, high resilience, and quick accumulation of potential – in the sense of capital, resources, knowledge, social networks of cooperation, leadership and social trust – all available for the system to shift into new state phases or initiate new dynamics. In short, the construction of new niches for populations to originate and develop. Associated processes include rapid movement into uninhabited or sparsely populated landscapes, population growth, and development of new technologies and food acquisition strategies (van der Leeuw, 2007, p. 215). An example of an r-phase situation may be technological innovation resulting in opportunities as a new economic niche opens. Let us take for example the development of internet-based companies after the creation of the world wide web. Initially, a high number of start-up firms competed to gain part of the market share. As competitive processes played out, the number of competitors winded down until a stable configuration with a small number of strong companies was formed (that is, the start of the K-phase, see infra). The r-phase can be considered to be highly resilient because of the abundance of available resources, high level of diversity, flexibility and connectivity, resulting in a robust system configuration in the face of perturbations (Holling and Gunderson, 2002; Walker et al., 2006). As the r-phase develops into K, system dynamics slow down and start to conserve existing properties rather than explore new avenues of development. Potential continues to accumulate, albeit more slowly and tightly bound to existing structures rather than being freely available for innovation and system development. The shift from r to K strategists therefore entails a shift from adaptation to external variability towards control of variability where increased efficiency is sought by minimising costs and streamlining operations (Holling and Gunderson, 2002, p. 44). K-phase systems exhibit less room for innovation and entrepreneurship. Internal system components become increasingly interconnected as they become increasingly mutually dependent within self-organised clusters of

110  Conceptualising social complexity relationships, sometimes resulting in extremely high levels of integration or hyper-coherence, where an increasingly smaller number of key productive strategies start to solely depend on one another, resulting in intensification of production strategies. Within the K-phase there is also increasing specialisation, efficiency, and process optimisation resulting in increasingly narrow avenues of development where these strategies may generate a multiplier effect induced by increasing returns to scale (Arthur, 2009; Krugman, 1991). These are key processes in the generation and accumulation of production surpluses needed for complex societies to store capital and resources as buffer for future perturbations. However, because of these strategies of intensification, fewer resources remain available as most resources tend to get ‘locked up’ over time, meaning they are more tightly controlled and more expensive, for example because of material accumulation by developing elite control mechanisms (Aimers and Iannone, 2013, pp. 23–24; Davidson, 2010, p. 1139). In other words, the cost of ‘getting things done’ grows higher over time. Efficiency and optimising behaviour, although theoretically desirable, can therefore be problematic in practice because in being efficient – as in optimising their behaviour – people, communities, societies, and other organisations often eliminate redundancies and emphasize a specific range of values and interests, resulting in a more homogenous system in terms of functions and response diversity, which can result in a dramatic decline in flexibility and hence resilience (Aimers and Iannone, 2013, pp. 23–24; Hegmon et al., 2008). As a system becomes increasingly interconnected, more and more energy and resources is reserved for maintaining existing structures (that is, maintaining its functional resilience). As noted earlier, measures to exploit energy and resources from the environment to maintain the structural integrity of complex societies are subjected to diminishing returns on investment, therefore requiring putting in more energy over time to get the same return output. As a result, overall efficiency resilience of the system continuously decreases to maintain functional integrity. This shows that system resilience is never infinite, but rather some sort of trade-off exists between maintenance of efficiency and maintenance of function. In analogy to ecological resilience, the concept of social resilience has been defined as “the ability of groups or communities to cope with external stresses and disturbances as a result of social, political, and environmental change” (Adger, 2000, p. 347). Strategies of social groups trying to cope with disturbance events have been elucidated earlier in the discussion on canonical loops of complexity development in the previous part. However, as has become clear by now, the gradual development of socio-political complexity may indeed provide short-term solutions, but need not necessarily be effective on the long term as increasingly elaborate structures require ever more maintenance and rob the system of the necessary flexibility to deal with new challenges.

Conceptualising social complexity 111 Rigidity is often considered to be characterised by low heterogeneity and high connectivity between system elements (Carpenter and Brock, 2008). However, high connectivity could also have benefits of decreasing response time of the system to certain disturbance events by mobilising a high number of agents for collective action, such as for example when dealing with famine or flood events. High connectivity therefore does not inherently result in a more rigid system as it allows information to flow more freely within the system (Kidder et al., 2016, p. 75). Highly connected system components do, however, allow disturbances to propagate throughout the entire system, whereas a less connected system might have contained certain disturbances in particular system components. This suggests that the trade-off in efficiency and function maintenance in complex coupled socio-environmental systems is tightly related to trade-offs in diversity and connectivity. These processes of increased interlocking of system components lead into a pathway of development where a system finds it increasingly difficult to break out of a set pattern because of associated sunk costs, which refer to a situation in which agents put more effort into continuing existing investments rather than exploring new ones, resulting in a tendency to undermine innovation (Janssen et al., 2003; Janssen and Scheffer, 2004). This tendency, often unintentionally, may therefore actually result in the inverse of resilience and lead to what has been called a ‘rigidity trap’ (Hegmon et al., 2008). As agents within a system have locked themselves into a certain way of doing things, the system itself begins to exhibit a path dependency (van der Leeuw, 2007, p. 215). It thus becomes brittle in the face of perturbations. At some point, the system may become too rigid to be able to deal with unexpected disturbances, either internally or externally induced – as posited by the canonical loop of complexity development described earlier – and the potential bounded with the interconnected system components can suddenly be released and lost from the disbanding organisational structure as the system moves into the Ω-phase. This event of slow accumulation leading up to an event of rapid destruction has also been called a tipping point (Gladwell, 2000), creative destruction (Schumpeter, 1942) or critical transition (Scheffer et al., 2012), describing phases of system transformation, both through incremental changes moving towards threshold values such as in the canonical loop of complexity development as well as through major disturbance events leading to system collapse (Aimers and Iannone, 2013, pp. 26–27). Such tipping points have also been defined as the temporary incapacities of a society’s information processing system to deal with the dynamics it is involved in, as a result of the accumulation of unintended consequences of earlier actions (van der Leeuw, 2012). Without the continued existence of a strong structural framework stifling novelty, the system transitions into a reorganisation phase (α) characterised by strong innovation. In this phase, connectivity is at its lowest point, allowing surviving but uncoupled system components to be re-used in novel combinations induced by the remaining system potential of the previous

112  Conceptualising social complexity cycle. Interestingly, this matches Prigogine’s observations that even in complex systems which are running down to simpler forms of dispersed low levels of activity, a concentration of remaining energy into focal points can create new elaborate phenomena. As such pockets of energy and information remain available, the system reorganizes itself and a new cycle develops. In this sense, sudden events traditionally associated with full societal collapse can in fact often indeed be considered more accurately as societal transformation (Faulseit, 2016; Schwartz and Nichols, 2006). This phase is commonly associated with increased system diversity, population migration, innovation and rapid restructuring and can be generally subsumed under the moniker of a regime shift (Filatova et al., 2016). This new cycle may resemble its precursor as uncoupled system components become rearranged in a system configuration strongly resembling the previous cycle, that is be in the same ‘basin of attraction’, or it may have fundamentally new functional characteristics in a system that has multiple stable states. The adaptive cycle consists of a ‘front loop’, from r to K, with slow and incremental growth and accumulation of potential (resources, capital, knowledge), punctuated by an episodic ‘back loop’, from Ω to α, of rapid reorganisation and renewal. Each part of the cycle thus results in one of two important elements of complex systems dynamics: the maximisation of production and accumulation, and maximisation of innovation (Holling and Gunderson, 2002, p. 47). The three properties of change – potential, connectedness, and resilience – set the limits of the system to, respectively, potential development, degree of control, and vulnerability to disturbance events that exceed that control. It is important to stress, however, that not all systems need necessarily pass through the different phases of the cycle in the same order (Aimers and Iannone, 2013, pp. 26–27). An r-phase may jump directly into an α-phase if a given socioecological system cannot sustain existing levels of development or is faced with an unexpected situational event that induces an impact of such a magnitude that exceeds the level of control the existing system structures can cope with. In other instances, an α-phase may stimulate additional reorganisation as the system is unable to settle on a new suitable configuration. Systems in a K-phase may also shift directly into an α-phase, thus avoiding an Ω-phase release, for example, in a sudden shift from a totalitarian to democratic regime. The adaptive cycle framework conceptualises the dialectic between shifting and stable human-environment interactions among multiple scales resulting in discontinuous structures exhibiting flexible and adaptive behaviour (Holling et al., 2002a). Processes of change operate at different rates, spanning several orders of magnitude, both on a spatial and temporal level. A socio-ecological system, therefore, typically consists of multiple interlinked adaptive cycles. A hierarchical sequence of semiautonomous, interconnected cycles has been termed a panarchy (Gunderson and Holling, 2002).

Conceptualising social complexity 113 Multi-scalar interactions: Panarchy One of the core questions in scientific research, is what the appropriate scales of observation are to approximate the structures and processes that constitute the object of study. An isolated scale of analysis hardly ever provides a sufficient explanation, as either the effects of processes in complex systems inherently unfold across multiple scales, or its properties and dynamics are influenced by processes on higher and/or lower scales. To study complex systems, we therefore need to consider the triadic structure as a key property of hierarchically ordered scales, which states that three adjacent hierarchical levels need to be considered for a parsimonious and sufficient description of the behaviour of the middle level (Salthe, 1985; Wu, 2013). An individual adaptive cycle corresponds to one particular level with its own logic and operates at a characteristic periodicity and spatiality. However, such cycles never operate in isolation. Individual cycles are integrated in two ways, basically corresponding to hierarchical and heterarchical structures. On the one hand through the so-called panarchy, a nested hierarchy of semi-autonomous levels, not necessarily subjected to strict top-down sequences of authoritative control (Holling et al., 2002b, p. 72). On the other hand, cycles of comparable size on a similar level can be interconnected across different systems or panarchies. Each adaptive cycle represents a functionally distinct level moving at specific speeds within an integrated system. While theoretically spanning a continuum of resolution scales, cycles in a panarchy tend to cluster around a few dominant frequencies or ‘lumps’ in relationships of space and time (Holling and Gunderson, 2002, p. 26). It has been argued that these discontinuities in size distributions can arise endogenously due to dynamic instabilities in the system (Rosser, 2011). In the adaptive cycle framework, these instabilities are caused by the interactions between cycles and variables operating at different speeds, leading to fluctuations in system dynamics. The system dynamics described by the adaptive cycle and panarchy framework inherently move at an uneven speed. We already discussed how within individual cycles system flows move in a slow ‘front’ loop from r to K, and rapidly in the back loop towards Ω, α, and back to r. Additionally, individual cycles inherently move at different speeds, with small and fast cycles integrated in a single panarchy. In particular, cycles that operate at small scales generally move quickly, whereas those at the top of the nested hierarchy move more slowly (Holling et al., 2002b). Multi-scalar episodic changes are therefore caused by non-linear interactions between small, fast, and large slow cycles and variables. Larger cycles provide the inertia and stability that permit lower scale cycles to pass through release and reorganisation while maintaining similar functions, that is, staying within the same basin of attraction, thus allowing adaptive cycles at one level to be repeated in the same or similar cycles of system configuration through the process of ‘memory’. The role of

114  Conceptualising social complexity memory is strongest when the higher level cycle is in the conservation phase. Conversely, coordinated release at small and fast scales may, in cascading fashion, trigger release at larger scale cycles, especially if these are at that time in the K-phase characterised by low resilience, a process called ‘revolt’, precipitating potential shifts into new basins of attraction at large scales through a phase of creative destruction (Walker et al., 2006). Widespread system shifts can occur either because lower-level cycles are synchronised, either through tight interconnectedness in the K-phase or when entering the back loop of system reorganisation simultaneously, triggering transitions in other cycles. Collapses in one cycle may stimulate changes in other adaptive cycles, both larger and smaller in size, such as increased mechanisms of exploitation (through immigration and a resulting larger labour force), conservation (the adoption of more productive and/or sustainable agricultural practices), or reorganisation (transformations aimed at making the system more resilient) (Aimers and Iannone, 2013, pp. 26–27). It should be remembered that a phenomenon that is considered to be a phase transition at one scale of analysis may be considered a ‘state flip’ at another, for example when an overarching socio-political unit collapses or is superseded this need not necessarily impact life in individual communities in a significant way. To assess to what extent a state flip occurred rather than a major perturbation within a consistent phase trajectory, it may prove useful to look at whether the unit of analysis has maintained the structures, controls, and members that are considered essential to its identity, and on what intersection between temporal, spatial, and organisational scales these changes developed. To this end, relevant scales of analysis should be identified, as well as the connections between them. In a socio-ecological system, human and environmental components are interlocked in mutually reinforcing ways, although they each operate on a characteristic spatial scale with specific temporal periodicities. In a seminal contribution, Sander van der Leeuw and James McGlade (1997; summarised in van der Leeuw 2020, 263–286) proposed a model of urban development, starting from an axiomatic dynamic of rural village communities embedded in a natural landscape, which can prove useful at this point. The human component of such a system generally consists of few superimposed rhythms, whereas the environmental component consists of a complexly integrated spectrum of biological and ecological rhythms (van der Leeuw and McGlade, 1997, pp. 338–339). The faster dynamics of social units have thus been embedded in and locked on to slower environmental dynamics. This is not to say that the environment should be reduced to an inert background stage for social dynamics to envelop. However, one of the most prominent differences between human and environmental systems is that the former consists of (potentially) knowledgeable agents who can coordinate and adapt their behaviour according to flows of information and communication with other people. Social agents are self-reflective and goal-oriented, and able to make decisions to move the system towards a

Conceptualising social complexity 115 desired state (Redman, 2005, p. 74). As a result, they can act and react with a markedly higher speed compared to the natural environment, which operates with a seasonal periodicity and therefore acts as a stabiliser for social dynamics. Some have questioned the explanatory potential of the adaptive cycle and panarchy framework and doubt its usefulness as an empirical framework. For example, the adaptive cycle emphasises that change is neither continuous nor chaotic. Rather, it is episodic and consists of periods of slow accumulation of capital and structures punctuated by sudden releases and reorganisation (Holling and Gunderson, 2002, p. 26). The link with Stephen Gould’s concept of punctuated equilibrium is easily made (Gould, 1999). However, the expression of the theory in general conceptual terms resulting in seemingly superficial parallelism with other models is exactly what has caused the adaptive cycle to be criticised as an oversimplification of complex system dynamics and no more than a general metaphor for system change (Gotts, 2007). This criticism misses the point in two ways. First, metaphors are highly useful heuristic devices that can provide novel ways of seeing and result in new insights (Gray and Macready, 2019, p. 129). Their explanatory potential should therefore not be dismissed out of hand. Second, it can be argued that the adaptive cycle framework provides more than a metaphor of system change. Holling and Gunderson themselves conceded that, at least for its general initialisation, the framework functioned more as a metaphor to help interpret events and their causes and consider the cycle in and by itself to be too general to be viewed as a testable hypothesis (Holling and Gunderson, 2002, p. 49). They do however suggest that it can still be highly valuable as an overarching framework to trace dynamics of change in coupled socio-ecological systems and develop more specific questions and hypotheses. This way, the adaptive cycle can be advanced beyond metaphorical uses to uncover and explain the underlying mechanisms at play in system development. The application of adaptive cycles in a wide variety of disciplines should not be interpreted as the transposition of a meaningless metaphor, but rather as an indication for the adaptability of this heuristic framework and its suitability in capturing a wide range of dynamics. Conceptual flexibility can just as well be an asset for any model or concept if applied with sufficient rigour (Weiberg, 2012). It has been argued that adaptive cycles are “ubiquitous in complex adaptive systems because they reflect endogenously generated dynamics as a result of processes of self-organisation and evolution” (Sundstrom and Allen, 2019). It should therefore not come as a surprise that scholars in different fields find value in them. Adaptive cycles and the panarchy offer a suitable high-level framework where various strands of theory can be coherently integrated to describe and understand processes, structures, and variables operating at discrete ranges of scale. The application of the notion of panarchy and its nested set of adaptive cycles possess the potential to further the necessary multi-evolutionary

116  Conceptualising social complexity and multi-trajectory approach advocated here, by providing a suitable epistemological framing tool to assess variable development within different domains of society as well as link these developments to the natural environment and overall socio-cultural matrix of a given society. It has been noted that the adaptive cycle model offers a useful heuristic for understanding well established archaeological patterns, however, this is only a first step to apply resilience theory to archaeological questions (Freeman et al., 2017, p. 85). Linking back to the introduction of this book, it must be realised that the adaptive cycle constitutes a high-level description of system dynamics which requires embedding in a multi-level, discipline-specific framework to attain empirical value. In other words, the concept must be operationalised. In the next chapter, I will show how the adaptive cycle framework can be applied to archaeological data in an extensive case study. In the final chapter of this book, I will also suggest some potentially fruitful ways forward to further operationalise this framework in future studies.

Notes 1. Although many teachers would probably dispute that and point towards the effectiveness of teaching in smaller groups. 2. Lucas draws from Husserl’s conceptualisation of the ‘temporal flux’ – the combination of progressions of subsequent points of time and an internal flow of time-consciousness – in two-dimensional time diagrams (Husserl, 1964). 3. The PoliGen model was developed on the MASON (Multi-Agent Simulator of Networks and Neighbourhoods) platform, an open-source Java simulation toolkit developed as a collaboration between the Evolutionary Computation Laboratory and the Center for Social Complexity at George Mason University (http://cs.gmu. edu/~eclab/projects/mason/). N2 4. The exponent of the curve approximates , with N being the total population 2 size (West, 2017, p. 317). 5. Although I would raise the point that if we truly wish to uphold that continuum, a different term for so-called non-urban sites should be defined and generally agreed upon in the literature. 6. These discourses can be traced back, respectively, to the works of the French political philosopher Jean-Jacques Rousseau (1712–1778) and his English counterpart Thomas Hobbes (1588–1679). 7 Dataset is openly available at: http://seshatdatabank.info/. 8. A small selection of papers: (Aimers and Iannone, 2013; Daems and Poblome, 2016; Gronenborn et al., 2014; Nelson et al., 2006; Peters and Zimmermann, 2017; Redman and Kinzig, 2003; Thompson and Turck, 2009; Widlok et al., 2012).

4

Social complexity trajectories in Anatolia

Introduction The main purpose of this chapter is to apply the conceptual model developed in the previous chapter to an extensive case study. This case study will focus on social complexity trajectories in (southwest) Anatolia from the Chalcolithic to Hellenistic period. The point is not necessarily to investigate all the nooks and crannies of the conceptual model, but rather to provide a general overview of its potential and explore the possibility space for future research. It is crucial to emphasise the unavoidable trade-off between scope and detail of coverage in such a study. This means that archaeological data will not necessarily be explored to the finest resolution possible, but will instead be used to illustrate points within this larger framework. The case study will cover all scales and dimensions of social complexity trajectories described earlier. These include interaction and information transmission in social groups, community formation through social reactor dynamics, networks of settlements, and macro polities. It also draws from different stages of complexity formation, its emergence and development as well as collapse, transformation, and reorganisation. The purpose of this case study is not to provide a quantification of social complexity and compare this value over time. This does not mean that quantification of social complexity is not a worthwhile approach. Indeed, below I will highlight two examples of macro-level quantified approaches to situate the present case study, and in the final chapter I will outline some lines of future research which will connect my model with potential quantification approaches. Here, I am in the first place interested in illustrating how the formal conceptual framework presented in the previous chapter can be made to work for archaeological studies, using a wide range of evidence and multiple scales of analysis. The aspect of scale is essential in this case study. This is why the multi-scalar concept of adaptive cycles plays an important role as the capstone piece of the model. To effectively apply the adaptive cycle framework in the present framework, it is essential to consider how to capture dynamics of social complexity through such a cycle. It has been

118  Social complexity trajectories in Anatolia noted that adaptive cycle studies are “confronted with similar challenges related to how to parametrize key variables and/or to establish a reliable age and stage model for the SES so that it can profitably be studied from an [adaptive cycle] perspective” (Bradtmöller et al., 2017, p. 4). In one of the earliest applications in archaeology, Charles Redman and Ann Kinzig (2003) noted that the nature of the adaptive cycle depends on the scale of interest. For example, they consider whether every societal reorganisation represents a cycle by itself, or whether such transformations represent specific phase transitions. These are valid considerations. The solution they offer however, can hardly be considered satisfactory. In their case study on centralisation and fragmentation of governmental structures in Mesopotamia, they argue whether the period between 3500 and 2000 BCE, spanning five major historical phases, consisted of one protracted cycle, or whether the individual phases and corresponding dynasties within this period each constitute their own cycle. Both proposals overly generalise the matter and disposes of all the strong points of this framework as societal dynamics operating on various scales across temporal, spatial, and organisational dimensions are reduced to one single dynamic of change. Yet, equating archaeological periods with stages in an isolated adaptive cycle appears the most frequent approach in various archaeological applications (Allcock, 2017; Nelson et al., 2006; Peters and Zimmermann, 2017; Rosen and Rivera-Collazo, 2012). Applied this way, the adaptive cycle indeed hardly transcends the level of a metaphor for change its detractors have made it out for as it becomes bereft of all potential for describing and explaining multi-scalar dynamics. What is needed is to describe patterns of development along a variety of domains, including economic, social, political, and environmental, within adaptive cycles operating on different scales, and then integrate each of these patterns into a multi-scalar framework, indicating relevant connections between different scales, dimensions, and domains. Using the framework outlined in the previous chapter, I will describe and explain social complexity trajectories through selection pressures – subsistence, cooperation, competition, interaction, distribution, production, governance, and demography – acting as input information for collective decision making, using complexity mechanisms – diversification, integration, and intensification – to generate pulling or pushing forces, respectively, concentrating or dissipating flows of energy, resources, and information within a specific social unit. If one would look at social complexity trajectories from a global perspective, one could only conclude that generally speaking, social complexity has been increasing over time. This global perspective, however, need not necessarily be in line with evidence on local or regional scales. I will highlight two examples of global perspectives to situate the approach outlined here, focusing on archaeological and palaeoenvironmental data to reconstruct micro- and meso-level complexity trajectories. These two examples are both

Social complexity trajectories in Anatolia 119 quantification-based approaches, one using the single proxy of social inequality, one using a composite proxy of complexity characteristics. The first pertains to a paper by Timothy Kohler and colleagues, comparing wealth inequality in Eurasia, North America, and Mesoamerica from Neolithic times onwards (Kohler et al., 2017). The main challenge for such an ambitious project of comparative research is to find a consistent, suitable proxy for all polities involved. In this study, wealth disparities were approximated through house size distributions. While house sizes only provide one way of displaying social inequality, other forms of material wealth have not been consistently retrieved from polities across the sample. Moreover, it has been argued from ethnographic and archaeological studies that house sizes provide a reliable estimator for household wealth.1 Using Gini coefficients – measuring the degree of concentration of any given quantity – to calculate social inequality, it was found that wealth disparities generally increased following the domestication of plants and animals, and scaled positively with socio-political scale. Inequality was higher in agricultural societies (median = 0.35) than with horticulturalists (median = 0.27) and hunter-gatherers (median = 0.17). Interestingly, the study also suggests that the greater availability of large domesticates in Eurasia allowed higher wealth disparities to emerge compared to North America and Mesoamerica (Figure 4.1). For each data point on the figure, time on the x-axis has been ‘normalised’ by taking the elapsed time since the emergence of domesticates in that particular area of the world.2 The regression lines in the reproduction of their data in Figure 4.1 do not exactly follow those of the original figure as Kohler and colleagues used a robust regression with locally weighted

Figure 4.1  Gini coefficients for New World (NW) and Old World (OW) polities (reproduced by author with data available from Kohler et al., 2017, Figure 4.3)

120  Social complexity trajectories in Anatolia scatterplot smoothing. Instead, I opted for a simple smooth regression as it suffices to bring across the point that I want to make here. The main difference between both figures is that the gap between Eurasian and American polities in the first part of the graph – before the point where the two lines meet – appears wider here compared to the figure in the original publication. Whether this observation has meaningful repercussions for (parts of) the interpretation of the graph is not my main point. It does raise a more general point that all choices made in the different steps of the data analysis can have an impact in how to represent and interpret our data and should therefore be considered carefully. It suffices to point out, however, that the bottom line of the study – the growing divergence in inequality between American and Eurasian polities – is still clearly observable. This divergence cannot be causally linked to the mere presence or absence of domesticates given that Eurasian polities only overtake their overseas counterparts more than 2,000 years later. The difference is generated because of a growing increase in inequality in Eurasian polities, whereas inequality declined or remained stable in North America and Mesoamerica. It has been suggested that the explanation for this inequality divergence can be found in the multiplier effects of domesticated draft animals for human labour resulting in agricultural extensification (Bogaard et al., 2019). Whereas the maximum potential yields plateaued in the Americas because of the absence of large domesticates, increasing returns to scale drove Eurasian polities forward on this axis of social development towards greater wealth disparities. Given that only the richest households could have afforded obtaining and maintaining draft animals (or at least that richer households could have more of them), potential wealth disparity was significantly higher. Moreover, the effects of positive feedback loops induced by labour intensification and secondary benefits of domesticated animals (manure, milk, cheese, meat, …) exacerbated this trend. Kohler et al. (2017, p. 3) argue that agricultural extensification set off a second process of increasing wealth disparities – observed in the right part of the graph – driven by the development of metallurgy and a (mounted) warrior elite. These innovations contributed greatly to successes in warfare and polity expansion, allowing Eurasian polities to markedly increase in scale. Let us now shift from single proxy measures to composite metrics of social complexity. Several attempts have been made to develop an encompassing measure of social complexity (Chick, 1997; Morris, 2013). Social complexity is a multi-dimensional and multi-scalar phenomenon that is not easily captured in a single quantitative metric. So far, however, no metric has found widespread acceptance. The question is what the relevant factors for social complexity would be and how to quantify them. To tackle these questions, the “Seshat: Global History Databank” project built a database, integrating data on 414 societies from 30 regions around the world spanning the last 10,000 years with coded information for 51 variables related to nine major

Social complexity trajectories in Anatolia 121 ‘complexity characteristics’: (1) polity population; (2) polity territory; (3) capital population (i.e. the size of its largest centre); (4) levels of hierarchy; (5) government; (6) infrastructure; (7) information systems; (8) specialised literature; and (9) monetary systems (Turchin et al., 2018a, 2015). PCA of these characteristics has shown strong evolutionary correlation, meaning that they co-evolved consistently across time and space. The first principal component captures as much as 77.2 ± 0.4% of the total variance among the variables. Turchin and colleagues suggest that it is indeed possible to capture social complexity with this metric. This does not mean that social complexity is actually uni-dimensional. Instead, this metric reflects “a composite measure of the various roles, institutions, and technologies that enable the coordination of large numbers of people to act in a politically unified manner” (Turchin et al., 2018a, p. 4). When plotting the values of the first component for each polity at 100-year-time intervals (Figure 4.2), the overall trend is one of increasing complexity over time, albeit slightly dipping towards the end. So, what are we to make from this graph? Does the upward trend of the graph really indicate increasing social complexity over time? And if so, what does that mean for the societies involved? It should be reiterated that the measure, first and foremost, captures the largest dimension of shared variability in societal development. The data suggest a strong covariation among these variables along the trajectory of social complexity across different societal configurations, from small-scale hunter-gatherers to largescale urban societies and polities (Shin et al., 2020). It does not, however, provide a direct answer as to how societies developed as they moved along this dimension of variability.

Figure 4.2  Composite complexity metric based on PCA from Seshat project (made by author with data available from Turchin et al., 2018a)

122  Social complexity trajectories in Anatolia Turchin and colleagues have suggested that the generally upward-directed trajectories of social complexity can be explained as being subjected to driven trends, that is, selection mechanisms that actively favour larger values of a specific function. More specifically, they suggest that group competition – most importantly through polity warfare – acted as the primordial driver of social complexity.3 Moreover, they see the observation of long periods of limited or slow change interjected with sudden, pronounced changes in social complexity – both upward and downward – as potentially consistent with the model of punctuated equilibrium in social evolution (Currie and Mace, 2011; Spencer, 1990). By their own admission, more formal testing, particularly on the mode and tempo of evolutionary change in the data, is needed before these claims can be confirmed. A potential explanation for the punctuated dynamics in complexity trajectories was suggested by Shin et al. (2020) who showed that an initial phase of increase in polity sizes is typically followed by an improved information processing apparatus and economic system, before additional scale increases can occur. They used this observation to hypothesise polity scale and information processing acting as thresholds for polity growth. It is from this interplay between thresholds that discontinuous trajectories of growth emerge, characterised by slow accumulation of innovations in information processing and transmission, punctuated by episodes of transformative change. Hierarchical cluster analysis of the first principal component in the Seshat data seems to indicate the existence of two clusters or types of societies, even though it is not yet clear what these actually mean (Turchin et al., 2018a). Interestingly, this duality is also attested in studies using other large datasets. Quantitative analysis using Guttman scaling and morphospace analysis4 based on data from the Atlas of Cultural Evolution (ACE),5 which includes 289 archaeological traditions6 across time and space, has suggested that the accumulation of social complexity is a cumulative process characterised by economies of scale, increasing functional diversity, higher rates of social connectivity, and more efficient energy capture (Ortman and Peregrine, 2018; Peregrine, 2018). These studies also suggested the existence of multivariate basins of attraction for social organisation – that is, two major types of societies. These basins generally revolve around aspects of scale and technology. The latter is a composite variable covering aspects of information capture, transport, social stratification, political integration, technological specialization, and money. Clearly, these are closely related to the nine complexity characteristics in the Seshat study. A recent study by Lux Miranda and Jacob Freeman (2020) used multi-dimensional clustering analysis to confirm the existence of two main clusters characterised by recurrent social structures. Additionally, they also identified five sub-clusters which provide a good fit to the Seshat data. Interestingly, their results also leave room for debate whether the first principal component of the Seshat analysis sufficiently captures variation in

Social complexity trajectories in Anatolia 123 social complexity. It is observed that the five sub-clusters do not match up with particular levels of complexity as societies from different clusters with markedly different complexity characteristics can still have strongly overlapping values in the first principal component. It must be noted that the fact that the first principal component captures a large amount of variation in the Seshat data, does not necessarily preclude the possibility for hidden variables to have influenced dynamics in the system. This suggests that, at the very least, certain nuances of diversity in complexity characteristics are not captured by the metric. The general picture painted from the Seshat databank points towards a global trend of increasing social complexity. Regardless of whether we use traditional archaeological approaches to complexity, or the approach based on flows of energy, resources and information supported here, this overall trend will hardly be disputed. Yet, this macro-scale observation will not necessarily be useful or interesting for a disciplinary specialist working on the archaeology of one specific region. To see what this might mean on a meso- and micro-level, I will focus on a case study that includes one area from the Seshat databank, the Konya Plain. I will not specifically focus on the Konya Plain per se, but rather contextualise this area using the archaeological evidence from the wider region of southwest Anatolia (modern-day Turkey) to discuss in some more detail the complexity trajectories, using the elements of the conceptual model highlighted in the previous chapter. This exercise should not be taken as a critique on the results of the Seshat databank, or its goals, but is rather intended to show that social complexity can be studied on different scales. Each of these scales offers different perspectives with its own potential for understanding the past.

Case study: Social complexity in Anatolia If we isolate the complexity trajectory of the Konya Plain area from the Seshat data, we see a generally upward-driven trend towards increasing complexity consistent with the overall picture (Figure 4.3). The full dataset covers the period from Early Neolithic to Ottoman times (9600 BCE–1838 CE). Discussing this entire period will not be possible in the given context. In this figure, we see a clear ‘leap’ in complexity from 5000 BCE onwards, that is, from the Middle Chalcolithic. I will take this as a starting point. For practical reasons, I will set the chronological endpoint of this discussion in Hellenistic times. This chronological range covers a sufficiently wide period that includes several phases of supposed increases and decreases of complexity. Looking at this figure, one thing that can immediately be noted is the sparsity of data points in the early periods. This is because the complexity value of any given period is reduced to a single data point and the chronological eras used to structure the history of these early times are fairly

124  Social complexity trajectories in Anatolia

Figure 4.3  Complexity trajectory of the Konya Plain (reproduced by author with data available from Turchin et al., 2018a)

broad. The more detailed periodisation of later times results from written sources providing far more information on changing dynasties and polities. This information is sorely lacking for the prehistoric period. This results in a more fine-grained view on the more recent stages of the complexity trajectory. Turchin and colleagues themselves noted in their supplementary information that “large changes in PC1 [the value plotted in this figure] could indicate a new polity taking control of an NGA rather than the kind of change within a society that is envisioned under the punctuational change hypothesis.” (Turchin et al., 2018a). Here, they touch upon an essential point. Because the Seshat data are coded on the level of the polity, changes in which polity controlled a given area are more likely to cause major changes in complexity values for that area than any internal dynamics, even though this might not mean much for the trajectory of social life in local communities which often continued unaltered regardless of which overarching polity was in control. For each period in the case study, I will describe the archaeological evidence that will support the interpretation of changes in complexity trajectories. I will then outline: (1) the main selection pressures for social complexity formation; (2) what complexity mechanisms were used; (3) whether these processes acted as pushing or pulling forces for complexity formation; (4) how this impacted flows of energy, resources and information; and (5) how the observed dynamics of stability and change relate to the adaptive cycle framework. Let us now turn to the evidence in more detail to see if we can elucidate what the upward trend in social complexity actually means in light of the archaeological data and what might have driven it. I will mainly discuss data from southwest Anatolia (see Figure 4.4), but I will frequently

Social complexity trajectories in Anatolia 125

Figure 4.4  Map of case study of social complexity trajectories in southwest Anatolia (made by author)

draw in evidence from elsewhere in Anatolia to support the argument where necessary. Southwestern Anatolia can be subdivided in three main topographical areas: the coastal zones, the Taurus mountains with intra-montane rivers and lake basins and the outskirts of the central Anatolian plateau. Large parts of the landscape are characterised by compartmentalisation through deep river valleys and high mountain chains. Some have suggested that this topography was one of the reasons for marked cultural regionalism, both in material culture and social organisation, from Neolithic times onwards (Çevik, 2007). At the same time, many scholars have noted strong connections and cultural interactions on a supra-­regional level, giving rise to the suggestions of a (Western) Anatolian socio-cultural koine at various points in time. Christoph Bachhuber (2015) suggested that the emergence of a citadel culture during EB III can be seen as a general phenomenon across Anatolia (see infra). The unprecedented homogenization of material culture and architecture attested at these sites was the result of groups of people across a vast area beginning to self-consciously to resemble one in another in a process that was at once emulative and competitive. Throughout the period under discussion here, the interplay between climate, environmental changes and human impact induced complex processes of landscape formation and degradation, resulting in a highly variable landscape that has radically changed over time. Inner Anatolia in the early

126  Social complexity trajectories in Anatolia Holocene, for example, was characterised by rich grasslands, which were replaced with oak parklands and low-diversity wood pastures during the mid-Holocene. I will not fully discuss this environmental and topographical diversity here, but some aspects will feature prominently at multiple points during the case study and will have profound impacts on complexity trajectories. Chalcolithic In his overview of Prehistoric Asia Minor, Bleda Düring laments the common practice of lumping early Chalcolithic times with the preceding Neolithic and Late Chalcolithic with the subsequent Early Bronze Age, leaving the Middle Chalcolithic “caught in the middle as an eventless span of time” (Düring, 2011, p. 201). This practice could be contributing to the trend observed in the Seshat data, creating the illusion of large complexity increases in the transition to the Early Bronze Age. As a result, the Middle Chalcolithic (5500–4000 BCE) in Anatolia is still poorly understood and only few settlements have been attested, especially in the southwest. Few settlements have been excavated in the area, leaving us with little evidence of structural remains. Excavations at other Chalcolithic sites in Anatolia have mainly yielded wattle-and-daub constructions and floors made out of gravel and sandy clay (Eres, 2003). Similar structures could be suggested for the southwest as well. Excavations at Hacılar yielded mostly

Figure 4.5  Main (late) chalcolithic sites discussed in the text (made by author)

Social complexity trajectories in Anatolia 127 Early Chalcolithic evidence, but according to the original excavator James Mellaart, the community consisted of about 150 people during the Middle Chalcolithic period (Mellaart, 1970). At Kulaksızlar, evidence of extensive stone production has been found, including blanks, production waste, stone working tools and manufacture rejects, along with stone objects such as figurines, bowls and beakers (Takaoglu, 2002). Raw materials for this production include marble, gabbro, basalt, and sandstone, whose sources were all located within walking distance of the site. The so-called ‘Kilia figurines’ are some of the most common artefacts produced at Kulaksızlar. This type of object has been found at a number of sites across western Anatolia. Despite the general dearth of excavated settlements, the general picture of the Middle Chalcolithic seems to point towards small-scale settlements which were to a certain extent integrated in interregional networks of socio-cultural interaction and exchange that have been described as a cultural koine (Schoop, 2011). Evidence of sites such as Kulaksızlar points to specialised production geared towards these networks, but it is not clear to what extent this can be extrapolated to other sites and it should perhaps be considered an exceptional case. For Late Chalcolithic times (4000–3000 BCE), more evidence is available. Excavations at Kuruçay Höyük uncovered 23 buildings dated to 3500 BCE (Duru, 1996). These buildings consisted of mudbrick walls on stone foundations and a flat roof. The limited surface are of these mounds likely posed strong restrictions and logistical challenges on the use of space in social life and community organisation. Evidence for the usage of rooftops for household activities or storage could be seen in this light as a strategy to maximise available space. The fact that only one structure was found with a built-in hearth suggests the usage of portable hearths, which also points towards a flexible use of space for various activities. Evidence for space maximisation is also found at Canhasan where houses were built almost contiguously (with merely 10 cm separation in between structures, possibly for drainage reasons) (French, 1998). The lack of obvious ground-floor doorways in many structures suggest that homes were accessed through rooftop doorways in similar fashion to Çatal Höyük. The structures at Kuruçay Höyük were mainly free-standing houses of one or more rooms, showing some differentiation in house sizes (Vandam et al., 2019). Even though we have little clear evidence, this could suggest the emergence of incipient social inequalities or wealth inequalities. The settlement was at this time no longer organised around larger sized open areas, but around streets/alleys with limited space and smaller open areas. Another architectural development at Kuruçay Höyük, as well as Canhasan, was the increased internal partitioning of rooms, the construction of buffer spaces and offset doorways, which impeded clear views from one room to another and increased (individual) control over spaces. This has been interpreted as

128  Social complexity trajectories in Anatolia an attempt to create greater privacy and space differentiation, shifting from community-oriented to household-oriented activities (Steadman, 2000; Vandam, 2019). The original excavator Refik Duru interpreted Kuruçay Höyük as a small urban centre, pointing towards the supposed existence of central buildings, shrines and extensive domestic structures identified as the residence of a leader. This interpretation is however not very convincing and has been rejected by others, who favour an interpretation as a large village (Düring, 2011, pp. 228–229; Vandam et al., 2019). In a later phase of the settlement at Kuruçay Höyük, previously isolated structures became increasingly interconnected, impeding easy access to buildings and passage through the village. While this could suggest a continued need for space maximisation, the walls of the Late Chalcolithic structures clearly became clearly thinner and lost their internal buttresses. This might indicate that rooftops were no longer used for household activities to the same extent as before. The apparent lack of ground floor doorways, however, suggests rooftop areas were still available and possibly still used to enter houses. Barbara Horejs (2014) described Late Chalcolithic community formation at Çukuriçi Höyük through a model of ‘proto-urbanisation’ as a phase of cultural transformation defined by community consolidation and functional differentiation in land use and resource exploitation. Before moving on to the model itself, it is worth pointing out the problematic nature of the term ‘proto’ as a societal classifier. Even though Horejs herself is quick to note that societal development is not a linear trajectory towards growth and increasing complexity, the term does retain teleological associations. Something which is considered ‘proto-urban’, is almost by default poised to evolve towards an urban state, even when that process is strictu sensu dissociated from notions of success and failure. The choice in terminology is at least unfortunate and should be reconsidered. The Late Chalcolithic has been described as an essential transitionary period towards more complex societies in a traditional archaeological sense, characterised by higher degrees of urbanism, monumental architecture, and long distance exchange networks (Horejs and Mehofer, 2014; Vandam et al., 2019). In her model, Horejs paints a trajectory of development from egalitarian farming villages in the Neolithic to differentiated settlement patterns in the Late Chalcolithic, followed by the emergence of urban communities within regional power structures in the Early Bronze Age. Even though no veritable urban centres existed in the Late Chalcolithic, certain processes of functional differentiation in settlements and practices took place. This differentiation is characterised by the emergence of social inequalities, a denser settlement pattern and intensified land use. Recent landscape surveys in the area surrounding Kuruçay Höyük in the Burdur Plain have indicated the existence of a differentiated settlement system consisting of larger höyüks (settlement mounds created by accumulation of detritus from prolonged habitation) and smaller sized village

Social complexity trajectories in Anatolia 129 settlements (often denoted as ‘flat’ sites in contrast with the höyüks), possibly of a more ephemeral nature (Vandam et al., 2019). Subsistence practices included farming crops such as emmer, barley, einkorn, and lentil, and rearing livestock such as cattle, goat, sheep and pig, as well as the exploitation of various ecological niches, covering both terrestrial and maritime ecosystems (Galik, 2014). A range of local and non-local raw material sources, including most notably obsidian, chert and copper, were exploited and used for the production of material goods. Technological innovations such as smelting developed in specialised production workshops. Along with existing crafts such as metal working and pottery production, textile production emerged as a specialised craft in Late Chalcolithic Anatolia as well (Schoop, 2014). Some of these goods continued to be geared towards regional and interregional exchange networks. It has been suggested that these networks covered large parts of (western) Anatolia and the eastern Aegean based on similarities in architectural forms and elements of material culture such as pottery shapes and decoration patterns (Horejs, 2014). We should be careful with interpreting these material similarities as signs of exchange networks given that there is no indication for common origins or the mechanism of diffusion. Still, it could point towards certain forms of connectivity and interaction resulting in the exchange of ideas at a macro level that are then implemented in local or regional production. Moreover, specific objects such as the aforementioned marble figurines from Kulaksızlar continued to be exchanged over a large area. From the preceding discussion we can conclude that reducing the extensive period covered by the Chalcolithic to a single data point (Figure 4.3) masks a great deal of variability and internal development over time. By placing this data point on the graph at the start of the period in 5000 BCE, developments that actually occur afterwards, over the course of a two millennium period, are compressed and shifted backwards in time. Most processes potentially contributing to a higher complexity value are moreover occurring in the Late Chalcolithic. The most considerable differences in complexity values are found in variables for hierarchical complexity, including a higher settlement hierarchy, more levels of administration and higher religious and military levels. It should be noted, however, that no clear indications exist for supra-local levels of organisation and administration at this time. It is interesting to note that Mellaart suggested a population size of about 150 people for Hacilar.7 It is tempting to link this to Dunbar’s number as a limit to the number of people forming a face-to-face community. The Chalcolithic in Anatolia saw many communities going through changes in socio-economic organization, most eminently expressed by increasing craft specialization, long-distance trade, and increasing wealth differentiation. Steadman has noted that increasing socio-economic complexity in Prehistoric Anatolia is driven by three interrelated behavioural processes: (1) increasingly complex or numerous performances of tasks and associated

130  Social complexity trajectories in Anatolia size increases; (2) (architectural) partitioning, and (3) segmentation of social structures (Steadman, 2000). These processes are part of a wider process of economic differentiation within local communities in southwest Anatolia (Eslick, 1988). Some of the most prominent indicators of increasing social complexity through these behavioural processes can be found in the evolution of the built architecture and use of space on the settlement mounds. These architectural changes are driven by the maximisation of available space and flexible usage of rooms. It is possible that these strategies were induced by population growth intensifying existing restrictions on space availability on settlement mounds. However, it remains difficult to prove this hypothesis without clear data for settlement sizes or population numbers. Regardless of whether population growth occurred and – if so – to what degree, I would like to suggest, that strategies of space maximisation such as the usage of roofs for household activities were part of attempts of local communities to grapple not only with the physical restraints on available space in settlement mounds, but also with the socio-cognitive limits on community sizes. By creating additional levels of household activity, people were able to reduce the cognitive load of increased interaction and information transmission, without losing the overall coherence within the community by expanding beyond the physical limits of the mound. In the Late Chalcolithic, we observe a markedly different strategy to deal with these cognitive loads, focused on internal partitioning of houses, the reduction of inter-household interactions and shifts towards household-based activities. This need not mean that people fell back on individual or family-based isolation. Socially integrative practices and interactions will have continued to play an important role in everyday life. The creation of designated areas and times for interaction may even have intensified the effect of its integrative nature. It should be noted that an extensive window of time separates the attestation of rooftop activities on the one hand, and the creation of private space on the other. In light of the limitations in available data, we should therefore be careful to paint the latter development as a reaction against, or development out of, the latter. As communities explored the possibilities for the creative use of space, they would still have encountered limits to expansion that became increasingly difficult to circumvent. At this point, accumulating scalar stress driven by energised crowding would have required a collective decision to be made. Either the community would fission and part of its members would relocate, or additional layers of social organisation needed to be built to act as information processing devices. The development of craft specialisation and inter-regional exchange networks are often linked to the emergence of local elites, creating a demand for exotic goods. However, despite some evidence for incipient social inequality, no clear indications exist for social stratification. It can therefore be expected that communities would generally have chosen to fission. We should not exclude the potential agency of the local community at large for the establishment of local craft production and participation in

Social complexity trajectories in Anatolia 131 inter-regional networks of exchange. Out of the four main categories of household functions (production, distribution, transmission and reproduction) identified in the seminal study of households by Wilk and Rathje (1982), Steadman focuses particularly on production as “the capacity to affect the use of space and architectural orientation” (Steadman, 2000). Production systems become more complex as the productive processes shift from linear to simultaneous sequences, that is from one person performing a series of sequential tasks to multiple people acting in a coordinated manner towards a common goal. She suggests that larger households, as a result of population growth and nucleation, would have had more potential to exploit economic opportunities through division of labour. Such opportunities may well have arisen as a result of innovations driven by energised crowding. As more tasks were to be performed in individual households, specific rooms may have been purposed for specific (set of) tasks. This could also be a potential explanatory factor for the observed architectural partitioning. At the same time, given the flexible usage of spaces, it is very likely that at least in parts of the house, multiple tasks were performed in the same spaces. In these cases, non-permanent material cues will likely have been placed – and removed afterwards – for the performance of specific tasks, for example the usage of portable hearths for cooking. Both craft specialisation and the emergence of these networks continued to develop during the subsequent Early Bronze Age and will be further discussed there. From Table 4.1, we can deduce that the main drivers of complexity trajectories in the Late Chalcolithic were pushing forces generated by Table 4.1  Social complexity trajectories in the Chalcolithic. *Energy (E), resources (R), information (I); **diversification (D), intensification (IS), integration (IG) Selection Period pressure LCH LCH LCH LCH LCH LCH LCH LCH

Flows* Process

Demography Population growth Interaction Energised crowding Interaction Energised crowding Interaction Energised crowding Distribution Demand for foreign goods Interaction Energised crowding Production Division of labour Demography Population growth

Outcome

Drivers

Mechanism**

E

R

I Pull Push D

IS

Population x nucleation Architectural partitioning Economic differentiation Social segmentation Inter-regional exchange

x

x

x

Space maximisation Craft x specialisation Group fission

x

x x x x

x

x x

r

x

x

r

x

x

r

x

x

r

x x

IG AC

x x

x

x

K x

x

x

r

K Ω

132  Social complexity trajectories in Anatolia diversification as their main mechanism. Limits to the processing of information flows acted as a restriction on growing community sizes. This could have resulted in repeated episodes of group fission which would gradually start filling up the landscape, driving up settlement numbers. When we integrate these processes in the adaptive cycle model (column AC in the table), we see that repeated processes of population growth, space maximisation and group fission in Chalcolithic times correspond to subsequent cycles of growth (r), the infilling of niches in the possibility space of the community (K) and release (Ω) to induce a new cycle. It is interesting to note that the limits of the system were met primarily on the level of information flows generated by social interaction, rather than potential limitations in flows of energy and resources. This means that the release phase might be triggered due to the accumulated effects of scalar stress generated by energised crowding, while resource exploitation might still be squarely in the r-phase. As a result, communities might have fallen apart due to social dynamics long before they faced any environmental or logistic problems induced by resource exploitation or energy consumption. This process skipped the reorganisation phase (α) given that communities remained in the same basin of attraction and new cycles (that is new communities founded elsewhere) retained the overall characteristics and dynamics of the existing system. Early Bronze Age Strong cultural continuity has been observed between the Late Chalcolithic and Early Bronze Age (EBA). As a result, these periods are often discussed together. Still, EBA (3100–2000 BCE) in Anatolia constituted another leap in the complexity trajectory of Figure 4.3. Most notably, it has been described as “a period of ‘urbanization’, or at least the age in which complex society emerged” (Steadman, 2011, p. 231). Some of the key factors of this transition were marked increases in community sizes and the establishment of power structures over a wider hinterland, resulting in the development of (micro-)regional polities (Çevik, 2007). Yet, the nature of these power structures and the ways in which central places exerted control over other sites remain unclear. Certain elements of this transformation such as the development of an elite class of leaders, public architecture and increasing settlement sizes may well have developed already in the Late Chalcolithic. A lot has been written about the problems of chronology related to EBA Anatolia, especially regarding its foundations in regionalised ceramic traditions and stratigraphic sequences of a limited number of major sites (Bachhuber, 2015; Düring, 2011). To keep things simple, I will retain the traditional threefold subdivision (EB I–II–III) to structure the overview and provide some more temporal resolution for the discussion.

Social complexity trajectories in Anatolia 133

Figure 4.6  Main EBA sites discussed in the text (made by author)

The two-tiered settlement structure already observed in Chalcolithic times, expanded in EBA with the emergence of (micro-)regional centres in addition to villages and farmsteads. Major sites were generally located on natural rises in the landscape, often along alluvial plains and with ample access to water (Bachhuber, 2015). Villages are typically considered to have been egalitarian communities. However, one has to differentiate between horizontal and vertical egalitarianism (Frangipane, 2007). In horizontally egalitarian communities, governance and decision-making power is spread across different groups. In vertical egalitarianism, social differentiation is downplayed through integrative practices to guarantee equal access to resources and social life, but decision-making power rests with one or more groups of people (often defined by kinship) who hold higher status. EBA villages were most likely characterised by a vertical egalitarian system. This means that status hierarchies existed within and between households but were not expressed in explicit ways on a settlement level. The farming household remained the essential arena of everyday life and the linchpin of the social fabric of the community. At the same time, social life would have largely taken place outside of the house itself in different types of public space such as open areas in the middle of the settlement, special-purpose central buildings, and extramural cemeteries. Christoph Bachhuber (2015) argued for the emergence of a ‘citadel’ culture in EBA Anatolia. Some have linked the emergence of fortified sites with a rise in organised violence and an increased preoccupation with site defensibility (Massa et al., 2020). Bachhuber, however, stressed that citadel sites such as Liman Tepe mainly emerged out of social differentiation and

134  Social complexity trajectories in Anatolia elite formation processes expressed through monumental architecture and central place functions related to administration and exchange networks. Two general types of citadels have been observed: (1) an early type emerging in EBA I–II characterised by a social-oriented logic based on integrative spaces and buildings; (2) a latter development from EBA III displaying more distinctly exclusionary practices. Both types represented the rewiring of flows of energy and resources into concentrated avenues of social competition and conspicuous consumption. The difference in governance structures between villages and citadels is mainly one related to scale rather than of clear qualitative distinctions. People with administrative functions in both types of communities held the power to extract production output. However, whereas village authorities probably only covered the individual households within the community, citadels acted as central places in the landscape exerting power as part of a projected information environment, integrating nearby villages as well. Not many sites dated to EB I (3100–2700/2600 BCE) are known from southwest Anatolia. At Beycesultan, only a single structure was unearthed, identified by the excavators as a ‘shrine’ (Lloyd and Mellaart, 1962). This identification is mainly derived from the interpretation of later building phases at this location which may have had this functionality. It has been argued, however that the proportions and spatial organisation of these structures did not differ significantly from domestic buildings elsewhere in Anatolia, and that the interpretation of a ‘shrine’ should not necessarily be preferred over that of a domestic building (Düring, 2011). The material culture associated with this structure includes marble figurines, bead necklaces, metal objects, two-handled jars, flat baking platters, and three-legged cooking pots which bear strong resemblance to Late Chalcolithic material, suggesting continuity of settlement. At Karataş in Lycia, a rectangular mudbrick structure was found at the centre of a mound, surrounded by circular wattle-and-daub structures on the slopes of a surrounding lower mound. The extension of the settlement towards the surrounding lower mound may suggest population growth resulting in the community extending beyond their original spot of habitation. In a next phase, a megaron structure, consisting of a central hall and entrance porch, was constructed on the central mound (Warner, 1994). It is easy to see how this central structure could be interpreted as the residence of a local leader. The presence of 12 clay-lined pits inside of the structure, two of which contained storage pithoi, could also suggest some form of storage facility for agricultural and/or horticultural produce. The building was also encircled by a buttressed enclosure, identified as a ‘court area’ that could have housed socially integrative events. Administrative activities – possibly related to the collective storage of goods – are hinted at through a bulla with stamped seal impression and terracotta stamp seals found in a deposit just outside of the structure. The presence of cylinder seals suggests either participation in extensive

Social complexity trajectories in Anatolia 135 exchange networks or the implementation of local administrative practices on the production, distribution and consumption of local produce (Bachhuber, 2015). The building as a whole likely played an important role in community activities and governance such as housing councils and commemorative events. Whereas these communities were unfortified, the citadel site of Liman Tepe had an elaborate fortification wall and glacis with gateways, curtain-wall effects and ‘horseshoe-shaped’ tower bastions. It has been suggested that this glacis not only had an important defensive function, but also signified permanence and stability in the power of the community that built it (Bachhuber, 2015).8 Extensive fortifications were also attested at Hacılar Büyük Höyük in the Lake District. It has been suggested that this site exerted an important central place functionality over the surrounding Burdur Plain (Umurtak and Duru, 2017, 2016). During the EB II (2600–2300 BCE), a general intensification of dynamics initiated in the Late Chalcolithic and EBA I is observed, including the emergence of complex techno-productive industries in metallurgy (most notably bronze production), the intensification of long-distance exchange networks, increased production of prestige goods in local material culture traditions, monumentalising architecture and settlement fortifications (Düring, 2011). In short, the first signs of a transition towards urban communities started to emerge. In the archaeological record of southwest Anatolia, however, these changes are only sparsely attested. At Karataş, the central building was damaged by erosion and a fortification wall was erected. Additionally, free-standing megaron buildings – some of which displaying similar structural features such as hearths, and iron and post stands as in Beycesultan – were constructed on the lower mound. The ceramic assemblage associated with these buildings (an estimated total of 128 structures, see Duru, 2008) showed strong parallels to the material from Beycesultan and Kuruçay Höyük, both in shape and decoration. To the north of modern-day Antalya, at the site of Bademağacı, a large number of structures were uncovered, lined along the slope of the mound in a radial settlement layout with all entrances facing a common central area. Although only about half of the site has been excavated, the settlement consisted of an estimated 60– 100 structures (Steadman, 2011). At Liman Tepe, a series of buildings with a possible administrative function were constructed within the fortified citadel. One large building with a central courtyard and rectangular storage rooms has been identified as the seat of a central authority with possible ritual functions (Erkanal, 1999). Finds from the building include fine red-and-black burnished wares possibly originating from Central Anatolia or mainland Greece, a bull-headed rython, cylindrical stone phalloi, and a clay ‘standard-­like object’ (Şahoğlu, 2005). Residential buildings have been found on the lower mound.

136  Social complexity trajectories in Anatolia Archaeological surveys in the Konya and Karaman Plains identified a range of large sites including Kanaç/Kıbrıs Höyük (ca. 42 ha), Seyithan Höyük (ca. 35 ha), Eminler Höyük (ca. 32 ha), and Sarlak Höyük (26 ha), which reached their largest extent between ca. 2800 and 2300 BCE (Massa et al., 2020). The site of Türkmen-Karahöyük was the largest settlement mound in the Konya Plain, consisting of a central mound of 35 ha and a surrounding lower town covering five secondary mounds over an area of about 40–50 ha. The oldest community at the site emerged around the middle of the third millennium BCE. At this time, habitation was likely concentrated on the central mound as the surrounding area was only settled from the late second millennium BCE onwards. The contemporaneous emergence of seven settlement aligned along the fringes of the May river delta in the Konya Plain could point towards the development of artificial irrigation channels of up to 25 km to deal with the increased aridification of the area. The development of water management practices in this period is confirmed by geoarchaeological studies conducted around the site of Çatal Höyük which indicate that the Çarşamba river was limited to seasonal flooding, likely by siphoning water flows through artificial channels, from the late third millennium BCE onwards (Boyer et al., 2006). Burials in EBA Anatolia generally took place in extramural cemeteries. Various types of burials have been noted, including pithoi graves – the most numerously attested type in southwest Anatolia – as well as cist graves, shaft graves, funerary jars, chamber graves and pit burials. Graves were generally not sealed permanently, but were built to be re-entered for secondary interment practices. Data from burial goods constitute some of the best evidence for the usage of material culture in social practices of these communities. Deposition of ‘foreign’ objects in burials from the site of Demircihöyük may have invoked a symbolically charged geographical distance, both in life and death. The attestation of ‘Syrianising’ lead bottles in particular could be seen as evidence of ‘performative representations of the external other’ (MacSweeney, 2011). They may have been associated with individuals adept at navigating the ritual interactions with ancestral people of the past as well as interactions with people from distant communities (Bachhuber, 2015). Regardless of their symbolic meaning, these objects embodied clear indications of interaction and exchange between peoples on an inter-regional scale. Even if we cannot take this as evidence for direct links between Anatolia and Syro-Mesopotamia, these people would have had no problem recognising the ‘foreignness’ of these objects and evaluated its meaning accordingly. In the Burdur Plain in the Lake District, an EBA pithos cemetery was discovered at the site of Gavur Evi Tepesi. The site associated with this cemetery was already discovered by the Hacilar excavation team but re-investigated during the intensive archaeological surveys conducted by the Sagalassos Project (Vandam et al., 2013). The site is located on a low

Social complexity trajectories in Anatolia 137 promontory overlooking the Burdur Plain and covers an area of about 1.9 ha. Even though the cemetery had been largely destroyed due to quarrying activities before an encompassing study could be conducted, 21 in situ pithos holes and a large amount of pithos fragments have been recovered, including one complete object. The famous ‘royal’ tombs of Alacahöyük in Central Anatolia merit a small digression here, even though the site is located outside of our study area. At this site, 18 burials – mainly shaft graves – dated to EBA I–II were discovered, built in monumental architecture and containing an impressive array of material goods, including many metal objects such as sun standards, weapons, serving vessels, figurines and personal ornaments. The monumental nature of these graves and richness of the associated objects are highly divergent from most of the known burials elsewhere in Anatolia at the time. This is especially surprising given the unremarkable nature of the settlement itself, which was not much different from the typical village at the time. The graves and objects themselves have been discussed extensively elsewhere (see Bachhuber 2015, for an overview of the literature). Here, I will focus specifically on the potential explanatory factors behind this divergence. The powerful evocation of animals attested in the iconography of the burial goods and the frequent attestation of cattle skulls and hooves in the graves suggests an important role of livestock in the local community. Combined with the village-like character of the settlement itself, this could point towards a local elite investing disproportionally into a pastoralist economy, potentially involving seasonal transhumance (Bachhuber, 2015). The question would then be whether this economic mode allowed this community to capture a disproportionate amount of resources and capital, or whether the cemetery represented a more focused investment into particular avenues of socio-economic life. Alternatively, Alacahöyük could have been a mixed-farming community not unlike Karataş and Demircihöyük, but with extreme focus on metallic forms of wealth deposited in the graves. This would again suggest a highly concentrated investment of resources into particular areas. The wealth in burial goods might be indicative of a prestige-goods political economic system geared towards exclusionary objects, produced by skilled craftsmen from materials that are difficult or expensive to procure. They act as powerful signifiers of social status and are indicative of contacts and interactions beyond the own community on a regional and interregional scale. Typologically similar standards to the ones found in the graves of Alacahöyük have been attested regionally in cemeteries elsewhere in the Halys Basin. The small number of graves found at Alacahöyük (14 rich burials and 4 modest ones) compared to Demircihöyük and Karataş (each with hundreds of burials) also suggest a more exclusionary mortuary practice where only a restricted part of the population could participate in these conspicuous rites. These burials might be the material

138  Social complexity trajectories in Anatolia expression of a hierarchical social organisation, driven by local elites participating in regional networks of production and exchange of precious metals. It has been suggested that the disproportionate reliance on conspicuous consumption of wealth at Alacahöyük might be indicative of a competitive interaction among members of this elite, resulting in an unsustainable social configuration (Bachhuber, 2015). It is perhaps telling that the cemetery was likely abandoned within a few generations of its first internments. The transition from EBII to EB III (2300–2000 BCE) is not well known and some have argued for a break in occupation at many sites in western Anatolia related to the arrival of new peoples, most notably Indo-Europeans (Mellaart, 1963). Recent studies of settlement patterns and material culture have disproven this thesis and pointed towards wider patterns of unrest across Anatolia suggesting rises in social conflict that seem unrelated to any external invasions (Massa and Şahoğlu, 2015). These studies have also shown a previously unrecognised continuity in occupation and cultural traditions. Archaeological surveys throughout Anatolia showed a marked increase in settlement numbers and site sizes during EB I–II, which is generally thought to have been the result of widespread population growth. In EB III, however, a sharp decline occurred, with settlement counts even ending up below Late Chalcolithic numbers (Bachhuber, 2015). Boyer and colleagues convincingly argue that, at least for the Konya Plain, this trend is not the result of biases introduced by alluvial masking and that the observed settlement distribution for EBA closely matches the expected patterns from their geomorphological model (Boyer et al., 2006). Instead, they argue that a new alluvial regime induced infilling of a topographical uneven landscape in the alluvial plains that reduced the availability of water and fertile sediments and resulted in a less-arable landscape with lower agricultural potential. It is possible that this induced a tendency towards population nucleation in central places. I will return to this point and highlight some other potential causal factors below. At Beycesultan, a gap in occupation of up to two centuries had been previously suggested. Ceramic assemblages, however, suggest a strong continuity with the preceding period. EB III structures at Beycesultan include three megaron-style buildings with a large hearth in the centre of the room and a platform adjacent to the entrance. In subsequent levels, a switch towards smaller rectangular structures has been observed (Düring, 2011). In Liman Tepe as well, the monumental administrative buildings within the citadel were abandoned in favour for smaller residential buildings and workshops. A common thread throughout the discussion of archaeological data from EBA Anatolia is the comparative lack of evidence for the south, resulting in an over-reliance on extrapolation from northern sites such as Troy. This observation should serve as a caveat when interpreting the Seshat databank, as much of the information for approximating the values

Social complexity trajectories in Anatolia 139 of the complexity characteristics is not derived from the Konya Plain itself or southwest Anatolia by extension, but rather from contemporary evidence found elsewhere. In and by itself, this is not necessarily problematic, but it needs to be acknowledged, and scholars using this data need to critically evaluate when this extrapolation is warranted or not. It is clear that the macro-level of the Seshat data does not cover the full extent of local and regional variability. If we are to understand why changes in social complexity took place at a given time at a given place, we have to go into the detail of archaeological (and environmental) data to untangle the diversity of drivers of change and localised responses by communities and polities. Urban communities with strong hierarchical organisation developed relatively late in Anatolia, almost a millennium after their emergence in the Syro-Mesopotamian region. This may appear surprising given the close connections between the two regions. Earlier, the spread of sedentary life in the Neolithic took off quickly after its first introduction in the Fertile Crescent (Düring, 2011). The material culture indicates continued close contacts between Anatolia and Syro-Mesopotamia during EBA. The delayed onset of urbanism has led Anatolia to be perceived as a ­‘cultural backwater’ and its eventual development as a second-order effect of the influence from Syro-Mesopotamia (Yakar, 1985). However, this diffusionist interpretation offers no real explanation or understanding of changing social dynamics in Anatolian communities that might have induced this process of urbanism. It should also be reiterated that most of the characteristics of urbanism identified in EBA III, were already initiated to some extent in the Late Chalcolithic, and had clear precursors in EB II. Factors of distinction between EB I–II and EB III are therefore not necessarily signs of a qualitatively different process, but rather of a quantitative difference of scale and intensity. This suggests a gradual process that was likely influenced but not driven by external contacts We must therefore understand what dynamics gave rise to the observed characteristics subsumed under the moniker of urbanism and define how this process is related to changes in social complexity in these communities. During EBA, the transition towards urban communities accelerated and resulted in the establishment of polities with extended spheres of influence. Özlem Çevik noted that BA communities in Anatolia did not undergo a single process of transformation but rather that local and regional drivers of change induced regionalised trajectories of development throughout Anatolia (Çevik, 2007). The oldest urban sites in Anatolia emerged around the middle of the third millennium BCE in the southeast at places such as Kazane Höyük, Titriş Höyük and Karababa. At this time, we observe increasing settlement sizes, public and monumental architecture, fortifications and acropolis structures, specialised economic activities and participation in inter-regional trade. Additionally, a four-tiered settlement system is identified with these urban centres at the apex.9 It should be noted,

140  Social complexity trajectories in Anatolia however, that the tiers are based on settlement sizes collected from survey data and that we have no clear indications for functional diversification or hierarchical relations across these tiers. Studies at Kurban, one of the secondary centres in the area, indicated that it may have functioned as a pottery production centre, whereas no indications of pottery production have been attested at the nearby urban centres. However, we should be careful to link these observations to postulate a functional diversification given that only 19 kilns have been identified over four phases (Arik et al., 1987, p. 205). It is also unclear whether locally produced pottery was distributed to other sites in the area. The absence of indicators for pottery production elsewhere is moreover possibly an artefact of the limited amount of archaeological research at some of these sites. According to Çevik, southeast Anatolia was characterised by more complex socio-economic and political patterns compared to the rest of Anatolia, at least until the second millennium BCE (Çevik, 2007). She argues that outside of the southeast, a two-tier settlement hierarchy existed with the large settlements reaching only a fraction of the sizes in the southeast. She explains this difference by distinguishing between processes of urbanism – which occurred in the southeast – and centralisation – which occurred elsewhere in Anatolia. Whereas the former is defined as a horizontal transformation benefitting all community members, the latter is seen as a vertical transformation expressed by the emergence of a ruling elite. This differentiation is problematic in several ways. First, the numbers for settlement sizes in the southeast seem overblown. It is said that the largest site in the area, Kazane Höyük, reached a massive 100 ha. When looking at the original report, however, it is clear that this figure derives from the results of archaeological surveys identifying EBA III pottery across the full extent of an artefact scatter of 100 ha (Wattenmaker and Misir, 1993). To conclude that this entire area was settled during this time based on this observation alone is surely strenuous at best and most likely untenable. Moreover, Çevik herself notes that Beycesultan reached around 30–40 ha in EBA, falling easily within the range of the larger sites in the southeast.10 Regardless of the validity of that number, this goes directly against her own claim of marked differences in settlement sizes. Second, the definition of urbanism as a process benefitting the entire community is highly problematic as the potential negative effects of urbanism have been extensively documented (Milgram, 1970). Moreover, even if we accept the definition, its claims remain largely unsubstantiated as no evidence is given that this was indeed the case. The definition of centralisation as the direct result of elite formation is equally problematic. The supposed difference between two-tiered and four-tiered settlement hierarchies based on settlement sizes remains meaningless without an understanding of the underlying drivers and functional differences within the settlement pattern that matches the size classes and could point toward emerging settlement hierarchies.

Social complexity trajectories in Anatolia 141 Ultimately, both urbanism and centralisation are driven by flows of energy, resources and information. More specifically, by the concentration of these flows in particular socio-spatial areas. In general, two sources of wealth were available to EBA communities in Anatolia: (1) agricultural production; and (2) participation in inter-regional networks of exchange. The first generates energy, the second resources (in the form of goods and capital). Each has its own, locally defined scope and potential. Local agricultural potential can be high or low, resources have uneven availability, access to distribution networks may be easy or not, and craft knowledge will be available based on previous experience. Each of these factors influences the scope of potential energy and resource exploitation, as well as the scope of actual exploitation. During the transition from the Late Chalcolithic to EBA, the introduction of traction and harnessing technology was an important innovation that would markedly alter the nature of day-to-day farming (Bogaard et al., 2019). It allowed to consistently use other sources of energy outside of human labour. This resulted in a marked increase in farming productivity and the onset of extensive horticulture practices. This had two main consequences. The additional energy surplus allowed rapid population growth during EB I–II, which is reflected in the increase of site number and sizes at this time. Once human labour was no longer the sole limiting factor of agricultural productivity, previously small differences and inequalities between members of the community could be exacerbated over time. Previously, wealth derived from physical prowess tended to converge to the mean over the span of generations. The usage of technology to increase farming yields could be transferred across generations, thus stimulating the development of social inequality and the rise of community elites, as well as sowing the seeds for elite competition within and among communities. During EB III, certain individuals and groups started to convert surpluses from increased farming productivity into social capital expressed through exchange and consumption of resources and goods as part of wider trade networks covering Anatolia, Syro-Mesopotamia, Thrace and the Aegean. The potential role of long-distance trade and information exchange in stimulating social dynamics in EBA Anatolia has long been recognised (Şahoğlu, 2005). Merely postulating the effect of this connectivity as a deus ex machina form of diffusionism does not suffice if we are to understand the causal factors and mechanisms of this process. The underlying pulling forces in these trade networks was the desire of people in Anatolian communities to obtain foreign objects and incorporate them into the own social structure and practices (Bachhuber, 2015). To meet this demand, portions of the available energy and resources needed to be invested towards building and maintaining these lines of exchange. The most important goods flowing through these networks were textiles and metals. Because of their durability, mainly the latter have been

142  Social complexity trajectories in Anatolia preserved. Bronze objects feature particularly frequently, even though bronze has no qualitative advantage over arsenic and copper alloys which were locally available. It has therefore been suggested that the adoption of bronze was not because of functional reasons, but rather as a prestige good (Düring, 2011). Besides metal objects, indirect evidence for the importance of metallurgy can be found in the increased attestations of skeuomorphic ceramic vessels at sites such as Beycesultan. Another tradition of skeuomorphism emerged in south Anatolia in the Konya Plain and Cilicia, named ‘Anatolian Metallic Ware’ characterised by features such as fluting, engraving, repoussé and other decorative techniques to evoke metal objects. Other goods traded between Anatolia, the northern Levant and northern Mesopotamia included ivory, lapis lazuli artefacts and ‘Syrian’ bottles (Massa and Palmisano, 2018). The overall picture of production and distribution patterns in EBA Anatolia consists of regional specialisation in production of goods, framed within a wider system of inter-regional exchange networks (Massa and Palmisano, 2018). It has been posited that one of the main mechanisms of exchange behind these interregional networks was gift-giving rather than entrepreneurial exchanges as part of trading networks (Bachhuber, 2015). Yet, the attestation of converging balance weights at a number of sites throughout Anatolia, including Aphrodisias and Çukuriçi Höyük, suggests some form of institutionalisation of exchange. Analysis on the weights of these objects indicated that multiple regional weight standards were used in Anatolia (Rahmstorf, 2006). Recent studies of the spatial distribution of weight shapes and different units of measurement indicated at least two – partially overlapping – exchange networks in Anatolia during the third and early second millennia BCE. A maritime network connecting the Levantine coast with the Aegean, leading by the southern Anatolian coast, and a terrestrial network consisting of a series of land routes through the Taurus mountains connecting northern Syria with central and western Anatolia (Massa and Palmisano, 2018). The spatial distribution of artefacts with Syro-Mesopotamian origin – concentrated mainly in the areas of origin and in major Anatolian centres – also indicates that the main mechanism of exchange was directional trade rather than down-the-line exchange. While elite demand will have constituted an important pulling force for ‘exotic’ objects into Anatolia, the use of weight standards is a strong indication for the commodification of metals beyond inter-elite gift-exchange. The defining characteristic for the value of metal was no longer (only) craftsmanship, personal relations in exchange, or biographical associations, but also its weight. The focus on weight as the main factor in defining value is also apparent in the appearance of ingots in the form of metal bars and rods, which likely represented standard values. The demand for these goods drove further development of metallurgy and textile production and these innovations acted as transformative industries (Muhly, 2011). Technological innovations in metallurgy included the

Social complexity trajectories in Anatolia 143 development of casting techniques; decoration in repoussé, filigree, granulation, polymetallicism and polychromatic effects. As the scale, intensity and extent of these productive processes increased, more resources were needed to produce objects for distribution over a larger area. This need could be fulfilled by more intensive strategies of resource exploitation or by exploiting a larger area containing more resources. Many regions of Anatolia are rich in metal ores – including copper, silver, lead, gold, zinc, antimony, arsenic, and iron – as well as forests, the two main resources needed for metallurgy. As a result, small-scale local metal production will likely have developed at several locations throughout the land. At Göltepe, for example, a specialised economy emerged focused on the exploitation of ores and metal production (Yener and Vandiver, 1993). It has been suggested that resource exploitation at Göltepe specifically targeted tin sources, but this hypothesis has been contested even though it has never been successfully refuted. It seems likely, however, that Anatolian communities had only limited access to tin and that this was primarily obtained through exchange with the SyroMesapotamian region and the Aegean. From EBA onwards, we see a shift in the main influences on (western) Anatolian metallurgy from the former to the latter (Muhly, 2011). Metallurgy and other production systems played an important role as transformative industries for Anatolian communities. Yet, production systems must be considered within a wider context and can only be considered part of the explanation for societal-wide changes in EBA Anatolia. One marked consequence of the development of productive industries is the need for more people to participate in the production process. This need could have acted as a clear pulling force for population concentration. Several sites in the Konya Plain started to expand beyond the central mound, as evidenced by the presence of ceramic scatters spread far beyond the initial locus of habitation (Massa et al., 2020). Expanding settlements could indeed point towards population nucleation in large centres. The nucleation of population in (fortified) centres during EBA III represented a clear break in settlement patterns from EB I–II where settlement numbers were increasing explosively. The stark reduction of settlement numbers in EB III has also been taken as an indicator of centralisation processes. Population nucleation and settlement growth could only have been possible due to the loosening of the restrictions on information processing constraining settlement sizes in Chalcolithic times. This indicates that some EBA communities moved well beyond the limitations of face-to-face communities by developing increasing levels of social organisation and hierarchical structures. Once this constraint was overcome, several centres started to increasingly take up central roles on a (micro-)regional level and extend their power structures and information environments over a larger area. These processes of central place formation were driven by a desire from local communities for control over alluvial plains with high agricultural potential, the availability of exploitable resources for production processes

144  Social complexity trajectories in Anatolia and access to trade networks for distribution of material goods. As a result, communities were able to direct flows of energy and resources from a larger area towards themselves. It is in this context of rewiring flows of energy and resources that we must see the emergence of citadel sites and central places in the Anatolian landscape. In contrast to Çevik’s suggested two-tier settlement hierarchies for most areas in Anatolia, the survey data from the Konya-Karaman Plains suggests a three-tier hierarchy, consisting of top-tier sites of up to 42 ha, a mid-tier of secondary centres of around 15–25 ha and a low-tier of villages 2–7 ha in size (Massa et al., 2020). Even smaller sites such as hamlets and farmsteads would likely have been present but are possibly not picked up because of the limitations of the extensive survey methodologies applied by the KRASP team. The contemporaneous emergence of fortified hilltop sites in the highlands surrounding the Konya-Karaman Plains – most of which were located in proximity to major thoroughfares with limited access to water and thus unlikely to have housed a normal residential settlement  – suggests a regionally implemented defensive strategy encompassing structures of control beyond the immediate hinterland of the major sites within the plains. This points towards clear functional differentiations between settlement classes. To what extent these patterns are the result of coordination between the main centres on a (micro-)regional level or whether one site obtained some form of primacy (possibly as a primus inter pares) remains unclear for now. Summarising the preceding discussion, EBA – and EB III in particular – Anatolia indeed appears to have been characterised by an increase in social complexity. Some processes had already been initiated in the Chalcolithic, allowing communities to break through the confines of face-to-face communities by developing internal social stratification evidenced by the emergence of a social elite and the formation of a vertical egalitarian social structure even in village communities. As population sizes grew in EBA, communities required more energy and resources to sustain themselves. On the one hand, they started to exploit larger areas as their main catchment. On the other hand, agricultural production was intensified to provide higher yields. The latter was made possible by technological innovations such as the introduction of traction and harnessing technology. EBA communities were increasingly able to rewire surplus flows of energy and resources into concentrated avenues of social competition and conspicuous consumption of goods. Demand for these goods stimulated the emergence of specialised craft production and participation in inter-regional exchange networks. Growing communities, concordant territorial increases and increasing economic opportunities also acted as pulling forces on a (micro-)regional level. This process of centralisation resulted in the establishment of a limited number of sites acting as central places in wider settlement hierarchies. A breakdown of the relevant selection pressures, flows, drivers and mechanisms of these processes can be found in.

Social complexity trajectories in Anatolia 145 Table 4.2  Social complexity trajectories in EBA. *Energy (E), resources (R), information (I); **diversification (D), intensification (IS), and integration (IG) Selection Period pressure

Flows* Process

EBA

Subsistence Harnessing technology

EBA

Demography Population growth

EBA

Governance Growing community sizes Governance Growing community sizes Interaction Growing community sizes Competition Increased social inequality Governance Centralisation

EBA EBA EBA EBA EBA EBA EBA EBA

Outcome

E R I

Farming productivity increase Growing community sizes Decisionmaking hierarchies Territory increase

x x

x

Projecting information environment Social stratification

Central place x formation Production Economic Specialised opportunities production Distribution Demand for Intra-regional foreign goods exchange Subsistence Resource Environmental x exploitation degradation Governance Environmental Community x degradation reorganisation

Drivers

Mechanism**

Pull Push D x

x

IS

IG AC

x

r

x

x

x

K

x

x

x

K

x

x

K

x

x

x

K

x

x

x

x

x

x

x

x

x

x

K x x

x

x x

x

x

x

K x

x

x x

K

K Ω α

From this table, we can clearly deduce a shift from pushing forces driven by diversification that characterised the Late Chalcolithic, to pulling forces driven mainly by intensification and integration during EBA. Once the threshold in limits to information processing was overcome, the development of additional social structures to facilitate information processing and decision making, made a new scale of social development potentially available. This scale increase allowed local communities to intensify the exploitation and processing of all three types of flows, taking up a wholly new role in the landscape as central places for local/regional power structures. During EBA, decision-making responses to selection pressures resulted in an overall phase transition from exploitation (r) to conservation (K) on the macro adaptive cycle. This entails that ongoing dynamics were intensified and became increasingly specialised in redirecting flows of energy, resources and information to fulfil their societal purposes. Once the informational threshold was overcome, a new possibility space for scale increases and associated processes of intensification and specialisation opened up.

146  Social complexity trajectories in Anatolia Towards the end of EB III, we see a first divergence from the unequivocally upward complexity trajectory identified in the Seshat data. At this point, we see a widespread disruption of settlement patterns and site occupation as well as a clear reduction of the characteristics of the social elites that had previously emerged. In the Konya Plain, 35 out of 135 EBA III sites showed evidence for destruction events and subsequent abandonment around 2300 BCE (Massa et al., 2019). James Mellaart famously attributed this disruption to the influx of Indo-European peoples, a hypothesis that has since been discarded. Geomorphological studies have shown that the reduction of site numbers in the Konya Plain can likely be associated with the reduction of agricultural potential due to alluvial infilling. Yet, this does not necessarily explain why we observe similar patterns across all of Anatolia. The observed social and cultural turmoil in many regions of Anatolia have been linked to the so-called 4.2 ka BP event, dated between ca. 2200 cal BCE and 1900 cal BCE (Massa and Şahoğlu, 2015). This large-scale climatic event entailed a short episode of cooling and aridification that affected large parts of the northern hemisphere and has been linked with a series of widespread collapses of agrarian communities. However, its effects would have been highly varied on the micro-scale, depending on local environmental factors and the various ways of decision making in human groups while dealing with these changes. Simulations of rainfall regimes based on modern data, for example, have shown that the coastal regions of south Anatolia would likely have been characterised by large variability in rainfall patterns, including increased risks for drought spells (Roberts et al., 2011). All palaeoenvironmental data proxies show a general trend towards warmer and drier conditions from the mid-Holocene onwards. This overall trend towards warmer conditions was at some points interrupted by significant cold phases. Stable isotope analysis on speleothems from the Kocain Cave in south Anatolia (in the modern-day region of Antalya) indicated the occurrence of a cold phase between ca. 2260 cal BCE and 2180 cal BCE (Göktürk, 2011). Palaeoclimatic studies of other regions in Anatolia likewise show a perturbation of climatic conditions during the late third millennium BCE, matching the timing of the 4.2 ka BP event (Ocakoğlu et al., 2019). But what did this event mean for local communities at the time? At many sites, destruction layers and other signs of violence dated to the end of EB III have been identified. The reduction of settlement numbers at this time  – which I previously explained as the result of strong centralisation ­processes  – has also been touted as a potential consequence of environmental disruption. Likely, several mutually reinforcing factors were at play. I have suggested here that ongoing social dynamics had already triggered a process towards nucleation and centralisation. Environmental stresses related to the 4.2  ka  BP might even have provided additional incentives towards ­centralisation – favouring for example settlements along water ­bodies – r­esulting in the disappearance of many (smaller) settlements in

Social complexity trajectories in Anatolia 147 less favourable topographical and environmental contexts (Massa and Şahoğlu, 2015). Continued population nucleation may have exacerbated the effects of environmental disruption by increasing pressure on the local landscape and its resources such as timber and grazing lands. Palynological data from the Lake District show a strong reduction in oak woodland and proliferation of anthropogenic species during EBA (Bakker et al., 2012). Elsewhere in the Taurus mountain range, we observe a phase of widespread forest clearance as well (van Zeist and Bottema, 1991). In the region of Gordion in central Anatolia, sediment samples from alluvial cores show the highest rates of environmental degradation during EBA (Marsh and Kealhofer, 2014). It should be emphasised, however, that human-­environment interactions are not always simple or straightforward. Despite the observed peak of environmental degradation, the intensity of land use at Gordion during EBA was not nearly as high as during the first millennium BCE. Yet, environmental impact at that time was far lower because of differential landscape vulnerability and sensitivity to disturbances. It can be suggested that the landscape during EBA was poised at a critical state, easily disturbed by initial agricultural expansion and exploitation of natural resources. Once this exploitation was initiated, natural vegetation covers were removed and large reservoirs of accumulated soils became prone to erosion. The evolution of communities during the Chalcolithic and EBA was part of an overall phase transition from exploitation (r) to conservation (K) on a macro adaptive cycle. Once a system transitions into a K-phase, it becomes increasingly vulnerable to shocks, both internal and external. It has been observed that communities participating most in inter-regional exchange networks were amongst the ones hit hardest by the changes at the end of EB III. Perhaps overreliance on these networks may have created higher levels of inter-dependency among groups, not only for luxury goods but also for metals, textiles, salt, agricultural produce, and livestock (Massa and Şahoğlu, 2015). As a result, even communities not directly impacted by environmental changes, may have seen severe disruption through connections with societies facing climatic instability. The fragility of an increasingly interconnected system is one of the most eminent characteristics of a system in the conservation (K) phase. As system rigidity increased, higher human impact on the landscape during EBA crossed the resilience threshold of the environment and triggered a release phase (Ω). At this stage, local communities – even though they themselves might still be in the r or K phases – were forced to undergo a regime shift towards a new adaptive cycle. This shift was associated with turmoil in many communities, driven by subsistence problems, the loss of economic opportunities, loss of legitimacy and authority for elite as ritual mediators and the disruption of gift-giving networks (Bachhuber, 2015). Still, existing elements of the development during EBA contained the seeds

148  Social complexity trajectories in Anatolia for new development in the subsequent period, albeit with some degree of innovation and change. It has been stated that the 4.2 ka BP event induced a bifurcation point, sending west and central Anatolia on different trajectories of development, the former geared more towards the Aegean, whereas the latter drew closer to the Syro-Mesopotamian area (Massa and Şahoğlu, 2015). This divergent development will be discussed in more detail in the next part on the Middle and Late Bronze Age. Middle and Late Bronze Age Following the widespread disruption encountered at many sites during EB III, many (but not all) communities started to recover and new centres emerged towards the end of the millennium. The second millennium BCE – covering the Middle Bronze Age (MBA) and Late Bronze Age (LBA) – saw the emergence of centralised polities such as the Hittite kingdom. In a textbook on Anatolian archaeology, it was even proclaimed that the archaeology of LBA “is an archaeology of imperialism” (Sagona and Zimansky, 2009, p. 266). Following the preceding discussions, we can reasonably expect the Seshat data to show another increase in social complexity at this point. Looking back at Figure 4.3, this expectation is confirmed. In the Seshat databank, the second millennium BCE is subdivided in MBA (2000–1700 BCE), the Hatti Old Kingdom period (1650–1501 BCE), LBA II (1500–1400 BCE), and the Hatti New Kingdom period (1344–1181 BCE). This periodisation hails from a combination of archaeological and historical evidence, using archaeological horizons such as MBA and LBA II, as well as polity-related terms such as the Old and New Kingdom of Hatti. During the Old Kingdom, the Hittites emerged as a main Anatolian powerhouse, whereas during the New Kingdom they expanded their hegemony over much of west Anatolia, Northern Levant, and parts of Upper Mesopotamia (Matessi, 2018). The distinction between these phases is therefore a valid one. It is, however, oftentimes very difficult to match information from textual sources with archaeological evidence. It has also rightfully been noted that “One of the biggest obstacles for the development of a successful archaeology of the Hittite state and empire has been a flawed methodology in which historical questions determine archaeological research agendas and simultaneously provide their principle interpretive framework” (Glatz, 2011, p. 879). I therefore decided not to uphold this subdivision and refrain from mixing archaeological and historical periodisation. Moreover, given that the four periods are plotted closely together at similar levels of complexity, it can be safely assumed that general patterns across these subdivisions will be representative for the overall period. I will therefore work with the general periods of MBA (2000–1700 BCE) and LBA (1700–1200 BCE) to structure the discussion.

Social complexity trajectories in Anatolia 149

Figure 4.7  Main MBA (circle) and LBA (diamond) sites discussed in the text (made by author)

MBA was characterised by the establishment of the Assyrian trade network. From Assyrian texts, we have some indications regarding conflicts between territorial polities in Anatolia. Unfortunately, the archaeological record from this period does not always match up with this level of detail from the textual sources. Several communities – in the scholarly record of this period often denoted as ‘city-states’ – appeared to have had their own local dynasties, headed by a ‘royal couple’. It is generally impossible, however, to link the dynasties mentioned in texts with the settlements attested in the archaeological record. Still, it seems that a limited number of independent polities centred on major settlements dominated many areas of Anatolia at this time, especially on the central plateau. Interregional trade networks continued to connect Anatolia and the Syro-Mesopotamian regions during MBA. However, compared to EBA, the exchange circuit started to shrink considerably and reached only the central Anatolian plateau, while no longer supplying many areas of western Anatolia (Massa and Palmisano, 2018). Whether this was due to a reduced reach of the merchants themselves or the decrease of demand for goods from local elites is difficult to assess. Given that some goods such as ivory still reached some islands in the Aegean (e.g. Crete and Skyros), we might suggest that trade routes were still able to reach these parts and that at least part of the maritime network between the Levantine coast and the Aegean was in use. Land routes through the Taurus mountains in west Anatolia, on the other hand, be it for lack of demand or other reasons, were no longer in use.

150  Social complexity trajectories in Anatolia The most important commodities were metal, wool, and textiles from the east in exchange for Anatolian silver and gold bullion. The driving forces behind this exchange network were Assyrian merchants organised in family firms rather than interpersonal gift-giving of EBA. These merchants settled in trade localities within local communities, called kārums, prompting the alternative term for this era as the Kārum period (Michel, 2011). The commercial and administrative centre of the Assyrian trade network was Kaniš (modern Kültepe) in central Anatolia, which was furbished with a large palace and several temples. Besides different driving actors, the main difference with EBA exchange networks was the increased institutionalisation and emergence of formal infrastructure for MBA networks (Dercksen, 1996). Caravans followed fixed routes using transport infrastructure such as roads, bridges, ferries, inns, relay stations and guards along major routes. Along the way, they paid tariffs to local kingdoms to ensure their safety and contribute to proper maintenance of the trade infrastructure. Standardised weight systems were used to facilitate inter-regional trade contacts. Local Assyrian trade enclaves were present in many Anatolian communities, ensuring a direct trading presence. Assyrian authorities at the centre of Kaniš also acted as mediator in conflicts between Assyrian merchants and local communities. This institutionalisation of inter-regional exchange networks is not necessarily an MBA innovation. It is likely that EBA exchange was supported by some forms of institutions as well. The regional weight standards used at that time indeed suggests some degree of institutionalisation beyond individual communities, but it does not seem to have reached the same level as in MBA. The presence of an extensive body of written documents from MBA Anatolia – most notably from Kaniš – goes a long way in explaining the seemingly higher level of institutionalisation compared to EBA for which we are fully reliant on archaeological evidence. On the other hand, the existence of a system of written documentation would have allowed more extensive institutionalisation of trade in the first place. The disconnect from the Assyrian exchange network was part of a wider trend of increasing divergence between west Anatolia and the central/eastern areas. It has been noted that in many of the main centres in the west we see a continued reduction in settlement size and disappearance of monumental public buildings. On the central plateau, on the other hand, we see a gradual transition from relatively small EBA polities to territorial citystates serviced by a complex administrative apparatus (Massa and Şahoğlu, 2015). Rank-size analysis of settlement patterns in central Anatolia indicated that, at the regional scale, the area was characterised by a convex distribution, suggesting a politically fragmented landscape of competing polities that were only loosely integrated (Palmisano, 2017). When breaking down the area on the local level, rank-size distributions follow a primate distribution, indicating that each of these polities was strongly politically

Social complexity trajectories in Anatolia 151 and economically centralised, focused on large centres dominating the surrounding hinterlands. In the Konya and Karaman Plains, many of EBA centres were abandoned towards the end of EB III. During MBA, we see the emergence of 14 new major sites (larger than 15 ha), mostly ‘flat-topped, steep-sloped mounds’, indicating the presence of soil-retaining fortification walls during MBA (Massa et al., 2020). Further evidence for such fortification walls has been found through ground walking and satellite imagery, but only at Konya-Karahöyük they were also attested through excavations. This site also contained several monumental public buildings. The attestation of two archives of door-sealings suggests that at least some of these buildings had an administrative function for tracking the storage and redistribution of food and other commodities. Possibly, a lower town spreading beyond the central mound was present as well, resulting in a total occupied area of up to 40 ha. Burials found at the top level of the site contained not only inhumantions in pithoi graves – as was common in Anatolia up to this time – but also cremations covered with pithos sherds were attested for the first time on the Anatolian plateau. The presence of two different burial systems within the same community has been suggested to indicate two groups with different belief systems but this remains hard to prove and it has been shown from ethnographic studies that this need not necessarily be the case. Other important sites in the Konya Plain reaching their maximum extent in MBA were Domuzboğazlayan Höyük, Büyük Aşlama Höyük, Karaman Kale Höyüğü, and Sırnalı Höyük. Most of these sites were located in the alluvial plains where water availability was higher compared to other areas in the region. On the Karaman Plain, no large sites have been identified that could be dated to MBA or LBA. In the surrounding highlands, 12 hilltop sites were attested, suggesting the continuation of EBA hierarchical settlement patterns aimed at maintaining control over the alluvial plains. Elsewhere in southwest Anatolia, MBA is less-well represented in the archaeological record. Soundings on the acropolis of Perge have yielded several walking levels dated to MBA based on associated pottery material and 14C samples (Martini and Eschbach, 2017). The presence of mudbrick walls suggests some form of settlement during MBA, but unfortunately the dense occupation during later phases makes it very difficult to reconstruct the settlement lay-out and organisation. The presence of cult objects such as libation jugs at the location of the later sanctuary of the acropolis points towards a long tradition of ritual activities at the site, starting from EBA and continuing into MBA as well. For the Burdur Plain in the Lake District, a study of the fabrics and typology of the pithos fragments and associated fine wares from EBA cemetery at Gavur Evi Tepesi suggests that this cemetery remained in use until the end of MBA. However, it is striking that elsewhere in the Lake District hardly any traces of human activity dated to MBA have been attested (Kaptijn et al., 2012; Vandam, 2019). The reason for this absence remains unclear for now.

152  Social complexity trajectories in Anatolia The transition from MBA to LBA (1700–1200 BCE) is poorly attested in the archaeological record of Anatolia, for a large part due to the difficulties in distinguishing material culture from both phases. The most notable event during this transition is the collapse of the Assyrian trade network, for reasons that are hitherto unknown. Following the end of the Kārum period, LBA is mostly known for the emergence of the Hittite kingdom in central Anatolia around 1650/1600 BCE, which lasted until the end of the 13th century. During the 14th and 13th centuries BCE, this kingdom covered most of the central Anatolian plateau, stretching from the Pontic Mountains in the north to the Taurus Mountains in the south. Its capital was located at Hattuša near the Halys river. Besides Hattuša, the Hittites established a series of provincial centres, such as Kuşaklı in the north and Kayalıpınar in the east. The spread of Hittite culture is most prominently reflected archaeologically in the distribution of monumental architecture and rock reliefs dated to the 13th century, ranging from Akpınar on the west coast to Kayalıpınar in the east. Hittite pottery shapes drew extensively from local Anatolian traditions, reflecting earlier patterns from EBA and MBA. A common form of material culture was distributed across large parts of central Anatolia. In addition to bronze metallurgy, iron objects were used as symbols of power by the Hittite kings (Seeher, 2011). Textual sources relate that taxes were levied in kind based on agricultural yields, which were siphoned off, stored in large silos and redirected towards the central administration, often on a regional level (Siegelová, 2001). Several specialised officials were tasked specifically with performing these functions. Interestingly, grain deposits retrieved from two monumental buildings at Kuşaklı-Šarišša may provide evidence for these taxation payments. It has been argued that the generally poor quality – small grain size and a relatively large amount of weeds intermixed with them – of these cereals could suggest that these yields came from fields that were designated for taxation and therefore less well tended for. This might suggest some form of local resistance against state control (Glatz, 2011). Most of the Hittite heartland falls beyond the scope of this case study and will not be discussed in detail. However, at its peak the Hittite kingdom likely included large parts of the south as well. Moreover, most information on polities and population groups in the west and southwest of Anatolia is derived from the archives at Hattuša. The most prominent example is that of the Arzawa, a term interchangeably used to denote a polity, a group of peoples and a geographical area. Multiple references to conflicts with kings of these peoples and conquests of their lands (or vice versa) have been attested (Bryce, 2011). Arzawa even stood on the brink of replacing the Hittites as the main power force in Anatolia at some point during the 14th century BCE, prompting the initiation of diplomatic relationships with the Egyptian pharaoh Amenhotep III, only to see their military advantage scuppered by a series of military counter expeditions

Social complexity trajectories in Anatolia 153 led by the Hittite prince Šuppiluliuma. Unfortunately, we only know about the Arzawa state from Hittite perspectives and it remains difficult to locate this state geographically and archaeologically. The lands mentioned in the Hittite archives generally stretched from the Troad in the northwest Aegean, along the coastal regions, possibly extending to the western edges of the Konya plain. An important royal seat of the Arzawa was located at Apaša, which could be identified with the site at Ayasuluk near Ephesos (Büyükkolancı, 2000). It is generally accepted that these lands were inhabited by groups of Luwian-speaking peoples during the second millennium BCE. To what extent this entire area can be considered a coherent geopolitical unity is difficult to assess given the liberal use of the term. After gaining a military victory, the Hittite king Muršili (ca. 1321–1295 BCE) dismantled the Arzawa kingdom and instated four Hittite client polities in west Anatolia: Mira, Šeha River Land, Wiluša, and Hapalla (Figure 4.8). The latter was likely located in central-west Anatolia, north of the Hittite Lower Land in the area of the Konya Plain. It has been suggested that the capital of Mira was located at Beycesultan and that its territory extended until a mountain pass 28 kilometres outside of Izmir, marked by an inscribed monument (Hawkins, 1998). The Šeha River Land was likely located towards the north of Mira, whereas Wiluša has been tentatively placed in the Troad and is associated with ancient Troy.11 Beyond direct Hittite control, at least three more polities existed in western Anatolia: (1) Maša in the northwest of Anatolia; (2) Karkiša, which can perhaps be etymologically linked to Caria in the southwest; and (3) the Lukka people

Figure 4.8  LBA polities in Anatolia (made by author)

154  Social complexity trajectories in Anatolia inhabiting an area extending from the western end of Pamphylia through Lycaonia, Pisidia, and Lycia. Many of these identifications are to some degree provisional and it remains difficult to conclusively link polities from textual sources with archaeological realities. The textual record presents us with some tantalising hints on policies of diplomacy and warfare as part of imperial interaction between these polities. It goes to show how much of the richness of social, political and economic interactions remains underexposed when focusing on the archaeological evidence alone. We can only wonder to what extent this world of competing kings and polities would have stretched back into earlier times. There is, however, no point in lamenting this, so let us have a look at what archaeological evidence is available. Regional archaeological surveys across Anatolia show a general pattern towards less densely occupied landscapes. Site numbers in MBA and LBA dropped to less than half compared to the earlier EBA level (especially compared to EB I–II) (Glatz, 2011). In much of central Anatolia, important MBA centres such as Acemhöyük and Alişar Höyük were replaced at the top of regional settlement hierarchies during LBA by previous secondary centres or newly established sites. Towards the south, in the Konya and Karaman Plains, a similar process occurs with Konya-Karahöyük experiencing a considerable contraction in size following a destruction event. However, beyond the decline of this prime centre, stable occupation patterns are observed and the other main MBA centres continued to be occupied in LBA. The major site at TürkmenKarahöyük, on the other hand, expanded from 30 ha to encompass an area of up to 125 ha, more than three times the size of the second largest site in the area at the time. Additionally, 14 hilltop sites were identified in the surrounding highlands. Outside of the decline of Konya-Karahöyük and emergence of Türkmen-Karahöyük, the general picture is one of continuity in settlement patterns between MBA and LBA. The Konya and Karaman Plains were situated at the periphery of the Hittite world during the first centuries of its emergence. It has been suggested that there was limited degree of Hittite intervention in this region up until ca. 1400 BCE (Matessi, 2018). The southernmost Hittite monument has been found at Eflatunpınar, a ritual pool complex located 85 kilometres west of the Konya Plain, furnished with a stone monument containing Hittite iconography. The monument is dated to the 14th or 13th centuries BCE based on art historical and architectural features. Stone monuments like these were most densely present in the urban centres of the Hittite heartland. Outside of these urban centres, however, they were mainly located in the peripheral areas of the Hittite sphere of influence. Using ‘costly signalling theory’, Glatz and Plourde (2011) argued that their use was not part of a commemoration of Hittite hegemony over these lands, but on the contrary, a medium to mediate ongoing territorial contestations along important communication routes and territorial boundaries. They were often not erected by the central authorities themselves, but rather by

Social complexity trajectories in Anatolia 155 local agents emulating Hittite iconography as part of the establishment of their own power strategies. It is in this regard interesting to note the spectacular growth of TürkmenKarahöyük at this time. More significant changes in material culture of the Konya Plain occurred during the 14th and 13th centuries BCE, roughly contemporaneous with the establishment of the ‘Lower Land’, a Hittite province which likely included the Konya and Karaman Plains as well as parts of the Lake District. It is possible that as part of the political reorganisation of local governmental structures, the Hittites directly intervened in the area to create new administrative centres. In this case, the marked increase in settlement size of Türkmen-Karahöyük may be an indication that it fulfilled this role. The attestation of central Anatolian pottery such as the Eggshell Ware and Drab Ware at the site further points towards close connections with the Hittite heartland (Osborne et al., 2020). The fact that no changes in settlement patterns are observed beyond the shift at the top from KonyaKarahöyük to Türkmen-Karahöyük piles further evidence on the suggestion of a top-down intervention. The site at Türkmen-Karahöyük has been tentatively identified as the capital of LBA polity of Tarhuntašša (Goedegebuure et al., 2020; Massa et al., 2020; Osborne et al., 2020). In the early 13th century, Muwatalli II moved the capital of the Hittite kingdom from Hattuša to Tarhuntašša, possibly as a reaction against shifting power balances in the Aegean and the Levant and the desire for a more centrally located power base (Matessi, 2018). If this identification is correct, it would mean that the observed centralisation and central place formation in the Konya and Karaman Plains would have been driven by this local polity, which could have extended all the way until the Göksu valley, the site of Kilise Tepe and the harbour at Ura to the south, and the Beyşehir and Seydişehir plains to the east. Türkmen-Karahöyük is the only LBA site somewhat approaching the size of Hattuša (180 ha). It is striking that other important LBA centres were fairly small compared to their EBA and MBA counterparts. Only a handful of sites grew larger than 20 ha, and these appear to have been some of the highest-order regional centres such as Kayalıpınar, which may be identified as Šamuha, the capital of the Upper Land. The restricted sizes of secondary centres within the Hittite sphere of influence may indicate that these centres were not able to fully exploit the potential of their hinterlands, instead having to redirect part of the flows of energy and resources towards the capital. The fact that Türkmen-Karahöyük breaks this trend my provide further evidence for it attaining a primary status at some point during the 13th century. Continuing the trend from MBA, virtually no material from LBA has been found in the Lake District (Vandam, 2019). The reasons behind this millennium-long absence of material evidence remain unclear.12 On the Pamphylian and Lycian coasts, LBA habitation has been attested in recent soundings at the acropoleis of Perge and Patara (Yildirim and Gates, 2007).

156  Social complexity trajectories in Anatolia At Perge, terraces had been laid out to allow construction of several structures, likely related to a sanctuary building as in later times. Evidence for cult activities is also found in material finds such as libation objects. Pottery material found in LBA layers points towards extensive contacts with the outside world, including Mycenaean and Cypriotic traditions such as the Red Lustrouw Wheelmade Ware, as well as material with parallels in Hattuša, Beycesultan and Demircihöyük (Martini and Eschbach, 2017). The nature of the relationship between the local community and the Hittite kingdom remains unclear. It has been suggested that Perge could be identified as the Hittite site of Parha, listed as a border town in a late 13th century bronze treaty table from Hattuša. While the identification remains tentative for now, the material culture of the site at least suggests that the community at Perge – as in later times – was well integrated in the wider eastern Mediterranean world. Unfortunately, our knowledge about the community itself is still relatively limited. At Beycesultan, two occupation levels (III–II) have been dated to LBA. Around 1300 BCE, major construction works were undertaken, resulting in the layout of several (domestic) buildings alongside a street network, possibly driven by collective planning strategies. A strong continuity has been observed in terms of ritual practices – both in location of cult architecture and style of material culture – suggesting a strong socio-cultural continuity with MBA community (Lloyd and Mellaart, 1958). At the centre of the mound, a central building with pillared supports and dressed masonry continued to be in use. The find of stamp seals at the building suggest an administrative function or other community-related purposes (Lloyd, 1972). Elsewhere at the site, finds of ornamented daggers, a decorated axe head, a serpentine bowl and a decorated horse harness can be considered prestige items, indicate the existence of a socially stratified society. In short, the community at Beycesultan seems to have sustained the same lines of community formation and complexity trajectories initiated in EBA and continued in MBA. The main difference with other sites at the time is the emphasis on an extremely localised sense of community identity (MacSweeney, 2011). Even though, imports from other areas in the eastern Mediterranean are still attested, material culture generally seems to have been embedded in local trends and shapes, having little affinity with material from the Hittite kingdom or the Arzawa polities. This has been interpreted as an expression of a strongly localised sense of community identity, possibly as a reaction against potential Hittite hegemonic strategies. In the next level, dated to the late 13th and early 12th centuries BCE, we observe an increased level of prosperity, social differentiation and centralised control over resources (MacSweeney, 2011). This phase marks a shift from collective organisation to increased social differentiation and inequality. At this point of time, the Hittite kingdom had already collapsed and its shadow would no longer have loomed over the area in general, and the Beycesultan community in particular. It is therefore interesting to note

Social complexity trajectories in Anatolia 157 that the general response to the resultant shift in power balances – perhaps induced by the lack of external threats – was to increase internal inequality and differentiation. At Aphrodisias, located in the same river system as Beycesultan, a markedly different trajectory has been observed, where initial collective action strategies did not give way for increased internal differentiation, but were rather emphasised in localised material culture and socially integrative events intended to strengthen community-structures and collectiveness (MacSweeney, 2011). The reasons behind this divergent trajectory are not clear, but has been suggested to be related to the smaller nature and lower-profile of the community at Aphrodisias compared to the larger site of Beycesultan. The latter was also located on a major east-west thoroughfare in the Anatolian landscape and therefore far more intensively connected to developments within the larger macro-region. It can be concluded that the main development in LBA was the emergence and expansion of the Hittite kingdom and that our knowledge from many areas in southwest Anatolia remains limited. The emergence of the Hittites constituted a transformation of the Anatolian socio-political landscape, from a host of independent local and micro-regional polities to a number of regional-level polities embedded in inter-regional structures of power under a macro-level polity (Matessi, 2018). It was probably no coincidence that the seat of this kingdom was located in central Anatolia, a region characterises by high centrality in existing communication networks between west Anatolia, the Levant and Mesopotamia. This central location would have allowed local power figures to gain control over extensive flows of energy, resources and information. Perhaps it can even be suggested that the Assyrian trade network collapsed because of a shift in power structures behind the actors involved. With the emergence of powerful regional and interregional polities in Anatolia, local actors may have professed stronger interests in having direct control over flows of goods across Anatolia, not in the least in function of taxes and state administration. The overall picture of the Hittite kingdom is one of a complex web of social, cultural, economic, and political connections across multiple scales of interaction, rather than a monolithic, centralised entity. This multi-dimensional configuration included extensive regional power structures and a series of neighbouring polities in west and south of Anatolia connected to the main state through a complex set of relationships characterised by varying degrees of dependence. This dependency only rarely entailed complete Hittite control, but rather what has been called a policy of ‘intensive hegemony’ as an intermediate stage between direct territorial control and indirect hegemonic rule (Glatz, 2011). During the 14th and 13th centuries BCE, the Hittite sphere of influence stretched from the Aegean coast in the west to the Levant and Upper Mesopotamia in the east. At this point, the Hittites established direct exchange relationships on equal terms with longstanding powers in Eastern Mediterranean such as Egypt (Matessi, 2018). Beyond royal exchange relationships, essential channels of communication opened

158  Social complexity trajectories in Anatolia up. For example, going from the Konya Plain through the Goksu valley towards the harbour site of Ura on the south coast with a direct maritime link to the Levant. Letters from the Hittite king to Ugarit suggest that Ura might have played the role of intermediary in flows of grain supplies during food shortages. This communication channel may also have funnelled the spread of Red Lustrous Wheel-made from Cyprus throughout Anatolia from the 15th century onwards (Grave et al., 2014). The disintegration of Hittite power around 1200 BCE marks the end of BA Anatolia. This relatively short period (1200–1150 BCE) is typically described in terms of widespread socio-political and cultural upheaval and collapse. Widespread destruction events are noted in sites throughout the eastern Mediterranean, including the Aegean, Egypt, Levant and Anatolia. Traditionally, the (partial) cause of this collapse is said to have been the raids of the infamous ‘Sea Peoples’ (Sandars, 1978). After this disruption, it would take several centuries before writing systems, urban communities, macro polity structures and long-distance exchange networks would reach the levels of the BA in the eastern Mediterranean. The period in-between the complex societies of LBA and Early Iron Age has been denoted as the ‘Dark Ages’. Recent studies, however, have increasingly identified rays of light in this supposedly dark time and the use of the term is now largely restricted to a historiographical one, referring to a historical period rather than imposing judgement on social structures and complexity at the time. I will not use the term Dark Ages as part of the periodisation in this book, but will rather refer to the Early Iron Age from 1200 BCE onwards. Before moving on to the Iron Age, however, let us first consider the potential causes for LBA ‘collapse’ in Anatolia and what this meant for the complexity trajectories of this period. Some have argued that the downfall of the Hittite kingdom was a gradual process rather than a sudden event. This is mainly supported by a re-analysis of destruction traces at Hattuša, showing that there is little evidence for a single synchronised event causing widespread destruction (Seeher, 2001). It remains of course difficult to hold this discussion in terms of absolute dates. Yet, even if it was not a single event, the available evidence does point towards a relatively rapid process in the order of decades rather than centuries. This observation also raises the question whether we should discuss this episode of change in terms of collapse or transformation.13 One of the more frequently cited causes are the consequences of ‘imperial overstretch’. Having extended their power into parts of the Levant and Mesopotamia, the Hittites clashed with Egypt over the control of Syria. This culminated in the famous battle of Kadesh during the reign of Ramses II the Great. Neither side walked away from that battle unscathed, however, Hittite power started to wane following the rise to power of the Assyrians. Having lost their footing in Mesopotamia, the Hittites were increasingly forced back into their Anatolian heartland, receiving the death blown from the combined efforts of the ‘Sea Peoples’, Assyrians, Phrygians, and others.

Social complexity trajectories in Anatolia 159 Clearly, this is the macro-level historical account which we find in most textbooks. While this accounts for some of the macro-level patterns of political history, it does not offer much of an explanation for a lot of the dynamics we see on local and regional scales. It has also been suggested that the cause for the demise of the Hittite kingdom could perhaps be found in the endemic instability of centralised economies of production and distribution and its failure to respond to the growth of decentralising ‘fringe’ economic activities. Scholars particularly point towards the growing role of small-scale Cypriot traders bypassing state-controlled long-distance exchange networks in the eastern Mediterranean from the 13th century BCE onwards (Sherratt, 2003). The impact of Cypriot commercial enterprises was most clearly visible in the increasing availability of iron prestige objects, resulting in a devaluation and widespread distribution of bronze for utilitarian goods. The problem with this argument, however, is that it rests squarely on the assumption that the wealth and power of LBA polities were largely derived from their direct participation in long-distance exchange networks. Given that the main economic sector for most of antiquity rested in agricultural production, this hypothesis seems to overstate the importance of trade networks. Some have highlighted changing environmental circumstances and their impact on agricultural potential as important causal factors. During the second millennium BCE, an ongoing overall drying trend culminated in the establishment of more arid conditions during the late second and early first millennia BCE (Roberts et al., 2011). This would have resulted in more frequent occurrences of major drought phases and declining water resources. Analysis of palaeoecological and archaeological data to reconstruct human-environment interactions in the past have identified the so-called “Beyşehir Occupation Phase” (BOP) (Eastwood et al., 1998). This episode of environmental change was characterised by the clearance of tree species such as deciduous oak, pine, cedar, juniper, and fir forests, and increases in open habitat and weedy species, along with the appearance of anthropogenic indicators such as cultivated trees, including olive and sweet chestnut (Boyer et al., 2006; Woodbridge et al., 2019). The BOP marks a phase of initial anthropogenic impact and the changes it induced on the landscape. It is crucial to note, therefore, that it was a phase of environmental change, not a climatic one. This means that its emergence and effects were established at different times at different places, depending on localised trajectories of human-environment interactions. It also means that we have to be careful with using environmental changes as causal factors for widespread effects on socio-political configurations. Whereas the overall climate trended towards drier circumstances, this need not necessarily have impacted every region in exactly the same way. So, what does all that mean for social complexity trajectories in MBA and LBA Anatolia? In the trajectory of Figure 4.3, both MBA and LBA constitute another increase in complexity, albeit not as much as EBA. I already noted in the previous part that this continuously upward trajectory does

160  Social complexity trajectories in Anatolia not correspond with the archaeological evidence on a local and regional level. At some point during EB III, many communities went through a process of transformation and reorganisation. By MBA, however, communities had largely recovered and returned on a process towards micro-regional polity formation driven by internal social stratification, craft specialisation and participation in inter-regional exchange networks. In terms of adaptive cycles, this development constituted the initiation of a new cycle and first exploitation of available resources in the r-phase (see Table 4.3). Higher social complexity is mainly found in the institutionalisation of trade networks. Institutionalisation and associated administration enhance the flow of information, standardises information transmission and facilitates information processing to streamline every individual act of exchange. We have to keep in mind, however, that some form of institutionalisation would likely have been present in earlier times as well but that we have no record of them comparable to the Assyrian documents that give us so much information for MBA. Increasing institutionalisation by elaborating existing system structures is a typical process for the K-phase of development. It is only in LBA, with the emergence of the Hittite kingdom, that we see a genuine increase in social complexity through the addition of a higher scale of socio-political organisation. On the macro level, complexity trajectories were driven by centralisation of economic and political power in territorial polities. However, rather than a single centralised entity, Hittite imperial policies were grafted on regional power structures. These power structures pulled regional flows of energy, resources and information from a larger territorial area towards the capital. These flows were driven by top-down implementation of policies rather than a bottom-up emergence of complexity structures. The identification of potential acts of local resistance towards the central authority is a case in point. The main selection pressure behind these complexity trajectories was governance (in all its facets) and the development of imperial policies, which functioned as pulling forces redirecting these flows. Imperialism, generally defined as the policies of empire, can be conceptualised as “processes that underlie recurring episodes of individual and collective interaction on a multitude of socio-political and cultural levels” (Glatz, 2011, p. 883). The emergence of a macro level of socio-political organisation therefore results in “­ multiple overlapping and intersecting socio-spatial networks.” (Mann, 1986, p. 1). It is through this intersection of networks that the surplus of energy, resources and information can be redirected. For example, the central administration only rarely acted directly on the local level, but mostly relied on the establishments of regional centres as an intermediate level in governance structures to implement its policies. It should be noted, however, that this observation is only valid for those areas under Hittite control. In the area of our case study, only the Konya and Karaman Plains show indications for higher complexity compared to EBA. However, it has been suggested that these lands were, if not under

Table 4.3  Social complexity trajectories in MBA and LBA. *Energy (E), resources (R), information (I); **diversification (D), intensification (IS), integration (IG) Flows*

Drivers

Process

Outcome

E

R

I

MBA MBA MBA

Interaction Distribution Governance

Energised crowding Inter-regional trade Centralisation

x

x

x

x

x x x

x x x

LBA LBA

Distribution Governance

Shifting power structures Centralisation

x

x x

x x

x

LBA LBA LBA LBA LBA LBA LBA

Governance Production Governance Governance Competition Competition Governance

LBA

Competition

Polity collapse

x

LBA LBA

Governance Governance

Polity policies Technological innovation Central place formation Taxation Peer-polity interaction Territorial contestation Expansion of control structures Internal and external stresses Polity collapse Polity collapse

Community formation Institutionalisation Micro-regional polity formation Collapse trade networks Supra-regional polity formation New settlement hierarchies Craft specialisation Secondary centres Redistribution of yields Regional polity formation Costly signalling Hegemonic policies

Internal stratification Community formation

x x x

x x x x x x

Pull Push

x x x

x x

x

x

x

x

x x x

D x

x

x x

x

x

x x

x x

Mechanism** IS

IG AC x x

r K K

x

Ω K

x x x x x x

x x

K K K K K K K Ω

x

α α

Social complexity trajectories in Anatolia 161

Period

Selection pressure

162  Social complexity trajectories in Anatolia direct Hittite control, at least part of the Hittite sphere of influence and part of another regional polity. One might raise a counterpoint and state that the other kingdoms in Anatolia, even though we have less information on them, interacted and competed extensively with the Hittite kingdom and could have created similar structures of governance. This would mean that we can indeed take information from the Hittite kingdom and extrapolate this to the rest of contemporary Anatolia. However, this would still be an inference that is not supported by the available evidence (at least for now). It can be suggested that the very driving forces of polity formation stimulating the emergence of the Hittite kingdom inherently also carried the seeds for its later dissolution. The highly regionalised structures of power did not only provide pulling forces to redirect energy, resources and information towards the central administration, but eventually also acted as pushing forces that destabilised the macro polity by stimulating the emergence of regional polities and secondary centres of power. This competitive interaction siphoned off energy and resources, for example through costly signalling strategies, that could have been spent to sustain polity formation. Polity formation was also constrained by the restrictions on carrying capacity and energy availability on the local level. The dwindling of settlement numbers during LBA and limited settlement sizes compared to earlier times suggests that, not only part of the available potential was redirected, but perhaps also that stricter environmental constraints on agricultural potential were starting to be felt. Local communities started to struggle to maintain the flows of energy and resources to sustain regional and macro-regional polities due to the effects of ongoing climate change culminating in aridification of the landscape, which markedly increased pressure on the landscape. The need for more arable and grazing land, and the use of timber as fuel resulted in widespread deforestation and intensified land use to make up for the losses in agricultural productivity. Lifestyles initiated in EBA were becoming seemingly harder to sustain in MBA and LBA (Allcock, 2017). The development of administrative structures, taxation and food redistribution strategies in the Hittite kingdom could suggest increasing efforts to mitigate the effects of drier climatic conditions and poorer harvests following droughts. Yet, over-reliance on natural resources, diminishing returns to investment and increased lock-in of resources to higher socio-political entities would have drastically impacted the resilience levels of local communities and reduced their capacity to adjust to internal and external stresses. Continued dependency on crops that became increasingly hard to farm in dryer circumstances likely resulted in more instances of crop failure. This is where a system firmly locked in the K-phase of conservation really started to reach its limits. Ultimately, this configuration proved to be unsustainable and a release phase resulting in the collapse of the Hittite kingdom ensued. The collapse of this macro-level polity resulted in the loss of social complexity related to those structures impacting inter-regional flows of energy, resources and

Social complexity trajectories in Anatolia 163 information. However, it need not necessarily have impacted existing complexity structures on lower scales. On the local level, we see communities responding to the collapse of higher polities in various ways. Some communities like Beycesultan intensified internal social stratification, possibly as a result of increased security following the loss of an external threat. Additionally, it is possible that certain members of the community managed to redirect part of the flows previously going towards the central administration towards fuelling their own position. Others, like Aphrodisias, responded by stressing collective identities and community formation, effectively reducing internal social inequality, very much along the same lines as the early communities in the Chalcolithic and EBA. This reorientation entailed the beginning of a new cycle, using different components of the previous cycle to reinvent themselves and continue their reorganisation as part of the brave new world known as the Iron Age. Iron Age In the complexity trajectory of Figure 4.3, the transition from LBA to Early Iron Age (EIA: 1200–800 BCE) is the first point where we see a clear loss of complexity. Given the collapse of the Hittite polity and wider turmoil in the eastern Mediterranean, there is some basis for this from a macro scale perspective. However, as we have seen – and will see in this part as well – this need not necessarily be relevant on other scales of analysis. In the Seshat data, the period between 1200 and 800 BCE is assigned to the Neo-Hittite kingdoms and the Tabal Kingdom. These Syro-Hittite polities emerged in the Syrian areas previously controlled by the Hittite kingdom. There is little indication, however, that these ever firmly established authority over the Konya Plain or other areas relevant for the present case study. I will therefore omit them from the discussion. Given that these phases are coded more or less on the same level of complexity as the subsequent Phrygian period, I believe the omission to be permissible and I will continue with a brief discussion of the Phrygians. The Phrygian kingdom was centred on the site of Gordion on the central Anatolian plateau and originated around the turn of the millennium. Its heyday is situated in the 9th and 8th centuries BCE. As with the Hittite kingdom, the core area of the Phrygian kingdom lies beyond the spatial demarcation of the present case study. However, at its point of maximum territorial extent the kingdom likely reached the area to the immediate north of the Konya Plain. I will therefore provide a brief overview of the main trends. In contrast to the Hittite kingdom, the Phrygian kingdom mainly remained a land-locked state, having no direct access to the Mediterranean, except for the Marmara region for a brief period in time during the 8th century BCE. Our understanding of how this kingdom emerged and reconfigured the political and economic structures of Anatolia after the Hittite collapse

164  Social complexity trajectories in Anatolia

Figure 4.9  Main EIA (diamond), MIA (circle), and LIA (triangle) sites mentioned in the text (made by author)

remains limited due to the lack of excavated sites. During the transition from LBA to EIA, Gordion transformed from the typical BA community sketched in the previous part to a small village community characterised by subsistence economy, household production and little external contacts. Even though there is no evidence for a hiatus in occupation, the shift in community structures have led some to suggest that LBA population of Gordion moved away and a new group occupied the site shortly afterwards (Voigt, 2011). However, there are no positive indicators that would support this suggestion over the simpler option of the local community adapting to changing circumstances. At the turn of the millennium, marked changes are observed in the architecture of Gordion, such as the construction of buildings with stone foundations around an open court, fortifications, stone pavements, and a series of public buildings. Most of the citadel (13.5 ha) was occupied by the end of the 9th century BCE. At this time, it was divided in two main zones, on the one hand, a palace complex with an administrative quarter and an area devoted to textile production and grain processing, and on the other a residential area (Rose, 2017). Some rooms in the industrial zone contained up to 600 loom weights and at full capacity this complex alone could have housed about 300 workers (Burke, 2005). Extensive imports of prestige goods are also attested, in the form of fine pottery wares and metals. The outer fortifications surrounding the citadel, Lower and Outer Towns, covered an area of about 103 ha. Surrounding the wider landscape of Gordion were more than 200 monumental tumuli.

Social complexity trajectories in Anatolia 165 Towards the end of the 9th century BCE, the height of parts of the citadel was raised 4–5 m above its previous level, increasing its monumental nature significantly. This massive construction project was estimated to have required the excavation and movement of more than half a million cubic meters of clay (Rose, 2017; Voigt, 2012). Clearly, this required the mobilisation of a massive labour force. The project was interrupted when the citadel was struck by a destructive event that has been linked to an invasion by the Cimmerians, a people from the Black Sea area, around 700 BCE. However, it is now generally accepted that the event should be dated to 800 BCE and was not related to an external threat but most likely the result of a fire that originated in the textile production area and had gotten out of control (Rose and Darbyshire, 2011). Following the destruction, a rebuilding phase was initiated. Throughout the remainder of the IA, the area would remain in a constant state of restructuration (Voigt, 2012). The extensive display of power and capital investment at Gordion during the 9th and 8th centuries BCE indicates that it must have been a powerful polity in central Anatolia at the time, able to derive energy and resources from a large area (Voigt, 2011). Geomorphological studies indicate that, even though settlement intensity seems to have increased,14 the degradation rate of the landscape decreased markedly to about 1/3 of its peak in LBA (Marsh and Kealhofer, 2014). It is likely that the initial rise in settlement intensity during EBA and MBA resulted in the primary loss of soils sensitive to erosion during LBA. Once these vulnerable soils were depleted, the remaining soils would have been less vulnerable, resulting in markedly decreased erosion rates. Land use intensity increased throughout IA, peaking during LIA, before declining again in Hellenistic times. At the same time, increasing site sizes throughout the IA indicate higher population nucleation compared to BA. This might indicate that less vulnerable soils were also less fertile and more land needed to be exploited to match increasing subsistence demands. It has also been suggested that environmentally destructive farming practices shifted to elsewhere in the larger region, possibly to areas upwards in the Sakarya drainage, which would explain the decrease in land degradation in the immediate surroundings of Gordion. The area of Gordion was generally not characterised by high agricultural potential. It has therefore been debated how the Phrygian kingdom could have been established without a firm agricultural foundation on this local level. Building large-scale irrigation systems to mitigate this situation would have required a large labour force. While we have seen that the elite in Gordion could indeed muster a large labour force for activities such as textile production if needed, there are little indications for widespread irrigation channels in the Gordian landscape. Climatic data seem to suggest a higher level of moisture within the area, which could indicate higher potential agricultural yields during EIA (Miller, 2011). Other potential explanatory factors are the location of Gordion at the juncture of major trade

166  Social complexity trajectories in Anatolia and transportation routes in central Anatolia (Voigt, 2011), not unlike the favourable location of Hattuša as the capital of the Hittite kingdom. The rise of the Phrygian kingdom can be situated within the wider framework of macro-level polities in Anatolia, starting already in LBA. However, the impact of this macro-level socio-political unit was not uniform across the area it controlled. For the Hittite kingdom, we saw that regionalised structures of power resulted in differential trajectories on the local and regional level. It has been noted that regions beyond the immediate core areas of these polities did not necessarily undergo changes in the same way as the centre (Bell, 2009). While many settlements did show signs of destruction at the end of LBA, a lot of these sites continued to be inhabited and communities seem to have recovered, whereas others were not affected at all in the first place. On the community level, we therefore see markedly divergent trajectories of development, not all of which entailed downscaling of social complexity. It is essential to differentiate between relevant scales of analysis to get a better understanding of these complexity trajectories. One of the most marked developments in EIA is an apparent decrease in cultural cohesion and material similarities between settlements across different regions. This suggests an increasingly localised and regionalised form of cultural identities and social practices expressed in material culture. Even though we have evidence of inter-regional exchange networks continuing to connect areas as widely apart as the Aegean, the Balkan, Anatolia and the Near East, these interactions no longer seem to converge in similarities in material culture (MacSweeney, 2011). Moreover, the impact of the Phrygian kingdom as a potential unifying force seems to have been relatively limited in west Anatolia as their interests seem to have been directed primarily eastward during the early first millennium BCE (Sams, 1993). Let us now focus on some communities in west Anatolia, particularly Beycesultan and Aphrodisias. In the previous part, I already discussed how both communities engaged in markedly different strategies of community formation following the collapse of the Hittite kingdom. Whereas the community at Beycesultan emphasised the articulation of individual identities and status differentiation, at Aphrodisias we saw an emphasis on collective community identities. Unfortunately, not much is known about the transition from LBA to EIA at Aphrodisias. The main feature of the site is a domestic complex, built with stone foundations and mudbrick superstructure. The large amount of storage vessels attested suggests a role as storage facility. The association of this central building with prestige objects such as serpentine axes and adzes and marine shells indicates some form of social differentiation and suggests a shift away from collective community identities. The presence of prestige goods also suggests a continued participation in – and perhaps even more extensive access to – long-distance trade networks (MacSweeney, 2011). Other features dated to this period are an area reserved for craft

Social complexity trajectories in Anatolia 167 production, most likely metal production, as indicated by the presence of a crucible, hearth and lumps of formless metal and slag. Additionally, other productive industries such as chipped stone production were attested as well. Cooking vessels made up a large part of the ceramic assemblage in this area, suggesting food production beyond the household scale, possibly in function of communal dining activities. In a subsequent level, an area covered with plaster on a pebble foundation was found. Even though no structures could be associated with this feature, the careful treatment of the plaster floor suggests that the area was roofed and that its walls perhaps lie outside of the excavated area. Associated ceramics consist almost exclusively of dining shapes, suggesting that this location functioned as an area for communal consumption. It is interesting to note that even though a wide range of pottery wares were attested throughout the settlement, material culture associated with the communal dining area seemed to consciously reflect localised choices in form and decoration (MacSweeney, 2011). At Beycesultan, a destruction even occurred in the mid-12th century, but habitation seemed to have resumed almost immediately, with a new settlement level following largely the same lay-out as before. All four houses preserved from this level, however, were relatively (and equally) small. There are also no public buildings attested that might suggest any form of central administration or stratified social organisation. Yet, the community continued to participate in long-distance exchange, as indicated by imports of luxury items such as a faience figurine, an ivory stamp seal, marine shells, metal vessels and a serpentine axe. A notably higher amount of jewellery and bodily adornments suggests an increased interest in personal differentiation and emphasis on individual status as part of a more active creation and contestation of status, possibly induced by a more unstable collective social structure (MacSweeney, 2011). Clearly, this is not an impoverished or isolated community as one might expect from traditional narratives of ‘Dark Age’ collapse and decline. Instead, we see an active community where individual status as an important arena of social competition trumped collectiveness. The subsequent settlement level, dated to the 11th century BCE, consists of only a handful of buildings, including a large megaron on the east mound and the central megaron flanked by a number of structures surrounding its courtyard. No other (residential) structures could be linked with his phase. To what extent this is caused by the extensive later building and quarrying activities that occurred at the site remains unclear. However, it has been suggested that this phase constituted a strong reduction in population size due to (part of) the community moving away (Lloyd, 1972). However, the interpretation of the original excavators of the site as a farmstead housing a single extended family seems unlikely. The presence of a huge circular baking oven (3 m diameter) indicates a scale of (food) production well beyond the level of the household (MacSweeney, 2011). The monumental nature of

168  Social complexity trajectories in Anatolia the central megaron building, moreover, suggests a continued form of public display. Access to the building was designed for inclusivity and collective participation, taking the form of an open pillared porch rather than a doorway. Ceramics found in the building consisted of only a narrow range of shapes. The low proportion of cooking and storage vessels is particularly notable. This could suggest that the building had a specialised use, perhaps linked to feasting and ritual activities such as libations. Perhaps, this phase can be linked to a return to a more centralised social structure preoccupied with collective participation in social practices, albeit possibly on a smaller scale than before. The transition from LBA to EIA did not seem to have had a marked impact on either Beycesultan or Aphrodisias. At the former, strategies of community formation geared towards internal stratification, differentiation and local identities – initiated already after the fall of the Hittite kingdom – were continued in EIA. However, we do see a shift towards more active contestation of status and personalised expressions of identity. This highly internally competitive structure was not maintained for long and some form of centralised social organisation performing integrative collective practices seem to have returned in the final EIA layer. At Aphrodisias, the emphasis on collective community identities seem to have continued throughout EIA. The suggestion of a short-lived phase of more increased importance of personal status and internal stratification, possibly in response to the increasing influence of Mira, seems unconvincing and is at any rate built on a limited amount of material evidence. More convincing is the assessment of Aphrodisias as consciously stressing localised identities through its usage of material culture in integrative communal events. It seems clear that despite the upheaval on a macro level, characterised by the rise and fall of polities, communities such as Beycesultan and Aphrodisias continued to operate. If anything, they even seem to have responded to the opening up of new economic niches and political opportunities and quickly moved in to exploit these, as evidenced by the increased range of external exchange contacts and richness of the later layers at these sites. To allow for a more comprehensive diachronic account, I will now first discuss the macro level dynamics for the Middle Iron Age (MIA) and Late Iron Age (LIA), respectively dominated by the Lydian kingdom and Achaemenid empire, before moving on to regional and local patterns in the Konya Plain and the Lake District. From MIA (800–540 BCE) onwards, the power of the Phrygian kingdom started to decline, possibly (partially) as a result of Cimmerian incursions. At the same time, the Lydian kingdom, centred on its capital Sardis, established control over much of west Anatolia. While the earliest habitation at Sardis dates back to LBA, it only developed into a sizeable settlement during the 8th century BCE. We do not know much about its earliest phases, but a series of architectural units were excavated that have been identified

Social complexity trajectories in Anatolia 169 as domestic workshop complexes (Roosevelt, 2009). The earliest evidence for settlement on the acropolis of Sardis dates to the early 7th century BCE. At this time, a monumental fortification wall enclosing an area of 108 ha was constructed. By the early 6th century BCE, Gordion, along with most of west Anatolia, was likely under Lydian control. At this time, the Lydian kingdom established itself as a macro-level polity, engaging diplomatic relationships with the Assyrian kings. The 6th century seems to have been a period of wealth and prosperity in western Anatolia in general but particularly so at Sardis. At this time, the site was densely occupied and the urban area was fully developed. Population estimates for the site range from 10,000–20,000, to even 50,000 people (Ratté, 2008). Houses consisted mainly of stone foundations, mudbrick superstructures and covered by either thatched roofs or architectural terracottas and tiles. Sardis was ruled by a royal family and a local elite providing social, political and religious officials. Specialised workshops existed for the production of certain foodstuffs, stone working, precious metals and luxury objects to supply these local elites as well as for distribution in long-distance exchange networks. Texts mention a host of public buildings in Sardis, including the palace complex, administrative buildings, temples and a later Achaemenid paradeisos (enwalled park) as the most visible landmarks (Roosevelt, 2009). However, most of these have not been attested in the archaeological record. At this time, major building works occurred on the acropolis, where terrace walls were constructed to create level areas. The monumentalisation of Sardis from the late 7th to mid-6th centuries BCE coincides with the most extensive phase of Lydian control over west Anatolia. The redirection of resources and capital from wider areas towards Sardis in the form of tribute will likely have sustained these building projects. The main settlement areas in Lydia were distributed along the edges of major river valleys, likely motivated by the need for subsistence, potential for communication and territorial control (Roosevelt, 2009). These locations offered easy access to valley floors, they were close to prominent routes of communication and provided access to fertile arable land and upland pastorage. They were therefore prime locations for exploiting and controlling flows of energy, resources and information. Sardis itself is an excellent example of a site with high locational benefits, situated in the fertile area of the Hermus River valley which also acted as a major corridor for trade and communication across Anatolia. The city was surrounded by large estates geared towards large-scale agricultural production and craft production to supply the capital. LIA Anatolia (540–330 BCE) is characterised by the advent of Persian rule. In the mid-540s, Sardis was conquered by the Persian king Cyrus II. Cyrus integrated Sardis in the structures of power of the Persian empire as the seat of the satrapy of Lydia. These Persians – or Achaemenids after their ruling dynasty – started out as one of many groups in what is now

170  Social complexity trajectories in Anatolia western Iran. They are first mentioned in the written record of the Assyrian kings in the second half of the 9th century BCE, but references are few and far between and we know very little of them at this time (Waters, 2014). For a long time, they remained at the fringes of the major powers of their time. The reign of Cyrus II – titled ‘the Great’ – (559–530 BCE) was an important catalyst in the history of the Persians. Under his rule, the Persians conquered their neighbours, the Medes, giving them an important power base in the Mesopotamian area, before turning their attention to Anatolia. At its peak in the early 5th century BCE, the Persian empire stretched from the Indus River Valley to the east until Thrace to the west. The emergence of the Persian empire evidently constitutes a major leap in power and scope of polities in antiquity, especially compared to its Anatolian predecessors. Through extensive (regionally driven) tributes, it was able to redirect flows of energy and resources towards the central administration on a scale that was altogether unimaginable before. An essential prerequisite for this innovative development was the construction of an extensive and efficient network of roads for the movement of goods, messengers and armies across the empire. However, within the current case study, the question is whether the integration in the Persian empire had any profound implications for the complexity trajectories of Anatolia. It is at any rate interesting to note that the Seshat analysis did not include a data point for the Achaemenid empire in the Konya trajectory. In Figure 4.3, the Achaemenid period seems to have been compressed together with the Lydian period. This strategy would be warranted if the Persian conquest did not notably alter dynamics within Anatolia compared to the Lydian period. To find out, we will have to take a closer look at data on the local and regional level. After the Persian conquest, the royal family of Sardis was replaced with a satrap (provincial governor) and other officials from the Persian elite to manage the new satrapy of Lydia. These satraps were the most important officials in the Persian empire, acting as the direct extension of the king himself on the regional level. However, besides this direct replacement at the top, local officials and elites generally retained high positions in the new administrative structure (Roosevelt, 2009). The Persian empire was generally built on strongly regionalised structures of power, even though these were far more strongly integrated in the central apparatus compared to Anatolian predecessors such as the Hittite kingdom. Yet, the fact that local elites retained their importance in governance has led to the suggestion that the transition into Achaemenid rule in Anatolia was characterised by strong continuity in local governmental structures. The imperial policies of the Achaemenid dynasty are generally difficult to detect in the archaeological record of Anatolia, which has prompted the suggestion of a ‘light touch’ of governance (Hornblower, 1994; Rojas, 2016). The establishment of satrapal capitals at cities such as Sardis and Kelainai (capital of the satrapy of Great Phrygia) was a noted exception as these

Social complexity trajectories in Anatolia 171 cities were essential nodes in the governmental apparatus, for example to collect taxes. Elspeth Dusinberre (2013) proposed an ‘authority-autonomy model’ to describe the interactions between imperial administration, local communities and the implementation of imperial authority on this local level in Achaemenid Anatolia. She concludes that Achaemenid governance was characterised by a strong overarching authority in aspects that mattered to the central administration, such as military control, taxation and tribute, but that otherwise local regions and communities were allowed considerable autonomy in matters such as mortuary and religious practices, (elite) education and personal identities. Socio-cultural cohesion was stimulated through (1) shared practices such as conspicuous consumption during banquets; (2) iconography, such as the use of Achaemenid symbols of power in personal items and reliefs; and (3) the development of extensive networks of roads and communication routes favouring inter-regional interactions. Let us now shift this discussion to local and regional levels and see how they stacked up against the macro polities at the time. In the Konya and Karaman Plains, 19 settlements were dated to IA, of which 13 had material dated to EIA, 15 to MIA, and 17 to LIA. Additionally, 17 fortified hilltop sites were identified, 3 dated to EIA, 7 to MIA, and 17 to LIA. These sites were located on isolated topographic features but generally retained easy access to the valleys. The fact that most of these sites were located along major passes into and out of the lowlands on hilltops with limited access to water, makes them unlikely locations for normal residential settlement, suggesting the continuation of a hierarchical settlement pattern centred on the major sites in the plains (Massa et al., 2020). At some of these sites, a dense ceramic scatter was identified outside of the fortifications, indicating the existence of a community beyond the walled area. Following the collapse of the Hittite kingdom, several major LBA sites shrunk in size or were abandoned during EIA. In turn, sites such as Alaattin Tepe and Zoldura rose to prominence. It is also interesting to note that the system of hilltop sites notably contracted in EIA, before expanding again in MIA and LIA. The site at Türkmen-Karahöyük – which was suggested to have been the capital of the kingdom of Tarhuntašša in LBA – continued to be inhabited, and likely maintained its immense LBA size (100–125 ha). A Luwian inscription discovered at the site features the title ‘Great King’ and other references to the royal dynasty from the Hittite period, suggesting a – real or imagined – claim of continuity between the rulers of LBA kingdom of Tarhuntašša and another polity centred on at TürkmenKarahöyük in MIA times (Goedegebuure et al., 2020; Massa et al., 2020). The inscription refers to a conflict with 13 kings, suggesting the existence of multiple rulers in a politically fragmented landscape. The territorial extent of these polities remains unclear for now but it has been suggested that three main spheres of influence can be identified in the region around the Konya Plain. The plain itself was associated with a polity centred on

172  Social complexity trajectories in Anatolia Türkmen-Karahöyük, whereas another was likely located at Tuwanuwa and one at Kayseri. In LIA, settlement mounds started to be abandoned in favour of a more dispersed settlement patterns focused on new locations on the surrounding plains. It has been suggested that this shift in settlement patterns might be the result of lesser needs for security and defensibility of individual sites, reflecting the secure incorporation of this area within larger territorial entities (Massa et al., 2020, 2019). At the same time, this seems somewhat contradicted by the observed expansion of the network of fortified hilltop sites. Perhaps it was perceived as necessary to extend the hilltop forts network to counterbalance reduced security of site locations in the plain. Some have suggested that in addition to a military function, these hilltop sites also obtained other functions as part of the Achaemenid structures of government such as exerting administrative control (as part of the exaction of taxation) and regulating trade. At this time, a number of new sites emerged in the northern part of the Konya Plain, an area that had previously remained largely uninhabited. Possibly, this is the result of the establishment of a coordinated program of irrigation in this area. Palaeobotanical analysis of finds from Kınık Höyük indicate an intensification of agricultural production during the 6th century BCE. It seems that, for the Konya Plain at least, the transition into Achaemenid rule was accompanied with a series of changes in economic and political structures of administration. In the Lake District, human activity is attested again from the 9th century BCE onwards in the area of Lake Burdur, after more than a millennium of hardly any archaeological data. EIA settlement pattern was centred on substantial fortified hilltop settlements (Figure 4.10). Outside of these hilltop sites, few other types of settlement are known. The overall view is one of a nucleated settlement pattern focused on a limited number of sizeable communities. An exception to this general picture of hilltop sites in the area is the settlement at Düver Yarımada, a site that has been identified as potential religious complex for the Phyrgian Mother Goddess (Kahya, 2015; Talloen et al., 2006). The site possibly acted as a central place for a series of smaller agricultural villages and hamlets in the Burdur Plain (Kaptijn et al., 2012). Its location at the edge of a large, fertile plain allowed the community of Düver Yarımada to sustain a significant settlement. Through this location, it also had prime access to a series of important avenues of communication, both in east-west direction between Lycia and the Konya Plain, and northsouth between the Anatolian highlands and the Pamphylian coast (Poblome et al., 2013b). As a potential node in these macro-level avenues of connectivity, the community might have been able to use the available agricultural potential to tap into wider inter-regional networks of exchange connecting the Mediterranean with the Anatolian inland (Daems, Accepted(a)). The considerable presence of painted geometric pottery (EIA) and black-on-red

Social complexity trajectories in Anatolia 173

Figure 4.10  MIA and LIA sites in the area of Lake Burdur, with indication of the study area of the Sagalassos project (made by the author)

pottery (MIA) typical for wider (southwestern) Anatolia indeed seems to indicate extensive participation in inter-regional networks of exchange.15 Whether or not the centrality of Düver Yarımada stretched until the area of the hilltop sites towards the east or was limited to the Burdur Plain itself remains unclear for now, even though it has been suggested that they could be considered part of an integrated settlement system (Poblome et al., 2013b). For now, more research is needed to test this hypothesis. During MIA and LIA, new communities started to emerge, sometimes competing over existing niches such as in the case of Kayiş Kale and Çatal Pınar, but mostly occupying additional niches in the landscape, such as Düzen Tepe and Sagalassos in the central parts of the Ağlasun river valley. Both sites were located on elevated positions overlooking the valley, fitting with the existing tradition of elevated sites located on hilltops, slopes or raised plateaus. Up to this point, settlements seem to have only taken off – be it out of concerns for security or other reasons – on such elevated positions in the landscape. From the late 5th century BCE onwards, we see a partial shift in settlement patterns towards areas in lower locations. At the same time, several smaller hamlets and farmsteads emerged across the Ağlasun river valley, indicating also a shift from a strongly nucleated settlement pattern towards a more dispersed population. There are no indications that this area in the Lake District was directly under control of an overarching polity at any point during the IA. Yet, while Achaemenid control is not clearly attested in the local area, its structures of government and control would likely have extended to include this area. Kelainai, the Achaemenid capital of Greater Phrygia and local administrative seat of the Achaemenid government, was located less than 50 km to the north. This would likely have had some influence on local dynamics. To what extent the observed trend towards settlement dispersal observed at

174  Social complexity trajectories in Anatolia this time can effectively be linked to the emergence of Achaemenid power remains unclear but the parallels with observations from the Konya Plain are clear to note. Of all the IA sites, Düzen Tepe has been studied most extensively. Through a combination of remote sensing, geophysical, topographical, and archaeological surveys, scattered structures were identified over an area of almost 75 ha, with a clear settlement nucleus of about 13 ha (Vanhaverbeke et al., 2010). Excavations determined that these structures generally consisted of domestic structures with stone foundations and a superstructure made from perishable materials, most likely mudbrick. Based on the material culture and 14C dating, it has been suggested that the settlement was inhabited from the late 5th to early 2nd centuries BCE, with a core occupation period in the 4th and 3rd centuries BCE (Daems et al., 2017; Vanhaverbeke et al., 2010). Recent calculations have yielded an estimated population size of about 1000 people,16 consisting mainly of farmers working in a smallholders-based subsistence system (Cleymans et al., In Preparation). Resource exploitation, artisanal production and distribution of material culture all occurred within a locally oriented economic system and subsistence-based production geared towards supplying the own community (Daems et al., 2017; Daems and Poblome, 2016). Clays used for pottery production and as building materials were targeted in the immediate vicinity of the settlement itself. Almost no pottery produced at Düzen Tepe was found beyond the settlement proper. Conversely, of the more than 30,000 sherds studied from the settlement, less than 1% was imported. Throughout the settlement, there are few overt signs of social stratification and social life was likely oriented towards the household or other family-based social units. The disordered layout of the settlement shows no clear indications for a centralised or public locus onto which communal life could have been oriented. Perhaps the community was governed through a vertical egalitarian system as noted for village communities in earlier times. Interestingly, the site is located less than 2 km away from another settlement, at the location of what in Hellenistic and Roman times would become the city of Sagalassos. Due to later building activities, we have only little information regarding the oldest phases of habitation there. Pottery studies have suggested, however, that the site originated roughly contemporaneously to Düzen Tepe in the 5th century BCE (Daems and Poblome, 2017). I have argued elsewhere that the highly comparable material culture at Düzen Tepe and Sagalassos could indicate that these were very similar settlements (Daems et al., 2017; Daems and Poblome, 2017). From this, it can be hypothesised that both communities would likely have had comparable energetic needs and required a commensurate territory to sustain themselves. For calculations of the energetic needs of the community, we used population estimates, diet reconstruction, yields, and agricultural strategies, to estimate that an area of 526 ± 280 ha (1σ) was needed to sustain the community of Düzen Tepe (Cleymans et al., In Preparation). The nearby

Social complexity trajectories in Anatolia 175 Ağlasun and Yeşilbaşköy valleys are known as fertile areas suitable to grow crops. GIS analyses of the area indicated that up to 1000 ha could have been available for both Düzen Tepe and Sagalassos to sustain their energetic needs. However, it should be noted that this calculation only takes into account the endosomatic energy needs required for biological survival and reproduction. Exosomatic energy needs such as resources for building houses and fuel for cooking and artisanal production have not (yet) been included and are the subject of ongoing research. Given that Düzen Tepe and Sagalassos managed to coexist for more than two centuries, we can assume that the carrying capacity threshold of the local landscape was not crossed, but more detailed analysis is needed to confirm this assumption. The general impression of Düzen Tepe – and by extension Sagalassos – is one of a self-sustaining community, using predominantly locally available raw materials and exploiting the immediate hinterland of the site to sustain its energetic needs and resource requirements. This need not necessarily mean that the community was isolated from the outside world. Local pottery shapes were embedded in wider trends of production preferences and styles, particularly geared towards southern and central Anatolia, as well as the Levant (Daems et al., 2017). Yet, some have interpreted this type of locally oriented village communities as ‘rudimentary’ and ‘old-fashioned’, especially compared with the later developments in Hellenistic times. (Vyncke and Waelkens, 2015).17 Earlier, I already indicated that such normative interpretations are not acceptable for a proper assessment of community formation and complexity trajectories. I will discuss this view in more detail in the part on the Hellenistic period. The relative longevity of Düzen Tepe – continuously inhabited for almost three centuries – should serve as a reminder that these people were not rudimentary or backward but were actually very successful in sustaining their way of life, managing the potential and limitations of the landscape surrounding them to create a suitable environmental niche. This way of life effectively constituted a local basin of attraction adapted to local circumstances. Elsewhere, I have suggested that communities such as Düzen Tepe operated within a local pathway of development, centred on basic selection pressures such as subsistence, habitation, defence, production, and exchange (Daems, 2019). These were predominantly expressed within household-based contexts of social life supplemented with a limited degree of (functional) inter-household or community-level organisation and collective action measures. So, how does this view of Düzen Tepe relate to other sites in the area, most notably the hilltop sites discussed earlier? Even though the site at Düzen Tepe emerged later (5th century BCE) compared to the earliest evidence from the hilltop sites (9th century BCE), settlement patterns from EIA to LIA were characterised by strong continuity, suggesting that they were driven by similar selection pressures. Extrapolation of the information

176  Social complexity trajectories in Anatolia from Düzen Tepe would then suggest that most of these hilltop sites were similar, self-sustaining communities geared towards the potential energy and resources from their immediate environment. This suggestion has been corroborated by geochemical and petrographic studies on the pottery material collected at these sites, which suggest the existence of locally oriented strategies of resource exploitation (Braekmans et al., 2017). The geochemical picture shows a clear ‘compartmentalization’ of the landscape as these hilltop sites exploited nearby clay sources and operated within their own environmental niche. At the same time, these communities were not isolated. Each of the identified petrographic groups encompassed several ware groups, occurring on multiple settlements, rather than being exclusively associated with specific sites. This suggests clear inter-valley connections in production and/or distribution systems. Even though some variability in site size can be noted, the largest of these hilltop sites – Kayiş Kale with a maximum size of 15 ha – is about the same size as Düzen Tepe, suggesting a comparable maximum catchment area. This does not seem to have been insurmountable considering that each of these sites acted as the primary locus of habitation for its surrounding valley and would have had access to sufficient arable land to sustain the community (Daems, In Press). The fact that these communities continued to exist for several centuries indicates that they were fairly successful in maintaining and reproducing themselves. Does this then mean that all of these sites were living in sustainable ways, in perfect harmony with their environment? That picture appears to be somewhat overly rosy. Palaeoenvironmental data suggests that human impact at this time was considerable. Non-metric multidimensional scaling analysis of palynological data18 indicates that human impact markedly increased during MIA (Broothaerts et al., 2019). This coincides with the onset of the Beyşehir Occupation phase (BOP) in the area. As noted earlier, the BOP is a highly localised event of environmental change and its onset can be traced in various parts of the area, between the 10th and 3rd centuries BCE (Bakker et al., 2012). These new conditions favoured agricultural and arboricultural production at higher altitudes, characterised by the increased appearance of cultivated tree species and other anthropogenic indicators (Bottema and Woldring, 1984). A model of soil erosion and sedimentation patterns developed for the Gravgaz valley, suggests that a decrease in pine forest cover induced a strong erosion phase around 700 BCE, followed by extensive soil accumulation in the valley bottom (Van Loo et al., 2017). It is likely that the energetic needs generated by a sustained population concentration in nucleated hilltop sites, focusing on available resources in the vicinity of the site, induced major forest clearance on the slopes. This led to a marked decrease of soil depths on the slopes, but at the same time increased soil depths and fertility in the lower valleys. As a result, the potential for habitation in the lower areas markedly increased, reflected in the emergence of hamlets and farmsteads in these areas. It seems clear that local communities took full

Social complexity trajectories in Anatolia 177 advantage of favourable conditions generated by warmer and more humid circumstances. In doing so, the impact generated by these communities markedly reshaped their environment and created the conditions and potential for new niches in the landscape to be exploited, resulting in a more diversified settlement pattern. At the same time, these communities did not completely deplete the potential of their catchment, given that many hilltop sites continued to be inhabited until Hellenistic times. The reorganisation of local system configurations should not be considered a collapse of local environmental niches, but rather a transition into a new equilibrium with a shifting focus towards the lower valley areas. This shift towards a new (adaptive) cycle laid the foundations for a series of developments in the Lake District during the Hellenistic period, of which the rise of Sagalassos to a position of regional primacy was the most marked one. I will discuss this development in more detail in the next part. Let us conclude this part with an overview of the discussion on IA Anatolia and link these observations to the interpretative framework of the adaptive cycle and social complexity trajectories. It must be stressed once more that we must carefully differentiate between different scales of analysis to discuss relevant system dynamics. On the macro level of territorial polities we see a continuation of dynamics and selection pressures already initiated in LBA with the Hittite kingdom. Supra-regional entities such as the Phrygian and Lydian kingdoms redirected energy and resources towards the capital through regional structures of administration built on taxation and tribute. These flows fuelled investment in the capital cities of Gordion and Sardis resulting in intensive urbanism and monumentalisation. As in LBA, smaller polities such as the kingdoms in the area of the Konya Plain continued to develop in the periphery of the central Anatolian polities. Even though a significant release (Ω) and reorganisation (α) event occurred towards the end of LBA, we see that in the subsequent cycle in EIA many of the existing system components are reused. Unfortunately, our knowledge of the transition from LBA to EIA is not fine-grained enough to allow a more detailed analysis of how these components were reorganised into the new cycle. On the micro-level of community formation, we can assess this transition with some more resolution, particularly at Beycesultan and, to a lesser extent, Aphrodisias. We see how these communities renegotiated their position within the wider world and reorganised internal organisational structures in response to the decline of polities such as the Hittite kingdom. The stories of these two settlements tell us that the rationale of widespread collapse of territorial polities at the end of LBA need not necessarily have affected individual communities negatively. To what extent elements from the preceding system configuration were being reused in the Lake District, or whether these can be considered in terms of a wholly new adaptive cycle, is unclear given the lack of evidence for human occupation in the preceding millennium. The observed trend of

178  Social complexity trajectories in Anatolia a gradual occupation of different parts of the landscape during EIA and MIA fits well with an ongoing exploitation (r) phase characterised by niche construction and resource diversification. Once the landscape started to fill up with strongly locally oriented hilltop sites, a transition towards a conservation (K) phase was initiated. The observed interconnectedness – both on a micro-regional and supra-regional level – can perhaps be considered indicative of such a transition as well. The increased human impact generated by the exploitation of additional environmental niches and associated resources put additional pressure onto the local socio-ecological system. A fairly wide range of selection pressures can be observed as driving forces for complexity trajectories in EIA (Table 4.4). It is, however, interesting to see that diversification can be considered the main mechanism of development in the complexity trajectories of EIA. It should be noted, however, that this mainly pertains to micro-level community formation and that the counterpart mechanism for macro-level polity formation remains unclear in light of the limitations of the current data. During MIA, a more pronounced focus on subsistence as main selection pressure can be observed, even though this is somewhat biased by the level of detail in information for human-environment interactions from the Lake District. Following the Achaemenid conquest at the start of LIA, Anatolia as a whole was for the first time integrated in a single socio-political unity that could be rightfully called an empire. Yet, this was not associated with any fundamental shifts on the local and regional level. Moreover, when changes occurred, they were not uniform across different regions. The observed dynamics seem to confirm to the authority-autonomy model of Achaemenid rule in Anatolia which suggests that Achaemenid intervention was based on regional and local circumstances and mainly limited to areas of direct interest to the central administration such as structures of administrative control and taxation. In the Konya Plain, shifts towards a more dispersed settlement pattern might be related to a decreased need for defensibility of individual sites, whereas the extension of the settlement network in the northern parts of the Konya plains and in the highlands surrounding the plain might be related to, respectively, an intensification of agricultural production and the imposition of a more structured network of taxation and tribute extraction. It is likely that intensification of agricultural production was geared towards meeting the systematic requirements of taxation. This shift towards increasing intensification of exploitation and production is typical for a sustained conservation (K) phase. In the Lake District, we likewise see a shift towards settlement dispersal in the form of hamlets and farmsteads popping up in the lower valley slopes in addition to nucleated hilltop settlements. It has been noted that the observed forest clearance and the subsequent soil erosion on the higher slopes as a result of the intense exploitation by hilltop sites, not only led to a decrease of soil depths on the slopes, but also in sediment accumulation in the lower valley areas. This resulted in the creation of large fertile areas

Table 4.4  Social complexity trajectories in IA. *Energy (E), resources (R), information (I); **diversification (D), intensification (IS), integration (IG) Flows*

Drivers

Process

Outcome

E

R

I

EIA EIA EIA EIA

Subsistence Subsistence Production Interaction

Population growth Community reorganisation Community reorganisation Community reorganisation

x x x

x x x

x

EIA EIA

Interaction Governance

Energised crowding Centralisation

x

x x

x x

EIA EIA EIA EIA

Production Governance Distribution Subsistence

x x x

x

EIA MIA MIA

Competition Subsistence Competition

Labour mobilisation Central place formation Central place formation Shifting agricultural practices External stresses Population growth Peer-polity interaction

Niche diversification Subsistence economy Household production Localised community identities Social stratification Supra-regional polity formation Specialised production Monumentalisation Inter-regional exchange Land degradation

MIA MIA MIA LIA

Subsistence Competition Subsistence Subsistence

LIA LIA

Subsistence Governance

Agricultural intensification Polity competition Soil erosion Diversified settlement patterns Agricultural intensification Structures of control

Polity collapse Niche diversification Regional polity formation Soil erosion Polity collapse Niche formation Niche diversification Niche diversification Dispersed settlement patterns

x x x x

x x

x x x

x x x x

x

x

x

x x x x

x

Pull Push x x

x x x x

x

x

Mechanism** D

x

x x x x

x

x

IS

IG

AC r r r r

x x x

x

K K K K K Ω

x

x

x x

x x

x x x x

x x x x

K Ω α r

x x

x x

K K

x

Ω r K

Social complexity trajectories in Anatolia 179

Period

Selection pressure

180  Social complexity trajectories in Anatolia suitable for crop cultivation. The strains imposed on the landscape by the hilltop sites exceeded local resilience, inducing a release (Ω) of the available potential (soil depth) and reorganization (α) into a new stable state (more fertile circumstances in lower areas). Whether the underlying selection pressures driving the local system beyond its threshold into a new stable state can be linked to Achaemenid rule, perhaps as an attempt to intensify agricultural production geared towards taxation similarly to the Konya Plain, or were the result of local drivers, for example as a result of population growth, remains unclear for now. The fact that the process of niche diversification was already observed during MIA seems to suggest that local drivers at least initiated the process, which was then perhaps intensified during LIA. It is clear by now that the integration of Anatolia into the Achaemenid empire did not constitute a major leap in complexity trajectories, but rather entailed a continuation and intensification of ongoing dynamics. The most notable aspect of this process is that the drivers of complexity formation in MIA and LIA Anatolia mainly acted as pushing forces, inducing an extensive diversification and dispersal of settlements across the landscape. This means that on a local and regional level, communities focused on strategies of diversification to meet the stimuli offered by the Achaemenid empire, rather than initiating pathways of growing settlement sizes or increasing complexity to enhance the exploitation of energy and resources from a larger area. Perhaps this can be related to restrictions imposed by the scale of energy and resources that could be exploited by these communities. Naturally, such strategies could only have continued as long as sufficient space for sustained diversification was available. Through the pressures exerted on the landscape at this time, some seeds were being planted that would grow into major drivers of complexity formation on a local and regional level during the Hellenistic period. Hellenistic period The final period under consideration in this case study is the Hellenistic period (334–325 BCE). In the complexity trajectory of Figure 4.3, this period constitutes a slight decrease in complexity value in comparison with the preceding LIA period, but overall remains on the same level. Let us again have a look how the data stacks up against this trajectory. The beginning of the Hellenistic period is of course heralded by Alexander III the Great (356–323 BCE) and his conquest of the Achaemenid empire. After his death in 323 BCE, his kingdom was divided amongst his generals. Large parts of Anatolia were ruled, respectively, by Antigonos Monophthalmos (333–301 BCE), Lysimachos (301–281 BCE), the Seleucids (281–189 BCE), and the Attalids (180–129 BCE). Throughout this period, other dynasties regularly disputed control over Anatolia, most notably the Ptolemeans, who at several points in time, controlled significant amounts

Social complexity trajectories in Anatolia 181 of the Anatolian coastal areas. Anatolia was never far from the minds and efforts of the various pretenders to the throne and often formed the stage for their competition. In the interest of space, I decided to focus particularly on the period of Seleucid rule in southwest Anatolia the period from the early 3rd to mid-2nd centuries BCE. This means that I will not discuss in detail the messy struggles and turmoil between the diadochi (the successors) in the aftermath of Alexander’s death. I will also not cover the late Hellenistic period, which was defined by the rise of Rome in Anatolia. Large parts of Anatolia were bequeathed to the Romans by Attalid III after his death in 133 BCE and the rise of Rome, while fascinating, constitutes a wholly different level of complexity trajectories. Including this account here would not be possible within the framework of the current case study. The Hellenistic period is commonly characterised as an age of change, indicated by an increased intensity and scale of human activity, including the development of larger cities, trade expansion, size of buildings, mobility of people, higher war frequency, as well as increasing differentiation in economic, political, and social spheres of human activity (Manning, 2018). This makes it an extremely relevant period for the study of social complexity trajectories. Before we can address this, however, we must first consider an important socio-cultural factor of this period, the perceived spread of Greek culture across the eastern Mediterranean and eastwards all the way up to the Indus valley. The spread of Greek cultural influences in Anatolia was not unique to the Hellenistic period but is generally said to have started in the western and southern coastal regions already from the 11th century BCE onwards. This so-called ‘Ionian colonisation’ is a concept that has largely been constructed on the basis of later Greek authors such as Herodotus, Thucydides, Pausanias, and Strabo. It has increasingly been realised that the concept of colonisation is essentially a modern construct based on the retrojection of imperial activity of European powers in the modern period onto the ancient world (Hurst and Owen, 2005). Recently, scholars have started to redefine the terms of the debate, shifting from a focus on direct cultural influences of Greek traditions towards locally driven renegotiation of socio-cultural and political traditions (Boehm, 2018; Chrubasik and King, 2017; Daems, Accepted(b)). The supposed diffusion of Greek culture was most eminently expressed in the spread of the Greek-style polis or city-state. The polis is inherently connected to Greek history and commonly considered the primordial unit of community organisation in the Aegean heartland of Greek culture in Archaic (MIA) and Classical (LIA) times. At the same time, the polis as an urban, cultural and political phenomenon has been used as a classificatory unit far beyond these temporal and spatial boundaries. The Copenhagen Polis Centre used a polythetic19 list of traits20 to identify 1035 communities in an inventory of MIA and LIA poleis throughout the Mediterranean (Figure 4.11) (Hansen and Nielsen, 2004). We can wonder to what extent

182  Social complexity trajectories in Anatolia

Figure 4.11  MIA and LIA poleis identified by the Copenhagen polis Centre in the Aegean and Anatolia (made by the author based on data from Hansen and Nielsen 2004)

such a large amount of settlements can truly be covered by a single moniker, disregarding most variability in community organisation. It can moreover be argued that, individually, many of the selected traits can be associated with a wide range of socio-political configurations, not necessarily limited to the specific case of the polis. Clifford Ando has recently defined the polis in a general sense, omitting the notion of Greekness, as “the combination of a population and set of institutions, centred on an agglomerated settlement and asserting dominance over a wider population and landscape” (Ando, 2018). Ando here follows Mogens Herman Hansen’s notion of city-state cultures, of which the Greek polis was but one example (Hansen, 2000). This approach does not solve the tension between the generic nature of individual traits and the specificity of identifying Greek polis communities per se. The only way of differentiating Greek poleis from other city-state cultures and other comparable modes of community organization within specific temporal and spatial parameters would then be, not through the traits themselves, but through their context (Daems, Accepted(b)). In this sense, the question is not ‘what is the polis’ but rather ‘what is Greek’. This means that we have to differentiate between, on the one hand, the polis as a political and urban phenomenon, and, on the other hand, the polis as a socio-cultural phenomenon. Failing to recognise this distinction means that we effectively revert back to socio-evolutionary approaches and biases imposed by problematic associations with Hellenocentric and Eurocentric discourses. This is all very relevant for the study of cultural identities and community formation in Anatolia in particular, and social complexity trajectories in general. Studying Anatolia through the lens of Greek culture is not just an

Social complexity trajectories in Anatolia 183 innocent perspective shift in scholarly works. All too often, the introduction of Greek cultural elements outside of the Aegean is associated with ‘open’ and ‘progressive’ mentalities, in contrast with the ‘backward’ and ‘reactive’ clinging to local traditions (Daems, Accepted(b)). Particularly on the level of communities and settlements there is an unequivocal association between the Greek polis and the development of urbanism and a ‘cosmopolitan’ urban way of life. Such associations are highly problematic given the strong associations between urbanism and social complexity, and therefore run the risk of reintroducing implicit social evolutionary perspectives. These (implicit and explicit) biases are commonly integrated in a narrative of Greek versus ‘Other’ cultures, a discourse deeply rooted in colonial and Eurocentric thinking (Vlassopoulos, 2007). This way, Greek cultural elements are commonly described as more complex, whereas socio-cultural developments in local communities are frequently ascribed to the integration of new peoples, mainly as colonists or veterans from the armies of the Hellenistic kings, or defined as a form of Hellenocentric aemulatio, influencing local tastes, preferences and styles because of a presupposed inherent desirability of Greek culture (Daems, Accepted(b)). Clearly, such straightforward normative assessments of cultural desirability can no longer be supported. The approach to social complexity presented here offers one way to move beyond such normative approaches by focusing on social complexity as a societal property defined by flows of energy, resources and information. Before addressing the micro-level of community formation in more detail, however, I briefly want to introduce the workings of macro-level polities and their mutual interactions. It should be noted that the Hellenistic kingdoms, while substantial, did not operate on the same scale as the Achaemenid empire at large. As we saw in the previous part, however, the unparalleled scale of the Achaemenid empire did not really translate in increasing social complexity on the local and regional level in Anatolia. During the Hellenistic period, a markedly different dynamic emerged from the competitive interactions between the Hellenistic kingdoms on the one hand, and the royal administration and local communities on the other hand. I will argue that from the combination of these horizontal and vertical lines of interaction and competition, stimuli for social complexity formation were generated. Three core functions of the Hellenistic kingdoms have been distinguished: warfare, gift-giving and conspicuous consumption (Ma, 2013). Due to the near-constant conflicts between different kingdoms, maintaining the military apparatus was probably the highest expenditure for central administration (Chaniotis, 2005). Under Seleucid rule, large parts of south and central Anatolia formed a frontier zone against the Ptolemies who controlled parts of the coastal areas. One of the most marked consequences of this ongoing military activity was the establishment of veteran colonies as a means of assuming control over the local and regional landscape (Cohen, 1978). It is indeed interesting to note that many – but not all – of these foundations are located more in the inland areas that were part of the aforementioned

184  Social complexity trajectories in Anatolia

Figure 4.12  Hellenistic city foundations in Anatolia (made by the author)

frontier zone and where urban density was lower (Figure 4.12). Some examples of Seleucid city foundations can be found in the northern parts of Pisidia, including Laodikeia, Apameia (formerly Kelainai), Apollonia (formerly Mordiaion), Antiocheia, Laodikeia Katakekaumene, and Seleukeia Sidera. These city foundations were important foci of control over the fertile agricultural lands on the northeast of Lake Burdur. These city foundations have also been described as “avatars of Hellenism in Pisidia as in much of inland Anatolia” (Cohen, 1995, p. 15) and are frequently considered to have served as examples for cultural emulation by local communities. However, the impact of the incorporation of inland Anatolia in the Hellenistic kingdoms has been debated. Some have argued that many areas such as Pisidia remained a ‘backward’ area until the reign of Augustus, ‘hardly touched by Greek culture and democratic institutions’ (von Aulock, 1977). Others have argued for a rapid, endogenously driven Hellenization as expressed in the use of Greek as the official language, municipal institutions and in material culture (Kosmetatou, 1997; Mitchell, 1991; Waelkens, 2004; Waelkens and Vandeput, 2007). From the preceding discussion it should be clear that both views need to be discarded for being inherently Hellenocentric. A more fruitful approach would be to contextualise community formation and its expression in material, political and cultural terms as part of an active dialectic between a community and its environment, which included not only city foundations but also other (neighbouring) communities and the administrative apparatus of the Hellenistic kingdoms. The policies of the latter can be considered particularly potent drivers of development. The second core dynamic of Hellenistic kingdoms – gift giving as an attestation of royal magnificence – was the main modus operandi to regulate the interactions between the Hellenistic kings and local communities

Social complexity trajectories in Anatolia 185 (Ma, 1999). Gift giving was part of a complex two-way flow of favours, services and gifts. Rather than imposing policies in top-down fashion, the Hellenistic kings and their administration engaged in ‘negotiations’ with local communities to find acceptable forms of power display and obedience, notably by combining royal gift-giving and concessions in return for (symbolic) honours representing factual submission to imperial control (Ma, 2013). Royal gifts could include territorial extension, city status, tax exemptions, buildings, and the right to strike coinage (Bringmann, 1993; Bringmann and Steuben, 1995). The two-way relationship between king and community based on reciprocal gift giving was skewed due to the uneven power distribution. These diplomatic interactions have even been described as “a complex freedom-charade” (Green, 1990, p. 196). Yet, it was not mere meaningless rhetoric. The ever-presence of rival kingdoms in Anatolia meant that alternative offers, both ideological and material, could be on the table and gift giving was necessary as a ‘bribe’ to maintain or gain control over local communities. Constant inter-polity competition, either directly through armed conflict or indirectly by competing for the favour of local communities, was not only an important driver of socio-political development, but also a major incentive to generate surpluses through monetization of the economy (Aperghis, 2004). In this hypothesis – which I have elsewhere named the ‘royal policy model’ (Daems, Accepted(b)) – it is argued that the Seleucids in particular were in urgent need for silver to pay their armies of mercenaries and consolidate power over their rivals. Whereas during the preceding Achaemenid period, taxes were generally paid in kind, the Seleucids initiated a shift in mechanisms of administration and extraction of economic surplus towards taxation in the form of silver. This placed the onus of extracting and selling agricultural surplus production onto local communities. This shift required the creation of additional markets where surplus production could be sold, inducing a widespread development of urban centres throughout the Seleucid kingdom. The extensive program of city foundations can be considered part of imperial strategies of urbanism which also included stimulating synoikismos21 or community merging to increase the economic base of local communities (Boehm, 2018). Monetary surpluses redirected towards central administration were then used to fuel conspicuous consumption at the royal courts as focal points of royal power and prestige (Strootman, 2014), as well as to fund military exploits and other expenditures such as gift giving (closing the loop in capital flows from the local level to the central administration and back). The picture painted above provides an extensive framework for the development of local communities in the Hellenistic period. Let us now take a look at some examples to see how some of these communities acted upon this new framework of economic and political incentives offered to them. For this part, I will draw extensively from my own research on Sagalassos as a pars pro toto example for local communities during the Hellenistic period.

186  Social complexity trajectories in Anatolia The emergence of Sagalassos as an urban centre and local and regional hub was one of the most notable developments in the area of Lake Burdur during the Hellenistic period. Elsewhere, I have argued that the rise of Sagalassos can be interpreted in light of an active participation in the policies and incentives offered by the Seleucids (Daems, 2019; Daems and Poblome, 2016; Daems and Talloen, In Review). This line of thought will be further explored here. In the wider area, the nucleated settlement pattern observed for most of the IA started to change already during LIA, most notably through increased settlement diversification and higher settlement numbers. This trend continued into the Hellenistic period. During the EH, new settlements started to emerge in or at the edge of the valleys, most likely as a result of the increased agricultural potential of these areas following the aforementioned processes of erosion and sedimentation. Given that many existing settlements continued to be inhabited, it can be suggested that the increase in settlement numbers could at least partially be the result of population growth (Poblome et al., 2013b). The transformation of Sagalassos into an urban community started in the late 3rd century BCE. At this point, it developed some form of political constitution and codified law system (Vandorpe and Waelkens, 2007), initiated the minting of civic coinage (Van Heesch and Stroobants, 2015), and constructed monumental public buildings (Poblome and Daems, 2019). At the beginning of the 2nd century BCE, the production of material culture intensified, became more specialised and geared towards external distribution (Daems et al., 2019). At this time, pottery production facilities were established in a spatially demarcated production quarter (Poblome et al., 2013a). Finally, a dependent political territory was established, possibly stretching as far as the Burdur Plain after 189 BCE (Daems and Poblome, 2016; Waelkens, 2004). Taken together, these processes can be subsumed under the moniker of polis formation, in the sense of the emergence of a political community with a certain territorial extent centred on an urban settlement. Per the discussion above, this shift did not involve direct Greek cultural influences. The usage of Greek at Sagalassos is attested in a single inscription detailing some form of political and judicial system. However, that very same inscription also seems to suggest that Greek language had not permeated deeply in the community. Of the 24 names mentioned in the decree, not a single one is Greek. Material culture dated to the Hellenistic period, particularly pottery, also shows close affinity with Anatolian and Levantine material culture, rather than the Aegean or the Greek mainland. Coinage struck at Sagalassos in the late 3rd century BCE of the well-known type of posthumous Alexander tetradrachms shows Herakles and Zeus on either side of the coin, as well as the Greek legends ΑΛΕΧΑΝΔΡΟY (“of Alexander”) and ΣΑΓΑ (“Saga”). It should be noted, however, that this type of coinage was widespread in the eastern Mediterranean at this time. The usage of Greek language seems

Social complexity trajectories in Anatolia 187 more closely connected to political considerations than any direct Greek or Hellenistic cultural influences. Nevertheless, the urban transformation of Sagalassos has been linked to its development as an explicitly Greek polis with associated open and progressive mentalities, whereas nearby Düzen Tepe was abandoned in the 2nd century BCE, supposedly because of the backward and isolated nature of its ‘Pisidian’ community (Vanhaverbeke et al., 2010; Vyncke and Waelkens, 2015). The urban transformation of Sagalassos and its associated processes can be more clearly and objectively interpreted through the lens of metabolic regimes and energy requirements. Several of the processes highlighted above represented a clear-cut increase in energy expenditures, such as the construction of monumental public buildings, which required a significantly larger investment in raw material exploitation, specialised labour for construction and higher capital investment. Part of the increased energy requirements to sustain these developments could have been met by the claim over an extended territory, including the fertile Burdur Plain. A total area of up to 900 km 2 was potentially available at this point. We have to consider, however, to what extent this full area could have readily been used as source for the necessary energy and resources. First of all, the area consists of several valleys interspersed in mountain ranges. Not only did this significantly reduce the amount of land available for agricultural production, it also impeded strong integration of different parts of the territory. Moreover, the asymmetric location of Sagalassos compared to its surrounding territory could have impeded flows of energy and resources from the full extent of the territory to have reached Sagalassos on a continuous or regular basis (Figure 4.13) (Daems, 2019). It can be suggested that Sagalassos would have focused more readily on the eastern areas of the territory centred on the Ağlasun and Çanaklı valleys in its immediate hinterland. We do know that from the Early Hellenistic period onwards, Sagalassos started to target clay sources from the neighbouring Çanaklı valleys – located approximately 7–8 km south of Sagalassos – to supply its specialised pottery production industry, indicating a more extensive raw material economy as part of a regionally oriented economic system (Daems, In Press). It might not have been very feasible to rely on continuous supplies from the Burdur Plain for basic subsistence needs instead of the immediate hinterland. Perhaps it could therefore be suggested that, despite the additional land potentially available to them, the community at Sagalassos continued to rely heavily on their immediate environment for necessary energy and resources. The fact that Düzen Tepe and Sagalassos coexisted for almost three centuries, but that the former was abandoned right at the time that the latter initiated its urban transformation might indicate that the additional energetic requirements to sustain this development left insufficient room for a community at Düzen Tepe to viably coexist with Sagalassos. Elsewhere, I have suggested that (part of) the population of Düzen Tepe

188  Social complexity trajectories in Anatolia

Figure 4.13  The territory of Sagalassos

might have moved to Sagalassos as part of a synoikismos event, possibly stimulated by Seleucid policies (Daems and Poblome, 2016; Daems and Talloen, Accepted). The resultant population increase and nucleation at Sagalassos would have had a major impact on social and economic dynamics within the community. Moreover, the disappearance of hamlets and farmsteads from the survey record in the surrounding Ağlasun valley might indicate a concentration of population coming from the rural areas towards the newly formed urban hub. Following the tenets of settlement scaling theory, energised crowding induced by this population nucleation acted as a major driver of community formation dynamics and economic growth associated with the urban transformation of Sagalassos. It is no coincidence that at this point in time the potters of Sagalassos started to congregate in the southern part of the town. The spatial proximity of these workshops was part of the wider establishment of agglomeration economies through economies of scale and knowledge spill-overs (Daems, In Press). Resource specialisation and diversification allowed these workshops to initiate a specialised production of fine table wares. This specialised production generated surpluses that were increasingly geared towards supplying a wider market as well. The pottery material of Sagalassos started to be exchanged on a wider spatial level, covering not only the immediate catchment area, but also gradually including neighbouring valley systems (Poblome et al., 2013b). At the same time, the appearance of amphorae – albeit initially in limited quantities – coming from Italy, Rhodos, Kos, and Chios from

Social complexity trajectories in Anatolia 189 200 BCE onwards, indicates participation in supra-regional exchange networks in addition to these local and regional networks (Monsieur, Daems, and Poblome 2017). In short, Sagalassos transitioned into an urban community that played an important role as a hub in economic and political networks across multiple scales. What is left to discuss are the selection pressures that initiated the development at Sagalassos and link these to wider dynamics in south­west Anatolia. Elsewhere, I have suggested three possible scenarios: (1)  an internally driven process of population growth and community formation; (2) a locally induced merging or synoikismos between Düzen Tepe and Sagalassos; (3) an externally stimulated synoikismos (Daems and Poblome, 2016). All three scenarios can account for the end-result of Sagalassos thriving and Düzen Tepe being abandoned. In each scenario, population growth, agglomeration economies and community formation can be considered the essential drivers as well. The only difference would be the initial catalyst. In the first two cases, it is difficult to pinpoint a specific causal factor (or set of factors) to offset the process of urban transformation from the local evidence alone. The difference between these two scenarios is that in the first the decline of Düzen Tepe and potential population movement towards Sagalassos was a protracted one, whereas in the second case it would constitute a (nearly) single event. Whether the potential effect of the second scenario would have been stronger is difficult to say, but its impact on the community would surely have required strong measures of collective integration to be implemented. The current state of the evidence, however, does not allow us to identify any such measures. The third scenario is part of the framework of Seleucid policies and stimuli outlined earlier. In this framework, it is suggested that the Seleucids were actively looking for suitable local partners to implement their policies of monetisation and economic stimulation, driving processes of urbanism and market formation. While it is impossible to reconstruct the individual and collective decision-making processes underlying the initial formulation of this strategy, it can be hypothesised that for some reason Sagalassos decided to engage with the incentives offered by the Seleucids whereas Düzen Tepe did not. Why exactly this happened and not the other way around is impossible to reconstruct. It can be noted that the former held certain locational advantages over the latter, most importantly easier access to the lower valley slopes and the potential for agricultural exploitation of these lands, as well as access to extensive water sources in the mountain ranges behind the site (Daems, 2019). On a general note, however, it should be remembered that the exact reason for potential divergent developments need not necessarily be very notable or profound given the sensitivity of initial conditions in complex systems such as human societies. This means that even small initial differences can

190  Social complexity trajectories in Anatolia lead to widely divergent trajectories of development. To further elucidate potential drivers of local developments, let us expand the scope of analysis towards wider patterns in southwestern Anatolia. Before the Hellenistic period, urban density in Pisidia was low and Pisidian communities appeared to have been organised mainly in fortified hilltop settlements (Hürmüzlü et al., 2009). For most parts of Pisidia, it was long believed that a widespread wave of urbanism occurred during the 2nd century BCE, as seen in sites such as Selge, Termessos, Adada, and Ariassos (Mitchell, 1991). This process has been linked to a phase of extensive economic prosperity under Attalid rule. Following the evidence from site such as Sagalassos, however, clear indications of urban change can be noted already towards the end of the 3rd century BCE. Other evidence such as coins, inscriptions and historical texts also indicate that urban communities had already emerged at this time at sites such as Etenna, Kremna, and Termessos (Daems, Accepted(b)). By contrast, in the coastal areas of Pamphylia and Lycia the potential impact of Seleucid policies of urbanism was less pronounced due to an existent tradition of urban communities. In Pamphylia, this tradition preceded most other regions in southwestern Anatolia, stretching back to EIA times. From MIA times onwards, a stable settlement pattern had emerged, centred on seven main centres that remained important until the Hellenistic period: Side, Magydos, Olbia, Aspendos, Sillyon, Perge, and Phaselis (Grainger, 2009). In Lycia, a coastal region to the southwest of the Lake District, a tradition of urban communities went back to MIA times. Some of its most characteristic features were an elaborate funerary architecture, the distinct Lycian language, and a shared coin standard (Hansen, 2002). The most

Figure 4.14  Main Hellenistic sites in southwest Anatolia mentioned in the text (made by the author)

Social complexity trajectories in Anatolia 191 important sites were Xanthos, Tlos, Limyra, Patara, Telmessos, and Avsar Tepesi. Already in the 6th century BCE, some of these centres developed into Herrensitzen, or urban centres serving as the political seats of local dynasts (Hansen, 2000; Kolb, 2008). The fortifications of these large settlements typically enclosed an area between 10 and 25 ha and housed between 1000 and 1500 people. During the Achaemenid period, Lycia was said to have been governed through a hierarchical political structure centred on a central dynast – operating under the suzerainty of the Achaemenid kings – who in turn ruled over a number of lesser dynasts, each with a certain degree of autonomy. The main Lycian settlements at this time were typically highly urbanised and fortified, with varying degrees of political independence, but displaying strong indicators of cultural cohesiveness. This has prompted the suggestion of an “indigenous”22 Lycian city-state culture dated from the second half of the 6th century to the first half of the 4th century BCE, which is differentiated from the Greek poleis on the basis of the absence of Greek cultural characteristics (Hansen, 2002). Clearly, this Hellenocentric criterium of differentiation does not really offer much help in distinguishing fundamental changes between the Achaemenid and Hellenistic period. In contrast to these settlements in the coastal regions, inland Lycia and neighbouring Kabalia were not extensively urbanised before the end of the 3rd century BCE, when sites such as Balboura, Boubon, Kibyra and Oenoanda started to emerge (Coulton, 2012). By the early 2nd century BCE, the differences between the coastal and inland areas of Lycia had largely disappeared, and communities in both areas entered into networks of political and economic cooperation, culminating in the formation of the Lycian League. It is interesting to note that in areas with no long-standing tradition of urban communities, such as Pisidia, northern Lycia, and Kabalia, a ‘second wave’ of urbanism occurred at the end of the 3rd century BCE. It can be hypothesised that this can be linked to the effects of Seleucid strategies of urbanism. The fact that the observed development of local communities coincided with the establishment of a series of Seleucid city foundations could corroborate this hypothesis. This means that even if these city foundations were not ‘cultural avatars of Hellenism’, they did offer a template of political and urban structures of community formation that might have resonated with nearby communities. The fact that hardly any city foundations have been attested in Lycia, which had a long history of urbanism, seems to support this suggestion. To confirm the validity of the royal policy model, however, we need to establish the economic effects of these policies and the degree of monetisation of local economies. Here it should immediately be noted that the direct link posited by the model between in military expenses, active urbanism and monetisation is not undisputed given the observed chronological

192  Social complexity trajectories in Anatolia gap between city foundations and the start of the minting of local coinage towards the end of the 3rd century BCE (de Callataý, 2019). Given that these early mints focused mainly on large silver denominations that were not suited for daily transactions on the local market, it can be suggested that the primary drivers of monetisation were military rather than economic policies. Clearly, the royal policy model might offer part of the explanation but by itself it does not provide sufficient explanatory power. In addition to these vertical lines of interaction, a second part of the explanation can be found in the horizontal interaction and competition between communities. Throughout the Hellenistic period, local communities increasingly participated in dense economic, political and social networks on local, regional and interregional scales. At Sagalassos, for example, amphorae from the Aegean and Italic world provides evidence for the participation in long-distance networks of exchange – in addition to existing local and regional patterns of exchange – from the late 3rd and 2nd centuries BCE onwards (Monsieur et al., 2017). Inscriptions relate many instances of sympoliteia (joint citizenship) as a valid institution to create strong social and political connections between communities, such as between Xanthos and Myra in Lycia (Bousquet and Gauthier, 1994). Through the establishment of political federations such as the Cybriatic tetrapolis and the Lycian League, local communities created close alliances and gained an important base of political power beyond that of the individual city. The formation of these federations can be considered a bottom-up phenomenon following from the interactive relationship with the Hellenistic kings as a way to bridge the gap between the micro- and macro-level (Davies, 2002). Finally, the institution of synoikismos offered an even more extreme option to create strong bonds among communities by effectively merging two or more communities together and create larger socio-political units. This was a particularly prevalent mechanism in Karia, attested at sites such as Aphrodisias and Plarasa (LaBuff, 2015). It is through the synergy between horizontal and vertical lines of interaction and competition that developments in local complexity trajectories, as exemplified by Sagalassos, started to take off in the Hellenistic period. Let us conclude this case study by once more linking the observations above to the interpretative framework of the adaptive cycle and social complexity trajectories. The focus in the Hellenistic period increasingly shifted away from relatively isolated hilltop sites towards more accessible settlements in the lower areas. On the one hand, these were often the most fertile locations in the local landscapes, offering opportunities for local communities to increase their agricultural potential. On the other hand, the creation of an increasingly dense multi-scalar and complex network of connections between local communities and overarching polities through a variety of social, political and economic links offered additional incentives and stimuli that favoured those communities better able to participate in these

Social complexity trajectories in Anatolia 193 networks of exchange and interaction. The combination of local potential and external opportunities created a new niche with its own set of selection pressures for sustained development. The availability of this niche does not mean that all communities had equal access or opportunity to profit from it. Whatever the underlying reasons for the creation of the niche, once the stimuli associated with the initial situational event were produced, local communities still had to perceive the opportunity, interpret it, and react accordingly to exploit its potential (Daems, 2019). None of these steps should be taken for granted, and indeed the noted divergent development between certain local communities could have been caused by different perceptions of potential opportunities or their responses to them. By responding to these opportunities, local communities were able to kick-start their own complexity trajectories onto a wholly new level, resulting in the significant size increases observed in settlements during the Hellenistic period. For most of these communities, this transformation was fuelled by transitioning into a K-phase of conservation built on production maximisation and specialisation, increasing internal and external connectivity, and increasing returns to scale generated by agglomeration economies (Table 4.5). During LIA, a new cycle had been initiated, focused on strategies of (niche) diversification to meet the stimuli offered by the Achaemenid Table 4.5  Social complexity trajectories in the Hellenistic period. *Energy (E), resources (R), information (I); **diversification (D), intensification (IS), integration (IG) Selection Period pressure

Flows* Process

Hell

Demography Population growth

Hell

Hell

Demography Population growth Governance Polity policies Interaction Peer-polity interaction Competition Peer-polity interaction Interaction Group fusion

Hell

Interaction

Hell

Production

Hell

Production

Hell

Distribution

Hell Hell Hell

Population nucleation Energised crowding Agglomeration economies Agglomeration economies

Outcome Higher settlement counts Population nucleation Urbanism

Drivers

Mechanism**

E R I Pull Push D x

x

IS IG AC

x

K

x x x

x

x

K

x x x

x

x

K

x x x

x

x

K

Polity x x x competition Population x x nucleation Energised x x crowding Agglomeration x x economies Specialised x x production (Inter-)regional x exchange

x

Urbanism

x

x

K x

K

x

x

K

x

x

K

x

x

K

x

x

K

194  Social complexity trajectories in Anatolia empire. From the Hellenistic period onwards, the focus shifted towards intensification and specialisation as main mechanisms of complexity formation to enhance the exploitation of energy and resources from larger areas. As a result of these strategies of intensification, clear increases in the scale of flows of energy and resources that were exploited and transmitted can be noted. This was only possible by overcoming previous limits to information processing by building additional layers of administration and information processing technologies through the complex network of interactions between polities and local communities. This innovation allowed Seleucid strategies of urbanism to take effect, in synergy with community agency responding to their incentives to increase the demographic and economic base of local communities. I started this case study in the Chalcolithic period and have since traversed almost 4000 years to end up at this point, in a wholly different Anatolia. By elucidating the complexity trajectories of this long period, I have attempted to shed light onto some of the underlying drivers of social complexity trajectories driven by flows of energy, resources and information in a multiscalar and multi-dimensional perspective. I hope to have shown that this approach to study social complexity using complex systems thinking as the overarching framework can be highly promising. I will leave the conclusions of the full case study to the next chapter, where I will also address possibilities for further operationalising the conceptual model outlined here, as well as offer some potential roads for the future of social complexity studies.

Notes 1. See for example Netting, 1982. For a more extensive literature overview, see Supplementary Information of Kohler et al., 2017. 2. This was simply done by subtracting the date of the introduction of domesticates from the absolute date of the polity. Note that some polities have been included that precede the introduction of domesticated plants and animals. 3. As discussed in the supplementary materials of Turchin et al., 2018b, pp. 23–24. 4. A morphospace is an n-dimensional space, defined by the all relevant variables to define a given phenomenon. 5. The Atlas of Cultural Evolution is a database modelled after the Human Relations Area Files (https://hraf.yale.edu/) compiled by Peter Peregrine and Carol Ember for the Encyclopedia of Prehistory (Peregrine and Ember, 2003). 6. An ‘archaeological tradition’ is defined as “a group of populations sharing similar subsistence practices, technology, and forms of sociopolitical organization across a contiguous area and over a long period” (Ortman and Peregrine, 2018, pp. 189–190). 7. It should be noted, however, that Mellaart makes this suggestion specifically for Middle Chalcolithic Hacılar, even though no archaeological remains have been conclusively dated to this period. 8. This is a commonly posited interpretation in light of the first emergence of monumental architecture and fortification structures. To what extent this would actually have been a mental strategy of the local community remains difficult – if not

Social complexity trajectories in Anatolia 195 impossible – to assess. One potential avenue that could provide some information would be to conduct an extensive analysis into the average duration of site habitation in a comparison between fortified and unfortified sites. Such an analysis goes beyond the scope of the present case study. 9. The classes of settlement sizes identified are: 0.1–1 ha, 2–5 ha, 6–15 ha, and 20–43 ha. 10. However, the extent of the settlement during the EBA is not very clear. Other estimates suggested a size of about 20 ha during EB III (Abay, 2011). 11. The ancient name of Troy was Ilion, which could be seen as derived from (W) Ilios. This thesis was first proposed by the Swiss scholar Emil Forrer in the 1920s and has been the topic of extensive debates ever since. A reference to Alakšandu, a king of Wiluša has also been associated with Alexandros, another name for the Trojan prince Paris. While no conclusive agreement has been reached, many scholars find the association plausible (Jablonka, 2011). 12. Recent material studies conducted by the author revealed that some of the pottery material collected during the intensive surveys in the Burdur Plain could potentially include some LBA material at two sites in the area. Further studies are needed, however, to confirm these tentative results. 13. This is the central question in a volume edited by Ronald Faulseit (2016), focusing on social resilience and transformation instead of collapse to explain the decline and reorganization of complex societies in response to various internal and external stresses. 14. Land use intensity was measured by the average amount of sherds per year by period found during archaeological surveys. This is of course a very crude metric which does not take into account breakage patterns and depositional factors. Without the underlying data or more details on the method it is also very difficult to assess how the calculations took into account the different lengths of chronological periods and the certainty of chronological assignment per period, which is not necessarily solved by dividing the total amount of sherds by the amount of years in a period. 15. Geometric and black-on-red pottery was also attested at some of the hilltop sites but in far less quantities and of lesser quality compared to sites in the Burdur Plain. One notable finding from a site in the Burdur Plain is a single sherd identified as Corinthian Ware, dated to the early sixth century BCE (Personal communication with Dr. Cornelis Neeft). 16. Using the shotgun method 2.0 (Cleymans, 2018), we arrived at an estimate of 958 with a 1σ standard deviation of 504. 17. The labelling of Düzen Tepe as a simple, backward society is particularly contrasted with that of Sagalassos which would in Hellenistic times go on to become a ‘Greek’ polis, interpreted as an independent, urbanised socio-political unit embracing Hellenistic culture. I will discuss the fallacies of this interpretation in more detail in the final part of the case study concerning the Hellenistic period. 18. Pollen were obtained from 8 core sequences derived from three valleys within the area: Bereket, Gravgaz and Ağlasun (Bakker et al., 2012). 19. This means that whereas the full conception of polis would be characterised and identified by all identified traits, specific instances can be said to fit the general category by matching only a number of characteristics, rather than necessarily combining all of them. 20. Traits included are: Tribal affiliation, federal membership, alliance membership, league membership, party to a treaty, subject of synoikism or comparable processes (metoikism, dioikism, refoundation, sympoliteia, etc.), attestation of exiles, military matters, envoys, proxenia, naturalization, theorodokoi, civic subdivision, constitution type, public enactments, manifestations of legal systems, officials, assembly, public architecture, acropolis, walls, urbanism, mint, control of land

196  Social complexity trajectories in Anatolia ownership, taxation, free non-citizens, cults, calendar, communal oracle consultation, participation/victors in games, communal dedications, colonizer, colonised, and foundation myth. 21. The term synoikismos denotes the merging of two separate entities into a single community, consisting of either one community being subsumed into the other, or the foundation of a wholly new community located at a new location. 22. I quote Hansen (Hansen, 2002) who uses the term “indigenous” to describe the Lycian city-state culture, as opposed to the Greek poleis. Throughout this chapter, I have critiqued the naive usage of Greek as a cultural qualifier, especially with regards to local cultures given the inherent risk of reverting back to Eurocentric and Hellenocentric heuristics.

5

Conclusions

In this final chapter, I will bring together the various strands of thought developed throughout this book in several short reflections. I will start where I left off in the previous chapter, by offering a brief conclusion to the case study focusing particularly on drawing out some lines of thought in a diachronic overview. I will then again broaden the perspective to consider the merits and potential of the conceptual model proposed in this book. Rather than merely summarising earlier statements, I will discuss some potential approaches to further operationalise the model in archaeological research. Important avenues of operationalisation that will be addressed are quantification, methodological operationalisation, and computational modelling. Finally, I will outline some musings on how I believe the study of social complexity in particular – and archaeology in general – must move forward in the (near) future.

Reflections on the case study In the introduction to the case study in the previous chapter, I stated that to understand social complexity trajectories we need to describe patterns of development along a variety of domains – including economic, social, political, and environmental – and then integrate each of these patterns into a multi-scalar framework, indicating relevant connections between different scales and dimensions. In that chapter, I have attempted to bring such a multi-dimensional and multi-scalar account of change and stability in social complexity trajectories and define relevant connections between scales in a comprehensive manner through the adaptive cycle framework. I have shown how we can present social complexity trajectories without reverting back to social evolutionary models by focusing on flows of energy, resources and information. This approach can provide a rich and deep understanding of individual cases while also allowing ground for comparison through time and space. The approach advocated here can also help to overcome the idiosyncrasies of highly specialised scholarship. On the one hand, many scholars focusing on a single period or limited timeframe tend to stress the importance of that period in developments deemed

198  Conclusions transformative, such as increasing social complexity, the emergence of urban communities, interregional trade, etc. This is not surprising given that these scholars often have the expertise to examine the archaeological evidence in great detail and compare with preceding and subsequent periods to piece out potentially meaningful changes. It does mean, however, that virtually every era will be considered (incipiently) transformative in its own right. While that is not necessarily wrong, it is also not particularly useful. For example, in the case study, almost every transition between the BA periods (EB I–II, EB III, MBA, and LBA) has in the scholarly literature been described as transformative or characterised by an episode of crisis and transition inducing cycles of collapse and renewal. By building and understanding long-term trajectories of social complexity trajectories, we can contextualise the degree of stability and change in each of these periods and assess them in a comparative perspective. It is absolutely essential, however, that such long-term accounts are built on the right premises. In the final chapter of his book on EBA Anatolia, Christoph Bachhuber posits an evolutionary trajectory of development, in which he states that villages and settlement mounds from EB I-II can be considered chiefdoms developing into proto-states (EBIII), city-states (MBA), and territorial states (LBA). Note that the MBA and LBA periods are not explicitly covered in his book and therefore not extensively substantiated. For EBA chiefdoms, the core of the argument is built on the role of community elites as ritual specialists and wealth accumulators by creating structures of power and authority. Bachhuber convincingly argues that many of the societal developments in EBA Anatolia are driven by local elites exercising roles within their communities. However, by integrating his analysis into a series of stages in a social evolutionary model, he reduces much of the richness of his own argument without clear gain. The single benefit of using broad classificatory terms such as chiefdoms is that it allows some ground for comparisons through time and space. Bachhuber admits as much by stating that his model “relies on theory and historical and ethnographical analogy…[to] offer more explanatory power and comprehensiveness.” (Bachhuber, 2015, p. 184). Yet, he does not attempt to flesh out potential comparisons to help understand EBA Anatolia by seeing it in a different light. Denoting EB I–II polities as chiefdoms therefore provides little added value. It seems that his integration of a social evolutionary model is mainly driven by an attempt to add an additional layer of theory, without deep understanding of its implications. This is regrettable as it diminishes his otherwise comprehensive account of EBA Anatolia. In the case study, I have consciously steered away – as much as possible – from contentious terms such as chiefdom and state, in favour of a more neutral term such as polity or – where possible – kingdom. In my opinion, relying on (strict) units of classification conflicts with an approach based on flows of energy, resources and information. This case study has shown that the complexity model presented here can help move beyond Eurocentric

Conclusions 199 and evolutionary approaches to social complexity. One of its strong suits is that it defines each transition phase on its own terms in accordance with its characteristics. The integration of the adaptive cycle in the conceptual framework of the model helps to offer a better understanding of change and continuity on multiple scales and domains, as well as define inter-scalar interactions. Moreover, using complex systems thinking for the epistemic scaffolding of this model allows to ‘shift gears’ and switch between scales in a comprehensive manner. Coarse-graining between scales as part of a multi-scalar exploratory strategy has been particularly useful to distil patterns and meaningful connections. On the one hand, it allows switching between levels of theory where needed, for example by integrating middle level theories such as settlement scaling and agglomeration economies into the high-level theory of adaptive cycles and phase transitions. On the other hand, it facilitates understanding complexity trajectories on multiple levels and allows information from one scale to inform understanding of dynamics on a another scale, such as when I used regional level patterns to inform the interpretation of micro-level dynamics at Sagalassos. One of the premises of the model and case study in this book is that transitions between different forms of socio-political organisation on multiple scales are defined by quantitative differences in capture, transmission and consumption of flows of energy, resources and information. The apparent qualitative leaps in some transitions, such as the transformation from village to urban communities and from micro-regional polities to supra-­ regional polities, mark the crossing of one of the dual thresholds in scale and information processing. The interplay between these thresholds induces discontinuous trajectories of growth, consisting of slow accumulation of innovations in information processing and transmission – oftentimes invisible in the archaeological record – punctuated by episodes of transformative change once the potential built up by these innovations is capitalised by the effective capture and transformation of flows of energy and resources. In the model, these transitions are marked by the establishment of an information environment on the community level and the projection of these information environments on the polity level. The establishment of an information environment is most eminently observed in the formalisation of community structures as part of urban transformations, for example in the establishment of a structured settlement layout or the construction of public buildings that serve as loci for collective practices. Communities can grow into regional or supra-regional polities by projecting this information environment through the extension of power structures across a wider territory. This can occur in a first phase by integrating other settlements into a hierarchical settlement pattern and assuming the role of a central place. In a second phase of power projection, other central places are incorporated in an extended hierarchical structure based on a primary centre and a series of dependent secondary centres on a regional level. We have seen several examples of this process in the development of the Hittite, Phrygian and

200  Conclusions Lydian kingdoms, respectively developing out of their capitals at Hattuša, Gordion and Sardis. As a critical remark, I should note that this transition from community to polity has not been covered in detail. On the one hand, the archaeological record does not always provide a sufficiently fine-grained perspective to capture this transition. On the other hand, in-depth discussions of such transformations are undertakings in their own right and were hardly feasible in light of the current book. Each of these cases undoubtedly deserves its own dedicated volume and I would like to see such applications take form in the future. Throughout the case study, I applied the model of complexity trajectories on five main blocks (Chalcolithic – EBA – MBA/LBA – IA – Hellenistic) covering a chronological window of almost 4000 years. For each period, I elucidated relevant selection pressures, flows, push-pull forces and complexity mechanisms. However, except for the immediate transitions from one period to the next, I did not really paint a diachronic perspective of patterns throughout this chronological window. To conclude these reflections on the case study, I will now briefly trace some diachronic patterns from the data summary given in Figure 5.1. It should be noted that, for the sake of brevity, I do not consider here the different scales (local-regional-supra-regional) of analysis separately. Throughout the case study itself, I have extensively discussed complexity trajectories on distinct scales, so revisiting them would take away from the general diachronic lines that I want to piece out here. I also do not include the concordant phases in subsequent adaptive cycles, given that these were extensively described as well and have a diachronic logic of their own in light of the inherent progression within the cycle. The focus lies therefore squarely on the flows, push-pull forces and mechanisms of complexity formation. There is a lot to unpack in this figure. Each stacked bar corresponds to one of the main chronological blocks. On the top row we see the flows of energy, resources and information. In the middle we see push and pull forces, and below the three complexity mechanisms (diversification, integration and intensification). All the percentages are given per period, per feature and are normalised to mitigate any biases introduced by different numbers of observations per period. Starting with the top row, the diachronic evolution of which flows were most important in fuelling complexity trajectories is perhaps the most muddled picture of the three. For the Chalcolithic, it is clear that information flows are predominating, mainly because of the limits to information processing acting as strong pushing factors in this period. After the Chalcolithic, the relative proportions of each of the three flows become more similar, with a small trend towards a stronger role for more extensive exploitation and distribution of resources in IA and Hellenistic times. For the push-pull dynamics, a more varied (and therefore more interesting) picture emerges. Most apparent is the predominance of pulling factors in EBA and Hellenistic times. These are the periods of highest increase in

Conclusions 201

Figure 5.1  Diachronic overview of flows of energy (E), information (I) and resources (R), push-pull dynamics and complexity mechanisms of diversification (D), integration (IG), and intensification (IS) (made by author)

social complexity trajectories, which are in both periods driven by a relatively equal distribution of underlying flows. There is a marked difference, however, in the mechanisms behind this complexity trajectory. The role of intensification in particular is fairly limited in EBA, whereas it is almost twice as important in the Hellenistic period. I have indeed noted that the Hellenistic period was eminently characterised by intensification of processes driving scale increases in processes such as urbanism, trade and polity competition. The role of integration is about equal in EBA and Hellenistic times, suggesting that in both periods we see an extensive degree of network formation and interactions between social units across scales. Conversely, only IA is characterised by a predominance of pushing forces. This is the period where complexity trajectories were most prominently restricted by thresholds in scale increase, resulting in communities and polities using diversification as the main mechanism of complexity formation. This mainly resulted in the exploitation of additional environmental and

202  Conclusions social niches. It is only with the increased integrative measures developed in the Hellenistic period that communities obtained a sufficiently extensive economic base to transcend the earlier thresholds in scale expansion. Regarding the complexity mechanisms, it is interesting to note that after the Late Chalcolithic, diversification gradually grows in importance until the Hellenistic period when it almost disappears, whereas intensification is only important in LCH before dwindling in importance – even disappearing altogether in MBA/LBA times, before again becoming more prominent in the Hellenistic period. This indicates a process in which additional niches (both social and environmental) were continuously created until the limits of a system based on extensive exploitation of flows of energy, resources and information were reached and a switch towards intensification of system dynamics was necessary to support sustained complexity trajectories. It is also apparent how integration is almost negligible in the Late Chalcolithic and IA compared to the other three periods. This corresponds well to the predominance of diversification in the former, indicating again the comparative lack of integration in wider networks of interaction, at least with regards to meaningful drivers of complexity formation. Overall, the case study developed in this book adds weight to recent findings suggesting that polity growth is constrained by thresholds of information processing and scale (defined here as the scope of energy and resources that can be captured) as an explanatory factor for punctuated trajectories of complexity formation. Once an important threshold in information processing was crossed in the Chalcolithic period, communities and polities were able to take a next step in scalar growth during EBA. At this point, however, a renewed threshold of information transmission restricted further growth until the Hellenistic period. It was only through the synergy between extensive horizontal and vertical structures of interaction and information exchange that this threshold was overcome and a new phase of scale increase built on intensification and integration of larger polities was made possible. At the same time, the case study also showed that it is essential to integrate multi-scalar perspectives on complexity trajectories by showing how micro and macro levels can both mutually reinforce each other or how one can remain wholly decoupled from major transitions on the other. In many respects, the present case study offers only the first steps towards deepening our understanding of complexity trajectories by applying this model. Further studies and further reflections are required for this research approach to gain maturity. In the next part, I will posit some reflections on this conceptual model and how to improve it towards the future.

Reflections on the conceptual model The basic purpose of the model presented here is to provide a way beyond social evolutionary approaches in conceptualising complexity trajectories. I have particularly argued for the suitability of the tenets of complex

Conclusions 203 systems thinking geared towards approximating flows of energy, resources and information. The model consists of a set of selection pressures that act as information input for problem-solving loops consisting of information processing and decision making, using complexity mechanisms to produce outcomes with a pulling or pushing effect, respectively concentrating or dissipating the flows of energy, resources and information that drive social complexity. This basic model of complexity formation was then applied to two main levels of social organisation: communities and polities (Figure 5.2). It is a trope that archaeologists are generally not very mathematically educated and are more often inclined to frame their analyses in narrative-based approaches. While narrative-based research is not inherently wrong, sometimes a different approach is required. Complex systems studies is a field where such formal approaches are particularly necessary given the non-­linear dynamics that characterise it. Therefore, the application of complex systems thinking in archaeology requires a formal approach as well. In recent years, archaeologists have increasingly been urged to start building systematically formalised approaches in theory building, hypothesis testing and data analysis (Smith, 2015). Formalised approaches have the advantage of being explicit and facilitate comparative analysis. Some people, however, mistake formalisation for building a set of mathematical equations. Equations indeed provide the ultimate form of formalisation, exactly defining how and how much each factor influences the overall system. While such equations can in principle be created by archaeologists if one has the inclination and relevant mathematical education, the archaeological record often does not offer the necessary resolution to empirically test and validate them. Luckily for us, mathematical equations are not the only type of formalisation. The model structure outlined in Figure 5.2 likewise constitutes a formal model, albeit one with less precision than an equation. This model explicitly defines the constituent components of complexity formation and how they relate to each other, even though these relationships are not defined in exact mathematical terms. Beyond the limitations of my own mathematical background, I feel that before we can start defining these equations, we need to obtain a better understanding of the conceptual and theoretical terms of the debate. The goal of this book was therefore to contribute specifically to formalisation in relevant theory building. This formal model, linking selection pressures and decision-making to outcomes on complexity trajectories through causal mechanisms of complexity formation therefore only constitutes a first step in the formalisation of social complexity studies in archaeology. We are only just getting started! As archaeologists, it is moreover, good to keep in mind that “even a model which uses formalization is in many cases necessarily a metaphor in archaeology” (McGlade and van der Leeuw, 1997, p. 22). I noted in the introduction of this book that moving from metaphors to algebra is the work of a discipline and cannot be accomplished in a single book, at least as long as the work is ongoing and the book is not a retrospective. Moreover, between

204  Conclusions

Figure 5.2  Basic model of complexity formation and its application on the level of communities and polities (made by author)

Conclusions 205 metaphor and algebra lies not a dark abyss or barren wasteland, but rather a rich possibility space of possible research methods and approaches with varying degrees of formalism. Likewise, the debate regarding the usage of quantitative versus qualitative methods is no dichotomy. Many different qualitative approaches exist, with high degrees of formalism and which are perfectly compatible with quantitative approaches (Ragin, 1989). In this part, I wish to elucidate three approaches to extend the current work both on a quantitative and qualitative level, by: (1) integrating quantified measures for some of the existing components of the model; (2) extending the model through further methodological operationalisation; (3) exploring new avenues of theory building and analysis through modelling approaches. Quantification Archaeologists have long been interested in developing quantitative measures of social complexity. The goal of these measurements is to order societies along a parameter of interest – for example, hierarchical levels of organisation – and outline subsequent stages of societal development along a single complexity trajectory. Frequently applied techniques include Guttman scaling (Carneiro, 1967), index ranking (Bowden, 1972, 1969), and factor analysis (Erickson, 1977). All of these have become staples in cross-cultural anthropological and archaeological studies. Yet, it has been noted that they all suffer from the same inherent limitations related to discarding data and minimising variance (McGuire, 1983). They reduce multiple variables to a single score, thus averaging out the effects of the individual components or coarse-grain continuous variables into presence/absence or discrete scores. This is especially problematic in the case of social complexity, which is a multi-dimensional and multi-scalar phenomenon that is not easily captured in a single quantitative metric. Archaeologists, also generally deal most frequently with qualitative data which are not easily captured or integrated in quantified approaches. Even though the methods of quantification outlined above can be (and have been) used to ‘quantify’ qualitative data, a certain degree of interpretation is required for the very preparation of the data. In the previous chapter, I discussed a recent approach by the Seshat project to build a composite metric of social complexity using PCA analysis of complexity characteristics. The same criticism can be levelled at this approach as it also requires discarding data and minimising variance. Yet, I still decided to use this metric throughout the case study as a baseline to situate the application of my model. Given that my approach does not offer a quantification of its own, this comparison may have come across as being not very explicit. Then again, quantifying should not necessarily be automatically equated with understanding. For example, what does it mean when we say that the complexity value for the Hellenistic period is slightly lower than that of Achaemenid times? Even if we set aside these

206  Conclusions general concerns regarding quantification, given that we are dealing with a composite metric it is not easy to grasp what exactly is causing the difference and to analyse the effects of every variable by itself would hardly have been feasible in the context of this book. Moreover, given that complex systems are generally driven by multi-causal processes that mutually reinforce each other, it is to be expected that attempts at isolating signals of change will not be very fruitful. Despite these misgivings, quantitative analyses have been part and parcel of the archaeological discipline and continue to gain prominence (Aldenderfer, 2018, 1998). This should not come as a surprise. Even though many archaeologists are hesitant to turn to mathematical tools, the advantages of quantification are legion. They allow for better comparability of case studies across time and space, as well as a formal and tractable analysis of factors relevant to the phenomenon under study. So, which quantitative measures are available and how can we apply them to reap these rewards? A rich literature exists on information, entropy and diversity measures which have been posited as suitable measures of (social) complexity (Lloyd, 2001). Three main groups of measurement can be discerned, related to measures of (1) difficulty of creation; (2) difficulty of description; and (3) degree of organization. The first group is mainly related to human-made or engineered complex systems and therefore not very relevant for ‘organically’ developing complexity in human systems. Many complexity measures from the second group hail from the field of cybernetics, based on measures of communication information and system entropy in description length of a given system. While entropy measures of information description work great in theory, they are often cumbersome to calculate and difficult to apply in practice. Still, some headway has recently been made in applying information measures to modelling social organisations as communication networks which seems promising (Wolpert et al., 2017). The third group of measurements capturing the degree of organisation is more closely related to the approach outlined in the conceptual model here, using relative complexity and mechanisms of change, and seems more relevant as a potentially fruitful extension of the model. To develop this approach, we could measure the intensity of causal factors in developing social complexity through these mechanisms. Elsewhere, I have preliminarily applied such an approach for calculating socio-economic complexity development (Daems, In Press). While much work is still needed to explore the full potential of this approach, preliminary results are promising. One way to improve the current status of the method could be by incorporating formal measures of the relative distance in subjective complexity measure proposed here. This distance could potentially be approximated by similarity measures such as Jaccard distance (Collins-Elliott, 2016; Rubio-Campillo et al., 2018), Hamming distance (Mertel et al., 2018), morphometric similarity (Coto-Sarmiento et al., 2018), and similarity networks (Östborn and Gerding, 2015).

Conclusions 207 A second approach to explore pertains to a fuller application of the social metabolism framework. In this book, I focused mainly on some calculations of endosomatic energy requirements in light of the discussion on the carrying capacity of the landscape between Düzen Tepe and Sagalassos. This posits only a first step towards the full implementation of the metabolism framework as the aspect of exosomatic energy requirements has, to my knowledge, so far not been (fully) applied yet. To do so, we need to supplement calculations of agricultural yields and energetic needs with approximations for the energy requirements of a variety of processes and activities such as artisanal production, architectural construction, transport, etc. The first steps towards this aim have already been undertaken, for example in a recent model of energetic requirements for heating, both in household contexts and for the upkeep of a public building such as the Roman bathhouse at Sagalassos (Janssen et al., 2017). Current work is aiming to expand this approach to approximate the fuel requirements of pottery production.1 Likewise, a vibrant field of architectural energetics has started to emerge in recent years which holds some potential to be integrated in the social metabolism framework (McCurdy and Abrams, 2019). Finally, another rapidly developing mathematic framework is building up in the field of settlement scaling. Even though the framework has not yet reached the same maturity as the field of metabolic scaling from which it is derived, the approach is gaining ground fast in archaeology (Lobo et al., 2019; Ortman et al., 2014; Smith, 2017). The appeal is easy to see given that, despite myriad potential biases (Schofield, 1991), site size approximations is one of the few consistent data sources that are available or fairly straightforward to compile.2 The main challenges to move the field of settlement scaling forward have been noted in Chapter 2: further theoretical maturation to explore the underlying dynamics of scaling laws in settlement systems and defining proper archaeological proxies to capture increasing returns to scale characteristic for super-linear scaling. Once these challenges are overcome, settlement scaling can march on to become one of the core methodological approaches in archaeology and help us understand the underlying patterns and dynamics of settlement systems in the past. Methodological operationalisation One aspect of the conceptual mode that particularly merits further exploration and that I want to highlight here, is the operationalisation of the adaptive cycle framework. Some analytical methods to study the multi-scalar dynamics of adaptive cycles have been proposed, most notably discontinuity analysis (Sundstrom et al., 2014). To translate this analytical framework into a quantitative and analytical approach, however, further operationalisation of the adaptive cycle is needed. Most notably it requires the identification of proper variables or archaeologically relevant proxies. The idea of discontinuity analysis is that when you take a continuous variable – say,

208  Conclusions population size of settlements for example – analysis of the distribution of this variable will show clear discontinuities, forming discrete classes if this variable corresponds to multi-scalar dynamics. These discontinuities can be analysed through a wide range of methods such as Bayesian classification and regression trees, Monte Carlo approaches such as the Gap Rarity Index, hierarchical cluster analysis, fractal dimensions, and time series analysis (Stow et al., 2007; Sundstrom et al., 2014). When constructing an adaptive cycle framework within an archaeological case study, it is essential to consider the challenging task of defining archaeologically valid parameters of connectedness and potential. Connectedness has, for example, been defined as the intensity of subsistence and exploitation strategies, mobility, and social organisation, whereas potential has been considered as ‘the potential for innovation’ (Rosen and Rivera-Collazo, 2012). Other definitions focus on connectedness as the level of vertical and horizontal social differentiation (Peters and Zimmermann, 2017), or consider the element of ‘integration’ to be crucial (Hegmon et al., 2008). A wide range of variables has been used in applications of resilience theory and adaptive cycles in archaeology. In a recent overview, Marcel Bradtmöller and his colleagues (2017, p. 5) identified four main proxies and related attributes to conceptualise/classify complexity within past social-ecological systems: subsistence, demography, social organisation, and technological innovation. Each of these can be studied through several proxy attributes. Subsistence, for example, has been captured through agricultural intensity (Weiberg, 2012), diversity and abundance of resources (Allcock, 2017; Nelson et al., 2006; Rosen and Rivera-Collazo, 2012), food storage (Allcock, 2017; Nelson et al., 2006), subsistence specialisation (Marston, 2015; Solich and Bradtmöller, 2017), variability in subsistence systems (Bicho et al., 2017; Gronenborn et al., 2017), and trading and redistributive networks (Cooper, 2012). Demographic trends can be studied through population size (Stiner and Kuhn, 2006), rate of population growth (Allcock, 2017; Marston, 2015), and limiting threshold populations (Gronenborn et al., 2017; Weiberg, 2012). Social organisation includes forms of social control (Allcock, 2017; Dunning et al., 2012; Nelson et al., 2011; Weiberg, 2012), social interaction networks (Cooper, 2012), and social mobility (Peters and Zimmermann, 2017; Zimmermann, 2012). Technological innovation is used less frequently as a proxy for adaptive cycle development, but if used, the rate of innovation is often equated with potential for system development as characteristic for the reorganisation or conservation domains (Rosen and Rivera-Collazo, 2012; Thompson and Turck, 2009). It has been suggested that to study dynamics of change social and ecological systems, at most 3–5 key variables should be used for analysis as to not overcomplicate the analysis, that is the so-called ‘rule of hand’ (Yorque et al. 2002; Walker et al. 2006). The main challenge is of course to compile proper datasets for any or several of these variables in order to capture the multi-dimensional and multi-scalar dynamics of the adaptive cycle framework (see infra).

Conclusions 209 Modelling complexity The main focus on complex systems thinking as the epistemic scaffolding of the conceptual model of social complexity developed here, inevitably requires a conceptualisation of bottom-up emergence in complexity trajectories. One potentially fruitful approach to further approximate the parameters of this bottom-up emergence could be to use computational modelling, more specifically agent-based modelling (ABM), to improve our theoretical understanding of this process, simulate potential scenario’s and test hypotheses. Earlier, I already discussed the different degrees of formalisation in methodological approaches. Different types of models are also characterised by diverging degrees of formalism. Verbal models are generally informal and define how processes relate to each other in loosely conceptual terms. On the other end of the spectrum, mathematical models consist of precise equations detailing exactly how each term relates to another and the effects of each component on the whole. Somewhere along the range in-between these both extremes, we find – among others – computational models. These types of models are highly formalised, meaning that all assumptions and relations between components need to be made explicit in order for a computational model to work. Yet, they also incorporate a degree of probability and uncertainty not present in mathematical equations that allows the modeller to derive a solution for an analytically intractable mathematical model. I already provided some background for ABM and its applications in archaeology in the second chapter so I will be rather brief here. To recapitulate, ABM simulates agents – depending on the question of interest these can be individuals, groups, cities or even polities – acting autonomously in heterogeneous ways, often based on decentralised information gained from interactions with other agents and their environment. Social complexity is a prevalent topic in ABM studies and various stages of complexity formation have been addressed through computer simulation, such as its emergence (White, 2013), development (Griffin, 2011), consolidation in settlement patterns (Altaweel, 2015; Griffin and Stanish, 2007), polity formation (Chliaoutakis and Chalkiadakis, 2016), and its collapse and transformation (Dean et al., 2000). The conceptual model developed here is therefore inheritably suited to be translated into an ABM. Such a model will need to be able to sufficiently capture essential emergent processes, particularly the concentration and dissipation of flows of energy, resources, and information on multiple levels. The main mechanisms behind this complexity formation are information transmission and uncertainty reduction. These are expressed on the micro level through the bundling of social interactions into practices and institutions, on the meso level as the establishment of information environments expressed in social and material domains, and, on the macro level through the establishment of power structures and the projection of information environments over a larger area. I have recently started to build an ABM

210  Conclusions based on these general premises, specifically applied to the origin of polis communities in Iron Age to Hellenistic southwest Anatolia, and will aim to develop this model further in the near future.3

Reflections on the future As noted earlier, the time for retrospectives in social complexity studies has not yet come. Instead, we must look at what the future might have in store for this exciting field and how we can prepare for it. In the introduction, I already referred to the extensive plateau of productivity from Gartner’s hype cycle that remains to be explored. I hope that the model and its application presented here offers a few first steps on the road to discovery. To aid our expedition, a few tools will be indispensable. These are: (1) capacity building; (2) data availability; (3) interdisciplinarity; and (4) inclu­ siveness. First, to continue developing the field of social complexity studies, especially from the perspective of complex systems thinking, we need to build a critical mass of theoretical knowledge and skills to support it. Especially for archaeologists, these are no trivial matters given that they are often not part of the core business of archaeological curricula. It requires us to gain understanding of the necessary background theories on complex systems thinking as well as knowledge of the relevant tools of complex systems science such as network analysis or ABM. If we ever want to move these conceptual and methodological approaches to the core of the archaeological discipline, a wide range of archaeologists needs to be, if not specialised in them, at least familiar with hem. It has been noted that “not all of us are trained in the chemistry of isotope analysis but most of us are likely to accept their validity. Instead of evaluating the validity of the method in these cases when we know little about it, we evaluate the logic of the argument and its supporting evidence. ABM ought to be held to this same standard” (Cegielski and Rogers, 2016, p. 292). Without integrating ABM or other aforementioned tools in the archaeological education, we cannot ask the average archaeologist to be able to evaluate the epistemological and empirical validity of such applications. This requires integration in archaeological curricula as well as more extensive training opportunities and workshops to build specialised expertise. Second, complex systems studies inevitably require the availability of extensive structured datasets. This cannot be dissociated from the push towards open science which is – rightfully – more and more gaining ground in academia. The importance of proper data collection, coding practices, and methods, as well working in a transparent framework geared towards openness and reproducibility cannot be overstated. Throughout this book, I have frequently noted the works of the Seshat Databank as an alternative approach to mine, comparing my data with theirs. This was only possible because of their serious commitment to open science. Regardless of the debates and controversies that have raged in social complexity studies in

Conclusions 211 recent years, one cannot deny that an important reason that these debates can be held in the first place, is the fact that data, methodologies and computer code are offered in full transparency and openness. If we really want to push archaeology forward to better understand the past, it is absolutely essential to start collaborating as a research community, rather than competing as splintered research groups or individuals. This brings me to the third point, interdisciplinarity. Pushing the limits of knowledge inevitably requires transcending disciplinary boundaries. By continuing to adhere to the prevalent silo mentalities in academia, we will never gather sufficient momentum to keep moving forward. If anything, the absolute lack of disciplinary boundaries is perhaps the single most important element that makes the Santa Fe Institute such a successful research institute in the study of complex systems thinking. Having had the privilege of experiencing a small taste of the atmosphere at the Institute, I cannot overstate the importance of establishing an open research environment where background disciplines are nothing more than an open invitation for curiosity, discussion and interaction rather than impermeable boundaries for isolated compartments. The need for open collaboration also brings out the fourth point I want to stress, inclusiveness. Too much inbreeding will render any species sterile and prone to extinction. Homo academicus is no exception to the rule. Ensuring sufficient creativity and ideas to sustain the growth of a field requires people of all backgrounds – social background, nationality, race, gender, etc. – to be allowed to participate in its cross-fertilisation. Even the most boundary-less research environments will remain inherently stillborn if they are founded on social or academic mono-cultures or other exclusionary practices. Only after we manage to unlock the inherent potential of these tools will social complexity studies – or any field of study for that matter – truly be able to blossom and move forward, pushing the boundaries of our knowledge to reveal truth and beauty hitherto unknown. How better to conclude this book than with words so elegant that they could not possibly have come from my own pen: “The time will come when the progress of research and prolonged study will reveal to sight the mysteries of nature that are now concealed. A single lifetime, though it were wholly devoted to the study…does not suffice for the investigation of problems of such complexity…The day will yet come when our successors will wonder how we could have been ignorant of things so obvious”.4 Words to live by and I hope to see the day.

Notes 1. Personal communication with Ph.D. student Stef Boogers of the Sagalassos Project (University of Leuven). 2. Straightforward but time consuming! 3. Source code of current and future versions is openly available at: https://github. com/driesdaems10/PolisABM. 4. Seneca, Naturalis Quaestionis VII.25.

Bibliography

Abay, E., 2011. Preliminary Report on the Survey Project of Çivril Baklan and Çal Plains in the Upper Meander Basin Southwest Anatolia. Ancient Near Eastern Studies 48, 1–87. Adams, R.M., 1966. The Evolution of Urban Society. Aldine Transaction, New Brunswick. Adger, W.N., 2000. Social and ecological resilience: Are they related? Progress in Human Geography 24, 347–364. https://doi.org/10.1191/030913200701540465 Adler, M.A., Van Pool, T., Leonard, R.D., 1996. Ancestral Pueblo Population Aggregation and Abandonment in the North American Southwest. Journal of World Prehistory 10, 375–438. Aimers, J., Iannone, G., 2013. The dynamics of ancient Maya developmental history, in: Iannone, G. (Ed.), The Great Maya Droughts in Cultural Context. Case Studies in Resilience and Vulnerability. University Press of Colorado, Boulder, pp. 21–51. Alberti, G., 2014. Modeling group size and scalar stress by logistic regression from an archaeological perspective. PLoS ONE 9, e91510. https://doi.org/10.1371/ journal.pone.0091510 Aldenderfer, M., 1998. Quantitative methods in archaeology: A review of recent trends and developments. Journal of Archaeological Research 6, 91–120. Aldenderfer, M., 2018. Quantification and social life, in: The Encyclopedia of Archaeological Sciences. American Cancer Society, pp. 1–6. https://doi.org/ 10.1002/9781119188230.saseas0485 Allcock, S.L., 2017. Long-term socio-environmental dynamics and adaptive cycles in Cappadocia, Turkey during the Holocene. Quaternary International, Adaptive Cycles in Archaeology 446, 66–82. https://doi.org/10.1016/j.quaint.2017.06.065 Altaweel, M., 2015. Settlement dynamics and hierarchy from agent decisionmaking: A method derived from entropy maximization. J Archaeol Method Theory 22, 1122–1150. https://doi.org/10.1007/s10816-014-9219-6 Altschul, J.H., Kintigh, K.W., Klein, T.H., Doelle, W.H., Hays-Gilpin, K.A., Herr, S.A., Kohler, T.A., Mills, B.J., Montgomery, L.M., Nelson, M.C., Ortman, S.G., Parker, J.N., Peeples, M.A., Sabloff, J.A., 2018. Fostering collaborative synthetic research in archaeology. Advances in Archaeological Practice 6, 19–29. https:// doi.org/10.1017/aap.2017.31 Anderson, P.W., 1972. More is different. Science 177, 393–396. https://doi.org/10.1126/ science.177.4047.393

Bibliography 213 Ando, C., 2018. The political economy of the Hellenistic polis: Comparative and modern perspectives, in: Börm, H., Luraghi, N. (Eds.), The Polis in the Hellenistic World. Rethinking the Polis in the Hellenistic Age. Franz Steiner Verlag, Stuttgart, pp. 9–26. Aoki, K., 2018. On the absence of a correlation between population size and ‘toolkit size’ in ethnographic hunter–gatherers. Philosophical Transactions of the Royal Society B 373, 20170061. https://doi.org/10.1098/rstb.2017.0061 Aperghis, G.G., 2004. The Seleukid Royal Economy: The Finances and Financial Administration of the Seleukid Empire. Cambridge University Press, Cambridge. Arik, O., Meriç, R., Korfmann, M., Neve, P., Cauvin, J., Erzen, A., Esin, U., Tuchelt, K., Baydur, N., Vetters, H., Minzoni-Déroche, A., Ferrero, D. de B., Berti, F., Bilgi, Ö, Sevin, V., Aslanapa, O., Yalman, B., Mikami, T., Yalçınkaya, I., Öğün, B., Arık, R., Duru, R., Cauvin, J., Hauptmann, H., Borchardt, J., Bingöl, O., Davesne, A., Müller-Wiener, W., Atalay, E., Alpagut, B., Radt, W., Greenewalt, C.J., Bass, G., Çilingiroğlu, A., le Roy, C., Özdogan, M., 1987. Recent archaeological research in Turkey. Anatolian Studies 37, 179. https://doi.org/10.2307/3642895. https://doi.org/10.2307/3642895 Arthur, W.B., 1989. Competing technologies, increasing returns, and lock-in by historical events. The Economic Journal 99, 116–131. https://doi.org/10.2307/ 2234208 Arthur, W.B., 2009. The Nature of Technology: What It Is and How It Evolves. Simon and Schuster. Arthur, W.B., 2015. Complexity and the Economy, 1st ed. Oxford University Press, Oxford; New York. Ashby, W.R., 1956. An Introduction to Cybernetics. Wiley, New York. Auban, J., Martin, A., Barton, C., 2013. Complex systems, social networks, and the evolution of social complexity in the East of Spain from the Neolithic to preRoman times, in: Berrocal, M., Sanjuán, L., Gilman, A. (Eds.), The Prehistory of Iberia: Debating Early Social Stratification and the State. Routledge, New York. Axtell, R.L., Epstein, J.M., Dean, J.S., Gumerman, G.J., Swedlund, A.C., Harburger, J., Chakravarty, S., Hammond, R., Parker, J., Parker, M., 2002. Population growth and collapse in a multiagent model of the Kayenta Anasazi in Long House Valley. PNAS 99, 7275–7279. https://doi.org/10.1073/pnas.092080799 Ayres, R.U., Kneese, A.V., 1969. Production, consumption, and externalities. The American Economic Review 59, 282.297. Bachhuber, C., 2015. Citadel and Cemetery in Early Bronze Age Anatolia. Equinox Publishing Limited. Bailey, K.D., 1990. Social Entropy Theory. SUNY Press. Bailey, K.D., 1994. Sociology and the New Systems Theory: Toward a Theoretical Synthesis. SUNY Press. Bak, P., 1996. How Nature Works: the Science of Self-Organized Criticality. Copernicus. Bak, P., Tang, C., Wiesenfeld, K., 1987. Self-organized criticality: An explanation of 1/f noise. Physical Review Letters 59, 381–384. https://doi.org/10.1103/ PhysRevLett.59.381 Bakker, J., Paulissen, E., Kaniewski, D., De Laet, V., Verstraeten, G., Waelkens, M., 2012. Man, vegetation and climate during the Holocene in the territory of Sagalassos, Western Taurus Mountains, SW Turkey. Vegetation History and Archaeobotany 21, 249–266.

214  Bibliography Ball, P., 2004. Critical Mass: How One Thing Leads to Another. Farrar, Straus and Giroux, New York. Bandy, M.S., 2004. Fissioning, scalar stress, and social evolution in early village societies. American Anthropologist106, 322–333. Bang, P.F., Scheidel, W., 2013. The Oxford Handbook of the State in the Ancient Near East and Mediterranean. Oxford University Press, Oxford. Barabási, A.-L., Albert, R., 1999. Emergence of scaling in random networks. Science286, 509–512. https://doi.org/10.1126/science.286.5439.509 Barton, C.M., 2014. Complexity, social complexity, and modeling. Journal of Archaeological Method and Theory; New York 21, 306–324. Barton, C.M., Ullah, I.I.T., Bergin, S.M., Mitasova, H., Sarjoughian, H., 2012. Looking for the future in the past: Long-term change in socioecological systems. Ecological Modelling 241, 42–53. https://doi.org/10.1016/j.ecolmodel.2012.02.010 Bausch, K., 2001. The Emerging Consensus in Social Systems Theory. Springer US. Beheim, B., Atkinson, Q., Bulbulia, J., Gervais, W.M., Gray, R., Henrich, J., Lang, M., Monroe, M.W., Muthukrishna, M., Norenzayan, A., Purzycki, B.G., Shariff, A., Slingerland, E., Spicer, R., Willard, A.K., 2019. Corrected analyses show that moralizing gods precede complex societies but serious data concerns remain (preprint). PsyArXiv. https://doi.org/10.31234/osf.io/jwa2n Bell, C., 2009. Continuity and change: The divergent destinies of Late Bronze Age ports in Syria and Lebanon across the LBA/Iron Age transition, in: Bachhuber, C., Roberts, R.G. (Eds.), Forces of Transformation: The End of the Bronze Age in the Mediterranean: Proceedings of an International Symposium Held at St. John’s College, University of Oxford 25 February–6th March 2006. Oxbow, Oxford, pp. 30–38. Bentley, A., Maschner, H., 2003. Complex Systems and Archaeology. Utah University Press, Salt Lake City. Bentley, R.A., 2003. An introduction to complex systems, in: Bentley, R., Maschner, H. (Eds.), Complex Systems and Archaeology. Utah University Press, Salt Lake City, pp. 9–23. Bentley, R.A., Maschner, H.D.G., 2007. Complexity theory, in: Bentley, R., Maschner, H. (Eds.), Handbook of Archaeological Theories. AltaMira Press, Lanham, pp. 245–270. Berger, P.L., Luckmann, T., 1966. The Social Construction of Reality: A Treatise in the Sociology of Knowledge. Doubleday. Berkes, F., Folke, C., Colding, J., 1998. Linking Social and Ecological Systems: Management Practices and Social Mechanisms for Building Resilience. Cambridge University Press, Cambridge. Bettencourt, L., 2013. The Origins of Scaling in Cities. Science 340, 1438–1441. https://doi.org/10.1126/science.1235823 Bettencourt, L., Lobo, J., Strumsky, D., 2007. Invention in the city: Increasing returns to patenting as a scaling function of metropolitan size. Research Policy 36, 107–120. https://doi.org/10.1016/j.respol.2006.09.026 Bettencourt, L., West, G., 2010. A unified theory of urban living. Nature 467, 912–913. https://doi.org/10.1038/467912a Bettinger, R.L., Eerkens, J., 1997. Evolutionary implications of metrical variation in Great Basin projectile points. Archeological Papers of the American Anthropological Association 7, 177–191. https://doi.org/10.1525/ap3a.1997.7.1.177

Bibliography 215 Bettinger, R.L., Garvey, R., Tushingham, S., 2015. Hunter-Gatherers: Archaeological and Evolutionary Theory, 2nd ed., Interdisciplinary Contributions to Archae­ ology. Springer US. Bicho, N., Cascalheira, J., Gonçalves, C., Umbelino, C., Rivero, D.G., André, L., 2017. Resilience, replacement and acculturation in the Mesolithic/Neolithic transition: The case of Muge, central Portugal. Quaternary International, Adaptive Cycles in Archaeology 446, 31–42. https://doi.org/10.1016/j.quaint.2016.09.049 Binford, L., 1977. For Theory Building in Archaeology: Essays on Faunal Remains, Aquatic Resources, Spatial Analysis, and Systemic Modeling. Academic Press, New York. Binford, S.R., 1968. New Perspectives in Archaeology. Aldine Publishing Company. Bintliff, J.L., 1991. The Annales School and Archaeology. Leicester University Press. Bintliff, J., 1997. Catastrophe, chaos and complexity: The death, decay and rebirth of towns from antiquity to today. The Journal of European Archaeology 5, 67–90. Bintliff, J., 1999. Structure and Contingency: Evolutionary Processes in Life and Human Society. Leicester University Press. Bintliff, J., 2003. Searching for Structure in the Past— Or Was It “One Damn Thing after Another”?, in: Bentley, R., Maschner, H. (Eds.), Complex Systems and Archaeology. Utah University Press, Salt Lake City, pp. 79–83. Bintliff, J., 2009. Catchment, settlement chambers and demography: Case studies and general theory in the Greek landscape from Prehistory to Early Modern Times, in: Favory, F., Nuninger, L. (Eds.), University of Burgundy, Dijon, pp. 107–117. Bintliff, J., Fellmeth, U., Guyot, P., Sonnabend, H., 2007. Emergent complexity in settlement systems and urban transformations, in: Historische Geographie Der Alten Welt. Grundlagen, Erträge, Perspektiven. Festgabe Für Eckart Olshausen. Hildesheim, pp. 43–82. Birch, J., 2013a. Between villages and cities. settlement aggregation in cross-cultural perspective, in: Birch, J. (Ed.), From Prehistoric Villages to Cities: Settlement Aggregation and Community Transformation. Routledge, Taylor & Francis Group, pp. 1–22. Birch, J., 2013b. From Prehistoric Villages to Cities: Settlement Aggregation and Community Transformation. Routledge, Taylor & Francis Group. Blanton, R.E., Fargher, L.F., 2016. How Humans Cooperate: Confronting the Challenges of Collective Action. University Press of Colorado, Boulder. Blau, P.M., 1964. Exchange and Power in Social Life. Transaction Publishers. Bobrowsky, P.T., Ball, B.F., 1989. The theory and mechanics of ecological diversity in archaeology. Quantifying diversity in archaeology 4–12. Boehm, R., 2018. City and Empire in the Age of the Successors - Urbanization and Social Response in the Making of the Hellenistic Kingdoms. University of California Press, Oakland. Bogaard, A., Fochesato, M., Bowles, S., 2019. The farming-inequality nexus: New insights from ancient Western Eurasia. Antiquity 93, 1129–1143. https://doi. org/10.15184/aqy.2019.105 Bogucki, P., 2003. The Neolithic settlement of Riverine interior europe as a complex adaptive system, in: Bentley, R., Maschner, H. (Eds.), Complex Systems and Archaeology. Utah University Press, Salt Lake City, pp. 93–103. Boix, C., 2015. Political Order and Inequality. Cambridge University Press.

216  Bibliography Boserup, E., 1965. The Conditions of Agricultural Growth: The Economics of Agrarian Change Under Population Pressure. Aldine, New York. Bottema, S., Woldring, H., 1984. Late quaternary vegetation and climate of southwestern Turkey. Palaeohistoria 26, 123–149. Boulton, J.G., Allen, P.M., Bowman, C., 2015. Embracing Complexity: Strategic Perspectives for an Age of Turbulence. Oxford University Press. Bourdieu, P., 1977. Outline of a Theory of Practice. Cambridge University Press, Cambridge. Bourdieu, P., 1986. The Forms of Capital. Bousquet, J., Gauthier, P., 1994. Inscriptions du Létôon de Xanthos. Revue des Études Grecques 107, 319–361. https://doi.org/10.3406/reg.1994.2621 Bowden, E., 1969. An index of sociocultural development applicable to precivilized societies. American Anthropologist 71, 454–461. https://doi.org/10.1525/ aa.1969.71.3.02a00040 Bowden, E., 1972. Standardization of an index of sociocultural development for precivilized societies. American Anthropologist 74, 1122–1132. https://doi. org/10.1525/aa.1972.74.5.02a00060 Boyd, R., Richerson, P.J., 1988. Culture and the Evolutionary Process, University of Chicago Press, Chicago. Boyd, R., Richerson, P.J., 2009. Culture and the evolution of human cooperation. Philosophical Transactions of the Royal Society B: Biological Sciences 364, 3281–3288. https://doi.org/10.1098/rstb.2009.0134 Boyer, P., 2018. Minds Make Societies Cognition Explains, 1st ed., Yale University Press, New Haven; London. Boyer, P., Roberts, N., Baird, D., 2006. Holocene environment and settlement on the Çarşamba alluvial fan, south-central Turkey: Integrating geoarchaeology and archaeological field survey. Geoarchaeology 21, 675–698. https://doi.org/10.1002/ gea.20133 Bradtmöller, M., Grimm, S., Riel-Salvatore, J., 2017. Resilience theory in archaeological practice – An annotated review. Quaternary International, Adaptive Cycles in Archaeology 446, 3–16. https://doi.org/10.1016/j.quaint.2016.10.002 Braekmans, D., Degryse, P., Neyt, B., Waelkens, M., Poblome, J., 2017. Reconstructing regional trajectories: The provenance and distribution of Archaic to Hellenistic ceramics in Central Pisidia (South-west Turkey). Archaeometry 59, 472–492. https://doi.org/10.1111/arcm.12259 Bringmann, K., 1993. The King as Benefactor. Some remarks on ideal kingship in the age of Hellenism, in: Bulloch, A.W. (Ed.), Images and Ideologies: Self-Definition in the Hellenistic World. University of California Press, Berkeley, pp. 7–24. Bringmann, K., Steuben, H. von, 1995. Schenkungen Hellenistischer Herrscher an griechische Städte und Heiligtümer: Zeugnisse und Kommentare. Akademie Verlag. Bronowski, J., 1973. The Ascent of Man. BBC Books, London. Broothaerts, N., Van Loo, M., Vandam, R., Cleymans, S., Daems, D., Poblome, J., Verstraeten, G., 2019. Human-environment interactions during the past 4000 years in Sagalassos and South-Western Turkey. Presented at the INQUA 2019, 25-31 July 2019, Dublin. Brughmans, T., 2010. Connecting the dots: Towards archaeological network analysis. Oxford Journal of Archaeology 29, 277–303. https://doi.org/10.1111/j. 1468-0092.2010.00349.x

Bibliography 217 Brughmans, T., 2011. Facebooking the past: A critical social network analysis approach for archaeology, in: Chrysanthi, A., Flores, P., Papadopoulos, C. (Eds.), Thinking Beyond the Tool: Archaeological Computing and the Interpretative Process. Archaeopress, Oxford. Brughmans, T., 2013. Thinking through networks: A review of formal network methods in archaeology. Journal of Archaeological Method and Theory 20, 623– 662. https://doi.org/10.1007/s10816-012-9133-8 Brughmans, T., Collar, A., Coward, F., 2016. The Connected Past: Challenges to Network Studies in Archaeology and History. Oxford University Press. Brughmans, T., Peeples, M., 2017. Trends in archaeological network research: A bibliometric analysis. 1 1, 1–24. https://doi.org/10.25517/jhnr.v1i1.10 Brughmans, T., Peeples, M.A., 2018. Network science, in: The Encyclopedia of Archaeological Sciences. American Cancer Society, pp. 1–4. https://doi. org/10.1002/9781119188230.saseas0402 Brughmans, T., Poblome, J., 2012. Pots in space: Understanding Roman pottery distribution from confronting exploratory and geographical network analyses, in: Barker, E., Bouzarovski, S., Pelling, C., Isaksen, L. (Eds.), New Worlds Out of Old Texts: Developing Techniques for the Spatial Analysis of Ancient Narratives. Oxford University Press, Oxford. Bryant, C.G.A., Jary, D., 1997. Anthony Giddens: Critical Assessments, Christopher G. A. Bryant and David Jary (Eds.). London: Routledge. Bryce, T., 2011. The Late Bronze Age in the West and the Aegean, in: Steadman, S.R., McMahon, G. (Eds.), The Oxford Handbook of Ancient Anatolia: (10,000–323 BCE). OUP USA, pp. 363–375. Bryce, T.R., Zahle, J., 1986. The Lycians: A Study of Lycian History and Civilisation to the Conquest of Alexander the Great. 1: The Lycians in Literary and Epigraphic Sources. Museum Tusculanum Press, Copenhagen. Buckley, W.F., 1967. Sociology and Modern Systems Theory. Prentice-Hall. Bunge, M.A., 1999. The Sociology-Philosophy Connection, Science and Technology Studies. Transaction, New Brunswick. Burke, B., 2005. Textile production at Gordion and the Phrygian economy, in: Kealhofer, L. (Ed.), The Archaeology of Midas and the Phrygians: Recent Work at Gordion. University of Pennsylvania Press, Philadelphia, pp. 67–100. Büyükkolancı, M., 2000. Excavations on Ayasuluk Hill in Selçuk/Turkey. A Contribution to the Early History of Ephesus, in: Die Ägäis Und Das Westliche Mittelmeer. Beziehungen Und Wechselwirkungen 8. Bis 5. Jh. v. Chr. Akten Des Symposions, Vienna, March 24–27, 1999, Archäologischen Forschungen. Österreichisches Archäologische Akademie der Wissenschaften, Vienna, pp. 39–43. Byrne, D., 1998. Complexity Theory and the Social Sciences: An Introduction. Routledge, London; New York. Cabrera, D., Colosi, L., Lobdell, C., 2008. Systems thinking. Evaluation and Program Planning 31, 299–310. https://doi.org/10.1016/j.evalprogplan.2007.12.001 Carballo, D., Roscoe, P., Feinman, G., 2014. Cooperation and collective action in the cultural evolution of complex societies. Journal of Archaeological Method and Theory 21, 98–133. https://doi.org/10.1007/s10816-012-9147-2 Carneiro, R.L., 1967. On the relationship between size of population and complexity of social organization. Journal of Anthropological Research 42, 355.364. Carneiro, R.L., 1970. A theory of the origin of the state. Science 169, 733–738. https://doi.org/10.1126/science.169.3947.733

218  Bibliography Carpenter, S., Brock, W., 2008. Adaptive capacity and traps. Ecology and Society 13. https://doi.org/10.5751/ES-02716-130240 Carter, B., New, C., 2004. Introduction: Realist social theory and empirical research, in: Carter, B., New, C. (Eds.), Making Realism Work: Realist Social Theory and Empirical Research. Routledge, London, pp. 1–20. Castellani, B., Hafferty, F.W., 2009. Sociology and Complexity Science: A New Field of Inquiry. Springer Berlin Heidelberg, Berlin, Heidelberg. Cavalli-Sforza, L.L., Feldman, M.W., 1981. Cultural Transmission and Evolution, 1st ed., Princeton University Press, Princeton, NJ. Cegielski, W.H., Rogers, J.D., 2016. Rethinking the role of agent-based modeling in archaeology. Journal of Anthropological Archaeology 41, 283–298. https://doi. org/10.1016/j.jaa.2016.01.009 Cesaretti, R., Lobo, J., Bettencourt, L.M.A., Ortman, S.G., Smith, M.E., 2016. Population-area relationship for medieval European cities. PLOS ONE 11, e0162678. https://doi.org/10.1371/journal.pone.0162678 Çevik, Ö, 2007. The emergence of different social systems in Early Bronze Age Anatolia: Urbanisation versus Centralisation. Anatolian Studies 57, 131–140. Chacon, R., Mendoza, R., 2017. Introduction, in: Chacon, R., Mendoza, R. (Eds.), Feast, Famine or Fighting?: Multiple Pathways to Social Complexity, Studies in Human Ecology and Adaptation, 8. Springer International Publishing: Imprint: Springer, Cham, pp. 1–14. Chaniotis, A., 2005. War in the Hellenistic World: A Social and Cultural History. Oxford Blackwell. Chapman, R., 2007. Evolution, Complexity and the State, in: Kohring, S., WynneJones, S. (Eds.), Socialising Complexity: Structure, Interaction and Power in Archaeological Discourse. Oxbow Books, Oxford, pp. 13–28. Chattoe-Brown, E., 2013. Why sociology should use agent based modelling. Sociological Research Online 18, 1–11. https://doi.org/10.5153/sro.3055 Cherry, C., 1957. On Human Communication. Wiley, London. Chick, G., 1997. Cultural complexity: The concept and its measurement. CrossCultural Research 31, 275–307. https://doi.org/10.1177/106939719703100401 Childe, G., 1950. The Urban Revolution, Town Planning Review. Chliaoutakis, A., Chalkiadakis, G., 2016. Agent-based modeling of ancient societies and their organization structure. Autonomous Agents and Multi-Agent Systems 30, 1072–1116. https://doi.org/10.1007/s10458-016-9325-9 Chrubasik, B., King, D. (Eds.), 2017. Hellenism and the Local Communities of the Eastern Mediterranean: 400 BCE-250 CE. Oxford University Press, Oxford, New York. Chudek, M., Henrich, J., 2011. Culture–gene coevolution, norm-psychology and the emergence of human prosociality. Trends in Cognitive Sciences 15, 218–226. https://doi.org/10.1016/j.tics.2011.03.003 Cilliers, P., 2001. Boundaries, hierarchies and networks in complex systems. International Journal of Innovation and Technology Management 05, 135–147. https://doi.org/10.1142/S1363919601000312 Cioffi-Revilla, C., 2005. A Canonical Theory of Origins and Development of Social Complexity. The Journal of Mathematical Sociology 29, 133–153. https://doi. org/10.1080/00222500590920860 Cioffi-Revilla, C., 2014. Introduction to Computational Social Science. Springer, Heidelberg.

Bibliography 219 Clark, A., 2016. Surfing Uncertainty: Prediction, Action, and the Embodied Mind. Oxford University Press, Oxford, New York. Clark, C., Haswell, M.R., 1964. The Economics of Subsistence Agriculture. Macmillan, Palgrave. Clarke, D.L., 1968. Analytical Archaeology. Methuen, London. Clauset, A., Shalizi, C., Newman, M., 2009. Power-law distributions in empirical data. SIAM Review 51, 661–703. https://doi.org/10.1137/070710111 Cleymans, S., 2018. The Shotgun Method 2.0: Estimating population numbers for second century A.D. Sagalassos. Ancient Society: Journal of Ancient History of the Greek, Hellenistic and Roman Worlds 263–304. https://doi.org/10.2143/ AS.48.0.3285204 Cleymans, S., Daems, D., Broothaerts, N., In Preparation. Sustaining people. Reassessing carrying capacity through the socio-ecological metabolism of the ancient community at Düzen Tepe, SW Turkey. Cohen, G.M., 1978. The Seleucid Colonies: Studies in Founding, Administration and Organization. Steiner, Wiesbaden. Cohen, G.M., 1995. Hellenistic Settlements in Europe, the Islands, and Asia Minor. University of California Press, Berkeley. Collar, A., Coward, F., Brughmans, T., Mills, B.J., 2015. Networks in archaeology: Phenomena, abstraction, representation. Journal of Archaeological Method and Theory 22, 1–32. https://doi.org/10.1007/s10816-014-9235-6 Collard, M., Buchanan, B., O’Brien, M.J., 2013. population size as an explanation for patterns in the paleolithic archaeological record: More caution is needed. Current Anthropology 54, S388–S396. https://doi.org/10.1086/673881 Collard, M., Kemery, M., Banks, S., 2005. Causes of toolkit variation among hunter-gatherers: A test of four competing hypotheses. Canadian Journal of Archaeology/Journal Canadien d’Archéologie 29, 1–19. Collins, R., 1995. Prediction in macrosociology: The case of the Soviet collapse. American Journal of Sociology 100, 1552–1593. https://doi.org/10.1086/230672 Collins-Elliott, S.A., 2016. Agglomerative clustering using cosine and Jaccard distances: A computational approach to Roman vessel taxonomy. Archeologia e Calcolatori 27, 7–25. Cooper, J., 2012. Fail to prepare, then prepare to fail: Rethinking threat, vulnerability, and mitigation in the Precolumbian Caribbean, in: Cooper, J., Sheets, P. (Eds.), Surviving Sudden Environmental Change. Understanding Hazards, Mitigating Impacts, Avoiding Disasters. University Press of Colorado, Boulder, pp. 91–116. Cote, M., Nightingale, A.J., 2012. Resilience thinking meets social theory: Situating social change in socio-ecological systems (SES) research. Progress in Human Geography 36, 475–489. https://doi.org/10.1177/0309132511425708 Coto-Sarmiento, M., Rubio-Campillo, X., Remesal, J., 2018. Identifying social learning between Roman amphorae workshops through morphometric similarity. Journal of Archaeological Science 96, 117–123. https://doi.org/10.1016/j. jas.2018.06.002 Coulton, J.J., 2012. The Balboura Survey and Settlement in Highland Southwest Anatolia. London British institute at Ankara. Coveney, P.V., Dougherty, E.R., Highfield, R.R., 2016. Big data need big theory too. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 374, 20160153. https://doi.org/10.1098/rsta.2016.0153

220  Bibliography Coward, F., 2010. Small worlds, material culture and ancient near Eastern social networks, in: Social Brain, Distributed Mind. pp. 453–484. Cowgill, G., 2004. Origins and development of urbanism: Archaeological perspectives. Annual Review of Anthropology 33, 525.549. Crabtree, S.A., Bocinsky, R.K., Hooper, P.L., Ryan, S.C., Kohler, T.A., 2017. How to make a polity (In the Central Mesa Verde Region). American Antiquity 82, 71–95. https://doi.org/10.1017/aaq.2016.18 Crabtree, S.A., Harris, K., Davies, B., Romanowska, I., 2019. Outreach in archaeology with agent-based modeling: Part 3 of 3. Advances in Archaeological Practice 7, 194–202. https://doi.org/10.1017/aap.2019.4 Creanza, N., Kolodny, O., Feldman, M.W., 2017. Cultural evolutionary theory: How culture evolves and why it matters. PNAS 114, 7782–7789. https://doi.org/10.1073/ pnas.1620732114 Creekmore, A.I., Fisher, K.D. (Eds.), 2014. Making Ancient Cities: Space and Place in Early Urban Societies. Cambridge University Press, New York, NY. Crumley, C.L., 1995. Heterarchy and the analysis of complex societies. Archeological Papers of the American Anthropological Association 6, 1–5. https://doi. org/10.1525/ap3a.1995.6.1.1 Currie, A., 2015. Philosophy of science and the curse of the case study, in: Daly, C. (Ed.), The Palgrave Handbook of Philosophical Methods. Palgrave Macmillan UK, London, pp. 553–572. https://doi.org/10.1057/9781137344557_22 Currie, A., 2018. Rock, Bone, and Ruin: An Optimist’s Guide to the Historical Sciences. MIT Press. Currie, T., Turchin, P., Bednar, J., Richerson, P., Schwesinger, G., Steinmo, S., Wacziarg, R., Wallis, J., 2016. Evolution of institutions and organizations, in: Wilson, D.S., Kirman, A. (Eds.), Complexity and Evolution: Toward a New Synthesis for Economics. The MIT Press, Cambridge, pp. 201–238. https://doi. org/10.7551/mitpress/9780262035385.001.0001 Currie, T.E., Mace, R., 2011. Mode and tempo in the evolution of socio-political organization: Reconciling ‘Darwinian’ and ‘Spencerian’ evolutionary approaches in anthropology. Philos Trans R Soc Lond B Biol Sci 366, 1108–1117. https://doi. org/10.1098/rstb.2010.0318 Daems, D., 2019. Building communities. Presenting a model of community formation and organizational complexity in southwestern Anatolia. Journal of Anthropological Archaeology 56, 101110. https://doi.org/10.1016/j.jaa.2019.101110 Daems, D., 2020. Social complexity and complexity economics. Studying socio-­ economic systems at Düzen Tepe and Sagalassos (SW Turkey), in: Verboven, K. (Ed.), Complexity: A New Framework to Interpret Ancient Economy Data. Palgrave Studies in Ancient Economies, pp. 163–201. Daems, D., Accepted(a). Adaptive cycles in archaeology. Indicators of change and stability in socio-economic systems, in: Vandam, R., Poblome, J. (Eds.), Once upon a River. Long-Term Change in the Social Metabolism of the Ağlasun Valley and its People. Daems, D., Accepted(b). Between cities and kings. Polis formation in southwest Anatolia during the Hellenistic period, in: Beyond Hybridity and CodeSwitching: Material Approaches to Cultural Concurrences in the Late Hellenistic Mediterranean. Daems, D., Braekmans, D., Poblome, J., 2017. Late Achaemenid and Early Hellenistic Pisidian Material Culture from Düzen Tepe (SW Anatolia). Herom 6, 11–47.

Bibliography 221 Daems, D., Poblome, J., 2017. The Pottery of Late Achaemenid Sagalassos: An overview. HEROM. Journal on Hellenistic and Roman Material Culture 6, 49–62. Daems, D., Poblome, J., 2016. Adaptive Cycles in Communities and Landscapes: The Case of Sagalassos and Düzen Tepe During the Classical/Hellenistic Period. Archaeological Review from Cambridge 31, 91.107. Daems, D., Talloen, P., Accepted. Moving in together? Synoikismos and modes of community development through push/pull dynamics in SW Anatolia. BASOR. Daems, D., Vandam, R., Cleymans, S., Broothaerts, N., Boogers, S., Matsuo, H., Mirhanoglu, A., In Press. The social metabolism of past societies. A new approach to environmental changes and societal responses in the territory of Sagalassos (SW Turkey), in: Erdkamp, P., Manning, J., Verboven, K. (Eds.), Climate and Society in Ancient Worlds. Diversity in Collapse and Resilience. Palgrave Macmillan. Daems, D., Van Der Enden, M., Poblome, J., Talloen, P., 2019. The Hellenistic Ceramics of Sagalassos. Presented at the International Association for Research on Pottery of the Hellenistic Period, in IARPOTHP conference proceedings Lyon, November 2015, Lyon. Davidson, D.J., 2010. The Applicability of the Concept of Resilience to Social Systems: Some Sources of Optimism and Nagging Doubts. Society & Natural Resources 23, 1135–1149. https://doi.org/10.1080/08941921003652940 Davies, B., Romanowska, I., Harris, K., Crabtree, S.A., 2019. Combining geographic information systems and agent-based models in archaeology: Part 2 of 3. Advances in Archaeological Practice 7, 185–193. https://doi.org/10.1017/aap.2019.5 Davies, J., 2002. The interpenetration of Hellenistic Sovereignties, in: Ogden, D. (Ed.), The Hellenistic World: New Perspectives. Classical Press of Wales and Duckworth, London, pp. 1–21. de Callataý, F., 2019. Did the Seleucids found new cities to promote coinage?, in: Oetjen, R. (Ed.), New Perspectives in Seleucid History, Archaeology and Numismatics: Studies in Honor of Getzel M. Cohen. De Gruyter, Berlin, pp. 561–573. de Molina, M.G., Toledo, V.M., 2014. The Social Metabolism: A Socio-Ecological Theory of Historical Change, Environmental History. Springer, Dordrecht, Dordrecht. de Ruiter, J., Weston, G., Lyon, S.M., 2011. Dunbar’s number: Group size and brain physiology in humans reexamined. Am Anthropol 113, 557.568. Dean, J.S., Gumerman, G.J., Epstein, J.M., Axtell, R.L., Swedlund, A.C., Parker, M.T., McCarroll, S., 2000. Understanding Anasazi culture change through agentbased modeling, in: Kohler, T.A., Gummerman, G.J. (Eds.), Dynamics in Human and Primate Societies. Oxford University Press, Inc., New York, NY, USA, pp. 179–205. Demarrais, E., Castillo, L.J., Earle, T., 1996. Ideology, Materialization, and Power Strategies. Current Anthropology 37, 15–31. https://doi.org/10.1086/204472 Dercksen, J.G., 1996. The Old Assyrian Copper Trade in Anatolia. Peeters Publishers & Booksellers. Derex, M., Beugin, M.-P., Godelle, B., Raymond, M., 2013. Experimental evidence for the influence of group size on cultural complexity. Nature 503, 389–391. https:// doi.org/10.1038/nature12774 DiMaggio, P.J., Powell, W.W., 1983. The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review 48, 147–160. https://doi.org/10.2307/2095101

222  Bibliography Drennan, R.D., Peterson, C.E., 2004. Comparing archaeological settlement systems with rank-size graphs: A measure of shape and statistical confidence. Journal of Archaeological Science 31, 533–549. https://doi.org/10.1016/j.jas.2003.10.002 Drennan, R.D., Peterson, C.E., 2008. Centralized communities, population, and social complexity after sedentarization, in: Bocquet-Appel, J.-P., Bar-Yosef, O. (Eds.), The Neolithic Demographic Transition and Its Consequences. Springer Netherlands, Dordrecht, pp. 359–386. https://doi.org/10.1007/978-14020-8539-0_14 Dunbar, R., Gamble, C., Gowlett, J., 2014. Thinking Big: How the Evolution of Social Life Shaped the Human Mind. Thames & Hudson, London. Dunbar, R.I.M., 1993. Coevolution of neocortical size, group size and language in humans. Behavioral and Brain Sciences 16, 681–694. https://doi.org/10.1017/ S0140525X00032325 Dunning, N.P., Beach, T.P., Luzzadder-Beach, S., 2012. Kax and kol: Collapse and resilience in lowland Maya civilization. PNAS 109, 3652–3657. https://doi. org/10.1073/pnas.1114838109 Düring, B., 2011. The Prehistory of Asia Minor. From Complex hunter-gatherers to Early Urban societies. Cambridge University Press, Cambridge. Düring, B., 2013. The anatomy of a prehistoric community: Reconsidering Çatal Höyük, in: Birch, J. (Ed.), From Prehistoric Villages to Cities: Settlement Aggregation and Community Transformation. Routledge, Taylor & Francis Group, pp. 23–43. Duru, R., 1996. Kuruçay Höyük II: Results of the Excavations 1978–1988: The Late Chalcolithic and Early Bronze Age Settlements. Türk Tarih Kürümü Basimevi, Ankara. Duru, R., 2008. From 8000 BC to 2000 BC: Six Thousand Years of the Burdur Antalya Region. Ege Yayinlari, Antalya Turkey. Dusinberre, E.R.M., 2013. Empire, Authority, and Autonomy in Achaemenid Anatolia. Cambridge University Press, New York. Eastwood, W.J., Roberts, N., Lamb, H.F., 1998. Palaeoecological and archaeological evidence for human occupance in southwest Turkey: The Beyşehir occupation phase. Anatolian Studies 48, 69–86. https://doi.org/10.2307/3643048 Efatmaneshnik, M., Ryan, M.J., 2016. A general framework for measuring system complexity. Complexity 21, 533–546. https://doi.org/10.1002/cplx.21767 Epstein, J.M., 2008. Why model? Journal of Artificial Societies and Social Simulation 11. Epstein, J.M., Axtell, R., 1996. Growing Artificial Societies: Social Science from the Bottom Up. Brookings Institution Press. Eres, Z., 2003. Die Huttenlehmreste von Aşağı Pınar, in: Karul, N., Eres, Z., Özdoğan, M., Parzinger, H. (Eds.), Aşağı Pınar I: Einführung, Forschungsgeschichte, Stratigraphie Und Archiktektur. Philipp von Zabern Verlag, GmbH, pp. 126–142. Erickson, E.E., 1977. Factors and patterns: The nature of sociocultural variation. Behavior Science Research 12, 227–250. https://doi.org/10.1177/106939717701200403 Erkanal, H., 1999. Liman Tepe. Kazı Sonuçları Toplantısı 20, 325–336. Eslick, C., 1988. Hacılar to Karataş: Social organization in South-western Anatolia. Mediterranean Archaeology 1, 10–40. Evans, T.S., Rivers, R.J., Knappett, C., 2012. Interactions in space for archaeological models. Advances in Complex Systems 15, 1150009. https://doi.org/10.1142/ S021952591100327X

Bibliography 223 Faulseit, R.K. (Ed.), 2016. Beyond Collapse: Archaeological Perspectives on Resilience, Revitalization, and Transformation in Complex Societies. Southern Illinois University Press, Carbondale, Illinois. Feinman, G.M., 2011. Size, complexity, and organizational variation: A comparative approach. Cross-Cultural Research 45, 37–58. https://doi.org/10.1177/ 1069397110383658 Feinman, G.M., 2012. The emergence of social complexity: Why more than population size matters. Cooperation and Collective Action: Archaeological Perspective 35–56. Feinman, G.M., 2017. Multiple pathways to large-scale human cooperative networks: A reframing, in: Chacon, R., Mendoza, R. (Eds.), Feast, Famine or Fighting?: Multiple Pathways to Social Complexity, Studies in Human Ecology and Adaptation, 8. Springer International Publishing: Imprint: Springer, Cham, pp. 459–478. Feld, S., Basso, K.H., 1996. Senses of Place. School of American Research Press. Feldman, M.W., Laland, K.N., 1996. Gene-culture coevolutionary theory. Trends in Ecology & Evolution 11, 453–457. https://doi.org/10.1016/0169-5347 (96)10052-5 Filatova, T., Polhill, J.G., van Ewijk, S., 2016. Regime shifts in coupled socioenvironmental systems: Review of modelling challenges and approaches. Environmental Modelling & Software 75, 333–347. https://doi.org/10.1016/j.envsoft. 2015.04.003 Fischer‐Kowalski, M., Haberl, H., 1997. Tons, joules, and money: Modes of production and their sustainability problems. Society & Natural Resources 10, 61–85. https://doi.org/10.1080/08941929709381009 Fiske, J., 1873. The progress from brute to man. The North American Review 117, 251.319. Flack, J., 2017a. Life’s information hierarchy, in: Walker, S.I., Davies, P.C.W., Ellis, G.F.R.E. (Eds.), From Matter to Life: Information and Causality. Cambridge University Press, pp. 283–302. Flack, J., 2017b. Coarse-graining as a downward causation mechanism. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 375, 20160338. https://doi.org/10.1098/rsta.2016.0338 Flack, J., 2019. Life’s information hierarchy, in: Krakauer, D.C. (Ed.), Worlds Hidden in Plain Sight. SFI Press, pp. 201–225. Flack, J., Erwin, D., Krakauer, D., 2013. Timescales, symmetry, and uncertainty reduction in the origins of hierarchy in biological systems, in: Sterelny, K., Joyce, B., Calcott, B., Fraser, B. (Eds.), Cooperation and Its Evolution. MIT Press, Cambridge, pp. 45–74. Flannery, K.V., 1968. Archaeological systems theory and early Mesoamerica, in: Meggers, B.J., Anthropological Society of Washington (Washington, D.C.) (Eds.), Anthropological Archeology in the Americas. The Anthropological Society of Washington, Washington, DC, pp. 67–87. Flannery, K.V., 1972. The cultural evolution of civilizations. Annual Review of Ecology and Systematics 3, 399–426. Fletcher, R., 1995. The Limits of Settlement Growth: A Theoretical Outline. Cambridge University Press, Cambridge. Fletcher, R., 2004. Materiality, space, time, and outcome, in: Bintliff, J. (Ed.), A Companion to Archaeology. Blackwell Publishing, Oxford, pp. 110–140.

224  Bibliography Forge, A., 1972. Normative factors in the settlement size of Neolithic cultivators (New Guinea), in: An, Settlement and Urbanism: Proceedings of a Meeting of the Research Seminar in Archaeology and Related Subjects Held at the Institute of Archaeology, London University. Duckworth Pub, London, pp. 363–376. Fox, R.G., 1977. Urban Anthropology: Cities in Their Cultural Settings, PrenticeHall Series in Anthropology. Prentice Hall, Englewood Cliffs. Frangipane, M., 2007. Different types of egalitarian societies and the development of inequality in early Mesopotamia. World Archaeology 39, 151–176. Freeberg, T.M., Dunbar, R.I.M., Ord, T.J., 2012. Social complexity as a proximate and ultimate factor in communicative complexity. Philosophical Transactions of the Royal Society B: Biological Sciences 367, 1785–1801. https://doi.org/10.1098/ rstb.2011.0213 Freeman, J., Hard, R.J., Mauldin, R.P., 2017. A theory of regime change on the Texas Coastal Plain. Quaternary International, Adaptive Cycles in Archaeology 446, 83–94. https://doi.org/10.1016/j.quaint.2016.12.029 French, D., 1998. Canhasan Sites 1: Canhasan I: Stratigraphy and Structures. British Institute of Archaeology at Ankara. Fried, M.H., 1967. The Evolution of Political Society: An Essay in Political Anthropology, 71st edition. ed. McGraw-Hill Humanities/Social Sciences/Langua, New York. Friston, K., 2010. The free-energy principle: A unified brain theory? Nature Reviews Neuroscience 11, 127–138. https://doi.org/10.1038/nrn2787 Froese, T., Gershenson, C., Manzanilla, L.R., 2014. Can government be selforganized? A mathematical model of the collective social organization of Ancient Teotihuacan, Central Mexico. PLoS ONE 9, e109966. https://doi.org/10.1371/journal. pone.0109966 Galik, A., 2014. Late Chalcolithic Subsistence Strategies on the Basis of Two Examples: The Çukuriçi Höyük in Western Anatolia and the Barcin Höyük in Northwestern Anatolia, in: Horejs, B., Mehofer, M. (Eds.), Western Anatolia before Troy. Proto-Urbanisation in the 4th Millenium BC?: Proceedings of the International Symposium Held at the Kunsthistorisces Museum Wien Vienna, Austria, 21-24 November, 2012, OREA. Verlag der Österreichischen Akademie der Wissenschaften, pp. 385–394. Garvey, R., 2018. Current and potential roles of archaeology in the development of cultural evolutionary theory. Philosophical Transactions of the Royal Society B: Biological Sciences 373. https://doi.org/10.1098/rstb.2017.0057 Gattiglia, G., 2015. Think big about data: Archaeology and the big data challenge. Archäologische Informationen 38, 113–124. https://doi.org/10.11588/ai.2015.1.26155 Gell-Mann, M., 1994. The Quark and the Jaguar: Adventures in the Simple and the Complex. St. Martin’s Griffin, New York. Giampietro, M., 2003. Multi-Scale Integrated Analysis of Agroecosystems. CRC Press. Giampietro, M., Mayumi, K., Sorman, A.H., 2011. The Metabolic Pattern of Societies: Where Economists Fall Short, 1 ed. Routledge, London; New York. Giddens, A., 1984. The Constitution of Society: Outline of the Theory of Structuration. University of California Press, Berkeley. Gintis, H., 2011. Gene–culture coevolution and the nature of human sociality. Philosophical Transactions of the Royal Society B: Biological Sciences 366, 878–888. https://doi.org/10.1098/rstb.2010.0310

Bibliography 225 Girvan, M., Newman, M.E.J., 2002. Community structure in social and biological networks. PNAS 99, 7821–7826. https://doi.org/10.1073/pnas.122653799 Gladwell, M., 2000. The Tipping Point: How Little Things Can Make a Big Difference. Abacus, London. Glaeser, E.L., Sacerdote, B., 1999. Why is there more crime in cities? Journal of Political Economy 107, S225–S258. https://doi.org/10.1086/250109 Glatz, C., 2011. The Hittite State and Empire from Archaeological Evidence, in: Steadman, S.R., McMahon, G. (Eds.), The Oxford Handbook of Ancient Anatolia: (10,000-323 BCE). OUP USA, pp. 877–899. Glatz, C., Plourde, A.M., 2011. Landscape monuments and political competition in Late Bronze Age Anatolia: An investigation of costly signaling theory. Bulletin of the American Schools of Oriental Research 361, 33–66. https://doi.org/10.5615/ bullamerschoorie.361.0033 Goedegebuure, P., Hout, T. van den, Osborne, J., Massa, M., Bachhuber, C., Şahin, F., 2020. TÜRKMEN-KARAHÖYÜK 1: A new Hieroglyphic Luwian inscription from Great King Hartapu, son of Mursili, conqueror of Phrygia. Anatolian Studies 70, 29–43. https://doi.org/10.1017/S0066154620000022 Goffman, E., 1972. Encounters: Two Studies in the Sociology of Interaction, Penguin University Books. Penguin books, Harmondsworth. Göktürk, O.M., 2011. Climate in the Eastern Mediterranean Through the Holocene Inferred from Turkish Stalagmites. Institut für Geologie und Oeschger Centre for Climate Change Research Universität Bern. Gotts, N., 2007. Resilience, panarchy and world-systems analysis. Ecology and Society 12, 24. Gould, S.J., 1989. Wonderful Life: The Burgess Shale and the Nature of History. W.W. Norton. Gould, S.J., 1999. Introduction: The scales of contingency and punctuation, in: Bintliff, J. (Ed.), Structure and Contingency: Evolutionary Processes in Life and Human Society. Leicester University Press, London, pp. ix–xxii. Gould, S.J., 2007. Punctuated Equilibrium, 1st ed. Belknap Press, Cambridge, Mass. Graham, S., 2006a. Networks, agent-based models and the Antonine Itineraries: Implications for Roman archaeology. Journal of Mediterranean Archaeology 19, 45–64. https://doi.org/10.1558/jmea.2006.19.1.45 Graham, S., 2006b. EX FIGLINIS The Network Dynamics of the Tiber Valley Brick Industry in the Hinterland of Rome., BAR International Series. John and Erica Hedges Ltd. Graham, S., 2008. The space between: The geography of social networks in the Tiber Valley, in: Coarelli, F. and Patterson, H. (Eds.), Mercator Placidissimus: The Tiber Valley in Antiquity New Research in the Upper and Middle River Valley, Edizione Qvaser, Rome, pp. 671–686. Graham, S., Milligan, I., Weingart, S.B., 2016. Exploring Big Historical Data: The Historian’s Macroscope. Imperial College Press, London. Grainger, J.D., 2009. The Cities of Pamphylia. Oxbow, Oxford. Grave, P., Kealhofer, L., Marsh, B., Schoop, U.-D., Seeher, J., Bennett, J.W., Stopic, A., 2014. Ceramics, trade, provenience and geology: Cyprus in the Late Bronze Age. Antiquity 88, 1180–1200. https://doi.org/10.1017/S0003598X0011539X Gray, D., Macready, M., 2019. Metaphors: Ladders of innovation, in: Krakauer, D.C. (Ed.), Worlds Hidden in Plain Sight. SFI Press, pp. 129–145.

226  Bibliography Green, P., 1990. Alexander to Actium: The Historical Evolution of the Hellenistic Age. University of California Press, Berkeley. Griffin, A.F., 2011. Emergence of fusion/fission cycling and self-organized criticality from a simulation model of early complex polities. Journal of Archaeological Science 38, 873–883. https://doi.org/10.1016/j.jas.2010.11.017 Griffin, A.F., Stanish, C., 2007. An Agent-based model of prehistoric settlement patterns and political consolidation in the Lake Titicaca Basin of Peru and Bolivia. Structure and Dynamics 2. Gringmuth-Dallmer, E., 1996. Kulturlandschaftsmuster und Siedlungssysteme. Siedlungsforschung 14, 7–31. Gronenborn, D., Strien, H.-C., Dietrich, S., Sirocko, F., 2014. ‘Adaptive cycles’ and climate fluctuations: A case study from Linear Pottery Culture in western Central Europe. Journal of Archaeological Science 51, 73–83. https://doi.org/10.1016/j. jas.2013.03.015 Gronenborn, D., Strien, H.-C., Lemmen, C., 2017. Population dynamics, social resilience strategies, and Adaptive Cycles in early farming societies of SW Central Europe. Quaternary International, Adaptive Cycles in Archaeology 446, 54–65. https://doi.org/10.1016/j.quaint.2017.01.018 Gumerman, G.J., Gell-mann, M., 1994. Understanding Complexity in the Prehistoric Southwest. Avalon Publishing. Gunderson, L.H., Holling, C.S., 2002. Panarchy: Understanding Transformations in Human and Natural Systems. Island Press, Washington, DC. Gyucha, A., 2019. Population aggregation and early urbanization, in: Gyucha, A. (Ed.), Coming Together: Comparative Approaches to Population Aggregation and Early Urbanization. State University of New York Press, New York, pp. 1–35. Hamilton, W.D., 1964. The genetical evolution of social behaviour. I. Journal of Theoretical Biology 7, 1–16. https://doi.org/10.1016/0022-5193(64)90038-4 Hansen, M.H. (ed.), 2000. A Comparative Study of Thirty City-State Cultures : An Investigation Conducted by the Copenhagen Polis Centre. Reitzel, Copenhagen. Hansen, M.H., 1998. Polis and City-State: An Ancient Concept and its Modern Equivalent. Munksgaard, Copenhagen. Hansen, M.H., 2002. Introduction, in: Hansen, M.H. (Ed.), A Comparative Study of Six City-State Cultures. Reitzel, Copenhagen, pp. 7–22. Hansen, M.H., Nielsen, T.H., 2004. An Inventory of Archaic and Classical Poleis. Oxford university press, Oxford. Harvey, D., 1969. Explanation in Geography. Edward Arnold. Harvey, D., 1973. Social Justice and the City. Johns Hopkins University Press. Hawkins, J.D., 1998. Tarkasnawa King of mira “Tarkondemos”, Boǧazköy Sealings and Karabel. Anatolian Studies 48, 1–31. https://doi.org/10.2307/3643046 Hawley, A., 1968. Human Ecology. International Encyclopedia of the Social Sciences. Haynes, P., 2017. Social Synthesis: Finding Dynamic Patterns in Complex Social Systems. Routledge. Hegmon, M., Peeples, M.A., Kinzig, A.P., Kulow, S., Meegan, C.M., Nelson, M.C., 2008. Social transformation and its human costs in the Prehispanic U.S. Southwest. American Anthropologist 110, 313–324. https://doi.org/10.1111/j.1548-1433.2008. 00041.x Henrich, J., 2004. Demography and cultural evolution: How adaptive cultural processes can produce maladaptive losses: The Tasmanian Case. American Antiquity 69, 197–214. https://doi.org/10.2307/4128416

Bibliography 227 Hidalgo, C., 2015. Why Information Grows: The Evolution of Order, from Atoms to Economies. Basic Books, New York. Hobson, E.A., Ferdinand, V., Kolchinsky, A., Garland, J., 2019. Rethinking animal social complexity measures with the help of complex systems concepts. Animal Behaviour. https://doi.org/10.1016/j.anbehav.2019.05.016 Holland, J., 1995. Hidden Order: How Adaptation Builds Complexity. Basic Books, Reading, Mass. Holland, J.H., 1998. Emergence: From Chaos to Order. Oxford University Press. Holland, J.H., 2014. Complexity: A Very Short Introduction, 1st ed., Oxford University Press. Holling, C.S., 1973. Resilience and stability of ecological systems. Annual Review of Ecology and Systematics 4, 1.23. Holling, C.S., Gunderson, L., 2002. Resilience and adaptive cycles, in: Gunderson, L.H., Holling, C.S. (Eds.), Panarchy: Understanding Transformations in Human and Natural Systems. Island Press, Washington, DC, pp. 25–62. Holling, C.S., Gunderson, L., Ludwig, D., 2002a. In quest of a theory of adaptive change, in: Gunderson, L.H., Holling, C.S. (Eds.), Panarchy: Understanding Transformations in Human and Natural Systems. Island Press, Washington, DC, pp. 3–22. Holling, C.S., Gunderson, L., Peterson, G., 2002b. Sustainability and Panarchies, in: Gunderson, L.H., Holling, C.S. (Eds.), Panarchy: Understanding Transformations in Human and Natural Systems. Island Press, Washington, DC, pp. 63–102. Horejs, B., 2014. Proto-Urbanisation without Urban Centres? A Model of Transformation for the Izmir Region in the 4th Millennium BC, in: Horejs, B., Mehofer, M. (Eds.), Western Anatolia before Troy. Proto-Urbanisation in the 4th Millenium BC?: Proceedings of the International Symposium Held at the Kunsthistorisces Museum Wien Vienna, Austria, 21–24 November, 2012, OREA. Verlag der Österreichischen Akademie der Wissenschaften, pp. 15–42. Horejs, B., Mehofer, M. (Eds.), 2014. Western Anatolia before Troy. Proto-Urbanisation in the 4th Millenium BC?: Proceedings of the International Symposium held at the Kunsthistorisces Museum Wien Vienna, Austria, 21–24 November, 2012, OREA. Verlag der Österreichischen Akademie der Wissenschaften. Hornblower, S., 1994. Asia Minor, in: Lewis, D., Boardman, J., Hornblower, S., Ostwald, M. (Eds.), The Cambridge Ancient History VI: The FOurth Century BC. Cambridge University Press, Cambrdige, pp. 209–233. Huggett, J., 2020. Is big digital data different? Towards a new archaeological paradigm. Journal of Field Archaeology 45, S8–S17. https://doi.org/10.1080/00934690 .2020.1713281 Hürmüzlü, B., De Giorgi, A., Iversen, P.A., 2009. New research in Northwestern Pisidia: Ancient Konane (Conana) and its territory. Colloquium Anatolicum 8, 235–256. Hurst, H., Owen, S., 2005. Ancient Colonisations: Analogy, Similarity and Difference. Bloomsbury Academic. Husserl, E., 1964. The Phenomenology of Internal Time-Consciousness. Indiana University Press, Bloomington. Ingold, T., 2000. The Perception of the Environment: Essays on Livelihood, Dwelling and Skill. Psychology Press. Irwin, G.J., 1978. Pots and Entrepots: A study of settlement, trade and the development of economic specialization in Papuan prehistory. World Archaeology 9, 299–319.

228  Bibliography Jablonka, P., 2011. Troy in regional and interregional context, in: Steadman, S.R., McMahon, G. (Eds.), The Oxford Handbook of Ancient Anatolia: (10,000–323 BCE). OUP USA, pp. 717–733. Janssen, E., Poblome, J., Claeys, J., Kint, V., Degryse, P., Marinova, E., Muys, B., 2017. Fuel for debating ancient economies. Calculating wood consumption at urban scale in Roman Imperial times. Journal of Archaeological Science: Reports 11, 592–599. https://doi.org/10.1016/j.jasrep.2016.12.029 Janssen, M.A., Kohler, T.A., Scheffer, M., 2003. Sunk-cost effects and vulnerability to collapse in ancient societies. Current Anthropology 44, 722. Janssen, M.A., Scheffer, M., 2004. Overexploitation of renewable resources by ancient societies and the role of sunk-cost effects. Ecology and Society 9. Jennings, J., 2016. Killing Civilization: A Reassessment of Early Urbanism and Its Consequences. University of New Mexico Press, Albuquerque. Johnson, A.W., Earle, T., 2000. The Evolution of Human Societies: From Foraging Group to Agrarian State, Second Edition, Stanford University Press, Stanford, Calif. Johnson, D.P., 2008. Contemporary Sociological Theory: An Integrated MultiLevel Approach. Springer Science & Business Media. Johnson, G.A., 1980. Rank-Size Convexity and System Integration: A view from archaeology. Economic Geography 56, 234. https://doi.org/10.2307/142715 Johnson, G.A., 1982. Organizational structure and scalar stress, in: Renfrew, C., Rowlands, M., Segraves, B. (Eds.), Theory and Explanation in Archaeology. Academic Press, Cambridge. Kahya, T., 2015. Düver - Yarımada Kaya Kabartmaları. Cedrus: Akdeniz Uygarlıkları Araştırma Dergisi 3, 129–139. Kaptijn, E., Vandam, R., Poblome, J., Waelkens, M., 2012. Inhabiting the plain of Burdur: 2010 and 2011 Sagalassos Project Survey. News of Archaeology from Anatolia’s Mediterranean Areas 10, 142–147. Kauffman, S.A., 1993. The Origins of Order: Self-Organization and Selection in Evolution. Oxford University Press, Oxford. Kendall, D.G., 1969. Incidence matrices, interval graphs and seriation in archeology. Pacific Journal of Mathematics 28, 565–570. https://doi.org/10.2140/ pjm.1969.28.565 Kepler, T., 2019. Picasso and perception: Attending to the higher order, in: Krakauer, D.C. (Ed.), Worlds Hidden in Plain Sight, Originally in SFI Bulletin, Summer 2001. SFI Press. Kidder, T., Haiwang, L., Storozum, M., Zhen, Q., 2016. New perspectives on the collapse and regeneration of the Han Dynasty, in: Faulseit, R. (Ed.), Beyond Collapse: Archaeological Perspectives on Resilience, Revitalization, and Transformation in Complex Societies. Southern Illinois University Press, Carbondale, pp. 313–341. Kintigh, K.W., Altschul, J.H., Beaudry, M.C., Drennan, R.D., Kinzig, A.P., Kohler, T.A., Limp, W.F., Maschner, H.D.G., Michener, W.K., Pauketat, T.R., Peregrine, P., Sabloff, J.A., Wilkinson, T.J., Wright, H.T., Zeder, M.A., 2014. Grand challenges for archaeology. Proceedings of the National Academy of Sciences 111, 879. https://doi.org/10.1073/pnas.1324000111 Klir, G., 2001. Facets of Systems Science. Springer Science & Business Media. Knappett, C., Evans, T., Rivers, R., 2008. Modelling maritime interaction in Agean Bronze Age. Antiquity 82, 1009.1024.

Bibliography 229 Knox, H., Savage, M., Harvey, P., 2006. Social networks and the study of relations: Networks as method, metaphor and form. Economy and Society 35, 113–140. https://doi.org/10.1080/03085140500465899 Kohler, T., Johnson, C., Varien, M., Ortman, S., Reynolds, R., Kobti, Z., Cowan, J., Kolm, K., Smith, S., Yap, L., 2007. Settlement Ecodynamics in the Prehispanic Central Mesa Verde Region, in: Kohler, T., van der Leeuw, S. (Eds.) The Modelbased Archaeology of Socionatural Systems. School for Advanced Research Press, Santa Fe, pp. 61–104. Kohler, T., Varien, M., 2012. Emergence and Collapse of Early Villages: Models of Central Mesa Verde Archaeology. University of California Press, Berkeley. Kohler, T.A., 2012. Complex systems and archaeology, in: Hodder, I. (Ed.), Archaeological Theory Today. Polity Press, Cambridge, pp. 93–123. Kohler, T.A., Bocinsky, R.K., Cockburn, D., Crabtree, S.A., Varien, M.D., Kolm, K.E., Smith, S., Ortman, S.G., Kobti, Z., 2012a. Modelling prehispanic Pueblo societies in their ecosystems. Ecological Modelling 241, 30–41. https://doi. org/10.1016/j.ecolmodel.2012.01.002 Kohler, T.A., Cockburn, D., Hooper, P.L., Bocinsky, R.K., Kobti, Z., 2012b. The coevolution of group size And leadership: An agent-based public goods model for prehispanic pueblo societies. Advs. Complex Syst 15, 1150007. https://doi.org/10.1142/ S0219525911003256 Kohler, T.A., Gumerman, G.G. (Eds.), 2000. Dynamics in Human and Primate Societies: Agent-Based Modeling of Social and Spatial Processes, 1st ed., Oxford University Press, New York. Kohler, T.A., Sebastian, L., 1996. Population aggregation in the prehistoric North American Southwest. American Antiquity 61, 597–602. https://doi.org/ 10.2307/281844 Kohler, T.A., Smith, M.E., Bogaard, A., Feinman, G.M., Peterson, C.E., Betzenhauser, A., Pailes, M., Stone, E.C., Marie Prentiss, A., Dennehy, T.J., Ellyson, L.J., Nicholas, L.M., Faulseit, R.K., Styring, A., Whitlam, J., Fochesato, M., Foor, T.A., Bowles, S., 2017. Greater post-neolithic wealth disparities in Eurasia than in North America and Mesoamerica. Nature 551, 619–622. https:// doi.org/10.1038/nature24646 Kolb, F., 2008. Burg, Polis, Bischofssitz: Geschichte der Siedlungskammer Von Kyaneai in Der Südwesttürkei. Philipp von Zabern Verlag, GmbH, Mainz. Kosmetatou, E., 1997. Pisidia and the Hellenistic Kings from 323 to 133 BC. Ancient Society 28, 5–37. https://doi.org/10.2143/AS.28.0.630067 Kosso, P., 2001. Knowing the Past: Philosophical Issues of History and Archaeology. Humanity Books, Amherst, N.Y. Kostof, S., 1999. The City Shaped: Urban Patterns and Meanings Through History. Thames & Hudson. Krugman, P., 1991. Increasing Returns and Economic Geography. Journal of Political Economy 99, 483–99. LaBuff, J., 2015. Polis expansion and elite power in Hellenistic Karia, Alexander the Great and the Hellenistic world. Lexington Books, Lanham. Lagopoulos, A., 2009. A History of the Greek City. Archaeopress. Lake, M.W., 2014. Trends in archaeological simulation. Journal of Archaeological Method and Theory 21, 258–287. https://doi.org/10.1007/s10816-013-9188-1 Laurence, R., Wallace-Hadrill, A., 1997. Domestic Space in the Roman World: Pompeii and Beyond. JRA, Portsmouth.

230  Bibliography Leonard, R.D., Reed, H.E., 1993. Population Aggregation in the Prehistoric American Southwest: A Selectionist Model. American Antiquity 58, 648–661. https://doi.org/10.2307/282200 Lewin, R., 1992. Complexity: Life at the Edge of Chaos. University of Chicago Press, Chicago. Lloyd, S., 1972. Beycesultan Vol. 3: Late Bronze Age Architecture. British Institute of Archaeology at Ankara. Lloyd, S., 2001. Measures of complexity: A nonexhaustive list. IEEE Control Systems 21, 7–8. https://doi.org/10.1109/MCS.2001.939938 Lloyd, S., Mellaart, J., 1958. Beycesultan excavations: Fourth preliminary report. Anatolian Studies 8, 93–125. Lloyd, S., Mellaart, J., 1962. Beycesultan Vol. 1: The Chalcolithic and early Bronze Age levels, Occasional publications of the British Institute of Archaeology at Ankara. British Institute of Archaeology at Ankara, London. Lobo, J., Bettencourt, L.M., Smith, M.E., Ortman, S., 2019. Settlement scaling theory: Bridging the study of ancient and contemporary urban systems. Urban Studies 57(4), 731–747. https://doi.org/10.1177/0042098019873796 Lorenz, E.N., 1963. Deterministic nonperiodic flow. Journal of the Atmospheric Sciences 20, 130–141. https://doi.org/10.1175/1520-0469(1963)020 2.0.CO;2 Lotka, A.J., 1956. Elements of Mathematical Biology. Dover Publications. Low, S.M., Lawrence-Zúñiga, D., 2003. Anthropology of Space and Place: Locating Culture. Wiley. Lucas, G., 2005. The Archaeology of Time. Routledge, London. Luhmann, N., 1995. Social Systems. Stanford Stanford university press. Lull, V., Micó, R., 2011. Archaeology of the Origin of the State: The Theories. Oxford university press, Oxford. Ma, J., 1999. Antiochus III and the Cities of Western Asia Minor. Oxford University Press, London; New York. Ma, J., 2013. Hellenistic Empires, in: Bang, P.F., Scheidel, W. (Eds.), The Oxford Handbook of the State in the Ancient Near East and Mediterranean. Oxford University Press, Oxford, pp. 324–357. MacSweeney, N., 2011. Community Identity and Archaeology: Dynamic Communities at Aphrodisias and Beycesultan. The University of Michigan Press, Ann Arbor. Maisels, C.K., 2010. The Archaeology 0f Politics And Power: Where, When And Why The First States Formed. Oxbow, Oxford. Malaina, A., 2014. Complex adaptive systems and global capitalism: The risk of a new ideology of global complexity. World Futures 70, 469–485. https://doi.org/10. 1080/02604027.2014.985950 Mann, M., 1986. The Sources of Social Power: Volume 1, A History of Power from the Beginning to AD 1760. Cambridge University Press. Manning, J.G., 2018. The Open Sea: The Economic Life of the Ancient Mediterranean World from the Iron Age to the Rise of Rome. Princeton University Press. Manning, P., 2020. A History of Humanity: The Evolution of the Human System. Cambridge University Press. Marchal, J.H., 1975. On the concept of a system. Philosophy of Science 42, 448–468. Marsh, B., Kealhofer, L., 2014. Scales of impact: Settlement history and landscape change in the Gordion Region, central Anatolia. The Holocene 24, 689–701. https://doi.org/10.1177/0959683614526937

Bibliography 231 Marston, J.M., 2015. Modeling resilience and sustainability in ancient agricultural systems. Journal of Ethnobiology 35, 585–605. https://doi.org/10.2993/ etbi-35-03-585-605.1 Martini, W., Eschbach, N., 2017. die Akropolis von Perge: Die ergebnisse der grabungen 1998-2004 und 2008. AKMED, Antalya. Maschner, H., Bentley, R.A., 2003. The power law of rank and household on the North pacific, in: Bentley, R., Maschner, Herbert (Eds.), Complex Systems and Archaeology. Utah University Press, Salt Lake City, pp. 47–60. Massa, M., Bachhuber, C., Şahin, F., Bostanci-Kolankaya, N., Tuna, Y., 2019. The Konya Regional Archaeological Survey Project: The 2017 and 2018 field seasons. Anatolica 45, 159–180. https://doi.org/10.2143/ANA.45.0.3287005 Massa, M., Bachhuber, C., Şahin, F., Erpehlivan, H., Osborne, J., Lauricella, A.J., 2020. A landscape-oriented approach to urbanisation and early state formation on the Konya and Karaman plains, Turkey. Anatolian Studies 70, 45–75. https:// doi.org/10.1017/S0066154620000034 Massa, M., Palmisano, A., 2018. Change and continuity in the long-distance exchange networks between western/central Anatolia, northern Levant and northern Mesopotamia, c.3200–1600 BCE. Journal of Anthropological Archaeology 49, 65–87. https://doi.org/10.1016/j.jaa.2017.12.003 Massa, M., Şahoğlu, V., 2015. The 4.2 ka BP climatic event in west and central Anatolia: Combining palaeo-climatic proxies and archaeological data, in: Meller, H., Arz, H., Jung, R., Risch, R. (Eds.), 2200 BC - A Climatic Breakdown as a Cause for the Collapse of the Old World? Landesmuseums für Vorgeschichte Halle, Halle, pp. 61–78. Matessi, A., 2018. The making of Hittite Imperial Landscapes: Territoriality and balance of power in South-Central Anatolia during the Late Bronze Age. Journal of Ancient Near Eastern History 3, 117–162. https://doi.org/10.1515/janeh-20170004 Maturana, H.R., Varela, F.J., 1980. Autopoiesis and Cognition: The Realization of the Living, 1st ed., D. Reidel Publishing Company, Dordrecht, Holland; Boston. Mayfield, J., 2016. Evolution and market complexity, in: Wilson, D.S., Kirman, A. (Eds.), Complexity and Evolution: Toward a New Synthesis for Economics. The MIT Press, Cambridge, pp. 47–63. https://doi.org/10.7551/mitpress/ 9780262035385.001.0001 McCurdy, L., Abrams, E.M., 2019. Architectural Energetics in Archaeology: Analytical Expansions and Global Explorations. Routledge. McGlade, J., van der Leeuw, S., 1997. Introduction: Archaeology and non-linear dynamics. New approaches to long-term change, in: McGlade, J., van der Leeuw, S. (Eds.), Time, Process and Structured Transformation in Archaeology. Routledge, London, pp. 1–31. McGuire, R.H., 1983. Breaking down cultural complexity: Inequality and heterogeneity. Advances in Archaeological Method and Theory 6, 91.142. Meadows, D., 2008. Thinking in Systems: A Primer. Chelsea Green Publishing. Meadows, D., Meadows, D., Randers, J., 1972. The Limits to Growth. Universe Books, New York. Meier, R.L., 1972. Communications Stress. Annual Review of Ecology and Systematics 3, 289–314. https://doi.org/10.1146/annurev.es.03.110172.001445 Meijers, E., 2007. From central place to network model: Theory and evidence of a paradigm change. Tijdschrift voor economische en sociale geografie 98, 245–259. https://doi.org/10.1111/j.1467-9663.2007.00394.x

232  Bibliography Mellaart, J., 1963. Early cultures of the South Anatolian Plateau, II. Anatolian Studies 13, 199–236. https://doi.org/10.2307/3642494 Mellaart, J., 1970. Excavations at Hacilar, Vol. 1. British Institute of Archaeology, Edinburgh. Mertel, A., Ondrejka, P., Šabatová, K., 2018. Spatial predictive modeling of prehistoric sites in the Bohemian-Moravian Highlands based on graph similarity analysis. Open Geosciences 10, 261–274. https://doi.org/10.1515/geo-2018-0020 Merton, R.K., 1968. Social Theory and Social Structure, Enl. ed., Free Press, New York NY. Meskell, L., Preucel, R.W., 2007. A Companion to Social Archaeology, Social Archaeology. Blackwell, Oxford. Mesoudi, A., 2016. Cultural evolution: A review of theory, findings and controversies. Evolutionary Biology; 43, 481–497. https://doi.org/10.1007/s11692-0159320-0 Michel, C., 2011. The Karum period on the Plateau, in: Steadman, S.R., McMahon, G. (Eds.), The Oxford Handbook of Ancient Anatolia: (10,000-323 BCE). OUP USA, pp. 313–336. Milgram, S., 1970. The experience of living in cities. Science 167, 1461. https://doi. org/10.1126/science.167.3924.1461 Miller, J.H., Page, S.E., 2007. Complex Adaptive Systems: An Introduction to Computational Models of Social Life: An Introduction to Computational Models of Social Life. Princeton University Press. Miller, N.F., 2011. Botanical Aspects of Environment and Economy at Gordion, Turkey. University of Pennsylvania Press. Mills, C.W., 1959. The Sociological Imagination. Oxford University Press, New York. Mingers, J., White, L., 2010. A review of the recent contribution of systems thinking to operational research and management science. European Journal of Operational Research 207, 1147–1161. https://doi.org/10.1016/j.ejor.2009.12.019 Minnis, P., 1996. Notes on economic uncertainty and human behavior in the prehistoric North American southwest, in: Tainter, J., Tainter, B. (Eds.), Evolving Complexity and Environmental Risk in the Prehistoric Southwest, Proceedings of the Workshop “Resource Stress, Economic Uncertainty, and Human Response in the Prehistoric Southwest,” Held February 25–29, 1992, in Santa Fe, NM. Addison-Wesley, Reading, pp. 57–78. Miranda, L., Freeman, J., 2020. The two types of society: Computationally revealing recurrent social formations and their evolutionary trajectories. PLoS ONE 15, e0232609. https://doi.org/10.1371/journal.pone.0232609 Mitchell, M., 2009. Complexity: A Guided Tour. Oxford University Press, Oxford; New York. Mitchell, S., 1991. The Hellenization of Pisidia. Mediterranean Archaeology 4, 119.145. Mithen, S., 1994. Simulating prehistoric hunter-gatherer societies, in: Gilbert, N., Doran, J. (Eds.), Simulating Societies. The Computer Simulation of Social Phenomena. Routledge, pp. 165–193. https://doi.org/10.4324/9781351165129-8 Monsieur, P., Daems, D., Poblome, J., 2017. Hellenistic and Italic Amphorae from Sagalassos. HEROM 6, 97–119. https://doi.org/10.11116/HEROM.6.1.4 Morgan, L.H., 1877. Ancient Society; Or, Researches in the Lines of Human Progress from Savagery, Through Barbarism to Civilization. H. Holt.

Bibliography 233 Morris, I., 2013. The Measure of Civilization: How Social Development Decides the Fate of Nations. Princeton University Press, Princeton. Morris, I., Scheidel, W. (Eds.), 2009. The Dynamics of Ancient Empires: State Power from Assyria to Byzantium. Oxford University Press, Oxford. Muhly, J., 2011. Metals and Metallurgy, in: Steadman, S.R., McMahon, G. (Eds.), The Oxford Handbook of Ancient Anatolia: (10,000-323 BCE). Oxford University Press, pp. 858–876. Murdock, G.P., Wilson, S.F., 1972. Settlement patterns and community organization: Cross-cultural codes 3. Ethnology 11, 254–295. https://doi.org/10.2307/3773219 Muthukrishna, M., Schulman, B., Vasilescu, V., Henrich, J., 2014. Sociality influences cultural complexity. Proceedings of the Royal Society B: Biological Sciences 281, 20132511. https://doi.org/10.1098/rspb.2013.2511 Muys, B., 2013. Sustainable development within planetary boundaries: A functional revision of the definition based on the thermodynamics of complex social-ecological systems. Challenges in Sustainability 1, 41–52. https://doi.org/10.12924/ cis2013.01010041 Nelson, M., 1996. Technological strategies responsive to subsistence stress., in: Tainter, J., Tainter, B. (Eds.), Evolving Complexity and Environmental Risk in the Prehistoric Southwest, Proceedings of the Workshop “Resource Stress, Economic Uncertainty, and Human Response in the Prehistoric Southwest,” Held February 25–29, 1992, in Santa Fe, NM. Addison-Wesley, Reading, pp. 107–144. Nelson, M., Hegmon, M., Kulow, S., Peeples, M., Kintigh, K., Kinzig, A., 2011. Resisting diversity: A long-term archaeological study. Ecology and Society 16. https://doi.org/10.5751/ES-03887-160125 Nelson, M.C., Hegmon, M., Kulow, S., Schollmeyer, K.G., 2006. Archaeological and ecological perspectives on reorganization: A case study from the mimbres region of the U.S. Southwest. American Antiquity 71, 403–432. https://doi. org/10.2307/40035359 Netting, R., 1982. Some home truths on household size and wealth. American Behavioral Scientist 25, 641–662. https://doi.org/10.1177/000276482025006004 Netto, V.M., 2017. The Social Fabric of Cities. Routledge, New York. Netto, V.M., Brigatti, E., Meirelles, J., Ribeiro, F.L., Pace, B., Cacholas, C., Sanches, P., 2018. Cities, from information to interaction. Entropy 20, 834. https:// doi.org/10.3390/e20110834 Netto, V.M., Meirelles, J., Ribeiro, F.L., 2017. Social interaction and the city: the effect of space on the reduction of entropy. Complexity. https://doi.org/ 10.1155/2017/6182503 Newman, M., 2010. Networks: An Introduction. Oxford University Press, Oxford. Newman, M., Barabási, A.-L., Watts, D.J., 2006. The Structure and Dynamics of Networks, 1 edition. ed. Princeton University Press, Princeton. Norenzayan, A., Shariff, A.F., Gervais, W.M., Willard, A.K., McNamara, R.A., Slingerland, E., Henrich, J., 2016. The cultural evolution of prosocial religions. Behavioral and Brain Sciences 39. https://doi.org/10.1017/S0140525X14001356 Nowak, M., Highfield, R., 2011. SuperCooperators: Altruism, Evolution, and Why We Need Each Other to Succeed. Free Press, New York, NY. O’Brien, M.J., Boulanger, M.T., Buchanan, B., Bentley, R.A., Lyman, R.L., Lipo, C.P., Madsen, M.E., Eren, M.I., 2016. Design space and cultural transmission: Case studies from Paleoindian Eastern North America. J Archaeol Method Theory 23, 692–740. https://doi.org/10.1007/s10816-015-9258-7

234  Bibliography O’Brien, M.J., Lyman, R.L., 2000. Applying Evolutionary Archaeology: A Systematic Approach. Springer US. Ocakoğlu, F., Çilingiroğlu, Ç, Erkara, İP., Ünan, S., Dinçer, B., Akkiraz, M.S., 2019. Human-climate interactions since the Neolithic period in Central Anatolia: Novel multi-proxy data from the Kureyşler area, Kütahya, Turkey. Quaternary Science Reviews 213, 1–17. https://doi.org/10.1016/j.quascirev.2019.04.016 Ortman, S., 2019. Imagining complex societies, in: Krakauer, D.C. (Ed.), Worlds Hidden in Plain Sight. SFI Press, pp. 181–191. Ortman, S., Cabaniss, A., Sturm, J., Bettencourt, L., 2015. Settlement scaling and increasing returns in an ancient society. Science Advances 1, e1400066. Ortman, S., Cofey, G., 2015. Universal Scaling: Evidence from Village-Level Societies. Santa Fe Institute Working Papers. Ortman, S., Peregrine, P., 2018. The contours of cultural evolution, in: Sabloff, J.A., Sabloff, P.L.W. (Eds.), The Emergence of Premodern States: New Perspec­ tives on the Development of Complex Societies. SFI Press, Santa Fe, NM, pp. 187–217. Ortman, S.G., Cabaniss, A.H.F., Sturm, J.O., Bettencourt, L.M.A., 2014. The pre-history of urban scaling. PLoS ONE 9, e87902. https://doi.org/10.1371/ journal.pone.0087902 Ortman, S.G., Coffey, G.D., 2017. Settlement scaling in middle-range societies. American Antiquity 82, 662–682. https://doi.org/10.1017/aaq.2017.42 Ortman, S.G., Davis, K.E., Lobo, J., Smith, M.E., Bettencourt, L.M.A., Trumbo, A., 2016. Settlement scaling and economic change in the Central Andes. Journal of Archaeological Science 73, 94–106. https://doi.org/10.1016/j.jas.2016.07.012 Osborne, J.F., Massa, M., Şahin, F., Erpehlivan, H., Bachhuber, C., 2020. The city of Hartapu: Results of the Türkmen-Karahöyük Intensive Survey Project. Anatolian Studies 70, 1–27. https://doi.org/10.1017/S0066154620000046 Osborne, R. (Editor), Cunliffe, B. (Editor), 1996. Mediterranean Urbanization 800–600 BC. Oxford University Press, Oxford. Osborne, R., 1996. Urban Sprawl: What is urbanization and why does it matter?, in: Osborne, R., Cunliffe, B. (Eds.), Mediterranean Urbanization 800–600 BC. Oxford University Press, Oxford, pp. 1–16. Östborn, P., Gerding, H., 2015. The diffusion of fired bricks in Hellenistic Europe: A similarity network analysis. Journal of Archaeological Method and Theory 22, 306–344. https://doi.org/10.1007/s10816-014-9229-4 Ostrom, E., 1990. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press. Ostrom, E., 2009. A general framework for analyzing sustainability of socialecological systems. Science325, 419–422. https://doi.org/10.1126/science.1172133 Owens, E.J., 1991. The City in the Greek and Roman World. Routledge, London. Packard, N.H., 1988. Adaptation Toward the Edge of Chaos. University of Illinois at Urbana-Champaign, Center for Complex Systems Research. Padgett, J.F., Powell, W.W., 2012. The Emergence of Organizations and Markets. Princeton University Press, Princeton. Palmisano, A., 2017. Confronting scales of settlement hierarchy in state-level societies: Upper Mesopotamia and Central Anatolia in the Middle Bronze Age. Journal of Archaeological Science: Reports 14, 220–240. https://doi.org/10.1016/j. jasrep.2017.05.029 Parsons, T., 1977. Social Systems and the Evolution of Action Theory. Free Press [etc.]cop.

Bibliography 235 Parsons, T., 1990. Prolegomena to a Theory of Social Institutions. American Sociological Review 55, 319–333. https://doi.org/10.2307/2095758 Peeples, M.A., 2019. Finding a place for networks in archaeology. Journal of Archaeological Research 27, 451–499. https://doi.org/10.1007/s10814-019-09127-8 Peregrine, P., 2018. Towards a theory of recurrent social formations, in: Sabloff, J.A., Sabloff, P.L.W. (Eds.), The Emergence of Premodern States: New Perspectives on the Development of Complex Societies. SFI Press, Santa Fe, NM, pp. 271–296. Peregrine, P.N., Ember, M. (Eds.), 2003. Encyclopedia of Prehistory Complete set of Volumes 1-8 and Volume 9, the index volume: Published in conjunction with the Human Relations Area Files. Springer US. Peters, R., Zimmermann, A., 2017. Resilience and cyclicity: Towards a macrohistory of the Central European Neolithic. Quaternary International, Adaptive Cycles in Archaeology 446, 43–53. https://doi.org/10.1016/j.quaint.2017.03.073 Pianka, E.R., 1970. On r- and K-selection. The American Naturalist 104, 592–597. https://doi.org/10.1086/282697 Pickett, S.T.A., Cadenasso, M.L., Grove, J.M., 2005. Biocomplexity in coupled natural–human systems: A multidimensional framework. Ecosystems 8, 225–232. https://doi.org/10.1007/s10021-004-0098-7 Pitts, F.R., 1965. A graph theoretic approach to historical geography. The Professional Geographer 17, 15–20. https://doi.org/10.1111/j.0033-0124.1965.015_m.x Plog, F.T., 1975. Systems theory in archeological research. Annual Review of Anthropology 4, 207–224. https://doi.org/10.1146/annurev.an.04.100175.001231 Pluciennik, M., 2005. Social Evolution. Bloomsbury Academic. Poblome, J., Braekmans, D., Music, B., Van Der Enden, M., Neyt, B., De Graeve, B., Degryse, P., 2013a. A pottery kiln underneath the Odeon of ancient Sagalassos, SW Turkey. The excavation results, the table wares and their archaeometrical analysis. Presented at the Networks in the Hellenistic World according to the pottery in the Eastern Mediterranean and Beyond, Archaeopress, pp. 193–204. Poblome, J., Braekmans, D., Waelkens, M., Firat, N., Vanhaverbeke, H., Martens, F., Kaptijn, E., Vyncke, K., Willet, R., Degryse, P., 2013b. How did Sagalassos come to be? A ceramological survey, in: Tekocak, M. (Ed.), Studies in Honour of K. Levent Zoroğlu. Suna & Inan Kiraç Research Institute on Mediterranean Civilizations, Antalya, pp. 527–540. Poblome, J., Daems, D., 2019. Once upon a Sagalassos: A story about origins and emergence, in: Poblome, J., Torun, E., Talloen, P., Waelkens, M. (Eds.), Meanwhile in the Mountains: Sagalassos. Yapı Kredi Yayınları, Istanbul, pp. 59–69. Powell, A., Shennan, S., Thomas, M.G., 2009. Late Pleistocene demography and the appearance of modern human behavior. Science 324, 1298–1301. https://doi. org/10.1126/science.1170165 Powell, L.E., Isler, K., Barton, R.A., 2017. Re-evaluating the link between brain size and behavioural ecology in primates. Proc. R. Soc. B 284, 20171765. https://doi. org/10.1098/rspb.2017.1765 Premo, L.S., 2006. Agent-based models as behavioral laboratories for evolutionary anthropological research. Arizona Anthropologist 17, 91.113. Price, D.D.S., 1976. A general theory of bibliometric and other cumulative advantage processes. Journal of the American Society for Information Science 27, 292–306. https://doi.org/10.1002/asi.4630270505 Price, T.D., Feinman, G.M. (Eds.), 2010. Pathways to Power: New Perspectives on the Emergence of Social Inequity, Fundamental Issues in Archaeology. Springer, Dordrecht.

236  Bibliography Purcell, N., 2007. Statics and dynamics: Ancient Mediterranean urbanism, in: Osborne, R., Cunliffe, B. (Eds.), Mediterranean Urbanization 800–600 BC. Oxford University Press, Oxford, pp. 249–272. Ragin, C.C., 1989. The Comparative Method: Moving Beyond Qualitative and Quantitative Strategies. University of California Press. Rahmstorf, L., 2006. In search of the earliest balance weights, scales and weighing systems from the east Mediterranean, the near and Middle East, in: Ascalone, E., Peyronel, L. (Eds.), Weights in Context: Bronze Age Weighing Systems of Eastern Mediterranean. Chronology, Typology, Material and Archaeological Contexts. Instituto Italiano di Numismatica, Rome, pp. 9–46. Rapoport, A., 1982. The meaning of the built environment: a nonverbal communication approach. Sage, Beverly Hills. Rapoport, A., 1988. Levels of meaning in the built environment, in: Poyatos, F. (Ed.), Cross-Cultural Perspectives in Nonverbal Communication, Ed. by Fernado Poyatos. Hogrefe, Toronto, pp. 317–336. Rapoport, A., 1990. Systems of activities and systems of settings, in: Kent, S. (Ed.), Domestic Architecture and the Use of Space: An Interdisciplinary CrossCultural Study. Cambridge University Press, New York, pp. 9–20. Rapoport, A., 2006. Archaeology and environment-behavior studies, in: Ashmore, W., Dobres, M.-A., Nelson, S., Rosen, A.M. (Eds.), Integrating the Diversity of 21st Century Anthropology: The Life and Intellectual Legacies of Susan Kent. UCLA Press, Los Angeles, pp. 59–70. Ratté, C., 2008. Reflections on the urban development of Hellenistic Sardis, in: Cahill, N.D. (Ed.), Love for Lydia. A Sardis Anniversary Volume Presented to Crawford H. Greenwalt Jr., Archaeological Exploration of Sardis, Report. Archaeological Exploration of Sardis, Cambridge, MA, pp. 125–133. Read, D., van der Leeuw, S., 2008. Biology is only part of the story …. Philosophical Transactions of the Royal Society B: Biological Sciences 363, 1959–1968. https:// doi.org/10.1098/rstb.2008.0002 Reckwitz, A., 2002. Toward a theory of social practices: A development in culturalist theorizing. European Journal of Social Theory. https://doi.org/10.1177/ 13684310222225432 Redman, C., Crumley, C., Hassan, F., Hole, F., Morais, J., Reidel, F., Scarborough, V., Tainter, J., Turchin, P., Yasuda, Y., 2007. Group report: Millennial perspectives on the dynamic interaction of climate, people, and resources, in: Costanza, R., Graumlich, L., Steffen, W. (Eds.), Sustainability or Collapse? An Integrated History and Future of People on Earth. MIT Press, Cambridge, pp. 115–148. Redman, C., Kinzig, A., 2003. Resilience of past landscapes: Resilience theory, society, and the longue durée. Ecology and Society; Vol. 7, No. 1 (2003). https:// doi.org/vol7/iss1/art14/ Redman, C.L., 2005. Resilience theory in archaeology. American Anthropologist 107, 70.77. Renfrew, C., 1986. Introduction: Peer polity interaction and socio-political change, in: Renfrew, C., Cherry, J.F. (Eds.), Peer Polity Interaction and Socio-Political Change. Cambridge University Press, pp. 1–18. Renfrew, C., 2008. The city through time and space. Transformations of centrality, in: Marcus, J., Sabloff, J.A. (Eds.), The Ancient City: New Perspectives on Urbanism in the Old and New World, School for Advanced Research Resident Scholar Series. School for Advanced Research Press, Santa Fe, pp. 29–52.

Bibliography 237 Renfrew, C., Bahn, P., 2012. Archaeology: Theories, Methods and Practice, 6 edition. ed. Thames and Hudson Ltd, London. Renfrew, C., Cherry, J.F., 1986. Peer Polity Interaction and Socio-political Change. Cambridge University Press. Rihll, T.E., Wilson, A.G., 1987. Spatial interaction and structural models in historical analysis: Some possibilities and an example. Hism 2, 5–32. https://doi. org/10.3406/hism.1987.1300 Robb, J., 2007. The Early Mediterranean Village: Agency, Material Culture, and Social Change in Neolithic Italy. Cambridge University Press, Cambridge; New York. Robb, J.E., Pauketat, T.R. (Eds.), 2012. Big Histories, Human Lives: Tackling Problems of Scale in Archaeology. SAR Press, Santa Fe. Roberts, N., Eastwood, W.J., Kuzucuoğlu, C., Fiorentino, G., Caracuta, V., 2011. Climatic, vegetation and cultural change in the eastern Mediterranean during the mid-Holocene environmental transition. The Holocene 21, 147–162. https://doi. org/10.1177/0959683610386819 Robertson, J., 1995. The social and economic organization of ancient Mesopotamian temples, in: Sasson, J.M. (Ed.), Civilizations of the Ancient Near East. Scribner. Rodman, M.C., 1992. Empowering place: Multilocality and multivocality. American Anthropologist 94, 640–656. https://doi.org/10.1525/aa.1992.94.3.02a00060 Rojas, F., 2016. Review of: Empire, Authority, and Autonomy in Achaemenid Anatolia (paperback reprint; first published 2013). BMCR. Romanowska, I., Crabtree, S.A., Harris, K., Davies, B., 2019. Agent-based modeling for archaeologists: Part 1 of 3. Advances in Archaeological Practice 7, 178–184. https://doi.org/10.1017/aap.2019.6 Roosevelt, C.H., 2009. The Archaeology of Lydia, from Gyges to Alexander. Cambridge University Press. Rose, C.B., 2017. Fieldwork at Phrygian Gordion, 2013–2015. American Journal of Archaeology 121, 135–178. https://doi.org/10.3764/aja.121.1.0135 Rose, C.B., Darbyshire, G., 2011. The New Chronology of Iron Age Gordion. University of Pennsylvania Press, Incorporated. Rosen, A.M., Rivera-Collazo, I., 2012. Climate change, adaptive cycles, and the persistence of foraging economies during the late Pleistocene/Holocene transition in the Levant. PNAS 109, 3640–3645. https://doi.org/10.1073/pnas.1113931109 Rosser, J.B., 2011. Complex Evolutionary Dynamics in Urban-Regional and Ecologic-Economic Systems. Springer. Rubio-Campillo, X., Montanier, J.-M., Rull, G., Bermúdez Lorenzo, J.M., Moros Díaz, J., Pérez González, J., Remesal Rodríguez, J., 2018. The ecology of Roman trade. Reconstructing provincial connectivity with similarity measures. Journal of Archaeological Science 92, 37–47. https://doi.org/10.1016/j.jas.2018.02.010 Sabloff, J.A., 2018. Extending our knowledge of premodern states, in: Sabloff, J.A., Sabloff, P.L.W. (Eds.), The Emergence of Premodern States: New Perspectives on the Development of Complex Societies. SFI Press, Santa Fe, NM, pp. 1–13. Sagona, A.G., Zimansky, P.E., 2009. Ancient Turkey. Routledge. Şahoğlu, V., 2005. The Anatolian trade network and the Izmir Region during the early Bronze Age. Oxford Journal of Archaeology 24, 339–361. https://doi. org/10.1111/j.1468-0092.2005.00240.x Salmon, M.H., 1978. What can systems theory do for archaeology? American Antiquity 43, 174–183. https://doi.org/10.2307/279242

238  Bibliography Salthe, S.N., 1985. Evolving Hierarchical Systems: Their Structure and Representation. Columbia University Press. Sams, G.K., 1993. Gordion and the near east in the early Phrygian Period., in: Mellink, M.J., Porada, E., Ôzgûç, T. (Eds.), Spects of Art and Iconography: Anatolia and Its Neighbours; Studies in Honor of Nimet Ôzgûç. Türk Tarih Kürümü Basimevi, Ankara, pp. 549–555. Sandars, N.K., 1978. The sea peoples: warriors of the ancient Mediterranean 12501150 BC. London Thames and Hudson. Sanderson, S.K., 1999. Social transformations: a general theory of historical development. Rowman & Littlefield, Totowa. Sanderson, S.K., 2001. Social evolution: Overview, in: Smelser, N.J., Baltes, P.B. (Eds.), International Encyclopedia of the Social & Behavioral Sciences. Pergamon, Oxford, pp. 14279–14286. https://doi.org/10.1016/B0-08-043076-7/03107-7 Savage, P.E., Whitehouse, H., François, P., Currie, T.E., Feeney, K., Cioni, E., Purcell, R., Ross, R.M., Larson, J., Baines, J., Haar, B. ter, Covey, A., Turchin, P., 2019. Reply to Beheim et al.: Reanalyses confirm robustness of original analyses (preprint). SocArXiv. https://doi.org/10.31235/osf.io/xjryt Schatzki, T.R., Knorr-Cetina, K., Savigny, E. von, 2001. The Practice Turn in Contemporary Theory. Psychology Press. Scheffer, M., Carpenter, S.R., Lenton, T.M., Bascompte, J., Brock, W., Dakos, V., Koppel, J., Leemput, I.A., Levin, S.A., Nes, E.H. van, Pascual, M., Vandermeer, J., 2012. Anticipating critical transitions. Science 338, 344–348. https://doi. org/10.1126/science.1225244 Scheidel, W., 2013. Studying the state, in: Bang, P.F. (Editor), Scheidel, W. (Editor) (Eds.), The Oxford Handbook of the State in the Ancient Near East and Mediterranean. Oxford university press, Oxford, pp. 5–57. Schofield, A.J., 1991. Interpreting Artefact Scatters: Contributions to Ploughzone Archaeology, Oxbow Monographs in Archaeology. Oxbow books, Oxford. Schoon, M., van der Leeuw, S., 2015. The shift toward social-ecological systems perspectives: Insights into the human-nature relationship. Natures-SciencesSocietes 23, 166–174. https://doi.org/10.1051/nss/2015034 Schoop, U.-D., 2011. The Chalcolithic on the Plateau, in: Steadman, S.R., McMahon, G. (Eds.), The Oxford Handbook of Ancient Anatolia: (10,000-323 BCE). OUP USA, pp. 150–173. Schoop, U.-D., 2014. Weaving Society in Late Chalcolithic Anatolia: Textile Production and Social Strategies in the 4th Millennium BC, in: Horejs, B., Mehofer, M. (Eds.), Western Anatolia before Troy. Proto-Urbanisation in the 4th Millenium BC?: Proceedings of the International Symposium Held at the Kunsthistorisces Museum Wien Vienna, Austria, 21–24 November, 2012, OREA. Verlag der Österreichischen Akademie der Wissenschaften, pp. 421–446. Schumpeter, J.A., 1942. Capitalism, Socialism and Democracy. Routledge. Schwartz, G.M., Nichols, J.J. (Eds.), 2006. After Collapse: The Regeneration of Complex Societies. University of Arizona Press, Tucson. Scott, A.J., Storper, M., 2015. The nature of cities: The scope and limits of urban theory. International Journal of Urban and Regional Research 39, 1–15. https:// doi.org/10.1111/1468-2427.12134 Seeher, J., 2001. Die Zerstörung der Stadt Hattusa, in: Akten Des IV. Internationalen Kongresses Für Hethitologie, Würzburg, 4-8 Oktober 1999. Harassowitz, Wiesbaden, pp. 623–634.

Bibliography 239 Seeher, J., 2011. The plateau: The Hittites, in: Steadman, S.R., McMahon, G. (Eds.), The Oxford Handbook of Ancient Anatolia: (10,000-323 BCE). OUP USA, pp. 376–392. Service, E.R., 1962. Primitive Social Organization: An Evolutionary Perspective. Random House. Shannon, C.E., Weaver, W., 1949. The Mathematical Theory of Communication. University of Illinois Press. Sherratt, S., 2003. The Mediterranean economy: Globalization at the end of the second millennium B.C.E., in: Dever, W.G., Gitin, S. (Eds.), Symbiosis, Symbolism, and the Power of the Past: Canaan, Ancient Israel, and Their Neighbors from the Late Bronze Age through Roman Palaestina ; Proceedings of the Centennial Symposium, W.F. Albright Institute of Archaeological Research and American Schools of Oriental Research, Jerusalem, May 29/31, 2000. Eisenbrauns, Winona Lake, Ind, pp. 37–62. Shin, J., Price, M.H., Wolpert, D.H., Shimao, H., Tracey, B., Kohler, T.A., 2020. Scale and information-processing thresholds in Holocene social evolution. Nature Communications 11, 2394. https://doi.org/10.1038/s41467-020-16035-9 Shove, E., Pantzar, M., Watson, M., 2012. The Dynamics of Social Practice: Everyday Life and How It Changes. SAGE. Siegelová, J., 2001. Der Regionalpalast in der Verwaltung des hethitischen Staates. Altorientalische Forschungen 28, 193–208. https://doi.org/10.1524/aofo.2001.28.2.193 Simmel, G., 2012. Soziologie des Raumes, Reprint. ed. Jazzybee Verlag. Simon, H., 1955. On a class of skew distribution functions. Biometrika 42, 425–440. Simon, H.A., 1962. The architecture of complexity. Proceedings of the American Philosophical Society 106, 467–482. Sindbæk, S.M., 2007. The small world of the Vikings: Networks in early medieval communication and exchange. Norwegian Archaeological Review 40, 59–74. https://doi.org/10.1080/00293650701327619 Sitte, R., 2009. About the predictability and complexity of complex systems, in: Aziz-Alaoui, M., Bertelle, C. (Eds.), From System Complexity to Emergent Properties. Springer, pp. 23–48. Sjoberg, G., 1960. The Preindustrial City: Past and Present. Free Press, New York (N.Y.). Slingerland, E., Monroe, M.W., Spicer, R., Muthukrishna, M., 2019. Historians Respond to Whitehouse et al. (2019), “Complex Societies Precede Moralizing Gods Throughout World History” (preprint). PsyArXiv. https://doi.org/10.31234/ osf.io/2amjz Smejda, L., Baumanova, M., 2015. Conceptual crossroads: Community and society in prehistory, in: Kristiansen, K., Smejda, Ladislav, Turek, J. (Eds.), Paradigm Found: Archaeological Theory - Present, Past and Future. Essays in Honour of Evzen Neustupny. Oxbow Books, Oxford, pp. 47–60. Smith, A.T., 2003. The Political Landscape: Constellations of Authority in Early Complex Polities. University of California Press, Berkeley. Smith, E., Morowitz, H., 2019. Searching for the laws of life: Separating chance from necessity, in: Krakauer, D.C. (Ed.), Worlds Hidden in Plain Sight. SFI Press, pp. 115–127. Smith, M., 2015. How can archaeologists make better arguments? The SAA Archaeological Record 15, 18.23. Smith, M., 2019. Energized crowding and the generative role of settlement aggregation and urbanization, in: Gyucha, A. (Ed.), Coming Together: Comparative

240  Bibliography Approaches to Population Aggregation and Early Urbanization. State University of New York Press, New York, pp. 37–58. Smith, M.E., 2014. Peasant mobility, local migration and premodern urbanization. World Archaeology 46, 516–533. https://doi.org/10.1080/00438243.2014.931818 Smith, M.E., 2017. Why Settlement Scaling Research is a Good Fit for Archaeology, in: SocArXiv. Presented at the “Settlement Scaling in Archaeology: Not Just Modern, Not Just Urban”, 82nd Annual Meeting, Society for American Archaeology, Vancouver, BC. https://doi.org/10.17605/OSF.IO/FWXFS Snodgrass, A., 1986. Interaction by design: The Greek city state, in: Cherry, J.F., Renfrew, C. (Eds.), Peer Polity Interaction and Socio-Political Change. Cambridge University Press, Cambridge, pp. 47–58. Solich, M., Bradtmöller, M., 2017. Socioeconomic complexity and the resilience of hunter-gatherer societies. Quaternary International, Adaptive Cycles in Archaeology 446, 109–127. https://doi.org/10.1016/j.quaint.2017.06.064 Spencer, C.S., 1990. On the tempo and mode of state formation: Neoevolutionism reconsidered. Journal of Anthropological Archaeology 9, 1–30. https://doi. org/10.1016/0278-4165(90)90004-W Spencer, C.S., 2014. Modeling the evolution of bureaucracy: Political-economic reach and administrative complexity. Social Evolution & History 13, 42–66. Spencer-Brown, G., 1972. Laws of Form. Julian Press. Steadman, S., 2011. The Early Bronze Age on the Plateau, in: Steadman, S.R., McMahon, G. (Eds.), The Oxford Handbook of Ancient Anatolia: (10,000-323 BCE). OUP USA, pp. 229–259. Steadman, S.R., 2000. Spatial patterning and social complexity on Prehistoric Anatolian tell sites: Models for mounds. Journal of Anthropological Archaeology 19, 164–199. https://doi.org/10.1006/jaar.2000.0363 Steward, J.H., 1949. Cultural causality and law: A trial formulation of the development of early civilizations. American Anthropologist 51, 1–27. Steward, J.H., 1972. Theory of Culture Change: The Methodology of Multilinear Evolution. University of Illinois Press. Stiner, M.C., Kuhn, S.L., 2006. Changes in the ‘connectedness’ and resilience of Paleolithic Societies in Mediterranean ecosystems. Hum Ecol 34, 693–712. https:// doi.org/10.1007/s10745-006-9041-1 Stonedahl, F., Wilensky, U., 2010. NetLogo Artificial Anasazi model. http://ccl. northwestern.edu/netlogo/models/ArtificialAnasazi. Storper, M., 2010. Why does a city grow? Specialisation, human capital or institutions? Urban Studies 47, 2027–2050. https://doi.org/10.1177/0042098009359957 Storper, M., Venables, T., 2004. Buzz: Face-to-face contact and the urban economy. Journal of Economic Geography 4, 351–370. Stow, C., Allen, C., Garmestani, A., 2007. Evaluating discontinuities in complex systems: Toward quantitative measures of resilience. Ecology and Society 12. https://doi.org/10.5751/ES-02055-120126 Strootman, R., 2014. Courts and Elites in the Hellenistic Empires: The Near East after the Achaemenids, c. 330 to 30 BCE, Edinburgh Studies in Ancient Persia. Edinburgh University Press, Edinburgh. Sundstrom, S., Angeler, D., Garmestani, A., Garcia, J.-H., Allen, C.R., 2014. Transdisciplinary Application of Cross-Scale Resilience. Nebraska Cooperative Fish & Wildlife Research Unit – Staff Publications. Sundstrom, S.M., Allen, C.R., 2019. The adaptive cycle: More than a metaphor. Ecological Complexity 39, 100767. https://doi.org/10.1016/j.ecocom.2019.100767

Bibliography 241 Tainter, J., 1996. Complexity, problem solving and sustainable societies, in: Costanza, R., Segura, O., Martinez-Alier, J. (Eds.), Getting Down To Earth: Practical Applications of Ecological Economics. Island Press, Washington DC, pp. 61–76. Tainter, J., 2006. Social complexity and sustainability. Ecological Complexity 91–103. Tainter, J., 2011. Energy, complexity, and sustainability: A historical perspective. Environmental Innovation and Societal Transitions 1, 89–95. https://doi. org/10.1016/j.eist.2010.12.001 Tainter, J., 2016. Why collapse is so difficult to understand, in: Faulseit, R.K. (Ed.), Beyond Collapse: Archaeological Perspectives on Resilience, Revitalization, and Transformation in Complex Societies. Southern Illinois University Press, Carbondale, pp. 27–39. Tainter, J.A., 1988. The Collapse of Complex Societies, Reprint ed., Cambridge University Press, Cambridge. Takaoglu, T., 2002. Chalcolithic marble working at Kulaksızlar in Western anatolia. Turkiye Bilimler Akademisi Arkeoloji Dergesi 5, 71–93. Talloen, P., Poblome, J., Waelkens, M., Vanhaverbeke, H., 2006. Matar in Pisidia: Phrygian influences in Southwestern Anatolia. Presented at the Religionsgeschichtliche Symposium location:Bonn date:19.-20. Mai 2005, Bonn, pp. 175–190. Teitz, M.B., Chapple, K., 1998. The causes of inner-city poverty: Eight hypotheses in search of reality. Cityscape 3, 33–70. Terrell, J., 1976. Island biogeography and man in Melanesia. Archaeology & Physical Anthropology in Oceania 11, 1–17. Terrell, J., 1977. Human biogeography in the Solomon Islands. Fieldiana. Anthropology 68, 1–47. Thompson, V.D., Turck, J.A., 2009. Adaptive cycles of coastal hunter-gatherers. American Antiquity 74, 255–278. https://doi.org/10.1017/S0002731600048599 Thurner, S., Hanel, R., Klimek, P., 2018. Introduction to the Theory of Complex Systems. Oxford University Press, Oxford. Tilly, C., 1992. Coercion, capital, and European states, AD 990–1992, Rev. ed., Studies in Social Discontinuity. MIT Press, Cambridge. Trigger, B.G., 1998. Sociocultural Evolution: Calculation and Contingency. Wiley. Turchin, P., 2003. Historical Dynamics: Why States Rise and Fall. Princeton University Press, Princeton. Turchin, P., 2013. The puzzle of human ultrasociality. How did large-scale complex societies evolve?, in: Richerson, P.J., Christiansen, M.H. (Eds.), Cultural Evolution: Society, Technology, Language, and Religion. MIT Press, Cambridge, pp. 61–74. Turchin, P., 2016. Ultrasociety: How 10,000 Years of War Made Humans the Greatest Cooperators on Earth. Beresta Books, Chaplin. Turchin, P., Brennan, R., Currie, T., Feeney, K., Francois, P., Hoyer, D., Manning, J., Marciniak, A., Mullins, D., Palmisano, A., Peregrine, P., Turner, E.A.L., Whitehouse, H., 2015. Seshat: The Global History Databank. Cliodynamics 6. https://doi.org/10.21237/C7clio6127917 Turchin, P., Currie, T.E., Whitehouse, H., François, P., Feeney, K., Mullins, D., Hoyer, D., Collins, C., Grohmann, S., Savage, P., Mendel-Gleason, G., Turner, E., Dupeyron, A., Cioni, E., Reddish, J., Levine, J., Jordan, G., Brandl, E., Williams, A., Cesaretti, R., Krueger, M., Ceccarelli, A., Figliulo-Rosswurm, J.,

242  Bibliography Tuan, P.-J., Peregrine, P., Marciniak, A., Preiser-Kapeller, J., Kradin, N., Korotayev, A., Palmisano, A., Baker, D., Bidmead, J., Bol, P., Christian, D., Cook, C., Covey, A., Feinman, G., Júlíusson, Á.D., Kristinsson, A., Miksic, J., Mostern, R., Petrie, C., Rudiak-Gould, P., Haar, B. ter, Wallace, V., Mair, V., Xie, L., Baines, J., Bridges, E., Manning, J., Lockhart, B., Bogaard, A., Spencer, C., 2018a. Quantitative historical analysis uncovers a single dimension of complexity that structures global variation in human social organization. PNAS 115, 144–151. https://doi.org/10.1073/pnas.1708800115 Turchin, P., Whitehouse, H., François, P., Hoyer, D., Nugent, S., Larson, J., Covey, A., Altaweel, M., Peregrine, P., Carballo, D., Feinman, G., Wallace, V., Bol, P.K., Korotayev, A., Kradin, N., Anderson, E., Savage, P.E., Cioni, E., Levine, J., Reddish, J., Brandl, E., Squitieri, A., 2019. Explaining the rise of moralizing religions: A test of competing hypotheses using the Seshat Databank (preprint). SocArXiv. https://doi.org/10.31235/osf.io/2v59j Turchin, P., Whitehouse, H., Korotayev, A., Francois, P., Hoyer, D., Peregrine, P., Feinman, G., Spencer, C., Kradin, N., Currie, T.E., 2018b. Evolutionary pathways to statehood: Old theories and new data. https://doi.org/10.31235/osf.io/ h7tr6 Turner, D., 2007. Making Prehistory: Historical Science and the Scientific Realism Debate, 1st ed. Cambridge University Press, Cambridge; New York. Turner, J., 2003. Human Institutions: A Theory of Societal Evolution. Rowman & Littlefield, Totowa. Umurtak, G., Duru, R., 2016. Hacılar Büyük Höyük Kazıları 2015. Kazı Sonuçları Toplantısı 14, 31–37. Umurtak, G., Duru, R., 2017. Hacılar Büyük Höyük Kazıları 2016. Kazı Sonuçları Toplantısı 15, 27–35. Van De Mieroop, M., 2004. A History of the Ancient Near East: Ca. 3000–323 BC. Wiley. van der Leeuw, S., 1981. Information flows, flow structures and the explanation of change in human institutions, in: Archaeological Approaches to the Study of Complexity. Universiteit van Amsterdam, Amsterdam, pp. 229–329. van Der Leeuw, S., 2004. Why model? Cybernetics and Systems 35, 117–128. https:// doi.org/10.1080/01969720490426803 van der Leeuw, S., 2007. Information processing and its role in the rise of the European World system, in: Costanza, R., Graumlich, L., Steffen, W.L. (Eds.), Sustainability or collapse?: An Integrated History and Future of People on Earth. MIT Press, Cambridge, pp. 213–241. van der Leeuw, S., 2012. Global systems dynamics and policy: Lessons from the distant past. https://doi.org/10.7564/12-coec3 van der Leeuw, S., 2016. Uncertainties, in: Brouwer Burg, M., Peeters, H., Lovis, W.A. (Eds.), Uncertainty and Sensitivity Analysis in Archaeological Computational Modeling, Interdisciplinary Contributions to Archaeology. Springer International Publishing, Cham, pp. 157–169. https://doi.org/10.1007/978-3-319-27833-9_9 van der Leeuw, S., 2019. Social Sustainability, Past and Future: Undoing Unintended Consequences for the Earth’s Survival, Cambridge University Press, Cambridge. van der Leeuw, S., McGlade, J., 1997. Structural change and bifurcation in urban evolution: A non-linear dynamical perspective, in: McGlade, J., van der Leeuw, S. (Eds.), Time, Process and Structured Transformation in Archaeology. Routledge, London, pp. 331–372.

Bibliography 243 van der Leeuw, S., Redman, C.L., 2002. Placing archaeology at the center of socio-natural studies. American Antiquity 67, 597. https://doi.org/10.2307/1593793 Van Heesch, J., Stroobants, F., 2015. The silver coinage of Sagalassos in Pisidia, in: Bland, R., Calomino, D. (Eds.), Studies in Ancient Coinage in Honour of Andrew Burnett. London, pp. 13–30. Van Loo, M., Dusar, B., Verstraeten, G., Renssen, H., Notebaert, B., D’Haen, K., Bakker, J., 2017. Human induced soil erosion and the implications on crop yield in a small mountainous Mediterranean catchment (SW-Turkey). CATENA, Special section on Geoarchaeology: Human-environment interactions in the Holocene 149, 491–504. https://doi.org/10.1016/j.catena.2016.08.023 Van Valen, L., 1973. A new evolutionary law. Evolutionary Theory 1, 1–30. van Zeist, W., Bottema, S., 1991. Late quaternary vegetation of the Near East. Reichert, Wiesbaden. Vandam, R., 2019. A very long time ago. Developments in prehistory, in: Poblome, J., Torun, E., Talloen, P., Waelkens, M. (Eds.), Meanwhile in the Mountains: Sagalassos. Yapı Kredi Yayınları, Istanbul, pp. 3–13. Vandam, R., Kaptijn, E., Poblome, J., Waelkens, M., 2013. The Bronze Age cemetery of Gâvur Evi Tepesi, Southwestern Turkey. Anatolica 39, 241–259. https://doi. org/10.2143/ANA.39.0.2990789 Vandam, R., Mušič, B., Medarić, I., 2019. Contextualizing Kuruçay Höyük: Assessing the unexplored Late Chalcolithic Landscape Near the Beginning of Early Social Complexity in SW Turkey. Journal of Field Archaeology 44, 1–18. https://doi.org/10.1080/00934690.2018.1558904 Vandorpe, K., Waelkens, M., 2007. Protecting Sagalassos’ fortress (Akra) and watchtowers: An early Hellenistic inscription. Ancient Society 37, 121–139. Vanhaverbeke, H., Waelkens, M., Vyncke, K., De Laet, V., Aydal, S., Mušič, B., De Cupere, B., Poblome, J., Braekmans, D., Degryse, P., Marinova, E., Verstraeten, G., Van Neer, W., Šlapšak, B., Medarič, I., Ekinci, H.A., Erbay, M.O., 2010. ‘Pisidian’ culture? The Classical-Hellenistic site at Düzen Tepe near Sagalassus (southwest Turkey). Anatolian Studies 60, 105–128. Vlassopoulos, K., 2007. Unthinking the Greek Polis: Ancient Greek History beyond Eurocentrism, Reissue ed., Cambridge University Press, Cambridge. Voigt, M.M., 2011. Gordion: The changing political and economic roles of a first millennium BCE city, in: Steadman, S.R., McMahon, G. (Eds.), The Oxford Handbook of Ancient Anatolia: (10,000-323 BCE). OUP USA, pp. 1069–1094. Voigt, M.M., 2012. The unfinished project of the Gordion Early Phrygian destruction level, in: Rose, C.B. (Ed.), The Archaeology of Phrygian Gordion, Royal City of Midas: Gordion Special Studies 7, Gordion Special Studies. University of Pennsylvania Press, Philadelphia, pp. 67–100. von Aulock, H., 1977. Münzen und Städte Pisidiens I. E. Wasmuth, Tübingen. Vyncke, K., Waelkens, M., 2015. The classical/hellenistic courtyard building at Düzen Tepe (SW Turkey). Hypotheses on the factors influencing house form and functions, in: Di Castro, A.A., Hope, C.A., Parr, B.E. (Eds.), Housing and Habitat in the Ancient Mediterranean: Cultural and Environmental Responses. pp. 161–169. Waelkens, M., 2004. Ein Blick von der Ferne: Seleukiden und Attaliden in Pisidien. Istanbuler Mitteilungen 54, 435–471. Waelkens, M., Vandeput, L., 2007. Regionalism in Hellenistic and Roman Pisidia, in: Elton, H., Bresson, A. (Eds.), Regionalism in Hellenistic and Roman Asia Minor. Ausonius, Bordeaux, pp. 97–105.

244  Bibliography Walker, B., Gunderson, L., Kinzig, A., Folke, C., Carpenter, S., Schultz, L., 2006. A handful of heuristics and some propositions for understanding resilience in social-ecological systems. Ecology and Society 11, 13. Walker, B.W., Salt, D., Reid, W., 2006. Resilience Thinking: Sustaining Ecosystems and People in a Changing World, Island Press, Washington, DC. Warner, J., 1994. Elmali-Karatas II: The Early Bronze Age Village of Karatas. Bryn Mawr College, Bryn Mawr. Wasserman, S., Faust, K., 1994. Social Network Analysis: Methods and Applications. Cambridge University Press, Cambridge. Waters, M., 2014. Ancient Persia: A Concise History of the Achaemenid Empire, 550-330 BCE. Cambridge University Press, New York, NY, USA. Wattenmaker, P., Misir, A., 1993. Kazane Höyük 1992. Kazı Sonuçları Toplantısı 15, 177–192. Watts, D.J., 2011. Everything Is Obvious: *Once You Know the Answer. Crown. Watts, D.J., Strogatz, S.H., 1998. Collective dynamics of ‘small-world’ networks. Nature 393, 440. https://doi.org/10.1038/30918 Weber, M., 1958. The City. Free Press. Weber, M., 1972. Wirtschaft und Gesellschaft: Grundriss der verstehenden Soziologie, 5., rev. Aufl./Besorgt von Johannes Winckelmann. ed. Mohr Siebeck, Tübingen. Weiberg, E., 2012. What can Resilience Theory do for (Aegean) Archaeology? Matters of scale. Processes and courses of events in archaeology and cultural history, N.M. Burström & F. Fahlander (eds). Weisz, H., 2007. Combining social metabolism and input-output analyses to account for ecologically unequal trade, in: Hornborg, A., McNeill, J., Martinez-Alier, J. (Eds.), Rethinking Environmental History. World-System History and Global Environmental Change. Altamira Press, Lanhan, pp. 289–306. West, G., 2017. Scale: The Universal Laws of Growth, Innovation, Sustainability, and the Pace of Life in Organisms, Cities, Economies, and Companies, 1st ed., Penguin Press, New York. West, G.B., Brown, J.H., Enquist, B.J., 1997. A general model for the origin of allometric scaling laws in biology. Science 276, 122–126. https://doi.org/10.1126/ science.276.5309.122 Westley, F., Carpenter, F., Brock, W., Holling, C.S., Gunderson, L., 2002. Why systems of people and nature are not just social and ecological systems, in: Gunderson, L.H., Holling, C.S. (Eds.), Panarchy: Understanding Transformations in Human and Natural Systems. Island Press, Washington, DC, pp. 103–119. White, A.A., 2013. Subsistence economics, family size, and the emergence of social complexity in hunter–gatherer systems in eastern North America. Journal of Anthropological Archaeology 32, 122–163. https://doi.org/10.1016/j. jaa.2012.12.003 White, D.A., 2016. Archaeology in the Age of Supercomputing, in: Forte, M., Campana, S. (Eds.), Digital Methods and Remote Sensing in Archaeology: Archaeology in the Age of Sensing, Quantitative Methods in the Humanities and Social Sciences. Springer International Publishing, Cham, pp. 323–346. https:// doi.org/10.1007/978-3-319-40658-9_15 White, L.A., 1943. Energy and the Evolution of Culture. American Anthropologist 45, 335–356.

Bibliography 245 White, L.A., 1949. The Science of Culture: A Study of Man and Civilization. Farrar, Straus. Whitehouse, H., François, P., Savage, P.E., Currie, T.E., Feeney, K., Cioni, E., Purcell, R., Ross, R.M., Larson, J., Baines, J., Haar, B. ter, Covey, A., Turchin, P., 2019a. A new era in the study of global history is born but it needs to be nurtured (preprint). SocArXiv. https://doi.org/10.31235/osf.io/t8hgu Whitehouse, H., François, P., Savage, P.E., Currie, T.E., Feeney, K.C., Cioni, E., Purcell, R., Ross, R.M., Larson, J., Baines, J., Haar, B. ter, Covey, A., Turchin, P., 2019b. Complex societies precede moralizing gods throughout world history. Nature 226–229. https://doi.org/10.1038/s41586-019-1043-4 Whitehouse, H., François, P., Turchin, P., 2015. The role of ritual in the evolution of social complexity: Five predictions and a drum roll. Cliodynamics 6. https://doi. org/10.21237/C7clio6229624 Widlok, T., Aufgebauer, A., Bradtmöller, M., Dikau, R., Hoffmann, T., Kretschmer, I., Panagiotopoulos, K., Pastoors, A., Peters, R., Schäbitz, F., Schlummer, M., Solich, M., Wagner, B., Weniger, G.-C., Zimmermann, A., 2012. Towards a theoretical framework for analyzing integrated socio-environmental systems. Temporal and spatial corridors of Homo sapiens sapiens population dynamics during the Late Pleistocene and Early Holocene, Quaternary International 274, 259–272. https://doi.org/10.1016/j.quaint.2012.01.020 Wilensky, U., 1999. NetLogo. http://ccl.northwestern.edu/netlogo/. Wilk, R., Rathje, W., 1982. Household archaeology, American Behavioral Scientist 25, 617–639. https://doi.org/10.1177/000276482025006003 Wilkinson, R.G., 1973. Poverty and Progress: An Ecological Model of Economic Development. Methuen. Wilson, A., Rees, P., Leigh, C., 1977. Models of Cities and Regions. John Wiley, Chichester and New York. Wilson, E.O., 1999. Consilience: The Unity of Knowledge, Reprint ed., Vintage, New York. Wimmer, A., 2006. Models, methodologies, and metaphors on the move, in: Wimmer, A., Kössler, R. (Eds.), Understanding Change. Models, Methodologies and Metaphors. Palgrave Macmillan, New York, pp. 1–33. Wimsatt, W.C., 2007. Re-Engineering Philosophy for Limited Beings: Piecewise Approximations to Reality. Harvard University Press, Cambridge. Winsborough, H., 1965. The social consequences of high population density. Law and Contemporary Problems 30, 120–126. Wirth, L., 1938. Urbanism as a way of life. American Journal of Sociology 44, 1–24. https://doi.org/10.1086/217913 Wobst, H.M., 1974. Boundary conditions for paleolithic social systems: A simulation approach. American Antiquity 39, 147–178. https://doi.org/10.2307/279579 Wobst, H.M., 1976. Locational relationships in Paleolithic society. Journal of Human Evolution 5, 49–58. https://doi.org/10.1016/0047-2484(76)90099-3 Wolman, A., 1965. The metabolism of cities. Scientific American 213, 179–190. Wolpert, D., Grana, J., Tracey, B., Kohler, T., Kolchinsky, A., 2017. Modeling social organizations as communication networks. arXiv:1702.04449 [physics]. Woodbridge, J., Roberts, C.N., Palmisano, A., Bevan, A., Shennan, S., Fyfe, R., Eastwood, W.J., Izdebski, A., Çakırlar, C., Woldring, H., Broothaerts, N., Kaniewski, D., Finné, M., Labuhn, I., 2019. Pollen-inferred regional vegetation

246  Bibliography patterns and demographic change in Southern Anatolia through the Holocene. The Holocene 29, 728–741. https://doi.org/10.1177/0959683619826635 Wright, H., 1978. Toward an explanation of the origin of the state, in: Cohen, R., Service, E. (Eds.), Origins of the State: The Anthropology of Political Evolution. Institute for the Study of Human Issues, Philadelphia, pp. 49–68. Wright, H.T., 1969. The Administration of Rural Production in an Early Mesopotamian Town. University of Michigan. Wright, H.T., Johnson, G.A., 1975. Population, exchange, and early state formation in Southwestern Iran. American Anthropologist 77, 267–289. Wu, J., 2013. Hierarchy theory: An overview, in: Rozzi, R., Pickett, S., Armesto, J., Callicott, J. (Eds.), Linking Ecology and Ethics for a Changing World, Ecology and Ethics. Springer, Dordrecht, pp. 281–301. https://doi.org/10.1007/ 978-94-007-7470-4_24 Wurzer, G., Kowarik, K., Reschreiter, H. (Eds.), 2015. Agent-based modeling and simulation in archaeology, Advances in Geographic Information Science. Springer International Publishing. Yakar, J., 1985. The Later Prehistory of Anatolia: The Late Chalcolithic and Early Bronze Age, BAR International Series. Yener, K.A., Vandiver, P.B., 1993. Tin processing at Göltepe, an Early Bronze Age Site in Anatolia. American Journal of Archaeology 97, 207–238. https://doi. org/10.2307/505657 Yildirim, B., Gates, M.-H., 2007. Archaeology in Turkey, 2004-2005. American Journal of Archaeology 111, 275–356. Yoffee, N., 2005. Myths of the Archaic State: Evolution of the Earliest Cities, States, and Civilizations, Digital print. ed., Cambridge University Press, New York. Yorque, R., Walker, B., Holling, C., Gunderson, L., Folke, C., Carpenter, S., Brock, W., 2002. Toward an integrative synthesis, in: Gunderson, L.H., Holling, C.S. (Eds.), Panarchy: Understanding Transformations in Human and Natural Systems. Island Press, Washington, DC, pp. 419–438. Youngblood, M., Lahti, D., 2018. A bibliometric analysis of the interdisciplinary field of cultural evolution. Palgrave Communications 4, 120. https://doi. org/10.1057/s41599-018-0175-8 Zimmermann, A., 2012. Cultural cycles in Central Europe during the Holocene. Temporal and spatial corridors of Homo sapiens sapiens population dynamics during the Late Pleistocene and Early Holocene. Quaternary International 274, 251–258. https://doi.org/10.1016/j.quaint.2012.05.014 Zimmermann, K.F., 1996. European migration: Push and pull. International Regional Science Review 19, 95–128. https://doi.org/10.1177/016001769601900211 Zipf, G.K., 1949. Human Behavior and the Principle of Least Effort: An Introduction to Human Ecology. Addison-Wesley Press.

Index

Italicized pages refer to figures and bold pages refer to tables. ABM see agent-based modelling adaptive cycles 108, 108–112; framework 117–118; individual 113; multi-scalar concept of 117 Adler, M. A. 88 agent-based modelling (ABM) 57–61, 209–210; for changing dynamical systems 58–59; in educational curriculum 61; for historical analysis and exploration 59 Anatolia, social complexity trajectories in 117–194, 197–202; Chalcolithic 126–132; Early Bronze Age 132–148; Hellenistic period 180–194; Iron Age 163–180; map of case study 125; Middle Bronze Age and Late Bronze Age 148–163 Anderson, Philip 33 Ando, Clifford 182 artefacts 30 Arthur, W. B. 36 Ashby, William 63 Atlas of Cultural Evolution (ACE) 122 autopoiesis 29 Bachhuber, Christoph 125, 134, 198 background theory 8 Barton, C. M. 20 Bausch, Kenneth 87 Berger, Peter 73, 79 Bettencourt, Luis 50 big data analysis, in archaeology 4 Big Gods theory 3 Binford, Lewis 8, 10 Blau, Peter 73 Boas, Franz 18 Bourdieu, Pierre 31, 70, 71

Boyer, P. 138 Bradtmöller, Marcel 208 brain, as predictive device 68–69 Brown, James 48 Buckley, Walter 29 Burgess, Ernest 21, 30 Carneiro, R. L. 26 Cavalli-Sforza, L. L. 55 centralisation, of governmental structures 118 Çevik, Özlem 139, 140 Chalcolithic Anatolia, social complexity trajectories in 126–132, 131; chalcolithic sites 126; Early Chalcolithic 127; excavated settlements 127–130; Hacilar 127; Kulaksizlar 127; Kuruçay Höyük 127–130; Late Chalcolithic 126, 127–132; Middle Chalcolithic 126, 127; settlements 126–127 change and stability, multi-scalar dynamics of 107–116; adaptive cycles 108, 108–112; multi-scalar interactions 113–116; panarchy 113–116 chaotic system 37 Cherry, Colin 32, 63 Childe, Gordon 22, 23, 24 Cioffi-Revilla, C. 82 Clarke, David 30–32 coarse-graining 11, 12, 190, 199 Cofey, G. 52 cognition 77 collective decision making: based on input information 84; canonical loop of 83; selection pressures and 81–86 communication 70

248  Index complex adaptive systems (CAS) 36, 39 complexity formation 81–93; basic model of 92, 204; complexity mechanisms 86–88; meso-level dynamics in social groups 82; potential drivers of 82; push-pull dynamics 88–96; selection pressures and collective decision making 81–86 complexity trajectories: in Anatolia (case study) 117–194, 125; causal factors 82; of Konya Plain 124; positive feedback loops reinforce 84, 85; punctuated dynamics in 122; selection pressures as drivers of 81–86; see also Anatolia, social complexity trajectories in complex social systems 2 complex societies 2, 3; large-scale cooperation, emergence of 55 complex systems thinking 4–6, 33–61; agent-based modelling 57–61; aims of 38; in archaeology 6–13, 40–61; and chaotic system 37; complex adaptive systems 36, 39; cultural evolution 53–57; definition of 5; dynamics in 38–39; information processing 35–36; information transmission 39–40; middle range theories 10; network science 40–47; properties of 34–40; settlement scaling 47–53; structure of 35, 37; system changes and phase transitions 38 composite complexity metric 121 conceptual model, of social complexity 202–210; agent-based modelling 209–210; methodological operationalisation 207–208; modelling complexity 209–210; quantitative measures 205–207 consilience, principle of 8 contingency 75 coupling 70 Coward, Fiona 44 Cowgill, George 21 cross-disciplinary concept migration 9 Crumley, Carole 27 cultural evolution 53–57; approaches in archaeology 54–55; bibliometric analysis of 54; demographic factors, correlation between 55–56; field of 54; goal of 54; issue pertaining to 56–57; large-scale cooperation, emergence of 55; in patterns identification of records 56 Currie, Adrian 6, 7

Darwin, Charles 16, 53 DiMaggio, P. J. 81 Dunbar, Robin 66, 67 Dunbar’s number 66, 67, 129 Duru, Refik 128 Dusinberre, Elspeth 170 diversification 87 Early Bronze Age (EBA) 132 EBA Anatolia, social complexity trajectories in 132–148, 145; Beycesultan 134, 135, 138; Burdur Plain 137; centralisation 140–141; ‘citadel’ culture, emergence of 134; decision-making responses to selection pressures 145–146; energy and resources, flows of 144, 145; evolution of communities 147; extramural cemeteries 136–138; human impact on landscape 147–148; Karatas 134–135; Konya and Karaman Plains 136; Liman Tepe 134, 135–136, 139; metallurgy and textile production, development of 142–143; population nucleation 143–144, 147; problems of chronology related to 132–133; production and distribution patterns in 142; ‘royal’ tombs of Alacahöyük 137; settlement growth 143–144; sites of 133, 133; social and cultural issues 146; TürkmenKarahöyük 136; urban communities 139–140; urbanism 139–141; wealth in burial goods 137–138 Efatmaneshnik, M. 86 energy flow, between society and nature 105 Engels, Friedrich 16 Enquist, Brian 48 epistemic resources 7 epistemic situation 7 Epstein, Joshua 58 Eurocentrism 17 Evans, Tim 45 Faulseit, Ronald 102 feedback loop 64, 104; of information processing and problem solving 85 Feinman, G. M. 2, 26 Feldman, M. W. 55 Fiske, John 17 Flannery, Kent 20, 32 Forge, Anthony 66 fragmentation, of governmental structures 118

Index 249 Freeman, Jacob 12, 122 Fried, Morton 19 Gartner’s hype cycle 12 generative entrenchmen 75 Giddens, Anthony 70–71, 76 Gini coefficients 119 Glatz, C. 154 Gould, S. J. 115 Graham, Shawn 44, 58 Griffin, A. F. 103 Gumerman, G. G. 59 Gunderson, L. 115 habitualisation 73, 79 Hansen, Mogens Herman 24, 182 Harvey, David 23, 76 Hawley, A. 81 Hegel, Georg 18 Hellenistic period Anatolia, social complexity trajectories in 180–194, 194; age of change 181; city foundations 183–184, 184; coinage 186; gift giving 184–185; Greek cultural influence 181–183, 186–187; Greek-style polis 181–183, 182, 191; inter-polity competition 185; ‘Ionian colonisation’ 181; Lycian settlements 191; macro-level polities and mutual interactions 183; material culture 186; Sagalassos 186–189, 188; Seleucid policies of urbanism 190–192; settlement patterns 186; sites in southwest Anatolia 190; urban transformation of Sagalassos 186–189 Herbert Mead, George 30 Historical Particularism 18 Hobson, Elizabeth 63 Holling, C. S. 104, 115 Horejs, Barbara 128 human-environment interactions 103–107; social-ecological systems, framework of 104; social metabolism, concept of 104–106 Ibn Khaldun 24 information: definition 64; flow 64; social practices as bundles of 69–79 information hierarchy 40 information processing, positive feedback loops of 85 information transmission: in complex systems 39–40; mechanisms of 68–69; social mechanisms of 69–79; through social interactions 65–69

institutionalisation 79–81 Ionian colonisation 181 Iron Age Anatolia, social complexity trajectories in 163–180, 179; Achaemenid empire 170–174, 178, 180; Düver Yarimada, settlement at 172–173; Düzen Tepe 174–176; Early Iron Age 163–168, 171; energetic needs of community 174–175; Gordion 163–166; Lake Burdur 172, 173, 177; Late Iron Age 169–172; macro-level polities 166; material culture 166; micro-level of community formation 177; Middle Iron Age 168–169, 171; monumentalisation of Sardis 169; Persian empire 169–170; Phrygian kingdom 163–166; polities in 168–173; population size 174; selection pressures 178; settlement intensity 165; settlements 168–169, 171–174; sites for 164, 164; soil erosion and sedimentation patterns, model of 176–177 Johnson, Gregory 66–67 Kent, Flannery 20 Kinzig, Ann 118 Kleiber, Max 48 Kohler, T. A. 59, 119, 120 large-scale cooperation 55 Late Bronze Age (LBA) 148 LBA Anatolia, social complexity trajectories in 151–163, 160; Arzawa 152–153; Beycesultan 156–157; Cypriot traders 159; disintegration of Hittite power 158; humanenvironment interactions 159; Konya and Karaman Plains 154–155; material traces 155–156; polities in Anatolia 152–159, 153; sites of 149; socio-political landscape, transformation of 157; TürkmenKarahöyük site 155 Lorenz, Edward 37 Lotka, Alfred 106 Lubbock, John 18 Lucas, Gavin 74 Luckmann, Thomas 73, 79 Luhmann, Niklas 29 Lyman, R. L. 56 Maisels, C. K. 26 Marx, Karl 16

250  Index material culture 30, 31, 32 Maturana, H. R. 29 MBA see Middle Bronze Age MBA Anatolia, social complexity trajectories in 148–151, 159–163, 160; Assyrian trade network 148–149, 150; commodities 150; inter-regional trade networks 149–150; Konya Plain, sites in 51; soil-retaining fortification walls 151 McGlade, James 114 Mead, George Herbert 30 Mellaart, James 129, 146 Merton, Robert 10 metaphors 9 Middle Bronze Age (MBA) 148 middle range theories (MRT) 10 Mingers, J. 28 Miranda, Lux 122 moralising gods hypothesis 3 Morgan, Lewis Henry 16, 17 MRT see middle range theories Netto, Vinicius 76 network science 40–47; preferential attachment 42–43, 43; proximal point analysis 43–44; scale-free networks 42, 44–45; small-world network 41–42, 42, 44–45; social network analysis 44, 46, 47; spatial interaction models 45–46 O’Brien, M. J. 56 Ortman, Scott 51, 52 Owens, E. J. 22 panarchy 113–116 Park, Robert 30 Parsons, Talcott 28–29 peer-polity interaction (PPI) 91–92 place 76 Pleistocene 6 Plog, F. T. 32 Plourde, A. M. 154 polythetic attribute system 31 Powell, W. W. 81 preferential attachment structure 42–43, 43 Price, Derek 26, 43 problem solving, positive feedback loops of 85 push-pull dynamics 88–96; central place functions 91; cognitive limits 90; framework of 91; group fission

90–91; local and (micro-)regional scale 90; on multiple scales 89, 89; peer-polity interaction 91–92 Rapoport, Amos 79 Rathje, W. 131 Redman, Charles 118 Red Queen effect 16 rigidity 111 rigidity trap 111 Rihll, T. E. 45 Rivers, Ray 45 Ryan, M. J. 86 Salmon, M. H. 27, 32 Sanderson, S. K. 26 Santa Fe Institute (SFI) 34 scale-free networks 42; small-world networks combined with 44–45 Schutz, Alfred 73 selection pressures: and collective decision making 81–86; for social organisation 82–84 self-organised criticality (SOC) 103 Service, Elman 19 SES see social-ecological systems Seshat: Global History Databank project 3–4, 120–123, 205 settlement chambers 100 settlements, nucleation of people in 89 settlement scaling 47–53; developed from biological organisms 48–49; foundational assumptions of 49–50; framework of, issues related to 52–53; mathematical model of urban scaling 50–51; rank-size distributions 47–48; seminal applications of 51; theoretical scaling graphs 49 SFI see Santa Fe Institute Shannon, Claude 63, 64 Shin, J. 122 Simon, Herbert 42 simple societies 2 Sjoberg, Gideon 21 Small, Albion 30 small-world networks 41–42, 42; scale-free network combined with 44–45; social network analysis combined with 44 Smith, Adam 17 Smith, Michael 89 SNA see social network analysis SOC see self-organised criticality

Index 251 social complexity 1–4; building blocks of 65–81; communities as social reactors 93–97; costs of 101–103; definition 2; development of, equation to measure 86; emergent social structures 79–81; and energy flow 103–107, 105; formation of 81–93; group size and 68; and humanenvironment interactions 103–107; information transmission 65–69; mechanisms 86–88; micro-level drivers of development 103; outcomes of 93–107; polity formation 97–101; social interaction 65–69; social practices as bundles of information 69–79; see also complexity formation social complexity trajectories 118; global perspective 118–119; see also complexity trajectories Social Darwinism 16 social-ecological systems (SES) 104, 118 social entropy 72; reduction, effects of space on 77 social evolution 16–20 social interaction 65–69 social knowing 70 social network analysis (SNA) 44, 46, 47 social practices: as bundles of information 69–79; components of 70; conceptualisation of 71; material dimension 77–79; micro-macro transitions, mechanisms of 72–79; recurrent iteration 71–72; social dimension 73; social entropy and 72; spatial dimension 76–77; temporal dimension 74–76 social subsystem 31–32 social systems: in archaeology 30–33; of communication 29; definition 28–29; essential attributes in 30–31; material attributes 30–31; Parsons’ model of 28–29; transfer of information 31 social systems thinking 28–30 societal collapse 102–103 spatial interaction models 45–46

Spencer, Herbert 16, 17, 53 Spencer-Brown, G. 29 state formation 23–27 Steadman, S. 129, 131 Steward, J. H. 18, 19 systems thinking: and complexity 27–33; definition 27–28; social systems in archaeology 30–33; social systems thinking 28–30; as structuring mechanism 28 Tainter, Joseph 88 Tilly, C. 26 traces 7–8 Trigger, Bruce 19 Turchin, P. 122 Turner, Derek 8 Turner, Jonathan 26, 82 Tylor, Edward 16, 18 uncertainty reduction devices 75 urbanisation 21–23 urban metabolism 50 urban scaling, mathematical model of 50–51 van der Leeuw, Sander 114 Varela, F. J. 29 Watts, Duncan 5 wealth inequality 119 Weber, Max 22, 23, 26 West, Geoffrey 48, 52 White, Leslie 18, 19, 28 Wilk, R. 131 Wilson, A. G. 45 Wimmer, Andreas 9 Wirth, Louis 21, 22 Wittgenstein, Ludwig 70 Wobst, H. M. 59, 67 Wright, Henry 20 Yoffee, Norman 10 Zipf, George Kingsley 47